Canine Stomach Research Articles

Cell proliferation and cell loss in intramucosal signet ring cell carcinoma of canine stomachs induced by N-ethyl-N'-nitro-N-nitrosoguanidine.

Signet ring cell carcinoma was induced in canine stomachs by N-ethyl-N'-nitro-N-nitrosoguanidine, and modes of cell proliferation and turnover in the carcinoma were studied by 3H-thymidine autoradiography in conjunction with morphometric analysis. From 2 to 15 months after the cessation of 8 months carcinogen treatment, carcinomas in an early stage were obtained. Most of the cancer tissues confined to the lamina propria showed a layered structure. This comprised three layers; the superficial and the deep layer were composed of signet ring cells, and the middle layer was composed of small round cells. The dogs were labeled with 3H-thymidine by s.c. injection and by local infusion of the celiac artery. Flash-labeled autoradiographs revealed that most 3H-thymidine incorporating cancer cells were located around the middle layer, with a small amount of mucin. Using a pulse labeling experiment, those labeled carcinoma cells were shown to migrate from the middle layer towards the surface. Morphometric analysis of the autoradiographs showed that the small cells in the middle layer migrated upwards and produced mucin to become full-blown signet ring cells by 5.5 days. In 15 days, most labeled cancer cells in the superficial layer had disappeared. This mode of cellular turnover appeared to mimic a cell renewal system of the normal gastric mucosa. If the cancer cells turn over in this way, the tumor must grow slowly, remaining as an intramucosal cancer for a relatively long period.

Canine gastric emptying of polycarbophil: An indigestible, particulate substance

We tested whether indigestible solids could empty from the canine stomach independently of gastric burst motor activity. Test meals contained polycarbophil, an indigestible particulate (1–3 mm diameter) substance. Test meals were slurries of 30 or 90 g of radiolabeled polycarbophil particles in saline. Meals were administered via an oral gastric tube. After 4 h, the stomach was drained via a gastric cannula, and the percentage of meal that had emptied into the duodenum was calculated. Antroduodenal motor responses to the meals were monitored with strain-gage force transducers. The motor responses to polycarbophil meals were compared with those after canned food and saline meals. Fifty percent of the 90-g polycarbophil meal emptied by 4 h; this occurred independently of gastric burst motor activity. Both polycarbophil and canned food elicited similar indexed motor responses and both delayed the postprandial reappearance of gastric burst motor activity. We conclude that small indigestible particles can stimulate fed state-like motility and empty from the stomach independently of gastric burst activity.

Gastric mucosal barrier: hydrophobicity of stretched stomach lining

Surface hydrophobicity of the luminal lining of the canine stomach has been studied as a very convenient means of following the adsorbed monolayer of surfactant believed to provide the gastric mucosal barrier. Hydrophobicity has been measured as the contact angle (theta) produced when a drop of saline is placed upon the surface. theta was found to decrease from 82 to 62 degrees upon 50% linear extension of samples of oxyntic mucosa from 10 dogs. When the phospholipid believed to cause the hydrophobicity was absorbed to glass slides, the contact angle was found to decrease with lower surface concentration. Thinning or "crazing" of the absorbed surfactant monolayer imparting the very hydrophobic nature of the luminal lining is discussed as a possible reason why ulcers tend to form at the crests of the folds, i.e., at points where the surface has been stretched and the monolayer disrupted.

DIFFERENTIATION OF THE ACTION POTENTIAL IN THE SMOOTH MUSCLE OF CANINE GASTRIC ANTRUM USING CALCIUM‐INHIBITORY DRUGS

Electrical and mechanical activity were recorded simultaneously in smooth muscle preparations from the antrum region of canine stomach by means of a single sucrose gap technique (SGT). The SGT was optimized to permit stable recording from multicellular smooth muscle preparations over several hours of electrical and mechanical activity with little disturbance of their normal properties. Acetylcholine (ACh, 10(-8) to 10(-6) M) induced or augmented dose-dependently the electrical and mechanical activity. The plateau of the action potential complex was elevated by ACh, while the contraction was increased in linear correlation to the magnitude of the plateau component. In spontaneously active (or in ACh-stimulated) preparations TEA (5 to 20 mM) magnified the plateau component, induced or strengthened spikes on the plateau ('secondary spikes'), and induced or strengthened phasic contractions. Nifedipine (10(-6) M) abolished secondary spikes, part of the plateau component of the action potential, and suppressed mechanical activity. The complex action potential of canine gastric antrum can be differentiated into (a) a basic action potential, consisting of an initial, primary spike and a plateau depolarization; this basic action potential is resistant to nifedipine and does not trigger any mechanical activity; and (b) a nifedipine-sensitive component (calcium component), which consists of an augmentation of the plateau depolarization and of secondary spikes, and which is responsible for the initiation of mechanical activity.

