BIP, AN INTESTINAL METABOLIC HORMONE
A Thesis submitted to the University of Surrey
for a Degree of
DOCTOR OF PHILOSOPHY
by
PIOTR KWASOWSKI
August 1986
Division of Clinical Biochemistry
Department of Biochemistry
University of Surrey
BuiIdford
BU2 5XH
UK
I W ' ProQuest Number: 27600360
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest
ProQuest 27600360
Published by ProQuest LLO (2019). Copyright of the Dissertation is held by the Author.
All rights reserved. This work is protected against unauthorized copying under Title 17, United States C ode Microform Edition © ProQuest LLO.
ProQuest LLO. 789 East Eisenhower Parkway P.Q. Box 1346 Ann Arbor, Ml 48106- 1346 SUMMARY
Using an established radioimmunoassay for Immunoreactlve
gastric inhibitory polypeptide (IR-GIP) the metabolic actions
of gastric inhibitory polypeptide (GIP) were investigated. It
was found that there was no circadian rhythm in basal plasma
TR-GIP and immunoreactlve insulin (IRI) concentrations in
normal human volunteers, but that there was in stimulated
IR-GIP and IRI plasma concentrations following a mixed meal.
It was shown that short-term adaptation of both humans and
rats, with high-fat diets, could cause moderate insulin
resistance as demonstrated by the removal of the feed-back
inhibition of exogenous insulin on fat-stimulated IR-GIP
release. It was shown that GIP via its role in the
enteroinsular axis, might play a part in the development of
the hyperinsulinaemia and obesity of the genetically obese
hyperglycaemic (ob/ob) mouse.
It was demonstrated that the pre-requisite for fat-
stimulated IR-GIP release in the ob/ob mouse is the absorption of fatty acids and glycerides capable of being esterified into triglyceride. Any physical, chemical or metabolic interference with this process resulted in a decrease of the amount of IR-GIP released. It was further shown, using a combination of both exogenous and endogenous lipid and GIP, that GIP has a role in the removal of chylomicron triglyceride from the circulation of the rat. It was not possible to confirm this in humans using exogenous lipid and endogenous glucose-stimulated IR-GIP release. It was shown in the insulinoma bearing NEDH rat that exogenous porcine GIP was able to augment the insulin stimulating ability of glucose when co-administered intravenously with glucose, thus providing strong evidence for GIP's role as an incretin.
II A.M.D.G,
III ACKNOWLEDGEMENTS
I should like to express my thanks to the following
people for their help over the last few years.
Firstly I must thank Profesor Vincent Marks for his
constant help and encouragement in the execution and (0 completion of this thesis. I must also express my djepest
gratitude to Dr Linda Morgan, who in the early days taught me
all I knew about GIP, who advised and encouraged me through
the initial experimental work and who, more recently, proof
read the first drafts of this thesis. I must also thank Dr
Shelagh Hampton for her immeasurable help and advice on
radioimmunoassays.
I would also like to thank Drs Peter Flatt and Clifford
Bailey for their help with the ob/ob mice. Dr Kim Tan for his
help with the insulinoma bearing rats, and Drs John Wright,
Bob Cramb, Mike Dunne and John Wong for their help with the
humans.
Finally, I would like to thank my parents for
encouraging me and making it possible for me to start, and my wife Christine, without whose proof reading and constant and untiring encouragement and help I would certainly not have finished.
IV CONTENTS
PAGE
SUMMARY I
DEDICATION III
ACKNOWLEDGEMENTS IV
CHAPTER ONE
GENERAL INTRODUCTION
1.1.0 General Background 2
1.1.1 Historical 2
1.1.2 Special Aspects 3
1.1.3 General Concepts 5
1.1.4 Recent Developments 8
1.2.0 Enterogastrone Concept 15
1.2.1 Endogenous Lines of Research 15
1.2.2 Exogenous Lines of Research 16
1.3.0 Incretin Concept 19
1.3.1 Endogenous Aspects 19
1.3.2 Incretin Candidates 21
1.4.0 Gastric Inhibitory Polypeptide (GIP) 24
1.4.1 Discovery of GIP 24
1.4.2 Isolation and Purification 26
1.4.4 Localisation 27
1.4.5 Physiological Actions of Exogenous GIP 28
1.4.6 Metabolic Effects of GIP 31
V 1.4.7 stimulation of Immunoreactive BIP (IR-GIP)
release 36
1.4.8 Inhibition of IR-GIP Release 42
1.5.0 Motilin 46
1.5.1 Historical Background and Discovery of
Motilin 46
1.5.2 Structure of Motilin 46
1.5.3 Cellular Localisation of Motilin 47
1.5.4 Physiological Actions of Motilin 47
1.5.5 Modulation of Circulating IRM 49
1.6.0 Summary of Introduction and Aims 52
CHAPTER TWO 53
METHODS
2.1.0 The Gastric Inhibitory Polypeptide (GIP)
Assay 54
lodination 54
Standards 56
Charcoal Stripped Serum 56
Assay 57
Antibody 57
2.2.0 The Motilin Assay 64
lodination 64
Antibody 65
Assay 65
2.3.0 Plasma Triglyceride Assay 73
VI 2.4.0 Plasma Glucose Assay 74
CHAPTER THREE 75
TWENTY-FOUR HOUR RHYTHMS IN THE ENTEROINSULAR AXIS
3.0.0 Introduction 76
3.1.0 GEP Hormones During a Twenty-four Hour Fast 77
3.1.1 Methods 77
3.1.2 Results 78
3.1.3 Discussion 79
3.2.0 Function of the Enteroinsular Axis at 1100 h
and 0200 h 87
3.2.1 Methods 87
3.2.2 Results 88
3.2.3 Discussion 88
3.3.0 Investigation of the Insulinaemic Responses
Following Intravenous Glucose at 1000 h and
0200 h 95
3.3.1 Methods 95
3.3.2 Results 96
3.3.3 Discussion 96
3.4.0 Function of the Enteroinsular Axis
Throughout the Day 107
3.4.1 Methods 107
3.4.2 Results 108
3.4.3 Discussion 108
VII Summary of Results - Chapter Three 117
CHAPTER FOUR 119
EFFECT OF DIETARY FAT ON THE ENTEROINSULAR AXIS
4.0.0 Introduction 120
4.1.0 Effect of a Short-term High Fat Diet on the
Enteroinsular Axis in Rats 123
4.1.1 Methods 123
4.1.2 Results 125
4.1.3 Discussion 131
4.2.0 Effect of Both High and Low Fat Diets on the
Enteroinsular Axis in Man 134
4.2.1 Methods 134
4.2.2 Results 137
4.2.3 Discussion 147
Summary of Results - Chapter Four 151
CHAPTER FIVE 152
EFFECT OF GIP ON THE RAT ADIPOCYTE
5.0.0 Introduction 153
5.1.0 Role of i.p. GIP in the Clearance of i.p.
Intralipid from the Plasma of the Rat 157
5.1.1 Methods 157
5.1.2 Results 158
5.1.3 Discussion 163
VIII 5.2.0 Role of Exogenous GIP in the Clearance of
i.v. Intralipid in the Rat 165
5.2.1 Methods 166
5.2.2 Results 167
5.2.3 Discussion 171
5.3.0 Effect of Anti-GIP Antibodies on the
Clearance of Plasma Triglyceride Following
Oral Triolein in Rats 174
5.3.1 Methods 174
5.3.2 Results 175
5.3.3 Discussion 180
Summary of Results - Chapter Five 182
CHAPTER SIX
THE ROLE OF GIP IN THE DIFFERENTIAL INSULIN RESPONSES
TO ORAL AND INTRAVENOUS GLUCOSE IN THE TRANSPLANTABLE
RAT INSULINOMA 183
6.1.0 Introduction 184
6.1.1 Methods 185
6.1.2 Results 187
6.1.3 Discussion 194
Summary of Results - Chapter Six 196
CHAPTER SEVEN 197
THE ROLE OF ENDOGENOUS CARBOHYDRATE-STIMULATED IR-GIP
IN THE REMOVAL OF EXOGENOUS TRIGLYCERIDE IN MAN
7.1.0 Introduction 198
IX 7.1.1 Methods 200
7.1.2 Results 201
7.1.3 Discussion 210
Summary of Ressults — Chapter Seven 212
CHAPTER EIGHT 213
IMMUNOREACTIVE GASTRIC INHIBITORY POLYPEPTIDE (IR-GIP)
IN OBESE (ob/ob) MICE
8.0.0 Introduction 214
8.1.0 Immunoreactlve Gastric Inhibitory Polypeptide
(IR-GIP) in the Plasma and Small Intestine of
the Obese Hyperglycaemic Mouse 216
8.1.1 Methods 216
8.1.2 Results 217
8.1.3 Discussion 220
8.2.0 Involvement of GIP and the Enteroinsular Axis
in the Metabolic Abnormlities of the Obese
Hyperglycaemic (ob/ob) Mouse 221
8.2.1 Methods 222
8.2.2 Results 225
8.2.3 Discussion 236
8.3.0 Effect of Chain-length and S^uration in
Fatty Acid Stimulated IR-GIP Release in
the Obese Hyperglycaemic (ob/bo) Mouse 239
8.3.1 Methods 239
8.3.2 Results 241
8.3.3 Discussion 246 Summary of Results - Chapter Eight 249
CHAPTER NINE 250
FINAL DISCUSSION AND FUTURE WORK
9.1.1 The Enteroinsular Axis 251
9.1.2 The Release of IR-GIP 256
9.1.3 The Enteroadipose Axis 257
9.1.4 Other Future Lines of Research 258
REFERENCES 263
PUBLICATIONS 280
XI CHAPTER ONE
GENERAL INTRODUCTION
paqe 1 1.1.0
General Background
1.1.1 Historical
The physiology, origin, chemistry, pharmacology and
clinical aspects of gastrointestinal hormones, or gut
hormones as they have been more succinctly labelled, has
become one of the most rapidly expanding and interesting
fields in contemporary endocrinology and gastroendocrinology.
One of the more eloquent pictures of the ever growing field
of gut hormones was painted by G.B.J. Glass.
"Tbis territory is reminiscent of a subtropical forest
full of proud and established trees, jungle-like bushes
inter-woven with each other, beautiful flowers not yet
classified in the Linnaen code, and also frail and weak
growths struggling for survival. In this new territory
the steps of the explorer are both risky and highly
rewarding, but certainly full of excitement." (Glass,
1980).
The study of gut hormones as we now know them, dates back to the very beginning of this century. Bayliss and
Starling published a paper entitled "On the causation of the so called 'peripheral reflex secretion' of the pancreas."
(Bayliss and Starling, 1902).
The "peripheral reflex" referred to in their paper was that propounded by Popielski (1901) who claimed that the secretions of the pancreas were under nervous control.
However, with Bayliss and Starling's discovery of secretin the age of "endocrinology" had begun and any serious thoughts
page 2 of the involvement of the central nervous system secretions of the pancreas were set aside for decades. In 1905 Edkins described the chemical mechanism of gastric secretion with the discovery of gastrin (Edkins, 1905). More than twenty years later it was discovered that yet another hormone contributed to the normal function of the gut. It was shown that when fat was introduced into the small intestine a hormone was released which caused gall bladder contractions. This hormone was named cholecystokinin (CCK) (Ivy and Oldberg, 1928), and thus the entire workings of the gut were explained by the actions of these three hormones; gastrin controlled gastric acid secretion, secretin that of pancreatic bicarbonate, and cholocystokin in-pancreozymin (as it became known) gall bladder contraction and pancreatic enzyme output. Even as late as 1970 it was argued that the hormonal regulation of the digestive tract and pancreas could be totally explained by these same three hormones (Grossman, 1970). Although the fuse was lit in 1902 the explosion in interest and "understanding" in the working of the gut did not occur until 60 - 70 years later. Although the hormonal actions of secretin have been known for over 80 years it was not until as recently as 1961 that the purification of porcine secretin was accomplished (Jorpes and Mutt, 1961). It then took a further 5 years before it was sequenced (Mutt and Jorpes, 1966). The situation was analagous for both gastrin and CCK. 1.1.2 Special Aspects When compared to other fields in endocrinology, progress page 3 in the study of hormones of the gastrointestinal tract has been hampered by some very real problems. Unlike hormones which are synthesized and secreted by small discrete glands such as the pituitary hormones, thyroid hormones or adrenal hormones (to name several). Populations of gut hormone secreting cells are distributed throughout the gastrointestinal tract according to their own individual freqencies (Bloom and Polak, 1978). This situation is complicated further by the fact that in any one section of the gut several populations of different gut hormone secreting cells co-exist. Thus the anatomical dissemination and intermingling of different endocrine cells within the gut and their interactions have made defining the role of a single gut hormone in (uncharacteristic) isolation a most difficult task. Returning to the development of the study of the glandular endocrine system, it was found that surgical removal of a gland gave rise to the symptoms of a deficiency syndrome, and the administration of. a glandular extract to those of over production. These cells are formed into a gland because they are required to respond to a blood borne stimulus. Reacting in this way to concentration changes of the stimulus they are only required to sample blood from one artery or to the stimulus from a single discrete nerve. This is in contrast with the situation in the gut where the stimuli to the diffuse endocrine system, as it has become known, are the irregular components of our diet; fat, carbohydrates and proteins, whether present as liquid or solid. The task of the diffuse endocrine system is to produce an integrated response that reflects the total relative proportions of the nutrient classes ingested, modulating the page 4 necessary physiological and metabolic processes as and when required. In this respect the diffuse endocrine system has a different role to the glandular endocrine system; it responds not only to just one stimulus but many, be they mechanical, osmotic or chemical. As the digestive tract has mechanically disruptive, secretory, digestive and absorbtive functions which are spatially distinct, the individual gut hormone responses reflect this spatial organisation to facilitate absorption and assimilation in a manner to the overall advantage of the organism. It is in this way, in some cultures, that gut hormones control the consequences of one of man's more time consuming occupations and via satiety its frequency and duration, as detailed in a short review (Koopmans, 1981). Finally, there have been very few simple and obvious connections between dramatic diseases and over- or under-production of gut hormones. Most disturbances in gastrointestinal function are not usually life threatening or sensational, and lack the "medical charisma" of diseases such as juvenile onset diabetes melitus or thyrotoxicosis. In all but their severest forms, diarrhoea, constipation and other gastrointestinal disturbances have been endured as part of man's lot in life. As a result of this lack of interest and clinical motivation, efforts in this area of research have quite naturally been limited. 1.1.3 General Concepts The term "gut hormone" has already been used several times in this introduction. However, as a definition of the substances in question it can be misleading as not all of page 5 these are hormones in the classical sense, nor are they all confined to the gut. Hence, the term regulatory peptides or regulatory peptides of the gut has found favour as a new title for these substances (Grossman, 1981), Various stimuli (physical or chemical) from within or without the body can cause the release of (in this context) regulatory peptides and are delivered to their target cells in one, or possibly a combination, of three modes. These are endocrine, classically the manner in which a regulator is borne by the blood to a distant target, neurocrine, in which the regulator diffuses across a narrow synaptic gap from neurone to target cell, and paracrine, in which the regulator diffuses through the intercellular space from the releasing cell to a nearby target cell without entering the blood stream. In table 1.1 there is a list of peptides which have been found in the digestive tract. The digestive tract, in this context however, is usually considered to include the pancreas in addition to the stomach and intestine, so to the list of names giving a blanket description to the peptides dealt with in this introduction (i.e. gut hormones / regulatory peptides of the gut) we may add gastroenteropancreatic hormones (Walsh, 1980). Most of the hormones in table 1.1 have been isolated and sequenced from tissues of the gut. Enkephalin and TRH were isolated and sequenced from brain tissue and subsequently found to be present in the gut by immunocytochemistry. Of the peptides present in both brain and gut not all were found to be identical in structure at these sites. Neurotensin (Kitagbi, Carraway and Leeman, 1976) and Substance P (Carraway and Leeman, 1979) were identical but somatostatin from the gut was found to be larger than that in the brain (Pradayrol et page 6 al., 1980). Several forms of CCK were found in the gut but only the smallest of these, CCK8 (Dockray, 1976), was found in any quantity in the brain. In recent years it has become evident that most, perhaps all, regulatory peptides exist in multiple molecular forms. There are two forms of heterogeneity. Micro-heterogeneity is where the overall length of the molecule is unchanged but individual amino acids in the sequence are altered. This is most readily seen in differing amino acid sequences of the same hormone in different species. The evolution of the gut hormones has been reviewed by Marks and Morgan (1982). Macro—heterogeneity, however, is where the hormone exists in several forms (within the same animal) consisting of fragments of various lengths of a precursor molecule. All these fragments include the biologically active part of the molecule. Macro-heterogeneity is seen following the synthesis of the preprohormone, which is first cleaved to the prohormone and then again to yield the circulating hormone. The cleaved fragments may have biological activities of their own, as has been shown for the connecting peptide of insulin (C-peptide) (Dryburgh, Hampton and Marks, 1980). This "circulating hormone" may be cleaved further yielding several molecular sizes of hormone. This has been found to be the case for gastrin, CCK, somatostatin and others (Marks and Morgan, 1982). A single precursor molecule may give rise to more than one regulatory peptide, a prime example being the synthesis of ACTH and endorphin. These originate from a common precursor found in the corticotrophin cells of the anterior pituitary (Mains, Eipper and Ling, 1977). Table 1.1 shows that all of the hormones detected in endocrine cells of the gut can be measured in the peripheral page 7 circulation by. radioimmunoassay (RIA). Whether this is also true for the peptides present solely in neurones is unclear. There are several criteria to be met if a peptide is to be considered a hormone (Grossman, 1981). It should be present in endocrine cells. It should be released by feeding or other stimuli, thereby increasing plasma concentrations and producing characteristic biological actions. Finally, what -Grossman (1977) called copying, where an infusion of exogenous peptide, mimicking the plasma concentrations observed with endogenous release, cause the normal biological action. The mechanism of paracrine and neurocrine "hormonal" action is as yet only a hypothetical concept, albeit generally accepted, and not an established mechanism. There does not yet exist a method by which inter-cellular or synaptic concentrations of peptide can be assesed, as can blood concentrations for endocrine hormones. 1.1.4 Recent Developments As has already been stated, following the discovery of gastrin and secretin, except for the discovery of CCK, very little happened in the field of gut hormones until relatively recently. Figure 1.1 shows these substances arranged by year of discovery and cumulative number up until September 1980 (Rehfeld, 1981). Most of the progress after the discovery of CCK until recently was in the description of new hormonal mechanisms based only on functional evidence such as the incretin and enterogastrone effects. Only about eight of the structurally characterized hormones shown in figure 1.1 are thought to act as circulating hormones; figure 1.2 shows their general page 8 locations. Table 1.2 gives the probable distribution and main pharmacological activities of the major gut hormones. There were several reasons for the explosion in the gut hormone field in the sixties and seventies. Most of these were technical, but the contributions of individuals at this time were also very important. It was the impetus provided by Roderick Gregory in Liverpool and Viktor Mutt in Stockholm in the purification and structural elucidation of gastrin, CCK and secretin in the early sixties which triggered the great upsurge in interest in the area of gut hormones. Profesor Mutt's laboratory continues to play a major role in the purification and sequencing of a large number of regulatory peptides. In addition to secretin (Jorpes and Mutt, 1961) and CCK (33 and 39) (Jorpes, Mutt and Toczko, 1964), gastric inhibitory polypeptide (GIP) (Brown, Mutt and Pederson, 1970), motilin (Brown, Mutt and Dryburgh, 1971), gastrin releasing peptide (GRP) (McDonald et al., 1978), PHI and PYY (Tatemoto and Mutt, 1980) have been sequenced. page 9 Tab le 1.1 Gastroenteropancreatic regulatory peptides and their Lbution in neurones, endocrine cel Is and blood pi Marks and Morgan, 1982). PEPTIDE BRAIN GASTROINTESTINAL TRACT Neural elements Endocrine cells Secretin - - + <31ucagon - - + €ilP - - + M otilin - - + Neurotensin + - + ^Gastrin + - + PP + ?+ + CCK + ?+ + Angiotensin + ?+ + ACTH + ?+ + Somatostatin + + : + Enkephalin + + + Bombesin + ?+ - Substance P + + - VIP + + - + = present - = absent page 10 Table 1.2 Nomenclature and classification of the major gut hormones (from Marks and Morgan, 1982). Substance Probable distribution Main pharmacological activity {il) Choiecystokinin Duodenum and jejunum Stimulation of gall-bladder contraction; (CCK) stimulation of pancreatic enzyme release (iii) Secretin Duodenum and jejunum Stimulation of pancreatic juice secretion | . j "(iv) Vasoactive Most of gastrointestinal Vasodilation i intestinal tract polypeptide (VIP) (v) -Gastric Duodenum and jejunum Stimulation of insulin secretion. inhibitory Inhibition of gastric acid secretion polypeptide (GIP) (vl) Motilin Duodenum Stimulation of gastric motor activity (v iii) Chymodenin Duodenum Stimulation of chymotrypsin secretion (ix) Somatostatin Antrum, duodenum Inhibition of many gastrointestinal and like and jejunum pancreatic hormones immunoreactivity (x) Met/leu Most of gastrointestinal Inhibition of GI motility and secretion enkephalins tracts (xi) Urogastrone Duodenum Inhibition of gastric acid secretion (xii) Pancreatic Most of gastrointestinal Inhibition of exocrine pancreatic polypeptide tract secretion (xiii) Neurotensin Lower part of small Hypertensive and hyperglycaemic intestine actions (xiv) PHI Duodenum and jejunum Stimulation of insulin, amylase and bicarbonate secretion. page 11 Figure 1.1 Gut hormones arranged by year of discovery and cumulative number up to the end of 1980 (from Rehfeld, 1981). • : a structurally characterized hormone Number 0 : a hormonal mechanism (only functional evidence) □ : a hormonal peptide also found in nerves 30 • PH I @ AC IH a T R H a Enkephohn a Neurolen*. ir. BH G lucagon a Bombe&in a PP a Somoloslolin 20 a Oubslonce P • Cbymodenin O vogogostronç • M t t i l i n O Enfero-0»yn|in @ V IP • G IP O Buiboga&trone O Antrol Cholonp • t nlerofjlucogon 10 • U'ogastronp O Gosifone O f nierocfinir O Q V.llik.n.r. O f nlcfoqoslfone P I n c f e l m ]# | CholeCySlokimr a G a s t r in S 1900 1910 1920 1930 1960 1950 1960 1970 1980 Year page 1: Figure 1»2 tides with a hormonal role Distribution of gut regulatory pep in man (from Bloom and Polak, 19B1). gastrin secretin GIF CCK motilin neurotensin enteroglucagon page 13 Figure 1.3 Distribution of IR-6IP in the human gut (from Bloom and Polak, 1981). G IF pmol/g <02 40 Î 6 < 0 2 11-30 « > 3 1 Number o( cells per mm' page 14 1.2 .0 , Enterogastrone Concept Ttie term enterogastrone was coined by Kosaka and Lim (1930) to describe a hypothetical hormone which inhibited gastric acid secretion. Its release from the upper small intestinal mucosa, was caused by the presence of fat or fatty acids within the upper small intestinal lumen. To study the mechanism .or mechanisms of humorally mediated gastric inhibitory processes two lines of research have been followed. The first was by using physiological stimuli such as fat, acid and hypertonic solutions in the small intestine, and observing their actions and relative potencies in both innervated and denervated preparations. The second mode was by attempting to isolate and purify the substances responsibe for the observed inhibitory mechanisms from the gut mucosa. 1.2.1. Endogenous Lines of Research It had been shown in the previous century that the addition of olive oil to a test meal of starch paste produced a delay in gastric emptying when given to human subjects (Ewald and Boas, 1886). It was subsequently found, that the addition of fat to a meal fed to dogs inhibited secretion of acid and pepsin, and that the reflex secretion of acid and pepsin produced by sham feeding was inhibited by prior feeding with olive oil (Pavlov, 1910). The same treatment inhibited motor activity in the stomach and brougj^ about a delay in gastric emptying, and it was the presence of fat in page 15 the duodenum rather than in the stomach itself which produced this inhibitory action. It was demonstrated in transplanted gastric pouches that fat could inhibit gastric motor activity, and that this was caused by a humoral agent (Farrell and Ivy, 1926). It was later shown that this inhibitory effect was not produced by absorbed fat and its digestion products, nor was it the result of the passage of bile into the intestine following fat digestion (Feng, Hou and Lim, 1929). Fat has to be in an absorbable form to cause the inhibition of gastric acid secretion; oil has to be pre-incubated with pancreatic juice (Sircus, 1958). Exclusion of bile and pancreatic juice produced a lesser degree of inhibition (Menguy, 1960). Duration of inhibition and rate of absorption was correlated positively, using oleic acid and triolein (Long and Brooks, 1965), and, using micellar fat mixtures, the inhibition could be produced from all levels of the small intestine. However, the degree of inhibition was the greatest from jejeunal loops, where the rate of absorption was fastest. Fat and fatty acids in the small intestine were not the only stimuli to the inhibition of acid secretion and motor activity studied. Hypertonic sugar solutions introduced into the duodena of conscious dogs inhibited the motor activity of transplanted fundic pouches, an effect shown not to be due to glucose absorption (Quigley and Phelps, 1934). 1.2.2. Exogenous Lines of Research Injections of saline extracts of the duodenal mucosa were shown to inhibit the acid respnses of Heidenhain pouches page 16 in animals following both meals and histamine. Purification of the extracts was attempted and they were reportedly free of secretin and vasodilatory agents (Kosaka and Lim, 1930). In later work extracts were shown to be active against meal- and histamine-stimulated gastric acid secretion and to inhibit motor activity from innervated gastric fistulae in dogs (Kosaka et al., 1932). An enterogastrone extract prepared from porcine duodenal mucosae demonstrated the ability to inhibit both hunger and distension induced contractions in the empty fasted stomachs of dogs (Gray, Bradly and Ivy, 1937). However, further purification attempts on the enterogastrone extract were unproductive (Greengard et al., 1946). An essentially pure preparation of secretin (Jorpes and Mutt, 1961) demonWrated an inhibitory effect on acid secretion stimulated by exogenous gastrin (Wormsley and Grossman, 1964). These studies were confirmed when it was found that a preparation of Jorpes et al. (1964) containing cholecystokin in-pancreozymin (CCK-PZ), a less pure preparation than the secretin, also inhibited acid secretion from an Heidenhain pouch stimulated by exogenous gastrin (Gillespie and Grossman, 1964). This preparation also inhibited low dose histamine-stimulated acid secretion. It was eventually shown that the inhibitory properties of synthetic secretin were the same as those for the natural material (Vagne et al., 1968). Secretin, not an impurity, was confirmed to be an acid secretion inhibitor from Heidenhain pouches of dogs when gastrin was the stimulus, but in Pavlov pouch dogs, in which acid secretion was stimulated by insulin hypoglycaemia, secretin was ineffective (Way, 1970). CCK-PZ has been suggested to behave like gastrin, stimulating acid page 17 secretion in low doses, but at high doses acting as an inhibitor (Magee and Nakamura, 1966). page 18 1.3.0. Incretin Concept 1.3.1. Endogenous Aspects The term "entero-lnsular axis" introduced by Unger and Eisentraut (1969), described a proposed regulatory mechanism which comprises all the stimuli from the small intestine influencing t+ie secretion of the pancreatic islet cell, including hormonal, neuronal and direct substrate stimulation of insulin, glucagon, somatostatin and pancreatic polypeptide. The existence of an inter— relationship between gut hormones and the pancreas was proposed, with a function of augmenting or accelerating islet cell hormone responses to ingested food. It was suggested that the islet cells release appropriate amounts of insulin and/or glucagon in response to messengers from the gut mucosa, in addition to the absorbed substances themselves. It had been shown many years earlier that the oral administration of porcine duodeno-jejunal extracts, containing secretin, improved the glycosuria of diabetics (Moore, Edie and Abram, 1906). This study stimulated further interest in the hypoglycaemic properties of duodenal extracts (Dixon and Wadia, 1926; Laughton and Macallum, 1932). The factor with hypoglycaemic activity within the duodenal mucosa has had a number of names, such as "hypoglycaemic secretin", "incretin", "duoden in" or "insulinotropic hormone". The hypog1ycaemic factor was shown not to be secretin (Zunz and Labarre, 1929; Labarre and Still, 1930), and Labarre (1932) coined the term "incretin" page 19 to identify thier factors which stimulated the internal secretion of the pancreas. The factor within their crude secretin preparation which stimulated the exocrine pancreas, however, he named "excretin". Loew, Gray and Ivy (1940) failed in their experiments with fasted, rather than hyperglycaemic, dogs to find a hypoglycaemic factor in the duodenum. They also showed that intraduodenal hydrochloric acid was ineffective in lowering blood sugar levels in both fasted and hyperglycaemic animals. Their conclusion "denied tbe entire evidence which has been advanced in support of the theory that the duodenum exerts a hormonal control over carbohydrate metabolism by producing a hypoglycaemic substance". Research on incretin did not resume for over twenty years. With the revival of interest in the differences of oral and intravenous glucose tolerance, spawned by the availability of an RIA for insulin, the estimation of plasma insulin concentration was applied to the problem. McIntyre, Holdsworth and Turner (1964) compared, in human subjects, the effects of glucose administered intravenously or by direct intra-jejunal infusion. Lower blood glucose levels were obtained during the intra-jejunal infusion, but the plasma insulin levels were higher than those following the administration of intravenous glucose. They concluded that a humoral substance was released from the jejunal wall during glucose absorption and that this, together with the hyperglycaemia, stimulated the greater release of insulin. Also in that year, it was shown that when glucose was administered orally a greater and more sustained insulin release was seen than with intravenous glucose (Elrick e^ al., 1964). page 20 Creutzfeldt (1979) considers that there are two criteria which must be satisfied for a candidate hormone to be considered the incretin of Labarre. It must be an endocrine transmitter produced in the gastrointestinal tract which is released by nutrients and especially by carbohydrates. It must also stimulate insulin secretion in the presence of glucose if exogenously infused in amounts not exceeding blood levels achieved after food ingestion. The dependency of the incretin effect upon glucose levels, at least at physiological concentrations of the hormone, is an important prerequisite to the prevention of hypoglycaemia. Thus, an incretin can be released by different intestinal stimuli, but elevated incretin blood levels stimulate insulin secretion only in the presence of elevated blood glucose levels, i.e. when insulin is required. 1.5.2 Incretin Candidates Injection of a crude preparation of secretin significantly increased the disappearance rate of intravenously administered glucose and elevated immunoreactive insulin (IRI) in humans (Dupré, 1964). Similar preparations, when administered as large single doses, were found to stimulate insulin both in vitro (Pfeifer et al., 1965) and in vivo (Dupré et al., 1966). It was suggested that IRI release by secretin was potentiated by glucose (Dupré ejt al^. , 1969). It has been proposed that glucose and secretin probably stimulate functionally separate storage pools of readily releasable insulin (Lerner and Porte, 1972). However the dose of secretin required to show an insulinotropic action far exceeds the amount released endogenously and has page 21 therefore been regarded as pharmocologleal (Buchanan et al., 1968). Increases in plasma IR-secretin following mixed meals had not been demonstrated (Bloom, Bryant and Cochrane, 1975; Chey et al., 1975) until it was shown that small and transient spikes of secretin could be detected after a meal (Chey et al., 1978). As glucose does not stimulate secretin release (Creutzfeldt and Ebert, 1985) the necessary criteria to establish secretin as an incretin have not been fully satisfied. Like secretin, gastrin has not been shown to satisfy the necessary criteria to be considered an incretin. Plasma gastrin levels show only transient elevation after oral glucose in man (Buchanan, 1973). There are conflicting opinions on the effect of exogenous gastrin administration; there have been reports of stimulation (Unger et al., 1967; Dupré et al., 1969; Iverson, 1971) and no effect (Buchanan, Vance and Williams, 1969). But to date gastrin is still not considered an incretin (Marks and Turner, 1977; Creutzfeldt and Unger, 1985). It has been shown that an intestinal mucosal extract with biological activity similar to CCK-PZ enhanced the increase in serum IRI produced by the intravenous infusion of glucose (Dupré and Beck, 1966). It was also shown that under the appropriate conditions this preparation was also g lucagonotropic (Unger et al., 1967; Buchanan et al., 1968). It was subsequently suggested, following isolated perfused rat pancreas experiments, that insulin was secondary to glucagon release (Füsganger et al., 1969). In all of these studies the CCK-PZ preparation was that of Jorpes and Mutt. It was from this preparation that gastric inhibitory polypeptide (GIP) was initially isolated (Brown et al., 1969; page 22 Brown, Mutt and Pederson, 1970), so that the activities attributed to CCK-PZ might be assigned, in fact, to GIP. On reviewing the literature, Marks and Turner (1977) concluded that CCK-PZ probably had some insulin-releasing effects in the presence of hyperamino-acidaemia and could be capable of potentiating the insulin stimulating effect of hyperglycaemia. They also thought that CCK-PZ was unlikely to be responsible for the insulinotropic effect of oral glucose, but they concluded that it made a contribution to the overall insulin response to a mixed meal containing fat^protein and carbohydrate. So, even though CCK is not stimulated by glucose (Creutzfeldt and Ebert, 1985), and cannot itself be an incretin, it might be part of a net incretin effect, as GIP and CCK-8 in low doses, which alone exert no effect, possess strong insulinotropic activity when injected together in mice (Ahren, Hedner and Lundquist, 1983). page 2ô 1.4.0. Gastric Inhibitory Polypeptide (GIP) Gastric inhibitory polypeptide has been the subject of diverse physiological and chemical studies since its isolation (Brown, Mutt and Pederson, 1970). However, since its isolation and characterisation the polypeptide has aquired new names from groups of workers who considered the old name to be misleading, and their "new" name to be more appropriate. These names include glucose-dependent insulino- tropic polypeptide and glucose-dependent insulin-releasing polypeptide. Fortunately, all these names can be shortened to the perfectly adequate GIP, which is how it will be refered to below. 1.4.1. Discovery of GIP While investigating the actions of gastrointestinal extracts containing CCK-PZ, Brown and Pederson (1970) found evidence that there might be another inhibitor of gastric acid secretion present in some of the preparations being used in their study which had been characterized as having inhibitory activity for both exogenous (Gillespie and Grossman, 1964) and endogenous gastrin-stimulated acid secretion (Brown and Magee, 1967). Preparations of similar purity had also been described as being inhibitors of basal (Johnson and Magee, 1965) and stimulated motor activity (Brown Johnson and Magee, 1967). However, apparently conflicting reports of an acid inhibitory effect (Magee and Nakamura, 1966) and the stimulation of basal acid secretion page 24 (Murat and White, 1966) of CCK-PZ preparations needed explanation. From these results with impure preparations it was inferred that the gastric actions might have been due to the presence of hormones other than CCK-PZ. To test this hypothesis the effects of the two preparations of CCK-PZ on intra-gall bladder pressure changes and on the gastric parameters of acid secretion, pepsin secretion and motor activity were studied (Brown and Pederson, 1970). The two preparations were described as "10% pure" and "40% pure". The model was used in the dog with a vagally and sypathetically denervated pouch of the body of tbe stomach. When the preparations were administered in doses that gave similar gall bladder-contracting effects (50% of maximum) no difference in antral motor activity was seen. However the "40% pure" preparation showed a greater stimulation of gastric acid than the "10% pure" material. Two possible explanations for the uncoupling of gall bladder and acid stimulatory effects were proposed. Either a stimulator of gastric acid secretion had been selectively concentrated, or an inhibitor of gastric acid secretion was removed during the purification procedure. Further experiments with another CCK-PZ preparation convinced the workers that there was an inhibitory factor for acid secretion in the CCK-PZ preparation other than CCK-PZ itself. At the same time as carrying out physiological studies in dogs (Brown and Pederson, 1970) Brown and his co-workers were pursuing the possible enterogastrone using tissue extraction techniques (Brown et al., 1969). Their starting material was an approximately 10% pure CCK-PZ prepared as described by Jorpes and Mutt (1961). After chromatography three fractions were obtained. One of the fractions showed page 25 little cholecystokin in activity but strongly inhibited I gastric acid secretion. They had achieved the separation of cholecystokin in from acid inhibitory activities. Purification procedures were amended and subsequently material of sufficient quantity and purity for qualitative amino acid analysis and bioassay was obtained (Brown, Mutt and Pederson, 1970). Hence, GIP came into recognition. 1.4.2. Isolation and Purification As mentioned above GIP was separated from CCK-PZ and purified by the method of Brown, Mutt and Pederson (1970). Its purity was assessed using polyacrylamide gel electrophoresis, and was found to be homogeneous. The amino acid sequence was determined using a highly purified preparation of GIP which yielded only one peak by polyacrylamide gel electrophoresis and high-voltage electrophoresis. It was determined by sequence studies with peptides produced by tryptic digestion of both the complete molecule and the c-terminal cyanogen bromide fragment yielding a 43 amino acid polypeptide (Brown and Dryburgh, 1971). A synthetic 1-38 GIP preparation was found to have poor acid inhibitory activity (Moroder et al., 1978), thus the primary structure of natural porcine GIP was reinvestigated (Jornval et al., 1981). It was shown that porcine GIP was actually only 42 amino acids long, and that in the original sequence the glutamine at reside 29 (shown below) had been duplicated. page 26 NH^î-Tyr-Ala-Glu-Gly-Thr-Phe-I le-Ser-Asp-Tyr-Ser-I le- -Ala-Met-Asp-Lys-Ile-Arg-Gln-Gln-Asp-Phe-Val-Asn-Trp- -Leu-Leu-Ala-Gln-Lys-Gly-Lys-Lys-Ser-Asp—Trp-Lys-His- —Asn—Ile—Thr—Gin—OH More recently, human GIP has been isolated and sequenced. It has been found to differ from porcine GIP at only two residues. Amino acid residues at 18 and 34 are His and Asn respectively in human GIP (Moody et al., 1984). 1.4.4. Localisation Cellular localisation of IR-GIP using indirect immuno fluorescence techniques have shown that the polypeptide is present in cells situated predominantly in the mid zone of the glands in the duodenum and upper jejunum in both dog and man (Polak et al., 1973). The cells demonstrated positive immunofluorescence using guinea-pig anti-porcine GIP serum and fluorescein labelled sheep anti-guinea-pig IgG. They were found to be more numerous than secretin cells found in similar areas. Cytochemical and other staining reactions characterised the GIP cells as lead haemotoxylin-positive, Grimelius positive, non argentaffin and weakly positive for tryptophan. The cells were considered to belong to the APUD series or endocrine polypeptide cells (Pearse, 1969). The GIP cell was tentatively considered, as a result of electron microscopy, to be the Di cell because the distribution most closely matched the small granule-containing cells first described by Vassallo, Capella and Solcia (1971). In further studies immunofluorescence preparations demonstrated the presence of numerous medium sized endocrine cells which reacted to anti-GIP sera in the duodenum and jejunum from page 27 man, pig and dog (Bu-ffa et al., 1975). After further investigations the GIP cell was eventually allocated to the K cell (Solcia et al., 1974) in all three species. Buffa et al. (1975) concluded that the immunofluorescent GIP cells were ultrastructurally indistinguishable from the K cells, and that the D % cell was not involved in GIP production. The distribution of IR-GIP has been investigated in the small intestine of the hamster (Gaginella et al., 1978). IR-GIP could not be detected in nervous or muscle tissue, but only the mucosal layer of the gut. Figure 1.3 shows the distribution of IP-GIP in the whole human gut. 1.4.5. Physiological actions of exogenous GIP In the early studies on the purification of GIP the parameter measured was the ability of the extracts to inhibit acid secretion. The animal model used throughout these studies was the vagally and sympathetically denervated pouch of the body of the stomach of the dog following pentagas^trin stimulation (Brown and Pederson, 1970). In the innervated stomach preparation in dogs it was found that, following pentagastrin stimulation of acid secretion, the inhibitory effect of GIP was only weak (Soon-Shiong, Debas and Brown, 1979a). In addition, in the innervated stomach in man, pentagastrin stimulated acid and pepsin secretion was only weakly inhibited (Maxwell et al., 1980). It has been suggested that GIP acts indirectly on the parietal cell via a mechanism which can be regulated by parasympathetic innervation (Brown et al., 1980). Exogenous somatostatin is a potent inhibitor of acid secretion from the stomach, and in an isolated perfused rat stomach preparation exogenous page 28 administration of GIP produced dose-related increases in somatostatin-1ike-immunoreactivity (SLI) (McIntosh et al., 1979). It has been demonstrated that the acid inhibitory action of GIP can be blocked by the administration of a cholinomimetic substance (Soon-Shiong et al., 1979b). The perfusion of a rat stomach with GIP released SLI and a plateau was maintained. When acetylcholine was introduced into the perfusate a prompt and complete inhibition of SLI release was seen. A similar effect was observed if vagal stimuation was introduced after the establishment of SLI release. The above would indicate that GIP may exert its enterogastrone activity via SLI, but that this effect can be overridden by the gastric innervation, so that in the innervated stomach the enterogastrone effect of GIP is very weak (McIntosh et al., 1979). In addition to the inhibitory effect of GIP on the parietal cell, an acid inhibitory function for the hormone, via the inhibition of gastrin release, has also been described (Vi liar et al., 1976). A standard meat meal was fed to dogs that had been prepared with Heidenhain pouches. GIP was perfused intravenously for two hours. It was found that gastrin release was suppresed when the food was taken one hour after the commencement of the infusion. The effect of different doses of GIP on pepsin output from both the vagally denervated gastric pouch and the vagally innervated gastric remnant has been studied (Pederson and Brown, 1972). Intravenous GIP inhibited pentagastrin stimulated pepsin secretion from the vagally innervated gastric remnant: hypoglycaemia-induced pepsin secretion (bolus insulin injection) could be reduced by co-administration and susequent infusion of GIP. However, page 29 inhibition of pepsin secretion has not been demonstrated in humans when using intravenous infusions yielding supraphysiological serum IR-GIP levels (Maxwell et al., 1980). It was demonstrated long ago that the presence of food in the upper small intestine stimulated secretion from isolated denervated intestinal loops (Nasset, Pierce and Murlin, 1935). A humoral mechanism named enterocrinin was postulated (Nasset, 1938). The existence of such a humoral mechanism was confirmed more recently using balloon exclusion experiments (Wright, Barbezat and Chain, 1979). When GIP was infused intravenously, at a dose now known to result in supraphysiological circulating levels of the peptide, secretion rates from both Thiry-Vella fistulae of the upper jejunum and lower ileum were increased (Barbezat and Grossman, 1971). There was a large net reduction in water absorption from the jejunum when human subjects were infused intravenously with GIP, at a dose shown to elevate IR-GIP to a level within the normal postprandial range (Helmon and Babezat, 1977) . The reduction in water absorption was associated with a change in chloride flux which was reversed from a net absorption, during the control period, to secretion. The absorption of sodium, potassium and bicarbonate ions was significantly reduced during the GIP infusion period, but returned to pre-infusion levels when GIP administration was ended. It was concluded that these effects were not due to an interaction with adenylate cyclase, a conclusion in agreement with the findings of Schwartz et al., (1974). Other actions attributed to GIP include various effects on electrolyte transport in the main excretory duct of the mandibular gland of the rabbit (Oenniss and Young, page 30 1978), and a dose-dependent increase in cat superior mesenteric blood flow (Fara and Salazar, 1978). 1.4.6. Metabolic effects of GIP Brown (1982) has reviewed the literature on the insulinotropic actions of GIP and other gut hormones. He drew attention to the equivocal nature of much of the evidence which suggested insulinotropic roles for gastrin, secretin and CCK-PZ. Particular attention was drawn to the lack of information concerning the purity of hormone preparations used, the magnitude of the dose, the method of administration and the prevailing serum glucose concentrations. There is much evidence to suggest that GIP is insulinotropic in man and other mammals. In human volunteers when GIP is infused to physiological concentrations concurrently with glucose, an improved glucose tolerance curve is observed together with an increase in IRI (Brown e^ al., 1975). However, in fasted normoglycaemic subjects a similar infusion of GIP yields only small and transient increases in IRI release. The insulinotropic action of GIP is clearly glucose dependent, the results having been shown repeatedly in humans (Andersen et a1., 1978). It has also been demonstrated in vitro in isolated rat pancreatic islets (Schauder et al., 1975) and the perfused isolated rat pancreas (Pederson and Brown, 1976). Again when GIP was infused to achieve serum levels which were supraphysiological, transient insulinotropic effects were obtained in the dog (Pederson, Schubert and Brown, 1975). Using a constant perfusion level of GIP, and increasing concentrations of glucose in an isolated perfused rat page 31 pancreatic preparation, Pederson and Brown (1976) found that there was a threshold glucose concentration below which physiological levels of GIP did not significantly stimulate IRI. This glucose concentration was approximately 5.5 mM. This compares favourably with the figure of 25 mg/dL above basal (Elahi et al., 1979). The potentiating action of GIP on insulin release in the isolated perfused rat pancreas was maximal at a glucose concentration of 16 mM. The insulin output, however, was several-fold higher than could be produced by glucose alone. At a fixed glucose concentration of 8.9 mM, increasing the GIP concentration of the perfusate produced a dose-dependent increase in insulin output. It has been reported that high levels of GIP can stimulate the release of glucagon from the in vivo rat pancreas (Brown et al., 1975). However, the concentration of GIP used in this experiment (100 ng/ml) can certainly be regarded as supraphysiological. Glucagon and insulin responses of the perfused rat pancreas to glucose in the presence and absence of GIP have been investigated (Pederson and Brown, 1978). A glucagonotropic effect of GIP was observed at low glucose concentrations, but this response was suppressed by increasing the glucose concentration. The potentiation of glucagon release by GIP was not significant at glucose concentrations above 5.5 mM, at which the insulinotropic effect of GIP became apparent. In man, however, when GIP was perfused at a concentration which achieved a post prandial-like elevation in plasma IR-GIP, no rise in IR-glucagon was observed (Elahi et al., 1979). Portal hyperinsulinaemia and hyperglycaemia both suppress glucose production by the liver. Other factors also contribute to this metabolic process. The role of GIP has page 32 been studied in this process in dogs with chronic portal venous catheters (Andersen et al., 1980). A GIP infusion was shown to reduce hepatic glucose output without concomitant increases in IRI, glucose or glucose disposal levels. It was suggested that GIP possesses a significant suppressive effect on hepatic glucose output in both the basal state and when already lowered by hyperinsulinaemia. It was also shown to potentiate the effect of insulin on net glucose disposal. The conclusion was that GIP may act synergistically to augment the hepatic glucose response as well as the B-cell response to ingested food. Possible interactions between GIP and glucagon in fat cells prepared by collagenase digestion of epididymal fat pads have been studied (Dupré et al., 1976). GIP was not only non-lipolytic but also inhibited glucagon-stimulated lipolysis. It had, however, no effect upon the lipolysis induced by secretin or VIP. GIP could displace bound isîtsj-g lycragon (biologically identical to "cold " glucagon) from fat cells. VIP and secretin had previously been shown unable to displace læoi—giucagon (Bataille, Freychet and Rosselin, 1974). GIP was itself demonstrated to have a weak lipolytic effect, but a stong anti-lipolytic action on glucagon-stimulated lipolysis by selectively blocking the action of adenylate cyclase, hence blocking the lipolytic action of glucagon (Ebert and Brown, 1976). GIP has been shown to stimulate lipoprotein lipase activity in 3T3-L1 cells, an established mouse fibroblast line resembling an adipocyte (Eckel, Fujimoto and Brunzell, 1979). These authors considered that the increased lipoprotein lipase activity produced by GIP could provide a mechanism for clearance of chylomicron triglyceride after □aae 33 feeding. The theory that GIP could provide a mechanism for the clearance of lipid from the circlation after feeding was soon put to the test in vivo. Chyle was collected from donor dogs via a thoracic duct fistula and then reinfused into normal recipient dogs during the infusion of either porcine GIP or normal saline (Wasada et al., 1981). It was found that in the GIP-infused animals the rise in plasma triglyceride was significantly below that of the control animals. The authors concluded that GIP does indeed have an effect upon the removal of chylomicron triglyceride from the blood. They also found that plasma glucagon rose significantly during the GIP infusion, and continued to rise during the infusion of chyle, while insulin levels remained unaffected. Thus, earlier work (Rouiller et al., 1980) which claimed a glucagonotropic action for GIP, was confirmed. Pro-insulin biosynthesis has been studied in collagenase-isolated rat pancreatic islets (Schafer and Schatz, 1979). Both incorporation of =*H-leucine and release of insulin were augmented by GIP when the islets were incubated with 1.0 and 2.0 mg/ml glucose for 3 h in the presence of 5 pg/ml GIP. This insulinotropic activity was not observed with 20 pg/ml of an amino acid mixture in the presence of 5 pg/ml GIP. Insulin levels were not augmented by GIP in the presence of the amino acid mixture when the glucose concentration was 1 mg/ml. However, using physiological concentrations of GIP, other workers have shown an enhanced insulin release in response to GIP in the presence of amino acids using monolayer cultures from islet cells of neonatal Wistar rats (Fujimoto et al., 1978). The authors confirmed, however, the glucose-dependent nature of page 34 the insulinotropic action of GIP. They showed that GIP doses as low as 1.0 ng/ml significantly enhanced the IRI release in the presence of 16.5 mM glucose, but at a glucose concentration of 1.7 mM a GIP concentration of 10 ng/ml or greater was required to elicit a response. Using short term cultured isolated rat islets, Siegel and Creutzfeldt (1985) have shown that, at physiological concentrations, GIP augments glucose stimulated (16.7 mM) insulin release. They also showed that under these conditions islet cyclic AMP concentrations were also increased. The authors conclude that GIP exerts its insulinotropic effect via cyclic AMP. Results obtained by other workers seem to indicate that insulin release is enhanced to a relatively greater extent than insulin-pro-insulin biosynthesis by GIP (Schafer and Schatz, 1979). As indicated above, collagenase isolated islets did not respond with the release of IRI until extremely high (and probably supraphysiological) concentra tions of GIP were added (Schrauder et al., 1975). The monolayer cultures on the other hand (Fujimoto et al., 1978) responded in a way similar to perfused rat pancreas preparations (Pederson and Brown, 1976) to GIP concentrations as low as 1.0 ng/ml. Brown (1982) considers that these apparently contradictory pieces of work support the hypothesis that collagenase treatment alters receptor sites for GIP on the islet cell membrane, which will recover following short term culture. Hence, non-cultured collagenase-isolated islets must be considered an unsuitable model for studies on the action of GIP and probably also for other insulinotropic peptides. page 35 1.4,7. Stimulation of immunoreactive GIP (IR-GIP) release A radioimmunoassay (RIA) for GIP was first described by Kuzio et al. (1974). It utilized an antiserum raised in guinea-pigs to pure porcine GIP conjugated to bovine serum albumin (BSA) by the carbodiimide condensation reaction (Goodfriend et al., 1964). Since then many other groups have developed their own GIP assays (Morgan, Morris and Marks, 1978; Lauritsen and Moody, 1978; Sarson, Bryant and Bloom, 1980; Burhol, Jorde and Waldum, 1980). Kuzio et al. (1974) used their RIA to study the physiological regulatory mechanism of GIP release. They reported that the serum IR-GIP levels in normal human subjects following an overnight fast were 237 + 14 pg/ml (mean + SEM) with a range of 75 - 500 pg/ml. After feeding a mixed meal, serum levels increased in a biphasic manner to approximately 1200 pg/ml within 45 minutes of food ingestion and remained significantly in excess of basal levels for over 4 hours. Similar results have been reported using the assays developed by Morgan, Morris and Marks (1978) and Lauritsen and Moody (1978). One group, however, have reported absolute fasting values five times lower than those of other groups (Sarson, Bryant and Bloom, 1980). There were, however, similarities in the degree of change observed in the response to a mixed meal. With the development of a RIA it was soon demonstrated that GIP played a significant part in the enteroinsular axis. It was shown that GIP could be released into the circulation by the ingestion of a secretogogue known to release insulin, namely glucose. Serum levels of IR-GIP, IRI and glucose were page 36 monitored following the ingestion of 75 g glucose (Cataland et al., 1974). Serum IR-GIP levels became elevated within 15 minutes of glucose ingestion and peaked at 30 minutes with a doubling of basal levels. The IR-GIP levels appeared to plateau from 30 to 60 minutes, before declining slowly over the next two hours. Intravenous glucose given as a single bolus injection (25 g) did not produce an increase in serum IR-GIP levels. By measuring portal venous levels of IR-GIP, IRI and glucose, these workers also demonstrated that portal IR-GIP levels were significantly elevated within 2 minutes of glucose ingestion, whereas insulin levels were increased at 5 minutes. It was further reported that several of the subjects demonstrated a fall in fasting levels of IR-GIP at 20 to 40 minutes after intravenous glucose. They suggested that glucose, insulin or some other factor may bring about regulation of IR-GIP by feed-back inhibition. The release of IR-GIP and its association with insulin release was studied after oral glucose in normal human volunteers using a hyperglycaemic clamp technique (Andersen et al., 1978). Little change in plasma IR-GIP levels occurred when intravenous glucose alone was administered (maintaining the hyperglycaemia at approximately 125 mg/dL above basal), but following oral glucose (40 g/m=^) IR-GIP levels doubled within 40 minutes. The oral glucose ingestion produced a highly significant increase in IRI levels (above those produced by intravenous glucose alone). These studies confirmed the work of McIntyre, Holdworth and Turner (1964) in demonstrating that blood glucose levels are rigidly controlled. In addition they demonstrated that the time course of the rise of IR-GIP and IRI were nearly identical. In further similar experiments Andersen et a 1. (1978) page 37 established an euglycaemic clamp in normal human volunteers to maintain basal blood glucose levels. A primed continuous insulin infusion of 120 mU/m^/min was given together with a servo controlled glucose infusion, producing a hyperinsulinaemia of approximately 300 pU/ml. On the administration of a similar dose of oral glucose at 60 minutes, plasma IR-GIP levels were found to increase in a manner similar to that in the hyperglycaemic clamp experiment. However, if the plasma glucose levels could be maintained at euglycaemia, no increase in endogenous insulin release was observed, despite the rise in IR-GIP; In several instances blood glucose concentrations could not be maintained at basal and it was noted that, if an increase of 20 mg/dl or more occurred, there was an increase in plasma IRI associated with the elevation in IR-GIP. The authors concluded that the effect of glucose-induced IR-GIP release on insulin secretion was dependent upon the prevailing degree of hyperglycaemia, and was not inhibited by the presence of marked hyperinsulinaemia produced by the exogenous insulin infusion. These studies did not, however, investigate the involvement of other gut hormones in the enteroinsular axis. It has been shown that IR-GIP release occurs in a dose related manner following oral glucose in man (Crockett et al., 1976) and in the dog (Pederson, Schubert and Brown, 1975). IRI release has been studied following oral plus intravenous glucose and also after exogenous porcine GIP plus intravenous glucose in dogs (Pederson, Schubert and Brown, 1975). Although the serum levels of IR-GIP achieved by the intravenous infusion of porcine GIP were higher than those achieved following oral glucose, the insulin output was also greater. Studies in dogs with Mann-Bollman fistulae showed page 38 that equal volumes of both 10% and 20% solutions of glucose, introduced directly into the duodenum, were potent stimuli for the release of IR-GIP (Martin et al., 1975). Naturally the higher concentration glucose solution was the more potent stimulus. No effect on IR-GIP was observed when hyperosmolar solutions of galactose and mannitol were used. The authors concluded that the IR-GIP response to glucose was dose related and that, in dogs, duodenal osmoreceptors did not play a primâtry role in the physiological release of IR-GIP. In man, however, oral galactose was found to release IR-GIP (Cleator and Gourlay, 1975; Morgan, Wright and Marks, 1979). Oral galactose was also found to be slightly insulinotropic and during hyperglycaemia, caused by an intravenous infusion of glucose, this affect was augmented. In the rat it has been shown that glucose, galactose and sucrose stimulate IR-GIP release while fructose does not (Morgan, Wright and Marks, 1979). It would appear that for the release of IR-GIP release by a sugar, the monosaccharide must be absorbed via a Na*^ dependent carrier system in the mucosa (Morgan, 1979). If the carbohydrate is present in a poorly-hydrolysable form IR-GIP release is impaired. The failure of lactose to stimulate IR-GIP release in the rat has been attributed to the low levels of lactase found in the adult (Sykes et al., 1980). It was also shown that other molecules which shared the Na^ dependent glucose carrier such as galactose, 3—0-methyl glucose and alpha- or beta- methyl glucoside, could also produce significant increases in IR-GIP. The transport of fructose (which does not stimulate IR-GIP) is generally considered to be passive or facilitated (Crane, 1968; Gray, 1975). The addition of sodium chloride to an oral glucose page 39 load significantly enhances the IR-GIP and IRI response. (Ebert and Creutzfeldt, 1980). The prevention of intestinal absorption of glucose by the use of an inhibitor such as phloridzin also prevents the release of IR-GIP (Creutzfeldt and Ebert, 1977; Sykes et al., 1980). In other situations where glucose absorption is impaired, such as patients with coeliac disease, the release of IR-GIP is also reduced (Creutzfeldt et al., 1976; Besterman et al., 1978). TRIS (tris-hydroxymethyl-aminomethane), a compound which competitively blocks intestinal brush border sucrase, when ingested with a glucose load significantly reduces the expected IR-GIP and IRI responses (Ebert and Creutzfeldt, 1978). Again, if a glucosidase inhibitor, such as BAY g 5421, is ingested with a starch load lower IR-GIP and IRI responses are observed (Creutzfeldt et a 1., 1979). Triglyceride has long been known to be one of the most potent stimuli to IR-GIP release. It was first described by Brown (1974) who claimed that the ingestion of 100 ml (66 g ) of a corn oil suspension (Lipomul) in normal human volunteers caused a large rise in serum IR-GIP levels, reaching a peak of 3 - 4 times basal within approximately 120 minutes of the load. These results were confirmed within a short time (Falko et al., 1975), and it was also reported that the ingestion of emulsified corn oil had no effect on glucose, IRI or non-ester ified fatty acid concentrations. It was clear that the endogenously released IR-GIP was not insulinotropic in the absence of hyperglycaemia. Work has been done to show that, in the same way as for glucose-stimulated IR-GIP release, triglyceride-stimulated IR-GIP release requires the hydrolysis and subsequent absorption of hydrolysis products to elicit an IR-GIP page 40 response. Again the mere presence of the nutrient within the small intestine is not enough to cause the release of IR-GIP. Groups of children with and without cystic fibrosis each ingested corn oil and serum IR-GIP was measured. The normal children had a ten-fold increase in serum IR-GIP following the triglyceride, but no increase in serum IR-GIP occurred in the children with cystic fibrosis (Ross and Shaffer, 1981). These authors then went on to assess the relative importance of the products of triglyceride hydrolysis and also fatty acid chain length in IR-GIP release. On separate occasions adult male humans consumed equimolar amounts (40 mmol) of corn oil, medium chain triglycerides, long chain fatty acids and glycerol. Long chain fatty acids caused a four— fold increase and corn oil a twelve-fold increase in IR-GIP levels. No effect was observed with medium chain triglycerides or glycerol. The volunteers also consumed a monoglyceride with no effect on serum IR-GIP levels. However, it seems unlikely that the monoglyceride used in this experiment would have been well absorbed as it is still solid at 35*C and thus unlikely to be in an appropriate state to stimulate IR-GIP release. Other workers published data from dog experiments which were in broad agreement with the above results, confirming that corn oil is better at eliciting an IR-GIP response than its components when equimolar doses are administered via gastric tube (Williams, May and Biesbroeck, 1981). It has been reported that the ingestion of a meat extract did not cause any measurable change in serum IR-GIP (Brown, 1974). It was also observed that the ingestion of 280 g fillet steak was also unable to elicit an IR-GIP response (Cleator and Gourlay, 1975). However, it was subsequently page 41 shown that the intraduodenal instillation of a mixture of amino acids caused a transitory elevation in serum IR-GIP concentrations together with a brief increase in serum IRI (Thomas et al., 1976). Quantitative differences were further demonstrated with respect to the IR-GIP releasing activity, the insulinotropic activity and the pancreatic exocrine stimulating effect of certain amino acids (Thomas et al., 1978). It was shown that a mixture of amino acids containing arginine, histidine, isoleucine, lysine and threonine caused a marked rise in integrated IR-GIP and IRI secretion but only a small exocrine pancreatic effect as indicated by trypsin output. It was also shown that a solution of amino acids containing methionine, phenylalanine, tryptophan and valine had only a slight effect on IR-GIP and IRI release, but there was a marked rise in bilirubin and pancreatic trypsin output. 1.4.8. Inhibition of IR-GIP release other interactions between glucose, GIP and insulin were sought (other than those where glucose stimulates IR-GIP release and they together stimulate IRI release). The role of exogenous insulin in the regulation of IR—GIP release was one of the first of these to be studied (Brown et al., 1975). Following an oral fat load, plasma IR-GIP levels were significantly diminished when the subjects also received a bolus injection of insulin (7^U) followed by an intravenous infusion of glucose (to prevent hypoglycaemia). Similar results were also obtained following an oral fat load when accompanied by only an intravenous infusion of glucose (Cleator and Gourlay, 1975). However, fat loads did not cause page 42 an IRI release. In the second experiment it was noted that intravenous glucose with simulataneous oral triglyceride produced a significantly greater IRI release than intravenous glucose alone, resulting in better serum glucose control. Many workers have found the concept of a feed-back inhibitory control for IR-GIP release involving insulin an especially attractive mechanism. It would help to explain the elevated IR-GIP levels found in patients with insulinoprivie diabetes and could also explain the higher levels found in patients with chronic pancreatitis (Botha, Vinik and Brown, 1976; Ebert et al., 1976). A significantly elevated IR-GIP release in response to a test meal in patients with chronic pancreatitis has been observed (Ebert et al■, 1976). When the patients were divided into three groups on the basis of their insulin responses, a significantly higher IR—GIP release was observed in the group with the intermediate IRI and glucose responses than those groups with the highest and lowest responses. The group with the lowest response had the greatest impairment of exocrine pancreatic function and therefore the greatest absorption problems. Although insulin has not been found to inhibit glucose- stimulated IR-GIP release (Brown, 1982), many workers have found that intravenous infusions of either glucose or insulin reduce the rise in fat-stimulated plasma IR-GIP release (Brown et al., 1975; Crockett et al., 1976; Cleator and Gourlay, 1975; Verdonk et al., 1980). But although obese subjects have been shown to have exaggerated IR-GIP responses after a mixed meal (Ebert et al., 1976), exogenous insulin fails to inhibit fat stimulated IR-GIP release in these subjects (Creutzfeldt et al., 1978). An intravenous infusion of glucagon has been shown to oaoe 43 depress -fasting IR-GIP levels, together with an increase in IRI and serum glucose (Ebert, Arnold and Creutzfeldt, 1977). It was also shown, in normal human volunteers, that prior intravenous administration of exogenous glucagon was able to completely suppress the serum IR-GIP response to a liquid test meal. Discontinuation of the glucagon infusion resulted in an immediate rebound increase in IR-GIP release above that seen with the test meal alone. In a further experiment the glucagon infusion was started 60 minutes after the ingestion of the test meal, resulting in a prompt and almost total inhibition of IR-GIP. The rapid fall in fasting and test meal-stimulated IR-GIP release, after the onset of glucagon infusion, may be indicative of a direct effect in the GIP-producing cells of the gastrointestinal mucosa, or modulation of its release by changes in serum levels of glucose or insulin (Ebert, Arnold and Creutzfeldt, 1977). Furthermore, it has been argued that glucagon was not exerting its effect by influencing absorption because the inhibition could not be observed in the fasting state. Hyperglycaemia, which was also observed in these studies, was considered not to be a cause of the inhibition because it has been reported (Creutzfeldt et al., 1976; Ebert, Frerichs and Creutzfeldt, 1976) that there was an exaggerated IR-GIP response to a test meal in patients with pathologically high glucose levels (diabetes). It was concluded that the most probable mechanism by which glucagon inhibited IR-GIP secretion was via direct action at the cellular level. Overal other substances have been shown to have an inhibitory effect upon the release of IR-GIP. Rat C-peptide II has been shown to inhibit fat stimulated IR-GIP release (Dryburgh et al., 1980). C-peptide released after stimulation page 44 of the pancreas by intravenous glucose and tolbutamide in combination with insulin antiserum, was also shown to significantly inhibit fat stimulated IR-GIP release. Intravenously administered atropine has been shown to inhibit intraduodenally perfused glucose-stimulated IR-GIP release in humans (Larrimer et al., 1978). The subcutaneous injection of atropine sulphate has also been shown to abolish completely the meal stimulated rise in IR-GIP; there was, however, no effect on basal IR-GIP release (Baumert et al., 1978). Intravenous administration of somatostatin has been found to delay the increase in serum concentrations of IR-GIP, IRI and glucose in dogs (Pederson, Dryburgh and Brown, 1975). When somatostatin was injected prior to triglyceride ingestion, the mean IR-GIP serum levels were suppressed for the entire course of the experiment. It was further shown that pretreatment with somatostatin produced a substantial reduction in the peak IRI response to a dose of porcine GIP usually found to be insulinotropic. The IR-GIP response to a liquid test meal in man has been shown to be suppressed completely by the infusion of somatostatin (Creutzfeldt and Ebert, 1977). IRI was also suppressed for as long as the somatostatin was infused. Blood glucose concentrations, however, sowed no difference, again indicating that the inhibition of IR-GIP release by somatostatin was not secondary to changes in glucose absorption. page 45 1.5.0 Mot!lin 1.5.1. Historical background and discovery of motilin The existence of motilin was postulated because of studies which demonstrated that alkalinization of the duodenum in dogs, with either fresh pancreatic juice or 0.3M TRIS buffer pH 9.0, induced an increase in the motor activity in pouches of the body of the stomach (Brown, Johnson and Magee, 1966). A very crude preparation of CCK-PZ was used in early purification attempts (Brown and Parkes, 1967). These workers chose this material as their starting point following the observations that, of several gut hormone preparations, it was only the "Pancreozymin" which was capable of reproducing the gastric motor activity stimulatory effect of duodenal alkalinization (Brown and Parkes, 1967). Motilin was finally purified from side fractions produced during the purificaton of secretin obtained from Mutt and Jorpes (Brown, Mutt and Dryburgh, 1971). 