PROGRESS IN GASTROENTEROLOGY

Transcription

1 GASTROENTEROLOGY Copyright 1971 by 'The Williams & Wilkins Co. Vol. 60, No.1 Printed in U. S. A. PROGRESS IN GASTROENTEROLOGY INTESTINAL HORMONES AS INHIBITORS OF GASTRIC SECRETION LEONARD R. JOHNSON, PH.D., AND MORTON 1. GROSSMAN, M.D., PH.D. Department of Physiology and Biophysics, University of Oklahoma School of Medicine, Oklahoma City, Oklahoma, and Veterans Administration Center, Los Angeles, California In the past decade, the field of gastrointestinal hormones, formerly the province of the physiologist, was profoundly altered by several major biochemical incursions. Three hormones have been isolated and chemically identified-gastrin by Gregory and Tracy, and secretin and cholecystokinin (CCK) by Jorpes and Mutt. The availability of pure gastrointestinal hormones has clarified much of the physiological data and in addition has spawned a plethora of physiological studies which have in turn revealed numerous exciting and unsuspected facts. The purpose of this article is to review the studies of the past few years which have led to a greater understanding of one particular aspect of gastrointestinal secretory control, namely, inhibition of gastric secretion by hormones of intestinal origin. To place the subject in historical perspective, we start with a brief review of some of the more important physiological studies which preceded the biochemical isolation of the hormones. There follow sections on the actions of exogenous secretin and CCK on gastric secretion stimulated by various means in different species, the effect of endogenously released intestinal hormones on gastric secretion, the significance of intestinal hormonal control of gastric secretion, and finally a look at pos- Received August 10, Address requests for reprints to: Dr. Leonard R. Johnson, Department of Physiology and Biophysics, 800 NE 13th Street, Oklahoma City, Oklahoma sible future directions that work in this field may take. The first observation of what later proved to be a duodenal mechanism for the inhibition of gastric secretion was probably made by Ewald and Boas in They added olive oil to a test meal of starch paste given to human subjects and observed inhibition of both secretion and gastric emptying. According to Babkin, the first demonstration of a duodenal inhibitory mechanism was made in 1904 by Sokolov,2 a pupil of Pavlov. He performed a more definitive experiment than that of Ewald and Boas by shqwing that the presence of HCI in the duodenum of a dog inhibited the Pavlov pouch response to a meal. These findings were confirmed by other investigators, but until 1925 the mechanism involved was generally regarded as being solely neural. The now classical observation of Farrell and Ivy,3 that feeding fat inhibited the motility of a completely transplanted gastric pouch, was the basis for the belief in a humoral substance from the intestine which inhibited gastric motility or secretion, or both. Much of the work of the early 1930's which proved the existence of a duodenal hormone that inhibits gastric secretion can be credited to R. K. S. Lim. It was our pleasure and great fortune that Professor Lim spent several months working in our laboratory just prior to his death in July The following brief account of that early work is taken in part from some of his notes outlining the development of

2 Jarnw.ry 1971 PROGRESS IN GASTROENTEROLOGY 121 the concept of enterogastrone. Soon after Farrell and Ivy's repore on inhibition of motility by fat, Feng et al. 4 furnished the first evidence for the hormonal inhibition of gastric secretion. They showed that the secretory response from a transplanted pouch was decreased when fat was fed with a normal meat meal. The intravenous injection of chyle from another dog did not inhibit secretion, proving that the observed inhibition was not due to the absorption of substances from the meal itself. The authors concluded that fat presumably released a hormone from the duodenum which inhibited secretion. Awareness of the existence of such a hormone naturally led to a search for it in various duodenal extracts. Kosaka and Lim 5 demonstrated that a crude intestinal extract presumably containing CCK inhibited the Heidenhain pouch response to a meal and to histamine but did not produce secretin-like effects on the pancreas. After finding similar properties in an extract made from duodenal mucosa which had been exposed to olive oil, Kosaka and Lim s named the active principle enterogastrone (from enteron, gastron, chalone). The work to isolate enterogastrone was continued by Lim and his co-workers and by Ivy and his group, but met with little success. Interest in this area waned until the discovery by Greenlee et al. 7 that a commercial secretin preparation inhibited secretion from a Heidenhain pouch that was being stimulated by the release of endogenous gastrin. These investigators, however, found no inhibition of gastric secretion stimulated by histamine. Since the extract used by Kosaka and Lim 5 was reported to have no secretion-like activity and to inhibit histamine-stimulated secretion, Greenlee et al. 7 concluded that their extract contained a different inhibitor from the one described by Kosaka and Lim. In 1964, Gillespie and Grossman B found that a preparation of Vitrum cholecystokinin-pancreozymin (CCK-PZ) was a potent inhibitor of gastrin-stimulated secre tion from the dog Heidenhain pouch. They confirmed the observation of Greenlee et al. 7 that secretion inhibited gastrin, but not histamine-stimulated secretion. In addition, they found that this CCK-PZ preparation inhibited only low doses of histamine. By 1964, several facts had been established regarding the inhibition of gastric secretion by duodenal hormones. First, the presence of acid, fat, or hypertonic solutions in the duodenum released a hormone or hormones which inhibited gastric acid secretion. Second, the duodenal hormones, secretin and cholecystokinin-pancreozymin, inhibited gastrinstimulated secretion and were, therefore, enterogastrones. (We use the term enterogastrone to mean any hormone released from the intestine which inhibits gastric secretion.) Many problems, however, remained unsolved and in order to clarify the mechanisms involved during duodenal inhibition of gastric secretion several important steps had to be taken. The actions of secretin and cholecystokinin-pancreozymin had to be fully described when given in conjunction with both gastrin and histamine. Inhibition by acid or fat in the duodenum was well documented but only a small amount of information existed as to which hormone (or hormones) was responsible in each case. Kosaka and Lim had found that their enterogastrone extract inhibited histamine-stimulated secretion, but secretin and CCK-PZ had been shown to have little effect on histamine. Therefore, the question arises: does enterogastrone exist as a distinct hormone, or could all of the actions of fat in the duodenum be explained by CCK and secretin? These were pressing problems, and meaningful attempts at their solutions awaited the existence of pure preparations of secretin and CCK-PZ. The biochemical epoch in the history of gastrointestinal hormones began in 1964 when Gregory and Tracy9 isolated gastrin and Gregory and his co-workers determined its structure lo and synthesized it. ll In 1966, Mutt and Jorpes 12 announced the