ENDOSCOPIC LOCAL INJECTION OF AN ANTICANCER AGENT INTO EXPERIMENTAL CANINE STOMACH CANCER

The emulsionized Methyl-CCNU (Me-CCNU-E) was injected endoscopically into normal canine stomachs. The emulsion was detected in the submucosal layer of the stomach wall and regional lymphnodes histologically by Oil-red-O stain. Using the Me-CCNU concentrated in higher than 20mg/ml, ulceration was made in the gastric mucosa where the emulsion was injected. But in case of the concentration of Me-CCNU was lower than 10mg/ml, no compicated lesion was found.Stomach cancers were induced in two Beagle dogs experimentally by N-ethyl-N'-nitro-N-nitrosoguanidine (ENNG), and Me-CCNU-E was injected directly into the tumors endoscopically.One dog had advanced stomach cancer which was classified into Borrmann type III. Me-CCNU (10mg/ml) was injected into the tumor 9 times within 13 weeks in total dose of Me-CCNU, 130mg.At the 17 weeks after the initial injection, the dog was sacrified and examined histologically. The cancer was not reduced, but degenerative findings were observed histologically in the tumor where the emulsion was injected. Regional lymphnode metastasis was not found. In the other dog, a II a + II c shaped advanced cancer and two II c shaped early cancers had been induced by ENNG. Total dose of 120mg of Me-CCNU-E (20mg/ml) was also injected into those tumors. At the 12 weeks after the treatment, this dog died for a perforation of the stomach.On the histological examination, cancer nests were still remained, however most of the tumors were vanished distinctly.It has been reported that the endoscopic injection of anticancer agents to the stomach cancer is useful method for the treatment of stomach cancer in the cases of high aged patients and poor risk patients. Through this experiment, it was indicated that the treatment of anticancer agents with emulsion was more effective to cure not only the cancer but the lymphnode metastasis.

STUDIES ON ENDOSCOPIC LOCAL INJECTION OF AN ANTICANCER AGENT INTO EXPERIMENTAL CANINE STOMACH CANCER

In order to establish a topical chemotherapy against stomach cancer, the following experiments were performed.1. Adriamycin (ADM) solutions were injected through an endoscope into the submucosal space along the lesser curvature of normal canine stomachs. The dogs were sacrificed at various time intervals after initial injection and the concentration levels of ADM in the tissue were measured for fundamental study. ADM levels in the tissue at the gastric wall where the drug had been injected were one hundred times or more higher than in cases which had received ADM intravenously. At the regional lymph nodes along the lesser curvature, the ADM level was almost equal to that in the gastric wall where the drug had been injected.2. Stomach cancers were induced in 4 dogs by oral administration of N-ethyl-N'-nitro-N-nitrosoguanidine. One milliliter of ADM, 1 mg/ml was injected directly into 3 lesions in 2 dogs endoscopically. Total doses of ADM were 4 to 15 mg against each tumor. Two out of 3 tumors were almost completely reduced by this procedure, but minute cancer nests still remained histologically. One lesion among those treated was not reduced in size, but severe degeneration was observed histologically. No complication has occurred in these two dogs. One milliliter of ADM, 10 mg/ml was injected into 6 lesions in 2 dogs. Total doses were 2.5 to 10 mg against each tumor. Three lesions disappeared completely, and the other three lesions were markedly reduced in size. Three of 6 elevated cancers completely disappeared, but remnant cancers were observed in all of 3 depressed cancers treated by this therapy. Ulceration occurred in 5 out of 6 lesions. Perforation was observed in one of the five ulcerations. No lymph node metastasis was found in any of 4 dogs.It was indicated that endoscopic local injection of ADM against stomach cancer was effective to some extent in curing not only the primary lesion but also its lymph node metastases.