1.5.2 Structure of motilin Following isolation and purification of the peptide hormone, it was sequenced and found to be 22 amino acid resides in length (Brown, Cook and Dryburgh, 1973). Various analogues were synthesised with substitutions at positions 13, 14 and 15 with little or no differences in immunological or biological activities (Wiinsch et al., 1973; Yana i hara et paqe 46 al., 1977; Brown and Dryburgh, 1978). Motilin has been found to exist in multiple molecular forms. In human gut extracts, subjected to gel filtration, two peaks of immunoreactive motilin (IRM) have been resolved, one peak co-eluting with pure porcine motilin (Christofides et al., 1981). 1.5.3 Cellular localisation of motilin IRM was first reported to be localised in the enterochromaffin (EC) cells of the dog, pig and baboon (Pearse et al., 1974). Subsequently IRM was shown to be present exclusively in the ECs* cell type (Heitz, Polak and Pearse, 1978). The presence of IRM has been described within nj^rones in the submucosae and muscle layers of human oesophagus, stomach, small intestine, colon and bladder (Chey and Lee, 1980). IRM has been found widely distributed in the rat central nervous system (O Donohue et al., 1981). 1.5.4 Physiological actions of motilin In vitro, the action of motilin is specific to smooth muscle of gastrointestinal origin (Domschke et al., 1974). There is a gradient of sensitivity to motilin down the alimentary tract, the highest susceptabi1ity being in gastric and duodenal muscle strips (Strunz et al., 1975). It has been found that the effects of motilin on isolated gastrointestinal muscles are not mediated by nervous pathways, and it was concluded that motilin causes contractions by stimulating receptors on or in the muscle cell (Strunz et al., 1976). page 47 Natural porcine motilin or a synthetic analogue, when given intravenously, has been shown to stimulate contraction of the stomach (Brown, Mutt and Dryburgh, 1971), lower oesophageal sphincter (Jennewein et al.,1975; Meisner et al., 1976) and small intestine in the dog (Lee, Hendricks and Chey, 1974; Jennewien et al., 1975). Motilinhas been shown to cause an increase in tone of the sphincter of Oddi in dogs (Neya et al., 1981). This was shown to be unaffected by denervation or atropinisation. Two groups of workers (Wingate et al., 1975; Itoh et al., 1975) reported independently that motilin produces a typical aborally propagated interdigestive myoelectric complex (IMC). The existence of a cyclical pattern of fasting gastrointestinal motor activity was originally demonstrated in the dog (Boldyreff, 1911: cited by Rees et al., 1982) to be confirmed many years later, again in the dog, as a migrating electric complex (Szurszewski, 1969). It was reported that 13-norleucine-moti1 in evoked phase III (activity front) of the IMC, starting in the antrum and proximal duodenum and slowly migrating along the small intestine to reach the terminal ileum in 60 to 80 minutes fWingate et al., 1975), It was shown that the intravenous bolus injection of synthetic motilin resulted in a prompt occurrence of phase III, thereby shortening the duration of phases I (quiescence) and II (occasional, irregular contractions). Following the premature phase III, phases IV (decline in activity front). I, II and III occurred at their expected regular time intervals (Chey and Lee, 1980). The effects of exogenous motilin on gastric emptying have been studied in man (Ruppin et al., 1975; Christofides et al., 1978) and in the dog (Debas, Yamagishi and Dryburgh, page 48 1977). In man synthetic 13-norleucine-moti1 in was found to delay gastric emptying of liquid meals, while intravenous natural porcine motilin accelerated the gastric emptying of solid meals. Comparisons of these two results are difficult as different preparations and doses of motilin were used. In dogs it has been observed that natural porcine motilin enhanced emptying of a liquid meal while the emptying of a solid meal was not affected (Debas, Yamagishi and Dryburgh, 1977). Apart from its many actions upon gastrointestinal smooth muscle a few other actions have also been ascribed to motilin. It has been shown to affect the secretory function of the stomach by stimulating pepsin secretion (Brown, Mutt and Dryburgh, 1971); Koch et al., 1976). It has also been shown to affect exocrine pancreatic function by weakly stimulating bicarbonate and protein secretion, and to act as an inhibitor to secretin-stimulated pancreatic secretion (Konturek et al., 1976). Finally, the most important action of motilin (in the context of the experimental work which will follow) is that it has been shown that exogenous motilin can cause an increase in the initial rises of IR-GIP and IRI following oral glucose (50 g ) (Long et al., 1982). 1.5.5 Modulation of circulating IRM Plasma concentrations of IRM have been determined using RIA by several groups of investigators (Dryburgh and Brown, 1975; Bloom, Mitznegg and Bryant, 1976; Itoh et al., 1978; Tai and Chey, 1978; Thomas, Kelly and Go, 1979). Basal IRM concentrations have been found to be subject to large page 49 inter-personal variations (Bloom, Mitznegg and Bryant, 1976). Fasting plasma IRM concentrations in 110 subjects showed a skew distribution, with levels ranging from 11 to 945 pg/ml. It was shown that duodenal alkalinization in dogs caused an increase in plasma IRM (Dryburgh and Brown, 1975; Lee e_t a_^. , 1978). Duodenal acidification has also been shown to increase plasma IRM levels in man (Mitznegg et al., 1976), dog (Lee et al., 1978; Tai and Chey 1978) and pig (Mitznegg et al,, 1978). A mixed meal has been shown to cause a significant, but transient, rise in plasma IRM levels in human volunteers where IRM levels returned to 'basal' within 60 minutes (Christofides et al., 1979a). The same authors went on to show that oral glucose was a potent inhibitor of IRM, whereas oral double cream was a potent stimulus to plasma IRM concentrations. Distension of the stomach, either by mechanical means or ingestion of a large volume of water, results in large increases of plasma IRM (Chr istof ides ejt a1., 1979b). Plasma IRM concentrations can also be modified in man by intravenous nutrients. Infusions of glucose and amino acids both result in a suppression of IRM, whereas intravenous fat provokes an increase (Christofides et al., 1979a). Release of IRM does not appear to be under cholinergic influence. Atropine was reported to be ineffective in Inhibiting the rise of IRM provoked by gastric distension (Christofides et al., 1979a). Basal motilin levels are, however, atropine sensitive; giving a large reduction lasting several hours (Christofides and Bloom, 1981). Other gastrointestinal hormones also appear to provoke large motilin response. Somatostatin, and pancreatic polypeptide (PP) infusions cause a substantial suppression of IRM page 50 concentrations, while bombesin appears to stimulate its release (Adrian et al., 1980). The role of motilin as an interdigestive hormone has, more recently, been questioned. In a study investigating the action of PP in stimulating intestinal activity, it was observed that PP substantially suppressed plasma IRM concentrations without inhibiting the initiation of motor complexes; activity fronts could be initiated even when peripheral plasma motilin levels were suppressed (Adrian ejt al., 1980). A large group of women were investigated at 16 and 36 weeks pregnancy and one week post partum. At both time points during pregnancy plasma IRM levels were very significantly lower than those observed at one week post-partum (Christofides et al., 1982). It is thus that motilin has been implicated, by the above authors, in the recurrent heartburn, epigastric discomfort and constipation often reported in pregnancy. page 51 1.6.0 Summary of introduction and aims The above describes how GIP seems to be a hormone of many parts. Although its originally characterised action of gastric inhibition seems fairly equivocal in humans, its role in the enteroinsular axis as an incretin seems more secure. In addition, important new metabolic roles for GIP have emerged in post prandial lipid metabolism. Plasma IR-GIP is stimulated by fats carbohydrates and proteins/amino acids. It seems fairly reasonable to postulate a general role for GIP in the humorally mediated management of the metabolic aspects of digestion via the enteroinsular axis and its actions on the adipocyte, even though the physiological management of digestion, such as gastric emptying, mesenteric blood flow, saliva compostition etc., seem less well established. Hence it will be the aim of this thesis to investigate the metabolic actions of GIP. The physiological actions of GIP will be considered as and when they occur, but will not be a prime objective. To this end some basic parameters must first be examined. Firstly, of importance in the timing and planning of experiments, is there a circadian rhythm in either basal IR-GIP release or in stimulated IR-GIP release? Also, as most of the experiments are carried out following oral stimuli, factors which may affect gastric emptying, such as motilin, should also be considered. Having established the basic parameters, the bulk of the experimental work will seek to further elucidate the metabolic roles of GIP. oaoe 52 CHAPTER TWO METHODS page 5^ Unless stated otherwise all reagents were obtained from BDH Ltd., Poole, Dorset, and were of AnalaR grade. All water was double glass distilled water, stored in glass until required. Pipettors and tips used were either Gilson (Anachem Ltd.) or Oxford (BCL Ltd.). They were used in accordance with manufacturers instructions and regularly cleaned and checked for accuracy (assuming 1.00 ml water = 1.00 g at 20*C). 2 . 1.0 The Gastric Inhibitory Polypeptide (GIP) Assay The IR-GIP assay was the main tool of the experimental work which follows. The assay was essentially that of Morgan et a 1. (1978) with only a few minor modifications. lod ination The iodination was a modification of that of Kuzio ejt a 1. (1974). Pure porcine GIP (EG III, Prof. J. Brown, Vancouver, Canada) stored dessicated at -20*C was used for the production of the iodinated tracer for the assay. The vial containing the GIP was allowed to equilibrate to room temperature before opening (10 - 20 minutes). Approximately 20 jjg was quickly removed from the vial (in order not to allow the stock to absorb too much atmospheric water) and weighed on a microbalance (Perkin Elmer AM2). It was dissolved in phosphate buffer (pH 7.5, 0.4M; 5 pg GIP per 10 pi buffer). Aliquots (10 pi) were pipetted into plastic Auto Analyzer cups (Sterilin) and used immediately or stored at page 54 -40*C until required. To the 10 pi of GIP solution was added ^^^I-sodium iodide (10 pi = 1 mCi; IMS 30, Amersham International) and chloramine-T (15 pg in 10 pi phosphate buffer pH 7.5, o.4M) in quick succession. The reactants wereagitated for 15 seconds with the pipette tip that was used for the addition of the chloramine-T. Sufficient mixing was achieved by drawing the reaction mixture up into the tip and expressing it 8-10 times. The reaction was stopped by the addition of sodium metabisulphite (40 pg in 20 pi phoshate buffer pH 7.5, 0.4M). The reaction mixture was immediately diluted with chromatography buffer (0.5 ml, acetate buffer 0.1 M, pH 5.0 containing 0.5 % human serum albumin CHSA, Lister Institute] and 500 KlU/ml aprotinin [Novo, Denmark]) and transferred to 1.1 X 15 cm column (Wright/Amicon) of superfine Sephadex G15 (Pharmacia) which had been previously swollen and equilibrated with the chromatography buffer. The column was then eluted with the chromatography buffer. 35 - 40 ten drop fractions (0.5 ml on a LKB Ultrorac 7000 fraction collector) were collected. 10 pi fractions were counted for 10 seconds on a LKB 1260 Multigamma II. The two fractions with most radioactivity were pooled and diluted 1:1 with acid/ethanol (1:100, concentrated HCl:absolute ethanol) and kept at -20*C until required. A typical elution pattern is shown in figure 2.1, the first peak being the iodinated-GIP and the second being free iodine. It can be seen that this protocol yields two "clean" peaks of free and bound iodine. Typically the yield of iodinated GIP is 50 - 60 % of the iodine used in the iodination. The maximum theoretical specific activity for mono-iodinated GIP is 390 nCi/ng. The specific activity of page 55 the peak -fraction shown in figure 2.1 is 211 nCi/ng. The specific activity is estimated by calculating the concentration of IR-GIP in increasing amounts of radioiodinated-GIP added to the assay (Walker, 1977). Standards Pure porcine GIP (EG III) was weighed out (approximately 40 pg, as above) and dissolved in a buffer as described by Sarson (1980) consisting of 0.14 M lactose, 0.04 M HSA, 11 nM citric acid, 6 nM cysteine hydrochlorise and 1600 KlU/ml aprotinin in 0.1 M formic acid. 100 pi aliquots containing 100 ng GIP were freeze dried to constant vacuum and stored at -20*C. Charcoal stripped serum Pooled fasting human serum was used to produce the charcoal stripped serum (CSS). 20 % (w/v) charcol (GSX, Sigma) was added to the serum pool. It was stirred overnight (18 h) at 4°C. The mixture was centrifuged 6000 g for 30 minutes (J6 centrifuge, Beckman), the supernatant was filtered through a Filtrox filter (grade W-Steril, H. Erben Ltd., Ipswich), and if necessary the filtration was repeated to remove any residual fines, until a pale coloured serum was obtained. The CSS was aliquotted into 5 ml portions and stored at -20*C. New batches of tracer, standard and CSS were always tested against the previous batch (except tracer) and quality controls, before use on samples, to ensure continuity of resu1ts. page 56 Assay The assay was unaltered from that of Morgan et al. (1978). Assay buffer was 0.04 M phosphate pH 6.5 containing 0.5 % HSA and 500 KlU/ml aprotinin. The assay protocol is shown in table 2.1. Standard curves were always set up in CSS. Donkey anti-rabbit antibody D41 (titre 1:16 initial) was obtained from Guildhay Antisera, Guildford. Normal rabbit serum (heat inactivated, Wellcome) was used at an original titre of 1:135. Following the addition of the normal rabbit serum and second antibody, 0.12 ml 12 7. PEG (mol. wt. 6000) was added, to accelerate the reaction, and the assay was left for 4 h before centrifugation (15 minutes at 5000 g , Beckman J6 centrifuge). The tubes were aspirated (Pasteur pipette and water pump) and counted for 2 - 4 minutes (to obtain 20 000 counts in the "total count" tubes) on a LKB 1260 Multigamma II. Data was processed by the on-board microprocessor using a Sp1ine-function best-line fitting program. Antibody All assays were performed with rabbit anti-porcine GIP antiserum (RIC 34/11 If, kindly provided by Dr. LM Morgan). The rabbit was immunized with a GIP/ovalbumin g lutaraldehyde conjugate. The antibody was used routinely at an initial dilution of 1:8000. The antiserum shows less than 1 7. cross-reactivity with glucagon, VIP, secretin, pancreatic polypeptide, insulin ot C-peptide. The antiserum recognises both the 5000 and 8000 molecular weight forms of GIP (Morgan et al., 1978). A typical standard curve is shown in figure 2.2. The limit of detection of the assay when calculated as 2 x SD from zero (n = 12) was found to be 80 pg/ml. The limit of page 57 detection of the assay, when taken as a 10 % drop from zero, gave a figure of 130 pg/ml. Both figures were in line with that obtained by Morgan et al. (1978). The intra-assay coefficients of variation are 14.9 % at 515 pg/ml, 6.2 % at 945 pg/ml and 5.0 7. at 1388 pg/ml (n =6). The inter-assay coefficients of variation are 20.9 % at 492 pg/ml, 9.0 % at 967 pg/ml and 8.8 7. at 1373 pg/ml (n = 6). These figures are in line with those of Morgan et al. (1978). Figure 2.3 shows that human, rat and mouse IR-GIP (pooled samples) run parallel in the assay. In chapter eight where serial blood sampling from mice was required it was neccessary to run the GIP assay with just 20 pi plasma samples. This was achieved by adding an extra 100 pi buffer to all tubes (except total counts) and adding 20 pi CSS or sample to the appropriate tubes. The standard values were thus increased by a factor of five. The limit of detection of the assay when calculated as 2 x SD from zero (n = 10) was found to be 392 pg/ml. The limit of detection of the assay, when taken as a 10 7. drop from zero, gave a figure of 692 pg/ml. These figures are in line with what would be expected from the standard assay. The intra-assay coefficients of variation are 38.9 7. at 711 pg/ml, 14.3 7. at 945 pg/ml and 8.2 7. at 4359 pg/ml (n = 10). The maximum binding of label in the presence of excess antiserum was 71.2 7.. Figure 2.4 shows that Scatchard analysis (Scatchard, 1949) of the antiserum gives a binding affinity of 3.23 x 10** L/mole. Again these values are in line with those obtained previously (Morgan et al., 1978). page 58 Table 2.1 GIP assay protocol Reagent (ul) Total NSB Zero Standards NSB Samples Counts Binding Samples Assay diluent 300 200 100 300 200 GIP-free 100 100 plasma Standard — — — 100 Sample 100 100 Antiserum 100 100 Incubate at 4°C for 24 hours 125 I-GIP 100 100 100 100 100 100 Incubate at 4°C for 24 hours NRS 50 50 50 50 50 DAR 50 50 50 50 50 Incubate (a) at 4°C for 24 hours or (b) for 4 hours. If (b) PEG 120 120 120 120 120 page 59 Figure 2. 1 A typical elution profile of radioiodinated (1 mCi) porcine GIP (5 pg) from a Sephadex G15 column, 10 p 1 aliquots of the 0.5 ml fractions were counted. The first (larger) peak is iodinated GIP the second (smaller) peak is free iodine. 60 z: 2 0 20 30 35 Fractions page 60 Figure 2.2 A typical IR-GIP standard curve, with porcine GIP standard and charcoal stripped serum. Total count tubes = 4500 cpm, non-specific binding = 3.8 (n = 6, mean + SD) . 30 ^ 2 0 (0 4-" c 3 P Ü *5 ■ + - » 0 H- •— , "O 1 10 o 0 12 3 4 GIP concentration (ng/ml) page 61 Figure 2.3 Standard curve demonstrating parallelism between porcine IR-GIP (n=6, mean + SD) and IR-GIP In human, rat and mouse plasma pools. % 30 o mouse a human « 20 + rat c 3 O Ü 75 4o-» H* u c 3 o “ 10 o'- 0 1 2 3 4 GIP concentration (ng/ml) page 62 Figure 2.4 Scatchard plot -for antlserum RIC34/IIIf. Slope = -3.23 X 10^‘^ L/mole, x-lncp. = 4.12 x moles/L, 1.5 1.0 M 1 - b/f 0.5 10 1 5 20 M/Lxio-n TOTAL GIP page 63 2. 2. 0 The Motilin Assay The IR-motilin (IRM) assay was set-up by myself. There were several reasons for setting-up this assay. Motilin may be a hormone of interest, as set out in chaprer 1.5. A quantity (1 mg) of synthetic porcine motilin (a gift from Prof. N. Yanaihara, Japan) was available. Also an important factor was that radioimmunoassay (RIA) was to be the main analytical tool in the experimental work, but the main RIA, that for IR-BIP, had already been set-up. Thus, this opportunity was taken to "work-up" a new assay, almost as an exercise in itself, but always in the knowledge that the IRM assay would form a very minor part in the research which was to follow. lodination Motilin was iodinated by a modification of Yanaihara ejb al. , (1980). In practice most of the conditions were identical to those of the GIP iodination, detailed above. 2 pg aliquots of synthetic porcine motilin (NY i.FS-111—58) were diluted in phosphate buffer (20 pi), and to this were added in quick succession chloramine-T (25 pg in 20 pi) and :^^“^I-sod ium iodide (1 mCi in 1 Op 1 ) . The reaction mixture was agitated for 15 seconds before being stopped with sodium metabisulphite (150 pg in 50 pi). The tracer was then chromatographed and stored as for the GIP tracer. Typically motilin iodination profiles were identical to those for GIP. page 64 An t ibody Two three month old half-lop rabbits (high antibody producing strain, Ranch Rabbits Ltd.) were each pre-primed with 20 pg unconjugated synthetic porcine motilin (NY^FS-III- 58) in 0.1 ml sterile water (Difco), 0.1 ml BCG (Difco) and 0.4 ml Marcol (mineral oil, Guildhay Antisera). They were injected intra-dermally (50 pi emulsion per site) on the upper back between the shoulder blades. This pre-priming procedure is not normally part of antiserum production, but I had been advised that in this situation it was desirable (Dr. J Dryburgh, personal communication). Two weeks later they were given a priming dose of 60 pg motilin conjugated to 120 pg bovine gamma-globulin (Koch-Light Laboratories) by the glutaraldehyde technique of Reichlin et a 1. (1968). To motilin in sterile water (0.125 ml) was added the bovine gamma-globulin in sterile water (0.125 ml), BGG (0.05 ml) and glutaraldehyde (0.05 ml at a 1:50 dilution). This was emulsified with Marcol (0.7 ml) and immediately injected using the same method as previously, into twenty sites on the mid backs of both animals. The animal producing the best antiserum was not boosted and after 5 months produced an antiserum with an initial titre of 1:15000. All the work below was carried out using this antiserum MF/R/PK2-Pr. Assay Standards (20 ng aliquots of synthetic porcine motilin) and charcoal stripped serum were prepared in an identical manner to those for GIP. Buffers and conditions for the motilin assay were essentially the same as for the GIP assay. However, aprotinin had been omitted from the assay buffer and the assay had a page 65 two day incubation with no pre-incubation with antibody, as this had not been found to increase binding. To a series of polystyrene LP3 tubes (Luckhams Ltd.) was added either 0.1 ml assay buffer, 0.1 ml motilin standard in the range 7.8 - 2000 pg/ml and 0.1 ml CSS or 0.2 ml buffer and 0.1 ml plasma sample. Next was added 0.1 ml antiserum at a 1:15 000 dilution. All tubes were set-up in duplicate (at least) and non-specific binding tubes were included as required. 0.1 ml tracer (5 000 - 10 000 cpm), was then added and the tubes were vortexed and incubated at 4®C for 48 h. The motilin assay protocol is shown in table 2.2. Phase separation, centrifugation, aspiration, counting and data processing were as for the GIP assay. A typical standard curve is shown in figure 2.5. Under these conditions the tracer has a specific activity of 386 nCi/ng. The antiserum when subjected to Scatchard analysis has an affinity constant of 1.0 x lO^i L/mole; see figure 2.6. The limit of detection of the assay when calculated as 2 X SD from zero (n = 12) has been found to be 10 pg/ml. The limit of detection of the assay when taken as a 10 % drop from zero gives a more modest figure of 42 pg/ml. The intra-assay coefficients of variation are 7.7 % at 120 pg/ml, 6.3 % at 400 pg/ml and 5.4 7. at 610 pg/ml (n = 7; all these values are within the physiological range). The inter-assay coefficients of variation are 11.7 % at 260 pg/ml and 14.6 % at 470 pg/ml (n =7). Figure 2.7 shows that human motilin runs parallel in this assay. Figure 2.8 shows that the antiserum, under the conditions of the assay, does not cross-react with glucagon, VIP, gastrin, GIP, C-peptide or insulin. page 66 In conclusion, the motilin assay described above is at least as rapid and sensitive as any other published motilin assay (Chey and Lee, 1980). It does not require an initial methanol extraction step (Tai and Chey, 1978). It requires only a two day incubation unlike that of Yanaihara et al. (1980) which is a four day disequilibrium assay. Finally, it is as sensitive as those of Bloom et al.. (1976) and Dryburgh et al. (1975). page 67 Table 2.2 Motilin assay protocol Reagent (>il) Total NSB Zero Standards Sample Sample Tube Tube NSB Day 1 Buffer 300 200 100 300 200 HFP 100 100 100 - - Standard — — - 100 - - Sample — — - - 100 100 1st Antlserum — — 100 100 - 100 125_ I-motllln 100 100 100 100 100 100 Incubate at 4°C for 48 hours. Day 3 NRS 50 50 50 50 50 DAR 50 50 50 50 50 12% PEG 120 120 120 120 120 Incubate at 4°C for 2—4 hours then centrifuge. aspirate and count. page 68 Figure 2.5 A typical IRM standard curve, with synthetic porcine motilin standard and charcoal stripped serum. Total count tubes = 8000 cpm, non-specific binding = 3.5 %. (n = 4, mean + SD). 60r 40 CO +-* c 30- D O Ü CO 4-* o H c20 o CD 10 0 1 Motilin concentration (ng/ml) page 69 Figure 2.6 Scatchard plot for antlserum MF/R/PK2-Pr. Slope = -1.0 X 1011 L/mole, x-incp. = 1.0 x 10"~ii mole/L, 0.8 0.6 b/f 0.4 0.2 2 M/Lx10-n Motili’n concentration page 70 Figure 2.7 Standard curve demonstrating parallelism between synthetic porcine motilin (n = 4, mean + SD) and IRM in a human plasma pool. 50 r 40 m 30 c D O Ü 03 •f-* 0 H 1 20 D O CO 10 0 ^ 1 Motilin concentration (ng/ml) page 71 Figure 2.8 Cross-reactivities of antiserum MF/R/PK2-Pr with some other gut hormones. 30 c-peptide 20 VIP gastrin b/t GIP insulin motilin 50 100 n g /m l Peptide concentration page 72 2. 3.0 Plasma Triglyceride Assay Plasma triglyceride concentrations were estimated using a manual enzymic colourimetric technique (H. Lange et a 1., unpublished), Peridochrom Triglycerides (Boehringer, London). The plasma sample (10 pi) was incubated with the reagent mixture (1 ml) for 30 minutes at 20 - 25*C and the absorbence read at 578 nm (Cecil 272 spectrophotometer) against a reagent blank within 900 minutes. The absorbence multiplied by a factor of 5.68 yielded the concentration of triglycerides within the sample in mM. The assay procedure was a multi—enzyme system utilizing the glycerol liberated on lipolysis of the triglyceride in a sequence of reactions which result in the formation of a chromogen via the reduction and subsequent oxidation of NAD. The reaction sequence is set out below. Triglyceride + 3 H.~>q (Lipase/Esterase) ------> glycerol + fatty acid glycerol + NAD^ (Glycerol Dehydrogenase) —------> dihydroxyacetone + NADH + NADH + MTT (Diaphorase) — ------> formazan + NAD"*" Formazan is the chromogen and MTT stands for 3-(4,5- d imethy-2-thiazoly1)-2,5-diphenhy tétrazolium bromide. page 73 2.4.0 Plasma glucose assay Plasma glucose concentrations were estimated using a semi-automated glucose oxidase technique (Stevens, 1971) using a Glucose Analyzer 2 (Beckman Ltd., Glenthroes). The sample (10 pi) was injected into aerated enzyme solution, and appropriate aeration was achieved by shaking the bottle containing the reagent several times with several changes of air. The method takes advantage of the fact that B-D-glucose from the sample reacts with dissolved oxygen in the solution by the action of glucose oxidase yielding gluconic acid and hydrogen peroxide. This reaction is set out below. Glucose + 2H-,o + 0^ (Glucose Oxidase) — ------> Gluconic acid + There is an equimolar consumption of glucose and oxygen. The change in oxygen concentration is monitored by an oxygen electrode located within the reaction cuvette giving a direct reading of the glucose concentration within the sample. The system is calibrated using a Beckman aqueous glucose standard (8.3 mM). page 74 CHAPTER THREE TWENTY-FOUR HOUR RHYTHMS IN THE ENTEROINSULAR AXIS page 75 3. 0. 0 It was long ago suggested that insulin injected at different times of day produces quantitatively different effects on circulating glucose levels (Forsegren et al., 1930). It has been shown that in normal subjects glucose tolerance was diminished and the plasma immunoreactive- insulin (IRI) response was delayed in the evening (Carroll and Nestel, 1972). A diurnal variation in insulin sensitivity has also been described in normal humans (Gibson and Jarrett, 1972; Gagliardiho et al., 1976) and in subjects with impaired glucose tolerance (Schulz et al., 1980). This chapter describes experiments concerning gastro- enteropancreatic (GEP) hormones. These were undertaken in order to shed some light on a previously neglected area, to see whether or not GEP hormones, and more specifically those involved in the entero-insular axis, show twentyfour hour rhythms in their response to a standard meal. In addition to plasma glucose and IRl, circulating levels of immunoreactive- gastric inhibitory polypeptide (IR-GIP) and immunoreactive- motilin (IRM) were measured. IR—GIP was measured as it is the chief incretin candidate. IRM was measured as it is a hormone which has been implicated in gastric emptying (Ruppin et al., 1975; Christofides, Bloom, Besterman, 1978; Christofides et al., 1981), bile flow and other factors influencing the rate of absorption of nutrients, hence affecting the magnitude and timing of post-prandial IR-GIP release. Furthermore, intravenous infusions of motilin have been shown to increase the initial rise in IRl and IR-GIP following an oral glucose load (Long et a 1. , 1982) via an effect on gastric emptying. page 76 3.1.0 GEP hormones during a twentyfour hour fast. Plasma concentrations of IR-GIP have been reported to have a diurnal rhythm (Jorde et al., 1980; Jones et al., 1985), but these variations appeared to be due solely to subjects eating meals during the day and fasting (i.e. sleeping) at night. Notice was given of preliminary results that indicated a diurnal rhythm in plasma IRM (Christofides, 1979b), but no data , save that IRM levels were higher in the morning than later in the day, were given at that time or since. IRM levels have been measured in just five subjects for 24 h on two occasions. However, the subjects were not fasted but fed meals of different compositions during the day (Shima et al., 1979), so that no clear picture of what happens to circulating IRM levels through 24 hours could be established. There is no published data on twenty-four hour variations in insulin or C-peptide release. However, an ultraradian variation has been described for the basal secretion of insulin and C-peptide, with a small amplitude and a period of 10 to 15 minutes (Lange et al., 1979). 3.1.1 Methods Twelve normal, healthy volunteers, four men and eight women aged 20 - 39 years (mean = 27.9 + 1.58 SEM), took part in the experiment. None of the group was under medication, except three of the women who were taking contraceptive steroids. Each volunteer ate a standard breakfast of a bowl of corn flakes, two slices of toast and two cups of tea or page 77 milk from 0710 h to 0730 h, followed at 0945 h by two jam doughnuts and two cups of tea or coffee (10% protein, 36% fat, 54% carbohydrate, 695 kilocalories total). Blood (10 ml) was sampled by indwelling venous catheter (antecubital vein) at hourly intervals for 24 hours at 1000 h. During the sampling period volunteers remained fasted but were allowed unlimited drinking water. During the day volunteers occupied themselves with light activities (reading, watching television etc.). During the night most of the volunteers slept between (and sometimes during) blood sampling. Blood was centrifuged immediately and plasma divided into aliquots for the individual assays, then frozen and stored at -20*C until assay. IRl estimations were performed by Dr. SM Hampton, using an antibody raised to pure porcine insulin, porcine standards and iodinated porcine insulin (Hampton, 1983). 3.1.2 Results The IR-GIP. IRM, glucose and IRl results are shown in figures 3.1 to 3.4 respectively. Areas under the curve and statistics for the IRM results are shown in table 3.1. In all the figures mean values are shown + standard errors of the mean (SEM). These results, with the exception of IRM results, but with the addition of melatonin, cortisol (free and bound) and immunoreactive-C-peptide appeared in Arendt et al. (1982). Owing to the large inter-personal variations in plasma IRM, all values were expressed as a percentage of the individual's own mean basal level from 1500 h to 0200 h (mean 314 pg/ml + i16 SEM, range 84 - 1436 pg/ml). This is a method page 78 which has been used by other workers to reduce raw IRM data to a form more amenable to meaningful statistical analysis (Shima et al., 1979). Areas under the curve were submitted to Student's t-test for paired observations. The results show that, following breakfast, IR-GIP concentrations were elevated and following the test meal were stimulated a little further and stayed elevated for a very long time, i.e. about 7 hours. Once IR-GIP levels returned to basal they remained there for the remainder of the time studied. Following the test meal IRM levels fell briefly, then rose and returned to an apparent "basal" level by about five hours following the test meal. They did not change significantly for about ten to eleven hours. Mean IRM levels fell to a nadir at 0400 h - 0500 h (151 + 40 and 172 + 48 pg/ml i.e. 62.0 and 65.5 % of basal at these time points respectively; p < 0.05). Basal levels were regained by 0900 h except for a brief peak at 0600 h. The mean area under the curve 0300 h - 0800 h was significantly lower than that for 2100 - 0200 h; p < 0.005. Glucose concentrations did not vary greatly over the 24 h period. IRI levels having been raised by the breakfast then rose still a little further after the test meal and returned to basal within about 3 hours of the test meal. IR-GIP levels at this point were still about twice basal levels. The decrease in circulating IRM concentrations coincided with a peak in melatonin release and the onset of a peak in plasma cortisol concentrations (Arendt et al., 1982). 3.1.3 Discusion The IR-GIP results are at least partly in agreement with those of Jorde et al. (1980). They found that in normal oaoe 79 individuals plasma IR-GIP levels remained elevated -from early morning until late at night, the subjects having eaten four "normal" meals during the day. IR-GIP levels took about 9 hours to return to basal levels after the last meal at 1930 h. The IR-GIP results from this experiment appeared to demonstrate that, following a meal, nutrients take several hours to be completely cleared from the stomach and small intestine, since it is the absorption of nutrients which causes the release of this hormone. No circulating venous glucose response was observed after the test meal at the time points studied. This is, perhaps, not surprising as IRI levels were already elevated from the breakfast eaten earlier. Neither plasma glucose nor IRI levels altered from basal in the remainder of the time studied in the experiment. So, although there may be an ultraradian rhythm in IRI release there would seem to be no overall circadian variation in its basal levels. In addition to Christofides' (1979b) preliminary results one other group of workers has monitored IRM levels through a twentyfour hour period, taking samples every 15 minutes (Shima et al., 1979). Five subjects were studied for 24 h on two occasions; the subjects were not fasted but fed meals of different compositions during the day. Their resuts, expressed as a percentage of the mean from 0045 h to 0815 h, showed that IRM levels continue to fluctuate rhythmically through the night with a period of approximately 90 minutes. Their results also showed (although they did not comment upon it) that on one occasion mean plasma IRM concentrations showed a marked fall from 0300 h to 0600 h. This fall would not appear to affect the migrating myoelectric complex (MMC) as this has been reported to continue with its approximately 90 minute cycle page 80 through the night (Finch et al., 1982). It is not clear what causes the fall in mean plasma IRM levels in the early morning. A clue as to whether the cause is neural or hormonal might come from further studies to investigate possible interactions with other hormones which have been suggested to share moti1 in's cyclical release and interaction with the MMC, such as pancreatic polypeptide (PP) (Janssens et al., 1R82) and gastrin (Peeters et al., 1980). Finally it is interesting to note that although IRI levels return to basal within 3 - 4 h of the meal, and therefore glucose is no longer being absorbed to any great extent, fat absorption and processing continues for a further few hours, as IR-GIP levels continue to remain elevated. It is possible, therefore, that in the first few hours after a mixed meal GIP's incretin activity is important, later to be superceded by its enterogastrone and/or possible lipid clearing activity. It might further be speculated, when considering the mean plasma IRM results, that the "dip" in mean values after the meal could be accounted for by glucose absorption and insulin release, and that the subsequent rise be due to continued fat absorption in the absence of elevated glucose or insulin concentrations. page 81 Tab le 5.1 Plasma IRM area under the curve data at two time periods following a test meal ant 0945 h, eaten within 15 minutes and subsequent fast. Subject Area 2100 h - 0200 h Area 0300 h - 0800 h (% basal.h> (% basal.h) A 590.8 478.9 B 567.2 402.7 C 740.5 360.8 D 654.1 471.5 E 706.2 453.6 F 555.6 433.2 G 626.0 461.2 H 602.5 854.4 I 696.4 468.3 J 704.3 358.5 K 610.0 338.6 t = 3.5556; p < 0.005 page 8: Figure 3.1 Plasma IR-GIP levels in healthy human volunteers following a test meal at 0 9 4 5 h, eaten within 15 minutes, and subsequent fast, (n = 6, mean + SEM). o o cn o o o m o o o o O LU CVJ o o CO a. o . I to o 3 o CL. cn o E cn C u J CO CVJ o page S3 Figure 3.2 Plasma IR| levels in healthy human volunteers following a test meal at 0 9 4 5 h, eaten within 15 minutes, and subsequent fast, (n = 6, mean + SEM). O O cr> o o o IT) o o o O L U o z C\J o o _J o o o o o CO a: LU o to o cr> —I o Q. Î K. J o o o o If) CO C\J o page 84 Figure 3.3 Plasma ; levels in healthy human volunteers following a test meal at 0945 h, eaten within 15 minutes, and subsequent fast. O o 1 CTl o o o cn o o o O L U o s: CM o o _J o o o CO LU ZD —I < LU o o C cn CO o m I— -J o o LO o o cn o page 85 Figure 3.4 Plasma^\ucc'S'Ç levels In healthy human volunteers following a test meal at 0 9 4 5 h, eaten within 15 minutes, and subsequent fast, (n = 6, mean + SEM). o o m o o o O L U o z: CsJ o o o o o o o CO LU (/) o o e> o g o (/) cn Lf) CO CVJ o page 86 3. 2.0 Function of the entero-insular axis at 1100 h and 0200 h. As no rhythms in basal plasma IR-BIP, glucose or IRI levels were observed in the previous experiment (3.1), this further experiment was carried out to see if any differences in stimulated IR-GIP, glucose or IRI could be observed at two different times of day. As Arendt et a1■ (1982) maintain, the pineal gland may have a role in the control of carbohydrate metabolism, and melatonin (an index of pineal function) most certainly does have a circadian rhythm (maximum concentration at 0200 h, and by 1000 h has reached a minimum concentration). It may also be of note that it is from about 0200 h that plasma IRM levels are seen to decline. 3.2.1 Methods Five normal healthy volunteers, four men and one woman, aged 23 - 28 years(mean = 25.4 +1.96 SEM) took part in the experiment. None of the group was under medication. At 1100 h after a 16 h overnight fast each volunteer consumed a standard meal consisting of two cups of coffee, one glass of milk, one bowl of cornflakes, two slices of toast, 25 g sucrose and two jam doughnuts (17 g protein, 37 g fat, 101 g carbohydrate, 779 kcal) within 20 minutes. For the next 15 h they again fasted, but were allowed free access to water. At 0200 h the following morning the volunteers again ate the same meal, within 20 minutes. Blood (10 ml) was sampled by indwelling venous catheter (antecubital vein) at hourly page 87 intervals for 24 h commencing at 1100 h. The remainder of the protocol was the same as that for the previous experiment (3.1). 3.2.2 Results The plasma IR-GIP, IRI and glucose results are shown in figures 3.5 to 3.7 respectively. Areas under the curve and statistics for the IR-GIP and IRI results are shown in tables 3.2 and 3.3 respectively. IRI estimations were performed by Dr. SM Hampton. The results show that the second meal gave a lower but broader IR-GIP response, having also a slightly greater area under the curve (p < 0.05). The glucose results were most striking. After the 1100 h meal there was practically no change in plasma glucose concentrations, but following the 0200 h meal glucose levels rose from about 4 mM to over 9 mM, and remained elevated for over 3 h after the meal. The meal at 0200 h gave a significantly greater IRl release (p < 0.005), the peak level being greater by 31 % and the response longer (about 50 %) when compared to the meal at 1100 h. 3.2.3 Discussion The glucose intolerance that was observed in the early morning following the mixed meal might be interpreted as insulin resistance. Although after the 0200 h meal the IRl response was greater, it had less effect on the lowering of plasma glucose concentrations than the IRl released following the 1100 h meal. However, it is not possible to tell, from this experiment, whether the augmented IRl release following page SB the 0200 h meal is due to an increased sensitivity of the pancreatic B-cell to GIP, to the hyperg1ycaemia, to a decreased clearance of IRI or to a combination of any of these factors. The duration and magnitude of the IR-GIP response following the two meals might be explained when considered together with the IRM results of the previous experiment (3.1). The reduction in the mean plasma IRM concentrations in the early morning might indicate an increase in the gastric emptying time, and this might help explain the slight blunting in the IR-GIP response. An increase in gastric emptying would result in a diminishing of the magnitude of the IR-GIP response, by increasing its duration. It may also be worth noting that although post-prandial IRI levels following the 0200 h meal are higher and elevated for a longer period, the IR-GIP levels do not appear to be much reduced by these higher levels. Although insulin does not inhibit glucose-stimulated IR-GIP release (Service et al., 1978) it does inhibit fat stimulated IR-GIP release (Brown ejt al., 1975), and the late phase IR-GIP release is probably almost entirely due to fat absorption. The results of this experiment, consistent with a degree of insulin resistance in the early morning, are interesting but inconclusive, since factors affecting gastric emptying cannot be excluded. The results of this experiment also beg the question of what happens at other times of day. page 89 Figure 3.5 Plasma IR-GIP levels in healthy human volunteers over 24 h, ■fasted from 1900 h the previous night, with two test meals o o o o o o C ÜJ o z: o CVJ o o o w oj s: CVJ I— I o o o 0 0 o o CD I q; (jj o o E o cn CO CVJ page 90 Figure 3.6 Plasma IRI levels in healthy human volunteers over 24 h, •fasted from 1900 h the previous night, with two test meals (eaten within 20 minutes) at 1100 h and 0200 h, and fasting following the meals, (n = 5, mean + SEM) o o o o o o o o CVJo ÜJ o CVJ •— < o o o o o 00 o o CO CO Q. o o o o o o VO CVJ o page 91 Figure 5.7 Plasma glucose levels in healthy human volunteers over 24 h, ■fasted from 1900 h the previous night, with two test meals (eaten within 20 minutes) at 1100 h and 0200 h, and fasting following the meals, (n = 5, mean + SEM) o o o o o CO o o o CM O o O UJ CVJ s: CVJ I— I o o o o o 00 o o UJ to o CD UJ o Q. o o E L _ CVJ CO CO CM page 9: Table 3.2 Plasma IR-GIP area under the curve data taken from figure 3.5 Subject Area 1045 h - 1845 h Area 0145 h - 0945 h pgml-ih pgml ~*-h L 9378.0 9467.5 M 10392.5 10842.0 N 5642.5 8000.5 P 8567.0 9854.5 S 6176.0 7629.0 Mean 8071.2 9158.7 S. D. 2065.8 1331.8 t = 2.3500; p < 0.05 page 93 Table 3.3 Plasma IRI area under the curve data taken from figure 3.6 Subject Area 1045 h - 1845 h Area 0145 h — 0945 h mUL“ ^h mUL~^h L 133 163 M 96 213 N 73 169 P lOl 218 S 133 197 Mean 107 192 S.D. 26 25 t = 5.0521; p < 0.005 page 94 3.3. 0 Investigation of the insulinaemic responses following intravenous glucose tolerance tests at 1000 h and 0200 h. This experiment follows on from experiment 3.2, and was designed to compare intravenous glucose tolerance in the early morning (0200 h) with that in the late morning (1000 h ) . This experiment seeks to investigate glucose tolerance with no direct intervention by the entero-insular axis. Thus any differences between the two time points cannot be attributed to the action of the entero-insular axis. Carroll and Nestel (1972) showed that in normal subjects glucose tolerance and plasma IRI response were diminished in the evening following oral glucose. The delay in plasma IRI response was significantly less, however, with intravenous glucose. 3.3.1 Methods Five normal healthy volunteers, three women and two men aged 22 to 33 years (mean = 27.4 + 1.96 SEM) took part in this experiment. None of the group was under medication. The volunteers fasted from 2000 h until 1000 h the following morning when, having had antecubital veins in both arms cannulated with intravenous catheters, two basal blood samples were taken (-15 and 0 minutes) from the left arm. Immediately upon the removal of the second sample, 25 g glucose (in 100 ml saline) was infused (over 2 - 3 minutes) into the right arm. Blood sampling continued from the left arm for 180 minutes. Subjects continued to fast until 0200 h page 95 the following morning when the above sampling and intravenous loading procedure was repeated. Subjects were allowed access to water at all times. Blood handling and the remainder of the protocol was the same as in experiment 3.1. 3.3.2 Results The IR-GIP, IRM, IRI, and glucose results are shown in figures 3.8 to 3.11 respectively, additional data for IR-GIP and IRM are shown in tables 3.4 to 3.7. The IRM results were expressed as a percentage of the basal value, where the basal value was calculated as the mean of the first two samples. Following the 1000 h i.v. glucose load a significant and rapid fall in IR-GIP concentrations, with a minimum of 68.4 % of basal at 5 minutes (P < 0.025) and remaining significantly lower for 45 minutes, was seen. N o such fall was observed following the 0200 h load. Plasma IRM levels following both 1000 h and 0200 h meals showed a very significant but slower decline than IR-GIP with minima of 62.6 % and 60.1 % of basal respectively. Plasma glucose levels reached slightly higher concentrations at 0200 h than at 1000 h and remained higher for longer, as did IRI, although to a lesser extent. 3.3.3 Discussion The lowering of plasma IR-GIP after the 1000 h load was most unexpected. Service et al. (1978) were unable to show any effect of hyperinsulinaemia induced by exogenous insulin on the basal secretion of IR-GIP in the absence of change in plasma glucose concentration, although Andersen et al. (1978) showed a diminution of the expected IR-GIP release following page 96 oral glucose in the presence of marked hyperinsulinaemia with mild or moderate hyperglycaemia. In this experiment, where marked and sustained hyperinsulinaemia was induced by hyperglycaemia, basal secretion of IR-GIP was inhibited following the 1000 h, but not the 0200 h, load. Thus, the results indicate that there is a possible feed-back inhibition of insulin on IR-GIP release, although the above data, when taken together with those in the literature (Service et al.. 1978; Andersen et al., 1978), suggest that a degree of tiyperg lycaemia must also be present for the Inhibition to occur. The apparent lack of a significant drop in plasma IR-GIP levels after the 0200 h load might be taken as yet another piece of evidence of insulin resistance at this time of day. The IRM results are in close agreement with those of Christofides, Bloom and Besterman (1978). They measured plasma IRM during an insulin stress test (0.02 lU/kg) and showed a gradual decrease in plasma IRM, reaching a nadir at 35 minutes (a fall of 58 %, p < 0.005). This result is very similar to the IRM results shown above. The authors claimed, however, with appropriate correlations, that the observed effect was due to a decrease in glucose concentrations. Obviously this cannot be the case in the above experiment. The only similarity between an insulin stress test and an intravenous glucose tolerance test is a sudden and rapid increase in plasma IRI concentrations. Therefore, the alternative and more probable explanation is that insulin has a negative feedback upon basal IRM release. It should be further noted that at 0200 h, following the i.v. glucose bolus, the time taken for IRM levels to reach a minimum level, similar in magnitude to that obtained following the page 97 0200 h load, was 60 minutes, whereas following the 1000 h load it took just 30 minutes. Could this be yet more evidence, albeit slight, of insulin resistance in the early morning? Finally, the plasma glucose results show a better glucose tolerance at 1000 h, and hence are strongly suggestive of insulin resistance at 0200 h, as the peak and subsequent glucose concentrations are consistently and significantly higher (at 15, 45 and 60 minutes) after the 0200 h compared to the 1000 h bolus. The relatively elevated glucose levels after the 0200 h bolus may be responsible for the slightly but significantly augmented IRI levels at 90 and 120 minutes post i.v. glucose. In conclusion, the plasma IR-GIP, glucose and IRM results when taken together are strongly suggestive of a degree of insulin resistance being present at 0200 h compared to 1000 h. This result is in agreement with that of the previous experiment (3.2), in confirming insulin resistance in the early morning, and excludes differences in gastric emptying as a possible alternative explanation. page 98 Figure 5.8 Mean plasma IR-GIP levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose In saline, in 2 - 3 minutes) at 1000 h (•) and 0200 h (o) (n = 5). ^00 fpg/ml PLASMA IR-GIP 500 400 300 200 BOLUS 100 L « « « «_____ I. .— .I. -15 0 15 30 45 60 90 120 150 180 TIME (h) page 99 Table 5.4 Plasma IR-GIP levels In healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 1000 h (n = 5). Significant differences shown are from zero time point (Student's paired t-test) Time (min) Mean (pg/ml) S.D. (pg/ml) Sig (p<) zero 418 161 5 286 172 0.025 10 326 154 0.010 15 355 154 0.050 30 307 131 0.025 45 361 170 0.050 60 384 173 n.s. 90 499 286 n.s. 120 378 187 n.s. 150 438 188 n.s. 180 423 157 n.s. page 100 Table 3.5 Plasma IR-GIP levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 0200 h (n = 5). Significant differences shown are from zero time point (Student's paired t-test) Time (min)______Mean (pg/ml) S.D. (pg/ml) Sig (p<) zero 349 214 5 323 195 n.s. 10 309 162 n. s. 15 285 163 n.s. 30 359 172 n.s. 45 350 88 n.s. 60 333 120 n.s. 90 316 213 n. s. 120 334 135 n . s. 150 284 215 n.s. 180 274 96 n. s. page 101 Figure 5.9 Mean plasma IRM levels in healthy human volunteers following an Intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 “ 3 minutes) at 1000 h (•> and 0200 h (o) (n = 5). Values are shown as a percentage of the mean zero concentration. 120 T % BASAL PLASMA IRM 100 i.v. BOLUS I I ... I .1 I -15 0 15 30 45 60 90 120 150 180 TIME (h) page 102 Table 3.6 Plasma IRM levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 1000 h (n = 5)- Values are shown as a percentage of the mean zero concentration. Significant differences shown are from zero time point (Student's paired t-test) Time (min) Mean (%) S.D. (7.) Sig (p<) zero 100.0 25.0 5 88.7 12.9 0.050 10 93.6 20.8 n.s. 15 84.5 19.2 n.s. 30 62.6 16. 1 0.0025 45 65.5 17. 1 0.005 60 79.6 23.3 O. 050 90 95.3 29.5 n.s. 120 76.7 16.3 O. 010 150 95.6 39.4 n.s. 180 104.5 25.5 n.s. page 103 Table 3.7 Plasma IRM levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 0200 h (n = 5). Values are shown as a percentage of the mean zero concentration. Significant differences shown are from zero time point (Student's paired t-test) Time (min)_____Mean (%)______S.D. (%)______Sig (p<) zero 100.0 29.5 - 5 98.7 40.4 n.s. 10 85.1 25.3 n.s. 15 80.7 24.4 n.s. 30 72.7 13.5 0.010 45 61.7 15.1 0.0025 60 60.1 12.8 0.0025 90 66.9 18.5 0.010 120 74.4 37.1 n.s. 150 63.9 25.6 0.025 180 56.4 6.3 0.0005 page 104 Figure 3.10 Plasma IRI levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 1000 h (•) and 0200 h (o) (n = 5, mean + SEM). mU/L PLASMA INSULIN 60 50 40 30 20 10 0 -15 0 15 30 45 60 90 120 150 180 TIME (min) page 105 Figure 3.11 Plasma glucose levels in healthy human volunteers following an intravenous glucose bolus (100 ml, 25 % glucose in saline, in 2 - 3 minutes) at 1000 h and 0200 h (o) (n = 5, mean + SEM). mM PLASMA GLUCOSE 12 9— f -15 0 15 30 45 60 90 120 150 180 TIME (min) page 106 3.4.0 Function of the entero-insular axis throughout the day. This experiment follows on from the previous experiments in this chapter. It has been shown that although the same oral and intravenous stimuli to IRI elicit a greater IRI response in the early morning, the glucose clearing ability of the IRI is impaired. However, the difference was more marked following oral stimuli and possible changes in gastric emptying were shown to be not the only possible reason for this difference. This experiment seeks to go one step further than experiment 3.2 and follow the function of the entero-insular axis at 4-hourly intervals throughout the day. 3.4.1 Methods Six normal healthy volunteers, three men and three women, aged 23 - 43 years (mean = 26.2 + 1.22 SEM) took part in the experiment. None of the group was under medication. The volunteers fasted for at least 10 h. They were fed a test meal of two cups of coffee, one glass of milk, one bowl of cornflakes, two slices of toast, 25 g sucrose and two jam doughnuts (17.1 g protein, 36.7 g fat, 100.9 g carbohydrate, 7799 kcal). The test meal was consumed within 20 minutes. This was done on six separate occasions over a six week period giving six time points of 1000 h, 1400 h, 1800 h, 2200 h, 0200 h and 0600 h. Blood samples were taken for 6 h following the meal. Blood sampling, handling and storage were the same as for the other experiments in this chapter. page 107 3.4.2 Results The plasma IRI, glucose and IR-GIP concentrations were measured for each individual at each time point for each occasion and the mean + SEM values are shown in figures 3.12 to 3.14 respectively. The area above the curve (increment above basal for the hormones) for each individual on each occasion was calculated. These results were then combined to give six mean areas under the curve at the different time points for glucose, IRI and IR-GIP shown in figure 3.15. Mean fasting plasma glucose, IRI and IR-GIP are shown in figure 3.16. Student's t-test for paired observations was applied to the resultant areas and means. No distinctive diurnal patterns were found in basal plasma concentrations of IRI or glucose, although basal concentrations of IRI and glucose were lower at 1400 h compared with 1000 h (p < 0.02). The integrated glucose response was less at 1000 h compared with all other times and increased during the day to a maximum at 2200 h. The integrated IRI results also show a minimum at 1000 h and a maximum at 2200 h. The greatest integrated IR-GIP response following the meal was at 1400 h, and the smallest at 2200 h. However, the integrated IR-GIP response at 1000 h was still greater than at 2200 h (p < 0.02). A large variation in basal IR-GIP concentrations was seen through the day. The lowest basal value was seen at 1400 h and the highest at 2200 h. 3.4.3 Discussion There is clearly a rhythm in the integrated glucose response following the test meal through the day. The page 108 smallest area under the curve is seen at the 1000 h time point, indicating the best glycaemic control; this coincides with the lowest IRI integrated area (i.e. the IRI is most effective in lowering glucose levels), together with a fairly low IR-GIP integrated area. The glucose results indicate that glycaemic control deteriorates through the day to a minimum at 2200 h. However, the increase in the amount of insulin secreted to remove the glucose from the circulation following the meal is roughly in step with the increase in post prandial glucose concentrations. This suggests that although more insulin is released through the day it is becoming less effective at lowering plasma glucose concentrations (losing its hypoglycaemic effect); insulin resistance appears to increase through the day. The interpretation of the IR-GIP results is a little less clear. Integrated increment over basal IR-GIP results appear roughly to mirror the glucose and IRI results in that they are fairly high in the late morning/early afternoon and low in the evening/early morning. Some of the apparent differences in IR-GIP results might, however, be attributable to the large variation in basal IR-GIP values observed. There are several complicating factors which affect the interpretation of the IR-GIP results. The first of these, already touched on, is the variation in basal IR-GIP concentrations through the day. If one considers the time point with the lowest IR-GIP concentration, 1400 h, and the highest, 2200 h, with the benefit of hindsight it is not too dificult to postulate an explanation for the difference. All the volunteers were instructed to fast for a minimum of 10 h, so that for the experiment with the test meal at 1400 h the volunteers should have fasted from 0400 h. However, none of page 109 the volunteers got up "in the middle of the nighf'to eat a meal immediately before 0400 h. Rather, they all had a late dinner, and some a snack at about 2300 h - 2400 h. One can see from the results that the late evening stimulated IR-GIP results are among the lowest observed through the day. When combined with at least a 14 h fast, by 1400 h basal TR-GIP values were very low. Then consider the basal IR-GIP values at the 2200 h time point; the volunteers ate an early, large lunch and fasted for just the minimum 10 h. If one looks at the stimulated IR-GIP values at around midday, they are considerably higher than those at around midnight, so that if circulating IR-GIP levels were going to "linger" on the / higher side of mean basal this is the time point at which it would occur. The above may also explain the observed slight variations in basal glucose and IRI levels. Should the experiment be redesigned and repeated, in the light of these results and assumptions, the volunteers should be instructed to fast for rather longer (at least 14 h), having just eaten a standard meal. However, in defence of the chosen experimental design, the preliminary experiments 3.1 and 3.2 indicated that 10 h would be sufficient to reduce basal IR-GIP values to a minimum, but the test meals used in these experiments were rather smaller than the type of meals that the volunteers would have consumed before a 10 h fast. Also possible effects of variations in gastric emptying can not be totally excluded from contributing to the variations in basal values. Finally, the above results strongly suggest a circadian rhythm in IR-GIP release following a mixed meal. But, any possible doubts about the interpretation of the IR-GIP data cannot diminish the strong evidence that the integrated page 110 glucose concentrations following a mixed meal increase through the day despite increasing IRI concentrations. Thus following mixed meals there is a circadian rhythm in insulin resistance. page 111 Figure 3.12 Variation in plasma IR-GIP response to a test meal given at 6 different times of the day over 24 h. Meal eaten within 20 minutes from O h . (n = 6, mean + SEM). 3 pg/ml 1000 h 1400 h 2 0 3 1800 h 2200 h 2 0 3 0200 h 0600 h 2 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 page 112 Figure 3.15 Variation in plasma IRI response to a test meal given at 6 different times of the day over 24 h. Meal eaten within 20 minutes from O h. (n = 6, mean SEM) • raU/L 1000 h 1400 h 90 1800 h 2200 h 70 50 30 10 90 0200 h 0600 h 70 50 30 10 0 1 2 3 4 5 6 0 1 2 3 4 5 6 h page 113 Figure 3.14 Variation in plasma glucose response to a test meal given at 6 different times of the day over 24 h. Meal eaten within 20 minutes from O h . (n = 6, mean + SEM). mM 1000 h 1400 h 1800 h 2200 h 0200 h 0600 h 0 } 2 3 4 5 6 0 1 2 3 4 5 6 h page 114 Figure 3.15 Integrated (6 h, area under the curve of increment above basal) response to glucose, IRl and IR-GIP to a test meal at times indicated (n = 6, mean + SEM). Marked significant differences are from 1000 h for glucose and IRl but 1400 h for IR-GIP. p < 0. 1 ** p < 0.05 12 mM.h GLUCOSE **** 8 ***** p 4 ** 0 3 mU/L.h X 100 PLASMA IR I 2 0 12 pg/ml.'h X 1000 PLASMA IR-GIP 8 4 0 1000 1400 1800 2200 0200 0600 TEST MEAL TIME (CLOCK TIME, h) page 115 Figure 3.16 Mean 10 h fasting plasma levels of glucose, IRI and IR--GIF normal human volunteers. (n = 6, mean + SEM). ** P < 0. 05 *** P < 0. 02 *** P < 0.01 7 r mM PLASMA GLUCOSE *** 0 10 fmU/L PLASMA IR I 8 - , 6 4 2 0 10 p g /tîil X 1000 PLASMA IR-GIP ** 1000 1400 18000 2200 0200 0600 TEST MEAL TIME (CLOCK TIME, h) page 116 SUMMARY OF RESULTS Following a mixed meal at 0945 and 24 h fast 1) plasma IR-GIP levels return to basal within 9 h and there is no circadian rhythm in basal levels. 2) no plasma glucose response or circadian rhythm is observed. 3) plasma IRI levels return to basal within 3 - 4 h and there is no circadian rhythm in basal levels. 4) plasma IRM levels fall and return to a mean "basal" within 5 h but later fall to a nadir at 0400 h - 0500 h returning to "basal" by 0900 h, exhibiting a circadian rhythm. 5) the late phase IR-GIP release is unlikely to be due to carbohydrate absorption but to fat absorption. Following a mixed meal at 1100 h then fasting until a further mixed meal at 0200 h 1) the meal at 0200 h gave a lower but broader peak in plasma IR-GIP, but with a slightly larger area under the curve, than the 1100 h meal. 2) the meal at 0200 h gave a large rise in plasma glucose; none was observed after the 1100 h meal. 3) the meal at 0200 h gave a larger and longer plasma IRI release than that following the 1100 h meal. Following i.v. glucose tolerance tests at 1000 h and 0200 h 1) basal plasma IR-GIP concentrations are only reduced following the 1000 h load. page 117 2) basal IRM concentrations are reduced more rapidly but to the same extent following the 1000 h load compared to the 0200 h load. 3) both plasma glucose and IRI reach slightly higher levels and are elevated for longer following the 0200 h load compared to the 1000 h load. Following a mixed meal at six times through the day 1) circadian rhythms in stimulated plasma glucose, IRI and IR-GIP were demonstrated. page 118 CHAPTER FOUR EFFECT OF DIETARY FAT ON THE ENTERO-INSULAR AXIS. page 119 4.0.0 Introduction It is now more than twenty years since the work of McIntyre et al. (1964) which re-awakened research interest in Incretin. They established that glucose administered intrajejunally, on a molar basis, resulted in an insulin release which was more than double that caused by intravenous glucose. Evidence for gastric inhibitory polypeptide's (GIF) candidature for incretin is discussed in chapter 1. Obesity in man is characterised biochemically by basal hyperinsulinaemia and an exaggerated insulinaemic response to glucose and most other insulinotropic stimuli (Karam et al., 1963). This cause and effect relationship between obesity and an overactive entero-insular axis is still vague and the actual cause of the hyperinsulinaemia controversial; however, it seems increasingly likely that an overactive entero-insular axis is involved to some extent. Obese subjects have been shown to have exaggerated IR-GIP responses after a mixed meal (Ebert et al., 1976) and exogenous insulin fails to inhibit fat stimulated IR-BIP release in these subjects (Creutzfeldt et al., 1978). In this context an overactive entero-insular axis would be defined as one where a normal enteric stimulus causes a supranormal release of incretin, thus causing the release of supranormal amounts of insulin. Quite naturally, although very little considered in the past, the nutrients and form of diet that an individual gut has biochemically (and perhaps to some extent, physically) adapted to optimally process will affect the associated gut page 120 hormone responses to a given meal. One again returns to the question of whether the overactive enteroinsular axis is a symptom of insulin resistance, or one of the causes. Insulin resistance is where normal (or supra-normal) concentrations of insulin produce a subnormal biological response at a target tissue. It is not unreasonable, however, to expect different tissues to display varying levels of resistance. Thus, Ni 11ms et al. (1978) showed that five days of calorie restriction abolished the exaggerated IR-GIP response to a fat or mixed meal in obese subjects, from which one is led to infer that previous high calorie intake might have been responsible for the observed excessive IR-GIP production and that diet rather than actual body weight is the significant factor. It has long been known that dietary composition influences glucose tolerance in healthy individuals (Sweeny, 1927). It was subsequently shown that a high fat diet reduced glucose tolerance and that this reduction was due to changes in susceptabi1ity to insulin (Himsworth, 1933 and 1934). More recently it has been shown that short term feeding of rats on a high fat diet both in vitro and in vivo resulted in an insulin resistant state (Zaragoza-Hermans and Felber, 1970; Susini and Lavau, 1978; Lavau et al., 1979). Reduced insulin receptor numbers (without change in affinity) have been demonstrated in the soleus muscle of rats fed a high fat diet for just ten days (Grundleger and Thenen, 1982). Other workers have found a greater than 40% decrease in insulin binding and receptor numbers, but no change in the affinity constant of insulin receptors in isolated fat cells or liver plasma membranes obtained from rats fed on a high fat, carbohydrate free diet (Ip et al., 1976; Sun et al., oaae 121 1977). Reports in the literature show basal plasma insulin levels in rats maintained on a high fat diet to be unchanged (Lavau et al., 1979), moderately decreased (Zarazoya-Hermans and Felber, 1970) or increased (Grundleger and Thenen, 1982). The apparent differences may possibly be explained by differences in age, rat strain, type of diet and length of time maintained on the diet. paqe 122 4.1.0 Effect of a short-term high fat diet on the entero- insular axis in rats. This series of experiments grew from experiments which were aimed at extending the work of Dryburgh et al (1980) who found that in rats exogenous C-peptide could inhibit the release of IR-GIP following an introduodenal fat load.In an attempt to repeat the experiment of Dryburgh et al. (but with a higher stimulated IR-GIP concentration from which it was hoped a more significant inhibition would arise), rats were pretreated with triolein for four days prior to the experiment. This time, however, endogenous C-peptide or insulin did not inhibit IR-GIP release (Hampton, 1983). Thus the following experiments in rats and in humans arose. 4.1.1 Methods 180 male Wistar Albino rats (240 - 290 g ) were used for the three parts of this experiment. They were randomly divided into two equal groups. Group I (fat pretreated/high fat diet) were dosed daily with 3 ml triolein (BDH, Technical grade) for four days prior to the experiment; they were also allowed free access to normal laboratoy food (Rodent Breeding Diet No. 1, Spratt's Laboratory Services, Barking, Essex). This resulted in a diet with a final fat content of 21%. Group II (normal diet/low fat diet) were only fed the normal laboratory food; the resultant diet contained Just 3.5% fat. The animals were weighed daily and fasted for 24 hours before each experiment. Three experimental protocols were performed page 123 on animals from groups I and II. They were all bled by cardiac puncture 10 minutes after an intraperitoneal injection of 0.4 ml/rat sodium pentobarbitone (60 mg/ml: Sagital; May and Baker, Dagenham). Blood was collected into lithium heparin tubes and centrifuged, and plasma was stored at -20*C until analysis for glucose and hormone concentrations. a) Time course of responses to oral fat. Rats from groups I and II were each divided into five sets (six rats per set). Set A from each group received no triolein. Animals in the other four sets (B-E) each received 1 ml triolein (Skcal) orally. Blood was collected from set A at O min and from sets B-E, after oral dosing, at 30, 60, 120, and 180 min respectively. b ) Effect of fat pretreatment on responses to oral fat with or without intraperitoneal insulin. Animals from groups I and II were each divided into 3 sets (lO rats per set). They were given the following: set A, 1 ml triolein orally 2 h before i.p. insulin (1 U/kg body weight); set B, 1ml triolein orally 2 h before i.p. saline (0.154 mol/L); set C , 1 ml saline (0.154 mol/L) orally 2 h before i.p. saline (0.154 mol/L). Blood was collected 20 minutes after the i.p. injection in all cases. c) Effect of fat pretreatment on the response to oral glucose. Rats from both groups I and II were divided into 2 sets (lO rats per set). Set A was given 3 ml glucose solution (2.78 mol/L; 5.8 Kcal) orally, and set B 3 ml saline (0.154 mol/L) orally. Blood was collected 60 min after oral dosing. Samples were assayed for IR-insulin, IR-GIP, glucose and triglycerides. The insulin samples were assayed by Dr. S. page 124 Hampton using an antiserum raised against bovine insulin, iodinated bovine insulin and a rat insulin standard, with a sensitivity of 0.5 pg/L (Hampton, 1983). Results were compared using Student's t-test for unpaired data and analysis of variance. 