3 122 PROGRESS IN GASTROENTEROLOGY Vol. 60, No.1 full amino acid sequence of secretin after spending many years on its isolation and purification Bodanzky and coworkers 15 soon synthesized the 27 -amino acid peptide, and secretin became the second gastrointestinal hormone to have been reproduced chemically. Jorpes and Mutt l6. 17 were also responsible for the isolation and purification of CCK-PZ. Except for the order of a few amino acid residues, the structure of CCK-PZ has been determined and the C-terminal heptapeptide amide of this hormone has been synthesized and shown to be more active than the whole molecule. 18 The C-terminal pentapeptide amides of CCK-PZ and gastrin are identical. 17 Until the work of Jorpes and Mutt, CCK and PZ had been considered to be separate hormones. We know now, however, that both of these activities are due to one hormone Throughout the rest of this article we refer to this hormone as CCK (cholecystokinin) since the cholecystokinetic property was described in while the ability to stimulate the secretion of pancreatic enzymes was not discovered until One of the major realizations of the past few years in this field is that the three known gastrointestinal hormones share many physiological actions. Therefore, it becomes awkward to name these hormones according to more than one action. Exogenous Administration of Secretin and Cholecystokinin In his concluding statements from a review on enterogastrone Gregory22 remarked that the source of the duodenal inhibition of gastric secretion could possibly be identified when pure preparations of secretin and CCK became available. These substances are now available and the next few pages summarize the results obtained with pure and, in the case of secretin, synthetic preparations of these hormones. Secretin Gastrin-stimulated secretion. The first suggestion that secretin was involved in the duodenal inhibition of gastric secretion was made by Greenlee et al. 7 in These investigators found that intravenous injections of impure secretin (Eli Lilly and Company, Indianapolis, Ind.) inhibited the Heidenhain pouch response to feeding and to antral pouch stimulation (i.e., endogenous gastrin) but not to histamine. Using two different secretin preparations, Jordan and Peterson 23 found a strong inhibition of the Heidenhain pouch response to a meal. Two results from Hallenbeck's laboratory were considered to be evidence that the inhibition noted by Jordan and Peterson was due to secretin supressing the release of gastrin First, the inhibitory effect of secretin was not due to a pancreatic product, for gastric secretion could be inhibited by secretin in the pancreatectomized animal. 24 Second, secretin was an ineffective inhibitor of alcohol-stimulated secretion, the mechanism of which was assumed not to involve the release of gastrin. 25 In 1964, Gillespie and Grossman 8 reported that single intravenous injections of Vitrum secretin inhibited the Heidenhain pouch response to continuous infusion of a gastrin extract. Soon afterward, Wormsley and Grossman 26 reported that a single intravenous injection of secretin inhibited gastrin-stimulated acid secretion from the gastric fistula as well as the Heidenhain pouch. Secretin produced a maximum inhibition of 75% of the response to a maximal dose of gastrin. There is, therefore, little doubt that secretin blocks the action of the hormone gastrin so it need not be postulated to interfere with its release (although the possibility that this is an additional action cannot be dismissed). The finding by Vagne et al. 27 that synthetic secretin was a potent inhibitor of canine gastric secretion meant that the inhibition of acid secretion produced by extracts of natural secretin was due, in fact, to secretin per se and not to an impurity. Secretin was also found to be a potent inhibitor of gastric secretion when administered as a continuous intravenous infusion 28 which simulates endogenous release more closely than does rapid intravenous injection. With near maximal stimulation by exogenous gastrin, a dose of secretin, which was submaximal for

4 January 1971 PROGRESS IN GASTROENTEROLOGY 123 pancreatic secretion, caused 90% inhibition of the acid response. These two findings were considered proof that the inhibition of gastric secretion was a physiological action of secretin. 28 A dose-response study29 revealed several characteristics of the inhibition by secretin of the Heidenhain pouch response to gastrin. We concluded that secretin is a stronger inhibitor of acid secretion than it is a stimulator of pancreatic secretion, based on the observation that the dose required for 50% inhibition of gastric secretion was 0.4 unit per kg-hr, whereas the dose required for 50% maximal pancreatic stimulation was 1.2 units per kg-hr. Inhibition actually occurred with as little as 0.06 unit of secretin per kg-hr. This dose is below the threshold for stimulating pancreatic secretion. Michaelis-Menten analysis of the dose-response curves to gastrin alone and to gastrin plus a fixed dose of secretin was typical of noncompetitive inhibition in that the calculated maximal response (CMR) was lowered by secretin and the dose of gastrin required for halfmaximal response (D50) remained the same. This was interpreted as evidence that secretin acts at a receptor site different from the one affected by gastrin. 29 The preceding experiments were done in dogs, but secretin has been shown to inhibit gastrin-stimulated secretion in other species as well. Stening et al. 30 found that gastric secretion in the cat was much less sensitive to inhibition by secretin than in the dog. Inhibition occurred but a higher dose of secretin was required and the effect was not as great. Konturek et al. 31 found somewhat more inhibition in that a dose of secretin which was maximal for pancreatic secretion inhibited by 50% acid secretion maximally stimulated by pentagastrin. However, in some earlier work, Konturek 32 found that secretin inhibited the acid response only to submaximal doses of pentagastrin in the cat. No explanation was given for the difference in the findings in these two studies. 