Focal gastric mucosal blood flow at the site of aspirin-induced ulceration

Focal gastric mucosal blood flow as measured by the hydrogen gas clearance method was compared with total gastric blood flow as determined by venous outflow in an isolated segment of canine stomach before, during, and after exposure to aspirin. Despite an increase in total gastric blood flow from 10.8 +/- 1.6 ml/min per chamber to 17.4 +/- 1.9 ml/min per chamber and mucosal blood flow at nonulcerated sites from 29.5 +/- 4.3 ml/min per 100 g to 83 +/- 14.4 ml/min per 100 g, mucosal blood flow at the site of aspirin-induced ulceration was significantly reduced from 29.5 +/- 4.3 ml/min per 100 g to 12.5 +/- 2.5 ml/min per 100 g. After the removal of aspirin, mucosal blood flow returned to control levels. Such a redistribution of mucosal blood flow in response to aspirin is consistent with the localized nature of acute aspirin-induced injury. The findings also explain the inability of previous methods measuring blood flow of the entire stomach to demonstrate mucosal ischemia during such injury.

Effects of calcium antagonists on smooth muscle membranes of the canine stomach

1. 1. In circular muscles of the canine stomach, diltiazem (> 3 × 10 −6 M) depolarized and nicardipine (> 10 −6 M) hyperpolarized the membrane. 2. 2. Diltiazem (> 5 × 10 −6 M) or nicardipine (> 3 × 10 −7 M) inhibited the plateau potential of spontaneously generated or electrically evoked action potential. 3. 3. In Na-deficient solution (between 137 and 30 mM Na), diltiazem and nicardipine selectively inhibited the plateau potential. In Na- (15 mM Na) or Ca-deficient (0.25 mM Ca) solution, spontaneously generated action potential ceased, and only graded responses were evoked by electrical stimulation. These graded responses were inhibited by diltiazem (> 5 × 10 −6 M) or nicardipine (> 3 × 10 −7 M). 4. 4. The ionic contribution for generation of action potential in this muscle cell was discussed in relation to actions of Ca antagonists.

Hemodynamic actions of nicotine on the canine stomach.

Acute effects of nicotine were determined in the circulation of the stomach in which measurements of gastric blood flow, systemic arterial and portal venous blood pressures, and the arteriovenous oxygen content difference were made. From these measurements gastric vascular resistance and oxygen consumption were calculated. Nicotine was infused for 10-min periods in two doses intraarterially (ia) and intravenously (iv) alone and following adrenergic receptor blockade with either propranolol or phentolamine or both agents. With ia nicotine alone, gastric blood flow and oxygen consumption increased. The effect of nicotine on blood flow was enhanced by prior treatment with phentolamine. The effects of nicotine on both blood flow and oxygen consumption were abolished in the presence of combined phentolamine plus propranolol. With iv nicotine alone, gastric blood flow, oxygen consumption, and arterial pressure increased transiently. Phentolamine enhanced the effect of nicotine on oxygen uptake and prevented the pressor response. Propranolol enhanced the pressor effect of nicotine but abolished the blood flow and oxygen consumption responses. Combined treatment with both adrenergic antagonists abolished nicotine effects on blood flow and oxygen consumption. These findings indicate that nicotine alters gastric hemodynamics and oxygen consumption during acute intravascular infusions of the drug. The vasodilator and metabolic effects of nicotine appear to be mediated via beta-adrenergic receptors. These findings provide little basis for implicating a local ischemic mechanism in the ulcerogenic effects of nicotine on the stomach.

Motor effects of gastrin releasing peptide (GRP) and bombesin in the canine stomach and small intestine: Program for study of control of smooth muscle function

Close intraarterial injections of synthetic porcine gastrin releasing peptide (GRP) or bombesin stimulated contractions in the stomach and inhibited ongoing contractile activity in the small intestine of anaesthetized dogs. Contractile activity of the circular muscle was recorded by serosal strain gauges and phasic activity when desired was elicited by local field stimulation or intraarterial motilin injections. In the stomach (corpus and antrum) following tetrodotoxin blockade of field-stimulated contractions, the contractile response to either peptide was not present, suggesting that stimulation of receptors on nerves initiated contractions in the stomach. Similarly, in the small intestine, the inhibitory response was eliminated by tetrodotoxin suggesting a neural receptor. Pre-treatment with reserpine did not alter the inhibitory response, either in the presence or absence of atropine, therefore, adrenergic inhibitory mechanisms did not appear to be involved. The concentration of bombesin producing 50% inhibition of field stimulation (ED 50) was increased following treatment with the putative M 1 muscarinic antagonist, pirenzipine suggesting activation of M 1 cholinergic inhibitory receptors by bombesin. After blockade by atropine of field-stimulated contractions and the contractile response to intraarterial acetylcholine, the ED 50 for bombesin inhibition of motilin contractions was increased. After muscarinic blockade, the residual inhibitory response of GRP/bombesin may involve activation of a neural non-cholinergic non-adrenergic inhibitory mechanism. These results suggest that GRP and bombesin act to alter motility in the dog in vivo by affecting neural activity.