4.1.2 Results Throughout the study unpretreated control rats (Group II) were slightly, but significantly heavier than the fat pretreated rats (Group I). There was, however, no significant weight gain in either group of rats on completion of the fat-pretreatment (control rats 297.0 + 4.14 g before pretreatment versus 297.5 + 4.13 g after pretreatment, p = n.s.; mean + SEM, n = 60). There were no significant differences in fasting plasma glucose, IRI or IR-GIP between the two groups on completion of the fat pretreatment. a) Time course of the réponse to oral fat. Mean IR-GIP levels were consistently higher in animals in the fat pretreated set, but the differences did not reach significance at any one time point (see figure 4.1). However, when the data was subjected to analysis of variance (two factor with replication) the two groups were found to differ significantly (p<0.05). Plasma triglyceride values are shown in figure 4.2. Concentrations were similar in both groups at the outset of the experiment, but were significantly higher in the fat-pretreated rats 120 and 180 min after oral fat. Basal plasma triglyceride concentrations were 0.56 + 0.04 and 0.42 + 0.04 mmol/L in the control and fat pretreated rats respectively. At 120 min plasma triglyceride levels were 2.08 page 125 ± 0.15 in the control group versus 4.91 + 0.84 mmol/L in the fat pretreated group, p<0.01; at 180 min they were 2.39 + 0.41 and 7.10 + 0.46 mmol/L respectively, p<0.01 (mean + SEM, n=6). Only plasma from fat pretreated animals showed a marked lipaemia following the oral fat. b) Effect of fat pretreatment on the response to oral fat with or without exogenous insulin. Figure 4.3 shows that there was a significantly smaller rise in triolein-stimulated IR-GIP release following i.p. insulin rather than saline in unpretreated rats (663 + 49 versus 853 + 93 pg/ml, p<0.025). No such difference was observed in the fat-pretreated animals showing similar fat-stimulated IR-GIP concentrations, whether they were given 1.p. insulin (1008 + 95 pg/ml) or saline (1116 + 100 pg/ml). In addition circulating plasma glucose levels in rats given i.p. insulin after oral fat fell significantly less in the fat-pretreated animals than in the fat pretreated controls (2.6 + 0.48 versus 1.57 + 0.15 mmol/L, p<0.05); this occured despite the fact that the plasma IRI levels 20 min after the the insulin injection were similar in the two groups (4.1 + 1.7 versus 4.7 + 1.2 pg/L, p = n.s.). c) Effect of fat-pretreatment on the response to oral glucose. Figure 4.4 shows that 1 h after oral glucose, the fat pretreated animals showed significantly higher plasma glucose levels (10.2 + 0.39 versus 8.9 + 0.41 mmol/L, p<0.025) and IRI levels (6.2 + 1.2 versus 2.5 + 0.59 pg/L, p<0.01) than the untreated controls. IR-GIP levels showed a tendency (p<0.10) to be higher in the fat pretreated group compared to their untreated controls. No such differences were observed in the unpretreated control group. page 126 Figure 4.1 Plasma IR-GIP concentrations following oral triolein (1 ml) in fat pretreated (o) and untreated control (•) rats, (n = 6, mean + SEM). O 00 O CM C 1 0) E H o (D i i page 127 Figure 4.2 Plasma triglyceride concentrations following oral triolein (1 ml) in fat pretreated (o) and untreated control (•) rats, (n = 6, mean + SEM). (* = p < 0.01). KH C 1 0) E >» page 128 Figure 4.3 Plasma glucose, IRI and IR-GIP concentrations following oral triolein (1 ml) and intraperitoneal insulin (1 U/kg body weight, 2 h after triolein) in fat-pretreated (non-hatched) and untreated control (hatched) rats, (n = 6, mean + SEM). (* = p < 0.05, ** = p < 0.025). 3 to o> Z 3. E Ç ♦ IIUJ page 129 Figure 4.4 Plasma glucose, IRI and IR-GIP concentrations following oral glucose (5 g/kg body weight) in fat-pretreated (non-hatched) and untreated control (hatched) rats. (n = 6, mean + SEM). (* = p < 0.025, ** = p < 0.01). (O s i lU ÏÎ (O page 130 4.1.3 Discussion The diet of the rats can be controlled to within fairly narrow limits. In our laboratory, rats' normal fat intake is approximately 3.5%, which is very low compared to the average western diet (about 40% fat). Oral dosing of rats using the above regimen yields a daily fat intake of 21%. The imposition of a high fat diet for four days resulted in significantly increased circulating triglyceride and IR-GIP levels following oral triolein. The increased triglyceride levels observed in the fat-pretreated group are almost certainly due to an increased rate of absorption of fat from the small Intestine. It has been shown that pancreatic lipase activity in the rat can be greatly augmented by a short high fat diet (Deschodt-Lanckman et al., 1971). These workers found, however, that saturated triglycerides were a very poor stimulus to pancreatic lipase; this is presumably due to the fact that at body temperatures fats like tricaprylin and tristearin are still solid and hence are poorly absorbed. Unsaturated triglycerides such as triolein (liquid at body temperatures) did increase pancreatic lipase by about three fold in just five days of adaptation. IR-GIP secretion relies on the absorption of nutrients across the small intestinal mucosa rather than their mere presence in the lumen (Sykes e^ al., 1980; Ross and Shaffer, 1981). An increase in lipase activity in the the fat pretreated group would account for the increased rate of triglyceride absorption seen, and hence augment the IR-GIP response to oral fat in these animals. The opposite of the above is also true; a reduction of nutrient absorption rate in certain diseases, for example, untreated coeliac disease, or severe malabsorption for any other cause, oaae 131 results In a reduced IR-GIP response to enteric stimuli (Creutzfeldt et al., 1976; Besterman et al., 1978). The addition of the unabsorbable carbohydrate guar slows down the rate of absorption and hence depresses the IR-GIP response (Morgan et al., 1979). An exaggerated IR-GIP response has been reported in obese subjects given a high calorie mixed meal (Creutzfeldt et al., 1978). The exaggerated response was lost after five days of dietary restriction, although weight loss was minimal. The above experimental results suggest that previous diet composition is more important than body weight in determining the magnitude of the IR-GIP response. Many workers have suggested the existence of a negative feedback effect of insulin upon IR-GIP release, and many groups have demonstrated that intravenous infusions of either insulin or glucose reduce the rise in fat stimulated plasma IR-GIP release (Brown et al., 1975; Crockett et al., 1976; Cleator and Gourlay, 1975; Verdonk et al., 1980). The above results show that the control group of rats on a low fat diet demonstrated an inhibition of fat stimulated IR-GIP release by insulin but that the fat pretreated group did not. This is in agreement with the results of Creutzfeldt et al (1978) where a similar defect in feedback of insulin in IR-GIP release was demonstrated in human subjects. Regrettably, this was not investigated further by this group in their study of the effects of dietary restriction on IR-GIP release in obese subjects (Willms et al., 1978). A decreased responsiveness of the GIP-secreting cells to insulin is probably responsible for the abolition of the negative feed-back control of insulin on fat stimulated IR-GIP release in the fat-pretreated rats. This phenomenon page 132 may be due to a reduction in insulin receptor numbers, since a study of insulin receptors on human monocytes has shown that these can be altered by dietary changes (Pedersen e^ al., 1980). It was subsequently shown that these changes could occur within 24 h (Schulter et al., 1980). The above results show that just four days of a high fat diet was sufficient to make the rats resistant to the hypoglycaemic action of exogenous insulin. Circulating IRI and glucose levels were also significantly higher after oral glucose in the fat pretreated animals, further supporting the idea that fat-pretreatment predisposes the rat to insulin resistance. One must be careful in extrapolating the above results to the human; as has already been stated, the rats' normal diet is composed of just 3.5% fat, and the fat-pretreated group received 21% fat in their diet, but this is still far short of the average western human dietary fat content of 40%. However, the above results do show that diet, and in particular fat, can be most important in the development of insulin resistance and hence changes in the entero-insular axis. page 133 4.2.0 Effect of both high and low fat diets on the entero-insular axis in man. This group of experiments follows on directly from the previous rat experiments. They seek to answer the question of whether insulin is able to inhibit fat stimulated IR-GIP release in humans when maintained on short term low or high fat diets. The low fat and high fat diets were not isocaloric; body weights of subjects were carefully maintained throughout the experimental period. The IR-GIP response to oral fat and glucose was also investigated on the high and low fat diets. 4.2.1 Methods Nine, normal, healthy volunteers, six men and three women, aged 23 - 31 years (mean = 28 + 1.9 SEM), took part in the experiment. All volunteers were within 20% of their ideal body weight (Metropolitan Life Assurance tables) and informed consent was obtained from all. The volunteers showed a large variation in their "normal" daily ditary fat intake: range 12 - 157 g (99 + 20.1 g>. The percentage of energy thus derived varied 20 - 50% (38 + 2.5%). The volunteers were split into two groups, the first of five and the second of four. One of the volunteers in the first group had ultimately to be eliminated when it was realised that their "normal" diet had a fat content higher than the high-fat diet of the experiment. The first phase of the experiment was carried out on five subjects (3 males, 2 page 134 females), who ate a low fat diet (<30 g/day) for nine days. The volunteers resumed their normal diets for fifteen days, followed by a further nine day period when they ate a high fat diet. The high fat diet was achieved by supplementing the low fat regimen with 250 ml double cream, daily. In this way their fat intake was maintained at about 150 g/day. The second phase of the experiment was carried out on the remaining four volunteers. Their normal dietary fat intake was checked to insure that it was below 130 g/day. They were then placed on an inverted dietary regimen to that of the first phase, i.e. high fat diet (9 days), normal diet (15 days), then finally low fat diet (9 days). All volunteers were weighed throughout the study. Each volunteer underwent three investigations durng the course of the experiment. a) Effect of high and low fat diets on responses to oral fat. On day nine of the dietary regimen, after an overnight fast, two of the volunteers from the first group (one male and one female) following the high fat diet only, and all of the volunteers from the second group after both the low fat and high fat diets, "drank" 200ml of double cream (96g fat). Two basal blood samples were taken (at -10 and 0 minutes) prior to the oral load. Blood samples were then taken at 30 minute intervals from 30 to 210 minutes. b) Effect of high and low fat diets on responses to oral fat with exogenous insulin. Following an overnight fast, on the seventh day of both high and low fat diets, all the volunteers were given an oral fat load followed by an intravenous bolus of insulin. Two basal blood samples (at -10 and 0 minutes) were taken prior to the ingestion of 200 ml double cream (96 g fat). Following the oral load, blood samples were taken at 30, 50 and 60 minutes. Immediately after the 60 minute blood sample, insulin (0.2 U/kg body weight) was administered intravenously (via the catheter - group one; into the other arm - group two). Further blood samples were taken at 30 minute intervals from 90 to 210 minutes after the start of the test. c) Effect of high and low fat diets on responses to oral glucose. After an overnight fast, on day nine of both high and low fat diets the members of the first group of volunteers (n=5) were given an oral glucose tolerance test. Basal blood samples were taken (-10 and O minutes) then each volunteer ingested the equivalent of 75 g glucose (Hycal) in 250 ml of water. Blood samples were taken at 30 minute intervals from 30 to 210 minutes after the oral load. As with previous human experiments, blood samples were collected via indwelling venous catheter from the antecubital vein and kept patent with 0.12 M sodium citrate solution. The samples were collected into lithium heparin and fluoride/oxalate (for glucose analysis) tubes. Samples were immediately centrifuged and aliquoted before being frozen and kept at -20"C until analysis. Samples for insulin were assayed by Dr. S. Hampton using an antiserum raised against bovine insulin, iodinated bovine insulin, and a bovine insulin standard (Hampton, 1983). LSI's were performed at the Clinical Biochemistry laboratory of St. Lukes Hospital, Guildford. Results were compared using Student's t-test for paired and unpaired data, as appropriate. 4.2.2 Results Table 4.1 shows that there was no significant weight change among the volunteers during the experiment. Basal triglyceride, IRI and IR-GIP were not significantly different on the two dietary regimens. However, basal LSI levels were elevated after the high fat diet compared to the low fat diet (29 + 2.9 versus 19 + 1.4 LSI; mean + SEM: p<0.025). The dietary intake of the subjects is shown in table 4.2. a) Effect of high and low fat diets on responses to oral fat. Figure 4.5 shows that plasma IR-GIP values after oral fat were not significantly different at any time point, or even with respect to the area under the curve (AUC) data, following the two dietary regimen's. Plasma IR-GIP AUC following high fat diet was 4235 + 715 versus a low fat diet value of 5490 + 734 pgml^^h (mean + SEM, n = 4). Plasma glucose and IRI showed no change following oral fat after either diet. b) Effect of high and low fat diets on responses to oral fat with exogenous insulin. Approximately 20 minutes after the i.v. insulin bolus all volunteers experienced some side effects of varying duration; they included (in more or less chronological order) sleepiness/disorientation, deterioration of vision, sweating and finally hunger. In both groups the manifestation of side effects was less pronounced following each subject's second insulin stress test. It was at this stage that one member of the first group was excluded as his "normal" fat intake was calculated to be 157 g fat per day. page 137 The results of this experiment are shown in figure 4.6. It can be seen that, after oral fat, plasma IR-GIP levels are inhibited by exogenous insulin following the low fat diet. The plasma IR-GIP AUC results were 5677 + 962 pgml^^h (mean + SEM) following the high fat diet versus 3979 + 597 pgml— ^h for the low fat diet (p<0.01, n=8, paired t-test). No such difference was observed in the IR-GIP response following oral fat alone after the two dietary regimens (see fig. 4.5). The volunteer with the normally high fat diet, (who had to be excluded), showed no suppresion of IR-GIP release with i.v. insulin following the low fat diet. At all time points plasma glucose levels were significantly lower on the low fat diet after insulin, compared to the high fat diet (figure 4.7). Basal glucose levels were elevated on the high fat diet, so the results were re-plotted as a percentage decrease of plasma glucose from basal levels (figure 4.8). Normoglycaemia was re-established more rapidly while on the high fat diet than on the low fat diet following the i.v. insulin. Plasma IRI levels were measured in four volunteers. No significant difference was observed in basal levels or in the rate of disappearance of the exogenous insulin on the two dietary regimens. Plasma triglyceride levels were significantly raised 180 minutes after oral fat on the high fat compared to low fat diet (1.6 + 0.2 vs. 1.14 + 0.12 mmol/L; p<0.025). The difference is demonstrated by calculation of the areas under the curves: AUC triglyceride 1.27 + 0.09 mmol/L.h on the high fat diet was greater than that on the low fat diet 0.38 + 0.12 mmol/L.h (mean + SEM) p<0.005 (figure 4.9) page 138 c) Effect of high and low fat diets on the response to oral glucose. Figure 4.10 shows that plasma IRI levels were not significantly different after oral glucose on the two dietary regimens. Figure 4.11 shows that plasma glucose was similar on both diets following oral glucose. Table 4.1 Effect of short term high an low fat fat diets on body weights of subjects (n = 9, mean + SD)- Low fat diet; start (Day 1): 66.4 + 9.6 kg end (Day 9): 66.1 + 10.1 kg High fat diet: start (Day 1): 66.4 + 9.3 kg end (Day 9): 66.4 + 10.1 kg Table 4.2 Daily energy and fat intakes for subjects on short term high (SHF) and low fat diets (SLF). (n = 9, mean + SD). Groups I and II Mean energy intake Mean fat intake kJ (Kcal) / day kJ (Kcal) / day Usual diet 523 + 228 (2188 + 952) 80 + 52 SLF 299 + 89 (1253 + 393) 21 + 10 SHF 556 + 79 (2325 + 331) 152 + 11 n a n o 1 Figure 4.5 Plasma IR-GIP réponse to oral fat (96 g) in volunteers on low (o> and high (•) fat diets (n = 6, mean + SEM). o OJ o 00 o C\J UJ o CD I a: ty) a. o o C7) I c I— ro c\j o 1 An Figure 4.6 Plasma IR-GIP response to oral fat (96g) followed at 60 minutes by i.v. insulin (0.2 lU/kg) in volunteers on low (o) and high (•) fat diets (n = 8, mean + SEM). o CM O 00 O CM UJ CD CM O O CD CO Qu CD O O CO _J Cl. Lu E O CD C I -I CO CM o page 141 Figure 4.7 Plasma glucose response to oral -fat (96g) followed at 60 minutes by i.v. insulin (0.2 lU/kg) in volunteers on low (o) and high (#) fat diets (n = 8, mean + SEM). o CM § o CM UJ O to O CD O O CO W) O o Ll_ CO O Q. O OO O page 142 Figure 4.8 Percentage change in plasma glucose response to oral fat (96g> followed at 60 minutes by i.v. insulin (0.2 lU/kg) in volunteers on low (o) and high (•) fat diets SEM) . -, o OJ o CO UJ CD UJ CO o o Q CD O CÛ CO CD _J D_ o _J OJ c o o Qi CO U_ _ l J o _J O o o cn page 143 Figure 4.9 Incremental plasma triglyceride response to oral fat (96g) followed at 60 minutes by i.v. insulin (0.2 lU/kg) in volunteers on low (non-hatched) and high (hatched) fat die.ts (n = 8, mean + SEM) . 20 r 1-6 12 '_i 0*8 o E £ 0*4 ## OL Oral fat Oral fat a lone + 1V Insulin page 144 Figure 4.10 Plasma IRI concentrât1 ions following oral glucose (75 g) in volunteers on low (o-o) and high (#-#) fat diets (n = 5, mean + SEM). O CM O 00 O C^J O CO OO D- O I o o o LO page 145 Figure 4.11 Plasma glucose concentratiions following oral glucose (75 g> in volunteers on low (o-o) and high (•-•) fat diets (n = 5, mean + SEM). o C\J o 00 o C\J c E U i o LU (/) o o CD ex. o I (_ __ I o o page 146 4.2.3 Discussion There is a considerable degree of variation in the amount of fat that humans normally ingest in their diets. In most prosperous western countries fat contributes 35 - 40 % of the total energy intake, while in poorer, Third World countries this figure is about 15% or lower. In order to keep a tighter control on the amount of fat consumed by the volunteers during the high fat diet of this study, it was decided that they should eat the same low fat diet throughout, but supplemented with one high fat substance, i.e. double cream. The amount of fat in the diet prior to the study appeared to affect the IR-GIP response to an oral fat load in the rat (section 4.1). However the human subjects in this experiment showed no change in IR-GIP response. This may perhaps be due to the large variation in normal dietary fat intake (12 - 157 g/day). Or it is perhaps that the metabolic and physiological chages caused even by short term adaptation in the human, when caused by chronic modrate/high fat intake, may not be so readily reversed in the short term. Indeed it may be postulated, that in the case of the one volunteer whose normal fat intake was very high that the reason why i.v. insulin made no difference on the fat-stimulated IR-GIP was that the relatively short term low fat adaptation was insufficient to reverse the chronic high fat adaptation which had occured. On the high fat diet the volunteers showed a significant increase in postprandial triglyceride concentrations compared to the low fat diet. As these changes in triglyceride concentrations were not associated with an increase in circulating IR-GIP concentration, and as an increased rate of page 147 nutrient absorption is associated with increased release of IR-GIP (Cataland et al., 1978; Creutzfeldt et al., 1978) one is led to two possible conclusions: that either the rate of fat absorption is increased, but because of increased GIP consumption at the adipocyte no apparent elevation of circulating IR-GIP is observed, or the rate of fat absorption is not greatly affected (nor, hence, the rate of IR-GIP release) and the increase in circulating triglyceride is due to reduced removal at the adipocyte. It has been shown in the rat that insulin is required for the incorporation of fatty acid into adipose tissue (Beck and Max, 1983). Thus insulin insensitivity may further maanifest itself in this manner. The fact that the above results were more pronounced in the rat experiment (section 4.1) further illustrates the dietary difference between the human (normally ingesting a high fat diet) and the rat (normally ingesting a very low fat diet). This probably, attleast in part, explains the differencees; other factors such as gastric emptying and intestinal motility should also be considered. There is evidence that shows that the rate of absorption of nutrients is a factor in determining the magnitude of IR-GIP response. There is a reduction in rate of nutrient absorption in diseases such as coeliac disease (Creutzfeldt et al., 1976) and marked malabsorption (Besterman et al., 1978), and, as stated previously, the addition to a test meal of the unabsorbable carbohydrate guar gum (Morgan et al., 1979; Jenkins, 1980) have all been shown to depress the IR-GIP response to oral nutrients. Exaggerated IR-GIP responses to a mixed meal have been reported in obese subjects (Creutzfeldt et al., 1978). These could, however, be reduced by dietary restriction (Willms et al., 1978). It has page 148 also been shown that maintenance on a high sucrose diet resulted in a greater IR-GIP response to sucrose compared to subjects on high starch diets (Reisner et al., 1980). It would, therefore, appear that diet rather than body weight is more relevant in determining the magnitude of the IR-GIP response to oral nutrients. A negative feed-back mechanism of insulin on IR-GIP release has been postulated as several groups have demonstrated that intravenous infusions of glucose or insulin reduce the IR-GIP response following oral fat (Brown et al., 1975; Crockett et al., 1976). It has also been observed that obese subjects have a defective insulin feed-back mechanism (Creutzfeldt et al., 1978), the above experimental findings would tend to agree with this. The results of this human study are also in agreement with the earlier rodent experiment (section 4.1). Although the results from that experiment were more clear cut, this is probably due the relatively much larger difference in the fat contents of the two diets, as the difference between low and high fat diets in the rat experiment was larger than that in the human experiment, also the "normal" human diet is higher in fat that of the rat. Also, stricter control of the rats' diet was Y possible. High fat diets have been shown to alter the insjlin sensitivity of a number of tissues, including the human monocyte (Pederson et al., 1980), and rat soleus muscle (Grundleger and Thenen, 1982). Grundleger and Thenen (1982) went on to demonstrate that although receptor numbers decreased (but with no loss in receptor affinity) on a high fat diet, post receptor metabolic changes were far more important in contributing to insulin sensitivity. page 149 The dose of exogenous insulin used in this experiment (0.2 U/kg body weight) in normal subjects results in mild to moderate neuroglycopenic symptoms, ‘ ^ commencing 20 to 30 minutes after insulin injection and lasting 10 to 30 minutes. This occured with all the subjects following both high and low fat diets. It was the first insulin stress test each volunteer undertook which caused the most severe symptoms with most prolonged after effects, irrespective of previous diet. The fact that normoglycaemia was re-established more rapidly following the high fat diet had no discernible effect on the severity of the symptoms or duration. One is, therefore , led to conclude that the manefestation of symptoms following an insulin stress test is atleast partly due to psycological factors. page 150 SUMMARY OF RESULTS Pretreatment of the rats with a high fat diet causes: 1) increased secretion of IR-GIP in response to an oral fat load which is probably due to an increased rate of fat absorption 2) abolition of the feed-back inhibition of exogenous insulin on fat-stimulated IR-GIP release 3) some degree of insulin resistance in both glycaemic ,n coj^rol and sensitivity of the mucosal IR—GIP secreting cell to insulin inhibition. Pretreatment of humans with a high fat diet causes: 1) increased concentrations of circulating triglycerides following an oral fat load but without concomittant increase in IR-GIP secretion, possibly indicating a slight degree of insulin resistance at the adipocyte. 2) a blunting of the feed-back inhibition of exogenous insulin on fat-stimulated IR-GIP release 3) allows normog1ycaemia to be re-established more rapidly following exogenous i.v. insulin, presumably as a result of the slight insulin resistance caused by short term high fat diet. In short, a high fat diet appears to lead to insulin resistance which in the long term one envisages could result in glucose intolerance, hyperinsulinaemia and associated problems. page 151 CHAPTER FIVE EFFECT OF GIP ON THE RAT ADIPOCYTE page 152 5. 0.0 Until recently most of the interest in the actions of GIP has been centred in its incretin role, and to a lesser degree in its enterogastrone role (Brown, 1982). It is only within the last decade that research has been directed towards elucidating any metabolic role that GIP may have in the metabolism of lipids. It is strange that this area of possible GIP action should be so neglected, as it was the action of fat on the small intestine, causing gastric acid inhibition (Kosaka and Lim, 1930), which prompted the research that ultimately led to the isolation of porcine GIP (Brown et al., 1969). With the benefit of hindsight, it would seem sensible to find out if there is a metabolic loop to accompany the physiological loop of fat absorption, which stimulates the release of IR-GIP and which in turn inhibits the emptying of fat from the stomach and its subsequent absorption. At the time nobody looked to see if GIP also helped in the clearance and metabolism of lipid once it has been absorbed. In intact rat adipocytes Dupré et al. (1976) investigat ed possible interactions between glucagon and GIP. They found that GIP, even at pharmacological doses, did not stimulate adenylate cyclase and was therefore non-lipolytic. It was also inhibitory for glucagon-stimulated lipolysis, but it was without effect, at pharmocological doses, on lipolysis induced by secretin or vasoactive intestinal peptide (VIP). At physiological doses, however, GIP could displace an excess of i==I-glucagon (identical to natural glucagon in action) bound to adipocytes. VIP and secretin had previously been shown not to displace ^==”I-g lucagon (Bataille et al., 1974). page 153 Ebert and Brown (1976) reported similar -findings in that GIP demonstrated a weak lipolytic effect but a strong anti-lipolytic action on glucagon-stimulated lipolysis by selectively blocking the activation of adenylate cyclase by glucagon, and to a lesser extent, secretin. It would thus appear that GIP competes with glucagon for specific receptors on adipocytes but does not activate adenylate cyclase, thereby blocking the lipolytic action of glucagon. Very recently, Starich et al. (1986) demonstrated that, in vitro, GIP increases adipocyte insulin receptor affinity. It also increases cellular insulin sensitivity, i.e. potentiates insulin-mediated glucose uptake. Cholecystokin in had no effect under these conditions. It has been shown that GIP can stimulate lipoprotein lipase activity in 3T3-L1 cells, an established mouse embryo fibroblast line resembling an adipocyte (Eckel et al., 1979). Cultured preadipocytes were incubated for 2 h at 37*C with GIP concentrations ranging from 0.005 to 5.0 ng/ml. GIP increased lipoprotein lipase activity measured in both the culture medium and in acetone/ether extracts of the cells. Eckel et al. (1979) were the first group to suggest that GIP could provide a mechanism for the clearance of chylomicron triglyceride after feeding by increasing lipoprotein lipase activity. The first in vivo demonstration of GIP's anti lipolytic action came in 1981 when Wasada et al. showed that exogenous GIP was able to promote the clearance of chylomicron triglyceride from the circulation of dogs. Chyle was collected from donor dogs via a thoracic duct fistula and infused i.v. into normal recipient dogs during an infusion of either GIP (Ipg/kg/h) or saline. In the GIP infused dogs the page 154 rise in plasma triglyceride was significantly below that of the control animals. During the experiment plasma glucose rose slightly but significantly. In the first phase, when GIP alone was infused, plasma glucagon concentrations almost doubled. They doubled again by the end of the chyle infusion. There was no significant increase in plasma IRI. The role of insulin in the regulation of adipose tissue lipoprotein lipase activity in humans has also been investigated (Sadur and Eckel, 1982). Normal subjects were infused with insulin to maintain serum levels of approximately 70 pU/ml while plasma glucose was clamped at euglycaemic levels. Free fatty acids fell to a minimum by 20 minutes and triglycerides by 80 minutes. However, lipoprotein lipase was not significantly raised for 6 h, when it was found that it was inversely related to basal lipase activity. In addition to insulin stimulating lipoprotein lipase, the authors concluded that adipose tissue itself is the main regulator of lipase response to insulin. Regrettably, IR-GIP concentrations were not measured in this experiment. The effect of GIP on fatty acid incorporation into adipose tissue has been studied in vitro using the rat epididymal fat pad model (Beck and Max, 1983). Without insulin in the incubation medium, GIP induced a slight but significant decrease in fatty acid incorporation. However, in the presence of rat insulin (100 pU/ml) it significantly enhanced the insulin-induced fatty acid incorporation in a dose dependent manner from 1 - 4 ng/ml GIP. The authors considered that the existence of such a phenomenon, along with that of a hyperactive entero-insular axis in obese subjects, could represent two important factors in the development of obesity. The same group later confirmed these oaae 155 results, and also went on to show that somatostatin has the opposite effect to GIP at the adipocyte, but that the action is not via the GIP mediated mechanism (Beck, Max and Villaume, 1985). Recently the role of endogenous GIP on the removal of triglyceride has been studied in dogs (Ohneda et al., 1983), with the removal of an intravenously administered trigyceride emulsion following oral glucose and galactose loads. The dogs were fasted overnight, the triglyceride was infused for 90 minutes and the glucose, galactose or water control was administered orally at 30 minutes. Blood glucose and insulin were only increased following the glucose load. Plasma glucagon did not change in any of the groups. There was a large IR-GIP response following the glucose load with a smaller one following the galactose load. Plasma triglyceride rose to the same level at 30 minutes in all three experimental groups. The peak levels of plasma triglyceride and integrated plasma triglyceride for 150 minutes did not differ in the three groups. Moreover, there was no difference in the removal rate of plasma triglyceride following withdrawal of the fat emulsion. The authors concluded that endogenously released GIP does not elicit any effect on triglyceride removal. The experiments below were designed to clarify the role of GIP in the removal of circulating triglyceride in the rat. page 156 5.1.0 Role of i.p. GIP in the clearance of i.p. Intralipid from the plasma of the rat. It appears that GIP has an antilipolytic effect in rat epididymal rat pads (Dupré et al., 1976), enhances lipoprotein lipase activity in mouse cultured preadipocytes (Eckel et al., 1979), enhances the insulin-stimulated incorporation of fatty acids into the adipose tissue of the rat (Beck and Max, 1983) and enhances the rate of clearance of chylomicron triglyceride from the plasma of dogs (Wasada et al., 1981). It was therefore decided that as a preliminary experiment it might be of benefit to investigate the effect of exogenous pure porcine GIP on the removal of a fat emulsion when both are administered intra-peritoneally (i.p.). 5.1.1 Methods 54 male wistar albino rats weighing approximately 150 g (range 135 - 165 g , mean = 154 + 8 g SD) were fasted for 24 h prior to the commencement of the experiment. On the morning of the experiment the animals were randomly divided into nine sets of six animals each and then weighed. The nine sets of animals were allocated to two groups, with four sets in each. The remaining set was the control. Group I received just 20 7. (w/v) soya bean oil emulsion (0.5 ml Intralipid 20 %, Kabi Vitrum, London). Group II received the same amount of Intralipid plus 120 ng pure porcine GIP (Quadra Logic Technologies, British Colombia; GIP III batch 26). Animals page 157 were bled by cardiac puncture (just once) having been anaesthetized 5 minutes previously with 0.15 ml/rat sodium pentobarbitone 60 mg/ml; Sagital, May and Baker, Dagenham). They were then killed immediately. Animals in the control set were anaesthetized and bled at the beginning of the experiment. Animals in groups I and II were then injected i.p. with Intralipid and Intralipid plus GIP respectively. Sets of animals from both groups were then anaesthetised and bled at 20, 40, 60 and 80 minutes post i.p. bolus. Blood was collected into chilled lithium/heparin tubes and stored at 0*C in ice until centrifugation (within 30 minutes). The plasma was stored at -20*C until analysis. The samples were assayed for triglycerides and IR-GIP. 5.1.2 Results The plasma IR-GIP results are shown in figure 5.1 and the plasma triglyceride results in 5.2. The weights of the rats used are shown in table 5.1. The IR-GIP results show that in group I (no GIP) plasma IR-GIP values varied very little from basal, particularly in the first 40 minutes of the experiment. In group II plasma IR-GIP values peaked at 20 minutes, with a mean approximately 3 times higher than that of the basal set of animals, but returned to basal by 60 minutes. At no time in either group did plasma IR-GIP values vary significantly from basal. The plasma triglyceride results show that, again at no time in either group did plasma triglyceride values vary significantly from basal, but group I did show an 18 % increase over basal at its peak at 40 minutes. However, the overall trend did show that in the 80 minute duration of the experiment in the group receiving page 158 the GIP in addition to the Intralipid, the apparent rate of increase in plasma triglyceride concentrations, the peak triglyceride values and overall amount of triglyceride absorbed were all lower than in the group which did not receive the i.p. GIP. page 159 Table 5.1 Weights (g) of animals used in experiment 5.1 20 min 40 min 60 min 80 min Group I 160 145 150 165 145 135 165 160 145 160 145 150 155 160 145 155 140 160 135 165 145 135 145 155 159 149 147 158 mean 2 7 10 6 SD Group 11 Control 160 155 155 155 1 160 145 160 140 160 ; 160 150 155 155 160 ; 160 160 165 150 155 160 160 165 165 155 1 155 155 145 145 160 ! 160 155 157 152 158 ! 159 mean 8 SD page 160 Figure 5,1 Plasma IR-GIP concentrations in normal 150 g rats following an intraperitoneal bolus of Intralipid (0.5 ml, 20 %> with (o) or without (#) pure porcine GIP (120 ng), (n = 6, mean + SEM) . o 00 (/) c B cn o O C O UJ o c\j Q. CD cn cn o CO Q_ CL cn c o o page 161 Figure 5,2 Plasma triglyceride concentrations in normal 150 g rats following an intraper1toneal bolus of Intralipid (0.5 ml, 20 %) with (#) or without (o) pure porcine GIP (120 ng). (n = 6, mean + SEM). T o 00 O CO to c CO o OO Ol UJ c UJ Q o CM CXl CO g CO o OO ox. Q. L CM page 162 5.1.3 Discussion As a preliminary experiment the results have been encouraging. Although neither of the parameters monitored achieved significance during the experimental period, they did tend to support the hypothesis that GIF has a role in the clearance of ingested fat, as set out in the introduction to this chapter. The main limitations in the experimental design would seem to be that the doses of GIF and Intralipid used were too small. However, results in chapter 4 show that even considerably larger doses of lipid, when given orally, give only modest rises in plasma triglyceride levels with poor stimulation of IR-GIP, indicating slow absorption rather than rapid removal. The dose of GXP given did cause a three fold increase over basal, but this just r e a c h e d the range expected for stimulated IR-GIP values following triglyceride stimulation in normal rats. The above conclusions assume that the rate of drainage and/or absorption of Intralipid from the intraperitoneal cavity was the same in both groups. A faster drainage in group I would account for a higher, earlier peak in plasma triglyceride as was observed in this group. However, exogenous GIP has been shown to increase mesenteric blood flow in the cat (Fara and Salazar, 1978). This would have the opposite effect to that observed in this experiment and could clearly not account for the differences observed. The accuracy of the triglyceride results relies on the assumption that the free glycerol in the Intralipid does not interfere to any great extent with the enzymic triglyceride assay; this was shown to be correct in chapter 7. In conclusion, although this experiment has not yielded any hard facts about the role of GIP in lipid clearance, and despite the limitations, as detailed above, it has been encouraging and provided the incentive to proceed with other experiments in this area. page 164 5. 2. O Role of exogenous GIP in the clearance of i.v. Intralipid in the rat. Following on from the previous experiment, it was decided that it might be more productive to repeat the experiment administering the GIP and/or Intralipid intravenously rather than use intraperitoneal loads. This would circumvent any possible problems with differing rates of drainage or absorption of material from within the peritoneal cavity. It was also decided that, rather than use individual animals for each time point, serial samples would be taken from individual animals. This would be far less wasteful in animals and reagents, although it would mean much smaller sample volumes. It would, however, allow the use of paired statistics. The use of Intralipid, rather than chyle collected from donor rats, was justified by its ready availability and its lack of any fat soluble enteric factors which might augment the action (if any) of GIP. It has been shown that Intralipid micelles, when incubated with rat plasma, rapidly incorporate free apo-proteins and become indistinguishable from endogenous chylomicra (Robinson and Ouardfort, 1979). Intralipid and chylomicra have identical kinetics for their plasma elimination in man and dog (Hallberg, 1965), and they both share similar kinetic behaviour as substrates in vitro for post heparin lipoprotein lipase (Boberg and Carlson, 1964). page 165 5.2.1 Methods 12 male Wistar albino rats weighing approximately 250 g (range 220 - 270 g, mean = 256 + 12 g SD) were -fasted for 24 h prior to the commencement of the experiment. On the morning of the experiment the animals were randomly divided into two groups of six, weighed and marked with a number from 1-6. All of the animals were anaesthetized by intraperitoneal injection of 0.3 ml Sagital (sodium pentobarbitone 60 mg/kg). Each animal was left for 15 minutes before the femoral vein was exposed (left side). The zero time point blood sample was taken (cut tip of the tail, 0.3 - 0.5 ml), and each animal then received a bolus injection via the femoral vein. Animals in group I received 0.5 ml of a 20 % (w/v) soya bean oil emulsion (Intralipid 20 %, Kabi Vitrum), while animals in group II received 500 ng pure porcine GIP (Quadra Logic Technologies) dissolved in the same amount of Intralipid. Animals were then bled at 5, 15, 30, 45 and 60 minutes after the i.v. bolus. The sample volumes were again 0.3 - 0.5 ml and were collected into chilled paediatric (2 ml) lithium heparin tubes and subsequently treated in the same way as the previous experiment. Aliquots of plasma were assayed for triglyceride, IRI and IR-GIP. The insulin assay was performed by Dr. KS Tan using a double-antibody radioimmunoassay using an antiserum raised against bovine insulin, iodinated bovine insulin and rat insulin standard; assay sensitivity was 0.5 ng/ml (Hampton et al., 1983, Hampton, 1983). Data was subjected to Student's t-test for statistical analysis (for paired observations where necessary). page 166 5.2.2 Results Plasma IRI results are shown in figure 5.3 and plasma triglyceride results in figure 5.4, weights of the animals are shown in table 5.2. Plasma GIP concentrations were found not to have risen above the limit of detection of the assay (500 - 600 pg/ml as 20 pi samples were used) in either group at any time point during the experiment. The plasma IRI results show that at no time in the experiment do mean plasma IRI levels reach "stimulated values" (basal rat plasma IRI levels being 1 - 2 ng/ml with this assay; KS Tan, personal communication). The areas under the curve were not significantly different. At no time do plasma IRI values vary significantly from the zero time value during the experiment in group I (no GIP). The only time point where plasma IRI values were different between groups was at 15 minutes past bolus (p < 0.05). Plasma triglyceride results show that, following the i.v. bolus, a peak is reached in both groups at 15 minutes. The plasma triglyceride values are significantly different at all time points, the largest statistical difference between groups being seen after the 5 minute time point (p <0.0005). The areas under the curve (incremental above basal) were also significantly different (p < 0.0025). It is clear that the triglyceride values up to and at 5 minutes for the animals in group II are smaller than those in group I. However, for the other time points studied from 15 - 60 minutes the mean triglyceride values parallel each other. page 167 Tab le 5.2 Weights (g) of rats used in experiment 5.2 Number Group I Group 11 1 265 265 2 275 250 3 268 265 4 243 255 5 243 235 6 248 254 mean 257 254 SD 14 11 oaae 168 Figure 5.3 Plasma IRI concentrations in normal 250 g rats following an intravenous bolus (femoral vein) of Intralipid (0.5 ml, 20 %> with (•> or without (o) pure porcine GIP (500 ng)- (n = 6, mean + SEM). o to o If) o o CO o CO UJ o H- C\J oo to o to 00 a. cn _I CO c\j o page 169 Figure 5.4 Plasma triglyceride concentrations in normal 250 g rats following an intravenous bolus (femoral vein) of Intralipid (0.5 ml, 20 %) with (#) or without (o) pure porcine GIP (500 ng). (n = 6, mean + SEM). o o LD O «51- l/) c CO o O CQ CO d: ui o f— CM Q Or: o CD O CO cc CL I J. CO page 170 5.2.3 Discussion It is clear from the IRI results of group I that an i.v. bolus of Intralipid (in human terms equivalent to an Intralipid bolus of 230 ml or about 46 g soya bean oil) has little effect on plasma IRI. The presence of BIP in the bolus as presented to group II did cause a slight but significant Increase in plasma IRI at all but the 15 minute time point. The plasma IR-GIP concentrations immediately after the i.v. bolus would probably have been supraphysiological, thus causing a slight peak in mean plasma IRI at the 5 minute time point, circumstantial evidence of GIP's presence. Although <3IP is only insulinotropic during hyperglycaemia (Elahi et al., 1979), infusions yielding supraphysiological levels of plasma IR—GIP in rats have been shown to cause transient insulin release (Pederson and Brown, 1976). There are undoubtedly differences in the plasma triglyceride levels in the two groups. The inclusion of GIP in the i.v. Intralipid bolus reduced the peak triglyceride value, however, as the two curves parallel each other after 5 minutes (in the last 30 minutes both groups show the same rate of decline in triglyceride values of 4 mmol/L/h). It would seem that the two groups react differently only in the first 5 minutes, during the possible hyperGIPaemia. This suggests that the GIP effect is short lived, and only effective while GIP levels are high. The fact that mean basal triglyceride levels were significantly higher in group II merely serves to heighten the difference between the two groups at 5 minutes. The main flaw in drawing the conclusion that GIP does augment the clearance of circulating lipid, is the lack of page 171 data, showing that plasma IR-GIP concentrations were in fact elevated immediately following the bolus. It was possible that the GIP was very quickly removed, or on the other hand, it is remotely possible that no GIP was injected at all, and that the triglyceride results were merely an artefact. Of the two possibilities, the transient and slight hyperinsulinaemia seen in group II at 5 miinutes would indicate that the pancreatic B-cell had been exposed to an insulinotropic stimulus such as a bolus of GIP. The detection of a slight rise in plasma IRI in group II, although it appears to confirm that GIP was in fact injected into the animals, does complicate the interpretation slightly since it becomes less clear whether the explanation for the observed results can be exclusively attributed to GIP. In the in vitro experiments of Eckel et al (1979)the stimulation of lipoprotein lipase activity could only be attributed to the presence of GIP. It may be that insulin has a part to play, as in the experiment of Wasada et al. (1901), where plasma IRI was observed to rise "modestly", and the work of Beck and Max (1983), who showed that GIP augmented the insulin induced incorporation of fatty acids into adipose tissue. The above conclusions are open to some dispute as they assume that GIP was actually injected, and that it is biologically active. The only way to resolve the question of whether GIP went in or not would be to repeat the experiment and sample more quickly after the GIP/Intralip id bolus, and perhaps run a GIP only control to see if the presence of Intralipid accelerates the removal of GIP. Also, it is still unclear what role, if any, insulin plays in the removal of lipid in this context. It must also be borne in mind that page 172 both this and the previous experiment rely on at least partial bio-activity of the GIP preparation used. It is assumed that there has been minimal degradation of the peptide (a new batch was used) and that porcine GIP is biologically active in the rat. page 173 5. 3.0 Effect of anti-GIP antibodies on the clearance of plasma triglyceride following oral triolein in rats. This experiment seeks to address the problem of whether GIP augments the rate of clearance of circulating triglyceride, using a different experimental approach to that of the previous two experiments. Rather than study the effect of exogenous porcine GIP on exogenous lipid, this experiment seeks to determine the effect of the immunological removal of endogenous GIP on the clearance of endogenous chylomicron triglyceride following an oral fat load. This is analogous to the work of Ebert et al. (1979) where anti—GIP antibodies were used to show that GIP was an incretin, although not the only one. It was also demonstrated by Wolfe et al. (1983), using antibodies, that GIP was an enterogastrone in dogs and that at least part of this effect was exerted via gastrin. It was subsequently shown that GIP may exert its effect on gastrin via somatostatin (Wolfe and Reel, 1986) 5.3.1 Methods 12 male Wistar albino rats weighing approximately 175 g (range 162 - 190 g , mean = 175 + 8 g SD) were fasted for 24 h prior to the commencement of the experiment. The animals were randomly divided into two groups of six, weighed and marked with a number 1 - 6. Each animal was then immobilised in a Tan/Bishop rat restrainer (Dr. KS Tan, University of Surrey), bled from the cut tip of the tail (0.3 - 0.5 ml), orally dosed (gavage) with 1 ml triolein (Sigma, London) and then page 174 Injected with 50 mg normal sheep IgG in 1 ml saline (group I) or 50 mg immune (anti-GIP) sheep IgG in 1 ml saline (group II). Each animal was then bled (cut tip of tail) at hourly intervals for the next 5 hours. Blood was collected and stored in the same manner as for the previous experiments. The normal sheep IgG and sheep anti-GIP IgG had been prepared previously by a modification of the procedure of Campbell et al. (1970). Sheep anti-GIP serum (S687/011580) was brought to 33 % saturation of ammonium sulphate solution (AnalaR, BDH, Poole). The precipitate was centrifuged, resuspended in a minimum of distilled water and dialysed extensively against distilled water. It was then aliquotted into 50 mg protein portions, freeze-dried and kept at -20"C until required. The protein concentration had been estimated by measuring the optical density at 280 nm and applying the factor that an Ea## of 1.4 equals 1.0 mg/ml protein. The same was done with normal sheep serum to give non-immune sheep IgG. In the standard radioimmunoassay the aliquot of sheep anti-GIP IgG gave a final titre of 1:6000. Samples from both groups were assayed for triglycerides and IRI. Samples from group I were assayed for IR-GIP and samples from group II for the presence of anti-GIP antibodies. Analytical data was subjected to analysis of variance or Student's t-test (for paired values where necessary). 5.3.2 Results The plasma triglyceride results are shown in figure 5.5 and plasma IRI results are shown in figure 5.6. Weights of animals used are shown in table 5.3. No GIP antibodies were oaae 175 detected in the zero time samples from the animals in group II. Thereafter all animals showed an excess of antibody present (on the plateau at the top of the antiserum dilution curve) when samples diluted 1:5 were assayed. Plasma IR-GIP concentrations remained in the undetectable range (limit 500 - 600 pg/ml) for the first 4 h of the experiment. However, by the fifth hour mean plasma IR-GIP concentrations had risen to 787 pg/ml + 140 SEM. If one refers to IR-GIP results obtained under similar conditions in experiment 4.1, one sees that these results are to be expected. The results show that both groups I and II exhibit a steady increase in plasma triglyceride values during the course of the experiment. Group II's mean plasma triglyceride concentrations are generally higher than those observed in group I, and the 2 and 3 h time points are significantly different (p < 0.10 and p < 0.0005 respectively). The mean areas under the curve for group II were significantly greater than those for group I (p < 0.05). Analysis of variance showed that behaviour of the two groups was significantly different; F(l,60) = 7.4708 (p < 0.01). Oral triglyceride does not appear to stimulate plasma IRI in either of the two groups and no difference is observed between the two groups by analysis of variance. oaae 176 Table 5.3 Weights (g) of rats used in experiment 5.3 Number Group I Group II 1 173 180 2 173 168 3 190 171 4 170 178 5 185 172 6______177_____ 162 mean 178 172 SD 8 7 oaae 177 Figure 5.5 Plasma triglyceride concentrations in normal 175 g rats following an oral triolein load (gavage, 1 ml) and intraperitoneal bolus (1 ml) of either anti-GIP (•) or non-immune sheep (o) immunoglobulin (50 mg). (n = 6, mean + SEM). (* = p < 0.1, **** = p < 0.0005) a: LÜ (/) CO UJ Q or o >- CD cn CM s o Q cn >- UJ CD Dd CO < cn cn CM CD 1 ~7 O Figure 5.6 Plasma IRI concentrations in normal 175 g rats following an oral triolein load (gavage, 1 ml) and intraperitoneal bolus (1 ml) of either anti-GIP (#) or non-immune sheep (o) immunoglobulin (50 mg), (n = 6, mean + SEM). Ê to Û- Q z UJ C O o > - CD cn h- CM cn o cn cn cn Q- ' i c L. CM O 5.3,3 Discussion The above results show that, in the absence of any interaction with IRI, endogenous GIP appears to have a role in the removal of post-prandial chylomicron triglyceride which can be partially abolished by the use of anti-GIP antibodies. Plasma IRI concentrations were not affected by the ingestion of fat or by the anti-GIP antibodies. However, there does appear to be a significant difference in plasma triglyceride concentrations between the two groups. The group which received the anti-GIP antibodies showed consistently Higher plasma triglyceride values for the first 4 h of the experiment, suggesting that endogenous GIP helps in the removal of endogenous triglyceride. It may also be possible that the anti-GIP antibodies had removed the inhibition on gastrin release and gastric emptying (Wolfe et al., 1983). This would mean that the difference shown in plasma triglyceride profiles may have been caused by the removal of the enterogastrone effect, causing dumping and increasing the rate of absorption of triglyceride, rather than affecting removal. However, it has been shown that GIP stimulates mesenteric blood flow (Fara et al., 1972; Fara and Salazar, 1978), and it may be argued that the removal of any possible enterogastrone effect may be partially, or totally, reversed by the possible reduced removal of lipid from the small intestine by the mesenteric veins and lymph vessels owing to the lack of stimulation of blood flow by GIP. It may not be surprising that plasma IR-GIP concentrations remained so low for so long in the above experiment. In experiment 4.1 one can see that the same oral oaae 180 dose of fat in similar rats gave mean plasma IR-GIP concentrations of only about 700 pg/ml in several hours from basal values of only about 200 pg/ml. In addition, one might expect delayed absorption of triglycerides since, following an i.p. injection, there might be short lived spastic ileus of the small intestine (KS Tan, personal communication). It has become clear that in the design of an experiment to investigate the role of GIP in the removal of lipid from the circulation, it is almost impossible to remove all potential interferences and ambiguities. Thus, the experiments in this chapter have sought to improve on previous work, in the light of the experimental data gained, but even this experiment (5.3) is flawed by possible unwanted interactions with gastric emptying and mesenteric blood flow. Interference from gastric emptying could probably be excluded if the experiment were repeated with intraduodenal instillation of a fat emulsion or an in situ gut perfusion (Sykes et al., 1980; Dryburgh et al., 1980) rather than an oral load, although such procedures begin to diverge from the physiological. In conclusion, I believe that the three experiments in this chapter, when taken together, suggest that GIP does have a role in the clearance of circulating lipid. oaae 181 SUMMARY OF RESULTS 1) There is a suggestion that in rats, exogenous porcine GIP administered intraperitoneally augments the removal of intraperitoneally administered fat in the form of Intralip id. 2) Intravenously administered porcine GIP significantly reduces the rise in plasma triglyceride following an intravenous Intralipid load. 3) Intraperitoneally administed sheep anti-GIP immuno globulin significantly reduces "endogenous" plasma chylomicron triglyceride following an oral triolein load. o a a e 1 8 2 CHAPTER SIX THE ROLE OF GIP IN THE DIFFERENTIAL INSULIN RESPONSES TO ORAL AND INTRAVENOUS GLUCOSE IN THE TRANSPLANTABLE RAT INSULINOMA. paqe 183 6.1.0 Introduction These experiments seek to shed some light on the role of GIP in the entero-insular axis. As in chapters 3 and 4, the assertion that GIP was an incretin (LaBarre and Still, 1930) was taken as fact (Creutzfeldt, 1979; Brown, 1902; Creutzfeldt and Ebert, 1985). This experiment seeks to elucidate the cellular processes involved in the entero-insular axis (Unger and Eisentraut, 1969). To this end the transplantable rat insulinoma (RIN) was used as a model. The development of the x-ray induced RIN in the NEDH rat (Chick et al., 1977) made available large amounts of material for pancreatic B-cell research (Duguid et al., 1976; Gazdar et al., 1980; Sopwith et al., 1981; Eisenbarth et al., 1981). The changes in glucose homeostasis induced by the RIN are similar in some respects to those produced by human insulinomas (Frerichs and Creutzfeldt, 1976; Flatt et al., 1982; Marks and Rose, 1981), suggesting its suitability as a model for the study of spontaneous insulinomas. However, the impaired insulin response to i.p. glucose and arginine suggests that RIN cells possess an intrinsic defect which makes them functionally quite different from normal rat and human B-cells. In the colony of animals studied the islet cell tumour is morphologically well differentiated and contains immuno- reactive insulin (IRI), significant amounts of immunoreactive somatostatin (1RS) but no detectable immunoreactive glucagon (IRG) (KS Tan, unpublished data). The insulinoma tissue has been found to contain detectable quantities of immunoreactive pancreatic polypeptide (IR-PP), although, the levels were much lower (about one thousandth) than in normal rat pancreas oaoe 184 (O'Hare et al., 1985). An impaired insulin response by the RIN to intraperitoneal glucose and arginine stimulation has been reported (Tan et al., 1981). It was found that both glucose and arginine failed to stimulate IRI concentrations above the high basal levels, whereas glucagon was able to elicit a potent release of IRI. To further understand the nature of the tumour B-cell, the responses of the insulinoma to intravenous and intragastric glucose was investigated with particular respect to the effects of GIP on glucose homeostasis in the tumour bearing animals. As has already been stated in the previous chapters, GIP is currently the strongest candidate for the glucose dependent insulinotropic agent released from the gut following oral carbohydrate stimuli (Creutzfeldt, 1979; Creutzfeldt and Ebert, 1985), termed incretin by LaBarre and Still (1930). GIP has been shown to be insulinotropic at elevated glucose levels in the rat (Rabinovitch and Dupré, 1974; Pederson and Brown, 1976); this effect can subsequently be inhibited, at least in part, by the use of anti-GIP antibodies (Lauritsen et al., 1981; Ebert and Creutzfeldt, 1982). 6.1.1 Methods The colony of insulinoma-bearing animals was established in the animal breeding unit of the University of Surrey in 1980. Breeding pairs were kindly provided by Professors WL Chick (Boston, Mass., USA) and CN Hales (Cambridge, UK). The insulinoma is maintained by serial subcutaneous transplantation within the colony. A donor rat is killed by cervical dislocation, the tumor (typically 1 - 2 g) is oaae 185 excised and trimmed of adherent fat and other tissue. The capsule is cut open and a slurry of blood and insulinoma cells are scraped from the fibrous capsule with the edge of a scalpel. The slurry is sucked-up into a plastic 1 ml syringe, and approximately 0.1 ml implanted subcutaneously into the scapular region of the rat, with a stainless steel trocar (2 mm outer diameter). The trocar is inserted some 40 - 50 mm posterior to the scapular region, thus forming a channel for the insulinoma in which to grow. A single donor rat was used for each experiment and recipients were male 12 -13 week old inbred NEDH rats (New England Deaconess Hospital, Boston, Mass., USA). Transplantion results in hyperphagia, increased weight gain, marked hyperinsulinaemia, and severe hypoglycaemia with death of the recipient by 22 - 27 days (Tan et al., 1981). All experiments were conducted 19 days post-transplantation. The animals were housed in an air-conditioned room at 22 + 2*C with a lighting schedule of 12 h light (0700 h - 1900 h) and 12 h dark. A standard pellet diet (Spratts Laboratoy Diet No. 1, Lillico, Reigate, Surrey) and tap water were available ad libitum. Six fed and concious insulinoma-bearing rats and 6 age matched control NEDH rats were bled (0.5 ml into 2 ml paediatric lithium heparin tubes) from the cut tip of the tail. They then received intragastric glucose (5 g/kg, as a 50 % solution by oral gavage). A further group of NEDH rats, twelve tumour bearing and six non-tumour bearing were anaesthetised by intraperitoneal injection of sodium pentobarbitone (Sagital, May and Baker, Dagenham; 60 mg/kg body weight). After basal, time -10 and 0 minutes blood samples (as above) all animals received an intravenous page 186 (femoral vein, as in chapter 5) bolus of glucose (0.5 g/kg, 25% in water). However, half of the insulinoma-bearing group together with the glucose received 1 pg/kg pure porcine GIP (Oadra Logic Technologies, GIP III, batch 26). All animals were further bled from the cut tip of the tail for the remainder of the experiment, at 20 minute intervals for one hour in the "intragastric" animals and at 2, 15, 30 and 40 minutes in the "intravenous" animals. Aliquots of plasma were assayed for plasma glucose and IRI (the latter by Dr. KS Tan, details in chapter 4). The results were subjected to Student's t-test for unpaired data. 6.1.2 Results Figure 6.1 shows that following intragastric glucose, plasma glucose concentrations were significantly elevated at the 20 minute time point with respect to basal (O minute time point) in both insulinoma-bearing rats (7.8 + 0.7 mmol/L vs. 1.7 + 0.2 mmol/L: mean + SEM; p < 0.001) and in control NEDH rats (10.0 + 0.5 mmol/L vs. 5.5 + 0.3 mmol/L; p< 0.001). Figure 6.2 shows that the rise in glucose levels was accompanied by an increase in plasma IRI concentrations compared with basal in both insulinoma-bearing rats (18.3 + 1.5 pg/L vs. 13.1 + 0.8 pg/L; p < 0.01) and controls (7.8 + 0.5 pg/L vs. 2.8 + 0.4 pg/L; p < 0.001). In contrast to the above, figure 6.3 shows that intravenous glucose failed to cause an increase in plasma IRI concentrations above high basal levels in the insulinoma-bearing rats. However, glucose, in combination with GIP, stimulated plasma IRI in the insulinoma-bearing rats significantly by 2 minutes (20.6 + 2.5 pg/L; p < 0.05) and 15 minutes (18.6 + 1.1 pg/L; p < page 187 0.02) compared to the control insulinoma animals given glucose alone. Figure 6.4 shows that there were no significant differences in plasma glucose levels between the two groups (insulinoma-bearing + i.v. glucose + i.v. GIF) at any time point; both groups showed significant rises above basal values for the two minute time point (p < 0.02). Figure 6.5 shows that in the control non-insulinoma-bearing NEDH rats i.v. glucose significantly stimulated plasma IRI ((10.3 +1.1 pg/L; p < 0.001) and plasma glucose (10.0 +0.8 mmol/L; p < 0.001) by the five minute time point vs. the zero time point. page 188 Figure 6.1 Effect of intragastric glucose (5 g/kg body weight) on plasma glucose concentrations in insulinoma-bearing (•) and control NEDH rats (o). (n = 6, mean + SEM). (* = p < 0.001 vs. zero time point for group). 1 0 £ 8 E S 6 O O Z) _i O 4 < CO < 2 CL 0 L i i 0 20 40 60 TIME ( MIN ) page 189 Figure 6.2 Effect of intragastric glucose (5 g/kg body weight) on plasma IRI concentrations in insulinoma-bearing (•) and control NEDH rats (o). (n = 6, mean + SEM). (* = p < 0.01 and ** = p < 0.001 vs. zero time point for group). 20 16 D) 3L. 12 Œ < CO 8 < —I CL 0 L i 1 0 20 40 60 TIME (MIN) page 190 Figure 6.3 Effect of intravenous glucose GXP (#) (1 pg/kg body weight) plus glucose (0.5 g/kg body weight) on plasma IRI concentrations in insulinoma-bearing NEDH rats, (n = 6, mean + SEM). (* = p < 0.05 and ** = p < 0.02 between groups). * * 20 1 8 O) 16 ÛC < 14 CO < _l CL 12 10 L I L J I 10 0 15 30 40 TIME ( MIN ) page 191 Figure 6,4 Effect of intravenous glucose (o> (0.5 g/kg body weight) and SIP (•) (1 pg/kg body weight) plus glucose (0.5 g/kg body weight) on plasma glucose concentrations in insulinoma- hearing NEDH rats, (n = 6, mean + SEM). O E E LU CO O O Z) _l CD < CO < _1 0_ 0 i 1 10 0 15 30 40 TIME ( MIN ) page 192 Figure 6.5 Effect of intravenous glucose (0.5 g/kg body weight) on plasma IRI and glucose concentrations in insulinoma-bearing NEDH rats, (n = 6, mean + SEM). (* = p < 0.001 vs. basal). 10 O) 3L CD < ± ± J 0 15 30 45 TIME ( MIN ) ^ 10 CL 1 0 15 30 45 TIME( MIN) page 193 6.1.3 Discussion The results demonstrate quite clearly that although intragastric (oral) glucose causes an increase in plasma insulin in the insulinoma-bearing animals, which is in line with the insulinotropic action of the glucose in the control, non-insulinoma-bearing animals, intravenous glucose has absolutely no effect on plasma insulin levels in the insulinoma-bearing rats. However, the addition of porcine GIP to the intravenous glucose bolus in the insulinoma-bearing rats restores the insulinotropic action of the glucose. The The results from the intravenous glucose + GIP experiment demonstrates nicely the insulin resistance of the insulinoma, as the large increase in plasma IRI caused by the GIP makes no difference to the plasma glucose response. Even higher doses of i.v. glucose have been found to be non-insulinotropic in these insulinoma-bearing rats (Tan et al., 1981; Flatt et al., 1982). This Insensitivity to glucose has also been reported with human insulinomas (Frerichs and Creutzfeldt, 1976; Marks and Rose, 1981). It would be interesting to speculate as to whether this insensitivity of the insulinoma "B-cel1" is an intrinsic function of the insulinoma or a loss or down regulation of surface receptors secondary to the hyperinsulinaemia, and whether this could also be demonstrated on normal pancreatic B—cells. Whilst responsiveness of the insulinoma to i.v. glucose is lost, its ability to respond to oral glucose is retained, as in many human insulinomas (Marks and Rose, 1981). The above results showing a differential response to oral and i.v. glucose provides strong evidence for the existence of the entero-insular axis and incretin, since any page 194 neural involvement in this situation is highly unlikely. GIP is the stongest candidate for incretin (Creutzfeldt, 1979; Creutzfeldt and Ebert, 1985). GIP is insulinotropic only in the presence of hyperglycaemia (Elahi et al, 1979). This glucose dependence suggests a degree of coopérâtivity between glucose and GIP at the receptor or post receptor level. However, it would appear that the cooperativity is not absolute and that a degree of dissociation exists between them. Jorde et al. (1986) have recently described a similar phenomenon. Normal human volunteers received intravenous gastric inhibitory polypeptide at mean blood glucose levels of 4.9 mmol/L following a priming infusion with glucose. A significant insulin release was seen during the GIP infusion. An effect the authors could not repeat without the priming glucose infusion. This further demonstrates the dissociation between glucose and GIP receptors. GIP may thus be implicated in the well characterised priming effect of glucose on insulin secretion (Grill, 1981). The interaction between GIP and its receptor was unaffected by the Insensitivity of the tumour to glucose stimulation, suggesting that the tumour B-cel1 defect is specific for the glucose, rather than GIP, mechanism of action. The glucagon receptor on the insulinoma cell surface has been demonstrated to be functionally intact (Tan et al., 1981). It is interesting to note that GIP was able to stimulate a large and sustained IRI secretion from the insulinoma during conditions of subnormal plasma glucose. However, for the insulinoma-bearing rats "euglycaemia" has become 1 - 2 mmol/L, and the 4 mmol/L obtained after i.v. glucose page 195 represents marked hyperglycaemia. This may indicate some sort of "recalibration" of the receptors, in order to take into account the ambient conditions of actual hypog1ycaemia. In order to confirm the conclusions drawn above, further work must be done. If the intragastric glucose load experiment were repeated in the manner of experiment 5.3, i.e. with and without anti-GIP antibodies, further weight might be added to the argument. It might also, in the case of an incomplete removal of the incretin effect, indicate the existence of other possible incretins. If it were possible to identify and locate the glucose and GIP receptors on the surface of the RIN cell, using differentially labelled agonists or monoclonal antibodies, the case for cooperativity between the two receptors would be strengthened if they were demonstrated to be adjacent to one another. In conclusion, the above results indicate that GIP is an incretin in the transplantable rat insulinoma, and that the tumour has lost its insulin responsiveness to intravenously induced hyperglycaemia, but it retains its ability to respond to intragastric glucose. SUMMARY OF RESULTS 1) GIP is an incretin in the transplantable NEDH rat insulinoma even under conditions of apparent hypoglycaemia. 2) The transplantable NEDH rat insulinoma has lost its responsivenes to intravenous but not intragastric glucose. page 196 CHAPTER SEVEN THE ROLE OF ENDOGENOUS CARBOHYDRATE-STIMULATED IR-GIP IN THE REMOVAL OF EXOGENOUS TRIGLYCERIDE IN MAN. page 197 7.1.0 Introduction This chapter follows on directly from chapter five in trying to elucidate the role of GIP in the clearance of chylomicron triglyceride, but this time in humans. The reasons for following this line of research are discussed in chapter five. As an overactive entero-insular axis might be implicated in insulin resistance and maturity onset diabetes (chapter four) an overactive "entero-adipose" axis might at least be implicated in the pathology of obesity fchapter eight). In an attempt to remove the possible problems of lack of incomplete biological activity of porcine GIP in humans, and ethical problems with the infusion of porcine GIP into humans (not to mention a certain amount of "volunteer resistance" at this prospect), it was decided to stimulate GIP endogenously. In order to differentiate any possible interactions with insulin, two different carbohydrate loads were used; glucose, which stimulates IR-GIP and insulin, and galactose, which stimulates IR-GIP but which does not, or only poorly, stimulates insulin (Morgan et al., 1979; Sykes et al., 1980). Again, since it would be more acceptable to volunteers. Intralipid was used as a "chylomicron" source as it has been demonstrated to be processed in a manner similar to endogenous chylomica (Robson and Quardfort, 1979; Hallberg, 1965; Boberg and Carlson, 1964). Intralipid is also free from any possible, unidentified, gut factors which might affect the removal of chylomicron triglyceride. The use of Intralipid as a "chylomicron" source does, however, present problems in terms of triglyceride analysis. Intralipid 20 7. is approximately 0.23 M triglyceride; it is oaoe 198 also 0.24 M free glycerol. The triglyceride assay used (Lange, unpublished; details in chapter 2), in common with most other enzymic triglyceride assays, hydrolyses the triglyceride to fatty acids and glycerol, and estimates the glycerol, thus giving an answer which includes both free and bound glycerol. It has been suggested that for accurate determination of plasma triglyceride concentrations separate, free glycerol determinations should be made for each sample in order to eliminate the free glycerol component (Stinshoff et al., 1977; Welle et al., 1984; Mott and Rogers, 1978). In an experiment where a safflower oil based emulsion was infused into children it was maintained that the free glycerol was quickly removed and could thus be ignored (Cooke and Burckhart, 1983). The alternative to measuring the free glycerol and/or triglyceride would be to monitor the actual removal of the Intralipid particles. This is done in the nephelometric measurement of the rate of removal of Intralipid particles from plasma in the routine clinical procedure for the assesment of intravascular lipolysis. The light scattering intensity (LSI) measured -by this technique is critically dependent on particle size. It is consequently most sensitive to changes in the concentration of the largest particles in the circulation following Intralipid infusion. Lipolysis produces a progressive reduction in the particle size to form remnant particles which are removed by the liver, so that a possible alternative to LSI measurement would be to monitor the rate of removal of the remnant particles. An enzymic methodolgy for this exists but has not yet been widely evaluated (Richmond et al., 1984). Thus, the experiments that follow seek to establish page 199 whether endogenously released GIP has any role in the removal of exogenously infused triglyceride, having established whether or not the method chosen to assay for triglyceride is appropriate or not. 7.1.1 Methods a) measurement of free glycerol and triglyceride following an intravenous Intralipid load. Four healthy volunteers (two men and two women) aged 22 - 41 years (mean = 30.8 + 5.1 SEM) were studied. They were all within 10 % of their ideal body weight and were not under any medication. After an overnight fast volunteers had antecubital veins in both arms cannulated. Immediately after a basal blood sample had been withdrawn from the left antecubital vein. Intralipid 20 % (50 ml, Kabi Vitrum) was infused at the rate of 5 ml/min into the right antecubital vein. Blood samples were taken from the left antecubital vein at 15 minute intervals, from 15 minutes before and until 2 h after the i.v. load. Blood for glucose analysis was collected into fluoride oxalate tubes, but lithium heparin tubes for glycerol, triglyceride and IR-GIP analysis. Plasma free glycerol analysis was performed by A. Thompson using a double enzyme technique (Wieland, 1984). b ) effect of endogenous GIP on the removal of exogenous triglyceride. Four healthy male volunteers aged 23 - 39 years (mean = 28.8 + 3.5 SEM) were studied. They were all within 10 % of their ideal body weight and were not under any medication. On three separate occasions (at one week intervals), after an overnight fast, volunteers had page 200 antecubital veins in both arms cannulated. Immediately after a basal blood sample had been withdrawn from the left antecubital vein, they were given either glucose (75 g in 150 ml water), galactose (75 g in 150 ml water) or water (150 ml) orally. 