3l, 32 There is, however, general agreement that secretin does inhibit in the cat but less than in the dog. In its response to secretin, the rat appears to be comparable to the dog. In one study,33 75 units per kg of secretin administered subcutaneously completely prevented the response to maximal and supramaximal doses of pentagastrin. "With regard to the inhibitory action of secretin on gastric acid secretion the human response lies between the extremes of canine sensitivity and feline resistance." This statement was taken from a recent paper by Brooks and Grossman 34 in which they reported the effect of secretin on pentagastrin-stimulated acid and pepsin secretion in human subjects. Reflux of duodenal contents into the stomach was prevented by a balloon placed in the duodenum and by suction. Natural secretion (2 units per kg-hr) produced a statistically significant 21% inhibition of the acid response to a maximal dose of pentagastrin. Konturek 35 did similar studies in human subjects but apparently did not succeed in preventing duodenogastric reflux because the decrease in acid output after secretin was attributable mainly to decrease in acid concentration. Wormslel 6 observed inhibition of pentagastrin-stimulated acid secretion by secretin in human subjects but, since his results are given only as acid output without information on acid concentration, the possible contribution of reflux cannot be judged. Chey et a1. 37 successfully controlled reflux and found that secretin inhibited the response to a low dose of pentagastrin (0.12 JJ.g per kg-hr) but had no effect on a higher dose (6 JJ.g per kg-hr). Further studies using a wide range of doses of secretin and pentagastrin are needed to give a more refined estimate of the sensitivity of man to inhibition by secretin as compared to other species. Histamine-stimulated secretion. Gastric acid secretion brought on by histamine is relatively resistant to inhibition by secretin as it is to other inhibitors such as atropine. Greenlee et a1. 7 found no inhibition by secretin of the acid responses from Heidenhain pouches during histamine stimulation. This result has been substantiated in many other studies employing single intravenous injections of secretin. 8, 26, 38, 39 Using continuous infusions of both histamine and secretin, Nakajima et a1. 40 reported that secretin inhibited the canine Heidenhain pouch response to low

5 124 PROGRESS IN GASTROENTEROLOGY Vol. 60, No.1 doses of histamine only. Whether this effect was actual inhibition by secretin is difficult to determine, for the investigators first constructed a dose-response curve to histamine and then repeated it on the same day in the presence of secretin. Therefore, there were no controls showing that the dogs' response to histamine had not merely decreased as the day progressed. Other studies in which secretin was infused for the middle hour of a 3-hr stimulation by continuous histamine infusions showed no change in acid response. 28,41 The conclusion was that histamine secretion in the Heidenhain pouch dog is not affected by secretin. Full exploration of a wide range of doses of histamine and secretin is needed but it is clear that histamine is more resistant than gastrin to inhibition by secretin. Wormsley and Grossman 26 found that secretin did inhibit the response of the vagally innervated main stomach to histamine. Information concerning vagal interactions with inhibitory hormones is lacking. However, since the vagus is important both in the release of gastrin and in sensitizing the parietal cell to the action of gastrin, it is likely that the acid response of the innervated stomach to histamine is dependent on the interplay of these factors. Johnston and Duthie 42 found no inhibition of histamine-stimulated secretin in 7 patients. Chey et al. 37 found in human subjects that secretin inhibited a low dose of histamine phosphate (0.02 mg per kghr) but not a high dose (0.04 mg per kg-hr). Recent studies have also shown that secretin does not inhibit histamine-stimulated secretion in cats 31 or rats. 43 Vagally stimulated secretion. In unpublished studies we found good inhibition with secretin against insulin, 2-deoxy-oglucose, and bethanechol stimulation from gastric fistulas in intact animals. Following antrectomy, however, the acid response was so reduced that it was impossible to determine whether or not secretin produced a change in secretion. These studies have recently been repeated by Way44 who used a continuous intravenous infusion of a threshold dose of histamine to produce a background level of activation which, when combined with insulin or 2-deoxy-o-glucose, resulted in reliable acid secretion from the Pavlov pouches of antrectomized animals. In this preparation, there was no evidence that secretin inhibited acid secretion in response to insulin or 2-deoxy-o-glucose. 44 In the same study, however, secretin did inhibit the response to bethanechol in the antrectomized dog. Before antrectomy, secretin strongly inhibited the responses to all three forms of stimulation. These experiments indicate that direct cholinergic stimulation of the parietal cell is not inhibited by secretin. It is unknown why bethanechol appears to be different. Previous reports of secretin inhibiting cholinergic stimuli in intact animals can probably be explained by the inhibition of endogenous gastrin released by vagal stimulation; even after antrectomy bethanechol probably acts by multiple mechanisms, and not just by cholinergic stimulation of parietal cells. Even though secretin acts at a different receptor site from gastrin, its receptor site must be intimately related to and interact with the gastrin site because secretin exerts its action selectively against gastrin and not against histamine and cholinergic stimuli. Cholecystokinin Effects on acid secretion. While studying the stimulation of pancreatic enzyme secretion, Preshaw and Grossman 45,46 noted that CCK when given alone in small amounts was a weak stimulant of gastric acid secretion. Other investigators have since confirmed that CCK is a partial agonist for gastric acid secretion. 47, 48 The stimulatory action of CCK led to some conflicting reports regarding its actions when administered in conjunction with other gastric secretory agonists. Gillespie and Grossman 8 reported that CCK inhibited the secretory response to gastrin in Heidenhain pouch dogs, and Bedi et al. 49 found that single intravenous injections of CCK (Cecekin) inhibited Heiden-

