The role of vasopressin in the pathogenesis of ethanol-induced gastric hemorrhagic erosions in rats

GASTROENTEROLOGY

1991;101:1242-1248

The Role of Vasopressin in the Pathogenesis of Ethanol-Induced Gastric Hemorrhagic Erosions in Rats Is Vasopressin an Endogenous Aggressor Toward the Gastric Mucosa? FERENC LfiSZL6, GIZELLA KAtiCSONY, ERZSliBET SZAB6, JENb LkNG, LAJOS BALfbPIRI, and FERENC A. LkZL6 First Department of Medicine, Department of Radiology, Department of Nuclear Medicine, and Department of Medical Chemistry, Albert Szent-Gyijrgyi Medical University, and Department of Comparative Physiology, Attila J6zsef University of Sciences, Szeged, Hungary

The role of vasopressin in the development of gastric hemorrhagic erosions induced by the oral administration of 1 mL of 75% ethanol in rats was studied. The area of the lesions in homozygous Brattleboro rats, having a defective vasopressin synthesis, was only 20% of that found in Wistar and heterozygous Brattleboro rats, which have normal vasopressin production. It is well known that vasopressin acts via the V, (pressor) and V, (antidiuretic) receptors. Administration of V, and V, vasopressin-receptor agonists and antagonists in this model showed that pressor-receptor activity is needed for the generation of all lesions in Wistar and heterozygous Brattleboro rats. Ethanol damage to the gastric mucosa was diminished by the V, antagonist with similar efficacy as in the case of a vasopressin deficiency. Administration of the V, antagonist and the absence of endogenous vasopressin were shown to protect the deeper layer of the gastric mucosa (assessed by histology) and to reduce significantly the ethanolinduced vascular injury and increase in vascular permeability (assessed by the monastral blue technique). Thus, endogenous vasopressin is clearly of great importance in the pathogenesis of gastric hemorrhagic lesions induced by ethanol. These results strongly suggest that vasopressin is an endogenous aggressor toward the gastric mucosa.

development of peptic ulcer disease (1,2). Gastric cytoprotection (3-5) means the prevention of chemical-induced (ethanol, HCl, NaOH) hemorrhagic erosions without inhibition of acid secretion. The number of known protective agents is steadily increasing, but their mode of action is not fully understood (4,~). The presence or absence of vascular endothelial damage and increased vascular permeability are wellknown controlling factors in gastric mucosal injury or protection (6,7). Protective agents exert direct or indirect effects on the vascular bed of the stomach (2). An intact mucosal microcirculation prevents the mucosal energy metabolism deficit caused by exogenous chemicals (8), allows rapid restitution of the mucous neck cells (9,10), and buffers acid back-diffusion into the mucosa (11,~). Physiological doses of vasopressin (VP), a potent pituitary nonapeptide, cause vasoconstriction in the intestinal tract (13). High doses of VP produce significant ischemia and secondary mucosal damage in the stomach (14). Gastric mucosal injury develops in a hemorrhagic shock model (11,12,15); VP apparently

G

Abbreviations used in this paper: dDAVP, deamino-o-argininevasopressin; DI rats, homozygous Brattleboro rats with diabetes insipidus; ED, effective dose; HZ rats, heterozygous Brattleboro rats; LVP, lysine-vasopressin; VP, vasopressin. o 1991 by the American Gastroenterological Association 0016-5085/91/$3.00

astric

acid

and

pepsin

are

aggressors

of the

gastroduodenal mucosa, whereas endogenous mucus and bicarbonate as well as an intact microvascular system are protectors. An imbalance between

these factors plays a pathophysiological

role in the

plays an essential role in generating mesenteric and splenic vasoconstriction following hemorrhage (16,17). teric