15 minutes later Intralipid 20 % (50 ml, Kabi Vitrum) was infused at the rate of 5 ml/min into t-he right antecubital vein. Blood samples were taken from the left antecubital vein at 15 minute intervals, from 15 minutes before the oral load until 2 h after the i.v. load. Blood for glucose analysis was collected into fluoride oxalate tubes, but lithium heparin tubes for IRI (details in chapter 3) and IR-GIP analysis. Serum was collected for triglyceride estimations. Samples were stored at -20"C until analysis. Results were subjected to Student's t-test for paired data or analysis of variance (ANOVA), as appropriate. 7.1.2 Results a) measurement of free glycerol and triglyceride following an intravenous Intralipid load. Figure 7.1 shows that intravenous Intralipid has no effect on plasma glucose concentrations (ANOVA F (15,3) = 1.2824, n.s.). Figure 7.2 shows that intravenous Intralipid has no stimulatory effect on plasma IR-GIP, but a slow and significant downward trend is seen during the course of the experiment (ANOVA F(3,27) = 22.13, p < 0.01). Figure 7.3 shows that, not surprisingly, following intravenous Intralipid there is an increase in plasma triglycerides. However, figure 7.4 shows that following intravenous Intralipid there is no change at all in plasma glycerol (ANOVA F(9,27) = 1.831, n.s.). There was, page 201 however quite a large inter-personal variation in the glycerol results which could, at least partially, account for the large standard errors (between subjects, ANOVA F (3,27) = 7.927, p < 0.001). b ) effect of endogenous GIP on the removal exogenous tr i glyceride. Figures 7.5 to 7.7 show the triglyceride, glucose, IRI and IR-GIP results for the oral water, galactose and glucose loads, respectively. The IRI response to glucose was maximal at 45 minutes (70.1 + 27.9 mU/L, mean + SEM) and was significantly greater (p < 0.05) than the modest peak at 15 minutes for galactose (19.9 + 10.3 mU/L). The insulin peak following galactose was not significantly elevated from basal concentrations, and was almost entirely accounted for by one of the subjects. The plasma glucose concentrations showed no significant changes following oral water or galactose. The mean peak IR-GIP response was significantly greater (p < 0.05) after glucose than it was after galactose (2597 + 557 pg/ml at 75 minutes vs. 1536 + 216 pg/ml at 15 minutes). The triglyceride results show that following oral water, glucose and galactose, peak serum triglyceride values of 3.63 + 0.42, 3.33 + 0.26 and 3.73 + 0.23 mmol/L (mean + SEM) are achieved respectively, from basais of 0.93 + 0.26, 0.83 + 0.20 and 1.13 + 0.09 respectively. The curves are not significantly different from each other, in any permutation, as assesed by paired t-test or ANOVA. page 202 Figure 7.1 Plasma glucose concentrations in normal human volunteers ■following an intravenous infusion of Intralipid 20 7. (50 ml at 5 ml/ml) at time = O minutes, (n = 4, mean + SEM). 4.9 4.7 ®4.5. 4.3 -15 0 30 60 90 120 Time (min) page 203 Figure 7.2 Plasma IR-GIP concentrations in normal human volunteers following an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = O minutes, (n = 4, mean + SEM). 450 O) Cl O 300 c o Ü a CD 150 -15 0 30 60 90 120 Time (min) page 204 F igure 7.3 Plasma triglyceride concentrations in normal human volunteers following an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = 0 minutes, (n = 4, mean + SEM). 2.5 g 2.0 o Ü m 2L- 1.5 (D Ü ” 1.0 1 1 J -15 0 30 60 90 1 2 0 Time (min) nanp 205 Figure 7.4 Plasma free glycerol concentrations in normal human volun teers following an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = 0 minutes, (n = 4, mean + SEM). .13 Ü c o Ü Ô.08 L_ O) O ,03 -15 0 30 60 90 120 Time (min) page 206 Figure 7.5 Plasma glucose, IRI , IR—GIP and serum triglyceride concentrations following oral water (150 ml at time = -15 minutes) and an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = O minutes in normal human volunteers, (n = 4, mean + SEM). mM Serum Triglyceride 0 8 mM Plasma CUucose -1 80 r mUL Plasma IR-InsuUn 40 I— --- 1----- 1---- - 4-----4- 2600 ngL-1 Plasma IR-GIP 1400 200 -30 -15 15 30 45 60 75 90 105 120 min. TIME AFTER ORAL LOAD page 207 Figure 7.6 Plasma glucose, IRI , IR-GIP and serum triglyceride concentrations following oral glucose (75 g in 150 ml water at time = -15 minutes) and an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = O minutes in normal human volunteers, (n = 4, mean + SEM). mM Serum Triglyceride mM Plasma Glucose r mUL Plasma IR-Insuiin 0 2600 Plasma IR-GIP 1400 200 -:50 - 1 9 15 10 45 f.U 75 91) 105 120 min. TIME AFTER ORAL LOAD page 208 Figure 7.7 Plasma glucose, IRI , IR-GIP and serum triglyceride concentrations following oral galactose (75 g in 150 ml water at time = -15 minutes) and an intravenous infusion of Intralipid 20 % (50 ml at 5 ml/ml) at time = O minutes in normal human volunteers, (n = 4, mean + SEM). U mM Serum Triglyceride 2 0 mM Plasma Glucose -1 80r mUL Plasma IR-lnsulin 40 4----- 1 0 -1 2600 ngL Plasma IR-GIP 1400 200 90 107 120 mm. -JO -15 15 30 45 60 75 TIME AFTER ORAL LOAD page 209 7.1.3 Discussion The results from experiment (a) are in accord with the results of Cooke and Buckhart (1983). They demonstrate that under the conditions of this experiment, following intra venous Intralipid no significant increase in plasma glycerol occurs. Basal glycerol concentrations were about 0.08 mmol/L, whereas basal triglycerides were about 0.85 mmol/L; thus, free glycerol represents less than 10 % of the apparent basal triglyceride concentration. At the plasma triglyceride peak at 15 minutes the contribution made by free glycerol drops to less than 4 %. One can, therefore, draw the conclusion that separate estimations for free glycerol are unnecessary when investigating the removal of intravenously infused Intralipid, as the correction obtained would probably be insignificant and perhaps smaller than the combined error of the triglyceride and glycerol assays. The results from experiment (b) disagree with the results of Wasada et al. (1981). They demonstrated that using re-infused chyle, exogenous GIP was able to reduce the peak and augment the removal of plasma triglyceride. The results are, in fact, in agreement with the results of Ohneda et al., (1983) who, in a very similar experiment, also demonstrated that endogenous IR-GIP stimulated by oral carbohydrate was ineffective in reducing plasma levels of intravenously infused triglyceride emulsion in dogs. In the experiment of Wasada et al., (1981) chyle collected from the thoracic duct was used as the chylomicron source. The chylomicra were not purified in any way, save filtration to remove "clots". The chyle was not assayed for the presence of gut hormones. Perhaps a gut hormone in the page 210 chyle synergised the action of GIP, or perhaps natural chylomicra contain an unidentified gut factor which can synergise the action of GIP. In this way one could explain why the removal of natural chylomicra, but not Intralipid derived chylomicra, is augmented by GIP. The lack of effect of IR-GIP on the removal of plasma triglyceride in this experiment could perhaps be due to a swamping of the GIP stimulatable lipid clearing system at the adipocyte, and that lipoprotein lipase had been stimulated by another means, so that the presence or absence of GIP was immaterial. Oral fat is known to produce a larger IR-GIP response than oral carbohydrate (Brown, 1982). The lack of OIP action might, therefore, be due to inadequate plasma IR-GIP levels. Indeed, if one looks at the glucose results, by the time plasma IR-GIP concentrations have peaked, triglyceride levels have already returned to basal. Perhaps the experiment would have been better designed if the Intralipid load had been delayed for a further 60 minutes, to coincide with the IR-GIP peak. The above results might be due to the intervention of another carbohydrate stimulated gut hormone such as, perhaps, glucagon-1ike-immunoreactivity (GLI). GIP inhibits the action of glucagon at the adipocyte (Dupré et al., 1976), so it is possible that GLI inhibits the action of GIP at the adipocyte. Finally, perhaps the early assertion that carbohydrate and fat stimulate different IR-GIP s which have different actions is correct (Brown et al.. 1975). In conclusion, under the conditions of this experiment, carbohydrate stimulated IR-GIP is ineffective in reducing serum triglyceride concentrations following an intravenous page 211 intralipid load. Under other conditions, such as when the peak IR-GIP concentrations and peak triglyceride concentrations coincide (as detailed above), the result might be more positive. Perhaps the best experiment would be if the investigation were repeated in pigs with purified pig chylomicra and exogenous porcine GIP. SUMMARY OF RESULTS 1) The separate estimation of free glycerol is unnecessary in the study of the removal of triglyceride from the circulation following intravenous Intralipid. 2) Under the conditions of this experiment endogenous carbohydate-stimulated GIP does not affect the removal of triglyceride following intravenous Intralipid. page 212 CHAPTER EIGHT IMMUNOREACTIVE GASTRIC INHIBITORY POLYPEPTIDE (IR-6IP) IN OBESE HYPERGLYCAEMIC (ob/ob) MICE. page 213 8. O. O The obese hyperglycaemic syndrome (gene symbol o^> arose in 1949 as a mutant within the V strain at the Jackson laboratory. Bar Harbor, USA (Ingalls et al., 1950). The syndrome is transmitted as a single autosomal recessive gene (chromosome 6, linkage group XI), and these obese mice have been widely studied as an animal model to investigate the aetiology and pathogenesis of obesity and diabetes (Herberg and Coleman, 1977; Bray and York, 1979). These mice exhibit hyperphagia, excessive fat deposition and hence obesity, the severity of which depend on interaction of the mutant gene with the background genome (Bailey et al, 1982). Increased insulin concentrations constitute a major pathogenic influence, promoting excessive triglyceride deposition and glucose intolerance through insulin resistance and pancreatic A-celi disfunction (Assimacopoulos-Jeannet and Jeanrenaud, 1976; Flatt and Bailey, 1981a; Flatt et al, 1982). In fed adult ob/ob mice the extent of the hyperinsulinaemia is related to both the quantity and composition of the food ingested (Bailey et al, 1982; Flatt and Bailey, 1982a, 1984). A direct insulinotropic effect of amino acids and possibly fatty acids contributes in this respect (Flatt and Bailey, 1982a,b, 1984). However, the importance of neural and endocrine components of the enteroinsular axis is highlighted by the presence of an insulin response to oral but not parenteral glucose, and the augmention of insulin response to feeding in conditioned mice (Flatt and Bailey, 1981a, 1983). The mice used were the product of a selective breeding programme to provide ob/+ and +/+ lean mice. The heterozygous page 214 breeding pairs were originally identified by unclassified lean (?/+) mice from litters containing obese individuals. The presence of obese animals among the progeny indicated that both parents were heterozygous. The mice were judged ob/+ from the matings then served for breeding purposes. Subsequently, ob/+ mice were identified by the presence of obese progeny after backcrossing to an established ob/+ mouse. The severity and developmental pattern of the diabetic syndrome in obese hyperglycaemic mice results from the interaction of the mutant gene with modifiers of the background genome (Flatt and Bailey, 1981b). The genetic background of the colony is detailed below. Original C57BL/6J heterozygous (ob/+) breeding pairs from the Jackson Laboratory, Bar Harbor, USA. were obtained in 1957 by Prof. D.S. Falconer of the Institute of Animal Genetics at the University of Edinburgh. Heterozygous mice were outcrossed at Edinburgh to two non-inbred local strains: to JH, selected for high litter size (Falconer, 1960a) and maintained as a closed non-inbred stock for five generations, and to CRL, selected for faster growth rate (Falconer, 1960b) and maintained for ten generations as a closed non-inbred stock. It was heterozygous mice from this latter stock which were outcrossed to two further non-inbred local strains, and the resultant ob/+ breeding pairs formed the colony at Aston. These mice were used to establish a closed non-inbred colony with a severity of diabetes between that of C57BL/6J and C57BL/KsJ ob/ob mice (Bailey et al., 1982; Herberg and Coleman, 1977). page 215 0 . 1.0 Immunoreactive Gastric Inhibitory Polypeptide (IR-GIP) in the plasma and small intestine of the obese hyperglycaemic mouse. Hyperphagia and hyperinsulinaemia are the prominent features in the pathogenesis of the syndrome in the mice (Flatt and Bailey, 1981a). Fed adult ob/ob mice show a marked insulin response to enteral but not parenteral glucose, indicating that the hyperinsulinaemia is due to the ingestion of nutrients (Flatt and Bailey, 1982a). In addition, the hyperinsulinaemia and hyperphagia follow similar age-related patterns, and plasma immunoreactive insulin (IRl) concentrations fall markedly when food is witheld (Flatt and Bailey, 1981a). The above observations direct attention towards the enteroinsular axis and therefore GIP (Creutzfeldt, 1979). Described below is an experiment to measure IR-GIP in the plasma and gut extracts of fed adult ob/ob mice. 8.1.1 Methods Obese hyperglycaemic (ob/ob) and lean (+/?) mice were provided by Dr. P.R. Flatt. They were bred in the animal breeding unit at the University of Surrey with stock from the Aston colony. The mice had been housed in an air conditioned room at 22 + 2*C with a light scedule of 12 h light (0800 h - 2000 h) and 12 h darkness. A standard pellet diet (Spratts laboratory diet no. 1, Lillico Ltd., Reigate, Surrey) and tap water were available ad libitum. Fed lean and obese mice page 216 (11tter-mates) were used at 5-6 months of age. They were anaesthetized under ether and bled by cardiac puncture. Blood was collected into paediatric Lithium heparin tubes, centrifuged immediately and the plasma aliquotted out for glucose, IRI and IR-GIP estimations. The mice were then killed (cervical dislocation) and the entire small intestine from the duodenal bulb to the ileo-caecal Junction (dudenum, jejunum and ileum) was excised, washed and perfused through with ice-cold normal saline (0.154 mM NaCl). Surplus liquid was removed (using fingers as a squegee), weighed, then the tissue extracted (scissor minced then sonicated for 10 minutes) with 5 ml/g acid ethanol (750 ml ethanol, 250 ml water and 15 ml HCl). Extracts and plasma samples were immediately frozen and kept at -20*C until analysis. Plasma glucose and IR-GIP were measured using the standard methods shown in chapter 2. The extracts were diluted with assay buffer and measured in an IR-GIP assay without the addition of hormone free plasma to the standard curve. Plasma IRl was measured by Dr. P.R. Flatt, using a dextran coated charcoal radioimmunoassay. Crystaline mouse insulin (Novo) was used as standard, having a sensitivity of 0.5 pg/ml and a within assay coefficient of variation for samples containing 0.1 to 24.5 ng/ml of 1.3 to 4.2% (Flatt and Bailey, 1981b). Results were subjected to Student's t-test for unpaired values. 8.1.2 Results The results in table 8.1 show that the obese mice when compared to their lean litter mates are significantly heavier 70.6 + 4.1 vs. 38.8 + 2.3 g (p < 0.001), moderately page 217 hyperglycaemic 13.3 + 1.8 vs. 7.4 + 0.3 mM (p < 0.02) and very hyperinsulinaemic 32.4 + 2.4 vs 2.1 + 0.05 ng/ml (p < 0.001; mean + SEM; n = 6-8). The obese mice had intestines which were 1.4 times heavier (p < 0.01) than lean mice; the intestinal concentration of IR-GIP was 1.9 times higher (p < 0.01) with the total intestinal IR-GIP being 2.8 times higher (p < 0.01). Consistent with these changes plasma IR-GIP concentrations were 15 times higher in the obese mice (p < 0.001). page 218 Table 8.1 IR-GIP in the plasma and small intestine of lean and obese mice. LEAN OBESE P< Body weight (g) 38.8 + 2.3 70.6 + 4.1 0.001 Plasma glucose (mM) 7.4 + 0.3 13.3 + 1.8 0.02 f^lasma IRI (ng/ml) 2. 1 ± 1.1 32.4 + 2.4 0.001 Plasma IR-GIP (pg/ml) 289 + 11 4350 + 877 0.001 Intestinal weight (g) 1.9 + 0.1 2.6 + 0.2 0.01 Intestinal IR-GIP (pg/g) 0.78 + 0.07 1.48 + 0.2 0.01 Total intestinal IR-GIP (pg) 1.45 + 0.14 4.12 + 0.61 0.01 [Values are mean + SEM, n = 6-83 page 219 8.1.3 Discussion The above results indicate that the small intestine of ob/ob mice have an enhanced rate of synthesis and secretion of IR-GIP. These data are consistent with previous reports suggesting an increased population of endocrine cells in the intestine of ob/ob mice (Polak et al., 1975; Best et al., 1977). As GIP is the prime incretin candidate (Creutzfeldt, 1979; Brown, 1982) the above results lead one to assert that GIP makes a substantial contribution to the hyperinsulinaemia and related syndromes of the ob/ob mouse. One must be mindful, t)owever, in the light of recent data that GIP is unlikely to be the sole incretin involved in an overactive enteroinsular— axis (Lauritsen et al.,1981; Ebert and Creutzfeldt, 1982). However, in view of the gross changes seen in IR-GIP concentrations, this hormone is at the very least a good candidate for further study in the enteroinsular-axis of the ob/ob mouse. page 220 8 .2.0 Involvement of GIP and the enteroinsular-axis in the metabolic abnormalities of the obese hyperglycaemic (ob/ob) mouse. Gastric Inhibitory Polypeptide (GIP) remains the most likely candidate for the role of incretin (La Barre, 1932) as the physiological component of the enteroinsular-axis (Unger and Eisentraut, 1969). GIP has been cited as the likely enteric messenger in the enteroinsular— axis in the dog (Pedersson and Brown, 1976; Takemura et al., 1982), the pig (Wolffbrandt et al., 1986) and in man (Elahi et al., 1979). In addition the use of anti-GIP sera has demonstrated that in the rat GIP is an incretin (Lauritsen et al., 1981) but it has also been shown, by adsorbing IR-GIP from rat gut extracts then testing for residual incretin activity, that it is not the only one (Ebert and Creutzfeldt, 1982). The only dissenting voices from those acclaiming GIP to be the most likely incretin candidate are those of Bloom and co-workers (Sarson et al., 1982) who claim that GIP is only an incretin in humans at pharmacological concentratons. However, work has recently been published that suggests that the antiserum used by Sarson et al (1980, 1982) only partially cross reacts with human GIP (Jorde, Burhol and Schulz, 1983; Krarup and Holst, 1984). Indeed, the stimulated levels of IR-GIP they obtain are approximately five times lower than those of Kuzio et al. (1974), Morgan et al. (1978), Lauritsen and Moody (1978) and Ebert, Ilmer and Creutzfeldt (1979). If one assumes that Sarson s porcine GIP assay cross reacts with human GIP only about one fifth as well as the other assays, then when they page 221 infuse porcine GIP into humans they infuse five times too little to reach normal stimulated levels. It is not surprising that they claim GIP to have only pharmocological effects as an incretin (Sarson et al., 1982). The experiments below seek to investigate the regulation of IR-GIP in ob/ob mice, and evaluate the involvement of GIP in the hyperinsulinaemia and associated features of the syndrome. 8.2.1 Methods Obese hyperglycaemic (ob/ob) mice and lean (+/?) mice as detailed above were used in the study. They were provided by Dr. P.R. Flatt from stock bred in the animal breeding unit at the University of Aston. The mice were housed in an air conditioned room at 22 + 2'C with a lighting scedule of 12h light (0800 h - 2000 h) and 12 h darkness. A standard diet (Mouse breeding diet, Heygate and Sons Ltd., Northampton) and tap water were available ad libitum. The standard diet comprised 2.5 % fat, 17.6 % protein and 46.8 7. carbohydrate (digestible energy 87.6 kJ/kg). The mice were adapted to the manipulative procedure for two weeks before the study. Blood samples were obtained from the cut tip of the tail in the conscious state at the appropriate time points. Plasma was separated and stored as in the previous experiment (section 8.1.1). a) Basal plasma IR-GIP concentrations with age. To investigate changes in plasma IR-GIP with age in relation to other features of the ob/ob syndrome, blood samples were taken from fed ob/ob mice at 3, 5, 10 and 20 weeks of age, page 222 and in lean mice at 3 weeks of age. b) The responses to fasting and refeeding. The effects of fasting for 24h and refeeding with a standard diet available ad libitum was examined in ob/ob mice at 10 weeks of age. The food intake was monitored in this experiment. Blood samples were taken at -24, 0, 2, 4, and 6h. c) Responses to fat, glucose and amino acids, ob/ob mice fasted for 18 h at 10 - 12 weeks of age were used. Food but not water, was witheld during the test. The following nutrients were administered orally (gavage) in a volume corresponding to 8 ml/kg body weight: (i) a fat emulsion (Intralipid; Kabi Vitrum Ltd., Ealing) containing 770 mg fractionated soya bean oil/kg, 87 mg glycerol/kg and 46 mg fractionated egg lecithin/kg body weight (32.2 kJ/kg); (ii) 2g glucose/kg body weight (32.2 kJ/kg); (iii) a mixture of essential and non-essential amino acids (Synthamin 17; Travenol Laboratories Ltd., Thetford, Suffolk) containing 0.8g L-amino acids/kg body weight (132.5 mg nitrogen/kg), and (iv) a fat/glucose/amino acid mixture (8 ml/kg body weight) comprising equal volumes of the three preparations. A basal blood sample was taken, and further blood samples at 30, 60, and 120 minutes following the oral load. d ) Effects of i.p. glucose on the response to oral fat. ob/ob mice fasted for 18h at 10-12 weeks of age were used. Food, but not water, was witheld during the test. Fat emulsion (as above in (c)(i)) was administered orally (gavage) at time zero and glucose (2 g/kg body weight) or an equivalent volume (5 ml/kg) of normal saline (0.154 mM NaCl) page 223 was injected i.p. at 60 minutes. A basal blood sample was taken at time zero and further samples at 30, 60 and 90 minutes. e) Effects of glucose and insulin on basal IR-GIP. ob/ob mice fasted for 18 h at 10 - 12 weeks of age were used. Food, but not water, was witheld during the test. Three groups of mice each received, by i.p. injection, in a volume of 5 ml/kg body weight either (i) glucose (2g/kg>, (ii) low dose insulin (5 units/kg) or high dose insulin (Actrapid; Novo: 100 units/kg). A basal blood sample was taken and further samples at 30 and 60 minutes. f) Effects of insulin on responses to fat and glucose. ob/ob mice fasted for 18 h at 10 - 12 weeks of age were used. Food, but not water, was witheld during the test. Sets of mice received oral doses (gavage) at time zero of either (A) glucose (2g/kg) or (B) fat emulsion (770 mg/kg) in a volume corresponding to 8 ml/kg body weight. Subsequently, at 30 minutes after the oral load, they received either (i) saline (0.154 mM NaCl), (ii) low dose insulin (5 units/kg) or (iii) high dose insulin (100 units/kg) by i.p. injection in a volume equivalent to 5 ml/kg body weight. A basal blood sample was taken and further blood samples at 30, 60 and 90 minutes after the oral load. g ) Effects of exogenous GIP with and without glucose. ob/ob mice fasted for 18 h at 10 - 12 weeks of age were used. Food, but not water, was witheld during the test. Two groups of mice received an i.p. injection of pure porcine GIP (40 pg/kg; J.C. Brown, batch EG III) either (i) alone or (ii) page 224 with glucose (2g/kg) in a total volume of 8 ml/kg body weight. Results were assessed using Student's paired or unpaired t-tests as appropriate. IRI and assays were performed as in the previous experiment (8.1). 20 pi plasma samples were assayed for IR-GIP as detailed in chapter 2. 8.2.2 Results a) Basal plasma IR-GIP concentrations with age. Plasma glucose, IRI and IR-GIP results are shown in figure 8.1. The results indicate that the onset of hyperinsulinaemia precedes the development of hyperglycaemia; the IRI levels were significantly elevated (p < 0.05) by five weeks but plasma glucose levels were not significantly increased (p < 0.05) until 10 weeks. Plasma IR-GIP concentrations, like those of IRI, were significantly higher at 10 - 20 weeks (p < 0.05) compared to the levels at 3 - 5 weeks. It was shown in the previous experiment (8.1) that the IR-GIP concentrations of adult ob/ob mice are considerably higher than those of age-matched lean (+/?) mice. However, in 3-week old weaning mice IR-GIP concentrations were not significantly different between the obese (782 + 163 pg/ml; mean + SEM, n = 9) and lean mice (752 + 188 pg/ml; n =7) , although plasma insulin concentrations were already raised in obese mice at this age (4.3 + 0.7 pg/L and 1.4 + 0.5 pg/L in obese and lean mice respectively). Interestingly, the IR-GIP concentrations of weaning lean mice were higher than those of adult mice from the previous experiment (8.1); this may reflect the relatively high proportion of fat and galactose in the milk. page 225 b) The responses to fasting and refeeding. Plasma glucose, IRI and IR-GIP, together with food intake results are shown in figure 8.2. Plasma IR-GIP concentrations fell during fasting, coinciding with a marked reduction of insulin and a decline in glucose. Refeeding with a standard diet available ad libitum evoked prominent and sustained increases in the plasma concentrations of all three parameters (p < 0.05) compared to the zero time point. c ) Responses to fat, glucose and amino acids. Plasma glucose, IRI and IR-GIP results are shown in figure 8.3. Oral fat proved to be the best stimulus to plasma IR-GIP. Plasma glucose and IRI concentrations tended to drift upward during the course of the experiment, but remained far below plasma levels expected in the fed state. Oral glucose was a moderate stimulus to IR-GIP, but neither IR-GIP nor IRI concentrations resembled those of the fed animal. The plasma glucose levels attained, on the other hand, were elevated to grossly hyperglycaemic levels. The rate of decline of the plasma glucose levels was very slow, and these remained elevated for the remainder of the experiment. Oral amino acids yielded a plasma IR-GIP response which was roughly equivalent to that of oral glucose. However, plasma glucose levels showed only a slight tendency to rise during the course of this experiment, resulting in only a modest rise in plasma IRI with a transient peak at the 30 minute time point. When all three nutrients were mixed and administered orally the resulting plasma IR-GIP levels were equivalent to that seen following oral glucose or amino acids (but still below peak fed' concentrations). There was also (for these animals) slight page 226 hyperglycaemia, together with a greatly augmented plasma IRI response which approached levels seen in the fed state. d ) Effects of i.p. glucose on the response to oral fat. Plasma glucose, IRI and IR-GIP results are shown in figure 8.4. Oral fat alone did not significantly elevate plasma glucose or IRI concentrations, although a slight upward trend in IRI was seen during the course of the experiment. Consistent with the results of the last section (c), the combined action of stimulated levels of IR-GIP with hyperglycaemia caused a large increase in IRI concentrations, to the fed level. Plasma IR-GIP concentrations were not significantly different at any time point. It would appear that neither the hyperglycaemia nor the hyperinsulinaemia in the second half of the experiment had any effect on plasma IR-GIP concentrations. e) Effects of glucose and insulin on basal IR-GIP. The plasma glucose and IR-GIP concentrations are shown in figure 8.5. The plasma glucose results show that both low and high dose i.p. insulin cause significant hypoglycaemia, and i.p. glucose causes marked hyperglycaemia. Neither hyperglycaemia nor hyperinsulinaemia together with hypoglycaemia had any effect on fasting IR-GIP concentrations. f) Effects of insulin on responses to fat and glucose. Plasma glucose and IR-GIP results are shown in figure 8.6. Consistent with the results of the previous experiment (e), i.p. insulin was effective in suppressing basal plasma glucose following oral fat, and also suppressed plasma glucose following an oral glucose load. However, together page 227 with the lack of effect on basal IR-GIP concentrations shown in the previous experiment (e), doses of insulin up to 100 units/kg failed to affect either fat- or glucose-stimulated IR-GIP release. g > Effects of exogenous GIP with and without glucose. Plasma glucose and IRI results are shown in figure 8.7. In the presence of a basal hyperglycaemia, exogenous GIP evoked a marked, but transient, IRI response without change in plasma glucose concentrations. In the presence of rising hyperglycaemia after glucose administration, GIP produced a similarly marked but considerably protracted IRI response. Concentrations of IR-GIP achieved during these experiments (mean values 654, 3383, 5175, 4955 and 2035 pg/ml at O, 5, lO, 15, and 30 minutes respectively) were similar to those observed in actively feeding obese mice in the experiments detailed in section 8.1. Plasma IR-GIP concentratons fell to 693 pg/ml by 90 minutes. page 228 Figure 8.1 Changes with age in plasma concentrations of glucose, IRI and IR-GIP in fed ob/ob mice. (n = 7 - 10, mean + SEM; * = p < 0.05 compared with 3 and 5 weeks of age, + = p < 0.05 compared with 3 weeks of age. Student's unpaired t—test >. O C\J O LO oo o o o o o o o LO o LO ( l / 6 u ) dI0 Biuseid o {/) C\J LO O) cn ro C o o oo CO CXI ([/5rf) UL[nsuL euiseLd O (XI --- LO CO oo o o 00 c n (XI (L P /6ui) asoon[6 emsc[d page 229 Figure 8.2 Effect of a 24 h fast and refeeding with a standard diet available ad libitum on plasma glucose, IRI and IR-GIP in 10 — 12—week—old obese ob/ob mice, (n = 8, mean + SEM). All values are significantly (p < 0.05) raised compared with time zero (Student's paired t-test). Mean values for food intake are shown. FASTING REFEEDING 270 (U in 180 O o Z3 nJ E roto 51 cn c c 3000 2000 CL O 1000 CL ro X> +J O CM C - Û •I— 4-> cnx: x j jkJ cn -24 Time (hours) page 230 Figure 8.3 Effects of orally administered (a) fat emulsion, (b) glucose, (c) mixture of a, b and c, on plama glucose, IRI and IR-GIP in 10 - 12—week-old obese ob/ob mice fasted for 18 h. (n = 6, mean + SEM; * = p < 0.05 compared with zero. Student's paired t-test). o o o o o o o o o O o 00 cvj o o § O u o o o o o o o o o o oo CO o o o o ([p/Biu) dsoon[B ewse[d ( l/6 it ) ui^nsui eiuseid ( l/6 u ) d I9 Btuseid page 231 Figure 8.4 Effect of I.p. glucose (•) or saline (o) (at 30 minutes) on fat-stimulated (oral, at zero minutes) plasma glucose, IRI and IR-GIP in 10 - 12-week-old obese ob/ob mice fasted for 18 h. (n = 6, mean + SEM; * = p < 0.05 compared with saline treated mice. Student's unpaired t-test). 720 "O CD E 540 360 CD 180 Q- CD 3 c =3 to C ro E CL 2000 cn c CL CD 1000 CL 0 30 60 90 Time (min) page 232 Figure 6.5 Effect of i.p, glucose (2 g/kg; #), low dose insulin (5 U/kg;A) or high dose insulin (100 U/kg; «) on plasma glucose and IR-GIP in 10 - 12-week-old obese ob/ob mice fasted for 18 h. (n = 6, mean + SEM). All plasma glucose levels at 30 and 60 minutes are significantly different from time zero (p < 0.05, Student's paired t-test). 720 ■o cn 540 E (D CO O 180 O 3 CD fO E CO 05 CL 1000 c CL 500 »— I CD 05 E CO fd Time (min) page 233 Figure 8.6 Effect of saline (0.154 mmol NaCl/L; •), low dose insulin (5 U / k g ; A > or high dose insulin (100 U/kg; # ) administered i.p. at 30 minutes, on plasma glucose and IR-GIP after stimulation by (a) fat emulsion and (b) glucose administered orally at time zero in 10 - 12-week-old obese ob/ob mice fasted for 12 h. (n = 6, mean + SEM). There were no significant differences of plasma IR-GIP values between the three groups of mice. - o m<— -o- - < M UDo o ro "■m- o oo o o o o c o o o o o o 00 o o C\J o -«A cn -O l o o o o o o o o o o o o o o (lP/6ui) 9S0Dn[6 eiusp[d (l/6u) dI9 ewseid page 234 Figure 8.7 Effect of I.p. porcine GIP (40 pg/kg) (alone and together with glucose (2 g/kg) on plasma glucose and IRI in 10 - 12-week-old obese ob/ob mice fasted for 18 h. (n = 4, mean + SEM; * = p < 0,05, Student's paired t-test). o ro If) O O O O O O O O CM CM 00 LO ro cu O ro LO LO o o o o o O o CM LO 00 ro ( lp/6iu) 9SODn[6 çiuse[d ([/6ir) uL[nsuL GmsPLd page 235 8,2.3 Discussion Raised plasma IR-GIP concentrations have been reported in some cases of human obesity, untreated type I (insulin- dependent) diabetes and type II diabetes (Bloom, 1975; Crockett, Mazzaferri and Cataland, 1976); May and Williams, 1978; Ross and Dupré, 1978; Deschamps et al., 1980; Brown 1982; Salera et al., 1982; Creutzfeldt et al., 1983), and in certain spontaneously obese-diabetic rodents, namely glycosurie Djungarian hamsters (Buchanan et al., 1982) and both ob/ob (Section 8.1) and db/db (Flatt et al., 1983) mice. Work is currently underway to confirm the role of GIP in obese Zucker fa/fa rats (Beck and Max, 1986). A plausible reason for raised IR-GIP concentrations in ob/ob mice rests, in part, with intestinal hypertrophy associated with an increase in the number and hormone content of GIP secreting cells (Polak et al., 1975; section 8.1). In addition it would appear that hypersecretion results from excessive alimentation and abnormal regulation of GIP cell function. The results in section (a) show that the onset of hyperglycaemia preceded the develpment of both hyperglycaemia and hyperGIPaemia. This is in agreement with previous observations in ob/ob mice (Bailey et al., 1982) and thus suggests the involvement of other factors in the aetiology of the hyperinsulinaemia and related disorders. In (b) the results again confirm previous data correlating feeding with hyperinsulinaemia in ob/ob mice (Bailey et al., 1982) and extends the relationship to include IR-GIP concentrations. The IR-GIP results demonstrate that in the ob/ob mouse, as in the normal human, in the fed state, plasma IR-GIP concentrations remain significantly elevated (Jorde et al., page 236 1980; Jones et al., 1985). Over all, fat is by far the best stimulus to IR-GIP (section c). It may be possible to infer, from the fat data in particular, that the ob/ob mice have a modified or defective gastric emptying, in that oral fat produces a peak in plasma IR-GIP by at least 30 minutes. This is faster than in the human (section 4.2) and certainly faster than in the rat (section 4.1) where, following oral fat, peak plasma IR-GIP concentrations are attained after 2 - 3 hours in both controls and fat-pretreated rats. It would thus appear that the ob/ob mouse has an accelerated gastric emptying and an enhanced fat absorptive capacity compared to the rat. However, as the results from the glucose/amino acids/mixed nutrients experiments do not appear to be particularly accelerated or atypical, then perhaps the fat results might be more effectively accounted for by a defective enterogastrone. Both glucose and amino acids were effective stimuli for IR-GIP. However, as only glucose caused marked and sustained hyperglycaemia, it was glucose which gave the most prolonged IRI release of the individual nutrients. These results are consistent with previous data (Brown, 1982; Creutzfeldt et al., 1983). On combining the nutrients a modest rise in IR-GIP was seen (equivalent to that after glucose or amino acids) and a hyperglycaemia, which was less than that following glucose but greater than that following amino acids. The IRI release, however, was greatly augmented, thus suggesting the action of an incretin. Studies in man and laboratory animals have suggested that under normal physiological conditions, feedback inhibition by insulin on intestinal K-cell function plays a role in the regulation of basal and nutrient stimulated IR-GIP concentrations (Brown, 1982; Creutzfeldt et al., 1983; page 237 Marks and Morgan, 1984). The results show that both -Fat and glucose stimulated and basal IR-GIP levels are totally refractory to any inhibition by either hyperglycaemia and/or hyperinsulinaemia (sections d, e and f). Indeed, in these ob/ob mice, glucose induced increases in endogenous insulin and amounts of exogenous insulin (25 to 500 times those used in the rat and human experiments in chapter 4) which would be fatal in lean mice, did not suppress either basal or fat- or glucose-stimulated IR-GIP release. The final section (g) confirmed that insulinotropic action of exogenous GIP is potentiated by glucose (Pederson and Brown, 1976; Elahi et al., 1979; Brown et al., 1980; Szecowka et al., 1982). Although the physiological action of GIP in normoglycaemic man is equivocal (Verdonk et al., 1980; Sarson et al., 1982), the above results show that the hypersecretion of GIP in response to feeding in ob/ob mice is a major insulinotropic stimulus in the presence of hyperglycaemia. Additional contributory factors, however, include direct and glucose-potentiated islet stimulation by amino acids and neuroendocrine components of the enteroinsular axis other than GIP (Bobbioni and Jeanrenaud, 1982; Ebert and Creutzfeld, 1982; Flatt and Bailey, 1983). Thus, GIP participates in the metabolic abnormalities of ob/ob through hyperinsulinaemia (Assimacopoulos-Jeannet and Jeanrenaud, 1976). However, other actions of GIP such as stimulation of glucagon secretion (Unger et al., 1967) inhibition of gastric acid secretion (Brown and Pederson, 1970) and augmentation of lipogenesis in adipose tissue (Eckel et al., 1979; Wasada e;t a 1., 1981) are also likely to be involved in the manifestation of the ob/ob syndrome. page 238 8. 3.0 Effect of chain-length and saturation in fatty acid stimulated IR-GIP release in the obese hyperglycaemic (ob/ob) mouse. While it was established some time ago that triglyceride was a most potent stimulus to IR-GIP release in many species including man (Brown et al., 1975; O Dorisio et al., 1976) and dogs (Falko et al., 1975), the mechanism of fat stimulated IR-GIP release has recg* ved only poor attention. Ross and Shaffer (1981) have carried out an investigation in humans and Williams et al (1981) in dogs. Both studies showed triglyceride to be a potent stimulus to IR-GIP but gave variable results with fatty acids and other digestion products. The aim of this experiment was to establish the precise requirements for fatty acid stimulated IR-GIP release. The previous set of experiments (8.2) showed fat to be an excellent and most rapid stimulus to plasma IR-GIP release in the ob/ob mouse, a much better stimulus than in the rat (section 4.1). In using the ob/ob mouse as the model for the following experiments one is assured as one can be of using a system where possible interactions of such variables as gastric emptying and fat absorption are optimized and thus their influence in modifying the results are at a minimum. 8.3.1 Methods Obese hyperglycaemic (ob/ob) mice were provided by Dr. P.R. Flatt. They were bred in the animal breeding unit at the University of Surrey with stock from the Aston colony. The page 239 mice had been housed in an air conditioned room at 22 + 2 °C with a light schedule of 12h light (0700 h - 1900 h) and 12h darkness. A standard pellet diet (Spratts Laboratory Diet No. 1., Lillico Ltd., Reigate, Surrey) and tap water were available ad libitum. The mice were adapted to manipulative procedures for two weeks before the study. Ater an overnight fast (18 h), 16 ob/ob mice, between 12 and 18 weeks of age, were randomly divided into three groups of 5 or 6 mice. Each group first received one fatty acid in one experiment and then after eight days another fatty acid in a further experiment. After a fasting blood sample each mouse received an aqueous emulsion/suspension (oral gavage) of a fatty acid (2.62 mmol/kg, at 7.5 ml/kg body weight). The amount of fatty acid administered was calculated to be equivalent to the dose used in previous mouse experiments (section 8.2). As fatty acids were being used, rather than triglyceride, a three-fold increase in molar dose was required in this experiment. The dose and nature of each fatty acid administered is shown below. Fatty acid Chain length:double bonds Dose mg/kg Prop ionic C3: 0 194 Capric CIO: 0 451 Stear ic C18: 0 746 Oleic C18: 1 741 Linoleic C18:2 735 Linolenic C18: 3 730 page 240 Blood samples were obtained from the cut tip of the tail in the conscious state at the appropriate time points (O, 30, 60, 90 and 120 minutes). Blood was collected into heparinized tubes. Plasma was separated and stored as in the previous experiments (section 8.1). Plasma was assayed for IR-GIP as detailed previously (section 8.2) and for triglyceride using the enzymic method described earlier (chapter 2). Results were subjected to Student's t-test for paired and unpaired observations, as appropriate. 8.3.2 Results Plasma IR-GIP and triglyceride results for propionic, capric, stearic, oleic, linoleic and linolenic acids are shown in figures 8.8 and 8.9. Oral propionic acid gave a slight but not significant increase in plasma IR-GIP and caused no change in plasma triglyceride concentrations. Oral capric acid caused no change in plasma IR-GIP concentrations but did cause a significant and sustained decrease in plasma triglycerides, the lowest concentration being at 30 minutes (p < 0.01), with the plasma concentrations again showing significance at 90 and 120 minutes (p < 0.05). Oral stearic acid had no significant effect on plasma triglyceride concentrations, although the results do show a slight downward trend over the course of the experiment. IR-GIP results show an almost biphasic response: highest at 120 minutes but only reaching significance (p < 0.05) at the 60 minute time point. Oral oleic acid did not have a significant effect on page 241 plasma triglyceride values although a peak was reached at 90 minutes (4.17 + 0.76 mM vs. 2.69 + 0.54 mM for the zero point; mean + SEN). The plasma IR-GIP results again showed a biphasic rise, reaching a peak at 60 minutes (2158 + 571 pg/ml) but showing significant elevations over zero only at 90 and 120 minutes (p < 0.02 and 0.01 respectively). The stimulated values obtained with oleic acid for the 90 minute point were significantly higher than those observed at the same stage, for propionic, capric and stearic acids (p < 0.05, 0.01 and 0.05 respectively). They were also higher than the 120 minute point for propionic and capric acids (p < 0.01). Following oral linoleic acid plasma triglycerides reach a peak at 30 minutes (p < 0.05) and then decline for the remainder of the experiment. Plasma IR-GIP concentrations reach a peak after 60 minutes (2336 + 792 pg/ml) but attain sigificance only at the 90 and 120 minute time point (p < 0.02). The plasma concentrations obtained at the 90 minute time point were significantly different to those obtained for capric acid (p < 0.05), and those at 120 minutes to capric and propionic (p <0.01 and 0.02 respectively). Plasma IR-GIP profiles following oral oleic and linoleic acids were practically superimposable. Following oral linolenic acid no significant change in plasma triglyceride occurred. Plasma IR-GIP concentrations showed a more gradual rise than for either oleic or linoleic acids. Plasma concentrations were significantly different from zero at 90 and 120 minutes (p < 0.05 and 0.01 respectively), and these time points were also significantly different to those observed for capric acid (p < 0.05). Mean incremental areas under the curve for IR—GIP page 242 -Following propionic, capric, stearic, oleic, linoleic, and linolenic acids were 691 + 350, 216 + 281, 1348 436, 2389 ^ 382, 2550 + 691 and 1508 + 665 (mean + SEM pg/ml. h) respectively. Thus the oleic acid IR-GIP response was significantly greater (p < 0.01) than that for stearic, capric or propionic acids but not significantly different to that for linoleic or linolenic acids. page 243 Figure 8.8 Plasma IR—GIP results following oral aqueous fatty acid emulsion (2.62 mmol/7.5 ml/kg) in IB h fasted ob/ob mice, (n = 5-6, mean + SEM, * = p < 0.05). o CM o cn -o o VjD o CO cz +-> CO # # o o CM o CD to OJ +->CJ o c o CO Q l c C_) o o CT> O CO o CO CL Qt- O O O O o o o o o O O o o o o O O o CVJ ro CM ( [iu/6d) ewseid page 244 Figure 8.9 Plasma IRI results -following oral aqueous fatty acid emulsion (2.62 mmol/7.5 ml/kg) in 18 h fasted ob/ob mice, (n = 5-6, mean + SEM, * = p < 0.05). o OJ TD o o ro ca c . o CO o OsJ cn (/) O) o 4-> •o 3 O) O l c o o o CO o cn o o CO cx O l. o i o LO CO OsJ ([/[ouiiu) emsPLd page 245 8.5.3 Discussion Ross and Schaffer (1981) showed using children suffering from cystic fibrosis, that triglyceride hydrolysis (and hence presumably absorption) is required for IR-GIP secretion. The same conclusions can be drawn from studies with patients suffering from chronic pancreatitis (Ebert and Creutzfeldt, 1980). Ross and Shaffer (1981) further demonstrated that medium chain triglycerides, monoglyceride (which was solid and hence probably poorly absorbed) and glycerol do not stimulate IR-GIP release. They also showed that equimolar amounts of triglyceride (corn oil; predominantly the glycerides of oleic and linoleic acids) were three times more potent in stimulating IR-GIP release than oleic acid (in terms of peak response). However, area under the curve data showed triglycerides to be just over twice as effective as the long chain fatty acid. Williams et al (1981) were unable to show IR-GIP stimulation by monoolein (probably the, 1-monoglyceride), palmitic acid, glycerol, oleic acid and linoleic acid! Creutzfeldt and Ebert (1985) have described experiments where the surfactant Pluronic L-81, an agent which does not impair triglyceride absorption, but blocks the formation of chylomicrons within enterocytes, completely abolished the IR-GIP response to intraduodenaal fat in rats. The above results show that the short chain propionic and medium chain capric acids do not stimulate IR-GIP release. The saturated long chain stearic acid is a moderate stimulus, whereas its unsaturated analogues (oleic, linoleic and linolenic acids) were all excellent stimuli to IR-GIP release. The results show, therefore, that only fatty acids which are ester if led to triglyceride within the enterocyte page 246 (Clément, 1980) are able to stimulate IR-GIP release. Short and medium chain fatty acids are not ester ified following absorption and are transported away from the small intestine, predominantly in the portal vein, as free fatty acids. The degree of saturation of the long chain fatty acids is important. Saturation influences the physical properties of the fatty acid, which in turn influence the rate of absorption and hence the magnitude of the TR-GTP response. Stearic acid has a melting point of 70*C which is above body temperature, and it is thus relatively less well absorbed than its unsaturated analogues which are all liquid at body temperatures. They are as triglycerides, more readily available to pancreatic lipase for lipolysis and micelle formation. The result that the series of unsaturated fatty acids (oleic, linoleic and linolenic : all CIS, but varying in degree of saturation from one to three double bonds, respectively) showed no difference in TR-GTP releasing potential suggests that following absorption no further differences, due to saturation, exist. Thus, once a long chain fatty acid has been rendered soluble enough to be readily absorbed across the mucosal surface of the enterocyte, individual chemical differences are no longer important, merely the ability to be incorporated into tr iglyceride. The main difference, in this context, between medium and long chain fatty acids is the estérification step. This estérification (or in the case of triglycerides, re-esterification) is an active, energy requiring process (Clément, 1980). Bearing all the above in mind, therefore, the primary requirement for the release of IR-GIP by fatty acids is the "metaboloic flux" associated with triglyceride page 247 estérification. This conclusion is not inconsistent with the findings of Creutzfeldt and Ebert (1985) with Pluronic L-81. Should chylomicron formation be halted, one would expect that the metabolic processes, within the enterocyte, which precede this step would also rapidly be halted through the build-up of intermediate products. The above is not too dissimilar to the situation shown to exist with carbohydrate-stimulated IR-GIP release. Glucose which is actively transported across the gut mucosa stimulates IR-GIP release (Morgan, 1979) but fructose which has "passive" or "facilitated" (Crane, 1968; Gray, 1975) transport across the gut mucosa does not (Morgan, 1979). page 248 SUMMARY OF RESULTS In obese hyperglycaemic (ob/ob) mice 1) there appears to be an enhanced rate of synthesis and secretion of IR-GIP in the small intestine. 2) the onset of hyperinsulinaemia prece^des the development of both hyperglycaemia and hyperGIPaemia. 3) fat is by far the best stimulus to plasma IR-GIP, and the peak in plasma concentrations is achieved very rapidly as compared to humans and rats. 4) fat augments the insulin release caused by other nutrients when combined with them. 5) both fat and glucose stimulated and basal IR-GIP levels are totally refractory to any inhibition by either hyperglycaemia and/or hyperinsulinaemia. 6) the insulinotropic action of exogenous GIP is potentia ted by glucose. 7) the pre-requisite for fat-stimulated IR-GIP release is the absorption of fatty acids and glycerides capable of being esterified into triglyceride, any physical, chemical or metabolic interference with this will result in decreased IR-GIP release. page 249 CHAPTER NINE FINAL DISCUSSION AND FUTURE WORK page 250 9.1.1. The enteroinsular axis The first experiment in chapter three (3.1) demonstrated that there was no diurnal variation in the basal levels of plasma glucose, immunoreactive insulin (IRI) and immunoreactive gastric inhibitory polypeptide (IR-GIP) in healthy human volunteers following a meal and 24 hour fast. There was, however, a diurnal variation in basal plasma immunoreactive motilin (IRM), but it was not clear why. It was decided in experiment 3.2 to look at whether there were any changes in the enteroinsular axis following a meal at two times during the day, 1100 h and 0200 h. This experiment showed a degree of insulin resistance at 0200 h compared with 1100 h. The intravenous glucose tolerance experiment (3.3) demonstrated that this difference could not be attributed to differences in rates of gastric emptying at these times. Thus the IRM results in experiment 3.1 seemed less relevant. For the final experiment in chapter three (3.4), observing the effects of a mixed meal at six time points through the day, plasma IRM levels were not measured as its role post-prandially in normal subjects was . not clear. The experiment showed the diurnal variation in the action of the enteroinsular axis throughout the day, and that the enteroinsular axis worked at its best at about 1000 h. Hence in subsequent chapters when the enteroinsular axis was being studied, this was borne in mind and the experiment timed accordingly. The experiments in chapter four stemmed from the work of Dr Jill Dryburgh (et al., 1980) on the inhibition of fat-stimulated IR-GIP release. In preliminary experiments we were seeking to augment fat-stimulated IR-GIP release by page 251 short-term fat pretreatment of the rats to be used in the experiment, as the plasma IR-GIP release in rats following oral fat was not very large. This was probably due to the fact that the rats normally received a low fat diet, so that the rate of fat absorption was not maximal. Although the fat-stimulated IR-GIP release was increased, we found we were unable to inhibit the response, as expected, with exogenous insulin. This led to experiment 4.1 where it was found that the short term pretreatment of rats with a high fat diet increased the rate of fat absorption and increased the magnitude of the IR-GIP release, but that this was associated with a degree of insulin resistance as demonstrated by the abolition of the feed-back inhibition of fat stimulated-IR-GIP release. The same was found in humans (experiment 4.2) although no increase in fat-stimulated IR-GIP was seen. Again, a degree of insulin resistance was demonstrated by the fact that the return to normoglycaemia after an insulin stress-test took longer after a fat load. The overall conclusion of the experiments in chapter four was that a high fat diet can, possibly via GIP, cause a degree of insulin resistance which in the long term could result in glucose intolerance, hyperinsulinaemia and the associated problems of obesity. The experiments in chapters three and four highlight the importance of diet in the function of the enteroinsular axis in both man and the rat. Even short-term changes in the diet can cause marked metabolic differences. This was demonstrated quite graphically in experiment 4.1, where just 4 days on a high fat diet was enough to cause insulin resistance. Even in man (experiment 4.2) fairly short-term changes in the amount of fat in the diet were enough to decrease the feed-back page 252 inhibition of insulin on fat-stimulated IR-GIP release and allow normoglycaemia to be re-established more rapidly following an insulin stress-test. Experiment 3.4 also showed that more care should be taken in the fasting of volunteers prior to experiments, as ample fasting in one situation is not always enough under other circumstances. So, when investigating the enteroinsular axis in the future, statements in the methodology such as "...normal human volunteers fasted overnight...", will no longer be sufficient. The subjects' usual diet for the preceding month(s), in terms of fat, carbohydrate, protein and fibre should be recorded. The subjects' meals immediately before the experiment should be the same, and be consumed at the same times, and they should all fast for exactly the same length of time. This should hopefully lead to more reliable results when dealing with "normal human volunteers", and perhaps produce less equivocal results when compared to animal experiments where all these factors are controlled as a matter of course. Further work in humans was beginning to get difficult as long-term dietary monitoring and manipulation was going to be required to extend the work. Thus it was to the obese hyperglycaemic hyperinsulinaemic (ob/ob) mouse that we turned to for additional data before future long-term human studies. In chapter eight a series of experiments were performed on obese (ob/ob) mice. In the development of the syndrome I there is an enhanced rate of synthesis of IR-(ÿ* in the small intestine. Hyperinsulinaemia precedes the development of both hyperglycaemia and hyperGIPaemia, suggesting that the hyper alimentation of the mice initially causes hyperinsulinaemia, which in turn causes insulin resistance; one of the features page 253 of this would be an overactive enteroinsular axis, which would then reinforce the hyperinsulinaemia and hence the development of the ob/ob syndrome. The experiments in chapter eight demonstrate that fat is the best stimulus to IR-GIP in the ob/ob mouse and that fat augments the insulin release caused by other nutrients. This demonstrates the possible role of fat in the development of hyperinsulinaemia. The insulin resistance of the GIP-cells of the ob/ob mice is demonstrated by the fact that both fat and glucose- stimulated IR-GIP release are not inhibited by either hyperglycaemia and/or hyperinsulinaemia. Thus, the first metabolic role for GIP which was investigated in chapters three, four and eight was that in the enteroinsular axis. The conclusion is that GIP may, via the enteroinsular axis, have a role in the development and perhaps maintainance of the hyperglycaemia of obesity. In chapter six the role of GIP in the differential insulin responses . to oral and intravenous glucose in insulinoma bearing NEDH rats was sought. It was found that the insulinoma bearing rats had lost their responsiveness to intravenous, but not intragastric, glucose, and that in this context GIP was apparently an incretin. The fact that GIP's insulin-stimulating ability is glucose dependent (Elahi et al., 1979) would indicate some cooperation between the glucose and insulin receptors at the pancreatic B-cell. This relationship appears to have broken down in the insulinoma- bearing animals. It is interesting to consider this finding together with the results of Jorde et al. (1986) where they found GIP to be insulinotropic at euglycaemia, following glucose priming. It would seem that in normal man and rat, hyperglycaemia alone and hyperGIPaemia together with page 254 hyperglycaemia are Insulinotropic. So glucose alone can initiate insulin release, but for GIP there is a "dual key" arrangement, where glucose is the other key. However, the results of Jorde et al. (1986) tend to indicate that "once the glucose key has been turned and then removed", there is a lag before the activated mechanism turns off. Indeed, the results with the insulinoma bearing rats show the relationship can be altered so that glucose alone can no longer initiate insulin release, but that GIP is also required. This is also the case with the ob/ob mouse where i.p. glucose is not insulinotropic without GIP (whether exogenous, or endogenously released by fat), but GIP ^ insulinotropic at basal glucose concentrations (figure 8.6), and oral fat alone is insulinotropic (figure 8.3). However, it is still not clear what the precise details of these defects, or even the original mechanisms, are. It is not clear if hyperinsulinaemia is indeed caused by an overactive enteroinsular axis (ob/ob mouse), in turn perhaps due to hyperphagia. But the result of the hyperinsulinaemia (insulinoma bearing rats) appears to be insulin resistance, increased weight, and at the B-cell the uncoupling of GIP stimulated insulin release from glucose dependence. These are also symptoms of the ob/ob syndrome. From the above conclusions one could formulate a theory about the development of obesity. Hyperphagia leads to an overactive enteroinsular axis, which in turn, via the causation of hyperglycaemia, uncouples GIP action from glucose dependence. Fat has a special part to play, as it seems to have a role in causing insulin resistance (at least in the feed back mechanism on GIP release). Now even fat (via GIP) can stimulate insulin release, so that following a mixed meal page 255 there is nothing to moderate a large and sustained hyperinsulinaemia giving rise to the sort of symptoms found in the ob/ob syndrome. 9.1.2. The release of IR-GIP The final experiment in chapter eight (experiment 8.3) shows the pre-requisite for fat-stimulated IR-GIP release in ob/ob mice is the absorption of fatty acids and glycerides capable of being esterified into triglyceride; any physical, chemical or metabolic interference with this will result in decreased IR-GIP release. But we still only know that the trigger for GIP release lies somewhere between the absorption of fat into the enterocyte and its transport out again as a chylomicron. There is still much work to be done in the future to elucidate the actual metabolic process which leads to IR-GIP release. Also we do not know what common link or links exist between fat, carbohydrate and amino acid stimulated IR-GIP release. So experiments such as 8.3 should be carried out for sugars and amino acids (most readily with the ob/ob mouse). With information on how fats, carbohydrates and amino acids stimulate IR-GIP release it might be possible to demonstrtate a common link in their metabolism in the enterocyte, and we may be able to get closer to the nature of the message sent to the GIP-cell to cause it to release GIP. It seems that GIP may be a gereral purpose, post prandial, metabolic hormone stimulated by nutrients, which augments the rate of removal from the circulation of glucose and probably fatty acids. Thus, as amino acids also stimulate GIP release (Brown, 1982) the role of GIP in the removal of amino acids by the liver should be investigated. page 256 9,1.5. The enterpadipose axis In chapter five work was begun to investigate metabolic roles of GIP other than as an incretin. The combined results of the three experiments, with combinations of endogenous and exogenous GIP and lipid, showed fairly convincingly that GIP probably has a role in the removal of chylomicron triglyceride in the rat. However, in chapter seven when the experiments switched to humans the results did not agree with those obtained in the rat. With the benefit of hindsight one can see that there were flaws in the experimental design. The hyper 1 ipidaemia following i.v. Intralipid was too short lived (about 1 hour) to effectively mimic the physiological response to oral fat, as was the hyperGIPaemia, being typical for a carbohydrate load (short-lived) and not a fat load (long-lived). Since everything else to do with the absorption and removal of fat from the circulation is done at a fairly slow pace, the possible stimulation of lipoprotein lipase by GIP may also take longer than was allowed for in the experimental design in chapter seven. It has been shown that insulin induced adipose tissue lipoprotein lipase levels take at least three hours to become significantly elevated (Sadur and Eckel, 1980). In the future, were the experiment to be repeated in humans, a continuous infusion of lipid should be used, ideally purified human chylomicra but probably Intralipid, mimicking physiological postprandial triglyceride levels, together with an infusion of human GIP (ideally non-synthetic). Such an experiment should yield results more page 257 in agreement with those in chapter five than those in chapter seven. Repeating the anti-GIP antibody experiment (5.1) in humans is not feasible for many reasons, including ethical as well as practical difficulties. 