6 Jarumry 1971 PROGRESS IN GASTROENTEROLOGY 125 hain pouch responses to gastrin and histamine. In antrectomized animals, however, CCK failed to inhibit the acid response of either the Heidenhain pouch or innervated gastric remnant to histamine. 49 In fact, in these animals, CCK augmented the acid response to histamine. The earlier study had shown that as the dose of histamine increased the inhibitory effect of CCK lessened. 8 In trying to explain the observed augmentation from the gastric remnant, Bedi et al. 49 suggested that the Cecekin contained a gastric secretagogue in addition to an inhibitor of the acid response. They further speculated that the secretagogue could be gastrin which would potentiate a histamine response. A large potentiation of the acid response had been demonstrated to result from the combination of histamine and a small amount of gastrin. 50, 51 Some of the apparent incongruities concerning the action of CCK were clarified by Stening et al. 52 who found that against a histamine background the action of CCK depended on the dose of histamine. In this study, CCK as a single rapid intravenous injection, produced unpredictable changes in the gastric acid response to low doses of histamine (0.01 and 0.02 mg per kg-hr of histamine di-hci). In some dogs the pouch response was enhanced while in others it was inhibited. Against 0.01 mg per kg-hr inhibition prevailed, whereas there was slight stimulation of the response to 0.02 mg per kg-hr of histamine. This was in accord with findings reported by Gillespie and Grossman 8 who found that CCK inhibited the response to mg of histamine di-hci per hr (approximately equal to 0.01 mg per kg-hr) and had no effect when the dose of histamine was doubled. Stening et al. 52 also showed that CCK enhanced the acid response to submaximal doses of histamine above 0.02 mg per kg-hr. CCK evoked a reproducible increase in the pouch responses up to the level of maximal histamine stimulation. In this same study, a continuous infusion of 8 units of CCK per kg-hr caused 90% inhibition of the acid response to a maximal dose of gastrin. 52 In a previous study, using dogs and experimental protocol identical with that of Stening et al., 52 Johnson and Grossman 41 showed that a continuous infusion of CCK (4 units per kg-hr) had no effect on the gastric acid response to 0.04 mg of histamine di-hci per kg-hr. Thus the enhancing effect of CCK on histamine-stimulated secretion occurred only with rapid intravenous injections, not infusions, whereas the inhibitory action of CCK on gastrin-stimulated secretion was seen with both modes of administration. We cannot explain this discrepancy. Since CCK strongly inhibited acid responses to gastrin, there is no doubt that it is a powerful enterogastrone. On the other hand, CCK enhanced acid secretion in response to submaximal doses of histamine. Stening et al. 52 explained all of these actions by assuming that CCK was a competitive inhibitor of gastrin. Because the C-terminal pentapeptide of CCK is identical with the active C-terminal pentapeptide of gastrin/ 7 it is reasonable to assume that CCK can occupy and compete for the gastrin receptor site on the parietal cell. We reasoned that, as a weak stimulant, CCK would increase secretion when given alone. When given in combination with gastrin, however, it would occupy receptors which would normally be available to the more potent gastrin molecules, producing inhibition of secretion. In the presence of histamine, the stimulatory effects of CCK would merely be added to those of histamine. 52 A recent dose-response study of the inhibition of gastric secretion by CCK leaves little doubt that CCK is a potent competitive inhibitor of the gastrin mechanism for the stimulation of acid secretion. 53 For near maximal rates of pentagastrin-stimulated acid secretion, about 0.5 unit per kg-hr of CCK was required for 50% inhibition and 2.0 units per kg-hr produced maximal (95%) inhibition. 53 Since the dose of CCK needed for halfmaximal pancreatic enzyme response is about 3 units per kg_hr,54 CCK, like secretin, is a more powerful gastrin antagonist than it is an agonist of pancreatic secre-

7 126 PROGRESS IN GASTROENTEROLOGY Vol. 60, No. 1 tion. Analysis of the dose-response curves in this studl 3 showed that CCK had no effect on the maximal response elicited by gastrin, but more than doubled the dose required to elicit that response. These are the characteristics of competitive inhibition and, therefore, the explanation given by Stening et a1. 52 for the various actions of CCK is probably correct. As with secretin, CCK produces its strongest inhibition of acid secretion in the dog. CCK inhibited the maximal response to pentagastrin in the gastric fistula rat but prolonged the over-all response. 33 In cats with gastric fistulas, Way and Grossman (unpublished data) showed that pentagastrin and CCK given separately produced the same maximal response. Both being full agonists, they would not be expected to inhibit each other. Over a wide range of doses CCK augmented pentagastrin-stimulated secretion; inhibition of pentagastrin by CCK was not observed. Wormsley55 found that CCK augmented the basal acid output in 10 of 16 human subjects, so it appears that CCK is a weak agonist in man as well as in the dog when it is administered by itself. In a carefully controlled study employing 5 human subjects in which duodenal reflux was prevented, Brooks and Grossman 34 found that a continuous intravenous infusion of 4 units per kg-hr of CCK inhibited pentagastrin (4.0 /-Lg per kg-hr) and stimulated secretion 33%. CCK has not yet been synthesized, but in the same study34 a pure preparation of CCK (Jorpes and Mutt) inhibited secretion in the single subject to which it was administered. Chey et a1. 37 also observed inhibition of pentagastrin stimulated acid secretion by CCK in human subjects. Fuller dose-response studies in man are needed to estimate the relative susceptibility of man and dog to inhibition by CCK. Since a synthetic preparation of CCK is not yet available, it cannot be stated that all actions of CCK extract are a property of the hormone itself. However, seven of the eight C-terminal amino acids of caerulein, a decapeptide isolated from the skin of an Australian frog, have been shown to be identical with those of CCK. 56 In addition, this substance has powerful cholecystokinetic properties in both its natural and synthetic forms. 54,57 Caerulein, like CCK, stimulated acid secretion when given alone and duplicated the inhibitory effects of CCK when administered with gastrin and the augmentary effects with histamine. 52 In addition, Brooks et a1. 