Pressure activity patterns in the canine proximal stomach: response to distension.

The stomach accommodates to distension with a minimal rise in intragastric pressure, but the effects of gastric accommodation on overall gut motor activity are unknown. In four conscious fasted dogs, constant gastric distension (0, 10, 20, 40, 80, 160, 320, and 640 ml) was produced by an air-filled intragastric bag; intrabag and gastrointestinal pressure activities (fundus, antrum, and proximal intestine) were recorded for 4 h. With the bag empty (or without bag), the proximal stomach exhibited a fasting cyclic motor pattern synchronized with that in the antrum and upper intestine. During periods of activity the fundus generated powerful phasic pressure waves. Increasing intrabag volumes up to 80 ml gradually lengthened phase II of the intestinal interdigestive motor cycle. At 160 ml or more, gut cyclic motor activity disappeared and was replaced by continuous irregular activity, although the intrabag pressure increased to less than 7 mmHg. We concluded that gastric distension markedly alters gut motor activity even though gastric accommodation prevents a major increase in intragastric pressure.

Multiple responses to electrical field stimulation in circular muscle of canine gastric corpus.

Innervation of circular muscle of the canine stomach studied in vitro was investigated by subjecting muscle strips to electrical field stimulation. Strips were cut from the lesser curvature of the gastric corpus and stimulated with 10-s trains of 0.5-ms pulses at 0.5-20 Hz, 40 V. Most responses were classified into one of three types. In general, field stimulation tended to elicit sequences of varying magnitudes of transient on-contraction, on-relaxation, off-relaxation, off-contraction. Responses were abolished by tetrodotoxin. On-contraction was almost abolished by atropine plus desensitization by 5-hydroxytryptamine (5-HT) or substance P. On-relaxation and off-relaxation were not affected by adrenergic blockade, methysergide, apamin, or 4-aminopyridine. ATP usually caused contraction and slightly diminished relaxation to field stimulation. Vasoactive intestinal polypeptide (VIP) had little effect on tone and response to field stimulation. Relaxation disappeared after scorpion venom treatment. This probably resulted from depletion of the transmitter which mediates relaxation. Off-contraction was reduced by atropine, desensitization by 5-HT or substance P, cromoglycate, indomethacin or ATP, but was not affected by adrenergic blockade, hexamethonium, methysergide, mepyramine, or VIP. The findings suggest that innervation of gastric corpus circular muscle included excitatory cholinergic and both excitatory and inhibitory noncholinergic, nonadrenergic innervation. However, the responses of circular muscle to field stimulation in vitro were drastically different from those obtained previously in vivo, suggesting damage or altered inputs to circular muscle when strips of circular muscle are studied.

Gastric influences on canine small intestinal myoelectric activity.

The aim was to determine whether the canine stomach modulates interdigestive and digestive small intestinal myoelectric activity. In four conscious dogs with electrodes implanted chronically on the duodenum and jejunum, enteric myoelectric activity was recorded during fasting and after feeding a 200-g liver meal. The dogs then underwent total gastrectomy and esophagoduodenostomy, after which they were restudied. Gastrectomy did not alter the pattern of the enteric interdigestive myoelectric complexes; the occurrence, period, duration of phase III, and consistency of distal propagation of enteric IMCs remained unchanged. Gastrectomy also preserved both the postcibal inhibition of the enteric IMCs and the postcibal induction of the fed myoelectric pattern in the small bowel. However, the onset of the fed pattern occurred more promptly after gastrectomy than before gastrectomy. We concluded that the pattern of canine duodenal and jejunal interdigestive myoelectric activity was largely independent of the stomach. The more rapid onset of the fed pattern with feeding postgastrectomy may relate to more rapid entry of chyme into the small intestine.