The presence

vascular

beds

of VP receptors

suggests

in the mesen-

its physiological

impor-

THE ROLE OF VASOPRESSIN IN GASTRIC CYTOPROTECTION

Novernher 1991

tance (18).These data prompted us to study the role of VP in gastric cytoprotection. For better comprehension of our experimental design, it should be noted that two different types of VP receptors, V, (pressor) and V, (antidiuretic), have been identified (19), and selective V, and V, receptor antagonists have been developed (20). Over the last 10 years, these antagonists have been used world-wide to (21)and to investigate the characterize VP receptors physiological and pathophysiological aspects of VP in animals (22) and, more recently, in humans (23). In studies concerning the role of VP, the Brattleboro strain of rats has generally been chosen because the homozygotes have hereditary defects in the synthesis and storage of VP and exhibit diabetes insipidus (24). Materials

and Methods

Animals and Chemicals Female Wistar rats, heterozgous rats, and homozygous Brattleboro rats with ous (DI rats) (originating from our breeding Their initial body weights were between

Brattleboro (HZ) diabetes insipidfarm) were used.

150 and 200 g. They were fasted for 24 hours before all experiments but received water ad libitum. Two hours before the studies, they wre deprived of water to be sure that their stomachs were empty. Lysine-vasopressin (LVP; Sandoz, Basle, Switzerland) (rat antidiuretic activity, 284 IU/mg; rat vasopressor activity, 270 IU/mg); deamino-D-arginine-vasopressin (dDAVP; SPOFA, Poland) (rat antidiuretic activity, 1200 IU/mg: rat vasopressor activity, 0.39 IU/mg); V, antagonist [d(CH,),Tyr(Me)AVP] [antidiuretic activity, 0.31 IU/mg; antivasopressor effective dose (ED). 0.16 nmol/kg]; and V2 antagonist [d(CH,),(D-Ile”,Ile’)AVP] (antiantidiureticED, 0.67 nmolikg; antivasopressor ED, 26 nmol/kg) were used in the experiments. The ED is defined as the dose that reduces the response to 2x units of agonist so that it is equal to the response to x units of agonist administered before the antagonist (20).

Gross Studies Twelve animals in each group received a single IP injection of the relevant substance 30 minutes before ethanol administration. Animals in the basic study received the solvent (0.9% NaCl), those in the agonist study received 0.1 &lOO g body wt LVP or 0.1 kg/100 g body wt dDAVP, and those in the antagonist study received 0.1 pg/lOO g body wt V, or 0.4 kg/100 g body wt V, antagonist. The doses were selected in pilot time- and dose-response studies to yield well-measurable changes: the doses of the V, and V, antagonists were in the same ratio as that of their EDs. More than one control group (Wistar rats) was used because the experiments were not performed at the same lime. Control groups received only the solvent (0.9% NaCl), in the same volume, by the same route, 30 minutes before ethanol administration. One milliliter of 75% ethanol was injected through a gastric tube into the stomach of pretreated and control rats. One hour later, the animals were killed by

1243

cervical dislocation, the abdomen was opened by a median laparotomy, and the stomach was removed, dissected along the greater curvature, stretched out, and photographed. The areas of the lesions (petechiae, hemorrhagic streaks, confluent hemorrhagic erosions) and the total mucosa were determined planimetrically (Tamaya Planix Digital Planimeter; Planix, Tamaya, Japan) on enlarged (15-20x) photographs, and the ratio of the two was expressed as a percentage.

Measurement of Gastric Acid Secretion in the Presence of a V, Antagonist The effect of the V, antagonist on basal gastric acid secretion was studied by the washout technique of Magee and Nakajima (25). Eight female Wistar rats initially weighing 200-250 g were fasted overnight and anesthetized with Nembutal(4 mg/lOO g body wt, IP). A tube was inserted into the stomach through the proximal side of the duodenum, and the pylorus was ligated. Secretion was collected continuously and sampled at lo-minute intervals. Gastric acid was titrated with a Radiometer (Copenhagen, Denmark] automatic titrater to pH 7 to measure the gastric acid output. The V, antagonist (0.1 ~$100 g body wt) was injected 40 minutes after the beginning of the experiment and its effect on secretion was monitored for 60 minutes.