58 have recently reported that caerulein or the C-terminal heptapeptide amide of CCK inhibited pentagastrin-stimulated secretion in man. That the effects of CCK extracts are mimicked by pure and synthetic preparations of ~ h structurally e similar caerulein and,y fragments of CCK itself is evidence ' I':it these actions of CCK preparations a due to the hormone itself and not to an impurity in the extract. Structure-activity relationships. Three naturally occurring peptides, gastrin, CCK, and caerulein, share the C-terminal amino acid sequence Gly-Trp-Met-Asp Phe-NH 2 They also share several physiological actions, but the relative potency of the compounds on a specific physiological effect depends on the rest of the molecule. 54, 57 Gastrin contains a tyrosyl residue at location 6 (for present purposes, it is simpler to number amino acid locations beginning with 1 from the C-terminal end in contrast with the conventional system of numbering from the N terminus) which mayor may not be sulfated in the natural state. Sulfation of the tyrosine in gastrin does not appear to alter its physiological activity in most test systems. 9, 54, 57 CCK and caerulein are more closely related to each other than to gastrin, for, in both, the tyrosyl residue occupies position 7 from the C terminus. A preliminary report indicated that desulfated caerulein is much less potent than its naturally occurring counterpart. 18 A detailed comparison of the relative effects of natural and desulfated caerulein showed that natural caerulein was 4 to 160 times more potent depending on the physiological action being studied. fi9 Also, desulfation of the C-terminal octapeptide of CCK resulted in.a 250-fold decrease in ability to contract the gallbladders of con-

8 January 1971 PROGRESS IN GASTROENTEROLOGY 127 scious cats. For gallbladder contraction and stimulation of pancreatic enzyme secretion sulfation of caerulein reduces the 050 but has no effect on calculated maximal response. By contrast, for gastric acid secretion, sulfation not only decreases 0 50 but it also markedly decreases calculated maximal response. The ability to behave as a competitive inhibitor of gastrin depends only on the change in "efficacy" (of which maximal response is an index) and not at all on the change in 050. It has also recently been shown that desulfation renders caerulein ineffective as an inhibitor of pentagastrin-stimulated acid secretion in both Pavlov and Heidenhain pouch dogs. 60 It can be assumed that de sulfated CCK probably would not inhibit gastrinstimulated secretion. CCK and caerulein contain the C-terminal tetrapeptide amide of gastrin. This is the minimal fragment of gastrin that produces all of the biological actions of the whole molecule. 61 Based on the effects produced by de sulfating caerulein and CCK fragments, the minimal fragment of these peptides includes the sulfated tyrosyl residue at position 7. In this regard it is of interest that, unlike gastrin, neither desulfated CCK nor desulfated caerulein has been reported to occur naturally. Unlike gastrin and CCK, secretin has failed to yield a portion of the molecule with biological activity. All 27 amino acids must be linked before any activity whatsoever is present. 62 Bodanszky et a1. 63 have recently concluded from studies of circular dichroism and optical rotatory dispersion of natural and synthetic secretin that the hormone exists in a helical tertiary structure. Therefore, if the helical conformation is required for physiological activity and if both ends of the molecule must be intact to form the helix, it is understandable why the entire molecule is required for activity. Inhibition by Endogenous Hormones By discussing the effects of exogenous secretin and CCK first, we have approached the subject of duodenal inhibition of gastric secretion in reverse order of historical chronology. Investigation in this area began with observations that certain constituents of chyme, especially acid and fat, inhibited gastric secretion when placed in the duodenum. This action was attributed to the release of enterogastrone. While secretin and CCK fulfill the requirements implicit in the meaning of the word enterogastrone, whether they are released in sufficient amounts after feeding to exert this action and whether enterogastrones other than these two hormones exist are uncertain. Acid in the Intestine Effect of intestinal acidification on gastric secretion. The first demonstration that intestinal acidification inhibited the gastric secretory response to a meal was carried out by Sokolov,2 one of Pavlov's pupils, in This phenomenon was, in keeping with tradition at that time, ascribed to a neural mechanism. Inhibition by acid has been documented many times since. Pincus et a1. 64 found that the degree of inhibition of the Pavlov pouch response to a meal depended upon the ph of the duodenal contents. Code and Watkinson 65 found that infusion of acid into the duodenum inhibited the response to- a meal from Pavlov pouches but not from vagally denervated pouches. There are a few other reports that the acid response from vagally innervated stomach is more susceptible to inhibition than that from vagally denervated pouches. Wormsley and Grossman,26 for example, found that, with histamine stimulation, secretin slightly inhibited secretion from the main stomach but had little or no effect on the Heidenhain pouch. These few results indicating that under some circumstances inhibition is dependent on vagal innervation are largely unexplained. Whether a vagal mechanism of inhibition exists in addition to the humoral one (see below) or whether vagal factors may modulate the humoral mechanism is an open question. However, it is well established that acid in the duodenum causes inhibition of gastric secretion via a humoral mechanism. Using dogs with either vagally innervated