9.1.4. Other future lines of research I have already mentioned above future lines of research which I believe stem directly from the work carried out in this thesis. The role of fat in the development of insulin resistance and hence the development of obesity requires long term studies in humans to generate additional data. The mechanism of GIP stimulated insulin release still needs clarification. The change in the relationship between the GIP and glucose receptors in the rat insulinoma when considered with the work of Jorde et al. (1986) on glucose priming of GIP-stimulation of insulin release in humans, suggests further areas of research. The most intriguing of these, I believe, would be to see if glucose priming could , at euglycaemia (arterial glucose levels), allow oral fat to stimulate insulin release, and whether this was achieved via GIP. The role of fat could also be incorporated into this set of experiments, as the theory expounded at the end of section 9.1.1 suggests that the higher the habitual fat intake of the individual the greater the potential for fat-stimulated insulin release. High and low fat diet groups of "normals" could be compared together with obese hyperinsulinaemic subjects. There is still a need for "a unifying theory" for the mechanism of nutrient-stimulated GIP release. Section 9.1.2 suggests experiments whose results may yield more information page 258 on this subject. Detailed studies, in parallel, of the processes involved in the absorption of carbohydrates, fats and amino acids, and the subsequent release of GIP, will not be an easy task. Further investigation of the mechanism of nutrient-stimulated GIP release might also explain the mechanism by which, in "normal subjects", insulin can inhibit fat- but not glucose-stimulated IR-GIP release. This also suggests an experiment to investigate possible feed-back inhibition of amino-acid-stimulated IR-GIP release. In section 9.1.3 there is an outline of an experiment which, hopefully, should unequivocally demonstrate whether GIP has a role in the removal of chylomicron triglyceride in humans. As in insulinoma bearing rats, glucose-stimulated insulin release appears to be GIP, or at least incretin, dependent. Further information on the relationship between the GIP and glucose receptors on the pancreatic B-cell might be gained by running the reverse of the Jorde et al. (1986) experiment in insulinoma bearing and normal rats. The rats would be primed with a large i.v. or i.p. dose of GIP, then when the arterial plasma GIP levels had returned to basal, the animals would receive an i.v. bolus of glucose. No-GIP controls would, of course, be required. It may be possible to demonstrate GIP-priming of glucose stimulated insulin release. If not, it would indicate a definite sequence of events for glucose/GIP- stimulated insulin release; more a "combination lock" than a "dual key". In chapters three, four and eight it has been argued that GIP may have a role in the development and maintainance of hyper insu1inaemia and obesity. An experiment which might provide additional data would be one in which an obese page 259 animal, or one which should become obese, had all its GIP cells removed. This could be achieved by immunologically destroying the GIP-cells with GIP antibodies. The ob/ob mouse would be an appropriate animal model as it becomes obese, and it is small, so that large amounts of anti-GIP antibodies would not be required. Control groups treated with non-immune serum would be required. The antiserum would have to be carefully screened in order to check for cross-reactions with other gut hormones as this would cause problems with the interpretation of results. It would probably be best to affinity purify the antibody as this would reduce the amount of antibody and foreign protein injected into the animals and reduce the likelihood of anaphylaxis. Should the affinity purified antibody still cross-react with any other gut or pancreatic hormones these could be adsorbed on an appropriate column. Two sets of experiments could be done. One would be to see if the mice can be prevented from becoming obese. One slight problem here is that only one quarter of the litter, from heterozygous matings, are ob/ob, whilst the other three quarters are lean ob/+ and lean +/+. Another experiment would be to see if the antibody treatment could arrest or even regress the development of obesity in the older mice. One of the problems with the interpretation of IR-GIP results in humans is that of cross-reactivity of the porcine GIP antibodies, that we all work with, with human GIP. Indeed Jorde et al. (1983) and Krarup and Holst (1984) have shown that antisera from different research groups have widely differing cross reactivities with human GIP. Human GIP has now been sequenced (Jdrnval et al., 1981), and synthetic human GIP has been made and is now commercially available. No data has yet been published confirming 100 % biological page 260 activity of the synthetic human GIP. A small research project is alrady underway in our laboratory to purify human GIP (both 5000 and 8000 molecular weight. Immuno-affinity purification procedures, developed in the last few years with my colleague Dr Kamran Beyzavi, will be used to purify GIP from fresh human intestinal tissue. We hope to produce material suitable for the production of monoclonal antibodies, and then go on to purify it further for sequencing. The monoclonal antibody will be used to develop immunoassays, hopefully, for both 5000 and 8000 molecular weight GIP. Finally, work that I would like to suggest for the near future would be to establish the existence, or otherwise, of other incretins. In order to achieve this a modification of the immunoaffinity techniques that we have developed would be used. Whole pancreatic tissue would be taken and from this a receptor fraction produced. This crude receptor fraction with receptors for both endocrine and exocrine functions, would be immobilised on our own rigid, inorganic, large pore affinity matrix. A crude extract (aqueous or acid/ethanol) of small intestine would be made and passed through the affinity column. Any hormone for which the pancreas has a receptor should remain on the column. The extract would then be fractionated (molecular seive, ion exchange, electrophoresis etc.) and assayed for incretin activity. A screening procedure for possible incretins would be required, and the easiest technique would probably be an in vitro technique using cultured rat insulinoma (RIN) cells. A likely candidate, once identified, would then be checked in vivo. This screening technique would require that rat pancreatic and intestinal tissue be used. The use of human islets or page 261 cultured insulinoma (should they be available) would allow the use of human tissues, but in vivo testing would be difficult both ethically and because of the amount of "incretin" that would be required. Should an incretin other than GIP be detected, purification, sequencing, monoclonal antibody production and immunoassay development would logically follow. Once the receptor column had been produced the investigation of other possible hormones in an "entero-pancreatic axis" would of course be possible. This technique has the potential to be used in a variety of situations where a hormonal tissue and a target tissue have been identified but where the messenger or messengers have not. The most important of these, with respect to future work from this thesis, would be the "entero-adipose ax is". page 262 REFERENCES Adrian TE, Greenberg GR, Barnes AJ, Chrlstofides ND, Alberti KGMM, Bloom SR (1980) Effect of PP on motilin and circulating metabolites in man. Eur J Clin Res 10: 235-240 Ahren B, Hedner P, Lundquist I (1983) Interaction of gastric inhibitory polypeptide (GIP) and cholecystokin in (CCK-8) with basal and stimulated insulin secretion in mice. Acta Endocrinol 102: 96-102 Andersen DK, Elahl D, Brown JC, Tobin JD, Andres R (1978) Oral glucose augmentation of insulin secretion. Interactions of gastric inhibitory polypeptide with ambient glucose and insulin levels. J Clin Invest 62: 152-161 Arendt J, Hampton S , English J , Kwasowski P, Marks V (1982) 24-hour profiles of melatonin, cortisol, insulin, C-peptide and GIP following a meal and subsequent fasting. Clin Endocrinol 16: 89-95 Assimacopoulos-Jannet F, Jeanrenaud B, (1976) The hormonal and metabolic basis of experimental obesity. Clin Endocr Metab 5: 337-365 Bailey CJ, Flatt PR, Atkins TW (1982) Influence of genetic background and age on the expression of the obese hyperglycaemic syndrome in Aston ob/ob mice. Int J Obes 6 : 11-21 Bataille D, Freychet P, Rosselin G (1974) Interactions of glucagon, gut glucagon, vasoactive intestinal polypeptide and secretin with liver and fat cell membranes: Bindind to specific sites and stimulation of adenylate cyclase. Enocrinol 95: 713-721 Baumert JE, Cataland S, Tetrick CE, Pace WG, Mazzaferri EL (1978) Effect of atropine on meat-stimulated gastrin and gastric inhibitory polypeptide (GIP) release. J Clin Endocrinol Metab 46: 473-476 Bayliss WM, Starling EH (1902) On the causation of the so-called "peripheral reflex secretion of the pancreas. Proc Royal Sos (London) 69: 352-353 Beck B, Max J-P (1983) Gastric inhibitory polypeptide enhancement of the fatty acid incorporationinto adipose tissue in the rat. Reg Pep 7: 3-8 Beck B, Max J-P (1986) Direct metabolic - effects of gastric inhibitory polypeptide (GIP): dissociation at physiological levels on insulin-stimulated fatty acid and glucose incorporation in rat adipose tissue. Diabetologia 29: 68 Beck B, Max J-P, Villaume C, (1985) Opposite effect of GIP and somatostatin on insulin-stimulated fatty acid incorporation into adipose tissue in the rat. IDF post-congress symposium on entero-insular axis. Salamanca (Spain) 30/9-1/10/1985 Best LC, Atkins TW, Bailey CJ, Flatt PR, Newton DF, Matty AJ (1977) Increased activity of the entero-insular axis in obese hyperglycaemic mice (ob/ob). J Endocrinol 72: 44P Besterman HS, Cook GS, Sarson DL, Christofides ND, Bryant MG, Gregor M, Bloom SR (1979) Gut hormones in tropical maladsorption. Brit Med J 1: 1252-1255 Besterman HS, Sarson DL, Johnson Dl, Stewart JS, Guerin S, Bloom SR, Blackburn AM, Patel HR, Modigiani R, Mallinson CH (1978) Gut-hormone profile in coeliac disease. Lancet 15 April, 785-788 Bloom SR (1975) GIP in diabetes. Diabetologia 11: 334 page 263 Bloom SR, Bryant MG, Cochralne JPS (1975) Normal distribution and post prandial release of gut hormones. Clin Sci Molec Med 49; 3p Bloom SR, Mitznegg P, Bryant MG (1976) Measurement of plasma motilin. Scand J Gastroent 11 (Supp 39); 47-52 Bloom SR, Polak JM (1978) Gut hormones overview. In: Bloom SR (ed) Gut hormones. Churchill Livingstone, Edingburgh, London, New York, pp 3-18 Bobbioni E, Jeanrenaud B (1982) Possible role of the CNS in regulating insulin secretion Via humoral factors, Int J Obes 6 CSuppl ID: 21-25 Boberg J, Carlson LA (1964) Determination of heparin-induced lipoprotein lipase activity in human plasma. Clin Chim Acta 10: 420-427 Botha JL, Vinik AI, Brown JC (1976) Gastric inhibitory polypeptide in chronic pancreatitis. J Clin Endocrinol Metab 42: 791-797 Bray GA, York DA (1979) Hypothalamic and genetic obesity in experimental animals: an autonomic and endocrine hypothesis. Physiol Rev 59: 719-809 Brown JC (1974) Gastric inhibitory polypeptide (GIP). In: Taylor S (ed) Endocrinology 1973. Heinemann, London, pp 276-284 Brown JC (1982) Gastric Inhibitory Polypeptide. Springer— Verlag. Berlin, Heidelberg, New York. Brown JC, Cook MA, Dryburgh JR (1973) Motilin, a gastric motor activity stimulating polypeptide: the complete amino acid sequence. Can J Biochem 51: 533-537 Brown JC, Dryburgh JR (1971) A gastric inhibitory polypeptide II: The complete amino acid sequence. Can J Biochem 49: 867-872 Brown JC, Dryburgh JR, Ross SE, Dupré J (1975) Identification and actions of gastric inhibitory polypeptide. Recent prog Horm Res 31: 487-532 Brown JC, Johnson LP, Magee DF (1966) Effect of duodenal alkalinization on gastric motility. Gastroenterology 50: 333-339 Brown JC, Koop H, McIntosh CHS, Otte SC, Pederson RA (1980) Physiology of gastric inhibitory polypeptide. Front Horm Res 7: 132-144 Brown JC, Magee DF (1967) Inhibitory action of cholecystokin in on acid secretion from Heidenhain pouches, induced by endogenous gastrin. Gut 8: 29-31 Brown JC, Mutt V, Dryburgh JR (1971) The further purification of motilin a gaastric motor activity stimulating polypeptide from the small intestine of hogs. Can J Physiol Pharm 49: 399-405 Brown JC, Mutt V, Pederson RA (1970) Further purification of a polypeptide demonstrating enterogastrone activity. J Physiol 209: 57-64 Brown JC, Parkes CO (1967) Effect on fundic pouch motor activity of stimulatory and inhibitory fractions separated from pancreozymin. Gastroenterology 53: 731-736 Brown JC, Pederson RA (1970) A multiparameter study on the action of preparations containing cholecystokin in pancreozymin. Scand J Gastroent 5: 537-541 Brown JC, Pederson RA, Jorpes E, Mutt V (1969) Preparation of highly active enterogastrone. Can J Physiol Pharmacol 47; 113-114 Buchanan KD (1973) Studies on the pancreatic enteric hormones. PhD thesis, Queens University of Belfast. page 264 Cited by Brown (1982) q.v. Buchanan KD, Alam MJ, Hanna KA, Banks IG, Herberg L (1982) Pancreatic and gut hormones in the DjUngarian hamster. Reg Pep 3: 67 Buchanan KD, Vance JE, Morgan A, Williams RH (1968) Effect of pancreozymin on insulin and glucagon levels in blood and bile. Am J Physiol 215: 1293-1298 Buchanan KD, Vance JE, Williams RH (1969) Insulin and glucagon release from isolated islets of Langerhans. Effect of enteric factors. Diabetes 18: 381-386 Buffa R, Polak JM, Pearse AGE, Solcia E, Grimelius L, Capella C (1975) Identification of the intestinal cell storing gastric inhibitory polypeptide. Histochemistry 43: 249-255 Burhol PG, Jorde R, Waldum HL (1980) Radioimmunoassay of plasma gastric inhibitory polypeptide (GIP), release of GIP after test meal and duodenal infusion of bile, and immunoreactive GIP components in man. Digestion 20: 336-345 Camble R, Cotton R^ Dutta AS, Gormley JJ, Hayward CF, Morley JC, Smithers MI (1973) On the synthesis of a porcine gastric inhibitory peptide. In:Hanson H, Jakubke HD (eds) Peptide 1972. North Holland, Amsterdam, pp 200-209 Carraway R, Leeman SE (1979) The amino acid sequence of bovine hypothalamic substance P. Identity of substance P from colliculi and small intestine. J Biol Chem 254: 2944-2945 Carroll KF, Nestel PJ (1972) Diurnal variation in glucose tolerance and in insulin secretion in man. Diabetes 22: 333-348 Cataland S, Crockett SE, Brown JC, Mazzaferri EL (1974) Gastric inhibitory polypeptide (GIP) stimulation by oral glucose in man. J Clin Endocrinol Metab 39: 223-228 Cataland S, O Dorisio TM, Crockett SE, Mekhjian HS (1977) Gastric inhibitory polypeptide (GIP) and insulin release in duodenal ulcer patients. J Clin Endocrinol Metab 47: 615-619 Chick WL, Warren S, Chute RN, Like AA, Lauris V, Kitchen KC (1977) A transplantable insulinoma in the rat. Proc Natl Acad Sci USA 74: 288-632 Chey WY, Lee KY (1980) Motilin. Clin in Gastroent 9: 645-656 Chey WY, Lee KY, Hendricks JG, Rhodes RA, Tai H-H (1978) Plasma secretin concentrations in fasting and postprandial state in man. Am J Dig Dis 23: 981-988 Chey WY, Rhodes RA, Lee KY, Hendricks J (1975) Radioimmunoassay of secretin:Further studies. In: Thompson JC (ed) Gastrointestinal hormones. University of Texas Press, Austin, pp 269-281 Christofides ND, Bloom SR (1981) Parallel release of two molecular forms of motilin in human plasma and suppression of big motilin by exogenous motilin infusion. Biomed Res 2: 109-112 Christofides ND, Bloom SR, Besterman HS (1978) Physiology of motilin II. In Bloom SR (ed) Gut hormones. Churchill Livingstone, Edinburgh, London, New York, pp 343-350 Christofides ND, Bloom SR, Besterman HS, Adrian TE, Ghatei MA (1979a) Release of motilin by oral and intravenous nutriments in man. Gut 20: 102-106 Christofides ND, Ghatei MA, Bloom SR, Boberg C, Gillmer MD (1982) Decreased plasma motilin concentrations in pregnancy. Brit Med J [Clin Res 3 285: 1453-1454 Christofides ND, Long RG, Fitzpatrick ML, McGregor GP, Bloom page 265 SR (1981) Effect of motilin on gastric emptying of glucose and fat in humans. Gastroenterology 80: 416-460 Christofides ND, Modi in IM, Fitzpatrick ML, Bloom SR (1978) Effect of motilin on gastric emptying of solid meals in man. Gut 19: A436-A437 Christofides ND, Modi in IM, Fitzpatrick ML, Bloom SR (1979b) Effect of motilin on the rate of gastric emptying and gut hormone release during breakfast. Gastroenterology 76: 903-907 Cleator IGM, Gourlay RH (1975) Release of gastric inhibitory polypeptide (IR-GIP) by oral ingestion of food substances. Am J Surg 130: 128-135 Clément J (1980) Intestinal absorption of triacylglycerols. Reprod Nutr Dévelop 20: 1285-1307 Cooke RJ, Buckhart GJ (1983) Hypertriglyceridaemia during the intravenous infusion of a safflower oil-based fat emulsion. Paediatrics 103: 959-961 Crane RK (1968) Absorption of sugars.. In: Code CF (ed) Alimentary Canal. American Physioogy Society, Washington, D.C., pp 1323-1358 Creutzfeldt W (1979) The incretin concept today. Diadetologia 16: 75-85 Creutzfeldt W, Ebert R (1977) Release of gastric inhibitory polypeptide (GIP) to a test meal under normal and pathological conditions in man. In: Bajaj JS (ed) Diabetes. Excerpta Medica, Amsterdam, pp 63-75 Creutzfeldt W, Ebert R (1985) New developments in the incretin concept. Diabetologia 28: 565-573 Creutzfeldt W, Ebert R, Arnold R, Frerichs H, Brown JC (1976) Gastric inhibitory polypeptide (GIP, gastrin and insulin: Response to test meal in coeliac disease and after duodenal pancreatectomy. Diabetologia 14: 15-24 Creutzfeldt W , Ebert R, Nauck M, Stockman F (1983) Disturbances of the entero-insular axis. Scand J Gastoenterol 18 CSuppl 823: 111-119 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 og serum levels. Diabetologia 14: 15-24 Crockett SE, Cataland S, Falko JM, Mazzaferri EL (1976) The insulinotropic effect of endogenous gastric inhibitory polypeptide in normal subjects. J Clin Endocrinol Metab 42; 1098-1103 Crockett SE, Mazzaferri El, Cataland S (1976) Gastric inhibitory polypeptide (GIP) in maturity onset diabetes mellitus. Diabetes 25: 931-935 Day JJ, Komarov SA (1939) Glucose and gastric secretion. Am J Dig Dis 6; 169-175 Day JJ, Webster DR (1935) The autoregulation of gastric secretion. Am J Dig Dis Nutr 2: 527-531 Debas HT, Yamagishi T, Dryburgh JR (1977) Motilin enhances gastric emptying of liquids in dogs. Gastroenterology 73: 777-780 Denniss AR, Young JA (1978) Modefication of salivary duct electrolyte transport in rat and rabbit by physalaemin, VIP, GIP and other enterohormones. Pfluegers Arch 376: 73-80 Deschamps I, Heptner W , Desjeux J-F, Baltakse V, Machinot S, Lestradet H (1980) Effects of diet on insulin and gastric inhibitory polypeptide levels in obese children. Pediatric Res 14: 300-303 Deschodt-Lanckman M, Robberecht P, Camus J, Chrisophe J page 266 (1971) Short-term adaptation of pancreatic hydrolases to nutritional and physiological stimuli in adult rats. Biochimie 53: 789-796 Dixon W, Wadia JH (1926) The action of intestinal extracts. Br Med J 1: 820 Dockray GJ (1976) Immunochemical evidence of cholecysto- kinin-like peptides in brain. Nature 264: 568-570 Domschke W, Strunz U, Mitznegg P, Ruppin H, Domshke S, Schubert E, Wünsch E, Jaeger E, Demling L (1974) 13-norleucine-moti1 in - Analyse der Wirkungen auf den Gastrointesinal-trakt. Naturwisseneschaften 61: 370 Dryburgh JR, Brown JC (1975) Radioimmunoassay for motilin. Gastroenterology 68: 1169-1176 Dryburgh JR, Hampton SM, Marks V (1980) Endocrine pancreatic control of the release of gastric inhibitory polypeptide: a possible physiological role for C-peptide. Diabetologia 19: 397-401 Dryburgh JR, Hampton SM, Morgan LM, Marks V (1980) C-peptide induced inhibition of GIP release. Régulât Pept [Supp] 1: 28 Duguid JR, Steiner DF, Chick WL (1976) Partial purification and characterisation of the mRNA for rat preproinsulin. Proc Natl Acad Sci USA 73: 3593-3543 Dupré J (1964) An intestinal hormone affecting glucose disposal in man. Lancet II: 672-673 Dupré J, Beck JC (1966) Stimulation of release of insulin by an extract of intstinal mucosa. Diabetes 15: 555-559 Dupré J, Curtis JD, Unger RH, Waddell RW, Beck JC (1969) Effects of secretin, pancreozymin or gastrin 6n the response of the endocrine pancreas to administration of glucose or arginine in man. J Clin Invest 48: 748-757 Dupré J, Greenidge N, McDonald TJ, Ross SA, Rubinstein D (1976) Inhibition of actions of glucagon in adipocytes by gastric inhibitory polypeptide. Metabolism 25: 1197-1199 Dupré J, Rojas L , White JJ, Unger RH, Beck JC (1966) Effects of secretin on insulin and glucagon in portal and peripheral blood in man. Lancet II: 26-27 Ebert R, Arnold R, Creutzfeldt W (1977) Lowering of fasting and food stimulated serun immunoreactive gastric inhibitory polypeptide (GIP) by glucagon. Gut 8: 121-127 Ebert R, Brown JC (1976) Effects of gastric inhibitory polypeptide (GIP) on lipolysis and cyclic AMP levels in isolated fat cells. Eur J Clin Invest 6: 327 Ebert R, Creutzfeldt W (1978) Aspects of GIP pathology. In: Bloom SR (ed) Gut hormones. Churchill Livingstone, Edinburgh, pp 294-300 Ebert R, Creutzfldt W (1980) Decreased GIP secretion through impairment of absorption. Front Horm Res 7; 192-201 Ebert R, Creutzfeldt W (1982) Influence of gastric inhibitory polypeptide antiserum on glucose-induced insulin secretion in rats. Endocrinology 111: 1601-1606 Ebert R, Creutzfeldt W, Brown JC, Frerichs H, Arnold R (1976) Response of GIP to a test meal in chronic pancreatitis - Relationship to exocrine and endocrine insufficiency. Diabetologia 12: 609-612 Ebert R, Frerichs H, Creutzfeldt W (1976) Serum gastric inhibitory polypeptide (GIP) response in patients with maturity onset diabetes and juvenile diabetes. Diabetologia 12: 388 Ebert R, Illmer K , Creutzfeldt W (1979) Release of gastric inhibitory polypeptide (GIP) by intraduodenal page 267 acidification in rats and humans and abolishment of the incretin effect of acid by GIP-antiserum in rats. Gastroenterol 76: 515-523 Eckel RH, Fujimoto WY, Brunzell JD (1979) Gastric inhibitory polypeptide enhanced lipoprotein lipase activity in cultured preadipocytes. Diabetes 28: 1141-1142 Edkins JS (1905) On the chemical mechanism of gastric secretion. Proc Roy Soc (London) 76: 376 Eisenbarth GS, Die H, Gazdar A , Chick W, Schultz JA, Scearce RM (1981) Production of monoclonal antibodies reacting with rat islet cell membrane antigens. Diabetes 30: 226-230 Elahi D, Andersen DK, Brown JC, Debas H, Hershkopf RJ, Raizes GS, Tobin JD, Andres R (1979) Pancreatic a- and G-cell responses to GIP infusion in normal man. Am J Physiol 237: 185-191 Elrick ML, Stimmler L, Hlad CJ, Arai Y (1964) Plasma insulin response to oral and intravenous glucose administration. J Clin Endocrinol 24: 1076-1082 Ewald CA, Boas J (1886) Beitràge zur Physiologie und Pathologie der Verdauung II. Virchows Arch [Pathol Anat] 104: 271-305 Falconer DS (1960a) The genetics of litter size in mice. J Cell Comp Physiol 56 [Suppl 13: 153-167 Falconer DS (1960b) Selection of mice for growth on high and low planes of nutrition. Genet Res 1: 91-113 Falko JM, Crockett SE, Cataland S, Mazzaferri EL (1975) Gastric inhibitory polypeptide (GIP) stimulated by fat ingestion in man. J Clin Endocrinol Metab 41: 260-265 Fara JW, Rubinstein EH, Sonnenschein RR (1972) Intestinal hormones in mesenteric vasodilation after intraduodenal agents. Am J Physiol 223: 1058-1067 Fara JW, Salazar AM (1978) Gastric inhibitory polypeptide increases mesenteric blood flow. Proc Soc Exp Biol Med 158: 446-448 Farrell JI, Ivy A (1926) Studies on the motility of the transplanted gastric pouch. Am J Physiol 46: 227-228 Feng TP, Hou HC, Lim RKS (1929) On the mechanism of the inhibition of gastric secretion by fat. Clin J Physiol 3: 371-378 Finch PM, Ingram DH, Henstridge JD, Catchpole BN (1982) Relationship of fasting gastroduodenal motility to the sleep cycle. Gastroenterol 83: 605-612 Flatt PR, Bailey CJ (1981a) Development of glucose intolerance and impaired plasma insulin response to glucose in obese hyperg1ycaemic (ob/ob) mice. Horm Metab Res 13: 556-560 Flatt PR, Bailey CJ (1981b) Abnormal plasma glucose and insulin responses in heterozygous lean (ob/+) mice. Diabetologia 20: 573-577 Flatt PR, Bailey CJ (1982a) Role of dietary factors in the hyperinsulinaemia of genetically obese hyperglycaemic (ob/ob) mice. J Nutr 112: 2212-2216 Flatt PR, Bailey CJ (1982b) Plasma glucose and insulin responses to glucagon and arginine in Aston ob/ob mice: evidence for a selective defect in glucose-mediated insulin release. Horm Metab Res 14: 127-130 Flatt PR, Bailey CJ (1983) Glucose and insulin response to conditioned feeding in lean and genetically obese hyperglycaemic (ob/ob) mice. Metabolism 32: 504-509 Flatt PR, Bailey CJ (1984) Dietary components and insulin responses to fasting and refeeding in genetically obese page 268 hyperg1ycaemic (ob/ob) mice. Brit J Nutr 51; 403-413 Flatt PR, Bailey CJ, Buchanan KD (1982) Regulation of plasma immunoreactive glucagon in obese hyperglycaemic (ob/ob) mice. J Endocr 95: 215-227 Flatt PR, Bailey CJ, Kwasowski P, Swanston-Flatt SK, Marks V (1983) Abnormalities of GIP in spontaneous syndromes of obesity and diabetes in mice. Diabetes 32: 433-435 Flatt PR, Tan K, Bailey CJ, Swanston-Flatt SK, Marks V, Webster JD (1982) Plasma glucose and insulin concentrations following implantation and surgical resection of a transplantable rat insulinoma. Biochem Soc Trans 10: 273-274 Frerichs H, Creutzfeldt W (1976) Hypoglycaemia, 1. Insulin secreting tumors. J Clin Enocrinol Metab 5: 747-767 Forsgren E, Holmgren H, Wilander O, Agren G (1930) The connection between the functional activities of the liver and the susceptibility of the organism to insulin. Acta Med Scand 73: 60-70 Fussganger K, Straub K, Goberna R, Jaros P, Schroder KE, Raptis S, Pfeifer EF (1969) Primary secretion of insulin and secondary release of glucagon from islated perfused rat pancreas following stimulation with pancreozymin. Horm Metab Res 1: 224-227 Fujimoto WY, Ensick JW, Merchant W, Williams RH, Smith PH, Johnson DG (1978) Stimulation by gastric inhibitory polypeptide of insulin and glucagon secretion by rat islet cell cultures. Proc Soc Exp Biol Med 157: 89-93 Gaginella TS, Mekhjian HS, 0 Dorisio TM (1978) Vasoactive intestinal peptide: Quantification by radioimmunoassay in isolated cells, mucosa, and muscle of the hamster intestine. Gastroenterology 74: 718 Gagliardino JJ, Bel lone CT, Angeletti AJ, Pessacq MT Notti HP (1976) Insulin-induced hypoglycaemia: 24 h variation. Acta Diabetol Lat 3: 134-137 Gazdar AF, Chick WL, Oie HK, Sims HL, King DL, Weir GC, Lauris V (1980) Continuous clonal, insulin-somatostatin- secreting cell lines established from a transplantable rat islet cell tumour. Proc Natl Acad Sci USA 77: 3519-3523 Gibson T, Jarrett (1972) Diurnal variation in insulin sensitivity. Lancet 947-948 Gillespie IE, Grossman MI (1964) Inhibitory effect of secretin and cholecystokin in on Heidenhain pouch responses to gastrin extract and histamine. Gut 5: 342-345 Glass GBJ (1980) Preface. In: Glass GBJ (ed) Gastrointestinal Hormones. Raven Press, New York, pp v-vi Goodfriend TL, Levine L , Fasman GD (1964) Antibodies to bradykinin and angiotensin: A use for carbodiimides in immunology. Science 144: 1344-1346 Gray GM (1975) Carbohydrate digestion and absorption: role of the small intestine. New Eng J Med 292: 1225-1230 Gray GS, Bradley WB, Ivy AC (1937) On the preparation and biological assay of enterogastrone. Am J Physio 118: 463-476 Greengard H, Atkinson AJ, Grossman MI, Ivy AC (1946) The effectiveness of parenterally administered "enterogastrone" in the prophylaxis of recurrences of experimental and clinical peptic ulcer. Gastroenterology 7: 625-649 Grill V (1981) Time and dose dependencies for priming effect of glucose on insulin secretion. Endocrinol Metab 3: page 269 E24-E31 Grossman MI (1970) Gastrin, cholecystokin in and secretin act on one receptor. Hypothesis. Lancet I: 1088-1090 Grossman MI (1977) Physiological effects of gastrointestinal hormones. Fed Proc 36: 1930-1932 Grossman MI (1981) General concepts. In: Bloom SR (ed) Gut hormones. Churchill Livingstone, Edinburgh, London, New York, PP 464-470 Grundleger ML, Thenen SW (1982) Decreased insulin binding, glucose transport and glucose metabolism in soleus muscle of rats fed a high fat diet, diabetes 31: 232-237 Halberg D (1965) Elimination of exogenous lipids from the blood stream. An experimental, methodological and clinical study. Acta Physiol Scand 65 CSuppl 2543: 1-23 Hampton SM (1983) The C-peptide of proinsulin its diagnostic use and a possible physiological role. PhD Thesis, University of Surrey. Heitz Ph, Polak JM, Pearse AGE (1978) Cellular origin of motilin. In: Bloom SR (ed) Gut hormones. Churchill Livingstone, Edinburgh, London, New York, pp 332-334 Helmon CA, Barbazat GO (1977) The effect of gastric inhibitory polypeptide on human jejunal water and electrolyte transport. Gastroenterology 72: 376-379 Herberg L, Coleman DL (1977) Laboratory animals exhibiting obesity and diabetic syndromes. Metabolism 26: 59-99 Himsworth HP (1933) The influence of diet on the sugar tolerance of healthy men and its reference to certain extrinsic factors. Clin Sci 1: 251-264 Himsworth HP (1934) Dietetic factors influencing the glucose tolerance and activity of insulin. J Physiol 81: 29-48 Ingals AM, Dickie MM, Snell GD (1950) Obese, a new mutation in the house mouse, J Hered 41: 317-318 Ip CH, Tepperman M, Holohan S, Tepperman J (1976) Insulin binding and insulin response of adipocytes from rats adapted to fat feeding. J Lipid Res 17: 588-599 Itoh Z, Aizawa I, Takeuchi S, Couch EF (1975) Hunger contractions and motilin. In: Vantrappen (ed) Proceedings of the 5th International Syposium on G I motility. Typoff-Press, Belgium, pp 48-55 Itoh Z, Takeuchi S, Aizawa I, Takayanagi R, Mori K, Taminato T, Seino Y , Imura H, Yanaihara N (1978) Recent advances in motilin research: its physiological and clinical significance. Adv Exp Med Biol 106: 241-257 Iverson J (1971) Secretion of glucagon from the isolated perfused canine pancreas. J Clin Invest 50: 2123-2136 Ivy AC, Oldberg E (1928) Hormone mechanism for gall bladdeer contraction. Amer J Physiol 85: 381-383 Janssens J, Hellemans J, Adrian TE, Bloom SR, Peeters TL, Christofides N, Vantrappen GR (1982) Pancreatic polypeptide is not involved in the regulation of the migrating motor complex in man. Reg Pep 1; 41-49 Jenkins DJA (1980) Influence of fibre and guar supplemented food on insulin secretion and glucose tolerance. In: Creutzfeldt W (ed) Frontiers of hormone research. S Karger, Basel Jennewein HM, Hummelt H, Seiwert R, Waldeck R (1975) The motor-stimulating effect of natural motilin on the lower oesophageal sphincter, fundus, antrum and duodenum in dogs. Digestion 13: 246-250 Johnson LP, Magee DF (1965) Cholecystokin in-pancreozymin extracts and gastric motor inhibition. Surg Gynecol Obstet 121; 557-562 page 270 Johnson LR, Grossman MI (1968) Secretin: The enterogastrone released by acid in the duodenum. Am J Physiol 215: 885-888 Jones IR, Owens DR, Sarsons DL, Bloom SR (1985) Day profiles of glucose dependent insulinotropic polypeptide (GIP) in normal subjects. Horm Metab Res 17: 660-662 Jorde R, Amland PF, Burhol PG (1986) The priming effect of glucose on the gastric-inhibitory polypeptide-induced insulin release. Scand J Gastroenterol 21; 47-50 Jorde R, Burhol PG, Schulz TB (1983) Fasting and postprandial GIP values in man measured with seven different antisera. Reg Pep 7: 87-94 Jorde R, Burhol PG, Waldum HL, Schulz TB, Lyngren I, Florholmen J (1980) Diurnal variation of plasma gastric inhibitory polypeptide in man. Scand J Gastroent 15: 617-619 Jornvall H, Carlquist M, Kwauk S, Otte SC, McIntosh CHS, Brown JC, Mutt V (1981) Amino acid sequence and heterogeneity of gastric inhibitory polypeptide (GIP). FEBS Lett 123: 205-210 Jorpes JE, Mutt V (1961) On the biological activity and amino acid composition of secretin. Acta Chem Scand 15: 1790-1791 Jorpes JE, Mutt V, Toczko K (1964) Further purification of cholecystokin in and pancreozymin. Acta Chem Scand 18: 2408-2410 Karam JH, Grodsky GM, Foksham PH (1963) Excessive insulin response to glucose in obese subjects as measured by immunological assay. Diabetes 12: 197-204 Kitagbi P, Carraway R, Leeman SE (1976) Isolation of a tridecapeptide from bovine intestinal tissue and its partial characterization as neurotensin. J Biol Chem 251: 7053-7058 Koch J, Domschke S, Belohalvek D, Domschke W, Wünsch E, Jaeger E, Demling L (1976) Gastric mucosal blood flow and pepsin secretion in dogs - stimulation by 13-nle- motilin. Scand J Gastoenterol 11: 93-96 Konturek S , Grossman MI (1965) Effects of perfusion of intestinal loops with acid, fat or dextrose on gastric secretion. Gastroenterology 49: 481-489 Konturek SJ, Krol R, Dembinski A, Wünsch E (1976) Effect of motilin on pancreatic secretion. Pflugers Archiv 364: 297-300 Koopmans HS (1981) Peptides as satiety agents: the behavioural evaluation of their effects on food intake. In: Bloom SR, Polak JM (eds) Gut hormones. Churchill Livingstone, Edinburgh, London, New York, pp 464-470 Kosaka T, Lim RKS (1930) Demonstration of the humoral agent in fat inhibition of gastric secretion. Proc Soc Exp Biol Med 27: 890-891 Kosaka T, Lim RKS, Ling SM, Liu AC (1932) On the mechanism of gastric secretion by fat. A gastric-inhibitory agent obtained from the intestinal mucosa. Chin J Physiol 6: 107-126 Krarup T, Holst JJ (1984) The heterogeneity of gastric inhibitory polypeptide in porcine and human gastrointestinal mucosa evaluated with five different antisera. Reg Pep 9: 35-46 Krarup T, Madsbad S, Regeur L, Faber OK, Tronier B (1984) Relation of immunoreactive gastric inhibitory polypeptide to changes in glycaemic control and B cell function in type 1 (insulin-dependent) diabetes page 271 mellitus. Acta Endocrinol (Copenh) 105: 221-225 Kuzio M, Dryburgh JR, Malloy KM, Brown JC (1974) Radioimmunoassay for gastric inhibitory polypeptide. Gastroenterology 66: 357-364 Labarre J (1932) Sur les possibilités d'un traitment d'un diabetes par l'incretine. Bull Acad R Med Belg 12: 620-634 Labarre J , Still EV (1930) Studies on the physiology of secretin. Am J Physiol 91: 649-653 Lang DA, Mathews PE, Phil D, Peto D, Peto J, Turner RC (1979) Cyclic oscillations of basal plasma glucose and insulin concentrations in human beings. New Eng J. Med 301: 1023-1027 Larrimer JN, Mazzaferri EL, Cataland S, Mekhjian HS (1980) Effect of atropine on glucose-stimulated gastric inhibitory polypeptide. Diabetes 27: 638-642 Laughton NM, Macallum AB (1932) The relation of duodenal mucosa to the internal secretion of the pancreas. Proc Roy Soc Lond (Biol) 111: 37-46 Lauritsen KB, Frederiksen H-J, Uhrenholdt A, Holst JJ, HUP (1982) The correlation between gastric emptying time and the response of GIP and enteroglucagon to oral glucose in duodenal ulcer patients. Scand J Gastroenterol 17: 513-516 Lauritsen KB, Holst JJ, Moody AJ (1981) Depression of insulin release by anti-GIP serum after oral glucose in rats. Scand J Gastroent 16: 417-420 Lauritsen KB, Moody AJ (1978) The response of gastric inhibitory polypeptide (GIP) and insulin to glucose in duodenal ulcer patients. Diabetologia 14: 149-153 Lavau M, Fried SK, Susini C, Freychet P (1979) Mechanism of insulin resistance in adipocytes in rats fed a high fibre diet. J Lipid Res 20: 8-16 Lee KY, Chey WY, Tai HH, Yajima H (1978) Radioimmunoassay of motilin. Validation and studies on the relationship between plasma motilin and interdigestive myoelectric activity of the duodenum of the dog. Am J Dig Dis 23: 789-795 Lee KY, Hendricks J, Chey WY (1974) Effects of gut hormones on motility of the antrum and duodenum in dogs. Gastroenterology 66: 729 Lerner RL, Porte D (1972) Studies of secretin-stimulated insulin responses in man. J Clin Invest 51: 2205-2210 Loew ER, Gray JS, Ivy AC (1940) Is a duodenal hormone involved in carbohydrate metabolism. Am J Physiol 129: 659—663 Long JF, Brooks FP (1965) Relation between inhibition of gastric secretion and absorption of fatty acids. Am J Physiol 209: 447-451 Long RG, Christofides ND, Fitzpatrick ML, 0 Shaughnessy DJ, Bloom SR (1982) Effects of intravenous somatostatin and motilin on the blood glucose and hormonal response to oral glucose. Eur J Clin Invest 12: 331-336 Mains RE, Eipner BA, Ling N (1977) Common precursor to corticotriphins and endorphins. Proc Natl Acad Sci USA 74: 3014-3018 Magee DF, Nakamura M (1966) Action of pancreozymin preparations on gastric secretion. Nature 212: 1487-1488 Maletti M, Portha B, Carlquist M, Kergoat M, Laburthe M, Marie JC, Rosselin G (1984) Evidence for and charecterization of specific high affinity binding sites for the gastric inhibitory polypeptide in pancreatic page 272 G-cells. Endocrinol 115: 1324-1331 Marks V, Morgan LM (1982) Gastrointestinal Hormones. In: Baum H, Gergely J, Fanburg BL (eds) Molecular aspects of medicine. Pergamon Press, Oxford, New York, Frankfurt, Vol 5 PP 225-292 Marks V, Rose FC (1981) Hypoglycaemia, 2nd edn. Blackwell Scientific Publications, Oxford, London. Marks V, Turner DS (1977) The gastrointestinal hormones with particular reference to their role in the regulation of insulin secretion. Essays Med Biochem 3: 109-152 Martin EW, Sirinek KR, Crockett SE, 0 Dorisio TM, Mazzaferri EL, Thomford NR, Cataland S (1975) Release of gastric inhibitory polypeptide: Composition of hyperosmolar carbohydrate solutions as stimuli. Surg Forum 26: 381-382 Maxwell V, Shulkes A, Brown JC, Solomon TE, Walsh JH, Grossman MI (1980) Effect of gastric inhibitory polypeptide on pentagastrin-stimulated acid secretion in man. Dig Dis Sci 25: 113-116 May JM, Wiliams RH (1978) The effect of endogenous gastric inhibitory polypeptide on glucose-induced insulin secretion in mild diadetes. Diabetes 27: 849-855 McDonald TJ, Nilsson G, Vagne M, Ghatei M, Bloom SR, Mutt V (1978) A gastrin releasing peptide from the porcine non-antral gastric tissue. Gut 19: 767-774 McGuigan JE (1968) Immunochemical studies with synthetic human gastrin. Gastroenterolgy 54: 1005-1011 McIntosh CHS, Pederson RA, Koop H, Brown JC (1979) Inhibition of GIP stimulated somatostatin-like immunoreactivity by acetylcholine and vagal stimulation. In: Miyoshi A (ed) Gut peptides. Kodansha, Tokyo, Elsevier/North Holland, Amsterdam, pp 100-104 McIntyre N, Holdsworth CD, Turner DS (1964) New interpretation of oral glucose tolerance. Lancet II: 20-21 Meissner AJ, Bowes KL, Zwick R, Daniel EE (1976) Effect of motilin on the lower oesophageal sphincter. Gut 17: 925-932 Menguy R (1960) Studies on the role of pancreatic and biliary secretions in the mechanism of gastric inhibition by fat. Surgery 48: 195-200 Mitznegg P, Bloom SR, Domschke W, Domschke S, Wünsch E, Demling L (1976) Release of motilin after duodenal acidification. Lancet I: 888-889 Moody AJ, Thim L, Valverde I (1984) The isolation and sequencing of human gastric inhibitory polypeptide (GIP). FEBS Lett 172: 142-149 Moore B, Edie ES, Abram JH (1906) On the treatment of diabetes mellitus by acid extract of the duodenal mucous membrane. Biochem J 1: 28-38 Morgan LM (1979) Immunoassayable gastric inhibitory polypeptide; investigations into its role in carbohydrate metabolism. Ann Clin Biochem 16: 6-14 Morgan LM, Goulder T, Tsiolakis D, Marks V, Alberti KGMM (1979) The effect of unabsorbable carbohydrate on gut hormones - modefication of post-prandial GIP secretion by guar. Diabetologia 17: 1-5 Morgan LM, Morris BA, Marks V (1978) Radioimmunoassay of gastric inhibitory polypeptide. Ann Clin Biochem 15: 172-177 Morgan LM, Wright JW, Marks V (1979) The effect of oral galactose on GIP and insulin secretion in man. page 273 Diabetologia 16: 235-239 Moroder L, Hallett A, Thamm P, Wilschowitz L, Brown JC, Wünsch E (1978) Studies on gastric inhibitory polypeptide: Synthesis of the octatricontapeptide QIpj.-»o with full insul inotropic activity. Scand J Gastroent (suppl 49) 13: 129 Mott GE, Rogers ML (1978) Enzymatic determination of triglycerides in human and baboon serum triglycerides. Clin Chem 24: 354-357 Murat JE, White TT (1966) Stimulation of gastric secretion by commercial cholecystokin in extracts. Proc Soc Exp Biol Med 123: 593-594 Mutt V, Jorpes JE (1966) Proc. IV International symposium on the chemistry of natural products. Stockolm, Sweden Section 2C-3 Nasset ES (1938) Enterocrin in, a hormone which excites the glands of the small intestine. Am J Physiol 121: 481-487 Nasset ES, Pierce HB, Mur 1 in JR (1935) Proof of humoral control of intestinal secretion. Am J Physiol 111: 145-158 Neya T, Mizutani M, Takaki M, Nekayama S (1981) Effect of motilin on the sphincter of Oddi in the dog. Acta Med Okayama 35: 417-420 O Donohue TL, Beifeld MC, Chey WY, Chang T-M, Nilaver G, Zimmerman EA, Yajima H, Adachi H, Poth M, McDevitt RP, Jacobowitz DM (1981) Identification, characterization and distribution of motilin immunoreactivity in the rat central nervous system. Peptides 2: 467-477 0 Dorisio TM, Cataland S, 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 0 Hare MMT, Shaw C, Swanston-Flatt SK, Marcelli M, Buchanan KD, Flatt PR (1985) Influence of a transplantable insulinoma on the pancreatic status of insulin and pancreatic polypeptide in the rat. Diabetologia 28: 157-160 Ohneda A, Kobayashi T, Nihei J (1983) Effect of endogenous gastric inhibitory polypeptide (GIP) on the removal of triacylglycerol in dogs. Reg Pep 6: 25-32 Pavlov IV (1910) The work of the digestive glands. Translated by Thompson WH. Griffin, London Pearce AGE (1969) The cytochemistry and ultrastructure of polypeptide hormone-producing cells of the APUD series and the embryologie, physiologic and pathologic implications of the concept. J Histochem Cytochem 17: 303-313 Pearce AGE, Polak JM, Bloom SR, Adams C, Dryburgh JR, Brown JC (1974) Enterochromaffin cells of the mammalian small intestine as the source of motilin. Virch Arch B Cell Path 16: 111-120 Pedersen 0, Hjelund E, Lindskov HD, Schwartz Sorensen N (1980) Increased insulin receptors on monocytes from insulin dependent diabetics after high starch high fibre diets. Diabetologia 19: 306 Pederson RA, Brown JC (1972) Inhibition of histamine-, pentagaastrin- and insulin-stimulated canine gastric sectetion by pure "gastric inhibitory polypeptide". Gastroenterology 62: 393-400 Pederson RA, Brown JC (1976) The insulinotropic action of gastric inhibitory polypeptide in the perfused isolated rat pancreas. Endocrinology 99: 780-785 page 274 Pederson RA, Brown JC (1978) Interaction of gastric inhibitory polypeptide, glucose and arginine on insulin and glucagon secretion from the perfused rat pancreas. Endocrinology 103: 610-615 Pederson RA, Dryburgh JR, Brown JC (1975) The effect of somatostatin on the release and insulinotropic action of gastric inhibitory polypeptide. Can J Physiol Pharmacol 53: 1200-1205 Pederson RA, Schubert HE, Brown JC (1975) Gastric inhibitory polypeptide. Its physiological release and insulinotropic action in the dog. Diabetes 24: 1050-1056 Peeters TL, Vantrappen G, Janssens J (1980) Fasting plasma motilin levels are related to the interdigestive motility complex. Gastroenterol 79: 716-719 Pfeifer EF, Telib M, Ammon J, Melani F , Ditschuneit H (1965) Direkte Stimulierung der Insulinsekretion in vitro durch Sekretin. Dtsch Med Wochenschr 90: 1663-1667 Pincus IJ, Thomas JE, Rehfuss ME (1942) A study of gastric secretion as influenced by changes in duodenal activity. Proc Soc Exp Biol 51: 367—368 Polak JM, Bloom SR, Kuzio M, Brown JC, Pearse AGE (1973) Cellular localization of gastric inhibitory polypeptide in the duodenum and jejunum. Gut 14: 284-288 Polak JM, Pearse AGE, Grimelius L, Marks V (1975) Gastrointestinal apudosis in obese hyperglycaemic mice. Virchows Arch [Cell Pathol] 19: 135-140 Popielski LB (1901) Uber das periferische reflektorische Nervencentrum das Pancreas. Pflug Arch des Physiol 86: 215-246 Pradayrol L, Jornvall H, Mutt V, Ribet A (1980) N-terminally extended somatostatin: The primary structure of somatostatin-28. FEBS Lett 109: 55-58 Quigley JP, Phelps KR (1934) The mechanism of gastric motor inhibition from ingested carbohydrate. Am J Physiol 109: 133-138 Rabinovitch A, Dupré J (1974) Effects of gastric inhibitory polypeptide present in impure pancreozymin- cholecystokin in on plasma insulin and glucagon in the rat. Endocrinology 94: 1139-1144 Rees WDW, Malagelada J-R, Miller LJ, Go VLW (1982) Human interdigestive and postprandial gastrointestinal motor and gastrointestinal hormone patterns. Dig Dis Sci 27: 321-329 Rehfeld JF (1981) Historical background. In Bloom SR, Polak JM (eds) Gut hormones. Churchill Livingstone, Edinburgh, London, New York, pp 10-16 Reichlin M, Schnure JJ, Vance VK (1968) Induction of antibodies to porcine ACTH in rabbits with non-steroido- genic polymers of BSA and ACTH. Proc Soc Exp Biol Med 128: 347-350. Reisner S , Michaelis OE, Cataland. S, 0 Dorisio TM (1980) Effect of isocaloric exchange of dietary starch and sucrose in humans on the gastric inhibitory polypeptide response to a sucrose load. Am J Clin Nutr 33: 1907-1911 Richmond W, Seviour PW, Teal TK, Elkeles RS (1984) A new approach to the intraavenous fat tolerance test with Intralipid. Clin Sci 67: 67p-68p Robinson SF, Quarfordt SH (1979) Apoproteins in association with intralipid incubations in rat and human plasma. Lipids 13: 343-349 Ross SA, Dupré J (1978) Effects of ingestion of triglyceride or galactose on secretion of gastric inhibitory page 275 polypeptide and on response 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. Gastroenterology 81; 108-111 Rouiller D, Schusdziarra V, Harris V, Unger RH (1980) Release of pancreatic and gastric somatostatin-like immunoreactivity in response to the octapeptide of cholecystokin in, secretin, gastric inhibitory polypeptide and gastrin-17 in dogs. Endocrinology 107: 524-529 Ruppin H, Domschke S, Domschke W, Wünsch E, Jaeger E, Demling L (1975) Effects of 13-nle-moti1 in in man. Inhibition of gastric evacuation and stimulation of pepsin secretion. Scand J Gastroent 10: 199-202 Sadur CN, Eckel RH (1982) Insulin stimulation of adipose tissue lipoprotein lipase. J Clin Invest 69: 1119-1125 Said SI, Mutt V (1970) Potent peripheral and splanchnic vasodilator peptide from normal gut. Nature 225: 863-864 Salera M, Giacomoni P, Pironi L, Cornia G, Capelli M, Martini A, Benefati F, Miglioli M, Barbera L (1982) Gastric inhibitory polypeptide release after oral glucose: relationship to glucose intolerance, diabetes melitus and obesity, J Clin Endocr Metab 55: 329-336 Sarson DL, Bryant MG, Bloom SR (1980) A radioimmunoassay of gastric inhibitory polypeptide in human plasma. J Endocrinol 85: 487-496 Sarson DL, Wood SM, Holder D, Bloom SR (1982) The effect of glucose-dependent insulinotropic polypeptide infused at physiological concentration on the release of insulin in man. Diabetologia 22: 34-36 Scatchard G (1949) The attraction of proteins for small molecules and ions. Annals NY Acad Sci 51: 660-672 Schafer R, Schatz H (1979) Stimulation of (pro-) insulin biosynthesis and release by gastric inhibitory polypetide in isolated islets of rat pancreas. Acta Endocrinol 91: 493-500 Schauder P, Brown JC, Frerichs H, Creutzfeldt W (1975) Gastric inhibitory polypeptide: Effects on glucose- induced insulin release from isolated rat pancreatic islets in vitro. Diabetologia 11: 483-484 Schülter KJ, Petersen KG, Kerp L (1980) Rapid change in insulin binding affinity of monocytes in female and male volunteers after breakfast. Diabetologia 19: 313 Schulz B, Ratzmann KP, Albrecht G , Bibergeil H (1980) Diurnal rhythm of insulin sensitivity (IS) in subjects with impaired glucose tolerance. Diabetes 39 [Supp ID: 160 Schwartz CK, Kimberg DV, Sheer in HE, Field M, Said SI (1974) Vasoactive intestinal peptide stimulation of adenylate cyclase and active electrolyte secretion in intestinal mucosa. J Clin Invest 54: 536-544 Service FJ, Nelson RL, Rubinstein AH, Go VLW (1978) Direct effect of insulin on secretion of insulin, glucagon, gastric inhibitory polypeptide and gastrin during maintainanceof normoglycaemia. J Clin Enocrinol Metab 47: 488-493 Shima K, Tanaka A, Shin S, Hashimura E, Kumahara Y (1979) Fluctuations of fasting plasma motilin in man. In: Miyoshi A (ed) Gut pepides. Kodansha, Tokyo, Elsevier/ North Holland, Amsterdam, pp 100-104 Siegel EG, Creutzfeldt W (1985) Stimulation of insulin release in isolated rat islets by GIP in physiological page 276 concentrations and its relation to islet cyclic AMP content. Diabetologia 28: 857-861 Sircus W (1958) Studies on the mechanisms in the duodenum inhibiting gastric secretion. 0 J Exp Physiol 43: 114-144 Solcia E, Capella C, Vasallo G, Buffa R (1974) Endocrine cells of the gastric mucosa. In Rev Cytol 42: 223-286 Soon-Siong P, Debas HT, Brown JC (1979a) The evaluation of GIP as an enterogastrone. J Surg Res 26: 681-686 Soon-Siong P, Debas HT, Brown JC (1979b) Cholinergic inhibition of gastric inhibitory polypeptide (GIP) action. Gastroenterology 76: 1253 Sopwith AM, Hutton JC, Naber SP, Chick WL, Hales CN (1981) Insulin secretion by a transplantable rat islet cell tumour. Diabetologia 21: 224-229 Starich GH, Bar RS, Mazzaferri EL (1985) GIP increases insulin receptor affinity and cellular sensitivity in adipocytes. Am J Physiol 249: E603-E607 Stevens JF (1971) Determination of glucose by an automatic analyzer. Clin Chim Acta 32: 199-201 Stinshoff K, Weisshaar D, Staehler F, Hesse D , Gruber W, Steier E (1977) Relation between concentrations of free glycerol and triglycerides in human sera. Clin Chem 23: 1029-1032 Strunz U, Domschke W, Domschke S, Mitznegg P, Wünsch E, Jaeger E, Demling L (1976) Gastroduodenal motor response to natural motilin and synthetic position 13-sustituted motilin analogues: a comparative in vitro study. Scand J Gastroent 11: 199-203 Strunz U, Domschke W, Mitznegg P, Domschke S, Schbert E , Wünsch E, Jaeger E, Demling L (1975) Analysis of the motor effects of 13-nor-leucine motilin on the rabbit, guinea-pig, rat and human alimentary tract in vi tro. Gastroenterology 68: 1485-1491 Sun JV, Tepperman HM, Tepperman J (1977) A comarison of insulin binding by liver plasma membranes of rats fed a high glucose diet or high fat diet. J Lipid Res 18: 533-539 Susini C, Lavau M (1978) In viv and in vitro responsiveness of muscle an adipose tissue to insulin in rats rendered obese by a high fat diet. Diabetes 27: 114-120 Sweeney JH (1927) Dietary factors that influence the dextrose tolerance test. Arch Intern Med 40: 818-830 Sykes S, Morgan LM, English J , Marks V (1980) Evidence for the preferential stimulation of gastric inhibitory polypeptide secretion in the rat by actively transported carbohydrates and their analogues. J Endocrinol 85: 201-207 Szecowka J , Grill V, Sandberg E, Efendic S (1982) Effect of GIP on the secetion of insulin and somatostatin and the accumulation of cyclic AMP in vittro in the rat. Acta Endocr 99: 416-421 Szurszewski JH, (1969) A migrating electric complex of the canine small intestine. Am J Physiol 217: 1757-1763 Tai HH, Chey WY (1978) Development of radioimmunoassay for motilin. Analyt Biochem 87: 350-358 Takemura J, Seino Y, Yamamura T, Tsuda K , Seino S, Itoh N, Imura H (1982) The role of endogenous gastric inhibitory polypeptide in the enteroinsular axis. J Clin Endocr Metab 54: 909-913 Tan K, Flatt PR, Webster JD, Marks V (1981) Glucose homeostasis during proliferation of a transplantable rat page 277 insulinoma. Diabetologia 21: 514 Tatemoto K, Mutt V (1980) Isolation of two novel candidate hormones using a chemicalmethod for finding naturally occuring polypeptides. Nature 285: 417-418 Thomas FB, Mazzaferri EL, Crockett SE, Mekhjian HS, Gruemer HD, Cataland S (1976) Stimulation of secretion of gastric inhibitory polypeptide and insulin by intraduodenal amino acid perfusion. Gastroenterology 70: 523-527 Thomas FB, Sinar S, Mazzaferri EL, Cataland S, Mekhjian HS, Caldwell JH, Fromkes JJ (1978) Selective release of gastric inhibitory polypeptide by intraduodenal amino acid perfusion in man. Gastroenterology 74: 1261-1265 Thomas PA, Kelley KA, Go VLW (1979) Does motilin regulate canine interdigestive gastric motility? Dig Dis Sci 24; 577-582 Unger RH, Eisentraut AM (1969) Entero-insular axis. Arch Intern Med 123: 261-266 Unger RH, Ketterer H, Dupré J, Eisentraut AM (1967) The effects of secretin, pancreozymin and gastrin on insulin and glucagon secretion in anaesthetized dogs. J Clin Invest 46: 1630— 1642 Vagne M, Stening GF, Brooks FP, Grossman MI (1968) Synthetic secretin: Comparison with natural secretin for potency and spectrum of physiological actions. Gastroenterology 55: 260-267 Vassallo G, Capella C, Solcia E (1971) Endocrine cells of the human gastric mucosa. Z Zellforsch 118: 48-67 Verdonk CA, Rizza RA, Nelson RL, Go VLW, Gerich JE, Service FJ (1980) Interaction of fat stimulated gastric inhibitory polypeptide on pancreatic alpha cell and beta cell function. J Clin Invest 65: 1119 Vi liar HV, Fender HR, Rayford PL, Bloom SR, Ramus NI, Thompson JC (1976) Suppression of gastrin release and gastric secretion by gastric inhibitory polypeptide (GIP) and vasoactive intestinal polypeptide (VIP). Ann Surg 184: 97-102 Walker (1977) An approach to immunoassay. Clin Chem 23: 384-402 Way LW (1970) Effect of secretin on gastric acid secretion in response to cholinergic stimuli. Gastroenterology 59: 510-517 Wasada T, McCorkle K, Harris V, Kawai K (1981) Effect of gastric inhibitory polypeptide on plasma levels of chylomicron triglyceride in dogs. J Clin Invest 68: 1106-1107 Welle HFT, Baartscheer T, Fiolet JWT (1984) Influnce of free glycerol on enymatic evaluation of triglycerides. Clin Chem 30: 1102-1103 Wieland OH (1984) Glycerol-UV method. In: Bergmeyer HU (ed) Methods in enzymatic analysis, Vol VI, 3rd edn., Verlag Chemie, Weinheim, pp 504-510 Williams RH, May JM, Biesbroeck JB (1981) Determinants of gastric inhibitory polypeptide and insulin secretion. Metabolism 30: 36-40 Willms B, Ebert R, Creutzfeldt W (1978) Gastric inhibitory polypeptide (GIP) and insulin in obesity: II Reversal of increased responsivenes to stimulation by starvation or food restriction. Diabetologia 14: 379-389 Wingate DL, Ruppin H, Thomson HH, Green WE, Domschke W, Wünsch E, Demling L, Ritche HD (1975) 13-norleucine motilin versus pentagastrin : contrasting and competitive page 278 effects on gastrointestinal effects on gastrointestinal myoelectrical activity in the conscious dog. Acta Hepatogastroenterologica (Stuttgart) 22; 409-410 Wolfe MM, Hocking MP, Maico DG, McGuigin JE (1983) Effects of antibodies to gastric inhibitory polypeptide on gastric acid secretiion and gastrin release in the dog. Gastroenterol 84: 941-948 Wolfe MM, Reel GM (1986) Inhibition of gastrin release by gastric inhibitory polypeptide mediated by somatostatin. Am J Physiol 250: G331-G335 Wolffbrandt KH, Damm Jorgensen K, Moody AJ, Pederson PC (1986) The effects of porcine GIP on insulin secretion and glucose clearance in the pig. Horm Metab Res 18: 159-162 Wormsley KG, Grossman MI (1964) Inhibition of gastric acid secretion by secretin and by exogenous acid in the duodenum. Gastroenterology 47: 72-81 Wright JP, Barbazat GO, Chain JE (1979) Jejunal secretiion in response to a mixed meal nutrient perfusion. Gastroenterology 76: 94-98 Wünsch E, Brown JC, Deimer K-H, Drees F, Jaeger E, Musiol J, Scharf R, Stocker H, Thamm P, Wendelberger G (1973) The total synthesis of norleucine-13-motilin. Z Naturforsch 28: 235-240 Yajima H, Ogawa H, Kubota M, Tobe T, Fujimura M, Henmi K, Torizuka K, Adachi H, Imura H, Taminato T (1975) Synthesis of the tricontapeptide corresponding to the entire amino acid sequence of gastric inhibitory polypeptide. J Am Chem Soc 97: 5593-5594 Yanaihara N, Yanaihara C, Nagai K, Sato H, Shimizu F , Yamaguchi K, Abe K (1980) Motilin-like immunoreactivity in porcine, canine, human and rat tissues. Biomed Res 1: 76-83 Yalow RS, Berson SA (1959) Assay of plasma insulin by immunological methods. Nature 184: 1648-1649 Zaragoza-Hermans N, Felber JP (1970) Studies on the metabolic effects induced in the rat by a high fat diet. 1. Carbohydrate metabolism in vivo. Horm Metab Res 2: 323-329 Zunz E, Labarre J (1929) Contributions à 1 'étude des variations physiogiques de la secretion interne du pancreas: Relation entre les secretion externe et interne du pancreas. Arch Int Physiol 31: 20-44 page 279 PUBLICATIONS Arendt J, Hampton S, English J, Kwasowski P, Marks V (1982) 24-hour profiles of melatonin, cortisol, insulin, C-peptide and GIP following a meal and subsequent fasting- Clin Endocr 16: 89-95 Bailey CJ, Flatt PR, Kwasowski P, Adams M (1986) Gastric inhibitory polypeptide and the entero-insular axis in streptozotocin diabetic mice. Diab Metab (in press) Bailey CJ, Flatt PR, Kwasowski P, Marks V (1985) GIP and the enteroinsular axis in hyperinsulinaemic and hypoinsulinaemic diabetic mice. Reg Pep 13: 90 (Abstract) Bailey CJ, Flatt PR, Kwasowski P, Powell CJ, Marks V (1986) Immunoreactive gastric inhibitory polypeptide and K cell hyperplasia in obese hyperglycaemic (ob/ob) mice fed high fat and high carbohydrate cafeteria diets. Acta Endocrinologica 112: 224-229 Flatt PR, Bailey CJ, Kwasowski P, Page T, Gee SM, Swanston- Flatt SK, Marks V (1983) Hypergipaemia in genetically obese hyperglycaemic (ob/ob) mice: Relationship to hyperphagia and intestinal K-cell insulin insensitivity. Scand J Gastroenterol 18 CSuppl 87 3: 110-111 Flatt PR, Bailey CJ, Kwasowski P, Page T, Marks V (1984) Plasma immunoreactive gastric inhibitory polypeptide in obese hyperglycaemic (ob/ob) mice. J Endocr 101: 249-256 Flatt PR, Bailey CJ, Kwasowski P, Page T, Marks V (1984) Effects of exogenous and endogenous GIP in obese hyperglycaemic (ob/ob) mice. Biochem Soc Trans 12: 790-791 Flatt PR, Bailey CJ, Kwasowski P, Page T, Marks V (1984) Pathological role of gastric inhibitory polypeptide in hyperinsulinaemia of obese hyperglycaemic (ob/ob) mice. In: Proceedings of the 7th International Congress of Endocrinology, Quebec, International Congress Series No. 652, Excerpta Medica, Amsterdam, p 444 (abstract) Flatt PR, Bailey CJ, Kwasowski P, Swanston-Flatt SK, Marks V (1983) Abnormalities of GIP in spontaneous syndromes of obesity and diabetes in mice. Diabetes 32: 433—435 Flatt PR, Bailey CJ , Kwasowski P, Swanston-Flatt SK, Marks V (1985) Glucoregulatory effects of caffeteria feeding and diet restriction in genetically obese hyperglycaemic (ob/ob) mice. Nutr Rep Int 32: 847-854 Flatt PR, Bailey CJ, Swanston-Flatt SK, Best L, Kwasowski P, Buchanan KD, Marks V (1984) Involvement of glucagon and GIP in the metabolic abnormalities of obese hyperglycaemic (ob/ob) mice. In: Lessons from animal diabetes. Shafir E and Renold AE (eds). John Libbey, London and Paris. p341-347 Flatt PR, Bailey CJ, Swanston-Flatt SK, Kwasowski P, Buchanan KD, Marks V (1982) Abnormalities of glucagon and GIP in page 280 ob/bo mice and their involvement in hyperinsulinaemia, hyperglycaemia and obesity. In: Proceedings of International Workshop on Lessons from Animal Diabetes, Jerusalem, p76 (abstract) Flatt PR, Kwasowski P, Swanston-Flatt SK, Marks V, Bailey CJ (1983) Immunoreactive GIP in plasma and intestine of obese hyperglycaemic (ob/ob) mice. Scand J Gastroenterol 18 CSuppl 833: 213-214 Hampton SM, Kwasowski P, Dunne M, Marks V (1981) Divergence between the glycaemic response to food ingestion and the function of the entro-pancreatic axis in volunteers fed a test meal at different times over 24 hours. Diabetologia 21: 6 (abstract) Hampton SM, Kwasowski P, Tan K, Morgan LM, Marks V (1982) Inhibitory effect of insulin on fat stimulated GIP release in rats. Reg Pep 3: 72 (abstract) Hampton SM, Kwasowski P, Tan K, Morgan LM, Marks V (1983) Effect of pretreatment with a high fat diet on the gastric inhibitory polpeptide (GIP) and insulin response to oral triolein and glucose in rats. Diabetologia 24: 278-281 Hampton SM, Tredger JA, Kwasowski P, Morgan LM, Wright J, Cramb R, Dunne M, Marks V (1983) Modification in control of the enteroinsular axis by a high fat’ diet. Diabetologia 25: 161 (abstract) Kwasowski P, Flatt PR, Burr KE, Marks V (1985) Effects of chain length and degree of saturation of fatty acids on IR-GIP release in genetically obese hyperglycaemic (ob/ob) mice. Diab Res Clin Prac CSuppl 13: S324 (abstract) Kwasowski P, Flatt PR, Bailey CJ, Marks V (1985) Effects of fatty acid chain length and saturation on gastric inhibitory polypeptide release in obese hyperglycaemic (ob/ob) mice. Biosci Rep 5: 701-705 Kwasowski P, Hampton SM, English J, Arendt J, Morgan LM, Marks V (1982) Circadian variations in plasma immunoreactive motilin concentrations. Reg Pep 4: 370 (abstract) Kwasowski P, Wright J, Marks (1984) Endogenous gastric inhibitory polypeptide (GIP) does not affect removal of triacylglycerol from plasma in man. Clin Sci 67: 18p (abstract) Kwasowski P, Tan KS, DeSilva M , Marks V (1984) Increased chylomicronaemia in fat fed rats given gastric inhibitory polypeptide antibodies. Diabetologia 27: 300A-301A (abstract) Lucey MR, Wass JA, Fairclough PD, 0 Hare M, Kwasowski P, Penman E, Webb J, Rees LH (1984) Does gastric acid release plasma somatostatin in man? Gut 25: 1217— 1220 Lucey MR, Fairclough PD, Wass JA, Kwasowski P, Medbak S , Webb page 281 J, Rees LH (1984) Response of circulating somatostatin, insulin, gastrin and GIP to nutrients in normal man. Clin Endocrinol (Oxf) 21; 209-217 Morgan LM, Tredger JA, Hampton SM, Kwasowski P, Wright J , Dunne M, Marks V (1983) Effect of diet upon response to oral fat and glucose in man; modification in control of the enteroinsular axis. Scand J Gastroenterol 18 CSuppl 873: 99-101 Morgan LM, Tredger JA, Madden A, Kwasowski P, Marks V (1985) The effect of guar gum on carbohydrate-, fat- and protein- stimulated gut hormone secretion: modification of postprandial gastric inhibitory polypeptide and gastrin responses. Brit J Nutr 53: 467-475 Morgan L, Tredger J, Wright J, Kwasowski P, Marks V (1982) The effect of guar supplementation on gastrin, motilin and GIP secretion following a high protein meal. Reg Pep 3: 78 (abstract) Salminen S, Salminen E, Porkka L, Kwasowski P, Marks V (1982) Effects of xylitol on gastric emptying, intestinal transit and motilin release. Reg Pep CSuppl 23: S90 (abstract) Tan KS, Kwasowski P, Marks V (1983) GIP-stimulated insulin secretion in a transplantable rat insulinoma. Reg Pep 7: 302 (abstract) Tan KS, Kwasowski P, Marks V (1984) A possible role for GIP in the differential response to oral and intravenous glucose in the transplantable rat insulinoma - A role for GIP. Clin Sci 67: 68p (abstract) Tan KS, Kwasowski P, MarksV (1985) Differential insulin responses to oral and iv glucose in the transplantable rat insulinoma - A role for GIP. Reg Pep 13: 163-168 page 282 249 Plasma immunoreactive gastric inhibitory polypeptide in obese hyperglycaemic {objob) mice P. R. Flatt, C. J. Bailey*, P. Kwasowski, T. Page* and V. Ma r k s Divisions of Nutrition and Fo o d Science and of Clinical Biochemistry, Department of Biochemistry, University of Surrey, Guildford g u 2 5XH *Department of Biological Sciences, University of Aston in Birmingham, Birmingham B4 7ET RECEIVED 31 August 1983 ABSTRACT Gastric inhibitory polypeptide (GIP), a recognized protracted in the presence of rising hyperglycaemia. component of the enteroinsular axis, is raised in the Orally administered fat, glucose and amino acids plasma and intestine of obese hyperglycaemic {objob) raised G I P concentrations with fat having a parti mice. T o evaluate the control of plasma G I P and its cularly strong effect. Glucose and amino acids also role in the hyperinsulinaemia of the objob syndrome, evoked prominent increases of insulin, but fat G I P and insulin were determined at different ages in produced only a small rise in insulin in the absence of fed mice, and at 10-12 weeks of age after increasing glucose concentrations. Consistent with fasting/refeeding and administration of GIP, different glucose-potentiation, a mixture of all three nutrients nutrients and insulin to mice fasted for 18 h. Plasma greatly augmented the insulin response without further G I P and insulin were raised in adult (10- and 20-week- increase of plasma GIP. Glucose-induced increase in old) compared with younger (3- and 5-week-old) mice, endogenous insulin and doses of exogenous insulin up although G I P was not increased in the presence of to 100 units/kg did not suppress basal, fat-stimulated hyperinsulinaemia at 3 weeks of age. Fasting or glucose-stimulated GI P release. The results indicate suppressed and refeeding promptly restored plasma that raised G I P concentrations m a k e an important G I P and insulin concentrations. Administration of contribution to the hyperinsulinaemia and related G I P to mimic postprandial concentrations evoked a metabolic abnormalities of the objob syndrome. marked but transient insulin response which was J. (1984) 101,249-256 INTRODUCTION and possibly fatty acids contributes in this respect (Berne, 1974; Flatt & Bailey, 1982a, 6, 1984). H o w The objob syndrome in mice is the most extensively ever, the importance of neural and endocrine c o m studied animal model of spontaneous obesity and ponents of the enteroinsular axis is highlighted by the diabetes (Herberg & Coleman, 1977; Bray & York, presence of an insulin response to oral but not 1979). Characteristically these mice exhibit hyper parenteral glucose, and the augmented insulin phagia, marked obesity, moderate hyperglycaemia response to feeding in conditioned mice (Flatt & and severe hyperinsulinaemia (Bailey, Flatt & Atkins, Bailey, 1981a, 1983). 1982). Increased insulin concentrations constitute a Studies in m a n and experirnental animals have major pathogenic influence, promoting excessive tri implicated gastric inhibitory polypeptide (GIP; also glyceride deposition and glucose intolerance through k n o w n as glucose-dependent insulinotrophic poly insulin resistance and pancreatic A-cell dysfunction peptide) as a physiological component of the entero (Assimacopoulos-Jeannet & Jeanrenaud, 1976; Flatt insular axis (Brown, 1982; Creutzfeldt, Ebert, Na u c k & Bailey, 1981a; Flatt, Bailey & Buchanan, 1982). In & Stockmann, 1983). Consistent with observations in fed adult o6/o6 mice the extent of hyperinsulinaemia is h u m a n obesity and type 2 (non-insulin-dependent) related to both the quantity and composition of the diabetes mellitus (Brown, 1982; Creutzfeldt et al. food ingested (Bailey et al. 1982; Flatt & Bailey, 1982a, 1983), increased G I P concentrations have been 1984). A direct insulinotrophic effect of amino acids reported in genetically obese-diabetic objob and dbjdb J. Endocr. ( 1984) 101,249-256 © 1984 Journal of Endocrinology Ltd Printed in Great Britain 0022-0795/84/0101-0249 $02.00/0 250 p. R. FLATT and others • Plasma GIP in o b /o b mice mice (Flatt, Bailey, Kwasowski et al. 1983). This study was not possible due to the large volume of plasma investigates the regulation of immunoreactive GI P in required for GI P determination (200 pi compared with objob mice, and evaluates the involvement of patho 20 pi for objob mice). Plasma was separated and stored logically raised G I P concentrations in the hyper as described previously (Flatt et al. 1982). Basal insulinaemia and associated features of the syndrome. plasma glucose, insulin and GI P concentrations were measured in fed mice at 3, 5, 10 and 20 weeks of age. The effects of fasting for 24 h and refeeding with a MATERIALS AND METHODS standard diet available ad libitum were examined at 10 weeks of age. F o o d intake was monitored in this Animals experiment. All other studies were conducted using Obese hyperglycaemic {objob) mice and lean (-f/?) mice fasted for 18 h at 10-12 weeks of age. Food, but mice from the colony maintained at the University of not water, was withheld during the test procedures. Aston in Birmingham were used. The origin and The following nutrients were administered by gastric characteristics of Aston objob mice have been intubation in a volume corresponding to 8 ml/kg body described in detail elsewhere (Flatt & Bailey, 19816; weight: {a) fat emulsion (Intralipid; Kabi Vitrum Ltd, Bailey et al. 1982). Mice were housed in an air- Ealing) containing 770 mg fractionated soybean oil/kg conditioned ro o m at 22 + 2 °C with a lighting schedule body weight, 87 mg glycerol/kg body weight and 46 mg of 12 h light (08.00-20.00 h) and 12 h darkness. A fractionated egg lecithin/kg body weight (32-2kJ/kg), standard pellet diet (Mouse breeding diet, Heygate & (6) 2 g glucose/kg body weight (32-2kJ/kg), (c) a mix Sons Ltd, Northampton) and tap water were available ture of essential and non-essential amino acids (Syn- ad libitum. The standard diet comprised 2-5% fat, thamin 17; Travenol Laboratories Ltd, Thetford, Suf 17-6% protein and 46-8% carbohydrate (digestible folk) containing 0-8 g L-amino acids/kg body weight energy 876 kJ/kg) with fibre, vitamins and minerals as (135-2m g nitrogen/kg) and {d) a fat-glucose-amino described elsewhere (Flatt & Bailey, 1982a). acids mixture comprising equal volumes of the three preparations. In addition, 2 g glucose/kg body weight, 40 pg porcine GIP/kg body weight (Professor J. C. Experimental procedures Brown, University of British Columbia, Canada), and Mice were adapted to the manipulative procedures for 5 and 100 units m o n o c o m p o n e n t porcine insulin/kg 2 weeks before study. Blood samples were obtained body weight (Actrapid; No v o Industria, Copenhagen) from the cut tip of the tail in the conscious state at the were given by intraperitoneal injection in a volume of times shown in the figures. Serial sampling of lean mice 5 ml/kg body weight. 15 30 1500 10 z 20 1000 E 10 500 3 5 10 20 3 5 10 20 3 5 10 20 Age (weeks) FIGURE 1. Changes with age in plasma concentrations of glucose, insulin and gastric inhibitory polypeptide (GIP) in fed obese {objob) mice. Values are means+S.E.M. of groups of seven to ten mice. *P<0-05 compared with 3 and 5 weeks of age; fP<0-05 compared with 3 weeks of age (Student’s unpaired r-test). J. Endocr. (1984) 101,249-256 Plasma G IP in ob/ob mice • p. r. fla tt and others 251 Analyses Fasting Refeeding ' 1 1 Plasma glucose was measured by an automated 15 glucose oxidase procedure (Stevens, 1971) and insulin was determined by dextran-coated charcoal radio if immunoassay using crystalline mouse insulin (Novo 10 ■ V r Industria) as standard (Flatt & Bailey, 19816). G I F |i was measured by radioimmunoassay (Morgan, Morris & Marks, 1978) using donkey anti-rabbit g a m m a globulin antiserum (Guildhay Antisera, University of --1_____ II_____ 1 —I------1— Surrey, Guildford) for separation of free from bound antigen. Porcine G I P (J. C. Brown) was used to prepare ^^^I-labelled tracer and as standard. Im m u n o - adsorbed hormone-free plasma was used to minimize non-specific interference and parallelism was de m o n strated between the standard curve and serially diluted objob mouse plasma. The GI P antiserum (RIC 34/111; I* Guildhay Antisera) was raised in a rabbit against a porcine GIP-glutaraldehyde-ovalbumin conjugate. This antiserum recognizes both 5000 and 8000 molecular forms of GIP, and exhibits negligible cross reactivity with other enteropancreatic hormones. 3000 Assay sensitivity was llOng/1 and the interassay coefficient of variation was 4-1% at 2677 ng/1 and S s 2000 22-5% at 138 ng/1. Groups of data were compared using Student’s paired and unpaired /-tests as 1000 - appropriate. Differences were considered to be significant at P < 0-05. Iff RESULTS Ilf Basal plasma GIF concentrations (Fig. 1) T o investigate changes of GI P with age in relation to Time (h) other features of the objob syndrome, plasma GIP concentrations were determined in fed objob mice at 3, FIGURE 2. Effect of a 24-h fast and refeeding with standard 5, 10 and 20 weeks of age. Consistent with previous diet available ad libitum on plasma concentrations of glu observations, the onset of hyperinsulinaemia preceded cose, insulin and gastric inhibitory polypeptide (GIP) in 10- the development of hyperglycaemia (Bailey et al. to 12-week-old obese {objob) mice. Values are means + S.E.M. of eight mice. All values are significantly (P < 0 05) 1982). Plasma G I P concentrations, like those of raised compared with time zero (Student’s paired /-test). insulin, were higher at 10-20 weeks of age compared Mean values for food intake are shown. with younger mice. As reported elsewhere (Flatt et al. 1983), the GI P concentrations of adult objob mice are substantially greater than age-matched lean ( -f- /?) mice. However, in 3-week-old weaning mice G I P Responses to fasting and refeeding (Fig. 2) concentrations were not significantly different between the objob mutants (782 + 163 ng/1 (s.e.m .), n = 9) and In view of the k n o w n relationship between feeding lean ( + /?) controls (752 +1 8 8 ng/1, n = 7), although activity and hyperinsulinaemia in objob mice (Flatt & plasma insulin concentrations were already raised in Bailey, 1982a, 1984; Bailey et al. 1982), GI P responses objob mice at this age (4 3 + 0 7 pg/1 and 1 4 + 0 5 pg/l to fasting and refeeding were examined. Plasma GI P in objob and +/? mice respectively). Interestingly, the concentrations fell during fasting, coincident with a G I P concentrations of weaning lean ( + /?) mice were marked reduction of insulin and a decline in glucose. higher than reported in adult life (Flatt et al. 1983), Refeeding with a standard diet available ad libitum possibly reflecting the relatively high proportion of fat evoked prominent and sustained increases in all three and galactose in the milk. parameters. J. Endocr. (1984) 101,249-256 252 P. R. FLATT and others • Plasma GIP in o b /o b mice (a) 30 STbSIs i 10 h 111 0 -I____ L 2000 0 ^ 1500 K, E&iwm t i 500 0 J I L 0 30 60 120 0 30 60 120 0 30 60 0 30 60 Time (min) FIGURE 3. Effects of orally administered (a) fat emulsion, (6) glucose, (c) amino acids and (d) a fat emulsion- glucose-amino acids mixture on plasma glucose, insulin and gastric inhibitory polypeptide (GIP) in 10- to 12-week-old obese {objob) mice fasted for 18 h. Values are means ± s.e.m, of six mice. *P < 0 05 compared with time zero (Student’s paired /-test). Responses to fat, glucose and amino acids(Fig. 3) insulin and possibly glucose as such (Brown, 1982; Creutzfeldt et al. 1983; M a r k s & Morgan, 1984), the Fat, glucose and amino acids were administered orally, effects of intraperitoneal glucose and insulin on basal alone and as a mixture to objob mice fasted for 18 h, to G I P concentrations were examined in objob mice. evaluate the role of individual nutrients in the GI P and Plasma GI P was not affected either by glucose or by insulin responses to feeding. All nutrients raised GI P insulin at a low or inordinately high dose (5 and 100 concentrations, although a particularly strong G I P units insulin/kg). response was generated by the fat emulsion. However, fat and amino acids produced smaller insulin Effects of insulin on responses to fat and glucose(Fig. 6) responses than glucose, while a combination of the three nutrients had a marked insulinotrophic effect. T o investigate the control of plasma GI P further, the effect of exogenous insulin on the GI P responses to fat Effects of glucose on response to fat(Fig. 4) and glucose was examined in objob mice. Consistent with the lack of effect on basal G I P concentrations, The extent to which glucose potentiates the insulin- doses of insulin up to 100 units/kg failed to affect either releasing effect of fat was examined by intraperitoneal fat- or glucose-stimulated G I P release, although the injection of glucose 30 min after oral administration of insulin was effective in suppressing plasma glucose. the fat emulsion. In this experiment the increase of plasma G I P was associated with raised glucose Effects of exogenous GIF(Fig. 7) concentrations and a marked insulin response which considerably exceeded that produced by intra The insulinotrophic action of exogenous G I P was peritoneal glucose alone (Flatt & Bailey, 1981a). evaluated in objob mice in the absence and presence of a glycaemic response induced by intraperitoneal glucose. Under conditions of basal hyperglycaemia, Effects of glucose and insulin on basal GIF(Fig. 5) G I P evoked a marked and transient insulin response Since there are reports that GI P secretion is suppressed without change in plasma glucose. In the presence of by endogenous insulin released by glucose, exogenous rising hyperglycaemia after glucose administration. J. Endocr. (1984) 101,249-256 Plasma GIP in ob/ob mice - p. R. flatt and others 253 G I P produced a similarly marked but considerably 2035 ng GIP/1 at 0, 5, 10, 15 and 30 min respectively) protracted insulin response which cannot be explained were similar to those observed in actively feeding objob by a direct effect of glucose alone (Flatt & Bailey, mice (Flatt et al. 1983). The concentration declined to 1981a). Concentrations of GI F achieved during these 693 ng GIP/1 by 90 min indicating a relatively short experiments (mean values 654, 3383, 5175, 4955 and half-life of this hormone in objob mice. 1000 Ü 500 2000 I I 1000 Time (min) n o u R E 5. Effect of intraperitoneally administered glucose (2 g/kg; #), a low dose of insulin (5 units/kg; A) or a high Time (min) dose of insulin (100 units/kg; ■) on plasma concentrations of glucose and gastric inhibitory polypeptide (GIP) in 10- to 12-week-old obese {objob) mice fasted for 18 h. Values FIGURE 4. Effect of glucose (#) or saline (O) on fat-stimu are means ± s.e.m. of six mice. All plasma glucose levels lated plasma concentrations of gastric inhibitory polypep at 30 and 60 min are significantly different {P < 0 05, tide (GIP) and on glucose and insulin in 10- to 12-week-old Student’s paired /-test) from time zero. obese {objob) mice fasted for 18 h. Fat emulsion was ad ministered orally at time zero and glucose or an equivalent volume (5 ml/kg) of saline (0 154 mm o l NaCl/1) was admini stered intraperitoneally at 30 min. Values are means+ s.e.m. of six mice. * P < 0 05 compared with saline-treated mice (Student’s unpaired /-test). J. Endocr. (1984) 101,249-256 254 P. R. FLATT and others • Plasma GIF in o b /o b mice 30. - W 20 10 0 2000 1500 Ü 1000 500 0 0 30 60 90 0 30 60 90 Time (min) FIGURE 6. Effect of saline (0-154 mm o l NaCl/1; O), a low dose of insulin (5 units/kg; A) or a high dose of insulin (100 units/kg; ■) administered intraperitoneally at 30 min, on plasma concentrations of glucose and gastric inhibitory polypeptide (GIF) after stimulation by (a) fat emulsion and (b) glucose administered orally at time zero in 10- to 12-week-old obese (ob/ob) mice fasted for 18 h. Values are means+s.e.m. of six mice. There were no significant differences of plasma GIF values between the three groups of mice. 40 10 •— 0 20 10 0 0 5 15 30 0 5 15 30 Time (min) J. Endocr. (1984) 101,249-256 Plasma GIP in ob/ob mice • p. r. fla tt and others 255 DISCUSSION the insulin-releasing actions of fat and amino acids are potentiated by dietary carbohydrate (Flatt & Raised plasma GI P concentrations have been reported Bailey, 1984) mediated partly through raised G I P in some cases of h u m a n obesity, untreated type 1 concentrations. (insulin-dependent) diabetes and type 2 diabetes Studies in m a n and laboratory animals have (Bloom, 1975; Crockett, Mazzaferri & Cataland, 1976; suggested that under normal physiological conditions, M a y & Williams, 1978; Ross & Dupre, 1978; feedback inhibition by insulin on intestinal K-cell Deschamps, Heptner, Desjeux et al. 1980; Brown, function plays a role in the regulation of basal and 1982; Salera, Giacomoni, Pironi èt al. 1982; nutrient-stimulated GIÇ,concentrations (Brown, 1982; Creutzfeldt et al. 1983), and in certain spontaneously Creutzfeldt et al. 1983; Hampton, Kwasowski, Tan et obese-diabetic rodents, namely glycosurie Djungarian al. 1983; Ma r k s & Morgan, 1984). A breakdown in this hamsters (Buchanan, Alam, Hanna et al. 1982) and mechanism has been suggested as a factor contributing both objob and dbjdb mice (Flatt et al. 1983). The to the increased GI P concentrations in certain cases of present study evaluates the raised plasma G I P h u m a n obesity and diabetes mellitus (Brown, 1982; concentrations of objob mice and demonstrates an Creutzfeldt et al. 1983; M a r k s & Morgan, 1984). In important physiological role of G I P in promoting objob mice, glucose-induced increase in endogenous insulin secretion, especially in the presence of rising insulin and amounts of insulin tested which would be hyperglycaemia. Thus, the insulinotrophic action of fatal in lean mice did not suppress basal, fat-stimulated exogenous GI P was potentiated by glucose as shown or glucose-stimulated G I P concentrations. Thus in elsewhere (Brown, 1982; Szecowka, Grill, Sandberg & sulin at the inordinately large dose used in the present Efendic, 1982), and a powerful insulin response to study was an ineffective inhibitor of G I P secretion fat-stimulated G I P release was unmasked by an in objob mice. accompanying increase in the level of glycaemia. Since Although the physiological action of G I P in glucose alone constitutes a poor insulinotrophic normoglycaemic m a n is equivocal (Verdonk, Rizza, stimulus in objob mice (Flatt & Bailey, 198In), these Nelson et al. 1980; Sarson, Wo o d , Holder & Bloom, observations indicate an important glucose-dependent 1982), this study has shown that hypersecretion of GI P role of GI P in the maintenance of hyperinsulinaemia. in response to feeding in objob mice is a major insulino The changes with age and the effects of fasting and trophic stimulus in the presence of rising hyper refeeding on plasma insulin and G I P concentrations glycaemia. Increased G I P concentrations therefore are consistent with this view. However, the onset of promote hyperinsulinaemia, although additional hyperinsulinaemia by 3 weeks of age, before the contributory factors include direct and glucose- increase of either basal GI P or glucose concentrations potentiated islet stimulation by amino acids and suggests the involvement of other factors in the neuroendocrine components of the enteroinsular axis aetiology of the hyperinsulinaemia. other than GI P (Bobbioni & Jeanrenaud, 1982; Ebert A plausible basis for raised GI P concentrations of & Creutzfeldt, 1982; Flatt & Bailey, 19826, 1983). objob mice resides in part with intestinal hypertrophy Accordingly, G I P participates in the metabolic associated with an increase in the number and abnormalities of objob mice through the hyperinsulin hormone content of G I P secreting K-cells (Polak, aemia and its associated pathological consequences Pearse, Grimelius & Marks, 1975; Flatt et al. 1983). In (Assimacopoulos-Jeannet & Jeanrenaud, 1976). addition, hypersecretion of GI P is thought to result However, other actions of GI P (Brown, 1982), such as from excessive alimentation and abnormal regulation stimulation of glucagon secretion, inhibition of gastric of K-cell function. Administration of fat, glucose and acid secretion and facilitation of lipogenesis in adipose amino acids confirmed that all nutrients raised G I P tissue, are also likely to be involved in the manifesta concentrations with a particularly marked response to tion of the objob syndrome. fat (Brown, 1982; Creutzfeldt et al. 1983). Whereas glucose and amino acids evoked prominent insulin responses, fat produced only a small rise in insulin ACKNOWLEDGEMENTS in the absence of an accompanying increase in glu cose. The effect of fat and the pronounced insulin re The authors acknowledge the excellent animal care sponse to the combined nutrient mixture indicate that provided by M. Gamble, P. Behan and B. Burford. FIGURE 7. Effect of intraperitoneally administered gastric inhibitory polypeptide (40 pg/kg) (a) alone and (6) together with glucose (2 g/kg) on plasma concentrations of glucose and insulin in 10- to 12-week-old obese (objob) mice fasted for 18 h. Values are means ±s.E.M. of four mice. *? < 0 05 compared with time zero (Student’s paired t-test). J. Endocr. (1984) 101,249-256 256 p. R. FLATT and others • Plasma GIP in o b /o b mice REFERENCES Flatt, P. R. & Bailey, C. J. (1984). Dietary components and plasma insulin responses to fasting and refeeding in genetically obese hyperglycaemic (objob) mice. British Journal o f Nutrition 51. Assimacopoulos-Jeannet, F. & Jeanrenaud, B. (1976). The hor Flatt, P. R., Bailey, C. J. & Buchanan, K. D. (1982). Regulation of monal and metabolic basis of experimental obesity. Clinics in plasma immunoreactive glucagon in obese hyperglycaemic Endocrinology and Metabolism 5, 337-365. Bailey, C. J., Flatt, P. R. & Atkins, T. W. (1982). Influence of (objob) mice. Journal o f Endocrinology 95, 215-227. Flatt, P. R , Bailey, C. J., Kwasowski, P., Swanston-Flatt, S. K. & genetic background and age on the expression of the obese Marks, V. (1983). Abnormalities of GIF in spontaneous syn hyperglycaemic syndrome in Aston objob mice. International dromes of obesity and diabetes in mice. Diabetes 32,433-435. Journal o f Obesity 6,11-21. Hampton, S. M., Kwasowski, P., Tan, K., Morgan, L. M. & Berne, C. (1974). Studies on the lipid metabolism in the pancreatic Marks, V. (1983). Effect of pretreatment with a high fat diet on B-cells of mice. Acta Universitatis Upsaliensis No. 215. the gastric inhibitory polypeptide and insulin responses to oral Bloom, S. R. (1975). GIF in diabetes. Diabetologia 11, 334. triolein and glucose in rats. Diabetologia 24,278-281. Bobbioni, E. & Jeanrenaud, B. (1982). Possible role of the CNS in Herberg, L. & Coleman, D. L. (1977). Laboratory animals ex regulating insulin secretion via humoral factors. International hibiting obesity and diabetes syndromes. Metabolism 26, 59-99. Journal o f Obesity 6, SappXA, 21-25. Marks, V. & Morgan, L. (1984). The entero-insular axis. In Recent Bray, G. A. & York, D. A. (1979). Hypothalamic and genetic advances in diabetes, vol. 1, pp. 55-71. Eds M. Nattrass & J. V. obesity in experimental animals: an autonomic and endocrine Santiago. Edinburgh: Churchill Livingstone. hypothesis. Physiological Reviews 59, 719-809. May, J. M. & Williams, R. H. (1978). The effect of endogenous Brown, J. C. (1982). Gastric inhibitory polypeptide. Monographs gastric inhibitory polypeptide on glucose-induced insulin secre on Endocrinology, vol. 24. Berlin: Springer-Verlag. tion in mild diabetes. Diabetes 27, 849-855. Buchanan, K. D., Alam, M. J., Hanna, K. A., Banks, I. G. & Morgan, L. M., Morris, B. A. & Marks, V. (1978). Radioimmuno Herberg, L. (1982). Pancreatic and gut hormones in the Djun assay of gastric inhibitory polypeptide. Annals o f Clinical Bio garian hamster. Regulatory Peptides 3,67. chemistry \5, 172-177. Creutzfeldt, W., Ebert, R., Nauck, M. & Stockmann, F. (1983). Polak, J. M., Pearse, A. G. E., Grimelius, L. & Marks, V. (1975). Disturbances of the entero-insular axis. Seandinavian Journal of Gastrointestinal apudosis in obese hyperglycemic mice. Virchows Gastroenterology 18, Suppl. 