Gastric Emptying Each experimental group consisted of eight rats. Female Wistar. HZ, and DI rats with initial body weights of 150-200 g were fasted for 24 hours but received water ad libitum. Before the experiment. the animals were deprived of water for 2 hours. The gastric emptying was determined by a radiographic technique using a radiographic imaging intensifier system (EDR 750; Medicor, Budapest. Hungary). Medifort RP (Forte, VBc. Hungary) radiographic films and an iodine-containing, nonionic. water-soluble, nonabsorbable, isosmotic contrast material (Omnipaque; Nycomed, Oslo, Norway; containing 240 mg/mL iodide) were used. Wistar. HZ, and DI rats received 1 mL of ethanol mixed with contrast material (0.75 mL of 100%1 ethanol + 0.25 mL of contrast material]. Immediately after the oral injection, the rats were placed into modified Bollman cages (the customary iron or plastic bars had been replaced by wooden ones to eliminate disturbing radiographic shadows). Three cages, each containing one Wistar, one HZ, or one DI rat, were situated in parallel under the intensifier, allowing simultaneous observation of the movement of the contrast material through the pylorus. The times of the start of gastric emptying. when the contrast material appeared in the duodenum, and the end, when the contrast material disappeared from the stomach, were measured in minutes after the oral administration (0 minute] of ethanol. The passage was checked on the monitor every minute by two investigators. For the documentation, radiographic photographs were taken: the first at 3 minutes after the oral administration of ethanol, the second at 10 minutes, and then at lo-minute intervals up to a maximum of 120 minutes.

1244

LASZLO ET AL.

GASTROENTEROLOGY

Measurement of Intragastric pH Under Basic, Agonistic, and Antagonistic Conditions In this study, each group consisted of eight rats. Female Wistar, HZ, and DI rats were used in the same experimental protocol as described above. Before the experiment, the animals received a single IP injection of Nembutal (Serva, Heidelberg, Germany; 2 mg/lOO g body wt] for premeditation. Intragastric pH was measured using an in vivo pH-sensitive system (OP 111; Radelkis, Budapest, Hungary; sensitive between pH 1 and pH 13). The flexible pH-sensitive electrode (OP 7223; Radelkis) was at most 3 mm in diameter. The indifferent electrode (OP 8387: Radelkis) was situated in a gauze pad, saturated with water to complete the pH-meter-indifferent electrode-rat-pH-sensitive electrode-pH-meter circuit. The pH-sensitive electrode was inserted transorally into the stomach of each rat three times: (a) before pretreatment (basal value); (b) 30 minutes after pretreatment, i.e., before ethanol administration; and (c) 60 minutes after ethanol administration.

Vascular Permeability Each group in this experiment consisted of eight to nine animals. All groups were prepared and treated as described above. Monastral blue (Sigma; 3 mg/lOO g body wt) was injected into the jugular vein 10 minutes before autopsy or 50 minutes after ethanol administration. Ten minutes before the monastral blue injection, the animals were anesthetized with Nembutal (Serva; 4 mg/lOO g body wt, IP). The autopsies were performed after cervical dislocation. The abdomen was opened by median laparotomy, and the stomach was removed, cut along the greater curvature, stretched out on cork, and fixed in formalin (6% in 0.9% NaCl solution) for 24 hours. After formalin fixation, the stomachs were put into 90% glycerol for a minimum of 24 hours. After this procedure, the stomachs were photographed. The areas of the total mucosa, the lesions, and the staining were measured planimetrically on enlarged, colored pictures. The ratios of lesioned/total and stained/total mucosa areas were expressed as percentages.