9 128 PROGRESS IN GASTROENTEROLOGY Vol. 60, No. 1 or vagally denervated gastric pouches, Andersson found that acidification of the excluded duodenum inhibited secretory responses from both types of pouches during fasting,66 a meal,67 and gastrin stimulation. 68 Many studies have confirmed that duodenal acidification inhibits the denervated pouch response to exogenous gastrin. 26, 27, 28, 69 Therefore, results showing that secretion stimulated by endogenous gastrin is strongly inhibited by duodenal acidification cannot be considered as evidence for a mechanism acting solely by suppression of gastrin release. Acid in the duodenum releases a hormone which inhibits gastric secretion at some point after the release of gastrin. Whether suppression of gastrin release also occurs is uncertain. Using histamine as a stimulant some investigators have found significant inhibition of gastric secretion after acidifying the duodenum. 26, 65, 70 However, in each case, inhibition varied and depended on the dose of histamine (low doses were inhibited), innervation of the gastric pouch, or the individual animal. In 1960, Andersson 71 investigated this problem and found "slight and irregular inhibition in some dogs, but no inhibition in others." Several extraneous factors may be operating to modify secretion during histamine infusion. The most obvious of these, endogenous gastrin, would potentiate a histamine response and be inhibited by duodenal acidification. Varying degrees of antral acidification in turn would account for different results from different animals or 1 animal on different days. With this problem in mind Andersson and Grossman 72 found that acidification of a vagally innervated antral pouch reduced the Heidenhain pouch response to histamine. That this effect was due to a suppression of gastrin release rather than the alleged "antral chalone" was shown by vagally denervating the antral pouch and demonstrating that the histamine response could no longer be inhibited. Endogenous gastrin released from the neutral innervated antrum, therefore, augments the stimulation by histamine; it is thus probably only this augmentation by endogenous gastrin that is inhibited by duodenal acidification during the infusion of histamine. Sircus 70 was unable to inhibit histamine-stimulated secretion by duodenal acidification in antrectomized Heidenhain pouch dogs. Even when precautions are not taken to exclude endogenous gastrin most investigators have reported that the duodenal mechanism triggered by acid does not inhibit a histamine response. 28, 69, 71 In the dog, therefore, histamine-stimulated secretion is relatively immune to inhibition by duodenal acidification just as it is to inhibition by secretin and CCK. Duodenal acidification inhibits gastrinstimulated Heidenhain pouch secretion in cats, but the effect is much less pronounced than in the dog. 30 Kontm '- et al. 3 1 found only 50% inhibition by oth secretin and duodenal acid infusion ' ~ a gastrin stimulus in the cat. In the.ne study, gastric secretion stimulated by histamine could not be inhibited by either acid or secretin. As early as 1942, Shay et al. 73 reported that duodenal acidification inhibited the acid response to a meal in man. This result was later confirmed and expanded by Johnston and Duthie. 42,74 They showed that duodenal acidification inhibited acid responses to both histamine and gastrin. These results of Johnston and Duthie could be explained by an inhibitory reflex originating from the duodenum, or the release of endogenous gastrin which would add to the histamine acid response and be inhibited by acid in the duodenum. However, they were unable to inhibit histamine with either secretin or CCK in the same patients. Johnston and Duthie 42 also apparently ruled out a reflex mechanism as the sole explanation for the phenomenon, for they were able to inhibit histamine-stimulated secretion in another group of patients by injecting blood taken from the original group during duodenal acidification at the time of the most pronounced inhibition. Therefore, they interpreted their results to indicate that inhibition was due to a hormone other than

10 JaTUlary 1971 PROGRESS IN GASTROENTEROLOGY 129 secretin or CCK. The use of transfusions to demonstrate humoral mechanisms is fraught with technical difficulties. Confirmation of these results by other methods is needed. Chey et al. 75 have recently demonstrated in the rat that the intraduodenal instillation of hydrochloric acid inhibited spontaneous secretion and secretion stimulated by a continuous infusion of pentagastrin. An acid-sensitive mechanism for the inhibition of gastric secretion appears, then, to exist in the duodenum of all species in which it has been sought. Location of the Hormone Released by Acid. Evidence that the hormone responsible for inhibition of gastric secretion during intestinal acidification originates in the duodenum comes from several types of experiments. Surgical resection 76, 77 or translocation 76 of the duodenum from the path of acid results in increased Heidenhain pouch secretion. Removal of acid-secreting mucosa from the stomach of a dog has been shown to increase acid output from a vagally denervated pouch.78 Presumably this procedure decreased the amount of duodenal acidification, resulting in decreased release of the humoral inhibitor. It has also been shown that acid perfusion of an isolated pouch of the duodenal bulb will effectively inhibit gastric secretion. 79, 80 Most investigators agree that the mechanism for acid inhibition of gastric secretion is primarily confined to the duodenum. There is, however, slight disagreement over the length of duodenum involved. Andersson et al. 69 were able to inhibit gastric secretion by acidifying pouches of proximal duodenum but not those made from distal duodenum. Kontruek and Grossman,80 however, obtained evidence for the involvement of the entire duodenum. Mter excision of the duodenal bulb, acidification of the remaining duodenum still resulted in inhibition. No inhibition was recorded after removal of the entire duodenum. The same investigators studied this question by irrigating loops from various portions of the intestine with solutions known to trigger the inhibition of gastric secretion. 8! Irrigation of a loop of proximal duodenum with HCI caused 75% inhibition of gastric secretion stimulated by a maximal dose of gastrin. Irrigation of a distal duodenal loop caused 40% inhibition, whereas irrigation of an ileal loop had no effect and a jejunal loop resulted in a 40% increase in gastric secretion. 8! Regardless of conflicting results concerning an inhibitory role for the distal duodenum all studies were in agreement that acidification of the duodenal bulb produced the greatest decrease in gastric secretion. Teleologically one would expect this to be the area for the receptors to be most abundant since it is the only portion of the gut exposed.. " gastric acid uncontaminated by ne\. ralizing secretions. Meyer et al. 82 have,. )cently shown that the pancreatic bice )onate response to intestinal acidification was as great when acid was introduced into the jejunum (45 cm from the pylorus, all reflux being prevented) as when put into the duodenal bulb. We would predict that under the same circumstances inhibition of gastric secretion would also occur. There are probably two factors responsible for these apparent contradictions when studies on pancreatic stimulation are compared with those on gastric inhibition. First, closing the gastric fistula may not provide enough endogenous acid to release maximal amounts of secretin from the more refractory distal gut. Second, long lengths of lower intestine may be required-hence the inability to show an effect by perfusing Thiry-Vella loops of jejunum and distal duodenum. Identification of the hormone released by acid. From the information currently available there are two hormones which are candidates for the role of the enterogastrone released by acid. These are secretin and CCK. A third possibility which cannot be ruled out is the existence of another enterogastrone separate and distinct from secretin and CCK. 4! A confusing, but plausible situation would be the release of a combination of these hormones by duodenal acidification. Since the immunoassayists have not yet applied