82, 111-119. Archiv. Section B: Cell Pathology 19, 135-150. Crockett, S. E., Mazzaferri, E. L. & Cataland, S. (1976). Gastric Ross, S. A. & Dupre, J (1978). Effects of ingestion of triglyceride inhibitory polypeptide (GIF) in maturity-onset diabetes mellitus. or galactose on secretion of gastric inhibitory polypeptide and on Diabetes 25, 931-935. response to intravenous glucose in normal and diabetic subjects. Deschamps, 1., Heptner, W., Desjeux, J.-F., Baltakse, V., Machi- Diabetes 27, 327-333. not, S. & Lestradet, H. (1980). Effects of diet on insulin and gas Salera, M., Giacomoni, P., Pironi, L., Cornia, G., Capelli, M., tric inhibitory polypeptide levels in obese children. Pediatric Marini, A., Benfenati, F., Miglioli, M. & Barbara, L. (1982). Research 14, 300-303. Gastric inhibitory polypeptide release after oral glucose: re Ebert, R. & Creutzfeldt, W. (1982). Influence of gastric inhibitory lationship to glucose intolerance, diabetes mellitus, and obesity. polypeptide antiserum on glucose-induced insulin secretion in Journal o f Clinical Endocrinology and Metabolism 55, 329-336. xsXs. Endocrinology 1601-1606. 111, Sarson, D. L., Wood, S. M., Holder, D. & Bloom, S. R. (1982). The Flatt, P. R. & Bailey, C. J. (1981a). Development of glucose in effect of glucose-dependent insulinotropic polypeptide infused at tolerance and impaired plasma insulin response to glucose in physiological concentration on the release of insulin in man. obese hyperglycaemic (objob) mice. Hormone and Metabolic Diabetologia 22,34-36. Research 13, 556-560. Stevens, J. F. (1971). Determination of glucose by an automatic Flatt, P. R. & Bailey, C. J. (19816). Abnormal plasma glucose and analyser. Clinica Chimica Acta 3199-201. 2 , insulin, responses in heterozygous lean (obj + ) mice. Diabetologia Szecowka, J., Grill, V., Sandberg, E. & Efendic, S. (1982). Effect 20, 573-577. of GIF on the secretion of insulin and somatostatin and the ac Flatt, P. R. & Bailey, C. J. (1982a). Role of dietary factors in the cumulation of cyclic AMP in vitro in the rat. Acta hyperinsulinaemia of genetically obese hyperglycemic (objob) Endocrinologica99,416-421. mice. Journal o f Nutrition 112, 2212-2216. Verdonk, C. A., Rizza, R. A., Nelson, R. L., Go, V. L. W., Gerich, Flatt, P. R. & Bailey, C. J. (19826). Plasma glucose and insulin J. E. & Service, F. J. (1980). Interaction of fat-stimulated gastric responses to glucagon and arginine in Aston objob mice: evi inhibitory polypeptide on pancreatic alpha and beta cell func dence for a selective defect in glucose-mediated insulin release. tion. Journal o f Clinical Investigation 65, 1119-1125. Hormone and Metabolic Research 14, 127-130. Flatt, P. R. & Bailey, C. J. (1983). Glucose and insulin response to conditioned feeding in lean and genetically obese hyperglycemic (objob) mice. Metabolism 32, 504-509. J. Endocr. (1984) 101,249-256 Bioscience Reports 5, 701-705 (1985) 701 Printed in Great Britain Effects of fatty acid chain length and saturation on gastric inhibitory polypeptide release in obese hyperglycaemic (ob/ob) mice Piotr KWASOWSKI!, Peter R. FLATJl, Clifford J. BAILEY2 and Vincent M ARKS! ^Department of Biochemistry, Divisions of Clinical Biochemistry and Nutrition & Food Science, University of Surrey, Guildford, U.K.; and '^Department of Molecular Sciences, Division of Biology, Aston University, Birmingham, U.K. (Received 12 August 1985) Pl a s m a gastric inhibitory polypeptide (GIP) responses to equimolar intragastrically administered emulsions of fatty acids (2.62 mmol/7.5 ml/kg) were examined in 18 h fasted obese hyperglycaemic (ob/ob) mice. Propionic acid (C3:0), a saturated short-chain fatty acid, and capric acid (C10:0), a saturated medium chain fatty acid, did not significantly stimulate GIP release. H o w e v e r , the saturated long-chain fatty acid stearic acid (C18;0), and especially the unsaturated long-chain fatty acids oleic (C18:l), linoleic (C18:2) and linolenic (C18:3) acids produced a marked GIP response. The results show that chain length and to a lesser extent the degree of saturation are important determinants of fatty acid-stimulated GIP release. T h e GIP-release action of long-chain, but not short-chain, fatty acids m a y be related to differences in their intracellular handling. Gastric inhibitory polypeptide (GIP) is secreted by the K-cells of the intestinal mucosa during the alimentary processing of food (Brown, 1982; Creutzfeldt, 1982; Marks & Morgan, 1983). In addition to its possible role in the control of gastric acid secretion (Brown, 1982), GIP exerts a glucose dependent insulin-releasing effect on the pancreatic islet B-cells, and constitutes a physiological c o m p o n e n t of the enteroinsular axis ( B r own, 1982; Creutzfeldt, 1982; Marks <5c Morgan, 1983). Studies in m a n y species, including man, have sho w n that GIP secretion is stimulated by carbohydrates, proteins and fats (Brown, 1982). Dietary fats provide an especially potent stimulus for G I P release (Brown et al., 1975; Falko et al., 1975; O'Dorisio et al., 1976), but the m e c h a n i s m is not established. Thus, in contrast to the prominent effect of triglycerides, variable GIP responses have been observed to fatty acids and mono- and di-glycerides (Ross & Shaffer, 1981; Williams et al., 1981). Recent studies have identified the genetically 702 KWASOWSKI ET AL. obese hyperglycaemic (ob/ob) m o u s e as a valuable model to investi gate G IP release. This model exhibits intestinal hypertrophy (Flatt et al., 1983) and enteroendocrine hyperplasia (Polak et al., 1975), with enhanced basal and nutrient-stimulated GIP concentrations (Flatt et al., 1983, 1984). T h e present study investigates the effects of fatty acid chain length and saturation on GIP release in ob/ob mice. Materials and Methods Groups of Aston obese hyperglycaemic (ob/ob) mice from the colony maintained at Surrey University were used at 12-18 weeks of age. The origin and characteristics of Aston ob/ob mice have been described elsewhere (Flat & Bailey, 1981; Bailey et al., 1982). Mice were housed in an air-conditioned room at 22 ± 2°C with a lighting schedule of 12 h light (0700-1900 h) and 12 h darkness. A standard pellet diet (Spratts Laboratory Diet 1, Lillico Ltd., Reigate, U.K.) and tap water were supplied ad libitum. This diet comprised 3.596 fat, 21. 5 % protein and 4 8 % carbohydrate (digestible energy 14.2 MJ/kg) with added fibre, vitamins and minerals. Fatty acids (Sigma Chemical Company Ltd., Poole, U.K.) were administered intragastrically as an aqueous emulsion (2.62 mmol/7.5 ml/kg) to conscious 18 h fasted mice. The dose corresponded with the overall composition of the fat emulsion (Intralipid, Kabi Vitrum Ltd., Ealing, U.K.) previously tested in ob/ob mice (Flatt et al., 1984). T h e fatty acids (with n umbers of carbon atoms: unsaturated carbon-carbon bonds, and dose in mg/kg) were: propionic acid (C3:0, 194 mg/kg), capric acid (C10:0, 451 mg/kg), stearic acid (C18:0, 746 mg/kg), oleic acid (C18:l, 741 mg/kg), linoleic acid (C18:2, 735 mg/kg) and linolenic acid (C18:3, 730 mg/kg). Blood samples (60 pi) were taken from the tail-tip of conscious mice immediately before and at 30, 60, 90 and 120 min after fatty acid administration. Plasma GIP was measured by radioimmunoassay (Morgan et al., 1978) using donkey anti-rabbit g a m m a globulin antiserum (Guildhay Antisera, University of Surrey, Guildford, U.K.) for separation of free from bound antigen. Porcine GIP (Professor J.C. Brown, University of British Columbia, Canada) was used to prepare 125i_Gip and as standard. Immunoadsorbed GIP-free human plasma was added to the tubes of the standard curve used to minimize non-specific interference, and parallelism was demonstrated between the standard curve and serially diluted ob/ob m o u s e plasma. T h e G I P antiserum (RIC 34/III; Guildhay Antisera) recognizes both 5000 and 8000 molecular forms of GIP. It exhibits negligible cross reactivity (less than 1 % ) with other entero-pancreatic hormones, including cholecystokinin, glucagon, gut glucagon-like immunoreactivity (enteroglucagon), insulin, pancreatic polypeptide, proinsulin C-peptide, secretin and vasoactive intestinal polypeptide. Groups of data we r e c o m p a r e d by analysis of variance ( A N O V A ) . Differences were considered to be significant if P < 0.05, and these were confirmed using Student's paired and unpaired t-tests as appropriate. Results Mean basal plasma GIP concentrations in the experimental groups of 18 h fasted ob/ob mice were 526-614 pg/ml (Fig. 1). Oral administration of propionic acid, a saturated short-chain (C3:0) fatty FATTY ACIDS AND GASTRIC INHIBITORY PEPTIDE 703 Propionic acid Capric acid Stearic acid 2000 (C10:0) (C18;0) A691 ± 350 A 216 ± 281 A1348 ± 436 j 1000 Oleic acid Linoleic acid Linolenic acid — 3000 (018:1) (018:2) (018:3) A2389 + 382 A2550 +691 A l 508 + 665 2000 1000 0 30 60 90 1200 30 60 90 120 0 30 60 90 120 Minutes Fig. 1. Effects of fatty acid chain length and saturation on plasma GIP concentrations in 18 h fasted ob/ob mice. Fatty acids were administered intragastrically as an aqueous emulsion at a dose of 2.62 mmol/7.5 ml/kg. Numbers of carbon atoms and unsaturated carbon-carbon bonds are shown in parentheses. A values represent the increment above basal for the GIP response expressed as pg*ml“^*b, calculated as described in the Results. Values are presented as means ± SEM of 5-6 mice. acid, and capric acid, a saturated medium-chain (Cl 0:0 ) fatty acid, did not significantly alter plasma GIP concentrations. However, the saturated and unsaturated long-chain fatty acids stearic acid (C18:0), oleic acid (C18:I), linoleic acid (C18:2) and linolenic acid (C18:3) produced a 3-4-foId elevation (P < 0.01) of plasma GIP concentrations between 60 and 120 min. Mean increments above basal areas for the GIP responses were calculated as the sum of values at 30-120 min minus 4 times the basal value, and divided by 2 to give final expres sion as pg*ml“l*h. T h e increments above basal were: propionic acid 691 ± 350 ( m e a n ± S E M ) , capric acid 216 ± 281, stearic acid 1348 ± 436, oleic acid 2389 ± 382, linoleic acid 2550 ± 691, and linolenic acid 1508 ± 665. T h e GIP response to oleic acid was significantly greater (P < 0.05-0.01) than the responses to propionic, capric and stearic acids, but not significantly greater than to linoleic and linolenic acids. Discussion Fat-stimulated GIP release is dependent on intraluminal hydrolysis of triglycerides and absorption of the digestion products. Accordingly, GIP release is compromised in conditions of impaired fat digestion such as cystic fibrosis and chronic pancreatitis (Ross Shaffer, 1981; Ebert <5c Creutzfeldt, 1980), and in the latter condition extracts of pancreatic enzymes partially restore the GIP responses to oral fat 704 KWASOWSKI ET AL. (Ebert & Creutzfeldt, 1980). In contrast, retardation of fat absorp tion with cholestyramine impairs GIP release (Ebert & Creutzfeldt, 1980). Previous studies on the relative importance of triglyceride digestion products in fat-stimulated GIP release have produced equivocal results. Thus, medium-chain triglycerides, monoglycerides, glycerol and both saturated and unsaturated long-chain fatty acids such as palmitic, oleic and linoleic acids have all been reported to lack stimulatory effects on GIP release (Ross & Shaffer, 1981; Williams et al., 1981). In the present study, oral administration of the short-chain and medium-chain fatty acids, propionic (C3:0) and capric (ClOtO) acids, failed to produce a significant G IP response. However, equimolar administration of a long-chain saturated fatty acid (stearic acid, C18:0) and especially the long-chain unsaturated fatty acids (oleic C18:l, linoleic C18:2, and linolenic C18.3 acids) produced a marked GIP response. Short-chain and medium-chain fatty acids are absorbed more rapidly than long-chain fatty acids, and the more potent GlP-releasing unsaturated long-chain fatty acids are absorbed faster than their saturated counterparts (Clement, 1980; Thomson & Dietschy, 1981). This indicates that G I P release is not related only to the rate of cellular uptake of fatty acids, an event which does not demand energy expenditure (Clement, 1980; Thomson & Dietschy, 1981). It is therefore likely that the stimulus for fat-induced GIP release is generated during the intracellular handling and metabolism of fatty acids. Short- and medium-chain fatty acids are transferred across the intestinal epithelium without estérification. However, long-chain fatty acids are conveyed to the smooth endoplasmic reticulum for estérifi cation prior to incorporation into chylomicrons and exocytosis into the intercellular c o m p a r t m e n t (Clement, 1980; T h o m s o n & Dietschy, 1981). T h e GlP-releasing action of fatty acids m a y be coupled therefore to the extent of estérification, an energy-consuming metabolic process confined to long-chain fatty acids (Clement, 1980; Thomson & Dietschy, 1981). Previous studies on the m e c h a n i s m of carbohydrate- stimulated GIP release have shown that only actively transported sugars, such as glucose and galactose, stimulate G IP secretion (Sykes et al., 1980). Thus there may be a common link between metabolic and secretory events responsible for nutrient-regulation of GIP release from the intestinal K-cells. References Bailey CJ, Flatt PR & Atkins TW (1982) Influence of genetic background and age on the expression of the obese byper- glycaemic syndrome in Aston ob/ob mice. Int. J. Obes. 6, 11-21. Brown JC (1982) Gastric Inhibitory Polypeptide. Monographs on Endocrinology, vol. 24, pp 1-88, Springer-Verlag, Berlin. Brown JC, Dryburgb JR, Ross SE & Dupre J (1975) Identification and actions of gastric inhibitory polypeptide. Recent Prog. Horm. Res. 31, 487-532. Clement J (1980) Intestinal absorption of triglycerols. Reprod. Nutr. Develop. 20, 1285-1307. FATTY ACIDS AND GASTRIC INHIBITORY PEPTIDE 705 Creutzfeldt W (1982) Castrointestinal peptides - role in pathophysiology and disease. Scand. J. Gastroenterol. 17, suppl. 77, 7-20. Ebert R & Creutzfeldt W (1980) Decreased GIP secretion through impairment of absorption. In: Frontiers in Hormone Research, vol 7, pp 192-201 (Creutzfeldt W, ed), Karger, Basel. Falko JM, Crockett SE, Cataland S & Mazzaferri EL (1976) Gastric inhibitory polypeptide (GIF). Intestinal distribution and stimulation by amino acids and medium chain triglycerides. J. Clin. Endocrinol. Metab. 41, 260-265. Flatt PR & Bailey CJ (1981) Abnormal plasma glucose and insulin responses in heterozygous lean (ob/+) mice. Diabetologia 20, 573-577. Flatt PR, Bailey CJ, Kwasowski P, Swanston-Flatt SK & Marks V (1983) Abnormalities of GIF in spontaneous syndromes of diabetes and obesity in mice. Diabetes 32, 433-435. Flatt PR, Bailey CJ, Kwasowski P, Page T & Marks V (1984) Plasma immunoreactive gastric inhibitory polypeptide in obese hyperglycaemic (ob/ob) mice. J. Endocrinol. 101, 249-256. Marks V & Morgan LM (1984) The enteroinsular axis. In: Recent Advances in Diabetes, vol 1, pp 55-71 (Nattrass M & Santiago JV, eds), Churchill Livingstone, Edinburgh. Morgan LM, Morris BA & Marks V (1978) Radioimmunoassay of gastric inhibitory polypeptide. Ann. Clin. Biochem. 15, 172-177. O'Dorisio TM, Cataland S, Stevenson M & Mazzaferri EL (1976) Gastric inhibitory polypeptide (GIF). Intestinal distribution and stimulation by amino acids and medium chain triglycerides. Am. J. Dig. Dis. 21, 761-765. Polak JM, Pearce ACE, Grimelius L & Marks V (1975) Castrointestinal apudosis in obese hyperglycaemic mice. Virchows Arch. (Cell. Pathol.) 19, 135-150. Ross SA & Shaffer EA (1981) The importance of triglyceride hydrolysis for the release of gastric inhibitory polypeptide. Gastroenterology 80, 108-111. Sykes S, Morgan LM, English J & Marks V (1980) Evidence for preferential secretion of gastric inhibitory polypeptide in the rat by actively transported carbohydrates and their analogues. J. Endocrinol. 85, 201-207. Thomson ABR & Dietschy JM (1981) Intestinal lipid absorption: major extracellular and intracellular events. In Physiology of the Castrointestinal Tract, vol 2, pp 1147-1220 (Johnson LR, ed). Raven Press, New York. Williams RJ, May, JM & Biebroeck JB (1981) Determinants of gastric inhibitory polypeptide and insulin secretion. Metabolism 30, 36-40. REPRINTED FROM: ELSEVIER SCIENCE PUBLISHERS B.V. BIOMEDICAL DIVISION Regulatory An International Journal Peptides Aims Now entering its sixth year of publication, Regulatory Peptides has realised its aim to handle all submitted manuscripts efficiently, wisely and fairly and to publish accepted articles as rapidly as possible. It is the policy of the Journal to expand to meet demands so that articles will not be delayed by an inflexible publication schedule. Scope REGULATORY PEPTIDES publishes reports of interdisciplinary studies on the physiology and pathology of peptides of the gut, endocrine and nervous systems which regulate cell or tissue function. Articles emphasizing these objectives may be based on either fundamental or clinical observations obtained through the disciplines of morphology, cytochemistry, biochemistry, physiology, pathology, pharmacology or psychology. EDITORS-IN-CHIEF Floyd E. Bloom (La Jolla, California, U.S.A.) Rolf Hakanson (Lund, Sweden) ASSISTANT EDITORS Marvin Brown (San Diego, California, U.S.A.) Frank Sundler (Lund, Sweden) EDITORIAL BOARD J.D. Baxter, San Francisco, California, U.S.A. R.A. NIcoll, San Francisco, California, U.S.A. S.R. Bloom, London, U.K. J.M. Polak, London, U.K. G.J. Dockray, Liverpool, U.K. J.F. Rehfeld, Copenhagen, Denmark J.D. Gardner, Bethesda, Maryland, U.S.A. E, Solda, Pavia, Italy V.L.W. Go, Rochester, Minnesota, U.S.A. L.O. Uttenthal, London, U.K. I.M. Modlln, Brooklyn, New York, U.S.A. N. Yanalhara, Shizuoka, Japan Manuscripts should be submitted to: Floyd E. Bloom Rolf Hakanson Division of Preclinical Neuroscience Department of Pharmacology and Endocrinology University of Lund Research Institute of Scripps Clinic Solvegatan 10 10666 North Torrey Pines Road S-223 62 Lund La Jolla, CA 92037, U.S.A. Sweden © 1985/86 Elsevier Science Publishers B.V. (Biomedical Division) Printed in The Netherlands Regulatory Peptides, 13 (1986) 163-168 163 Elsevier RPT 00444 Differential insulin responses to oral and intravenous glucose in the transplantable rat insulinoma - a role for GIP K.S. Tan, P. K w a s o w s k i a n d V. M a r k s Division o f Clinical Biochemistry, Biochemistry Department, University o f Surrey, Guildford, Surrey GU2 5XH, United Kingdom (Received 18 April 1985; revised manuscript received 19 June 1985; accepted for publication 12 October 1985) Summary W e have previously demonstrated an impaired insulin response to intraperitoneal glucose and arginine by the transplantable N E D H rat insulinoma. The nature of this tumour B-cell defect has been further studied by investigating the response of insu linoma-bearing rats to intravenous and intragastric glucose. Intravenous glucose failed to stimulate plasma immunoreactive insulin (IRI) above high basal levels (14.5 ± 1.1 /tg/L). However, significant elevation of the plasma IRI concentration was observed following an intragastric glucose load (17.1 ± 1 . 5 jUg/L; P < 0.02). In view of the different effects of oral and intravenous glucose on insulin secretion in the RIN, implicating an involvement of incretin factors from the gut, the response of the tumour to GI F was investigated. Plasma IRI concentrations rose significantly in these animals (20.6 ± 2.5 jUg/L at 5 min, P < 0.02). W e conclude that (a) the transplantable rat insulinoma is responsive to GIP, and (b) that whilst the tumour B-cell has lost its insulin responsiveness to hyperglycaemia produced by intraperitoneal or intravenous glucose, it retains its ability to respond to intragastric glucose. This could be due to incretin factors from the gut of which G I P is currently the strongest candidate. insulinoma rats; incretin; G I P Address for correspondence: Dr. K. Tan, Biochemistry Department, University of Surrey, Guildford, Surrey GU2 5XH, U.K. 0167-0115/86/503.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division) 164 Introduction The development of an X-ray induced transplantable insulinoma in the N E D H rat [1] has ma d e available large amounts of material for B-cell research [2-5]. This islet cell tumour is morphologically well differentiated and contains immunoreactive in sulin (IRI), significant amounts of immunoreactive somatostatin (1RS) but not de tectable immunoreactive glucagon (IRG) ([!]; Tan, unpublished data). W e have previously reported an impaired insulin response by the rat insulinoma (RIN) to intraperitoneal glucose and arginine stimulation [6]. Both glucose and ar ginine failed to stimulate IRI concentrations above high basal levels whereas gluca gon produced a potent release of IRI. To understand further the nature of the tumour B cell defect, we have investigated the reponses of the insulinoma to intravenous and intragastric glucose and studied the effect of gastric inhibitory polypeptide (GIP) on glucose homeostasis in these tumour animals. GI P is currently the strongest candidate for the glucose-dependent insulinotropic agent released from the gut following oral carbohydrate stimuli [7] termed ‘incretin’ by L a Barre and Still [8]. G I P has been shown to be insulinotropic at elevated glucose levels in the rat [9,10]; this effect was subsequently inhibited, at least in part, by the use of anti-GIP antibodies [11,12]. The response of the R I N to G I P m a y throw further light on the nature of the tumour B-cell defect. Materials and Methods Animals The insulinoma was maintained by serial subcutaneous transplantation within a colony of inbred albino N E D H strain of rats [1]. Transplantation resulted in hyper- phagia, increased weight gain, marked hyperinsulinaemia (36.0 ± 2.7 fig/L), severe hypoglycaemia (0.8 ± 0 . 1 mmol/L) with death of the recipient by 22-27 days [6]. All experiments were conducted 19 days post-transplantation. Experimental protocol Intragastric glucose (5 g/kg) was administered by trocar in conscious fed animals into six insulinoma-bearing rats and six age-matched control N E D H rats. Another group of six fed N E D H rats were given intravenous glucose (0.5 g/kg) via the femoral vein following anaesthesia with sodium pentobarbitone (Sagatal, Ma y & Baker, Da g enham, U.K., 6 mg / 100 g body wt.; intraperitoneal). Blood sampling was performed by tail bleeding. T w o groups of six insulinoma rats were further anaesthetised. Glu cose (0.5 g/kg) was administered intravenously via the femoral vein into one group, whilst the other group of animals received glucose (0.5 g/kg) plus GI P (1 jug/kg; Dr. John Brown, Vancouver, Canada). Blood samples were obtained from the tail at different time intervals and assayed for plasma IRI and glucose. Chemical analyses Plasma IRI was measured by radioimmunoassay using guinea-pig insulin anti serum (Guildhay Antisera, Guildford, U.K.), rat insulin standards (Novo, C o p e n 165 hagen, Denmark) and ^^^I-insulin (A mersham International pic, Amersham, U.K.). Plasma glucose was determined by a Cobas Bio Analyser using glucose oxidase (Roche Diagnostica, Herts., U.K.). Statistical analyses A two-tailed Student’s t-test for unpaired data was used. Results are expressed as mean ± S.E.M. (n = 6). Results Following the intragastric administration of glucose, the plasma glucose concen tration was significantly elevated at 20 min in the insulinoma rats (7.8 ± 0.7 mm o l / L compared with basal = 1.7 ± 0.2 mmol/L; P < 0.001) and in the control N E D H rats (10.0 ± 0.5 m m o l / L compared with basal = 5.5 ± 0.3 mmol/L; P < 0.001) (Fig. 1). It was accompanied by an increase in plasma IRl concentration in the in sulinoma rats (18.3 ± 1.5 fig/L compared with basal = 13.1 ± 0.8 fig/L; P < 0.01) and the control N E D H rats (7.8 ± 0.5 /rg/L compared with basal = 2.8 ± 0.4 /rg/L; P < 0.001) (Fig. 2). In contrast to this, intravenous glucose failed to raise plasma IRl concentrations above high basal levels in the insulinoma rats (Fig. 3). However, glucose in combination with GIF, significantly stimulated plasma IRl in the insuli n o m a rats by 2 min (20.6 ± 2.5 jUg/L; P < 0.05) and 15 min (18.6 ±1.1 /rg/L; P < 0.02) compared with the control insulinoma rats given glucose alone (Fig. 3). N o significant differences in plasma glucose levels were observed between the 2 groups at all time points, although both groups were significantly higher than basal values (P < 0.02) at 2 min (Fig. 4). In control N E D H rats, intravenous glucose significantly stimulated plasma IRl (10.3 ± 1.1 /zg/L; P < 0.001) and plasma glucose (10.0 ± 0.8 mmol/1; P < 0.001) by 5 min (Fig. 5). 10 20 8 16 6 4 8 2 4 0 20 40 60 0 20 40 60 TIME ( MIN ) TIME (MIN ) Fig. 1. Effect of intragastric glucose (5 g/kg body wt.) on plasma glucose concentrations in insulinoma (#---- @) and control NEDH rats (O O). Results are mean ± S.E.M. (n = 6). * = P < 0.001 compared with values at zero time of each group. Fig. 2. Effect of intragastric glucose (5 g/kg body wt.) on plasma immunoreactive insulin (IRI) in insuli noma ( # ------#) and control NEDH rats (O O). Results are mean ± S.E.M. (« = 6). * = P < 0.01 and ** = />< 0.001 compared with values at zero time of each group. 166 20 8 4 6 3 4 2 12 1 -10 0 15 30 40 -10 0 15 30 40 T1ME( MIN) TIME ( MIN ) Fig. 3. Effect of intravenous glucose (0.5 g/kg body wt.; O O) and GIP (1 ^g/kg body wt.) plus glucose (0.5 g/kg body wt.; # # ) on plasma immunoreactive insulin (IRI) in insulinoma rats. Results are mean ± S.E.M. {n = 6). ** = P < 0.05 and * = P < 0.02 compared between the two groups. Fig. 4. Effect of intravenous glucose (0.5 g/kg body wt.; O O) and GIP (1 ng/kg body wt.) plus glucose (0.5 g/kg body wt.; # ------# ) on plasma glucose concentrations in insulinoma rats. Results are mean ± S.E.M. (n = 6). 10 10 8 6 6 2 I 15 30 45 0 15 30 450 TIME ( MIN ) TIME( MIN) Fig. 5. Effects of intravenous glucose (0.5 g/kg body wt.) on (A) plasma immunoreactive insulin (IRI) and (B) plasma glucose concentrations in control NEDH rats. Results are mean ± S.E.M. (« = 6). * = P < 0.001 compared with basal values at zero time. Discussion Changes in glucose homeostasis induced by the transplantable rat insulinoma are similar in some respects to those produced by insulinomas in ma n [13-15] suggesting their suitability as model for the study of spontaneous insulinomas. T he impaired insulin response to i.p. glucose and arginine suggests, however, that RI N cells possess an intrinsic defect which makes them functionally quite different from normal rat and h u m a n B-cells [6]. This defect has been further investigated in the present study. A n impaired insulin response to glucose was observed in the RIN. The inability to increase plasma insulin 167 levels after i.v. glucose could be due to too low a plasma glucose increment since a glucose threshold might exist in these rats which have been hypoglycaemic for some time. However, at higher glucose doses (2 g/kg body wt.), w e were still unable to demonstrate insulin release from these rats [6,13]. It would appear therefore that the inability of the R I N to respond to i.v. glucose is a genuine defect. This insensitivity to glucose has also been reported in hu m a n insulinomas [15,16] and in the hamster insulinoma [17,18]. N o defects in glucose uptake or glucose phosphorylation were found in the hamster insulinoma cell but there was a failure by glucose to stimulate calcium influx. Whether this is also true of the RI N remains to be established. Whilst responsiveness of the RI N to intravenous glucose is lost, its ability to respond to oral glucose is retained as in ma n y hu m a n insulinomas. This differential response provides further evidence for the existence of an ‘enteroinsular axis’ in which incretin factors, released from the gut in response to nutrient ingestion, stimulate insulin secretion since neural involvement is extremely unlikely in this situation. G I P is currently the strongest candidate for the role of incretin and its role in mediating the insulin response of oral glucose has been suggested by several authors [7,12]. Whilst intravenous glucose failed to stimulate insulin secretion, the R I N was found to be responsive to GI P wh e n administered together with glucose. This diver gence is interesting because G I P is insulinotropic only in the presence of hypergly caemia [19]. This glucose dependence suggests a degree of cooperativity between glucose and GI P at the receptor or post-receptor level. However it would appear that the cooperativity is not absolute and that a degree of dissociation exists between them. The interaction between GI P and its receptor was unaffected by the insensivity of the tumour to glucose stimulation, suggesting that the tumour B-cell defect is specific for glucose rather than the G I P mechanism of action. W e have previously demonstrated an insulin response to glucagon in the R I N [6] indicating that the glucagon receptors on the insulinoma cells are functionally intact. Since the addition of GI P to i.v. glucose also results in insulin release, it is suggested that the receptor for this peptide is also preserved, and that the tumour B-cell defect is not related to these peptide receptors. These findings are consistent with data obtained from in vitro studies using dispersed cells from the RIN. Although glucose failed to stimulate insulin secretion from these R I N cells, responsiveness to G I P was retained however (Tan, unpublished observation). Further work, however, is needed to understand the exact nature of this defect. The ability of GI P to stimulate insulin secretion from the insulinoma under the condition of a lower plasma glucose level than in normal rats was observed. This is interesting but the reasons for this are not clear and further studies on the glucose sensitivity for the GI P effect in the insulinoma and control rats are needed to under stand the underlying mechanisms. Glucose clearance in the insulinoma rats after intravenous glucose with or without GI P was almost identical despite a huge differ ence in their insulin responses. This is probably due to the marked insulin insensitivity in the insulinoma rats as previously reported [6]. Insulin administration failed to lower the plasma glucose concentrations in these hypoglycaemic insulinoma rats. W e conclude that care ought to be exercised wh e n using the R I N as a source of B-cells for metabolic studies in view of findings reported here. However, it is un- 168 doubtedly a useful model for studies relating to G I P and other incretin factors and also studies in the mechanism of insulin secretion. Acknowledgement This work has been supported by the M R C . References 1 Chick, W.L., Warren, S., Chute, R.N., Like, A.A., Lauris, V. and Kitchen, K.C., A transplantable insulinoma in the rat, Proc. Natl. Acad. Sci. USA, 74 (1977) 628-632. 2 Duguid, J R., Steiner, D.F. and Chick, W.L., Partial purification and characterisation of the mRNA for rat preproinsulin, Proc. Natl. Acad. Sci. USA, 73 (1976) 3539-3543. 3 Gazdar, A.F., Chick, W.L., Oie, H.K., Sims, H.L., King, D.L., Weir, G.C. and Lauris, V., Continuous, clonal, insulin-somatostatin - secreting cell lines established from a transplantable rat islet cell tumour, Proc. Natl. Acad. Sci. USA, 77 (1980) 3519-3523. 4 Sopwith, A.M., Hutton, J.C., Naber, S.P., Chick, W.L. and Hales, C.N., Insulin secretion by a trans plantable rat islet cell tumour, Diabetologia, 21 (1981) 224-229. 5 Eisenbarth, G.S., Oie, H., Gazdar, A., Chick, W., Schultz, J.A. and Scearce, R.M., Production of monoclonal antibodies reacting with rat islet cell membrane antigens. Diabetes, 30 (1981) 226-230. 6 Tan, K., Flatt, P R., Webster, J.D. and Marks, V., Glucose homeostasis during proliferation of a transplantable rat insulinoma, Diabetologia, 21 (1981) 514. 7 Creutzfeldt, W., The incretin concept today, Diabetologia, 16 (1979) 75-85. 8 La Barre, J. and Still, E.U., Studies in the physiology of secretion. Ill further studies on the effects of secretin on the blood sugar. Am. J. Physiol., 91 (1930) 649-653. 9 Rabinovitch, A. and Dupre, J., Effects of gastric inhibitory polypeptide present in impure pancreo- zymin-cholecystokinin on plasma insulin and glucagon in the rat. Endocrinology, 94 (1974) 1139- 1144. 10 Pederson, R.A. and Brown, J.C., The insulinotropic action of gastric inhibitory polypeptide in the perfused isolated rat pancreas. Endocrinology, 99 (1976) 780-785. 11 Lauritsen, K.B., Holst, J.J. and Moody, A.J., Depression of insulin release by anti-GIP serum after oral glucose in rats, Scand. J. Gastroent., 16 (1981) 417-420. 12 Ebert, R. and Creutzfeldt, W., Influence of gastric inhibitory polypeptide antiserum on glucose-induced insulin secretion in rats. Endocrinology, 111 (1982) 1601-1606. 13 Flatt, P R., Tan, K., Bailey, C.J., Swanston-Flatt, S.K., Marks, V. and Webster, J.D., Plasma glucose and insulin concentrations following implantation and surgical resection of a transplantable rat in sulinoma, Biochem. Soc. Trans., 10 (1982) 273-274. 14 Frerichs, H. and Creutzfeldt, W., Hypoglycaemia, 1. Insulin secreting tumours, J. Clin. Endocrinol. Metab., 5 (1976) 747-767. 15 Marks, V. and Rose, F.C., Hypoglycaemia, Blackwell Scientific Publications, Oxford, London, 2nd edn., 1981. 16 Yip, C.C. and Schimmer, B.P., Human pancreatic islet tumour cells maintained in vitro, Diabetologia, 9 (1973) 251-254. 17 Shapiro, E., Eto, S., Fleischer, N. and Baum, S.G., Regulation of in vitro insulin release from a transplantable Syrian hamster insulinoma. Endocrinology, 97 (1975) 442-447. 18 Shapiro, S., Kaneko, Y., Baum, S.G. and Fleischer, N., The role of calcium in insulin release from hamster insulinoma cells. Endocrinology, 10 (1977) 485-493. 19 Elahi, D., Andersen, D.K., Brown, J.C., Debas, H., Hershcopf, R.J., Raizes, G.S., Tobin, J.D. and Andres, R., Pancreatic a- and )?-cell responses to GIP infusion in normal man. Am. J. Physiol., 237 (1979) 185-191. INSTRUCTIONS TO AUTHORS Submission of a paper to Regulatory Peptides is understood to imply that it has not previously been published (except in an abistract form), and that it is not being considered for publication elsewhere. Manuscripts should be submitted in quadrupiicate, and be complete in all respects (including 4 copies of all illustrations). Only manuscripts which deal with original research not previously published and not under consideration for publication in other journals will be considered. All papers submitted for review will be subjected to an editorial decision to accept, revise or reject within one month. It is expected that all issues will be published within two months of receipt of the last paper of that issue. All manuscripts will be considered on an equivalent basis of scientific rigour and peer evaluation. In no case should articles exceed 12 printed pages. (Mini-)reviews will be by invitation only. Manuscripts should be submitted to either (Chief) Editor at the above addresses. Manuscripts, written in English, must be submitted in double-spaced typing on pages of uniform size. For paper 21.5 X 28 cm, with 4-cm margins, the following estimations provide a guide to final article length: 2 typewritten pages or 3 ‘average’ figures with legends o r 25 references = 1 printed page. S.I. units should be used throughout. Papers which will be longer than 5 printed pages should be divided into the following sections, headed by captions (Introduction; Materials and Methods; Experimental Results; Discussion; Acknowledgements). All manuscripts should be accompanied by a Summary of 200 words or less which describes the experimental objectives, results and conclusions. Authors should provide a list of not more than 6 key words, not included in the title of the manuscript, for subsequent indexing. Authors’ full names, academic or professional affiliations, and complete addresses should be included, on a separate title page. The name and address of the author to whom correspondence may be sent should be indicated, including telephone number for immediate enquiry. Literature references. Citation of literature references in the text should be given at the appropriate places as numbers in square brackets. All references cited in the text should be listed at the end of the paper on a separate page (also double-spaced), arranged in numerical order of citation in the text, not in alphabetical order. Literature references must be complete, including initials of the authors cited, title of paper referred to, abbreviated title, volume, year, and first and last page numbers of the article in a periodical. The abbreviations of journal titles should conform to those adopted by Chemical Abstracts Service, 1974. (See example 1 below.) The form of literature references to books should be: author(s), initials, title of book, pubUsher and city, year and page number referred to. (See example 2 below.) References to authors contributing to multi-author books or to proceedings printed in book form should be similar to those for books. (See example 3 below.) Examples: 1 Pelletier, G. and Leclerc, R., Localization of Leu-enkephalin in dense core vesicles of axon terminals, Neurosci. Lett., 12 (1979) 159-163. 2 van Ree, J.M. and Terenius, L. (Eds.), Characteristics and Functions of Opioids, Developments in Neurosci ence, Vol. 4, Elsevier/North-Holland, Amsterdam, 1978, 520 pp. 3 Renaud, L.P., Electrophysiology of brain peptides. In A.M. Gotto, Jr., E.J. Peck, Jr. and A.E. Boyd, III (Eds.), Brain Peptides: A new Endocrinology, Proceedings of the 4th Argenteuil Symposium, Waterloo, Belgium, October 29-31, 1978, Elsevier/North-Holland Biomedical Press, Amsterdam, 1979, pp. 119-138. Numbered references to personal communications, unpublished data and manuscripts either “ in preparation” or “ submitted for publication” should not be included. If essential, such material may be incorporated at the appropriate place in the text. Illustrations. All illustrations must be submitted in quadruplicate, and reach the editors in a form and condition suitable for reproduction. The illustrations should bear the author’s name and be numbered in Arabic numerals according to the sequence of their appearance in the text, where they are to be referred to as Fig. 1, Fig. 2, etc. Line drawings should be in black ink on drawing or tracing paper. Lettering should be clear and of adequate size to be legible after reduction. Professional labelling is preferable but, if this is not possible, lettering should be done in fine pencil. Photographs, including radiographs and electron micrographs should be supplied as clear black and white prints on glossy paper. Figures will usually be reduced to fit into the normal type area of 12.5 X 19 cm. However, larger reproduction up to a maximum of 17.5 X 26 cm is possible on request. Reproduction in colour will have to be approved by the Editors and Publishers. The extra cost of colour reproduction will be charged to the author(s). Each illustration must have a legend which should be typed with double-spacing on a separate page and begin with the number of the illustration it refers to. T ables. Tables of numerical data should be typed (double-spaced), each on a separate page, numbered in sequence in Roman numerals (Table I, II, etc.), provided with a heading, and referred to in the text as Table I, Table II, etc. Proofreading. In the interest of speed, proofreading will be usually undertaken by the Publisher, and proofs will be sent to the author only if publication schedules permit. At this stage no changes to the original manuscript will be accepted. R eprints. A total of 50 reprints of each paper will be provided free of charge to the author(s). Additional reprints can be ordered at prices shown on the reprint order form which will be sent to the author upon acceptance of the paper. Page charges will not be made. Book reviews will be published as space permits. Books for review in this journal should be sent to one of the Editors-in-Chief. RPeptides ^ t t l a t o r y Editors-in-Chief: Fioyd E. Bioom, Division of Preclinical Neuroscience and Endocrinology, Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, C A 92037, U.S.A. Roif Hâkanson,Department of Pharmacology, University of Lund, Solvegatan 10, S-223 62 Lund, Sweden Assistant Editors: Marvin Brown, The S A L K Institute, P.O. Bo x 85800, San Diego, C A 92138, U.S.A. Frank Sundier, Department of Histology, University of Lund, Biskopsgatan 5, S-223 62 Lund, Sweden ORDER FORM REGULATORY PEPTIDES Send your orders to: 1986 - Vols. 14-16 (3 volumes in 12 issues) ELSEVIER SCIENCE PUBLISHERS B.V. (Biomedical Division) P.O. Box 211, Please enter an institutional subscription for 1000 AE Amsterdam □ 1986 at U S $ 282.25/Dfl. 762.00 (including postage and handling) The Netherlands For further information and/or free □ Please send m e a free sample copy specimen copy, write to: In the USA and Canada: JOURNAL INFORMATION CENTER □ Please send m e a proforma invoice Elsevier Science Publishing Co. Inc., 52 Vanderbilt Ave., New York, NY 10017, USA N a m e (please print) Address ______In the rest of the world write to the promotion manager in Amsterdam Journals are sent by surface delivery to all I enclose m y personal cheque bank draft U N E S C O coupons countries, except the following countries where SAL air delivery (Surface Airlifted Mail) is ensured: U.S.A. South Africa Please bill my credit card Canada Malaysia American Express _____ Japan Singapore * Mastercharge/* Eurocard / * Access Australia South Korea * Issuing bank New Zealand Taiwan P R. China Pakistan Yisa/Bankamericacard/Barclaycard India Hong Kong Expiry Date ______Israel Brazil Signature ______Date Airmail rates for other countries are available upon request. The Dutch guilder price is definitive. US $ prices are subject to exchange rate fluctuations. Prices are excl. B. T. W. for Dutch customers. Diabetologia (1983) 24: 278-281 Diabetologia © Springer-Verlag 1983 Effect of Pretreatment with a High Fat Diet on the Gastric Inhibitory Polypeptide and Insulin Responses to Oral Triolein and Glucose in Rats S. M. Hampton, P. Kwasowski, K.Tan, L. M. Morgan and V. Marks Division of Clinical Biochemistry, Department of Biochemistry, University of Surrey, Guildford, Surrey, UK Summary. Male Wistar rats were pretreated with 3 ml triolein controls (plasma glucose nadir 2.6 ±0.48 versus 1.6 ±0.15 orally for 4 days in addition to their normal diet. A similar mmol/1, p<0.05). Fat-pretreated rats showed significantly control group were allowed free access to normal laboratory higher insulin and glucose levels compared with the untreated food. Wh e n given an oral fat load (1 ml triolein) plasma gas rats when given oral glucose (plasma insulin: 6.2 ± 1.2 versus tric inhibitory polypeptide (GIP) and triglyceride levels were 2.5 ± 0.59 pg/1, /7< 0.01; plasma glucose: 10.2±0.39 versus significantly higher in the fat pretreated group. Inhibition of 8.9±0.41 mmol/1, /><0.025). Pretreatment of rats on a high fat-stimulated GI P release by exogenous insulin was de m o n fat diet causes (1) increased GI P secretion in response to an strated in the untreated control group (plasma GIP: 663 ±4 9 oral fat load, (2) abolition of the feed-back inhibition of versus 853 ±9 2 ng/I, mean ± SE M p < 0.025), but pretreat exogenous insulin on fat-stimulated GI P release, and (3) some ment with an oral fat load abolished this effect (plasma GIP: degree of insulin resistance. 1008 ±9 5 versus 1116 ± 100 ng/1, p NS). Plasma glucose levels were significantly higher in fat pretreated rats given oral Key words: GIP, insulin, triolein, dietary adaptation. fat and intraperitoneal insulin compared with untreated The concept that the gut can modify insulin secretion tor in the altered GIP response. Willms et al. [11], for ex has been well established since experiments in 1964 by ample, found that 5 days of calorie restriction abolished McIntyre et al. [1] demonstrated that intrajejunally ad the exaggerated GIP response to a fat or mixed meal in ministered glucose led, on a molar basis, to more than obese subjects, suggesting that a previously high calorie twice as much insulin release as intravenous glucose. intake might have been responsible for the excessive The intestinal hormone gastric inhibitory polypep GIP production. tide (GIP) is a potent stimulator of glucose-induced in We have therefore investigated the effect of pretreat sulin secretion [2] and is currently thought to be one of ment with fat supplements on fat and glucose-induced the major endocrine components of the entero-insular GIP release in rats. The inhibition of fat-stimulated GIP axis. GIP is released after ingestion of glucose [2,3] ami release by exogenous insulin was also investigated. no acids [4] or fat [2,5], the latter being especially potent. Exogenous insulin [2, 6] and C-peptide [7] have been shown to inhibit fat-induced GIP release and may play Materials and Methods a physiological role in the regulation of GIP secretion. Obesity in man is characterised biochemically by Male Wistar rats weighing between 240 and 290 g were used. The basal hyperinsulinaemia and an exaggerated insulin- 160 rats were divided into two equal groups. The first (fat-pretreated aemic response to glucose and most other insulino group) was given 3 ml triolein (24 kcal) a day orally for 4 days before the experiment as well as being allowed free access to rodent breeding tropic stimuli [8]. The cause of the hyperinsulinaemia is diet no. 1, expanded (Spratt’s Laboratory Services, Barking, Essex). still controversial but it could be due, at least in part, to This resulted in a diet with a final fat content of 21%. The second an overactive enteroinsular-axis. Exaggerated GIP re group (untreated, low fat diet group) was fed rodent breeding diet sponses have been observed after a mixed meal in obese alone, the resultant diet containing 3.5% fat. Animals were weighed each day and fasted for 24 h before each experiment. Three experi subjects [9] and exogenous insulin fails to inhibit fat- mental protocols were performed on fat-pretreated and untreated stimulated GIP in this group [10]. However, the compo rats, during which they were anaesthetised with IP pentobarbitone sition of the diet preceding the test appears to be a fac (100 mg/kg body weight). Blood was collected by cardiac puncture s. M. Hampton et al.: GIP and Insulin Responses to High Fat Diet 279 1200 Plasma glucose was measured using a glucose oxidase method (Beckman glucose analyser) and plasma triglycerides using a fully en zymatic UV kit (Boehringer, Lewis, Sussex). 800 Statistical Analyses O) Results were compared using Student’s t-test for unpaired data and a analysis of variance. 400 Results o 60 120 180 The unpretreated control rats were slightly, but signifi Time (min) cantly heavier than the fat-pretreated rats throughout Fig.l. Plasma GIP levels following oral triolein (1 ml) in fat-pretreat the study. There was, however, no significant weight ed (O — O) and untreated control rats (#— —#) (mean ± SEM, gain in either group of rats on completion of the fat-pre- n = 6 ) treatment (control rats 297.0 ±4.14 before pretreatment versus 297.5±4.13 g after pretreatment, p NS; fat-pre treated rats 280.2 ±4.96 before pretreatment versus 279.0±10g after pretreatment, p NS; mean±SD, into lithium heparin tubes and the samples immediately centrifuged. n = 60). There were no significant differences in fasting Plasma for hormone and glucose estimation was divided and stored at plasma glucose, insulin or GIP levels between the two -2 0 °C . groups on completion of the fat-pretreatment. Experimental Protocols Time Course of the Response to Oral Fat Time course of responses to ora! fat: Fat-pretreated and untreated groups of rats were each divided into five sets (six rats/set). Set A Mean GIP levels were consistently higher in animals in from each group received no triolein. Animals in the other four sets the fat-pretreated sets, but the differences did not reach (B-E) of each group received 1 ml triolein (8 kcal) orally. Blood was statistical significance at any single timepoint (Fig. 1). collected by cardiac puncture at Omin (set A) and at 30, 60,120 and However, when the data was subjected to analysis of 180 min after oral dosing (sets B-E respectively). variance (two-factor with replication) the two groups Effect o f fat-pretreatment on responses to oral fa t with or without IP in were found to be significantly different at the 5% level. sulin: Fat-pretreated and untreated rats were each divided into three Plasma triglyceride concentrations were similar in sets (10 rats/set) and given; set A: 1 ml triolein orally 2 h before IP in the two groups at the beginning of the experiment but sulin (1 U/kg body weight); set B: 1 ml triolein orally 2 h before IP sa were significantly higher in the fat-pretreated rats than line (0.154mol/1); set C: 1 ml saline (0.154mol/1) orally 2 h before IP saline. Blood was collected by cardiac puncture 20 min after the IP in in the controls 120 and 180 min after receiving oral fat jections in all cases. (basal plasma triglyceride concentrations were 0.56 ± 0.04 and 0.42 ± 0.04 mmol/1 in the control and fat-pre- Effect of fat-pretreatment on the response to oral glucose: Groups of fat- treated groups respectively). At 120 min plasma triglyc pretreated and untreated rats were each divided into two sets (10 rats/ eride levels were 2.08 ±0.15 in the control group versus set). Set A was given 3 ml 2.78 mol/1 glucose (5.8 kcal) orally and set B 3 ml saline (0.154 mol/1) orally. Blood was collected by cardiac punc 4.91 ± 0.84 mmol/1 in the fat-pretreated group,p < 0.01 ; ture 60 min after oral dosing. at 180 min they were 2.39 ±0.41 and 7.1 ±0.46 mmol/1 respectively < 0.01 : mean ± SEM, n=€). The plasma in the fat-pretreated group had a marked lipaemic ap Chemical Analyses pearance which was not apparent in the controls. Immunoreactive insulin was measured by a double-antibody radioim munoassay using an antiserum raised against bovine insulin (Guild hay Antisera, Guildford, UK, MF/GP/10-VA), iodinated bovine in Effect of Fat Pretreatment on the Response to sulin (Radiochemical Centre, Amersham, Bucks., UK) and a rat in Oral Fat with or without Exogenous Insulin sulin standard (Novo, Copenhagen, Denmark, lot no. RC 791009). The sensitivity of the assay was 0.5 pg/l and the interassay coefficient Intraperitoneal insulin was associated with a signifi of variation 15.6% at 0.5 pg/1 and 6.7% at 6.0 pg/1. [Insulin values are cantly smaller rise (Fig. 2) in triolein-stimulated GIP se expressed as pg /1 as the rat standard used was calibrated on a weight cretion in unpretreated rats than in those given IP saline basis (1 |ig = approximately 25 mU)]. Immunoreactive GIP was measured by a double-antibody radio (663±49 versus 853±93 ng/1, /><0.025). This differ immunoassay [12]. The antiserum recognizes the larger molecular ence was not observed in the fat-pretreated rats which form of GIP (mol. wt. approximately 8,000 daltons) as well as stand showed similar triolein-stimulated GIP levels whether ard GIP (mol. wt. 4,977). A porcine GIP standard was used, but paral they were given IP insulin or saline (1008 ±95 versus lelism was demonstrated between the standard curve and a serially di luted sample of rat plasma containing high levels of endogenous GIP. 1116±100 ng/1). Plasma glucose levels in rats given IP The assay sensitivity was 110 ng/1 and the interassay coefficient of insulin after oral fat fell significantly less in the pretreat variation 4.1% at 2,677 ng/1 and 22.5% at 138 ng/1. ed than in the unpretreated controls (2.6 ±0.48 versus 280 S. M. Hampton et al.: GIP and Insulin Responses to High Fat Diet GLUCOSE INSULIN _ G IP (mtnol/l) 800- 600- Fig. 2. Plasma glucose, insulin and GIP levels following oral triolein (1 ml) in fat-pretreated Q and untreat Saline Triolein Triolein Saline Triolein Triolein Saline Triolein Triolein ed control ^ rats (mean ± SEM, Insulin Insulin insulin n= 10; V < 0.05, 0.025) GLUCOSE INSULIN r. GIP (mmol/l) (pg/i) (ng/l) I 10 1000 800 - 600 400 Fig. 3. Plasma glucose, insulin and G IP levels following oral glucose (5 mg/kg body weight) in fat-pretreated Q and untreated control @ Saline Glucose Saline Glucose Saline Glucose rats (mean±SEM, n=10; *p<02S, **/?<0.01) 1.57 ±0.15 mmol/l, p<0S)5) although plasma insulin GIP and triglycerides after dosing with oral triolein. levels 20 min after insulin injection were similar in the The increased triglyceride levels seen in the two groups (4.1 ±1.7 versus 4.7 ±1.2 pg/1,p NS). fat-pretreated group probably reflect their increased ca pacity to absorb a large oral fat load. Deschodt-Lanck- man et al. [13] showed that pancreatic lipase activity in Effect of Fat Pretreatment on the Response to the rat is stimulated by a high fat diet. An increase in li Oral Glucose pase activity in the fat-pretreated group of rats would Following oral glucose, fat-pretreated rats had signifi result in an increase in the rate of absorption of the cantly higher levels of plasma glucose (10.2 ±0.39 ver large oral triolein load. The increased GIP response to sus 8.9 ± 0.41 mmol/l, p < 0.025) and insulin (6.2 ±1,2 oral triolein seen in the fat-pretreated group could be versus 2.5 ± 0.59 pg/1, p<0.01) levels 1 h after dosing due to their increased rate of fat absorption. GIP secre than their untreated controls (Fig. 3). Plasma GIP levels tion apparently depends upon absorption of nutrients showed a tendency, significant at the 10% level, to be and not their mere presence in the intestine [14, 22]. higher in the fat-pretreated group compared with their There is some evidence that the rate of absorption is al untreated controls. so a factor in determining the magnitude of the GIP re sponse. A reduction of nutrient absorption rate in cer tain diseases, untreated coeliac disease or marked mal Discussion absorption from any other cause, for example, causes a reduced GIP response to enteric stimuli [15,16] and the addition of the unabsorbable carbohydrate guar to a The administration of a high fat diet to rats for 4 days meal both slows down its rate of absorption and de resulted in significantly increased levels of circulating presses the GIP response [17]. s. M. Hampton et al.: GIP and Insulin Responses to High Fat Diet 281 Creutzfeldt et al. [10, 11] reported an exaggerated 4. Thomas FB, Mazzaferri EL, Crockett SE, Mekhjian HS, Gruemer GIP response in obese subjects given a high calorie HD, Cataland S (1976) Stimulation of secretion of gastric inhibito mixed meal. This was not, however, observed in their ry polypeptide and insulin by intraduodenal animo acid perfu sion. Gastroenterology 70: 523-527 subjects after 5 days of dietary restriction, although 5. Falko JM, Crockett SE, Cataland S, Mazzaferri EL (1975) Gastric weight loss was minimal. Our data support the concept inhibitory polypeptide (GIP) stimulated by fat ingestion in man. J that composition of the previous diet is more relevant Clin Endocrinol Metab 41: 260-265 than body weight in determining the magnitude of the 6. Crockett SE, Cataland S, Falko JM, Mazzaferri EL (1976) The in GIP response to oral nutrients and is supported by pre sulinotropic effect of endogenous gastric inhibitory polypeptide in normal subjects. J Clin Endocrinol Metab 42:1098-1103 liminary results obtained in healthy human subjects 7. Dryburgh JR, Hampton SM, Marks V (1980) Endocrine pan eating diets of widely different composition. creatic control of the release of gastric inhibitory polypeptide. Various workers [2,6,18,19] have demonstrated that Diabetologia 19:397-401 intravenous infusions of either insulin or glucose reduce 8 . Karam JH, Grodsky GM, Forsham PH (1963) Excessive insulin response to glucose in obese subjects as measured by immuno the rise in plasma GIP provoked by fat ingestion and chemical assay. Diabetes 12:197-205 have suggested the existance of a negative feedback 9. Ebert R, Willms B, Brown JC, Creutzfeldt W (1976) Serum gastric control of insulin upon GIP release. In the present inhibitory polypeptide (GIP) levels in obese subjects and after study the control group of rats on a low fat diet showed weight reduction. Eur J Clin Invest 6:327 inhibition of fat-induced GIP release by insulin but the 10. Creutzfeldt W, Ebert R, Willms B, Frerichs H, Brown JC (1978) Gastric inhibitory polypeptide (GIP) and insulin in obesity: in fat-pretreated ones did not. A similar defect in feedback creased response to stimulation and defective feedback control of of insulin on GIP secretion has been demonstrated in serum levels. Diabetologia 14:15-24 obese human subjects by Creutzfeldt et al. [10] but was 11. Willms B, Ebert R, Creutzfeldt W (1978) Gastric inhibitory poly not investigated by Willms et al. [11] in their study of the peptide (GIP) and insulin in obesity: II. Reversal of increased re sponses to stimulation by starvation or food restriction. Diabeto effects of dietaiy restriction on GIP secretion in obese logia 14:379-387 subjects. 12. Morgan LM, Morris BA, Marks V (1978) Radioimmunoassay of The abolition of the negative feedback control of gastric inhibitory polypeptide. Ann Clin Biochem 15:172-177 insulin on fat-induced GIP release in fat-pretreated rats 13. Deschodt-Lanckman M, Robberecht P, Camus J, Christophe J may be explained by a decreased responsiveness of the (1971) Short-term adaptation of pancreatic hydrolases to nutri tional and physiological stimuli in adult rats. Biochimie 53: GIP-secreting cells to insulin due to a reduction in in 789-796 sulin receptor numbers, though other explanations are 14. Sykes S, Morgan LM, English J, Marks V (1980) Evidence for possible. Investigation of insulin receptors on human preferential stimulation of gastric inhibitory polypeptide secretion monocytes has shown that insulin receptor number in the rat by actively transported carbohydrates and their ana logues. J Endocrinol 85:201-207 can be altered by dietary changes [20] and that these 15. Creutzfeldt W, Ebert R, Arnold R, Frerichs H, Brown JC (1976) changes can occur rapidly - within 24 h [21]. Even a Gastric inhibitory polypeptide (GIP), gastrin and insulin: re short period of pretreatment made the rats in the pres sponse to test meal in coeliac disease and after duodeno-pan- ent study resistant to the hypoglycaemic action of exo createctomy. Diabetologia 12:279-286 genous insulin. Circulating insulin and glucose levels 16. Besterman HS, Cook GS, Sarson DL, Christofides N, Bryant MG, Gregor M, Bloom SR (1979) Gut hormones in tropical malabsorp were also significantly higher after oral glucose in the tion. Br Med J II: 1252-1255 fat-pretreated animals providing further evidence that 17. Morgan LM, Goulder TJ, Tsiolakis D, Marks V, Alberti KGMM the high fat diet predisposed to insulin resistance. (1979) The effect of unabsorbable carbohydrate on gut hormones: The normal laboratory rat diet contains very little modification of post prandial GIP secretion by guar. Diabeto logia 17:85-89 fat, unlike the typical diet consumed by people in the 18. Cleator IGM, Gourlay RH (1975) Release of immunoreactive gas Western World and caution must, therefore, be exer tric inhibitory polypeptide (IR-GIP) by oral ingestion of food sub cised when assessing the relevance of these findings to stances. Am J Surg 130:128-135 man. Nevertheless, the data do provide evidence that 19. Verdonk CA, Rizza RA, Nelson RL, Go VLW, Gerich JE, Service diet can be of considerable importance in the develop FJ (1980) Interaction of fat-stimulated gastric inhibitory polypep tide on pancreatic alpha and beta cell function. J Clin Invest 65: ment of insulin resistance and changes in activity of the 1119-1125 entero-insular axis. 20. Pedersen O, Hjollund E, Lindskov HO, Schwartz Sorensen N (1980) Increased insulin receptors on monocytes from insulin de Acknowledgements. P. Kwasowski and K.Tan acknowledge financial pendent diabetics after high starch high fibre diets. Diabetologia support from the Medical Research Council and L. M. Morgan from 19:306 (Abstract) Guildhay Antisera, University o f Surrey. 21. Schluter KJ, Petersen KG, Kerp L(1980) Rapid changes in insulin binding affinity of monocytes in male and female volunteers after breakfast. Diabetologia 19: 313 (Abstract) References 22. Ross SA, Shaffer EA (1981) The importance of triglyceride hydro lysis for the release of gastric inhibitory polypeptide. Gastro 1. McIntyre N, Holdsworth DJ, Turner DS (1964) New interpreta enterology 80:108-111 tion of oral glucose tolerance. Lancet 2:20-21 2. Brown JC, Dryburgh JR, Ross SA, Dupre J (1975) Identification Received: lOJune 1982 and actions of gastric inhibitory polypeptide. Recent Prog Horm and in revised form: 13September 1982 Res 31:487-532 Mrs. S. M. Hampton 3. Cataland S, Crockett SE, Brown JC, Mazzaferri EL (1974) Gastric Department of Biochemistry inhibitory polypeptide (GIP) stimulation by oral glucose in man. J University of Surrey Clin Endocrinol Metab 39: 223-228 Guildford GU2 5XH, UK Abnormalities of GIP in Spontaneous Syndromes of Obesity and Diabetes in Mice PETER R FLATT CLIFFORD J BAILEY. PETER KWASOWSKI. SARA K SWANSTON FIATT. AND VINCENT MARKS pend on the mutant gene and its interaction with the back SUMMARY The role of GIP in the pathogenesis of spontaneous ground genome.® ® This study examines immunoreactive GIP syndromes of obesity-diabetes was examined in ob/ob in the plasma and intestine of ob/ob and db/db mice of mice of the Aston stock and db/db mice of the C57BL/ different genetic backgrounds. KsJ background. Compared with lean controls, fed adult ob/ob and db/db mice, respectively, exhibited 1.8- MATERIALS AND METHODS fold and 2.1-fold increases in body weight, 1.8-fold and Groups of Aston ob/ob mice derived from C57BU6J stock 2.8-fold elevations of plasma glucose, and 15.4-fold outcrossed to four noninbred local strains^ and C57BI_/KsJ and 5.6-fold elevations of plasma insulin. As indicated db/db mice from a background in repulsion to the misty {m) by the relative magnitude of the hyperglycemia and hy- perinsulinemia, db/db mice displayed a particularly se coat gene® were used at 17-26 wk of age. Their respective vere form of diabetes. Plasma GIP concentrations of age-matched lean littermates were used as controls. The ob/ob and db/db mice were elevated 15.1-fold and 6.2- mice were maintained on a standard pellet diet (Spratts lab fold, respectively; the increments closely corre oratory diet no. 1, Lillico Ltd., Reigate. United Kingdom) and sponded with the degrees of hyperinsulinemia. Small tap water ad libitum. Blood was obtained from fed mice by intestinal weight was increased 1.4-fold and 1.8-fold in cardiac puncture under ether anesthesia for determination ob/ob and db/db mice, respectiveiy, but the intestinal of plasm a glucose," insulin,® and GIP.'^ Immunoreactive GIP GIP content expressed as p.g/g intestine or jxg/intes- was also measured in the small intestine of these mice. The tine was raised only In ob/ob mice (1.9-fold and 2.8- duodenum, jejunum, and ileum were rapidly excised, washed fold, respectively). Since glucose stimulation of insulin with ice-cold 0.154 mmol/L NaCI, weighed, and extracted release is defective in both mutant strains, the results strongly implicate pathologically raised GIP concentra with 5 ml/g acid ethanol (750 ml ethanol, 250 ml water. 15 tions in the hyperinsulinemia and related metabolic ab ml concentrated hydrochloric acid). normalities of the obesity-diabetes syndromes. It is GIP was measured by radioimmunoassay using donkey suggested that hypersecretion of GIP results in part anti-rabbit gamma globulin antiserum (Guildhay Antisera. from loss of normal feedback inhibition by endoge University of Surrey, United Kingdom) for separation of free nous insulin. DIABETES 32:433-435, May 1983. from bound antigen Porcine GIP (Dr. J. C. Brown, University of British Columbia, Canada) was used for the preparation of '‘®l-tracer and as the standard. Charcoal-treated hormone- free plasma was used to minimize nonspecific interference. he possible involvement of gastric inhibitory poly T The GIP antiserum used (RIC 34/111. Guildhay Antisera) was peptide, commonly referred to as glucose-depend raised in the rabbit against a porcine GlP-glutaraldehyde- ent insulinotropic polypeptide (GIP), in the m eta ovalbumin conjugate. This antiserum exhibited negligible bolic disturbances of human obesity and diabetes mellitus has recently attracted considerable interest.'