Histopathological

Vol. 101. No. Fi

inflammatory cell reaction in the submucosa. The most severe lesion in each section was taken, and the data of the appropriate groups were pooled as a histopathological distribution of lesions in percent. Gastric Blood Flow The gastric blood flow was measured by a radiolabeled rat erythrocyte microsphere technique after LVP treatment. For labeling, 300-350-g male Wistar rats received 0.31 mg/lOO g body wt of tin(I1) pyrophosphate (prepared in our laboratories) into the tail vein 15 minutes before decapitation. Seven to eight milliliters of blood was collected into heparinized tubes from each rat. 99mT~0; (Sorin Biomedica, Vienna, Austria), 150-200 MBq, was added to 15 mL of blood, which was washed three times with physiological saline (0.9% NaCl). The radiochemical purity (> 99%) of the labeled erythrocytes was checked radiochromatographically. For blood-flow measurements, female Wistar rats (initial body weights were 200-210 g) were fasted for 24 hours but received water ad libitum. Each group consisted of 8-12 rats. The animals received injections in the tail vein of 0.5 mL of radiolabeled blood 10 minutes before the administration of LVP (0.1 &lOO g body wt, IP) or physiological saline (the solvent, at the same time, by the same route these were the controls). The rats were killed by cervical dislocation 5, 15, or 30 minutes after LVP or saline injection, and the stomach was removed by median laparotomy, cut along the greater curvature, rinsed gently in saline, and weighed. The radioactivity of each stomach was measured for 20 seconds using a gamma counter (Gamma, Budapest, Hungary; 140 keV; window, 50%), and the gastric radioactivity was calculated as the number of counts per gram stomach wet weight. One-half milliliter of radiolabeled blood was dissolved in 1000 mL of distilled water and used as a standard in each experiment. The radioactivities of l-mL aliquots of this solution were measured before and after the appropriate gastric radioactivity measurements, and the injected radioactivity was calculated. The blood-flow data were expressed in terms of the ratios of measured gastric radioactivity to injected radioactivity.

Evaluations

Eight rats were tested in each group. They were treated and prepared as described above. Sixty minutes after the oral administration of 75% ethanol, the stomachs were fixed in formalin (6% in 0.9% NaCl solution) for 24-hours, and six rectangular pieces of the glandular mucosa (about 3 x 15 mm) were then cut out from standard sites of the fundic area (four pieces) and the antrum (two pieces) and embedded in paraffin, H&E staining was used. All sections were assessed according to the scale of Lacy and Ito (26): 0, no damage; I, luminal surface damage, the swelling of the foveolar and superficial parietal cells involving an edematous lamina propria and superficial erosions caused by disruption of damaged foveolar cells; II, luminal surface and gastric pit cell damage, with erosions involving the upper half of the mucosa. surrounded bv dilated vessels: and III. damage to all cells’ from the lum&al surface to the gastric gland cells and severe necrosis of the whole mucosa with an

Statistics The results were analyzed using the nonparallel Mann-Whitney U test, and differences were considered significant when the probability was < 5%. Results Gross Study The ratio of lesioned to total mucosa area in DI of that in Wistar and HZ rats (Figure 1). The gastric lesions were significantly more extensive in LVP-pretreated Wistar and HZ rats; the DI rats showed the basic control picture of hemorrhagic erosions. No change in the basic situation was caused by dDAVP (Figure 1). The V, antagonist was

rats was 20%-25%

November

THE ROLE OF VASOPRESSIN

1991

IN GASTRIC CYTOPROTECTION

1245

Gastric Emptying J-J

Wistar

H

HZ

m

01

There were no significant differences in the gastric emptying of 1 mL of 75% ethanol between DI, HZ, and Wistar rats. The gastric emptying started immediately after the oral administration of ethanol and lasted for at least 120 minutes. The pylorus was temporarily closed from 12-13.25 until 49.13-53.38 minutes after ethanol injection in all three strains. During this time, no contrast material was observed in the duodenum and the initially emptied volume passed slowly through the jejunum. Intragastric pH Changes

LVP Y

txlslc

skcly 1.

”

antngonist study3.

agonist study 2.

Figure 1. Role of VP in the development of 75% ethanol-induced gastric hemorrhagic erosions. Protective effects of a VP deficiency and the V, antagonist against gastric mucosal damage induced by 75% ethanol in rats. n = 12; mean 2 SEM expresses the extent of ethanol-induced gastric hemorrhagic erosions; P < 0.001 vs. controls or Wistar rats in the basic study.

found to protect the gastric mucosa against 75% ethanol-induced damage in Wistar and HZ rats, whereas the V, antagonist had no effect in this model (Figure 11.