11 130 PROGRESS IN GASTROENTEROLOGY Vol. 60. No.1 their skills to this problem, our conclusions, for the present, must be based on comparing the various gastrointestinal effects of secretin and CCK with those of duodenal acidification. The classical work of Bayliss and Starling S3 demonstrated that duodenal acidification released a hormone which stimulated a copious flow of pancreatic juice rich in bicarbonate. This hormone was extracted and named secretin. It is now established that the rate of pancreatic bicarbonate output is directly related to the rate of introduction of acid into the intestine. 82, 84 Secretin is the only substance in the body which is a strong stimulant of pancreatic volume flow and bicarbonate output. The others, CCK, gastrin, and acetylcholine, are weak stimulants of flow and strong stimulants of enzyme output. 85 Secretin weakly stimulates enzyme secretion. s5 Pascal et a1. 86 studied bicarbonate and volume flow of the canine pancreas in response to duodenal acidification and exogenous secretin. The two were found to produce identical results within the physiological dose range. This result has recently been confirmed by one of US. 87 Although CCK acting alone is only a weak stimulant of volume and bicarbonate, it potentiates the pancreatic response to secretin. s8 If substantial amounts of endogenous CCK were released by duodenal acidification, the maximal pancreatic response should be greater than that produced by secretin alone, whereas observation shows them to be essentially equal. Therefore, there is no evidence that the pancreatic volume-bicarbonate response to duodenal acidification is mediated by CCK or any hormone other than secretin. A study comparing the effects of different doses of secretin with those of different doses of acid in the duodenum on both the aqueous and enzymatic components of pancreatic secretion would be of considerable help in deciding whether or not there is significant release of CCK during duodenal acidification. It should be mentioned that Preshaw et al. 84 found evidence that at high rates of duodenal acid infusion pancreatic enzyme output was increased over that stimulated by exogenous secretin and concluded that under these conditions acid released CCK. Unfortunately, enzyme output was determined only for a single low dose of exogenous secretin which gave a response higher than that produced by a moderate amount of duodenal acidification. Recent unpublished studies of Meyer and Grossman support the view that high rates of acid infusion release small amounts of CCK. In 1968, Johnson and Grossman 28 found that continuous intravenous infusions of a dose of secretin submaximal for pancreatic secretion could completely account for all of the inhibition of secretion seen from the canine Heidenhain pouch during duodenal acidification. We concluded, first, that secretin was a physiological inhibitor of gastric acid secretion in the dog and, second, that secretin was probably the only enterogastrone which was released from the dog duodenum in significant quantities by the presence of acid. 28 Several recent findings substantiate the second part of this conclusion. Secretin has been shown to inhibit gastrin-stimulated secretion t ~ r oa unon competitive mechanism. 29 CCK, on the g h other hand, is a competitive inhibitor of gastrin-stimulated secretion. 53 The kinetics of the inhibition of secretion produced by duodenal acidification are noncompetitive and identical with those of secretin. 89 If the inhibition of gastric secretion caused by duodenal acidification is mediated by secretin, there should be a relationship between the rate of pancreatic secretion and the degree of inhibition of gastric secretion caused by a given amount of acid in the duodenum. Using dogs with chronic pancreatic and gastric fistulas, Preshaw 90 simultaneously studied both types of secretion in response to duodenal acidification during a continuous background infusion of gastrin. He concluded that the data s\lpported the hypothesis that both the pancreatic stimulatory and gastric inhibitory mechanisms activated by duodenal acidification are mediated by the hormone secretin. 90 Pure natural secretin, synthetic secre-

12 January 1971 PROGRESS IN GASTROENTEROLOGY 131 tin, and duodenal acidification stimulate pepsin secretion in the dog and cat.30 In a similar study in man, Brooks et a1. 91 demonstrated that intravenous injection of secretin and perfusion of the duodenum with 100 mm HCl strongly stimulated pepsin secretion. Magee and his co-workers 92 have recently reported that, while irrigation of isolated duodenal pouches with citrate buffer at ph 3.0 and 5.0 stimulated pepsin secretion from the Heidenhain pouch, introduction of a solution at ph 1.0 inhibited pepsin output. They also found that CCK inhibited the pepsin response to secretin and on this basis have concluded that at high levels of duodenal acidification the major hormone released is CCK. This result is puzzling and difficult to explain considering other results from similar studies. In the cat, closing the gastric fistula during near maximal stimulation of gastric secretion, and thereby diverting acid into the duodenum, caused a 400% increase in pepsin secretion from the Heidenhain pouch.31 Irrigation of the human duodenum with 400 ml of 100 mn HCl per hr increased gastric pepsin outputs 300%.91 We have recently found that irrigation of the intact dog duodenum with 160 mn HCl at a rate of 160 ml per hr results in a 3- to 6-fold stimulation of pepsin output (Harrison and Johnson, unpublished results). The amount of acid entering the duodenum in these cases equals and probably exceeds the maximum entering under physiological conditions. The best evidence that duodenal acidification can release CCK is derived from early studies showing that acid in the gut causes gallbladder contraction. 