Inhibitory effect of dopamine on canine gastric fundus.

The mechanism of the inhibitory effect of dopamine on canine stomach fundus was studied in longitudinal and circular muscle fundus strips, contracted by transmural electrical stimulation or by methacholine. Results obtained for longitudinal and circular strips were similar. Dopamine (1 X 10(-6)-1 X 10(-4) M) concentration-dependently inhibited frequency-response curves to electrical stimulation; these concentrations did not change the resting tone of the strips. Dopamine (1 X 10(-4) M), tested on contractions of similar amplitude induced in the same strips by electrical stimulation at 0.5 Hz and by methacholine, inhibited the electrically induced contractions but had little influence on the contractions induced by methacholine. The inhibition of the electrically induced contractions by dopamine 1 X 10(-4) M was not influenced by the presence of cocaine 3 X 10(-5) M or hydrocortisone 3 X 10(-5) M. The alpha 1- and alpha 2-adrenoceptor antagonist phentolamine and the alpha 2-adrenoceptor antagonist rauwolscine markedly antagonized the inhibitory effect of dopamine on the response to electrical stimulation at 0.5 Hz. The alpha 1-adrenoceptor antagonist prazosin and the dopamine receptor antagonists haloperidol and domperidone had no effect. The dopamine receptor antagonist metoclopramide decreased the inhibitory effect of dopamine but had a similar effect on the inhibition caused by noradrenaline. These results indicate that the inhibitory effect of dopamine in the dog gastric fundus is mainly mediated by an interaction with alpha 2-adrenoceptors on the intramural cholinergic neurons; this effect is largely direct since it was not influenced by cocaine.(ABSTRACT TRUNCATED AT 250 WORDS)

A possible explanation for effects of Sr 2+ on contraction-relaxation cycle in canine stomach

1. 1. Effects of Sr 2+ on smooth muscle activity and Ca 2+ movement by plasma membrane (PM) were examined in preparations isolated from canine stomach to determine how Sr 2+ modulates the contraction-relaxation cycle in smooth muscle. 2. 2. Sr 2+ caused contraction in Ca 2+-free solution. 3. 3. In the presence of Sr 2+ KCl caused a response similar to that in normal Ringer's solution, this contraction was inhibited by both nitroprusside and verapamil. 4. 4. Sr 2+ inhibited Ca 2+ uptake by PM in the presence and absence of ATP. 5. 5. Release of Ca 2+ from loaded PM vesicles was accelerated by Sr 2+. 6. 6. Results suggest that Sr 2+ could replace Ca 2+ in Ca 2+ binding sites in the cell.

Measurement of regional gastric mucosal blood flow by hydrogen gas clearance

Gastric mucosal blood flow as measured by the hydrogen gas clearance method was compared with total gastric blood flow as determined by venous outflow in an isolated segment of canine stomach. During rest and histamine stimulation, hydrogen gas clearance correlated lineally with total gastric blood flow (p < 0.001). Correlation analysis revealed a slope of 1.21 and a correlation coefficient of 0.97. The slope of 1.21 indicates that the ratio of mucosal to total gastric blood flow was approximately 82 percent. This fractional distribution of blood flow is in agreement with that previously reported by microsphere technique. In the intact stomach of anesthetized rats and dogs, antral mucosal blood flow was significantly higher than that of the corpus. This finding may, in part, contribute to the observation that acute stress erosion occurs predominantly in the corpus of the stomach. Histamine stimulation selectively increased the mucosal blood flow of the corpus, whereas antral mucosal blood flow remained unchanged. As a result, there was no significant difference in mucosal blood flow between the antrum and the corpus during histamine stimulation.

Regulation of pepsinogen release from canine chief cells in primary monolayer culture