Vascular Injury, Visualized by the Mona&al Blue Technique

Gastric Acid Measurement The V, antagonist gastric acid secretion. Table

1.

applied

The intragastric pH was in the range of 1.74in each study before pretreatment (basal value): no significant differences were found between the different strains (Table 1). The pH values in the pretreated groups, with the exception of the LVPpretreated group, did not differ significantly from the basal values; LVP pretreatment caused a significant elevation of the intragastric pH from 1.74-1.8 to 2.81-3.33 (Table 1). Sixty minutes after ethanol administration, the intragastric pHs were similar, in the range of 3.48-4.48, in all groups investigated (Table 1). The pH of the 75% ethanol was 7.8-8. 2.2

does not influence

Hemorrhagic lesions were found in all stomachs in each group (Table 21, but the incidence of

Intragastric pH Data Before and Ajier Pretreatment With Various Vasopressin Agonists and Antagonists and After

Oral Administration of 75% Ethanol 30 minutes after pretreatment

60 minutes after ethanol administration

Group

Basal values

Pretreatment

DI HZ Wistar

2.15 2 0.26 2.06 k 0.28 2.08 2 0.28

Saline Saline Saline

2.08 2 0.24 1.95 + 0.30 2.05 + 0.31

4.26 -t 0.44” 4.19 ” 0.36” 3.78 ” 0.52”

DI HZ Wistar

1.80 + 0.21 1.80 k 0.12 1.74 -r 0.14

LVP LVP LVP

3.33 f 0.23” 2.97 k 0.27” 2.81 2 0.31”

3.58 k 0.51” 3.55 + 0.55” 3.48 2 0.62”

DI HZ Wistar

1.89 + 0.10 1.94 2 0.08 2.02 2 0.15

dDAVP dDAVP dDAVP

1.88 + 0.11 1.92 f 0.11 1.99 + 0.08

4.48 -t 0.41” 4.28 k 0.44” 4.17 + 0.47”

DI HZ Wistar

2.12 + 0.15 2.16 2 0.19 2.20 r 0.12

V, antagonist V, antagonist V, antagonist

2.18 2 0.21 2.33 ? 0.23 2.14 ” 0.11

4.11 f 0.48” 4.00 + 0.56” 3.87 2 0.57”

DI HZ Wistar

1.92 2 0.25 1.88 k 0.25 1.78 f 0.25

V, antagonist V, antagonist V, antagonist

1.95 k 0.21 1.81 -t 0.14 1.79 ” 0.25

3.83 + 0.51’ 3.86 _C0.47” 3.61 + 0.42’

NOTE. n = 8; data presented as mean f SEM. “P < 0.001 vs. basal pH.

1246

LASZLO

ET AL.

GASTROENTEROLOGY

Vol. 101, No. 5

Table 2. Effects of a Vasopressin Deficiency and the V, Antagonist of Vasopressin on 75% Ethanol-Induced Vascular Injury and increased Vascular Permeability of Rat Gastric Mucosa, Investigated Using the Monastral Blue Technique Group investigated

Incidence of damage

Wistar HZ Wistar + V, antagonist HZ + V, antagonist DI

919 919

Damaged area (%) 13.88 11.95

Incidence of staining

+ 1.15 t 2.66

919 919

Stained area (%) 12.00 ? 1.15 9.40 k 1.32

919

2.71 k 0.42”

319

0.39 2 0.21b

818 818

2.21 2 0.62” 3.75 ” 0.89”

418 318

0.75 f 0.28b 0.47 + 0.23b

NOTE. n = 8-9; data presented “P < 0.001 vs. control damage. *P < 0.001 vs. control staining.

as mean

k SEM.

staining was much lower in the DI and V, antagonistpretreated Wistar and HZ rats (Table 2). In the untreated Wistar and HZ rats, 12% and 9.4% of the mucosa was stained, compared with 0.39%-0.75% in the other groups (Table 2). Histopathological Mucosal

Evaluation

of the

Injury

Based on a standard severity scale (26), 98%showed very severe damage (grades II-III) with 0%2% superficial disruption (grade I) in the untreated HZ and Wistar rats. Intact mucosa could not be observed in this group (Figure 2). In untreated DI rats and in V, antagonist-pretreated HZ and Wistar rats, the damage to the gastric mucosa was 20%25% grade 0, 35%-45% grade I, and only 25%-45% grades II-III (Figure 2). 100% of the stomach

Gastric Blood Flow

Five, 15, and 30 minutes after the IP injection of LVP, the gastric blood flow was reduced by 23%, 26.7%, and 19.9%, respectively (Table 3).