^ How ever. the role of GIP in the pathogenesis of spontaneous Presented in pari at the Glaxo International Symposium on Gut Hormones syndromes of obesity and diabetes in mice has not been and Disease. London. September 1982 investigated. The most extensively studied examples are the From the Department of Biochemistry. Division of Nutrition and Food Science, and Division of Clinical Biochemistry. University of Surrey (PR F . P.K . S K S obese-hyperglycemic {ob/ob) and diabetic-obese {db/db) F. and VM ). Guildford. United Kingdom, and the Department of Biological syndromes.^^ These mutant mice exhibit hyperphagia, ex Sciences (C J.B ). University of Aston in Birmingham. United Kingdom Address reprint requests to Dr. P. R Flatt. Department of Biochemistry. Uni cessive fat deposition, hyperglycemia, and hyperinsulin versity of Surrey. Guildford. Surrey GU2 5XH. United Kingdom emia. The severity and duration of these abnormalities de Received for publication 1 November 1982 DIABETES. VOL 32. MAY 1983 433 ABNORMALITIES OF GIP IN OB OB AND DB DB MICE cross-reactivity wittr ottier enteropancreatic hormones and Release of GIP from the small intestine in response to provided an assay sensitivity of 110 pg/ml with an interassay nutrients' ® can be envisaged to provide an important stim coefficient of variation of 4,1-22.5% over the concentration ulus for insulin secretion in both ob/ob and db/db mice. range of 138-2677 pg/ml Although the physiologic action of GIP in normoglycemic man is equivocal,®® the insulinotropic cyclic AMP-mediated RESULTS effect demonstrable in rodents®®®® undoubtedly constitutes Both groups of mutant mice exhibited the characteristic fea a major stimulus for insulin secretion at the pathologically tures of obesity (reflected by enhanced body weight), hy raised concentrations of GIP and glucose encountered in perglycemia and hyperinsulinemia (Figure 1 ). The db/db mu both strains of mutant mice. Consistent with this view, plasm a tation gave rise to a particularly severe form of diabetes on GIP concentrations were higher in the ob/ob mutant, cor the C57Bl_/KsJ bact 100 - 450 rPO.OOli ^ 360 - 80 II rh § ®'® 3 20 * 40 ^ 180 g I ?0 I 90 I'” 0 0 J1 ob/ob db/db ob/ob db/db ob/ob db/db colony colony colony colony colony colony ob/ot d: dr I olonv deny FIGURE 1. Characteristics of lean and otw se dl- at>etic mice. Vaiues are mean * SEM of groups of rP O.OS -p-O.OOIt seven mice from the ob/ob colony and groups of □ : eight mice from the db/db coiony. The genotype r+i of lean mice from the ob/ob colony was undeter mined ( + /?; expected genotype frequencies 1 □ + / + :2 +lob). Lean mice from the db/db colony consisted of equal numbers of -L/+ and +/db mice, distinguished by the misty (m) coat gene.* rj £ Û Statistical comparisons were made by Student's f ob/ob db/db ob/ob db/db ob/ob db/db test: *P < 0.05; **P < 0.01; ***P < 0.001 compared colony colony colony colony colony colony with lean mice from the same colony. 434 DIABETES. VOL 32. MAY 1983 PETER R FI ATT AND ASSOCIATES ACKNOWLEDGMENTS to glucagon and arginine in Aston ob/ob mice: evidence lor a selective defect in glucose-mediated insulin release. Horm. Metab. Res. 1982; 14:127-30. Mutant colonies were descended from heterozygous ob! + '* Flatt, P. R , Bailey, C J , and Buchanan, K. D : Regulation of plasma and C/D/+ breeding pairs kindly provided by Dr. I. W. Atkins immunoreactive glucagon in obese hyperglycemic (ob/ob) mice. J. Endocri (Department of Biological Sciences. University of Aston in nol 1982; 95:215-27. '• f^latt, P. R., and Bailey. C. J : Development ol glucose intolerance Birmingham, United Kingdom) and Dr. P. Trayhurn (Dunn and impaired plasma insulin response to glucose in otiese hyperglycemic Nutrition Unit. Cambridge. United Kingdom), respectively. (ob/ob) mice. Horm. Metab. Res. 1981; 13:556-60. " Flatt, P. R., and Bailey, C J : Effect of age and glucose priming on the insulin-secretory response to glucose in fed obese hyperglycemic (ob/ ob) mice. Biochem. Soc. Trans. 1982; 10:29-30. '* Flatt, P. R , and Bailey, C. J : Relationship between the nutritional components of the diet and the plasma insulin response to fasting and re REFERENCES feeding in ot>ese hyperglycemic (ob/ob) mice. Submitted for publication. Br. ’ Brown. J. C: Gastric Inhibitory Polypeptide. Berlin. Springer Verlag, J. Nutr. 1982:88. '® Flatt, P. R., and Bailey, C. J.: Importance of the enteroinsular axis for * Creutzfeldt. W: Gastrointestinal peptides—role in pathophysiology the insulin-secretory response to glucose in otiese hyperglycemic (ob/ob) and disease. Scand J. Gastroenterol. 1982; 17(Suppl. 77):7-20. mice. Biochem. Soc. Trans. 1981; 9:220-21. ’ Herkierg, L.. and Coleman. D. L.; Laboratory animals exhibiting obe ** Flatt, P. R., and Bailey, C. J.: Role ol dietary factors in the hyperin sity and diabetes syndromes. Metabolism 1977; 26:59-99. sulinemia of genetically obese hyperglycemic (ob/ob) mice. J. Nutr. 1982; * Bray, G. A., and York, D. A.: Hypothalamic and genetic obesity in 112:2212-16. experimental animals: an autonomic and endocrine hypothesis. Physiol. Rev. Marks, V., and Morgan. L. M.: Gastrointestinal hormones. Mol. As 1979; 59:719-809. pects Med. 1982; 5:225-92. * Coleman. D. L.: Diabetes-obesity syndromes in mice. Diabetes 1982; “ Sarson, D. L.. Wood, S. M.. Holder, D , and Bloom. S. R.: The effect 31 (Suppl. 1):1-6. of glucose-dependent insulinotropic polypeptide infused at physiological con «Hummel, K. P., Coleman, 0. L., and Lane, P. W.: The influence of centration on the release of insulin in man. Diatietologia 1982; 22:34-36. genetic background on expression of mutations at the diabetes locus in the “ Ebert, R . Illmer, K., and Creutzfeldt, W.: Release of gastric inhibitory mouse. I. C57BI_/KsJ and C57BI_/6J strains. Biochem. Genet. 1972; 7:1-13. polypeptide (GIP) by intraduodenal acidification in rats and humans and abol ' Coleman, D. L , and Hummel, K. P.: The influence of genetic back ishment of the incretin effect of acid by GIP-antiserum in rats. Gastroenterology ground on the expression of the ot>ese (ob) gene in the mouse. Diabetologia 1979; 76:515-23. 1973: 9:287-93. Lundquist, I., and Ahren, B.: Cholinergic-peptidergic actions of glu * Boquist, L., Heilman, B , Lernmark. A., and Taljedal, I. B.: Influence cagon and insulin release in mice. Diabetologia 1982; 23:184. of the mutation "diatjetes" on insulin release and islet morphology in mice of “ Szecowka, J., Grill, V., SandlJerg, E.. and Efendic, S.: Effect of GIP different genetic backgrounds. J. Cell Biol. 1974; 62:77-89. on the secretion of insulin and somatostatin and the accumulation of cyclic AMP ® Flatt, P. R., and Bailey. C J.: Abnormal plasma glucose and insulin in vitro in the rat. Acta Endocrinol. 1982; 99:416-21. responses in heterozygous lean (ob/ -L ) mice Diat>etologia 1981; 20:573-77. “ Polak, J. M., Pearse. A. G. E., Grimelius. L., and Marks, V: Gastroin '“ Bailey, C. J., Flatt, P. R , and Atkins. T. W: Influence of genetic testinal apudosis in obese hyperglycemic mice. Virchows Arch. (Cell Pathol.] background on the expression of the ot>ese hyperglycemic syndrome in Aston 1975; 19:135-50. ob/ob mice. Int. J. Okies 1981; 6:11-21. Best, L. C , Atkins, T. W, Bailey, C. J , Flatt, P. R . Newton, D. F, and " Stevens, J. P.: Determination of glucose by automatic analyser. Clin. Matty, A. J.: Increased activity of the enteroinsular axis in obese hypergly Chim. Acta 1971; 32:199-201. cemic mice (ob/ob). J. Endocrinol. 1977; 72:44P. Morgan, L. M„ Morris, B. A., and Marks, V.: Radioimmunoassay of “ Soil, A. H., Kahn, C. R., Neville. D. M.. and Roth, J.: Insulin receptor gastric inhibitory polypeptide. Ann. Clin. Biochem. 1978; 15:172-77. deficiency in genetic and acquired obesity. J. Clin. Invest. 1975; 56:769-80 Coleman. D L.: Thermogenesis in diatietes-obesity syndromes in “ Flatt, P. R., Swanston-Flatt, S. K.. and Bailey, C. J.: Glucagon anti mutant mice. Diabetologia 1982; 22:205-11 serum: a tool to investigate the role of circulating glucagon in obese hyper " Flatt. P R., and Bailey. C J : Plasma glucose and insulin responses glycemic (ob/ob) mice. Biochem Soc Trans. 1979; 7:911-13. DIABETES. VOL 32. MAY 1983 435 79(1 (ilOC'HF.MIC AL SO( 11:1 Y I KANSACTIONS Fernânclc/., F... Valcarcc.( . Cuc/.va. J. M. A Medina. J. M. (IVKj) Lorcn/o, M.. ( aidés. T.. Bcnito, M. & Medina. J. M. (19X1) Biochem. J. 214, 525 552 Biochem. J 198. 425 428 iyncdjian, P. B., Ballard. F J. & Han.son. R. W. (1975) J. BiuL Stalmans. W.. De Wulf. H.. Mue, L. & Hcrs, H. G. (1974) Eur. J. Chem. 280. 5596-5603 Biochem. 41. 127-134 Kalhan. S. C . D'Angclo. L. J.. Savin. M. S. & Adam, P. A. (1979) Tilghman, S. M.. Hanson. R. W.. Rcshcf. L.. Hopgood. M. F. & J. dm . Inivsi. 63. 388-394 Ballard, F. J. (1974) Proc. Nall. Acad. Sci. U.S.A. 71. 1304-1308 Effects of exogenous and endogenous gastric inhibitory polypeptide in obese hyperglycaemic (ob/ob) mice PETER R. FLATT.* CLIFFORD J. BAILEV.t glucose, a similarly marked but considerably protracted PETER KWASOWSKL* TONY PACEf and insulin response was observed. Plasma glucose concen VINCENT MARKS* trations increased steadily and exceeded 30 mmol/l at * Divisions of Nutrition and Ftwd Science, and Clinical 30min. The insulin response cannot be attributed to the B iochemistry. Department of BÙK'hemi.stry. Unicer.sity of hyperglycaemia perse, since administration of glucose alone Surrey, Guildford. SurreyGU2 5XH. U.K.. and produces only a small insulin response (Flatt & Bailey, ] Department of Biological Sciences. University of Aston in 19816). The dose of GIP administered resulted in plasma Birmingham, Birmingham B4 7ET, U.K. Gastric inhibitory polypeptide, also commonly referred to as glucose-dependent insulinotropic polypeptide or GIP, is a 42 amino acid hormone synthesized and released by the K- 40 cellsof the small intestine in response to nutrients. Previous studies have shown that plasma GIP concentrations are inordinately raised in obese hyperglycaemic {objob) mice, 30 and that oral administration of carbohydrate, protein and especially fat evokes marked plasma G IP responses (Flatter 3 Ü al., 1983a; P. R. Flatt, C. J. Bailey, P. Kwasdwski, T. Page rs 2 0 & V. Marks, unpublished work). Overfeeding, K-cell hyperplasia and iniscnsitivity of these cells to suppression by endogenous insulin are envisaged to generate the hypergi- paemia (Flatt et al., 1983a,6). Since GIP is a recognized component of the entero-insular axis (Brown, 1982), the raised concentrations in objob mice suggest an important pathogenic role in the promotion of hyperinsulinaemia and 30 60 90 180 related metabolic abnormalities of the objob syndrome Time (mill) (Flatt et ai, 1983c). The present study examines the insulinotropic action of exogenous and endogenous GIP in 40 relation to the prevailing hyperglycaemia inobjob mice. Obese (objob) mice from the Aston colony were supplied with a standard pellet diet and maintained as described previously (Flatt & Bailey, l98lo; Bailey c/ o/.. 1982). The c 30 mice were used at 10-12 weeks of age and starved for I8h before experimentation. The insulin response to exogenous 3 Ü GIP was examined by intraperitoneal administration of « 20 porcine GIP (40/ig/kg) (Professor J. C. Brown, University of British Columbia, Canada) together with either glucose (2g/kg) or an equivalent volume of saline (0.154 mmol/l NaCI). Blood samples for determination of plasma concen trations of glucose, insulin and GIP were taken from the cut tail-tips of conscious mice at 0, 5, 15. and 3()min. Plasma disappearance of GIP was followed until 180min. The 0 30 60 90 insulin response to endogenous GIP was examined by oral Time (min) administration of fat (Intralipid. 870mg/kg) followed at 30min by an intra peritoneal injection of either glucose Fig. I. Pla.sma insulin re.spon.ses to e.xogenous and endogenous (2g/kg) or an equivalent volume of saline (0.154mmol/1 ga.stric inhibitory polypeptide (GIP) in 18 h .starved objob mice NaCI). Blood samples for plasma glucose, insulin and GIP (a) Plasma insulin concentrations after intraperitoneal determination were taken at 0, 30, 60. and 90min. Insulin administration of GIP (40//g/kg) with cither saline (O------and GIP were measured by radioimmunoassay (Flatt et ai. 0)or 2g/kg glucose (------# #). (6) Plasma insulin concen 1983a). Statistical evaluation was made by Student's /-test. trations after oral administration of fat (Intralipid, As shown in Fig. la. intra peritoneal administration of 870mg/kg) followed at 30min by intraperitoneal injection GIP with saline evoked a marked but transient plasma of either saline (O- * O) or 2g/kg glucose (---- # #). insulin response. Plasma glucose concentrations (not illus Each graph shows plasma GIP concentrations (□------□ ) trated) were not significantly changed, being within the for the saline-treated group only; GIP values for glucose- range 7-10mmol/l. When GIP was administered with treated mice were similar. Values are means+ S.E.M. of groups of four to six mice. *f <0.05 compared with insulin Abbreviation used : GIP. gastric inhibitory polypeptide. concentrations of glucose-treated group. 1984 607th MF.RTING. I.ONDON 791 GIP concentrations similar to the maximum values ob plasma glucose concentrations. Thus it is envisaged that served in actively feeding ohiab mice (Flatt et al., 1983a). elevated plasma GIP concentrations in the presence of post Plasma GIP concentrations declined quickly indicating a prandial hyperglycaemia make an important contribution relatively short half-life of the hormone in objob mice. to the hyperinsulinaemia of objob mice. Fig. 16 shows the plasma insulin responses to endogenous GIP released by oral administration of fat. A substantial increase in plasma GIP concentrations was achieved Bailey. C. J.. Flatt. P. R. & Atkins. T. W. (1982) Int. J. Obesity ft, similar to values in fed objob mice. In mice treated with 11-21 saline at 30min, basal plasma glucose concentrations (not Brown. J. C. ( 1982) Ga.uric Inhibitory Polypeptide, Springer Vcrlag. illustrated) were maintained within the range 7-10mmol/l Berlin throughout. In these mice there was only a very small Flatt. P. R. & Bailey. C. J. (l9 8 lo ) Diabetologia 20, 573-577 increase in plasma insulin concentrations. However, mice Flatt, P. R. & Bailey. C. J. (19816) Biochem. Soc. Trans. 9,220-221 receiving an intraperitoneal injection of glucose at 30min Flatt, P R.. Bailey. C. J.. Kwasowski, P., Swanston-Flatt, S. K. & showed a rise in plasma glucose concentrations exceeding Marks. V. (I983 Short-term effects of wheat bran supplementation on glucose and lipid metabolism in man DAVID J. MOORE, FIONA J. WHITE, medication. Each volunteer’s normal diet was supple PETER R. FLATT and DENNIS V. PARKE mented with 0.3 g of unprocessed wheat bran/kg body Divisions of Nutrition and Food Science, and Biochemistry, weight per day for 6 weeks. At the onset, and after 5 weeks Department of Biochemistry, University of Surrey, Guildford, of wheat bran supplementation, 4-day semi-quahtitative Surrey GU 2 5 X H . U.K. records of dietary intake were made to determine if the con sumption of wheat bran had altered the diet. Fasting blood Extensive studies in man and experimental animals have samples were taken before and 6 weeks after the addition of demonstrated that the addition of certain plant fibres to the wheat bran to the diet. A 0.5 ml aliquot of whole blood was diet is accompanied by a significant decrease in plasma used for glycosylated haemoglobin determination (Kynoch cholesterol, triglyceride and glucose concentrations (Kay & & Lehmann, 1977). Aliquots of plasma were used for the Truswell, 1977; Jenkins er al., 1978; Anderson & Chen, determination of cholesterol (Rudel & Morris, 1973), HDL- 1979). Several water-soluble fibres such as pectin and guar cholesterol (Allen et al., 1979), LDL-cholesterol (Friedewald gum have hypocholesterolaemic and hypoglycaemic effects etal., 1972), triglycerides(Soloni, 1971) and glucose concen (Jenkins et al., 1977) but the effects of the water-insoluble trations (Stevens, 1971). fibre wheat bran have been variable in this respect (Chen & An increase in dietary fibre consumption by 34% was Anderson, 1979). Mucilaginous fibres, although effective associated with only minor changes in the intake of protein, modulators of glucose and lipid metabolism, are of limited fat, carbohydrate, alcohol and cholesterol. As shown in palatability. The present study was therefore undertaken to Table I, there were no significant changes in plasma clarify the possible short-term effects of dietary wheat bran cholesterol, triglyceride or glucose concentrations after supplementation on lipid and glucose metabolism in man. 6 weeks of wheat bran supplementation. Plasma HDL- Seven healthy University students (three males and fbur choleslerol was significantly increased while both LDL- females, aged 18-22 years) volunteered to take part in this cholesterol concentrations and glycosylated haemoglobin study. None of the subjects smoked or received any form of were significantly decreased. The lack of effect of wheat bran supplementation on plasma cholesterol, triglyceride and glucose concentrations is in agreement with previous studies (Jenkins et al., 1975; Truswell & Kay, 1976). Abbreviations used: HDL, high-density lipoprotein; LDL, low- However, the increase in plasma HDL-cholesterol after density lipoprotein. 6 weeks is interesting with regard to the suggestion that this Table I. Short-term effects of wheat bran supplementation on gluco.se and lipid metabolism in man Values arc the means+S.E.M. for seven subjects. Statistical significance was evaluated by Student’s paired /-test : *P<0.05. **P<0.001 compared with the value before wheat bran supplementation. Duration of wheat bran supplementation 0 weeks 6 weeks Plasma total cholesterol (mg/100 ml) 186+17 193+11 Plasma HDL-cholesterol (mg/100ml) 72 + 5 I0 5 ± 7 * Plasma LDL-cholesterol (mg/100 ml) 8 5 ± 7 64 ± 6 * Plasma total triglyceride (mg/IOOmI) 73 + 9 5 3 ± 7 Plasma glucose (mg/IOOmI) 89 + 3 96 + 4 Glycosylated haemoglobin (%) 11.0 + 0.6 6 .0 + 0.2* Vol. 12 » Clinical Endocrinology, (1982) 16, 89-95 24-HOUR PROFILES OF MELATONIN, CORTISOL, INSULIN, C-PEPTIDE AND GIP FOLLOWING A MEAL AND SUBSEQUENT FASTING J. A R E N D T , S. H A M P T O N , J. E N G L I S H P. KWASOWSKI AND V. M A R K S Department of Biochemistry, Division of Clinical Biochemistry, University of Surrey, Guildford, Surrey G U 2 5 X H (Received 3 June 1981; revised 13 August 1981; accepted 20 August 1981 ) SUMMARY Melatonin, free and total cortisol, insulin, C-peptide and glucose-dependent insulin-releasing peptide (GIP) were measured in the plasma of twelve normal volunteers (eight wo m e n and four men), at hourly intervals for 24 h following a meal and subsequent fasting. On e volunteer was excluded from calculations due to a possible effect of stress on melatonin secretion. Melatonin and cortisol showed the normal 24-h variation with peak values at 0200-0500 h, and 0900 h respectively. Following post-prandial stimulation, gut hormones remained basal throughout the sampling period. N o significant relationship was found between 24-h melatonin secretion and basal, or stimulated gut hormone secretion. Melatonin secretion did relate significantly to body weight, suggest ing that data concerning pineal effects in endocrine physiology and pathology, and affective disease, should be reviewed in the light of these observations. In experimental animals, several lines of evidence indicate that the pineal gland ma y have a role to play in the control of carbohydrate metabolism, possibly via an effect on insulin secretion (Bailey et al., 1974; Gorray et al., 1979). Fe w data are as yet available in man, although several cases of familial insulin-resistance with hyperinsulinaemia and pineal hyperplasia have been reported (Barnes et al., 1974; West et al., 1975). Plasma melatonin is at present the best peripheral index of pineal function in ma n and in-vitro inhibitory effects of melatonin on insulin secretion have been described, albeit using pharmacological dose levels (Bailey et ai, 1974). We have therefore investigated the relationship of circulating melatonin in normal volunteers, to body weight, and basal and post-prandial insulin, C-peptide and GIP. Free and bound plasma cortisol were determined as an index of stress during the experiment. W e have also determined the Correspondence: Dr J. Arendt. 0300-0664/82/0100-0089S02.00 © 1982 Blackwell Scientific Publications 89 90 J. Arendt et al. extent to which the concentration of plasma melatonin at particular time points during the night is indicative of the total secretion during 24 h, METHODS Sampling Twelve normal, healthy volunteers, four m e n and eight wo m e n , aged 20-39 years, (x =27-9 ±1-58 S E M ) took part in the experiment. Three of the w o m e n were taking contraceptive steroids, otherwise no volunteer was under medication. Following a standard breakfast of cornflakes, toast, tea, coffee or milk at 0710-0730 h, followed by doughnuts, tea or coffee at 0945 h (695 kilocalories total, 1 0 % protein, 3 6 % fat, 5 4 % carbohydrate), 10 ml blood was sampled by indwelling venous catheter at hourly intervals for 24 h starting at 1000 h. During this period subjects were fasted but allowed unlimited drinking water. During the day subjects occupied themselves with light activities (reading, calculating, watching television). During the night most subjects slept between (and sometimes during) blood sampling. Plasma was immediately decanted, frozen and stored at — 20°C until assay. Volunteers were weighed and their height measured immediately before the sampling period. Sampling was performed between the 5.3.80 and 10.3.80 with the exception of our volunteer (female) wh o was sampled on 29.4.80. Night-time samples were taken in dim artificial light. Assays All hormones were measured by previously published radioimmunoassay (RIA) techniques (Arendt et al., 1977; Mo r g a n et al., 1978; Ha m p t o n et al., 1980; Elliot et al., 1980). All samples from any one individual were measured in the same assay. Assay coefficients of variation were similar to those quoted in the above references. Analysis Basal secretion of the gut hormones was taken to be the baseline values following post-prandial stimulation at the following times: 1600-0900 h (insulin and C-peptide), 1900-0900 h (GIP). ‘Total’ secretion of each hormone was determined for the sampling period by integration of the area under the secretion profile after subtraction of the assay limit of detection in the case of melatonin and cortisol, or after subtraction of me a n basal values in the case of the gut hormones. Integration was accomplished by cutting out and weighing the area under the curve. Correlation coefficients were determined by simple linear regression. ‘Total’ melatonin secretion and individual peak plasma melatonin concentration was compared with body weight, obesity index (height/weight^ x 100, Baird et al., 1974) and ‘total’ and basal secretion of the other hormones. The plasma melatonin concentration at different time points during the night (from 2300 h to 0600 h) was compared with the ‘total’ secretion during 24 h. RESULTS M e a n plasma levels of all hormones throughout the sampling period are shown in Fig. 1. Melatonin and hormones of the entero-insular axis 91 8 0 ~ 4 0 • 70 130 • I 6 0 Ë 20 ■ 3 5 0 Ç 4 0 o 3 0 II 13 15 17 19 21 23 01 0 3 0 5 07 0 9 h •5 2 0 ^ 10 8 0 6 II 13 15 17 19 21 2 3 01 0 3 0 5 0 7 0 9 h 4 250 r 2 200 15 II 13 15 17 19 21 23 01 0 3 0 5 0 7 0 9 b 2000 150 - 10 100 O' 1000 5 5 0 li. II 13 15 17 19 21 2 3 01 0 3 0 5 07 0 9 b II 13 15 17 19 21 2 3 01 0 3 0 5 0 7 0 9 h Fig. 1. Mean plasma levels of melatonin (MT), free (— ) and total (—) cortisol, insulin, C-peptide and GIP in twelve normal volunteers, at hourly intervals for 24-h following food ingestion. See text for sampling conditions. SEM shown, n = 12. 0 -3 0 ■ 0-25 ■ 0*20 ■ 0-14 0-10 0-08 0-04 0-02 20 40 60 80 100 120 140 150 250 200 300 MT (pg/ml) 0 2 0 0 h Fig. 2. The relationship of total melatonin (MT) secretion during 24-h to the plasma concentration at 0200 h in twelve normal volunteers. r = 0-9667. 92 J. Arendt et al. Plasma melatonin showed a night-time rise and low to undetectable values during the day. Melatonin did not show episodic secretion during the day, although seven of the subjects showed evidence of two or more secretory episodes at night. Only one volunteer gave two spikes of secretion during the day comparable with night values. Mean maximum levels reached a plateau between 0200 and 0500 h. Melatonin concentration (pg/ml) between 2300 h and 0600 h correlated well with ‘total’ secretion over 24 h: correlation coefficients varied from 0-9004 (P<0-01, 0600 h) to 0-9667 ( f <0 01, 0200 h). Fig. 2 illustrates the 0200 h relationship and indicates the usefulness of single time-point determination of night-time melatonin, as an index of total secretion. The mean plasma levels of the pancreatic and gut hormones peaked 75 min after food ingestion (Fig. 1). In view of the sampling conditions, we were unable to estimate the individual peak values: in view of the rapid post-prandial rise, more frequent sampling would have been necessary. Following stimulation, values remained basal for the rest of the sampling period and cyclic oscillations were not detectable. Plasma free and bound cortisol showed the well-known 24-h rhythm with mean peak values at 0900 h. Serious problems with the cannula during the night in one volunteer were reflected in a massive increase in free, and to a lesser extent, bound cortisol, and very high melatonin was recorded: (volunteer S.P. 0400 h, MT =151 pg/ml, group mean = 54-1 ±7-8 pg/ml,x±SEM , «=11; free cortisol = 45-2 ng/ml, group mean = 5-7+1-1 ng/ml, X SEM, « = 11). In view of a possible effect of stress on melatonin secretion (Lynchet al.. Table 1. Significant relationships between MT secretion and different variables in eleven normal volunteers, three men and eight women Coefficient of No. of correlation Significance Variable I Variable II pairs (r) level ‘Total’ MT secretion body weight (kg) 11 0-6482 P<0-05 during 24-h sampling period (arbitrary units)* 8 0-8279 P < 0-01t height ^ - ^ ^ 2 x 100 11 -0-6080 f <0 05 weight 8 -0-8530 P < 0-01t Peak MT concentration body weight (kg) 11 0-6681 P<0-05 (pg/ml) 8 0-8880 P < 0-01t height — 1^ 2 x 100 11 -0-5950 P<0-05 weight 8 -0-887 P < 0-01t ‘total’ insulin secretion 11 0-6134 P<0-05 during 24-h sampling period (arbitrary units)* * Defined in text, t Women only. Melatonin and hormones o f the entero-insular axis 93 1973; Leone et al., 1979), this volunteer was excluded from further calculations. In the remaining eleven volunteers, melatonin did not relate to cortisol secretion. The significant relationships between melatonin, body weight and other hormones in the remaining eleven volunteers are shown in Table 1, No significant correlations were found between melatonin and C-peptide, or GIP, or between ‘total’ melatonin and insulin secretion. It was noted however that peak plasma melatonin concentration correlated with ‘total’ insulin secretion. Body weight in these volunteers ranged from 46 5 to 97-0 kg (7 M 8 ± 5 04, x ± SEM, «=11) and was found to relate significantly to both ‘total’ and peak plasma melatonin values with greater significance when women were considered separately (Table 1). The individual obesity index also correlated to ‘total’ plasma melatonin with greater significance when women were considered separately (Table 1). In women only, a weak relationship was noted between total insulin and body weight (r= 0*5819, f <0 05). In a further group of eleven male volunteers, aged 39-70 years (53-7 ±2*6, x +SEM), who served as controls for a recent schizophrenia study (Perrier et al., in preparation) melatonin concentration at 2400 h again related to body weight (r=0 66, f <0 05). Neither peak plasma melatonin concentration nor ‘total’ plasma melatonin related to age in the present group of subjects. DISCUSSION Previous reports of episodic melatonin, insulin and C-peptide secretion in man have employed shorter sampling intervals than those reported here. In the case of insulin and C-peptide, Langet al. (1979) have reported cyclic oscillations of basal values, with small amplitude and a period of 10-15 min in ten volunteers, which would not be detectable with our sampling conditions. In the case of melatonin, the daytime plasma values reported by Weinberget al. (1979) in five subjects and described as episodic secretion are so high (up to 20-fold higher than those reported here) that the probable contribution of cross-reactants to their values should be considered. A further report of episodic daytime melatonin secretion by Linsell et al. (1979) using gas-chromatography-mass-spectro- metry, concerns one subject. If episodic daytime melatonin secretion occurs, it is probably below the limit of detection (7-14 pg/ml) of our assay. In view of the possible non-specific elevation of plasma melatonin in response to psychological stress, it is clearly important to control for stress when assessing pineal function. However, the lack of a significant melatonin-cortisol relationship in the majority of subjects is in agreement with the observations of Vaughanet al. (1978; 1979). Plasma melatonin may be affected by stage of the menstrual cycle (Arendt, 1978; Tapp et al., 1980) and possibly by oral contraceptives (Tapp et ai, 1980) in the morning, but the evidence is inconclusive. Total secretion as assessed by 6-hydroxy melatonin sulphate determination did not change during the cycle in five women (Fellenberget al., 1980) but this study provides no information on possible changes in the timing of melatonin secretion during the menstrual cycle. For the present experiment it is assumed that menstrual cycle does not significantly affect total secretion, and it is not possible to assess any contribution of contraceptive steroids due to the small number of subjects. This is the first report to our knowledge of a relationship of body weight to melatonin in man. It is possible that melatonin is related more to body fat than size; four of the female volunteers in the present study exceeded ideal body weight by more than 10%, and all weight calculations were more significant in females. Clearly, however, more information 94 J. Arendt et al. is required in order to confirm or deny this suggestion, and all deductions must be tentative in view of the small number of subjects in this study. It is conceivable that there is a cause-effect relationship between melatonin secretion and body weight. Evidence for a pineal involvement with growth indicates that in rats (Relkin 1976) and sheep (Forbes et al., 1981) it ma y mediate the effects of photoperiod on growth. Also, brown adipose tissue (BAT) is currently thought to be of importance in weight control (Rothwell & Stock, 1979): melatonin administration increases B A T in hamsters (Heldmaier & Hoffman, 1974) and it is possible that B A T is modified by pineal secretory products in man. Obesity is kn o w n to be associated with high circulating insulin levels (K a r a m et al., 1963) and it is possible that the association of peak melatonin concentration to stimulated insulin secretion results from both hormones being related to body fat. However the general lack of association between stimulated and basal levels of the entero-insular hormones and melatonin production suggests that pineal indoleamines are not concerned with long-term regulation of pancreatic and gut-hormone secretion. An y acute effects of melatonin on pancreatic and gut hormone secretion might be sh own by feeding during the night under suitably controlled conditions. M e a n plasma melatonin values at 2400 h under fasting conditions reported here (51-7+15 pg/ml, X + SE M ) were similar to those of a group of eleven volunteers with the same range of age and weight sampled at 2400 h at the same time of year as this study following an evening meal (68-0 + 8 pg/ml, x SE M , Arendt et ai, 1980). Thus it is unlikely that short-term fasting affects melatonin secretion in man. Possible effects of the pineal gland on growth and maintenance of body weight in ma n merit further investigation, inter alia with reference to anorexia nervosa and the associated endocrine abnormalities. Data concerning melatonin levels in the course of puberty ma y prove mo r e rewarding if considered in conjunction with the gross changes in body weight that occur at this time. Furthermore, when comparing control and experimental or pathological populations, matching for body weight is clearly of importance. ACKNOWLEDGEMENTS W e thank the M R C for financial support during this study. REFERENCES A r e n d t , J., W etterberg, L., H e y d e n , T., Sizonenko, P.C. & Paunier, L. (1977) Radioimmunoassay of melatonin: human serum and cerebrospinal fluid. Hormone Research, 8 , 65-75. Bailey, C.J., A tkins, T.W. & M alty, A.J. (1974) M elatonin inhibition of insulin secretion in the rat and mouse. Hormone Research, 5, 21-28 B a ird , I., Silverstone, J.T., G rim s h a w , J.J. & A s h w e ll, M. (1974) Prevalence of obesity in a London borough. Practitioner, 212, 706-714. B a rn e s , N.D., P a lu m b o , P.J., H a y le s , A.B. & Folgar, H. (1974) Insulin resistance, skin changes and virilization: a recessively inherited syndrome possibly due to pineal dysfunction. Diabetologia, 10,285-289. E l l i o t , P R., Powell-Tuck, J., Gillespie, P.E., L a id lo w , J.M., Lennard-Jones, J.E., E n g l is h , J., Chakraborty, j. & M a r k s , V . (1980) Prednisolone absorption in acute colitis. Gut, 21, 49-51. Fellenberg, A. J., P h illip o u , G . & S e a m a rk , R. F . ( 1980) Specific quantitation of urinary 6-hydroxy melatonin sulphate by gas-chromatography mass spectrometry. Biomedical Mass Spectrometry, 1, 84-87. Forbes, J.M ., Brown, W.B., Al-Banna, A.G.M. & Jones, R. (1981) The effect of daylength on the growth of lam b s. Animal Production, 32, 2 3 -2 8 . Melatonin and hormones of the entero-insular axis 95 G o r r a y , K.C.. Q u a y , W.B. & E w a r t , R.B.L. (1979) Effects of pinealectomy and pineal incubation medium and sonicates on insulin release by isolated pancreatic islets in vitro. Hormone and Metabolic Research, II, 432-436. H a m p to n , S.H. & M a r k s , V. (1979) Development, validation and use of a human C-peptide radioimmuno assay. British Diabetic Association Meeting, Bristol, 1979. Diabetologia, 17, Heldmaier, G. & Hoffmann, K. (1974) Melatonin stimulates growth of brown adipose tissue. Nature, 247, 224-225. K a ra m , J.S., Grodsky, G.M. & Forsham, P.H. (1963) Excessive insulin response to glucose in obese subjects as measured by immunochemical assay. Diabetes, 12, 197-204. L a n g , D.A., M a tt h e w s , D R., P e to , J. & T u r n e r , R.Ç. (1979) Cyclic oscillations of basal plasma glucose and insulin concentrations in human beings. New England Journal of Medicine, 301, 1023-1027. L eo n e, R.M., S ilm a n , R.E., H o o p e r, R.J.L., C a r t e r , S.J., F in n ie , M.A., E d w a r d s , R., S m ith , I., F r a n c is , P. & M u l l e n , P.E. ( 1979) A routine assay for methoxytroptophol and melatonin in the peripheral circulation using gas-chromatography-mass-spectrometry. Progress in Brain Research, 52 , 263-265. Linsell, C., M ullen, P.E., S ilm a n , R.E., L e o n e , R.M., F in n ie , M.A., C a r t e r , S.J., H o o p e r, R.J.L., S m ith , 1. & Francis, P. (1979) The measurement of the daily fluctuations of 5-methoxytryptophol in human plasma. Progress in Brain Research, 52, 501-505. L y n c h , H.J., E n g ., J.P. & W u r tm a n , R.J. (1973) Control of pineal indole biosynthesis by changes in sympathetic tone caused by factors other than environmental lighting. Proceedings o f the National Academy o f Science, 70, 1704-1707. M o r g a n , L.M., M o r r is , B.A. & M a r k s , V. (1978) Radioimmunoassay of gastric inhibitory polypeptide. Annals o f Clinical Biochemistry, 15, 172-177. Relkin, R. (1976) The Pineal. Eden Press, Montreal. R o t h w e l l , N.J. & S to c k , M.J. (1979) A role for brown adipose tissue in diet-induced thermogenesis. Nature, 281, 31-35. T a p p , E., S t. J o h n , J.G. & S k in n e r, L. (1980) Serum melatonin levels during the menstrual cycle. Abstract, International Symposium on Melatonin, Bremen, 1980. V a u g h a n , G.M., A l l e n , J.P., T u l l i s , W., Siler-Khodr, T.M., L a P e n a , A. D e & S a c k m a n , J.W. (1978) Overnight plasma profiles of melatonin and certain adenohypophyseal hormones in men. Journal o f Clinical Endocrinology and Metabolism, 47, 566-571. V a u g h a n , G.M., M c D o n a ld , S.D., J o r d a n , R.M., A l l e n , J.P., B e l l , R. & Stevens, E.A. (1979) Melatonin, pituitary function and stress in humans. Psychoneuroendocrinology, 4, 351-362. W e in b e rg , U., D ’E l e t t o , R.D., W e itz m a n , E.D., E r l i c h , S. & H ollander, C.S. (1979) Circulating melatonin in man: episodic secretion throughout the light-dark cycle. Journal of Clinical Endocrinology and Metabolism, 48, 1 14-118. W e s t, R.J., Lloyd, J.K. & T u r n e r , W.M.L. (1975) Familial insulin-resistant diabetes, multiple somatic anomalies, and pineal hyperplasia. Archives o f Disease in Childhood, 50, 703-708. NUTRITION REPORTS INTERNATIONAL GLUCOREGULATORY EFFECTS OF CAFETERIA FEEDING AND DIET RESTRICTION IN GENETICALLY OBESE HYPERGLYCAEMIC (ob/ob) MICE Peter R. Flatt*, Clifford J. Bailey’*’, Peter Kwasowski*, Sara K. Swanston-Flatt* and Vincent Marks* *Divisions of Nutrition and Food Science, and Clinical Biochemistry, Department of Biochemistry, University of Surrey, Guildford, Surrey, GU2 5XH, U.K., and ’^’Department of Molecular Sciences, University of Aston in Birmingham, Birmingham, B4 7ET, U.K. ABSTRACT In a 42 day study, energy intake of young adult (12-16 weeks old) obese hyperglycaemic (ob/ob) mice was increased by 45% with a supplementary cafeteria diet, and decreased by 47% when the feeding period was restricted to 2 hours daily. Compared with control mice (8% increase in body weight), body w eight was increased by 20% in cafeteria fed m ice and decreased by 18% in diet restricted mice. Cafeteria feeding considerably increased (up to 138%) plasma concentrations of gastric inhibitory polypeptide (GIP), associated with a smaller increase (50%) in plasma insulin, and no significant changes in basal plasma glucose or oral glucose tolerance. Dietary restriction substantially reduced plasma GIP, insulin and glucose concentrations (up to 33%, 25% and 30% of control values respectively), but only glucose reached the concentrations observed in lean (+/+) mice. Insulin resistance, assessed by insulin hypoglycaemia tests, was decreased in diet restricted mice but not significantly altered in cafeteria fed mice. Diet restricted mice also showed a marked reduction in the glycaemic response to a test meal. The results indicate that in young adult ob/ob mice, energy dissipation is increased during extra energy intake, while energy conservation is increased when intake is reduced. The results are also consistent with the view that additional food consumption enhances nutrient-induced GIP secretion which promotes the hyperinsulinaemia. INTRODUCTION The inordinate adiposity of genetically obese hyperglycaemic (ob/ob) mice has been attributed to the combined effects of increased metabolic efficiency and hyperphagia (1,2). Increased metabolic efficiency is evident during suckling, when energy intake is normal, implicating this disturbance as a causative influence in the genesis of obesity. Hyperinsulinaemia, a prominent endocrine abnormality of the ob/ob syndrome, also arises before weaning (1). However, it is the hyperphagia which grossly exacerbates the hyperinsulinaemia after weaning. The hyperinsulinaemia promotes the obesity, insulin resistance, hyperglycaemia and multiple metabolic disorders of the ob/ob mutant (3). Amelioration of the glucoregulatory abnormalities of ob/ob mice during dietary restriction has been documented previously (1,4), but little attention has been given to the effects of overfeeding. A recent study has shown that the hyperphagia and obesity of ob/ob mice can be increased using a varied diet of palatable cafeteria food items (5). The effect of extra energy consumption on OCTOBER 1985 VOL. 32 NO. 4 847 NUTRITION REPORTS INTERNATIONAL the hyperinsulinaemia and impaired glucose homeostasis of ob/ob mice has not been examined. This issue is addressed in the present study by comparing the changes in plasma glucose and insulin concentrations in ob/ob mice overfed with a cafeteria diet or subjected to dietary restriction. The study includes measurements of plasma gastric inhibitory polypeptide (GIP), a glucose-dependent insulinotropic hormone released from the intestine during feeding (6). GIP concentrations are markedly raised in ob/ob mice (7,8) and appear to constitute an important part of the overactive enteroinsular axis linking the hyperphagia and hyperinsulinaemia. MATERIALS AND METHODS Animals. Aston obese hyperglycaemic (ob/ob) mice and lean (+/+) mice bred at the University of Surrey were used at 12-16 weeks of age. The origin and characteristics of the Aston strain of mice have been described previously (9,10). Mice were maintained in an air-conditioned room at 22 + 2°C with a lighting schedule of 12 hours light (0700-1900 h) and 12 hours dark. Diets. Mice were supplied ad libitum with tap water and a stock laboratory pellet diet (Spratts laboratory diet no. 1, Lillico Ltd., Reigate, U.K.) comprising 48% carbohydrate, 21.5% protein, 3.5% fat, 2.7% fibre; metabolizable energy 3400 kcal/kg. The cafeteria diet used in the present study consisted of 14 different items supplied as two different items daily in addition to the stock diet. Table I lists the food items and the amounts consumed. The mice readily consumed each of the cafeteria food items, and a particular preference for tuna fish was observed from amongst the selection provided. The overall composition of the cafeteria diet consumed during 1-38 days of study was 38.3% carbohydrate, 13.8% protein, 22.2% fat, 3.0% fibre; metabolizable energy 4035 kcal/kg calculated from food tables (11). TABLE I Composition of cafeteria diet æ d amounts of cafeteria food items consumed by (ob/ob) mice Food Item Amount consumed Energy consumed (g/mouse/24 h) (kcal/mouse/24 h) Tuna fish 6.9 + 0.4 19.8 + 1.2 Peanuts 5.3 + 0.4 30.1 + 2.4 Madeira cake 5.1 + 0.5 20.1 + 1.9 Muesli 4.6 + 0.2 17.0 + 0.7 Cheese 4.4 + 0.3 18.0 + 1.2 Milk chocolate 4.1 T 0.4 21.7 + 2.1 Chocolate digestives 3.7 + 0.3 18.2 + 1.4 Crisps 3.7 + 0.2 19.6 + 1.2 Luncheon meat 3.5 + 0.4 11.0 + 1.2 Mars bar 2.7 + 0.3 12.0 + 1.2 Frankfurter 2.2 + 0.2 6.0 + 0.5 Macaroni 2.1 T 0.3 7.9 + 1.2 Gingernuts 1.5 + 0.2 6.7 + 1.0 Dates 1.4 + 0.2 3.6 + 0.5 Values are mean + SEM of 6 mice receiving each cafeteria food item on 5 occasions. Two different cafeteria food items were supplied ad libitum each day as a supplement to stock diet, also suppiied ad libitum. 848 OCTOBER 1985 VOL. 32 NO. 4 NUTRITION REPORTS INTERNATIONAL Experimental Procedure. Equal numbers of male and female ob/ob mice were assigned to each of 3 groups. Control mice were supplied stock diet ad libitum. Cafeteria fed mice were supplied stock diet plus two cafeteria food items daily ad libitum. Diet restricted mice were subjected to progressively shorter feeding periods during which stock diet was supplied ad libitum. Access to food outside of these periods was prohibited. The feeding periods were 7 hours/day (1000-1700) for days 1-7; 4 hours/day (1000-1400) for days 8-14; 2 hours/day (1000-1200) thereafter. The study was continued for 42 days, and body weight and food intake were monitored daily for 1-38 days. Blood samples (100 pi) for plasma glucose, insulin and GIP analyses were taken from the tail tip of conscious mice at the times shown in Figure 2. For comparison at the end of the study these parameters were also measured in age and sex matched lean (+/+) mice supplied stock diet ad libitum. Test Procedures. The following tests were performed at 1000 hours on days 39-42. Insulin hypoglycaemia tests were undertaken on the three groups of mice. Control and cafeteria fed mice were studied in the fed state and received an intraperitoneal injection of monocomponent porcine insulin (Actrapid, Novo, Copenhagen, Denmark), 100 U/kg body weight. Diet restricted mice, which had been deprived of food for 22 hours received 20 U insulin'/kg body weight. Glucose tolerance tests (2 g glucose/kg body weight, administered orally in a 40% w/v solution) were conducted on freely fed control and cafeteria fed mice. Food was withheld during both the insulin hypoglycaemia and glucose tolerance tests. A 2 hour test meal of stock diet ad libitum was given to control and diet restricted mice. The control mice were fasted from 1200 hours on the previous day (i.e. fasted for 22 hours), corresponding to the period when food was withheld from the diet restricted mice. In each of the three tests, blood samples (20 pi) were taken from the tail tip for plasma glucose analysis at the times shown in the Figures. Analyses. Plasma glucose, insulin and GIP were measured as described previously (7,8). Data are presented as mean £ SEM of 6 mice. Student's paired and unpaired £-tests were used as appropriate for statistical comparison. Differences were considered to be significant for P < 0.05. RESULTS Body weight and energy intake of control, cafeteria fed and diet restricted ob/ob mice are illustrated in Figure 1 and summarised in Table II. Supplementation of the stock diet for 38 days with a variety of palatable cafeteria food items increased energy intake by 45% compared with control mice. The overall partition of energy intake was 90% from the cafeteria diet and 10% from the stock diet, with daily variations of 65% to 98% from the cafeteria diet, depending upon the items supplied. Body weights of the two groups of mice did not significantly differ, although the cafeteria fed mice showed a significantly (P < 0.05) greater increase in body weight than control mice (20.4% and 7.6% respectively). Conversely, a restricted feeding period of 2 hours daily reduced total energy intake by 47%, accompanied by a progressive fall in body weight. In cafeteria fed mice, plasma GIP concentrations were increased (56% and 138% respectively) at 28 and 35 days, and the level of hyperinsulinaemia was increased (50%) at 35 days (Figure 2). The level of hyperglycaemia in cafeteria OCTOBER 1985 VOL. 32 NO. 4 849 NUTRITION REPORTS INTERNATIONAL 100 O) 80 60 40 « ■q ./ Cafeteria ^ 20 Stock ••t’* 20 Days FIGURE 1. Panels a and b illustrate body weigtit and energy intake in control (------), cafeteria fe3 (------) and diet restricted (•-•-•) ob/ob mice. Panel c illustrates ttie partition of energy intake between cafeteria food items and stock diet in the cafeteria fed mice. Values are mean + SEtvl of 6 mice. TABLEn Body Weight and E^rqy Intake of Control, Cafeteria Fed and Diet Restricted ob/ob Mice Body Weight Energy intake Initial Final Change Change Total (day 0) (day 38) (days 0-38) (days 0-38) (days 0-38) 9 9 g/mouse % kcal/mouse Control 79.5 + 4.7 65.6 + 3.2 + 6.1 + 3.3 + 7.6 865 + 50 Cafeteria fed 76.3 + 5.2 91.9 + 5.3+ +15.7 + 2.3® +20.4*= 1257 + 20*= Diet restricted 75.0 + 5.3 61.2 + 4.1+^* -13.8 + 1.4*=* -18.4*=* 460 + 10®* Values are mean + SEtvl of 6 mice. P < 0.05, P < 0.01, ^ P < 0.001 compared with control mice. ♦ P < 0.001 compared with cafeteria fed mice. ^ P < 0.001 compared with initial body weight. 850 OCTOBER 1985 VOL. 32 NO. 4 NUTRITION REPORTS INTERNATIONAL fed mice was not significantly altered. Diet restricted mice consumed substantially less food than control ob/ob mice (for example 3.9 + 0.1 compared with 7.5 + 0.2 g/mouse/day on day 37, mean + SEM, P < 0.001). THis amount was slightly less than that consumed by lean mice fed ad libitum (5.4 + 0.5 g/mouse/day, n = 6, P < 0.05). The diet restricted mice showed reduced concentrations of the three plasma constituents measured. Plasma glucose concentrations fell to 5.9 + 0.7 mmol/l. This was within the range for fed lean mice 7.3 + 0.4 mmol/l. Plasma insulin and GIP concentrations, although markedly reduced in diet restricted ob/ob mice, remained in excess of values in fed lean mice (1.6 + 0.3 ng/ml and 289 + 11 pg/ml respectively). Insulin hypoglycaemia tests conducted at the end of the study are illustrated in Figure 3. Due to the severe resistance of ob/ob mice 100 U insulin/kg was administered intraperitoneally to control and cafeteria fed mice. The insulin induced fall in plasma glucose was similar in these groups. However, a smaller dose of insulin (20 U/kg) produced a much greater fall in plasma glucose in the diet restricted mice, demonstrating a less severe state of insulin resistance. Consistent with these data, oral glucose tolerance was similar in control and cafeteria fed mice (Figure 4). Moreover, the glycaemic response to a 2 hour test meal was reduced in diet restricted mice compared with control mice (Figure 4), 30 20 g l u c o s e 100 10 0 40 S Pla sm a insulir 20 ng/ml 0 30 60 M i n u t e s 0 FIGURE 3. Percentage change in plasma glucose concentrations of control ( • ------•), cafeteria fed (■- - -■) and diet restricted 4 (a a ) ob/ob mice during intraperitoneal P la s m a insulin hypoglycaemia tests conducted at the GIP end of the study. Control and cafeteria fed ng/ml mice were freely fed and received 2 100 U insulin/kg. Diet restricted mice were deprived of food for 22 hours and received 20 U insulin/kg. Values are mean + SEM of 6 0 mice. 0 7 14 21 28 35 Days FIGURE 2. Plasma glucose, insulin and GIP concentrations in control ( • ------•), cafeteria fed (■------■) and diet restricted ( a a) ob/ob mice. Values are mean + SEM of 6 mice. OCTOBER 1985 VOL. 32 NO. 4 851 NUTRITION REPORTS INTERNATIONAL 40 30 P l a s m a 30 20 g l u c o s e mmol/l 20 10 0 0 0 30 60 0 2 5 M i n u t e s H o u r s FIGURE 4. Left panel; plasma glucose concentrations of freely fed control (• ------•) and cafeteria fed (■ m) ob/ob mice during oral glucose tolerance tests (2 g/kg body weight). Right panel; plasma glucose concentrations of 22 hour fasted control (• ------•) and diet restricted ( a a ) ob/ob mice during a 2 hour test meal Values are mean + S E M of 6 mice. despite the consumption of more food by the diet restricted group (3.8 + 0.1 and 1.5 + 0.1 g/mouse/2 hours) in diet restricted and control mice respectively, P < 0:001). DISCUSSION The results confirm that hyperphagia in ob/ob mice can be increased by supplementing the stock diet with a variety of palatable cafeteria food items (5). The mice consumed much greater quantities of the cafeteria foods than the stock diet, and generally exhibited a preference for the more fatty food items (12). Without knowledge of the energy content of the mice at the beginning and end of the study, and of energy expenditure during the study, an accurate determination of metabolic efficiency cannot be made. However, the relatively small increase in body weight of the cafeteria fed mice (20%) compared with control mice (8%), despite the 45% greater energy intake in cafeteria fed mice, indicates that mechanisms for energy dissipation were increased during extra energy intake. Preadult ob/ob mice examined previously (5) stored a larger proportion of the additional energy consumed during cafeteria feeding than the young adult mice in the present study. This is consistent with the more rapid accretion of adipose tissue at the younger age (10). Although the capacity for energy dissipation is severely compromised in ob/ob mice (1,2), older ob/ob mice may possess a greater capacity than preadult ob/ob mice. This cannot be attributed to increased energy expenditure in physical activity, since ob/ob mice become less active with age. Whether there is an age-related alteration in the thermogenic activity of brown adipose tissue, or the activity of other energy consuming processes in ob/ob mice is not clear (1,2,13). The effect of overfeeding in ob/ob mice contrasts with lean animals, which show a much greater capacity for dietary induced adiposity at a younger age (14). As previously noted (1,4,13), diet restricted ob/ob mice are even more highly efficient energy conservers than freely fed ob/ob mice, although the biochemical basis of this adaptation awaits elaboration. In the presence of hyperglycaemia, GIP is a potent insulinotropic agent (6), and excessive production of GIP particularly in response to dietary fat. 852 OCTOBER 1985 VOL. 32 NO. 4 NUTRITION REPORTS INTERNATIONAL contributes to nutrient induced hyperinsulinaemia in ob/ob mice (3,7,8). Raised GIF concentrations may promote the adiposity of ob/ob mice by directly stimulating lipogenesis as well as via insulin (6,7). Thus the increased quantity of food and the greater proportion of fat consumed by the cafeteria fed mice (22.2% fat compared with 3.5% fat for stock diet) could account for the additional concentrations of GIP. This in turn would contribute to increased hyperinsulinaemia and obesity (15). Consonent with this view, dietary res-triction markedly reduced plasma GIP and insulin concentrations. However, the concentrations of these hormones in diet restricted ob/ob mice did not reach those of lean mice, despite consumption of less food than lean mice. This emphasises that hyperphagia alone is not entirely responsible for the raised insulin and GIP concentrations in ob/ob mice, although the hyperplasia of islet B cells and intestinal K cells associated with hyperphagia represents a major contributory factor. The fall in glucose concentrations in diet restricted ob/ob mice to within the range of lean mice, and the reduced glycaemic response to meal ingestion, supports other evidence that chronic hyperphagia is important in the development of hyperglycaemia and glucose intolerance in ob/ob mice (1,3). This may be related in part to the reduced insulin resistance. However, the insulin dose (20 U/kg, ip) required to produce a marked fall of plasma glucose in diet restricted mice was still some 80 fold greater than required in lean mice (16). Although hyperplagia strongly promotes the pathogenesis of the ob/ob syndrome, the additional intake of energy induced by cafeteria feeding did not cause a further significant deterioration of hyperglycaemia, glucose intolerance and insulin resistance. This is likely to reflect the short duration of the study and the age of the ob/ob mice used. Also, the inordinate severity of these abnormalities in young adult mice may obscure the measurement of subtle changes by the tests employed. However, the coexistence of similarly raised glucose concentrations but greater hyperinsulinaemia in the cafeteria fed mice indicates an aggravation of the ob/ob syndrome. REFERENCES 1. Bray, G. A. and York, D. A. Hypothalamic and genetic obesity in experimental animals: an autonomic and endocrine hyoothesis. Physiol. Rev. 59, 719 (1979). 2. Himms-Hagen, J. Brown adipose tissue thermogenesis in obese animals. Nutr. Rev. 41, 261 (1983). 3. Flatt, P. R., Bailey, C. J., Swanston-Flatt, S. K., Best, L., Kwasowski, P., Buchanan, K. D. and Marks, V. Involvement of glucagon and GIP in the metabolic abnormalities of obese hyperglycaemic (ob/ob) mice. In: Lessons from Animal Diabetes (E. Shafrir and A. E. Renold, eds.), John Libbey, London, 1984, p. 341. 4. Flatt, P. R. and Bailey, C. J. Glucose and insulin response to conditioned feeding in lean and genetically obese hyperglycaemic (ob/ob) mice. Metabolism 32, 504 (1983). 5. Trayhurn, P., Jones, P. M., McGuckin, M. M. and Goodbody, A. E. Effects of overfeeding on energy balance and brown fat thermogenesis in obese (ob/ob) mice. Nature 295, 323 (1982). OCTOBER 1985 VOL. 32 NO. 4 853 NUTRITION REPORTS INTERNATIONAL 6. Brown, J. C. Gastric inhibitory polypeptide. Monographs on Endocrinology, Vol. 24, Springer-Verlag, Berlin, 1982. 7. Flatt, P. R., Bailey, C. J., Kwasowski, P., Page, T. and Marks, V. Plasma immunoreactive gastric inhibitory polypeptide in obese hyperglycaemic (ob/ob) mice. J. Endocr. 101, 249 (1984). 8. Flatt, P. R., Bailey, C. J., Kwasowski, P., Swanston-Flatt, S. K. and Marks, V. Abnorm alities of GIP in spontaneous syndromes of obesity and diabetes in mice. Diabetes 32, 433 (1983). 9. Flatt, P. R. and Bailey, C. J. Abnormal plasma glucose and insulin responses in heterozygous lean (ob/+) mice. Diabetologia 20, 573 (1981). 10. Bailey, C. J., Flatt, P. R. and Atkins, T. W. Influence of genetic background and age on the expression of the obese hyperglycaemic syndrome in Aston ob/ob mice. Int. J. Obes. 6, 11 (1982). 11. Paul, A. A. and Southgate, D. A. T. McCance and Widdowson's "The Composition of Foods", 4th edition, HMSO, London, 1978. 12. Romsos, D. R. and Ferguson, D. Self-selected intake of carbohydrate, fat and protein by obese (ob/ob) and lean mice. Physiol. Behav. 28, 301 (1982). 13. Coleman, D. L. Thermogenesis in diabetes-obesity syndromes in mutant mice. Diabetologia 22, 205 (1982). 14. R othwell, N. J. and Stock, M. J. Thermogenesis: com parative and evolutionary considerations. In: The Body Weight Regulatory System: Normal and Disturbed Mechanisms. (L. A. Cioffi, W. P. T. James and T. B. Van Itallie, eds.). Raven Press, New York, 1981, p. 335. 15. Bailey, C. J. and Flatt, P. R. Animal models of diabetes. In: Recent Advances in Diabetes (M. Nattrass, ed.). Vol. 2, Churchill Livingstone, Edinburgh, in press. 16. Bailey, C. J. and Flatt, P. R. Hormonal control of glucose homeostasis during development and aging in mice. Metabolism 31, 238 (1982). Accepted for publication: July 11, 1985. 854 OCTOBER 1985 VOL. 32 NO. 4 Acta Endocrinologica 1986,112: 224-229 Immunoreactive gastric inhibitory polypeptide and K cell hyperplasia in obese hyperglycaemic (ob/ob) mice fed high fat and high carbohydrate cafeteria diets C. J. Bailey^, P. R. Flatt^ P. Kwasowski C. J. PowelF and V. Marks^ Department o f Molecular Sciences^, Aston University, Birmingham and Department of Biochemistry^ University of Surrey, Guildford, UK Abstract. The effect of diet composition on plasma as a physiological component of the enteroinsular and intestinal concentrations of immunoreactive gastric axis (Brown 1982; Creutzfeldt et al. 1983), and inhibitory polypeptide (GIP) and intestinal K cell den hypersecretion of GIP has been suggested as a sity was examined in obese hyperglycaemic (ob/ob) factor contributing to the hyperinsulinaemia of mice. The mice were reared from 3 to 11 weeks of age ob/ob mice (Flatt et al. 1984a,b). The raised GIP on either stock diet, a high fat (HF) cafeteria diet or a concentrations of ob/ob mice render this mutant a high carbohydrate (HC) cafeteria diet. The HF cafete convenient model for studies of GIP physiology. ria diet increased the concentration of GIP in plasma (75%) and in the intestine (118%) and increased the As noted in other species (Brown 1982), orally density (54%) of GIP-secreting K cells in the upper administered fat elicits a greater acute plasma GIP jejunum compared with the stock diet. Plasma and response than other nutrients in ob/ob mice (Flatt intestinal GIP concentrations were not significantly et al. 1984a), but the long-term effects of different altered by the HC cafeteria diet, although the density of dietary components on plasma and intestinal GIP K cells in the upper jejunum was increased (45%). The remain to be established. extent of hyperglycaemia and hyperinsulinaemia in The present study examines the chronic effects ob/ob mice was not significantly altered by the HF of excess fat and excess carbohydrate on the de and HC cafeteria diets. The results indicate that an velopment of raised GIP concentrations in young increased amount of dietary fat chronically stimulates ob/ob mice fed for 8 weeks on high fat and high the production and secretion of GIP, and enhances intestinal K cell density in ob/ob mice. carbohydrate cafeteria diets. Obese hyperglycaemic (ob/ob) mice exhibit a par ticularly severe obesity and hyperinsulinaemia, Materials and Methods Animals marked hyperphagia and moderate hyperglycae mia (Bray & York 1979; Bailey et al. 1982). Recent Obese hyperglycaemic (ob/ob) mice and lean (-(-/+) mice on the Aston background were housed as previ studies have identified excessive concentrations of ously (Flatt et al. 1984a). The origin and characteristics immunoreactive gastric inhibitory polypeptide of these mice have been described elsewhere (Flatt & (GIP) in the small intestine and plasma of these Bailey 1981; Bailey et al. 1982). mice (Flatt et al. 1983, 1984a) associated with hyperplasia and increased hormone content of Diets the intestinal GIP-secreting K cells (Polak et al. Groups of mice were fed either stock diet (Mouse 1975). Considerable evidence has implicated GIP breeding diet, Heygate & Sons Ltd., Northampton, UK) 224 or approximately isoenergetic high fat (HF) and high mouse plasma. Porcine GIP (J. C. Brown, University of carbohydrate (HC) cafeteria diets. The diets and tap British Columbia, Canada) was used to prepare *^®1- water were available ad libitum. The metabolisable labelled tracer and as standard. The GIP antiserum energy contents of the diets were; stock diet 12.2 MJ/kg (RIG 34/111, Guildhay Antisera), raised in rabbit against (26% protein, 8 % fat, 6 6 % carbohydrate); HC cafeteria a porcine GIP-glutaraldehyde-ovalbumin conjugate, diet 13.2 MJ/kg (10% protein, 7% fat, 83% carbohy recognises the 5000 and 8000 molecular forms of GIP, drate). The HF and HC cafeteria diets were supplied in and exhibits negligible cross-reactivity with other en- the form of 2 food items daily. Five pairs of food items teropancreatic hormones. Details of the assay sensitivity were used in rotation in each cafeteria diet. The pairs of and specificity have been described previously (Morgan HF cafeteria foods were chocolate and fish cake, fish etal. 1978; Flatt etal. 1984a). paste and corned beef, sardines and pork pie, marzipan and roast beef, beefburger and sausage. For the HC /mmunocytochemistry cafeteria diet the food pairings were white bread and Rehydrated paraffin sections (5 pm) were cut trans matzo, rice krispies breakfast cereal and malt bread, versely from the segments of upper jejunum. The cornflakes breakfast cereal and jelly, sugar puffs and sections were immunostained by the unlabelled per- weetabix breakfast cereals, madeira cake and rye crisp- oxidase-antiperoxidase (PAP) technique (Sternberger bread (Bailey et al. 1985). Nutrient composition and 1979) using rabbit anti-porcine GIP antiserum (RIG metabolisable energy values were calculated from food 34/111) described above. Sections were incubated with tables (Paul 8c Southgate 1978) or manufactures infor the antiserum (1:1500 dilution) for 16 h at 4°C, fol mation. lowed by incubations with donkey anti-rabbit antiserum (Guildhay Antisera) for 40 min at 4°C and with rabbit Experimental procedure PAP complex (Dakopatts, Glostrup, Denmark) for 40 Three groups of ob/ob mice were matched for sex and min at 4°C. Peroxidase activity was visualised using body weight immediately prior to weaning (3 weeks of 0.05% 3,3-diaminobenzidine (British Drug Houses, age), and weaned onto either the stock diet, HF or HC Poole, UK) in citrate-acetate buffer, pH 5, and sections cafeteria diet. A group of lean mice was also weaned were counterstained with 0.5% Harris’ haematoxylin onto the stock diet. The mice were maintained on their (British Drug Houses, Poole, UK). Control sections respective diets until 11 weeks of age, when 24 h food were treated with normal rabbit serum instead of rabbit intake was monitored (Bailey et al. 1985) for 4 consecu anti-porcine GIP antiserum. Sections were washed twice tive days. Body weight was recorded, and a blood for 10 min with saline in 0.05 M Tris buffer, pH 7.6, sample (150 pi) was taken from the tail tip for determi between each incubation. Thirty-five representative nation of plasma glucose, insulin and GIF. sections were selected from throughout each tissue Mice were killed by cervical dislocation at 12 weeks segment. Positively stained GIP cells (K cells) were of age. The intestine (duodenum-jejunum-ileum) was counted per whole transverse section at x 250 magnifi rapidly removed, cleaned by perfusion with ice cold cation. Sections were focused up and down to ensure saline, weighed and the length measured. One cm inclusion of all cells throughout the thickness of the segment was taken exactly 15 cm distal to the duodenal- sections. Only sections with villi at right angles to the pyloric junction (corresponding to upper jejunum), lumen were used, and only cells with a surface area fixed for 24 h in Bouin's fluid, dehydrated through equal to or greater than a nucleus were counted. The graded ethanols, cleared in toluene and embedded in area of each section was determined using a grid, and paraffin wax. The remaining tissue was re-weighed, the number of K cells was expressed per mm^, to extracted with 5 ml/g acid ethanol (Flatt et al. 1983) and accommodate variations in area. assayed for GIP. Statistical analysis Assay Data were compared using Student’s unpaired ( test. Plasma glucose was measured by an automated glucose Differences were considered to be significant for oxidase procedure (Stevens 1971), and plasma insulin P < 0.05. was determined by double antibody radioimmunoassay using crystalline mouse insulin standard (Bailey & A hm ed-Sorour 1980). GIP was measured by double Results antibody radioimmunoassay (Morgan et al. 1978) using At 11 weeks of age ob/ob mice fed stock diet were donkey anti-rabbit gamma globulin antiserum (Guild- characteristically obese (as indicated by greater hay Antisera, University of Surrey, Guildford, UK) to separate bound and free antigen. Immuno-adsorbed body weight), hyperphagic, hyperglycaemic, hy- hormone-free plasma was used to minimise non-speci perinsulinaemic and exhibited raised plasma GIP fic interference, and parallelism was demonstrated be concentrations compared with lean mice fed the tween the standard curve and serially diluted ob/ob same diet (Fig. 1). The ob/ob mice fed a HF 15 Acta endocr. 112. 2 225 Body weight Energy intake Plasma glucose Plasma insulin Plasma GIP (g) (kJ/mouse/day) (mmol/l) (ng/ml) (ng/ml) 100 20 b e 150 -X. 30 1 , 1 a • • a e ac -UjI- _i_ 100 _L e 4" e rL 50 1 e # 10 20 # e — 1— e 50 - e e e e # e e r I I lean stock diet i m obese stock diet 0 obese HF cafeteria diet Q obese HC cafeteria diet fig./. Body weight, energy intake, and plasma concentrations of glucose, insulin and gastric inhibitory polypeptide (GIP) in lean (+/+) mice fed stock diet, and in obese (ob/ob) mice fed either stock diet, a high fat (HF) cafeteria diet or a high carbohydrate (HC) cafeteria diet. Values are mean ± s e m o f 6 mice. Energy intake values were calculated during the last 4 days of the study. All parameters measured were significantly (P < 0.05) higher in obese mice than in lean mice; a: P < 0.05 compared with obese HF cafeteria diet; b: P < 0.05 compared with obese HC cafeteria diet; c: P < 0.05 compared with obese stock diet. cafeteria diet, but not those fed a HC cafeteria mice, and the HF cafeteria diet increased the diet, showed a greater body weight than ob/ob intestinal GIP concentration (118%) and content mice fed a stock diet. However, energy intake (71%) compared with stock fed ob/ob mice. The during the last 4 days of the study was similar in intestinal GIP concentration and content were not ob/ob mice fed the stock diet and the HF cafeteria significantly altered by the HC cafeteria diet, diet, and slightly lower in ob/ob mice fed the HC although the mean values were higher (45% and cafeteria diet. The extent of hyperglycaemia and 25% respectively) than in stock fed ob/ob mice. hyperinsulinaemia was not significantly different Quantitative evaluation of histological sections in the three groups of ob/ob mice, but plasma GIP of upper jejunum confirmed that the density of K concentrations were considerably raised (75%) in cells in this region of intestine is similar in lean the group of ob/ob mice receiving the HF cafete and ob/ob mice fed a stock diet (Fig. 3). However, ria diet. the density of K cells was increased in ob/ob mice Intestinal length and weight were characteristi receiving the HF (33% and 54%) and HC (25% cally greater in stock fed ob/ob mice than in lean and 45%) cafeteria diets compared with the lean mice (Fig. 2). The HF and HC cafeteria diets did and ob/ob mice fed the stock diet. The distribu not significantly alter intestinal length in ob/ob tion of K cells was similar in all sections, being mice, but intestinal weight was lower in ob/ob mainly at the neck of the crypts and near the base mice receiving the HF cafeteria diet. The intesti of the villi, with very few cells in the apical region nal concentration and content of GIP were about of the villi. There were no apparent differences in 200% greater in stock fed ob/ob mice than in lean the intensity of staining. 