93 Secretin when given alone does not contract the gallbladder, although it potentiates the actions of a small dose of CCK or gastrin. 94 Although Stening and Grossman 94 showed that duodenal acidification enhanced the gallbladder response to gastrin, they did not test the effects of acid in the duodenum in the absence of other stimuli. Using dogs with Heidenhain pouches and gastric and pancreatic fistulas, Way and Grossman 95 studied the effects of acid diverted into the duodenum on pancreatic and gastric secretion. They concluded that their results strongly support the concept that the inhibition of acid secretion by duodenal acidification is largely due to the release of secretin. However, they also found evidence for the release of another hormone which could contribute to the inhibition by acid. During periods of equal pancreatic secretion duodenal acidification caused greater inhibition than did exogenous secretin; the difference was small but statistically significant. This mechanism was ascribed to the duodenal bulb, for the difference disappeared after excision of the proximal 5 cm of duodenum. 95 There is some evidence that acidification may release small amounts of CCK, but, considering the large amount of secretin released, there is at present no evidence that the CCK plays a significant role in the inhibition of gastric secretion. On the contrary, there is some evidence that CCK may not be an effective inhibitor of acid secretion in all species. In the cat, porcine CCK (the natural CCK currently available) does not inhibit gastrinstimulated secretion and is a full agonist when administered alone (Way and Grossman, unpublished data). In the.rat, a single subcutaneous injection of CCK inhibited but prolonged the secretory response to an injection of pentagastrin,33 and, when given as a continuous infusion, CCK stimulated but did not inhibit gastrinstimulated secretion. 75 The presence of amino acids in the canine duodenum provokes a pancreatic response typical of CCK, but does not inhibit gastrin-stimulated secretion (Spingola and Grossman, unpublished results). In conclusion, recent studies show: (a) that stimulation of pancreatic secretion and inhibition of gastric secretion are both quantitatively related to amount of duodenal acidification,90 (b) that acid in the intestine and secretin both stimulate pepsin secretion,31.91 and (c) that secretin 29 and acid in the duodenum 89 both inhibit gastrin-stimulated secretion via a noncompetitive mechanism. Therefore, there is little reason to depart from our 1968 conclusion that secretin is the

13 132 PROGRESS IN GASTROENTEROLOGY Vol. 60, No.1 only enterogastrone of physiological significance released in dog by acidification of the intact duodenum. 28 However, the strong inhibition produced by acidification of just the duodenal bulb has not yet been fully studied to determine whether it has these characteristics of secretin release just enumerated (noncompetitive kinetics, stimulation of pepsin secretion, parallel stimulation of pancreatic flow with little stimulation of pancreatic enzymes). Andersson et al. 96 found that acidification of a pouch of the duodenal bulb inhibited pentagastrin-stimulated acid secretion from Heidenhain pouches without stimulating flow from a pancreatic fistula in the same animal. These workers 96 have extracted from the mucosa of the duodenal bulb of pigs a substance that inhibits pentagastrin-stimulated gastric acid secretion without stimulating pancreatic secretion; they call the active principle of this extract "bulbogastrone." Significance of inhibition produced by duodenal acidification. To assign a physiological role to the facts that acid releases secretin from the duodenum and secretin inhibits gastric secretion, one must demonstrate that amounts of acid sufficient to release secretin normally occur within the duodenum. From detailed studies of pancreatic secretion during intestinal acidification Meyer et al. 82 have proposed a model in which the rate of secretin release is a function of the length of intestine acidified below a critical ph. They found that secretin release was independent of ph below ph 3, depending instead on the rate of entry of acid into the intestine. Above ph 3, secretin release decreased until at ph 4.5 to 5.0 a threshold occurred above which secretin could not be released. 82 Studies of duodenal ph required for inhibition of gastric secretion correlate well with those required to stimulate pancreatic secretion. An intrabulbar ph of 3.1 to 4.1 results in inhibition of the acid response to exogenous gastrin 97 and sham feeding. 98 A quantitative study of the effects of various doses of exogenous secretin on near maximal gastrin-stimulated acid secretion revealed that the threshold dose of secretin for inhibition was approximately 0.06 unit per kg_hr. 29 Approximately 5 times this dose is needed to stimulate pancreatic secretion. 82 Therefore acidification sufficient to stimulate the pancreas should also inhibit the stomach. A study of the effects of acid on gastric inhibition similar to the one that Meyer et al. 82 did on pancreatic stimulation would add important new information for resolving this problem. The definitive experiment of measuring duodenal ph during feeding has recently been done by Brooks and Grossman. 99 They recorded the ph of the contents of the duodenal bulb and the mid-duodenum continuously during the response to a meat meal in dogs. The bulbar ph fell from 5 to 3.5, 30 min after feeding and remained below 4 for 3.5 hr. The ph of the mid-duodenum never fell below 5. Infusion of 3 meq of titratable acid per 15 min at ph 3.5 into the canine duodenum increased pancreatic bicarbonate output to approximately half of the maximum obtained with secretin, and infusion of only 2 meq of ph 3.5 buffer per 15 min stimulated the pancreas to about one-quarter of its capacity.82 Therefore, assuming that the volume of fluid passing the pylorus was sufficient in Brooks and Grossman's study,99 the ph of the duodenal contents was low enough to release secretin. It should be emphasized, however, that it is only in the duodenal bulb that the ph dropped sufficiently to release enough secretin to inhibit acid secretion. As pointed out above, studies 96 of the simultaneous effects of duodenal acidification on pancreatic and gastric secretion in which the acid was confined to the duodenal bulb suggested that secretin was not the humoral agent. The important question of the significance of,secretin release for the inhibition of gastric secretion will not be entirely settled until reliable estimates of the amount of secretin released bya meal are available. However, from the information currently availab.le the release of se-