Discussion This study evaluated the possible pathogenetic role of VP in the development of gastric hemorrhagic erosions induced by 75% ethanol in rats. The gross mucosal lesions in the VP-deficient DI rats were significantly less severe than in the control rats, suggesting the pathogenetic role of endogenous VP in the development of ethanol-induced extensive damage of the stomach. Exogenous LVP significantly aggravated the ethanol-induced lesions both in the HZ and Wistar rats and in the rats with presumed endogenous VP deficiency, providing further evidence of such a role of VP. A V, antagonist conferred

Figure 2. Histopathological evaluation of the protective effects of a VP deficiency and the V, antagonist of VP against 75% ethanolinduced gastric erosions in rats. n = 8; columns represent the histopathological distribution of the severity (scale, O-III) of 75% ethanol-induced lesions in the appropriate groups. 1,Wistar; 2, HZ; 3, Wistar with V, antagonist; 4, Hz with V, antagonist; 5, DI.

significant protection in the control rats without affecting gastric acid secretion. The ineffectiveness of the V, agonist dDAVP and the V, antagonist showed that the V, receptor activity is unrelated to these changes. Two questions arose after the gross evaluation of the gastric lesions. The first potential problem originated from the well-known fact that ethanol reduces the plasma VP level and this hormone acts via the circulation. However, it has been shown that a high dose of orally administered ethanol causes antidiuresis (27) and an elevation of the plasma VP level (28). These results are in agreement with our recent findings, in which we have shown a significant increase in the plasma VP level after the oral administration of 1 mL of 75% ethanol to rats (29), indicating that a high dose of ethanol releases VP. The second question was whether the protective effect of an endogenous VP deficiency in DI rats may be explained by differences in their hemodynamic parameters, rates of gastric Table 3. Gastric Blood Flow 5, 15, and 30 Minutes After Lysine-Vasopressin Treatment, investigated Using a Radiolabeled Erythrocyte Microsphere Technique in Rats Time after treatment Group Saline-treated LVP-treated

control

(min)

5

15

30

9.81 + 0.48 7.51 t 0.57”

9.52 + 1.13 6.85 ? 1.11”

11.20 zk 0.61 9.21 ? 0.24”

NOTE. Data are expressed in terms of measuredgastric injected radioactivity ratios (X 104); n = 8-12: mean “P < 0.05 vs. saline-treated groups.

radioactivity/ 2 SEM.

November

1991

THE ROLE OF VASOPRESSIN

emptying, and intragastric pH. The hemodynamic parameters of DI rats differ from those of HZ and other types of “normal” rats, i.e., they have a mildly hyperkinetic circulation (30). This possibly allows a better perfusion of the organs, including the stomach, and an elevated oxygen supply may be responsible for the protective effect. No significant differences in gastric emptying and intragastric pH were found for any of the strains, suggesting that the protective effect of a VP deficiency is not related to these parameters. Monastral blue and histopathological studies showed that in control rats receiving a V, antagonist and in rats with an endogenous VP deficiency, the increase in vascular permeability and the deep mucosal lesions induced by ethanol were significantly reduced. These findings indicate that the gastric microvascular system is a possible “end-target;” the endogenous VP deficiency and a V, antagonist exert their protective effects against ethanol damage at this level. Exogenously administered LVP reduces the gastric blood flow. The diminished oxygen supply possibly sensitizes the gastric mucosa, resulting in aggravation of the ethanol-induced lesions and elevation of the intragastric pH (31) in all three study groups. As concerns how exogenously administered or possibly endogenously released VP can act on the gastric blood flow, Laycock and Lightman (32) have recently shown that VP in a dose closely similar to that used in this study elevates the arterial blood pressure in both DI and “normal” rats. This elevation is associated with a significant increase in total peripheral resistance and a decrease in cardiac output. Gardiner et al. (33) have shown that VP in doses that do not affect the systemic blood pressure causes vasoconstriction of the superior mesenteric artery in rats; this effect is blocked by a V, antagonist (33). No data are available yet as to whether VP acts directly on the gastric small arteries or whether VP has receptors there. Hemodynamic changes, especially vasoconstriction in the superior mesenteric artery, may explain the gastric blood flow-reducing effect of LVP in the present experiment. Vasopressin-induced platelet aggregation (34) and increased vascular permeability (35,36) may play additional roles in the development of the gastric injury described above. In conclusion, endogenous VP has been found to have an important role in the pathogenesis of ethanolinduced gastric hemorrhagic erosions in rats via a V, receptor-mediated pathway. Administration of the V, antagonist and the lack of endogenous VP were shown to protect the deeper layer of the gastric mucosa and to reduce significantly the ethanolinduced vascular injury and increased vascular permeability. Thus, in addition to gastric acid and pepsin, endogenous VP is strongly suggested as an aggressive