226 Intestinal Intestinal Intestinal GIF Intestinal GIF length weight concentration content (cm) (g) (ug/g) (ug) 80 ■ 4 • 8 ■ 4^bc 4FO 1 60 - _ 1 3 • 6 - JJ - • • • 4- • _L 40 • 2 • • 4 - • y^ • 4- à # • 4 20 e y^i- 1 • 2 e y!t • y^ • e y^ 0 n I I lean stock diet i m obese stock diet @ obese HF cafeteria diet Q obese HC cafeteria diet Fig. 2. Intestinal length and weight, and intestinal concentration and content of gastric inhibitory polypeptide (GIP) in lean (+/+) mice fed stock diet, and in obese (ob/ob) mice fed either stock diet, a high fat (HF) cafeteria diet or a high carbohydrate (HC) cafeteria diet. Values are mean ± s e m of 6 mice. < 0.05 compared with lean mice fed stock diet, a: P < 0.05 compared with obese HF cafeteria diet; b: P < 0.05 compared with obese HC cafeteria diet; c: P < 0.05 compared with obese stock diet. K cell density of upper jejunum (cells/mm^) Discussion 4Fc 10 Obese (ob/ob) mice readily consumed the varied JL. I I lean stock diet yk- and palatable cafeteria diets, showing a greater 4- m obese stock diet preference for the HF than the HC cafeteria 4- foods (Bailey et al. 1985). Although energy intake Q obese HF cafeteria diet from the HF cafeteria diet and the stock diet was similar, body weight was greater in mice receiving obese HC cafeteria diet the former diet, supporting previous evidence that metabolic efficiency is increased by this diet (Bailey et al. 1985). Whereas the extent of basal Fig. 3. hyperglycaemia and hyperinsulinaemia in ob/ob Density of K cells in the upper jejunum of lean (+/+) mice was not significantly changed by the cafete mice fed stock diet, and of obese (ob/ob) mice fed either ria diets, plasma GIP concentrations were raised stock diet, a high fat (HF) cafeteria diet or a high by the HF cafeteria diet. This cannot be attributed carbohydrate (HC) cafeteria diet. Values are mean to hyperalimentation since ob/ob mice consumed ± SEM of 35 transverse sections from each of 6 mice. ■¥■ P < 0.05 compared with lean mice fed stock diet, similar amounts of the HF and stock diets, thus a: P < 0.05 compared with obese HF cafeteria diet; substantiating acute studies showing a potent b; P < 0.05 compared with obese HC cafeteria diet; GIP-releasing effect of fatty acids (Brown 1982; c: P < 0.05 compared with obese stock diet. Flatt et al. 1984a). A sustained insulinotropic 227 effect of GIP in ob/ob mice appears to be depend Acknowledgments ent upon a rise in the extent of hyperglycaemia The authors gratefully acknowledge the excellent ani (Flatt et al. 1984a), which may explain why the mal care provided by Melvin Gamble and Wayne raised GIP concentrations in the HF cafeteria fed Fleary. ob/ob mice were not associated with a further increase in basal hyperinsulinaemia. Consistent with previous observations, the stock fed ob/ob mice exhibited an increased intestinal References concentration and content of GIP compared with Bailey C J & Ahmed-Sorour H (1980): Role of ovarian lean mice (Flatt et al. 1983). This was associated hormones in the long-term control of glucose homeo with a similar density of K cells in the upper stasis: effects on insulin secretion. Diabetologia 19: jejunum, although increased numbers of K cells 4 7 5 -4 8 1 . have been observed in more distal regions of the Bailey C J, Flatt P R & Atkins T W (1982): Influence of small intestine in older ob/ob mice (Polak et al. genetic background and age on the expression of the 1975). Feeding a HF cafeteria diet to ob/ob mice obese hyperglycaemic syndrome in Aston ob/ob mice. Int J Obesity 6: 11-21. produced jejunal K cell hyperplasia with a further Bailey C J, Flatt P R & Radley N S (1985): Effect of high increase of the intestinal GIP concentration and fat and high carbohydrate cafeteria diets on the content. Thus, chronic stimulation of GIP release development of the obese hyperglycemic (ob/ob) syn by increased consumption of fats appears to pro drom e in mice. N utr Res 5: 1003 — 1010. mote K cell proliferation and function. K cell Bray G A & York D (1979): Hypothalamic and genetic hyperplasia of the upper jejunum was also ob obesity in experimental animals: an autonomic and served in ob/ob mice fed the HC cafeteria diet. A endocrine hypothesis. Physiol Rev 59: 719—809. concomitant increase of the intestinal GIP concen Brown J C (1982): Gastric inhibitory polypeptide. tration and content was not observed, suggesting Monographs on Endocrinology. Vol 24. Springerr that the GIP content of K cells in HC cafeteria Verlag, Berlin. C reutzfeldt W, Ebert R, Nauck M 8c Stockman F (1983): fed ob/ob mice may be less than in HF cafeteria Disturbances of the entero-insular axis. Scand J fed ob/ob mice. This coincides with the relative Gastroenterol. Suppl 82, 18: 111-119. potencies of carbohydrate and fat for the stimula Flatt P R & Bailey C J (1981): Abnormal plasma glucose tion of GIP release (Flatt et al. 1984a), but differ and insulin responses in heterozygous lean (ob/4-) ences in the intensity of K cell staining in mice fed mice. Diabetologia 20: 573—577. the cafeteria diets were not apparent with the Flatt P R, Bailey C J & Buchanan K D (1982): Regula present technique. The lack of a significant effect tion of plasma immunoreactive glucagon in obese of the HC cafeteria diet (83% metabolisable hyperglycaemic (ob/ob) mice. J Endocrinol 95: 215 — energy from carbohydrate) on plasma and inte 227. stinal GIP compared with stock fed ob/ob mice Flatt P R, Bailey C J, Kwasowski P, Page T 8c Marks V (1984a): Plasma immunoreactive gastric inhibitory may reflect the already high carbohydrate com polypeptide in obese hyperglycaemic (ob/ob) mice. J ponent of the stock diet (66% metabolisable Endocrinol 101: 249-256. energy from carbohydrate). Flatt P R, Bailey C J, Kwasowski P, Swanston-Flatt S K & It has been suggested that GIP increases glu Marks V (1983): Abnormalities of GIP in spontane cagon secretion in states of glucose intolerance ous syndromes of obesity and diabetes in mice. Dia (Salera et al. 1982). Thus GIP might contribute to betes 32: 433-435. the hyperglycaemia in ob/ob mice by maintaining Flatt P R, Bailey C J, Swanston-Flatt S K, Best L, inappropriately raised plasma glucagon concen Kwasowski P, Buchanan K D & Marks V (1984b): trations (Flatt et al. 1982). GIP has also been Involvement of glucagon and GIP in the metabolic reported to stimulate lipogenesis in adipose tissue abnormalities of obese hyperglycemic (ob/ob) mice. In: Shafrir E 8c Renold A E (eds). Lessons from (Brown 1982). Hence, in combination with the Animal Diabetes, pp 34 1—347. John Libbey, London. hyperinsulinaemia, raised GIP levels might pro Morgan L M, Morris B A & Marks V (1978): Radio mote obesity in ob/ob mice. The present study immunoassay of gastric inhibitory polypeptide. Ann indicates that in addition to enhanced GIP secre Clin Biochem 15: 172-177. tion, dietary fat is also an important determinant Paul A A & Southgate D A T (1978): The composition of K cell number and intestinal GIP content in of Foods, 4th edn. Her Majesty’s Stationery Office. ob/ob mice. London. 2 2 8 Polak J M, Pearse A G E, Grimelius L & Marks V ( 1975): Sternberger L A (1979): Itnmunocytochemistry, 2nd Gastrointestinal apudosis in obese hyperglycaemic edn. Wiley & Sons, New York. mice. Virchows Arch B Cell Pathol 19: 135-150. Stevens J F (197 I): Determination of glucose by an Salera M, Giacomoni P, Pironi L, Cornia G, Capelli M, automatic analyser. Clin Chim Acta 32: 199-201. Marini A, Benfenati F, Miglioli M & Barbara L (1982): Gastric inhibitory polypeptide release after oral glucose: relationship to glucose intolerance, dia------betes mellitus, and obesity. J Clin Endocrinol Metab Received September 16th, 1985. 55: 329—336. A cceptedjannuary 13th, 1986. 229 SIX: HORMONAL PATTERNS (4) Involvement of glucagon and GIP in the metabolic abnormalities of obese hyperglycemic (ob/ob) mice Peter R. F L A T T S Clifford J. BAILEY^, Sara K. SW A N S T O N - F L A T T * , Leonard BE S T \ Peter KW A S O W S K I * , Keith D. B U C H A N A N \ and Vincent MA R K S * ^Department of Biochemistry, Divisions of Nutrition & Food Science and Clinical Biochemistry, University of Surrey, Guildford, UK; ^Department of Biological Sciences, University o f Aston in Birmingham, UK; ^Laboratory of Experimental Medicine, University o f Brussels Medical School, Belgium and * Department of Medicine, Queen's University o f Belfast, UK. Abnormalities of glucagon and GIP in genetically obese hyperglycemic(ob/ob) mice are described. Evi dence is presented that glucagon contributes to the hyperglycemia and that GIP contributes to the hyper- insulinemia of the obese mutant. A scheme is proposed in which unchecked food consumption leads to hyperplasia and hyperactivity of GIP secreting cells and other components of the enteroinsular axis, result ing in hyperinsulinemia. The hyperphagia and hyperinsulinemia in turn promote excessive fat deposition with the production of insulin resistance, pancreatic A-cell dysfunction and hyperglycemia. Introduction The obese hyperglycemic syndrome (gene symbol ob) arose in 1949 as a mutation within the V strain of mice at the Jackson Laboratory, Bar Harbor, M E , USA*. The syndrome is transmitted as a single autosomal recessive gene (chromosome 6 , linkage group XI), and obese mice have been adopted widely as an animal model to study the etiology and patho genesis of obesity-diabetes^’^. These mice exhibit hyperphagia, hyperinsulinemia, hypergly cemia and obesity, the severity of which depend upon interaction of the mutant gene with the background genome. W e report some recent observations concerning abnormalities of glucagon and GI P (glucose dependent insulinotropic polypeptide) in relation to the meta bolic disturbances in the Aston strain of objob mice. Experimental procedures The origin, breeding and characteristics of lean and obese mice from the Aston colony have been described elsewhere^’®. Blood for determination of plasma glucose®, insulin^, and C-terminal immunoreactive gluca gon (GGLI)^ was obtained from the cut tip of the tail of conscious mice. Plasma GIP was determined using blood obtained by cardiac puncture from anesthetised mice®. GIP and N-terminal immunoreactive glucagon (N —GLI)® were also measured in duodenaljejunal-ileal extracts prepared with acid ethanol***. The insulin releasing activity of intestinal extracts was measured using perihised obese mouse islets. The same proce dure was used to investigate the insulinotropic action of factors released by nutrient stimulation of obese mouse jejunum. The jejunum was everted and placed in a perifusion system in series with isolated islets. The mucosal surface was bathed in medium containing glucose, phenylalanine, methionine and valine at concen trations of 18.0, 7.0, 5.5 and 12.8 mmol/l respectively. The final glucose concentration of the perifusion medium was less than 4 mmol/l in all experiments. Values presented in the figures are means ± s.e. of groups of 5-14 mice. Data were compared using Student’s f-test. Results and comment Inappropriate hyperglucagonemia. Glucagon is an important regulatory hormone, recog nized for its stimulatory action on hepatic glycogenolysis, gluconeogenesis and insulin secre tion**. Thus, inappropriately raised glucagon concentrations m a y contribute to hyperglyce mia in conditions of insulin deficiency and/or resistance*^. Indeed, hyperglucagonemia has been noted in hu m a n diabetes mellitus** and spontaneously diabetic rodents, including fatty (fa/fa) rats*®, K K mice*\ db /d b mice*®, B B rats*® and djungarian hamsters*’. As shown in Fig. 1, Aston obese mice exhibit hyperglucagonemia relative to the prevailing glucose and insu 341 Fig. 1. A: Plasma glucose (mmol/l); B: Plasma insulin (ng/ml); C: plasma C- terminal immunoreactive glucagon (C- GLI [pg/ml]) in fed and 24-h fasted lean (open columns) and obese (hatched col umns) mice between 5 and 40 wk of age. j r Æ *P i L Æ ± J ^ \ m \ m Fed F a s te d lin concentrations, as noted with this mutation on other genetic backgrounds*®"*®. Plasma glucagon concentrations were paradoxically raised in these mice irrespective of age, nutri tional status and prominent age-related changes in the severity of the hyperglycemia and hyperinsulinemia. Abnormal regulation o f glucagon secretion. The co-existence of hyperglycemia, hyperinsu linemia and hyperglucagonemia in obese mice suggests a defect in a glucose - or insulin- mediated mechanism to suppress glucagon secretion. Plasma glucagon responses to i.p. glucose load in lean and obese mice are shown in Fig. 2. In the fed state, glucose suppressed glucagon in both genotypes, although glucose tolerance was impaired and a positive insulin response was lacking in obese mice. After a 24 h fast, i.p. glucose paradoxically increased glucagon in obese mice. This effect was associated with the restoration of a small positive insulin response, but glucose tolerance was further impaired. The action of glucose to sup press glucagon in fed but not fasted obese mice suggests a dependence on the marked hyper insulinemia associated with feeding. Moreover, heparin-induced elevation of free fatty acids reduces glucagon in fasted obese mice, thereby precluding a general defect to suppression by metabolic substrates’. The ability of exogenous insulin to suppress glucagon was examined in fasted mice (Fig. 3). The insulin dose effective in lean mice (0.25U/kg) was not effective in the obese mutant. A 400-fold greater dose of insulin (100 U/kg) was required to suppress both glucagon and glucose in obese mice. This demonstrates that the pancreatic A-cells of obese mice are in sensitive to the normal inhibitory action of insulin. The abnormality could account, in part, for both the impaired hypoglycemic action of insulin and the exaggerated glucagon response to arginine in these mice’'*’’**'**. Receptor and postreceptor faults have been identified in other insulin sensitive tissues of obese mice, including muscle and adipose tissue*®. The nature of the lesion responsible for impaired insulin suppression of glucagon secretion is un determined. According to current theory’’*^'*®, the defect m a y involve reduced A-cell glucose metabolism mediated through alterations of insulin receptor binding and postrecep tor events such as adenylate cyclase activity. Significance o f circulating glucagon. The relative hyperglucagonemia of obese mice, together with severe insulin resistance*® and a marginal reduction in glucagon-receptor binding*’, implicate glucagon in the hyperglycemia and hyperinsulinemia. Consonant with this view. 342 Fig. 2 Fig. 3 Plasma glucose «mol/I Plasma insu I in ng/ml 4 o- Plasma C-GLl pg/ml Plasma C-CLI pg.m! 150 i 100 Time (nln) Fig. 2 (above, left). Plasma glucose, insulin and C-GLI responses to up. glucose (2 g/kg) in 20-wk-old fed and 24-h fasted lean\(o — o) and obese (* — mice. Fig. 3 (above, right).Plasma glucose and C-GLI responses to i.p. porcine insulin (Actrapid^, 0.25 and 100 U/kg) in 24-h fasted 20-xuk-old lean and obese mice, (o—o, lean 0.25 U/kg; • — • obese, 0,25 U/kg; ■— "obese 100 U/kg). exogenous glucagon markedly increased plasma glucose and insulin concentrations in fed obese mice (Fig, 4), T he physiological significance of these findings was directly evaluated by injection of specific C-terminal glucagon antibodies’ in a quantity sufficient to neutralize > 1,5 times the total pancreatic glucagon content^®. As shown in Fig, 5, induction of gluca gon deficiency normalized the glucose concentration of fed obese mice, with only a small decrease in plasma insulin. These observations indicate that glucagon makes a substantial contribution to the hyperglycemia of obese mice, but factors additional to glucagon are un doubtedly of greater importance in the hyperinsulinemia. Abnormalities of gut GLI. The existence of extrapancreatic glucagon-like immunoreactivity (GLI) is well established^’, but the physiological significance of this heterogeneous group of peptides remains to be clarified, G L I is located predominantly in the gut and is sometimes termed enteroglucagon. Postulated physiological actions of GL I include stimulation of insu- ■0 g Time (nin) Time (min) *•0 Fig, 4 (above, left). Plasma glucose and insulin responses to i.p. glucagon (1 mg/kg) in 10-wk-old fed lean (O— O) and obese (•— •) mice. Fig, 5 (above, right).Plasma glucose and insulin responses to an up. injection of specific C-terminal glu cagon antiserum in 20-wk-old fed lean (open columns) and obese (hatched columns) mice. 343 lin secretion and gut mucosal growth^°’^‘. In view of the hyperinsulinemia and enlarged intestine of obese mice®, N GL I was determined in duodenal-jejunal-ileal extracts after separation of multiple molecular forms by gel chromatography®*. Consistent with a possible pathological effect, the yield of dried extract from obese mouse intestine was 18 times greater and contained an increased amount of 3500-12 000 mol wt N-GLI per unit of extract (Fig. 6 ). However, the obese mouse intestine was apparently devoid of >2 0 00 0 mol wt N-GLI. This could reflect rapid cleavage into smaller N-GLI components. Since there is a link between glucagon and feeding®®, it is noteworthy that parabiosis experiments have seeded the hypothesis that hyperphagia in obese mice results from absence of a putative satiety factoi^^. Importance of the enteroinsular axis. A case for an important role of the gut in the hyper insulinemia of obese mice does not rest merely with the possible insulinotropic influence of increased amounts of small to m e d i u m mol wt N-GLI. Compelling evidence that the entero insular axis is involved in the hyperinsulinemia has accumulated in recent studies®’*®’®®"®*. For example, age changes in hyperinsulinemia show a similar pattern to that of the hyper phagia, and plasma insulin concentrations decrease markedly w h e n food is withheld. In addition, parenterally administered glucose fails to produce a positive plasma insulin response in the fed state, whereas orally administered glucose evokes a pronounced insulin response. Voluntary ingestion of protein, fat and especially carbohydrate also evokes a marked eleva tion of insulin, and conditioning to intensify neuroendocrine pathways associated with feed ing augments the insulin response to food ingestion. Furthermore, there is hyperplasia of enteroendocrine cells in obese mouse intestine®®’®’. Indeed, both an extract of obese mouse intestine and intestinal factors, released by a glucose-amino acid mixture, stimulate insulin release from perifused islets of fed obese mice (Fig 7). Obese mouse intestine Lean mouse intestine M3 H4 M3 M4 N-GLI N-GLI 20 10 C 45 Elution volume (ml) Elution volume (ml) Time (min) Fig. 6 (above, left).N-terminal glucagon-like immunoreactivity (N-GLI) in acid ethanol extracts o f intestine of 10-wk-old fed lean and obese mice after separation o f multimolecular forms by gel filtration on a 45 x 1 cm Sephadex G-50 column. 50 mg of each extract was applied to the column and eluted at 8.5 ml/h with 0.1 mol/1 acetic acid. Elution positions of molecular markers are: Ml, blue dextran; M2, cytochrome C; M3, I-insulin; M4, ‘*® I-glucagon. Fig. 7 (above, right). Insulin release (ng/mllmg) from perifused islets: (a) islets of fed lean (0.—0) and obese — *) mice exposed to 3 mmol/l glucose and 100 pg/ml dried extract of obese mouse duodenum- jejunum; (b) islets of fed obese mice exposed to 3 mmol/l glucose and factors released from obese mouse jejunum by a glucose-animo acid mixture{•— «Test: everted jejunal segment exposed to glucose- amino-acid stimulation on mucoasl surface and perifused in series with islets; o o Control: islets perifused without jejunal segment. 344 Abnorm alities o f GIP. Although several gastrointestinal hormones have been reported to possess insulinotropic activity^®, only GIP (glucose-dependent insulinotropic polypeptide) satisfies the necessary criteria to be ascribed a physiological role in the enteroinsular axis^*. As shown in Fig. 8 , the concentration and total intestinal content of GIP were raised in fed obese mice. However, most notable was the 15-fold elevation of plasma GIP concentrations, an increment similar in magnitude to the degree of hyperinsulinemia®. GIP and glucagon stimulate insulin secretion b y a c o m m o n cyclic-AMP-mediated mechanism^*’^®. Since glucagon is a potent insulin secretagogue in obese mice (Fig. 4)**, pathologically raised GIP concentrations are likely to m a k e a major contribution to the hyperinsulinemia, especially in the prevailing environment of hyperglycemia. Hypersecretion of GIP might be related to hyperalimentation or enteroendocrine hyperplasia. However, GIP is normally subject to feedback inhibition by insulin^®’^‘ and C-peptide^'*. Thus raised GIP concentrations could well result from insulin insensitivity of GIP secreting K-cells^' analagous to the situation observed with pancreatic A-cells. Moreover, C-peptide formation and/or action m a y be defective in obese mice. Hypothalamus Intestinal Intestinal Plasma # \ GIP GIP GIP Appetite Satiety concentration content concentration M9/g H9 ng/ml Hyperphagia Enteroinsularjj^^^ntestlne Aris Insulin -Insulin Hyperglycemia Hypersecretion Resistance ^ Hyperglucagonemi Adipose Tissue i O b esity OL oL Fig. 8 (above, left). Glucose-dependent insulinotropic polypeptide (GIP) in acid ethanol extracts of intest ine and in plasma of 17-26-wk-old fed lean (open columns) and obese (hatched columns) mice. Fig. 9 (above, right).Proposed involvement of glucagon and GIP in the metabolic abnormalities of obese (ob/ob) mice. Insulin hypersecretion can result from (1) direct action of insulinotropic amino acids on the pancreatic B-cells, and (2) indirect action of nutrients (especially dietary carbohydrates) mediated by neural and hormonal components of the enteroinsular axis. GIP has been ascribed a q)ecial importance in the enteroinsular axis, and its action is likely to be potentiated in the presence of hyperglycemia. Involvement of glucagon and GIP in the metabolic disturbances of obese mice. There is accumulating evidence that glucagon and GIP contribute to the metabolic disturbances of h u m a n obesity and diabetes mellitus**’^‘. The role of these hormones in the pathogenesis of spontaneous obesity-diabetes syndromes in rodents has received comparatively little atten tion. It is evident from the foregoing sections that obese mice provide an animal model in which abnormalities of both glucagon and GDP are implicit in the metabolic derangements. Figure 9 proposes a scheme indicating the possible involvement of these hormones. It is envisaged that hyperphagia promotes hypersecretion and hyperplasia of GIP secreting cells and other hormonal and neural components of the enteroinsular axis. This provides an indirect mechanism for stimulation of B-cell hyperplasia and insulin hypersecretion which markedly augments the direct stimulatory effects of nutrients such as amino acids. A direct stimulatory effect of glucose is not important in this respect. Hyperinsulinemia in turn exacerbates insulin resistance which lifts the normal feedback restraint imposed by insulin on GIP-secreting K-cells and glucagon-secreting A-cells. T he hyperglucagonemia contributes to the hyperglycemia, which potentiates the insulinotropic action of GIP. In an environment of hyperphagia, hyperglycemia, hyperinsulinemia and mounting insulin resistance, excess calories are stored by adipose tissue thereby leading to obesity. 345 References 1 Ingalls, A.M., Dickie, M.M., and Snell, G.D.: Obese, a new mutation in the house mouse./.Hered. 41: 317-18, 1950. 2 Herb erg, L., and Coleman, D.L.: Laboratory animals exhibiting obesity and diabetic syndromes. Metabolism 26: 59-99, 1977. 3 Bray, G.A., and York, D.A.: Hypothalamic and genetic obesity in experimental animals: an autonomic and endocrine hypothesis. P/iy«'o/. Rev. 59: 719-809, 1979. 4 Flatt, P.R., and Bailey, CJ.: Abnormal plasma glucose and insulin responses in heterozygous lean (ob/+) m ice. Diabetologia 20: 573-77, 1981. 5 Bailey, CJ., Flatt, P.R. and Atkins, T.W.: Influence of genetic background and age on the expression of the obese hyperglycaemic syndrome in Aston ob/ob mice. fnt. J. Obesity 6: 11-21,1982. 6 Stevens, J.F.: Determination of glucose by an automatic analyser.Clin. Chim. Acta 32: 199-201, 1971. 7 Flatt, P R., Bailey, C.J., and Buchanan, K.D.: Regulation of plasma immunoreactive glucagon in obese hyperglycaemic (ob/ob) m ice./. Endocrinol 95: 215-27, 1982. 8 Flatt, P.R., Bailey, CJ., Kwasowski, P., Swanston-Flatt, S.K., and Marks, V.: Abnormalities of GIP in spontaneous syndromes of obesity and diabetes in mice. D iabetes 32: 433-35, 1983. 9 Flatt, P.R., and Swanston-Flatt, S.K.: Stimulation of antiglucagon antibodies in rabbits and guinea pigs using a glucagon-carbodiimide-albumin conjugate.Endocr. Experimentalis 15: 3-16, 1981. 10 Kenny, AJ.: Ex tractable glucagon of the human pancreas. /.Clin. Endocr. Metab. 15: 1089-105,1955. 11 Felig, P., Sherwin, R.S., Soman, V., Wahren, J., Hendler, R., Sacca, L., Eigler, N., Goldberg, D., and Walesky, M.: Hormonal interactions in the regulation of blood glucose.Rec. Prog. Horm. Res. 35: 501-29, 1979. 12 Unger, R.H., and Orel, L.: Glucagon and the A cell. Physiology and pathophysiology. Part 1.New Engl J. Med. 304: 1518-24; 1575-80, 1981. 13 Seino, S., Seino, V., Takemura, J., Tsuda, K., Kuzuya, H., Ishikawa, K., Shimazu, T., and Imura, H.: Somatostatin insulin and glucagon secretion from isolated perfused pancreas of obese rats.Am . J. P hysiol 241: El46-50, 1981. 14 Ohneda, A., Kobayashi, T., Nihei, J., and Nishikawa, K.: Glucagon in spontaneously diabetic KK mice. Horm. Metab. Res. 13: 207-11, 1981. 15 Herb erg, L., Berger, M., Buchanan, K.D., Gries, F.A., and Kem, H.: Tiermodelle in der Diabetesfors- chung: metabolische und hormonelle Besonderheiten. Z. Versuchstierkunde 18: 91-105, 1976. 16 Nakhooda, A.F., Like, A.A., Chappel, C.I., Murray, F.T., and Marliss, E.B.: The spontaneously diabetic Wistar rat. Metabolic and morphologic studies.Diabetes 26, 100-12, 1976. 17 Herb erg, L., Buchanan, K.D., Herbertz, L.M., Kem, H.F., and Kley, H.K.: The djungarian hamster, a laboratory animal with inapropriate hyperglycaemia. Comp. Biochem. Physiol Series A 65: 35-60, 1980. 18 Lavine, R.L., Voyles, N., Perrino, P.V., and Recant, L: The effects of fasting on tissue cyclic AMP and plasma glucagon in the obese hyperglycemic mouse.Endocrinology 97: 615-20,1975. 19 Dubuc, P.U., Mobley, P.W., Mahler, RJ., and Ensinck, J.W.: Immunoreactive glucagon levels in obese- hyperglycemic (ob/ob) mice Diabetes 26, 841-46, 1977. 20 Dunbar, J.C., and Walsh, M.F.: Glucagon and insulin secretion by islets of lean and obese (ob/ob) mice. Horm. Metab. Res. 12: 39-40, 1980. 21 Laube, H., Fussganger, R.D., and Pfeiffer, E.F.: Paradoxical glucagon release in obese hyperglycaemic mice. Horm. Metab. Res. 6: 426, 1974. 22 Flatt, P.R., and Bailey, CJ.: Plasma glucose and insulin responses to glucagon and arginine in Aston ob/ob mice: Evidence for a selective defect in glucose-mediated insulin release.Horm. Metab. Res. 14: 127-30, 1982. 23 Bailey, CJ., Lord, J.M., and Atkins, T.W.: The insulin receptor and diabetes. InRecent advances in diabetes. Nattrass, M., and Santiago,/., Eds. Edinburgh, Churchill Livingstone, pp. 27-44, 1984. 24 Itoh, M., Reach, G., Furman, B., and Gerich, J.: Secretion of glucagon. InIslets o f Langerhans. Cooper- stein, SJ., and Watkins, D., Eds, New York, Academic Press, pp. 225-55, 1981. 25 Ostenson, C-G: Regulation of glucagon release: Effects of insulin on the pancreatic A-cell of the guinea pig. Diabetologia 17: 325-30, 1979. 26 Flatt, P.R., and Bailey, CJ.: Development of glucose intolerance and impaired insulin response to glucose in obese hyperglycaemic (ob/ob)mice. Horm. Metab. Res. 13: 556-60, 1981. 27 Freychet, P.: Interactions of polypeptide hormones with cell membrane specific receptors: Studies with insulin and glucagon.Diabetologia 12: 83-100, 1976. 28 Flatt, P.R., Swanston-Flatt, S.K., and Bailey, CJ.: Glucagon antiserum: A tool to investigate the role of circulating glucagon in obese hyperglycaemic (ob/ob)mice. Biochem. Soc. Trans. 7: 911-13, 1979. 29 Perez-Castillo, A., and Blazquez, E.: Tissue distribution of glucagon, glucagon-like immunoreactivity, and insulin in the rat. Am. J. Physiol 238: E258-66, 1980. 30 Marks, V., and Turner, D.S.: The gastrointestinal hormones with particular reference to their role in the regulation o f insulin secretion. Essays Afed. Biochem. 3: 109-52, 1977. 346 31 Barrowman.J.A.: The trophic action of gastrointestinal hormones. Digest/on 12:92-104, 1975. 32 Flatt, P.R., Bailey, CJ., and Swanston-Flatt, S.K. Heterogeneity of glucagon-like immunoreactive peptides (GLI) in the intestine of obese hyperglycaemic (ob/ob) mice. Horm. Metab. Res. 33 Mayer, J.: Physiology of hunger and satiety. In Modem nutrition in health and disease, 6th Edn, Goodhart, R.S., and Shils, M.E., Eds. Philadelphia, Lea ic Febiger, 1980, pp. 560-77. 34 Coleman, D.L.: Obesity and diabetes: Two mutant genes causing diabetes-obesity syndromes in mice. Diabetologia 14: 141-8, 1978. 35 Flatt, P.R., and Bailey, CJ.: Importance of the enteroinsular axis for the insul-secretory response to glucose in obese hyperglycaemic (ob/ob) mice.Biochem. Soc. Trans. 9: 220-1,1981. 36 Flatt, P.R., and Bailey, CJ.: Role of dietary factors in the hyperinsulinaemia of genetically obese hyperglycaemic (ob/ob) mice./. Nutr. 112: 2322-36, 1982. 37 Flatt, P.R., and Bailey, CJ.: Glucose and insulin responses to conditioned feeding in lean and gene tically obese hyperglycaemic (ob/ob) mice Metabolism 32: 504-09,1983. 38 Polak, J.M., Pearse, A.G.E., Grimelius, L., and Marks, V.: Gastrointestinal apudosis in obese hypergly caemic mice. Virchows Arch B Cell Path. 19: 135-50, 1975. 39 Best, L., Atkins, T.W., Bailey, CJ., Flatt, P.R., Newton, D.F., and Matty, AJ.: Increased activity of the enteroinsular axis in obese hyperglycaemic (ob/ob) m ice./. Endocr. 72: 44P, 1977. 40 Creutzfeldt, W.: The in cretin concept today. Diabetologia 16: 75-85, 1979. 41 Brown, J.C.: Gastric inhibitory polypeptide. Berlin, Springer Verlag, p.88, 1982. 42 Szecowka, J., Grill, V., Sandberg, E., and Efendic, S.: Effect of GIP on the secretion of insulin and somatostatin and the accumulation of cyclic AMP in vitro in the rat. Acta Endocrinol 99: 416-21, 1982. 43 Sharp, G.W.G.: The adenylate cyclase-cyclic AMP system in islets o f Langerhans and its role in the control of insulin release. Diabetologia 16: 287-96, 1979. 44 Dryburgh, J.R., Hampton, S.M., and Marks, V: Endocrine pancreatic control of the release of gastric inhibitory polypeptide: a possible physiological role for C-peptide. Diabetologia 19: 397-401, 1980. 45 Flatt, P.R., Bailey, CJ., Kwasowski, P., Page, T., and Marks, V.: Plasma immunoreative gastric inhibitory polypeptide in obese hyperglycaemic (ob/ob) m ice./. Endocrinol. 101: 249-256, 1984. 347 à . SHORT COMMUNICATIONS Effect of Diet Upon Response to Oral Fat and Glucose in Man; Modification in Control of the Enteroinsular Axis L. M. MORGAN, J. A. TREDGER, S. M. HAMPTON, P. KWASOWSKI, J. WRIGHT, M. DUNN & V. MARKS Department of Biochemistry (Divisions of Clinical Biochemistry and Nutrition and Food Science), University of Surrey and St. Luke’s Hospital, Guildford, Surrey, U.K. The GI hormone Gastric Inhibitory Polypeptide their fat intake to approximately 150 g/day by (GIP), whose secretion is stimulated in response supplementing the low fat diet with 250 ml double to oral carbohydrate and to fat, augments cream/day. Each subject underwent two tests— glucose-induced insulin secretion (1) and is cur one on day 7 and one on day 9 of each modified rently considered to be an important endocrine diet period. The tests were: (a) 100 g oral fat (b) component of the entero-insular axis. Exogenous 100 g oral fat -f- insulin iv 0.2 U/kg body weight insulin (1) and C-peptide (2) inhibit fat-stimulated (c) 75 g oral glucose. Venous blood samples were GIP release and may play a role in the regulation collected for 3& h following each test and assayed of GIP secretion. Dietary modification can effect for plasma immunoreactive GIP (5), IR-insulin, G IP secretion. In man, a previously high sucrose IR-C-peptide, glucose and triglycerides. Volun diet increases the GIP response to oral sucrose teers did not show any significant weight change (3). Feeding rats a high fat diet for 4 days increases during the study period. Insulin significantly GIP secretion in response to oral fat, abolishes inhibited GIP secretion when subjects were eating the feedback inhibition of exogenous insulin on a low fat diet [Incremental area under the post fat-stimulated GIP release and increases insulin prandial plasma GIP curve 3979 ± 547 pg ml"' h resistance (4). W e have, therefore, investigated (x ± SE M ) after fat + insulin (n = 8) versus the plasma GIP response to glucose and to fat 5490 ± 599 pg ml"'h (n = 4) after fat alone p < and the feedback control of exogenous insulin on 0.01]. This did not occur when subjects ate the fat-stimulated GIP release in human subjects high fat diet (Incremental area under the post maintained on low and high fat dietary regimens. prandial plasma GIP curve 5652 ± 962 pg ml"' h Nine healthy volunteers (normal dietary fat (n = 8) versus 4503 ± 654 pg ml"' h (n = 6), intake <130g/day), participated in the study. p = n.s.) (Fig. 1). Plasma triglycerides were sig They ate a low fat diet (LED) providing approxi nificantly higher following oral fat, whilst the mately 30 g fa^day for 9 days. They resumed their subjects were on the high fat diet (Incremental usual diet for 15 days and then consumed a high area under the curve plasma triglycerides fat diet (HFD) for a further 9 days, increasing 1.27 ± 0.09 mmol/l"' h (n = 6) than whilst on the 100 Short Communications 3000 2000-°;S'og‘ 1000 a FAT ALONE Cl 3000 IV insulin 0 2 U/kg oral fat Û. 2000 lOOg 1000 FAT + INSULIN 0 60 120 180 Time (min) Fig. 1. Plasma GIF levels in healthy volunteers on either a high (# ------# ) or low ------(O O) fat diet following 100 g oral fat with or without iv insulin. Mean ± SEM, * p < 0.Ô5, ** p < 0.025. low fat diet ^0.38 ± 0.12 mmol/r* h (n = 4), ing triglyceride levels following oral fat and (iii) p < 0.005) (Fig. 2). Plasma glucose fell sharply causes faster recovery of plasma glucose levels after iv insulin on both dietary regimens. H o w following hypoglycaemia produced by iv insulin. ever, the subjects recovered significantly faster It did not affect the metabolic response to oral when they were maintained on the high fat diet glucose of any of the parameters measured. This (plasma glucose 150 min. post insulin 3.9 ± impairment in feedback control of GIF could lead 0.23 mmol/l on the high fat diet (n = 8) versus to inappropriately high GIF secretion and an 3.2 ± 0.26 mmol/l on low fat diet (n = 8), p < increase in the contribution of the enteroinsular 0.025). There were no significant differences axis to insulin secretion. A high fat diet has been between the two diets in plasma glucose, insulin, associated with glucose intolerance and hyper C-peptide or GIF following oral glucose. insulinaemia in experimental animals (6). It is W e conclude that a high fat diet for 9 days (i) possible that an early event in the development abolishes negative feedback control of insulin on of this type of insulin resistance is a change in the fat-stimulated GIF release (ii) increases circulat secretion of one or more hormones of the Short Communications 101 T 2 0 o E 1-6 3 ID zz 0) 1*2 J. O) (0 E 0 8 (0 jg a 75 ^ 0*4 o> E Q) k. Q B 0 FAT ALONE FAT + INSULIN Fig. 2. Incremental plasma triglyceride levels in 9 healthy volunteers on either a high ( #------# ) or low (O O) fat diet following 100 g oral fat with or without iv insulin (0.2 U/kg body weight). (Mean ± SEM), * p < 0.05, ** p < 0.005. enteroinsular axis component of insulin secre 3. Reiser S, Michaelis OE, Cataland S, O’Dorisio TM, tion, leading to excessive insulin production. (1980) Am J Clin Nutr 33, 1907-1911 4. Hampton SM, Kwasowski P, Tan K, Morgan LM, Marks V. (1983) Diabetologia 24, 278-281 REFERENCES 5. Morgan LM, Morris BA , Marks V. (1978) Ann Clin 1. Brown JC, Dryburgh JR, Ross SA, Dupre J. (1975) Biochem 15, 172-177 Recent Prog Horm Res 31, 487-532 6. Grundleger ML, Thenen SW. (1982) Diabetes 31, 2. Dryburgh JR, Hampton SM, Marks V. (1980) Dia 232-237 betologia 19, 397-401 110 Short Communications Hypergipaemia in Genetically Obese Hyperglycaemic {ob/ob) Mice: Relationship to Hyperphagia and Intestinal K-Cell Insulin Insensitivity P. R. FLATT*, C. J. BAILEYt, P. KWASOWSKI*, T. PAGEt, SUSAN M. GEE*, SARA K. SWANSTON-FLATT* & V. MARKS* * Department of Biochemistry, Divisions of Nutrition & Food Science, and Clinical Biochemistry, University of Surrey, Guildford, Surrey, GU2 5XH, and t Department of Biological Sciences, University of Aston in Birmingham, B4 7ET, UK The ob/ob syndrome in mice is characterized by Obese {ob/ob) mice derived from the Aston obesity, hyperphagia, insulin resistance, hyper colony (1-4) were used at 12-16 weeks of age. glycaemia, hyperinsulinaemia and pancreatic A- Mice were maintained for 5 weeks on a standard and B-cell dysfunction (1-4). The obese mutant pellet diet provided either (a) ad libitum for 24 has also recently been shown to exhibit markedly h/day; (b) ad libitum for 24 h/day with a varied raised GIP concentrations (5); an abnormality daily selection of cafeteria foods (7), or (c) ad proposed to contribute to the hyperinsulinaemia libitum but for a restricted period of as little as and related metabolic disturbances through the 2 h/day (8). Mice fed (a) and prefasted for 18 h stimulation of insulin secretion and the promotion were used to assess the effects of exogenous of lipogenesis in white adipose tissue. The factors insulin on basal, fat-stimulated and glucose- responsible for the hypergipaemia remain to be stimulated GIP concentrations. Blood for plasma established, although hyperalimentation can be GIP determination (5) was obtained from the cut envisaged to make an important contribution. To tail-tip. evaluate this issue, we have examined the effects Obese {ob/ob) mice characteristically exhibited of dietary restriction and cafeteria feeding on GIP hyperphagia and overweight (Table I). Imple concentrations of ob/ob mice. Furthermore, since mentation of time-dependent restricted feeding GIP-secreting K-cells are normally subject to produced the anticipated changes in energy intake feedback inhibition by insulin (6), the effects of and body weight (8). However, hyperphagia exogenous insulin were investigated. evoked by a cafeteria diet did not increase body Table I. Effects of dietary manipulation on plasma GIP concentrations in obese (ob/ob) mice. Energy intake Plasma GIP Dietary manipulation Body weight (g) (kJ/mouse/day) (pg/ml) Freely fed 85.6 ± 2 .1 5.45 ±0.11 2041 ± 462 22-hour fasted 85.4 ± 3.0 7.92 ± 0.06t 820 ± 125* Dietary restricted 6 1 .2 ± 2 .8 t 51 ± It 845 ± 115* Cafeteria fed 92.0 ± 3.2$ 2.90 ± 0.29td 5512 ± 373t$ Dietary restriction and cafeteria feeding were imposed five weeks previously by reduction of the feeding period to 2 h/day (8) and by supplementation of the stock diet with palatable food items (7. such as chocolate and cheese) respectively. Energy intake was calculated from the daily food consumption values of the previous five weeks. Dietary restricted mice were fasted for 22 hours prior to blood sampling. Values are means ± SEM of six mice. Statistical significance was evaluated by unpaired Student’s t test: * p < 0.05, t p < 0.001 compared with fed mice; + p < 0.001 compared with dietary restricted mice. Short Communications 111 Table II. Effects of insulin on basal, fat-stimulated and glucose-stimulated plasma GIP concentrations in obese (ob/ob) mice. Plasma GIP (pg/ml) Dose of Oral insulin Insulin-treated administration (U/kg) Control mice mice None basal 5 469 ±46 475 ± 77 None (basal) 100 501 ± 111 615 ± 152 Fat 100 1882 ± 297 1510 ± 206 Glucose 100 612 ± 40 729 ± 95 Monocomponent porcine insulin was administered intraperitoneally to untreated mice or to mice treated 30 min previously with either oral fat (Intralipid, 32.2 kJ/kg) or oral glucose (2 g/kg, 32.2 kJ/kg). Blood was collected 30 min later. Values are means ± SEM of six mice. Fat and glucose significantly raised GIP in control mice (p < 0.001 and p < 0.05, respectively by unpaired Student’s t test). GIP conentrations were not modified by insulin treatment. weight as expected (7), indicating retained modified by relatively short periods of food dep capacity for diet-induced thermogenesis in adult rivation or excess. Oral administration of glucose ob/ob mice. Plasma GIP concentrations of the and particularly fat provoke large increases in obese mutant were substantially higher than those GIP which are not suppressed by insulin. The of lean mice (289 ± 11 pg/ml, mean ± SEM, K-cells should therefore be added to the growing n = 8), and were markedly reduced but not nor list of target cells which exhibit insulin insensitiv malized by a 22 h fast. Extended dietary restric ity in ob/ob mice (1, 3). The results support the tion did not produce a further disease, but view that hyperphagia and K-cell insulin insen cafeteria feeding markedly raised GIP concen sitivity promote the hypergipaemia of obese trations. Since nutrient-stimulated GIP secretion (ob/ob) mice. is normally suppressed by 1 U/kg insulin in rodents (6), the effects of insulin were examined REFERENCES in ob/ob mice. As shown in Table II, doses of up 1. Flatt PR, Bailey CJ. Horm Metab Res 1981, 13, to 100 U/kg insulin failed to suppress either basal, 556-560 fat-stimulated or glucose-stimulated GIP concen 2. Flatt PR, Bailey CJ. Horm Metab Res 1982, 14, trations. This indicates that the K-cells of the 127-130 3. Flatt PR, Bailey CJ, Buchanan KD. J Endocrinol obese mutant are grossly insensitive to the normal 1982, 95, 215-227 inhibitory action of insulin, and that the marked 4. Bailey CJ, Flatt PR, Atkins TW. Int J Obesity 1982, GIP releasing action of nutrients such as fat are 6 , 11-21 5. Flatt PR, Bailey CJ, Kwasowski P, Swanston-Flatt responsible for the massive GIP response to SK, Marks V. Diabetes 1983, 32, 433-435 feeding. 6. Hampton SM, Kwasowski P, Tan K, Morgan LM, To conclude, the present study confirms the Marks V. Diabetologia 1983, 24, 278-281 7. Trayhum P, Jones PM, McGuckin MM, Goodbody existence of hypergipaemia in ob/ob mice and AE. Nature 1982, 295, 323-325 shows that circulating GIP concentrations are 8. Flatt PR, Bailey CJ. Metabolism 1983, 32, 504-509 Short Communications 213 Immunoreactive GIP in Plasma and Small Intestine of Obese Hyperglycaemic (ob/ob) Mice P. R. FLATT, P. KWASOWSKI, S. K. SWANSTON-FLATT, V. MARKS & C. J. BAILEY* Divisions of Nutrition & Food Science, and Clinical Biochemistry, Dept, of Biochemistry, University of Surrey, Guildford, Surrey, GU2 5XH, UK Hyperphagia and hyperinsulinaemia are promi dependent on factors generated by the ingestion, nent features in the pathogenesis of the ob/ob enteral presence and absorption of nutrients. For syndrome in mice (1,2). As shown in Fig. 1, example, fed adult ob/ob mice exhibit a marked unchecked food consumption with resulting insulin response to enteral but not parenteral hypersecretion of insulin promotes excessive fat glucose (4-6). Furthermore, the hyperinsulinae deposition through the production of severe mia and hyperphagia follow similar age-related insulin resistance linked to the development of patterns, and plasma insulin concentrations fall pancreatic A-cell dysfunction and the manifes markedly when food is withheld (1,2). These tation of hyperglycaemia (1-3). Several lines of observations focus attention on the enteroinsular evidence indicate that the hyperinsulinaemia is axis, in which GIP has been ascribed special importance. T o assess the possible involvement • On research leave from the University of Aston in Birmingham. of GIP in the hyperinsulinaemia of the ob/ob mutant, we have examined immunoreactive GIP Hypothalamus in the plasma and small intestine. / \ Fed lean and ob/ob mice derived from the Appetite Satiety Aston colony (1,7) were used at 5-6 months of Hyperphagia age. Blood samples for analysis of plasma glucose, j insulin and GIP (7,8) were obtained by cardiac Intestine puncture under ether anaesthesia. The small GIP and other lac tors* intestine (duodenum, jejunum and ileum) was excised, washed with ice-cold 0.154 mmol/l NaCl, Insulin ^ Insulin Hyperglycaemia weighed and extracted with acid ethanol (750 ml Hypersecretion Resistance* ethanol, 250 ml water, 15 ml hydrochloric acid). As shown in Table I, ob/ob mice characteristically exhibited overweight, moderate hyperglycaemia Hyperglucagonaemia and pronounced hyperinsulinaemia. In ob/ob mice; the intestine was 1.4 times heavier than Adipose Tissue lean mice, the intestinal concentration of GIP was 1.9 times higher, and the total intestinal GIP Obesity content was 2.8 times greater. Consistent with * Insulin hypersecretion can result from (a) direct these changes, the plasma GIP concentration was action of amino acids on the pancreatic B-cells, and 15 times higher, indicating an enhanced rate of (b) indirect action of nutrients (especially dietary carbo synthesis and secretion of GIP by the small intes hydrates) mediated by neural and hormonal compo nents of the enteroinsular axis in which GIP has been tine of ob/ob mice. The results confirm the ascribed special importance. enlargement of the intestine in ob/ob mice and demonstrate that the concentration and total con Fig. 1. Proposed relationship between hyperphagia and hyperinsulinaemia in the metabolic abnormalities of tent of GIP are increased. This is consistent with ob/ob mice. previous reports suggesting an increased popu- 214 Short Communications Table I. Immunoreactive GIP in plasma and small linaemia. However, recent studies suggest that intestine of lean and obese (ob/ob) mice GIP is unlikely to be the sole component of the Lean mice Obese mice enteroinsular axis in ob/ob mice (6), and other insulin-releasing gut hormones are likely to con Body weight (g) 38.8 ± 2.3 70.6 ± 4.1$ tribute to the severity of the hyperinsulinaemia. Plasma glucose (mmo(/l) 7.4 ± 0.3 13.3 ± 1.8* Plasma insulin (ng/ml) 2.1 ± 0.5 32.4 ± 2.4$ Plasma GIP (p^ml) 289 ± 11 4350 ± 877$ Intestinal weight (g) 1.9 ± 0.1 2.6 ± 0.2$ REFERENCES Intestinal GIP (ug^g) 0.78 ± 0.07 1.48 ± 0.20$ Total intestinal GIP (ug) 1.45 ± 0.14 4.12 ± 0.61$ 1. Bailey CJ, Flatt PR, Atkins TW. Int J Obesity 1982, 6 , 11-12 Values are means ± S.E.M. for 6-8 mice. Statistical 2. Flatt PR, Bailey CJ. Horm Metab Res 1981, 10, significance was evaluated by Student’s t test: *p < 556-560 0.02, tp < 0.01, $p < 0.001, compared with lean mice. 3. Flatt PR, Bailey CJ, Buchanan KD. J Endocrinol 1982, 95, 215-227 4. Flatt PR, Bailey CJ. Biochem Soc Trans 1981, 9, lation of endocrine cells in the intestine of ob/ob 220-221 5. Flatt PR, Bailey CJ. Biochem Soc Trans 1982, 10, mice (9,10). Since GIP is viewed as a powerful 29-30 potentiator of nutrient-induced insulin release, 6. Flatt PR, Bailey CJ. J Nutr 1982, 112, 2332-2336 the magnitude of the increase in plasma GIP 7. Flatt PR, Bailey CJ. Diabetologia 1981, 20, 573- 577 concentrations suggests that this hormone makes 8. Morgan LM, Morris BA , Marks V. Ann Clin a substantial contribution to the hyperinsulinae Biochem 1978, 15, 172-177 mia and related abnormalities of the syndrome 9. Polak JM, Pearse AGE, Grimelius L, Marks V. Virchows Arch Cell Pathol B 1975, 19, 135-150 (Fig. 1). Gastrointestinal factors therefore appear 10. Best LC, Atkins TW, Bailey CJ, Flatt PR, Newton to represent the major stimulus to hyperinsu DF, Matty AJ. J Endocrinol 1977, 72, 44P ^ British Journal o f Nutrition (1985), 53, 467-475 4 67 The effect of guar gum on carbohydrate-, fat- and protein-stimulated gut hormone secretion: modification of postprandial gastric inhibitory polypeptide and gastrin responses B y L. M. M O R G A N , ! J. A. TREDGER,^ A. MADD E N , ^ P. K W A S O W S K P A N D V. M A R K S ! Divisions of ^Clinical Biochemistry and ^Nutrition, Department of Biochemistry, University of Surrey, Guildford, Surrey GU2 5XH {Received 18 June 1984 - Accepted 30 November 1984) 1. The effect of incorporating guar gum into predominantly single-component meals of carbohydrate, fat or protein on liquid gastric emptying and on the secretion of gastric inhibitory polypeptide (GIP), gastrin and motilin, was studied in healthy human volunteers. 2. Volunteers were given either 80 ml Hycal (carbohydrate meal), 150 g cooked lean minced beef (protein meal) or 200 ml double cream (fat meal) either with or without 5 or 6 g guar gum. Liquid gastric emptying was monitored in the fat and protein meals by taking 15 g paracetamol, consumed in water, with the meals and monitoring its appearance in circulation. 3. Postprandial insulin and GIP levels were both significantly reduced by addition of guar gum to the carbohydrate meal. Postprandial GIP secretion was also reduced by addition of guar gum to the protein meal, but protein-stimulated gastrin secretion was enhanced by guar gum. There was a significant negative correlation between peak circulating gastrin levels and the corresponding GIP levels. Postprandial GIP secretion and plasma motilin levels were unaffected by addition of guar gum to the fat meal. 4. 5 and 10 g guar gum/1 solutions in water possessed buffering capacities between pH 2-75 and 5-5. 5. Guar gum at 5 g/1 caused no detectable change in liquid gastric-emptying time. 6. The observed augmentation of gastrin secretion by guar gum following a protein meal could be due either to the buffering capacity of guar gum or to the attenuation of GIP secretion. It is possible that the chronic use of guar gum could be associated with changes in gastric acid secretion. Gu a r g u m is an unabsorbable carbohydrate derived from the seeds of the Indian cluster bean {Cyamopsis tetragonoloba). W h e n incorporated into a meal it reduces postprandial hyperglycaemia (Jenkins et al. 1976, 1911 a, 1978; M o r g a n et al. 1979) and has been used successfully as an adjunct therapy for both insulin-dependent and non-insulin-dependent diabetics (Jenkins et al. 19776, 1980). Used on a regular daily basis it can also lower plasma cholesterol levels, and has been used in the treatment of patients with hypercholesterolaemia (Jenkins et al. 1979). The mechanisms by which guar gu m exerts its effect on glucose tolerance and circulating cholesterol levels are not yet fully understood. Previous investigations have centred primarily on the effect of guar g u m on carbohydrate absorption and metabolism, where it has been shown to slow do w n the rate of absorption of glucose from the small intestine (Caspary et al. 1980) and to diminish postprandial insulin secretion (Jenkins et al. 1977 a). Secretion of the insulin-stimulating gastrointestinal hormone, gastric inhibitory polypeptide (GIP), is also reduced by the addition of guar gu m (Morgan et al. 1979) and this reduction has been postulated as being partly responsible for the reduction in insulin secretion observed. In addition to its insulin-stimulating properties, GI P has the ability to inhibit gastric acid secretion (Brown et al. 1975). Gu a r gu m has also been shown to delay the gastric emptying of a liquid test meal in animals (Leeds et al. 1979; Rainbird et al. 1983, 1984) and in ma n (Wilmshurst & Crawley, 1980; Blackburn et al. 1984). Gu a r gu m ma y exert an effect either directly or indirectly, via changes in GIP, on the secretion of other gastrointestinal 'X 468 L. M. Morgan and others ^ hormones, namely gastrin and motilin, hormones involved in gastric acid secretion and gastric motility (Christofides et al. 1919 a). The effects of guar gu m on motilin and gastrin secretion have not previously been studied. W e have therefore investigated the effect of incorporating guar gu m with predominantly single-component meals of fat, protein or carbohydrate on the secretion of the gastrointestinal hormones GIP, gastrin and motilin. The effect of guar gu m on liquid gastric emptying was also monitored. MATERIALS AND METHODS Clinical studies Nineteen healthy male volunteers took part in the study. They were aged between 19 and 35 years and were within 1 0 % of their ideal body-weight. Each subject gave his informed consent and the study was approved by the Ethical Committees of St Luke’s Hospital, Guildford, and the University of Surrey. Subjects attended after an overnight fast on two separate occasions at least 1 week apart and were assigned to one of the test meal protocols. Carbohydrate meal.VwQ subjects consumed 80 ml Hycal (Beechams, Brentford) diluted to 250 ml with water. This provided the equivalent of 50 g glucose in the form of glucose and glucose polymers. O n one of the occasions the solution contained 5 g guar g u m (Norgine, London) of a high viscosity grade (O ’Connor et al. 1981) which was whisked into it immediately before consumption. Protein meal. Six subjects were given a meal of 150 g lean minced beef, providing 30 g protein and 7 g fat, either with or without 5 g guar gu m which was incorporated into the meat before cooking. Paracetamol (15 g) was consumed simultaneously on each occasion with 300 ml water and plasma paracetamol levels measured as an index of gastric emptying (Holt et al. 1979). Fat meal. Eight subjects consumed 200 ml double cream which provided 96 g fat, 3 g protein and 4 g carbohydrate, flavoured with 5 g Kool Aid (General Foods Corp., N e w York), a soft-drink flavouring of no nutritional or energy value. G u a r g u m (6 g) was whisked into the test meal immediately before consumption. In an attempt to provide two meals of a similar consistency, the cream in the control meal was whipped to a firm texture. Paracetamol (15 g) was consumed with each of the meals, as before, and plasma paracetamol levels measured as an index of gastric emptying. Venous blood samples were collected from an indwelling venous cannula kept patent with 0-123 M-sodium citrate whilst the subjects were at rest in the fasted state, and at intervals for 150 or 180 min from the start of the meals. In the initial experiment (carbohydrate meal), blood was collected for 180 min. As the maximal changes in circulating hormone levels occur between 30 and 100 min, postprandial sampling time was reduced to 150 min in subsequent experiments. Chemical analyses Blood glucose was measured using a glucose oxidase {EC 1.1.3.4) technique. I m m u n o reactive plasma insulin, GIP, gastrin and motilin were measured by double-antibody radioimmunoassay techniques (Dryburgh & Brown, 1975; M o r g a n et al. 1978) using antibodies supplied by Guildhay Antisera (University of Surrey, Guildford). Plasma tri glycerides were measured using a fully enzymic u.v. kit (Boehringer, Mannheim). Plasma paracetamol was measured by high performance liquid chromatography using a 100 x 5 m m O D S (C^g) Hypersil (5 fim) column and a running solvent consisting of 0-05 M-sodium acetate (pH 4-6) in methanol (80:20, v/v). Benzoic acid was used as an internal standard and the eluate monitored with a u.v. detector at 249 nm. Modification of gastrin and GIP by guar gum 469 Light-scattering index Light-scattering indices were measured on a Thorpe Nephalometer (Thorpe Instruments, Cheshire) as an index of postprandial chylomicronaemia. Buffering capacity of guar gum Solutions containing 5 and 10 g guar gum/1 in 50 ml water were prepared. The pH change was measured when they were titrated, in triplicate, against a solution of 0-1 M-hydrochloric acid and compared with a control of 50 ml water. Statistical analyses Results were compared using Student’s t tests for either paired or unpaired data, where appropriate. Areas under the curve were calculated using the trapezoidal rule. P values of < 0 05 were accepted as statistically significant. Correlation coefficients were calculated by Spearman’s rho. RESULTS Carbohydrate meal. Plasma insulin and GI P levels following oral glucose are shown in Fig. 1. Postprandial insulin and GI P levels were both significantly reduced by addition of guar g u m to the meal (areas under the plasma insulin curve 0-180 min: 63 3 (se 17-4) compared with 21 3 (SE 3 6) mU / 1 per h for control and guar gu m meals respectively, P < 0 01 ; areas under the plasma GI P curve 0-180 min: 2710 (se 416) compared with 1760 (se 256) ng/1 per h, P < 0 01). Addition of guar g u m to the carbohydrate meal did not significantly affect peak blood glucose levels (5 7 (se 0 3) mmol/l compared with 5 0 (se 0 3) mmol/l for control and guar g u m meals respectively). In the second phase of the experiment (90-180 min) the addition of guar gum, in each case, prevented the fall in venous blood glucose to below fasting levels, thus effectively ‘smoothing’ the postprandial glucose curve compared with that of the control. Plasma motilin was not measured after the carbohydrate meal, as circulating motilin levels fall after oral carbohydrate to levels below the sensitivity of the assay (<50 ng/1). Protein meal. Plasma gastrin and GI P levels following ingestion of the protein meal are shown in Fig. 2. Postprandial GI P secretion was significantly decreased by addition of guar g u m to the meal (area under the plasma GI P curve 0-150 min: 1744 (se 118) compared with 1265 (sE 100) ng/1 per h for control and guar gu m meals respectively, P < 0 05). In contrast, postprandial gastrin secretion was enhanced by the addition of guar gu m (area under the plasma gastrin curve 0-120 min: 43 (se 11) compared with 68 (se 14) mU / 1 per h for control and guar g u m meals respectively, P < 0-005). There was a highly significant negative correlation between peak plasma gastrin levels and the corresponding plasma G I P levels (Fig. 3) in both control and guar gu m meals (r 0-80, P < 0-01). Plasma motilin levels rose postprandially but were similar after control and guar g u m meals (basal motilin levels 108 (se 18) ng/1; peak levels for control and guar g u m meals 155 (sE 37) ng/1 and 140 (se 22) ng/1 respectively). The area under the plasma motilin curve 0-150 min (control meal 142 (se 34) ng/1 per h) was not significantly affected by addition of guar g u m to the meal (guar g u m meal 121 (se 11) ng/1 per h). Peak circulating levels of paracetamol were obtained 20 min after ingestion of both control and guar gu m meals (Fig. 4). Gastric emptying, as assessed by the appearance and concentration of paracetamol in the circulation, was not affected by adding guar g u m to the meal. Plasma glucose and insulin were not significantly elevated above basal levels at any time point after consumption of the protein meal. A 470 L. M. M o r g a n a n d o t h e r s 70 3 E I E 1200 1000 g I 800 (3 i GOO " - a 51 400 200 0 30 60 90 120 150 180 Period after administration of meal (min) Fig. 1. Plasma insulin and gastric inhibitory polypeptide (GIP) concentrations (mean values with their standard errors represented by vertical bars) in five healthy subjects following a 50 g glucose meal with (O ) or without ( # ) 5 g guar gum (* P < 0 05,**P < 0-025). Fat meal. Plasma G I F levels rose following oral fat but were not significantly affected by the presence of guar gu m (area under plasma GI F curve 0-150 min: 1775 ( s e 240) ng/1 per h compared with 1707 ( s e 169) ng/1 per h for control and guar gu m meals respectively). Plasma motilin levels were similarly unaffected - they rose from a me a n basal level of 299 (sE 58) ng/1 to 356 ( s e 108) and 463 ( s e 108) ng/1 for control and guar g u m meals respectively. The area under the plasma motilin curve 0-150 min (control meal 248 (sE 83) ng/1 per h) was not significantly affected by addition of guar g u m to the fat meal (guar gu m meal 257 ( s e 67) ng/1 per h). Plasma triglyceride levels did not rise significantly until 60 min following the fat load. The me a n plasma triglyceride level at 120 min was significantly higher after the control than after the guar g u m meal (2 29 ( s e 0*6) mmol/l compared with 1 49 ( s e 0*3) mmol/l for control and guar g u m meals respectively, P < 0 05) but the area under the plasma triglyceride curve 60-150 min (control meal 0 78 (sE 0 6) mmol/l per h) was similar on the two occasions (guar gu m meal 0-27 ( s e 0 25) mmol/l per h). Plasma light-scattering indices (as an index of postprandial chylomicronaemia) rose significantly following the meal but were unaffected by addition of guar gum. Plasma paracetamol levels following the two fat meals were unaffected by guar g u m Modification of gastrin and GIP by guar gum 471 3 E to a E ?- ^ ---- % Û. 1200 1000 E: 800 0 1 I 600 Î Î Î 400 0 30 60 90 120 150 Period after administration of meal (min) Fig. 2. Plasma gastrin and gastric inhibitory polypeptide (GIP) concentrations (mean values with their standard errors represented by vertical bars) in six healthy subjects following a meal containing 30 g protein with (O ) or without (@) 5 g guar gum (* P < 0 05,** P < 0 025). (Fig. 4). However, circulating paracetamol levels 20 min after ingestion of the fat meal were significantly lower than after the protein meal, and peak circulating levels were not attained until 40 min after meal ingestion. Plasma glucose and insulin levels were not significantly elevated above basal levels at any time point after consumption of the fat meal. Buffering capacity of guar gum. The titration curves of a 5 and 10 g guar gum/1 solution in water are shown in Fig. 5. Both concentrations of guar gu m exhibited a buffering capacity between p H 2-75 and 5-5. DISCUSSION Addition of guar gu m to an oral glucose load reduces the postprandial secretion of both G I F and insulin. These findings are in agreement with previous studies where addition of guar gu m to a high carbohydrate solid or liquid meal was found to reduce GI F and insulin secretions in both normal and diabetic subjects (Morgan ct al. 1979; O ’Connor ct al. 1981). The release of GI F from the gut in response to food m a y be related to the rate of active absorption of carbohydrates (Ebert & Creutzfeldt, 1978; Sykes ct al. 1980). Because guar g u m increases the viscosity of solutions, it has been suggested that addition of guar g u m to meals reduces the rate of diffusion of nutrients in the gut towards the absorptive surface (Caspary ct al. 1980), although the total absorption of carbohydrate, as measured by urinary xylose excretion or breath hydrogen studies, is unchanged (Gassull ct al. 1976; Leeds ct al. 1978). GI F secretion is also related to the rate of gastric emptying and conditions where rapid gastric emptying occurs result in an increased G I F response (Creutzfeldt, 1981). 472 L. M. M o r g a n a n d o t h e r s 160 140 120 3 1 100 I “ 80 E a I 0 200 400 600 800 1000 2000 Plasm a GIF (ng/1) Fig. 3. Correlation between peak plasma gastrin concentrations and the corresponding plasma gastric inhibitory polypeptide (GIP) concentration in six healthy subjects following a meal containing 30 g protein with (O ) or without (@) 5 g guar gum (r 0 80,P < 0 01). 200 _ 160 - 120 80 40 0 20 40 60 80 100 120 Period after administration of meal (min) Fig. 4. Plasma paracetamol levels in healthy volunteers given fat (A , A ) or protein (© , O) meals either with (A , O) or without (A , ® ) guar gum. Mean values with their standard errors represented by vertical bars, n 6 for protein meal, n 8 for fat meal. Modification of gastrin and GIP by guar gum 473 5-5 5 0 4 5 4 0 3 5 3 0 2 5 20 - 4 M 0 10 50 100 500 1000 HCI (jumol) Fig. 5. Titration curves of suspensions of guar gum in 50 ml tap water titrated against a solution of 0-1 M-hydrochloric acid. Mean values for triplicate estimations. O, Tap water; # , 5 g guar gum/1; A , 10 g guar gum/1. Delayed gastric emptying in humans has been shown by adding guar gu m (Wilmshurst & Crawley, 1980) and a mixture of pectin and guar gu m to a liquid meal (Holt et al. 1979), and guar g u m alone delays gastric emptying of a liquid glucose meal in rats (Leeds et al. 1979) and pigs (Rainbird et al. 1984). GI P stimulates insulin secretion during hyperglycaemia (Brown et al. 1975), and the diminution in insulin response to test meals by the addition of guar gu m is probably, at least in part, due to diminished GI F secretion. Addition of guar gu m to the protein meal also resulted in diminished postprandial GI P secretion. Intestinal absorption of free amino acids, as with glucose, takes place via active transport systems and therefore guar gu m is also likely to slow do w n the rate of amino acid absorption, as has been shown in vitro (Elsenhans et al. 1980). In contrast, gastrin secretion was enhanced by adding guar gu m to the protein meal. The mechanism by which guar gu m exerts this effect is not clear. The products of partial protein digestion provide the main stimulus to gastrin secretion in response to feeding. Gastrin secretion can be affected by both the rate of gastric emptying and the p H of stomach contents. Gu a r gu m did not affect the rate of gastric emptying of the liquid phase of the protein meal as assessed by the appearance of paracetamol in the circulation. This method of assessing liquid gastric emptying has previously been found to correlate well with sequential scintiscanning techniques using ^^^™In D T P A (Holt et al. 1979). In a subsequent study (J. A. Tredger et al. 1984) using ®®™Tc-labelled filter paper incorporated into minced beef, we have shown that guar gu m does not affect solid-phase gastric emptying of the protein meal. The results in hu m a n s are in broad agreement with work carried out in pigs which showed that guar gu m had no effect on the rate of gastric emptying of dry matter after consumption of a high-solid mixed meal (Rainbird et a/. 1983). 16 NUT 53 474 L. M. M o r g a n AND OTHERS The differences in gastrin secretion observed cannot, therefore, be explained in terms of the length of time the food was in contact with the G cells. In the present study, suspensions of guar g u m in water exhibited a significant buffering capacity at p H values above 2-75. G u a r g u m m a y possibly affect gastrin secretion by virtue of this buffering capacity since gastrin release is inhibited by acid acting directly on the antral G-cells. A m i n o acid-stimulated gastrin release is 80 % inhibited in normal ma n (Walsh et al. 1975) by a luminal p H of 2-5. Fasting antral p H is usually in the region of 1-3 (Dotevall, 1961) and rises after ingestion of a meal, the exact extent depending on the diluting and buffering capacity of the food, the volume of gastric acid secreted and the rate of gastric emptying. G u a r g u m m a y raise postprandial intragastric pH, as has been found with other fibres (T adesse, 1982), by virtue of its native buffering capacity thereby diminishing the contribution of the negative feed-back loop in inhibiting gastrin release. Modulation of GI P secretion by guar g u m provides an alternative explanation for the observed differences in gastrin secretion provoked by the control and guar gu m meals. GI P was initially na m e d for its capacity to inhibit gastric acid secretion (Brown et al. 1975) and exogenous G I P infusions have been shown to inhibit gastrin release (Villar et al. 1976; Arnold et al. 1978), the effect seemingly being mediated via an increase in gastric somato statin secretion (McIntosh et al. 1979). Studies in dogs by Wolfe et al. (1983) using anti- G I P antibodies have shown that GI P can function as a physiological inhibitor of gastric acid secretion through its effect on gastrin release. It is possible that attenuation by guar gum, of the GI P response to the protein meal, leads to an attenuated somatostatin response and hence unrestrained gastrin secretion. The highly significant negative correlation (P < 0-01) between circulating gastrin and GI P levels is consistent with this hypothesis. G I P secretion was unaffected by the addition of guar gu m to the fat meal. Although serum triglyceride levels are not the ideal indicators of fat absorption, especially in the first hour following ingestion, plasma triglyceride levels were similar after both guar-gum-supplemented and control meals, suggesting that other factors are more important in determining the rate of fat absorption. This is further supported by the failure to find any differences in the plasma light-scattering indices, a measure of chylomicronaemia, between the control and guar gu m meals. These findings are in contrast to those of Jenkins (1978), w h o reported that guar g u m was associated with an elevation in postprandial triglyceride levels after a mixed meal which he postulated was due to an effect of guar gu m on postprandial chylomicronaemia. Plasma paracetamol levels (which are an index of liquid gastric emptying rates) were similar after control and guar gu m fat meals, although levels 20 min after ingestion of the fat meal were significantly lower than after ingestion of the protein meal. This suggests a slower rate of empyting for the fat meal and is consistent with previously published studies on the effect of meal composition on gastric emptying (Hunt & Stubbs, 1975). It seems probable that guar gu m is unable to exert any additional effect in delaying gastric emptying in this situation. Motilin is released in response to oral fat and protein and is implicated in altering the rate of solid gastric emptying (Christofides et al. 19796). However, in the present study, guar gu m did not affect the secretion of motilin in response to either fat or protein nor were any changes in gastric-emptying rate of the liquid phase detected. The present study has indicated that guar gum, in amounts that would normally be consumed at each meal, affects the secretion of G I P and gastrin, both of which are implicated in the control of gastric acidity. As the use of guar gu m in the therapy of diabetes and hyperlipidaemia is, of necessity, a long-term one, it would be desirable to ascertain whether the chronic use of guar g u m is associated with changes in gastric acid secretion and whether guar gu m can affect the rate of solid phase gastric emptying in humans. Modification of gastrin and GIP by guar gum 475 REFERENCES Arnold, R., Ebert, R., Creutzfeldt, W., Becker, H. D. & Borger, H. (1978). Scandinavian Journal of Gastroenterology 13 (Suppl. 48), 11. Blackburn, N. A., Redfern, J. S., Jarjis, H., Holgate, A. M., Hanning, I., Scarpello, J. H. B., Johnson, I. T. & Read, N. W. (1984). Clinical Science 66, 329-336. Brown, J. C., Dryburgh, J. R., Ross, S. A. & Dupre, J. (1975).Recent Progress in Hormone Research 31,487-532. Caspary, W. P., Elsenhans, B., Siifke, U., Ptok, M., Blume, R., Lembcke, B. & Creutzfeldt, W. (1980). In Frontiers in Hormone Research, vol. 7, pp. 202-217 [W. Creutzfeldt, editor]. Basel: Karger. Christofides, N. D., Bloom, S. R., Besterman, H. S., Adrian, T. E. & Ghatei, M. A. (1979a). Gut 20, 102-106. Christofides, N. D., Modlin, I. M., Fitzpatrick, M. L. & Bloom, S. R. (19796). Gastroenterology 76, 903-907. Creutzfeldt, W. (1981). In Gut Hormones, 2nd ed., pp. 533-540 [S. R. Bloom and J. M. Polak, editors]. London: Churchill Livingstone. Dotevall, G. (1961). Acta Medica Scandinavica 170, 59-63. Dryburgh, J. R. & Brown, J. C. (1975).Gastroenterology G8, 1169-1176. Ebert, R. & Creutzfeldt, W. (1978). In Gut Hormones, pp. 294-300 [S. R. Bloom, editor]. London: Churchill Livingstone. Elsenhans, B., Siifke, U., Blume, R. & Caspary, W. F. (1980). Clinical Science 59, 373-380. Gassull, M. A., Goff, D. V., Haisman, P., Hockaday, T. D. R., Jenkins, D. J. A., Jones, K., Leeds, A. R. & Wolever, T. M. S. (1976). Journal o f Physiology {London) 259, 52-53. Holt, S., Heading, R. C., Carter, D. C., Prescott, L. F. & Tothill, P. (1979).Lancet i, 636-639. Hunt, J. N. & Stubbs, D. F. (1975). Journal of Physiology {London) 245, 209-225. Jenkins, D. J. A. (1978). In International Conference on Atherosclerosis, pp. 173-176 [L. A. Carson, R. Paoletta, C. R. Sirton and G. Weber, editors]. New York: Raven Press. Jenkins, D. J. A., Hockaday, T. D. R., Howarth, R., Apling, E. C., Wolever, T. M. S., Leeds, A. R., Bacon, S. & Dilawari, J. (19776). Lancet ii, 779-780. Jenkins, D. J. A., Leeds, A. R., Gassull, M. A., Goff, D. V., Wolever, T. M. S. & Alberti, K. G. M. M. (1977a). Proceedings o f the Nutrition Society 36, 44A. Jenkins, D. J. A., Leeds, A. R., Slavin, B., Mann, J. & Jepson, E. M. (1979).American Journal of Clinical Nutrition 32, 16-18. Jenkins, D. J. A., Leeds, A. R., Wolever, T. M. S., Goff, D. V., Alberti, K. G. M. M., Gassull, M. A. & Hockaday, T. D. R. (1976). Lancet ii, 172-174. Jenkins, D. J. A., Wolever, T. M. S., Leeds, A. R., Gassull, M. A., Haisman, P., Dilawari, J., Goff, D. V., Metz, G. L. & Alberti, K. G. M. M. (1978). British Medical Journal i, 1392-1394. Jenkins, D. J. A., Wolever, T. M. S., Taylor, R. H., Reynolds, D., Nineham, R. & Hockaday, T. D. R. (1980). British Medical Journal29Xi, Leeds, A. R., Bolster, N. R., Andrews, R. & Truswell, A. S. (1979). Proceedings o f the Nutrition Society 38, 44A. Leeds, A. R., Bolster, N. R. & Truswell, A. S. (1978). Proceedings of the Nutrition Society 37, 89A. McIntosh, C. H. S., Pederson, R. A., Koop, H. & Brown, J. C. (1979). InGut Peptides, pp. 100-104 [A. Miyoshi, editor]. Amsterdam: Elsevier North Holland. Morgan, L. M., Goulder, T. J., Tsiolakis, D., Marks, V. & Alberti, K. G. M. M. (1979).Diabetologia 17, 85-89. Morgan, L. M., Morris, B. A. & Marks, V. (1978). Annals of Clinical Biochemistry 15, 172-177. O’Connor, N., Tredger, J. & Morgan, L. M. (1981).Diabetologia 20, 612-615. Rainbird, A. L., Low, A. G. & Sambrook, I. E. (1983). Proceedings of the Nutrition Society 42, 24A. Rainbird, A. L., Low, A. G. & Sambrook, I. E. (1984). Proceedings of the Nutrition Society 43, 28A. Sykes, S., Morgan, L. M., English, J. & Marks, V. (1980).Journal of Endocrinology 85, 201-207. Tadesse, K. (1982). Journal of Physiology 322, 102P-103P. Tredger, J. A., Morgan, L. M., Peke, J. & Marks, V. (1984).Regulatory Peptides 350.9, Villar, H. V., Fender, H. R., Rayford, P. L., Bloom, S. R., Ramus, N. I. & Thompson, J. C. (1976). Annals o f Surgery 184, 97-102. Walsh, J. H., Richardson, C. T. & Fordtran, J. S. (1975).Journal of Clinical Investigation 55, 462-468. Wilmshurst, P. & Crawley, J. C. W. (1980). British Journal of Nutrition 44, 1-6. Wolfe, M. M., Hocking, M. P., Maico, D. G. & McGuigan, J. E. (1983).Gastroenterology 84, 941-948. Printed in Great Britain 16-2