To study the regulation of pepsinogen secretion by chief cells, we have developed techniques for the isolation, enrichment, and short-term culture of chief cells from canine stomach. The fundic mucosa was enzyme dispersed and chief cells were enriched to a content of about 70% using an elutriator rotor. After 36 h in culture confluent monolayers formed that were highly enriched in chief cells. Carbachol induced a time-dependent release of pepsinogen into the medium, with about a threefold increase in pepsinogen secretion over controls found after 60 min of incubation. Carbachol stimulation of pepsinogen secretion was dose dependent, with 5 microM producing 50% of the maximal response found at a carbachol concentration of 100 microM. Atropine (100 microM) produced a rightward shift of the dose-response curve, indicating the presence of a muscarinic receptor. Dibutyryl cAMP, 8-bromo-cAMP, and forskolin also markedly stimulated pepsinogen secretion. Secretin and vasoactive intestinal peptide (VIP) stimulated pepsinogen secretion, but the response were of smaller magnitude than found with carbachol or the cAMP analogues. The phosphodiesterase inhibitor isobutylmethylxanthine also caused a small stimulation of pepsinogen secretion but did not enhance the response to secretin or VIP. These findings indicate that epithelial monolayers can spontaneously form from isolated canine chief cells and retain functional differentiation evident by a response to stimulation. Canine chief cells in culture possess muscarinic and secretin receptors and respond to cAMP.

Pentagastrin potentiates nonadrenergic inhibitory neuromuscular transmission in orad stomach of the dog.

Mechanical and intracellular electrical activities were recorded separately from circular muscle of the orad corpus of the canine stomach. In normal Krebs solution, transmural electrical nerve stimulation produced excitatory and inhibitory responses. The excitatory response was blocked by atropine; the inhibitory response was unaffected by phentolamine or propranolol alone or in combination. Both responses were blocked by tetrodotoxin. In the presence of pentagastrin, the inhibitory response to electrical nerve stimulation was potentiated in a dose-dependent manner. The threshold concentration for this potentiating effect was 3 X 10(-11) M. Intracellular recordings showed that hyperpolarization of the resting membrane potential in response to inhibitory nerve stimulation was greater in the presence of pentagastrin than in its absence. These data suggest that, in addition to a direct stimulatory effect on the muscle, pentagastrin also potentiates the effect of ongoing activity of intramural inhibitory nerves. This latter effect is only apparent when the inhibitory neural network is active. Thus, the in vivo effect of pentagastrin on motor activity of the corpus will depend on the degree of ongoing activity of intramural inhibitory nerves.

Relationship between acid secretion and blood flow in the canine stomach: Role of oxygen consumption

The relationship of acid secretion to blood flow and oxygen consumption was investigated in the isolated canine stomach during intraarterial infusion of pentagastrin. Blood flows above and below control values were achieved by pump perfusion and partial arterial occlusion, respectively. When blood flow was increased from a control value of 38 ml · min−1 · 100 g−1 to 85 ml · min−1 · 100 g−1, both acid secretion and oxygen consumption remained constant at 0.08 mEq · min−1 · 100 g−1 and 2.1 ml · min−1 · 100 g−1, respectively. A reduction in blood flow to 9.0 ml · min−1 · 100 g−1 caused acid secretion and oxygen consumption to decrease to 26% and 44% of control values, respectively. The results indicate that acid secretion is independent of blood flow at flow rates above 38 ml · min−1 · 100 g−1. A model of the interrelationships between blood flow, oxygen consumption, and acid secretion is proposed that may explain much of the conflict in the literature concerning these three parameters.

Prostaglandin-induced bicarbonate secretion in the canine stomach: Characteristics and evidence for a cholinergic mechanism

Using a canine chambered stomach preparation, the effect of topical 16,16-dimethyl prostaglandin E 2 (dmPGE 2) in neutral solution (150 m N NaCl) on gastric mucosal bicarbonate (HCO 3 −) secretion was assessed. Compared to control studies in which neutral solution alone bathed the epithelium, dmPGE 2 (0.5, 1.0, 2.0, and 4.0 μg/ml), when applied to gastric mucosa, significantly increased the output of gastric HCO 3 − in a stepwise and dose-related fashion. Accompanying these effects in HCO 3 − output were similar increases in sodium, potassium, chloride, and gastric perfusate volume. In other studies, the effects of intravenous atropine and close intraarterial tetrodotoxin on this PG-induced HCO 3 − secretion were evaluated. Both agents completely prevented the stimulation of HCO 3 − output induced by dmPGE 2 (2 μg/ml). It is concluded that dmPGE 2, when topically applied to canine gastric epithelium, is a potent stimulant of bicarbonate output that is dose-dependent. The ability of atropine and tetrodotoxin to prevent this secretion suggests that a cholinergic mechanism may be involved and that dmPGE 2 mediates its effects on HCO 3 − output through acetylcholine release which in turn stimulates cholinergic nerve endings.