IN GASTRIC

CYTOPROTECTION

1247

factor toward the gastric mucosa. Investigation of the incidence of peptic ulcer disease among subjects with central diabetes insipidus appears to be indicated.

References and prostaglan1. Fiske SC. Peptic ulcer disease, cytoprotection dins. Arch Intern Med 1988;148:2112-2113. basis for a role for sulfhydrils and 2. Szab6 S. Experimental dopamine in ulcerogenesis: a primer for cytoprotectionorganoprotection. Klin Wochenschr 1986;64(Suppl VII):ll6122. by prostaglandins. Gastroenterology 3. Robert A. Cytoprotection 1979;77:761-767. 4. Silen W. What is cytoprotection enterology 1988;94:232-235.

of the gastric

mucosa?

Gastro-

KL, Miller TA. Response to an 5. Robert A. Guth PH, Schmidt editorial on gastric cytoprotection. Gastroenterology 1989;96: 548-554. 6. Szabd

S, Trier JS, Brown A, Schnoor J. Early vascular injury and increased vascular permeability in gastric mucosal injury caused by ethanol in the rat. Gastroenterology 1985;88:228236. dose of 7. Leung FW. Robert A, Guth PH. Effect of cytoprotective 16,16-dimethyl prostaglandin on gastric mucosal blood flow (abstr). Gastroenterology 1982;88:1472

R. Menguy

R. Masters YF. Gastric and ‘stress ulceration.’ Ann Surg of 9. Ito S, Lacy ER. Morphology defense. and restitution in the Gastroenterology 1985;88:250-260,

mucosal energy metabolism 1974;180:538-548. rat gastric mucosal damage, presence of luminal ethanol.

10. Tarnawski

A, Hollander D. Stachura J, Krause WJ, Gergely H. Prostaglandin protection of the gastric mucosa against alcohol injury a dynamic time-related process. Gastroenterology 1985: 88:334-352. 11. Moody FG. McGreevy J, Zalewsky C. (:heung LY. Simons M. The cytoprotective tal erosive gastritis.

effect of mucosal blood flow in experimenActa Physiol Stand 1978:35(Suppl):35-43.

12. Kivilaakso

E. Fromm D. Silen W. Relationship between ulceration and intramural pH of gastric mucosa during hemorrhagic: shock. Surgery 1978;84:70-78.

13. Liard

JF. DBriaz 0, Schelling P, Thibonnier distribution during vasopressin infusion

M. Cardiac output or dehydration in

conscious

dogs. Am J Physiol 1982:243:H663-H669. WP. Acute gastric mucosal damage induced by bile salts, acid and ischemia. Gastroenterology 1975;68:699-707. K. Guth PH. Role of blood flow 15. Leung FW, Itoh M, Hirabayashi in gastric and duodenal mucosal injury in the rat. Gastroenterology 1985;88:281-289, CV. Intestinal vasoconstric16. McNeil1 JR, Stark RD, Greenway tion after hemorrhage: roles of vasopressin and angiotensin. 14. Ritchie

Am J Physiol

1970;219:1342-1347.

GreenwayCV. Sympathetic and hypophyseal roles in the splenic response to hemorrhage. Am J Physiol 1971;220:837-840. 18. St-Louis J, Schiffrin EL. Biological action and binding sites for vasopressin on the mesenteric artery from normal and sodium depleted rats. Life Sci 1984;35:1489-1495. 19. Michell RH, Kirk CJ. Billah MM. Hormonal stimulation of phosphatidyl-inositol breakdown, with particular reference to the hepatic effect of vasopressin. Biochem Sot: Trans 1979;7: 861-865. 20. Manning M. Sawyer WH. Synthesis and receptor specificities of vasopressin antagonists. J Cardiovasc Pharmacol 1986: (Suppl 7):S29-s35. 17. Stark RD, McNeil1 JR,

1248

tiSZL6

GASTROENTEROLOGY Vol. 101, No. 5

ET AL.

21. Jard S. Mechanisms

22. 23. 24. 25.

26.

27. 28.

29.

30.

of action of vasopressin and antagonists. Kidney Int 1988;34(Suppl26):S38-S42. Liard JF. Vasopressin antagonists and their use in animal studies. Kidney Int 1988;(Suppl26):S43-S47. Thibonnier M. Use of vasopressin antagonists in human diseases. Kidney Int 1988;(Suppl26):S48-S51. Sokol HW, Valtin H (eds). The Brattleboro rat. Ann NY Acad Sci 1982;394:1-828. Magee DF, Nakajima S. The effects of antral acidification on the gastric secretion stimulated by endogenous and exogenous gastrin. J Physiol 1968;196:713-721. Lacy ER, Ito S. Microscopic analysis of ethanol damage to rat gastric mucosa after treatment with a prostaglandin. Gastroenterology 1982;83:619-625. Kissin B, Begleiter H. The biology of alcoholism. London: Plenum, 1971. Linkola J, Ylikahri R, Fyhrquist F, Wallenius M. Plasma vasopressin in ethanol intoxication and hangover. Acta Physiol Stand 1978;104:180-187. L&z16 F, Karacsony G, NBfr&di J, Rengei B, Vecsernyes M, Laczi F, Szarvas F, Lengyel Z, L&z16 FA. A high dose of ethanol releases vasopressin and simultaneously induces lesions in the rat stomach-the role of vasopressin in gastric cytoprotection. In: Thorn NA, Vilhard A, Treiman M, eds. Proceedings of the Fourth International Conference on the Neurohypophysis. London: Oxford University, 1989:159-161. Zicha J, Karen P, Krpta V, Dluohd H, Krecek J. Hemodynamics

31. 32.

33.

34.

35.

36.

of conscious Brattleboro rats. Ann NY Acad Sci 1982;394:409413. Jacobson ED, Swan KG, Grossman MI. Blood flow and secretion in the stomach. Gastroenterology 1967;52:414-420. Laycock JF, Lightman SL. Cardiovascular interactions between vasopressin, angiotensin and noradrenaline in the Brattleboro rat. Br JPharmacol 1989;96:347-355. Gardiner SM, Compton AM, Bennett T. Regional hemodynamic effects of vasopressin in conscious, unrestrained, Brattleboro rats. Br J Pharmacol 1989;97:147-152. Siess W, Stiefel M, Binder H, Weber PC. Activation of V,receptors by vasopressin stimulates inositol phospholipid hydrolysis and arachidonate metabolism in human platelets. Biochem J 1986;233:83-91. Rosenberg GA, Kyner WT, Fenstermacher JD, Patlak CS. Effect of vasopressin on ependymal and capillary permeability to tritiated water in cat. Am J Physiol 1986;251:F485-F489. D6czi T, Szerdahelyi P, Gulya K, Kiss J. Brain water accumulation after the central administration of vasopressin. Neurosurgery 1982;11:402-406.

Received May 12,1989. Accepted May 1,1991. Address requests for reprints to: Ferenc LBszl6, M.D., First Department of Medicine, Albert Szent-GyBrgyi Medical University, Szeged, Hungary, H-6701, P.B. 469. The authors are grateful to Professor Maurice Manning for the vasopressin antagonists.