Gastric epithelial restitution at low luminal pH during influence of pentagastrin or cimetidine in the cat

GASTROENTEROLOGY 1987;93:753-64

Gastric Epithelial Restitution at Low Luminal pH During Influence of Pentagastrin or Cimetidine in the Cat JON ERlK GRijNBECH, KETIL GRONG, JON LEKVEN, and KNUT SVANES Surgical Research Laboratory,

Department

JAN ERlK VARHAUG,

of Surgery, University

This study examines restitution of the gastric surface epithelium at JuminaJ pH 1 after damage to the gastric mucosa by instillation of 2 M NaCJ for 10 potential difference min. After mucosal damage, decreased to zero and H+ back-diffusion increased. Gastric blood flow increased after exposure to 2 M NaCJ, irrespective of the secretory state of mucosa. Neither pentagastrin nor cimetidine showed any protective effect against mucosal damage produced

by 2 M NaCJ. Ninety minutes after mucosal exposure to 2 M NaCJ there was an almost complete restitution of the gastric surface epithelium in pentagastrin-stimulated and cimetidine-treated animals, whereas about 15% of the gastric surface still showed damage in animals in which acid secretion was not interfered with. These findings show that restitituion of the surface epithelium, in vivo, is a rapid and resistant process even at a very low JuminaJ pH. The secretory state of mucosa may have some influence on this process. In vitro observations in frog and guinea pig preparations have demonstrated rapid restitution of the gastric surface epithelium after damage of the surface by hypertonic saline (1,~). In vivo exposure of the rat gastric mucosa to 100% ethanol was followed by restitution of >i’g% of the gastric surface 30 min later (3). We have recently demonstrated functional and structural recovery of the gastric surface epithelium in an in vivo cat model 90 min after exposure to Received July 11,1986.Accepted March 31,1987. Address requests for reprints to: Dr. Jon Erik Griinbech, Department of Surgery, Haukeland Hospital, N-5016 Bergen, Norway. This work was supported by a grant from the Norwegian Society for Fighting Cancer. The authors thank Inger Vikbyr, Elly Soltvedt, Hilde Killingstad, Bente Fredriksen, and Randi Sundsfjord for skilled technical assistance. 0 1987 by the American Gastroenterological Association 0016-5085/87/$3.50

of Bergen, Bergen, Norway

2 M NaCl, keeping 7.4 (4).

the gastric luminal

content

at pH

Knowledge of the regulatory mechanisms of this recovery is largely incomplete. It has been suggested that calcium is required for both cell migration and the formation of intercellular junctions, and intact microfilaments may be required as indicated by inhibition of restitution with cytochalasin B in vitro (5).Also, the acidity in the stomach appears to influence recovery inasmuch as lowering the luminal pH to 3 in frog and to 1 in guinea pig gastric preparations inhibited epithelial restitution in vitro (87). The present study was undertaken to establish whether restitution takes place, or not, in the presence of a strongly acidic luminal solution. Furthermore, we examined whether stimulation or inhibition of acid secretion under such circumstances influenced the recovery process of the gastric surface epithelium. As gastric mucosal blood flow is known to be influenced by cimetidine and pentagastrin (8), we found it necessary to study the influence of these drugs on the blood flow response after mucosal injury.

Materials and Methods Animal Preparation Fifty-three cats of both sexes, weighing 2.3-5.4 kg, were used. During the last 24 h before the experiment, food was withdrawn, but the access to water was free. Anesthesia was induced with sodium pentobarbital, 40 mg/lcg i.m., and additional doses were given intravenously to preserve anesthesia and spontaneous respiration. Tracheostomy was performed. Three cannulas, 1.02 mm in diameter, were inserted. One cannula was inserted via the right carotid artery into the left cardiac ventricle, the second via the left femoral artery to the lower portion of the abdominal aorta, and the third into the inferior caval vein from the groin. Correct position of the cardiac cannula was assured by

754

GRQNBECH ET AL.

obtaining a typical pressure trace with low diastolic pressures. Mean aortic pressure was continuously recorded by means of a Statham P23De pressure transducer (Hato Rey, P.R.). The blood pressure was allowed to fluctuate, but animals with mean aortic pressure below 95-100 mmHg due to bleeding, too deep anesthesia, or pneumothorax were not used in this study. Ringer-acetate solution was infused intravenously at a rate of about 10 ml kg-’ h-’ by a constant-rate infusion pump throughout the experiment, Blood gas (pH, Pao*, and PacoJ analyses (BMS 3Mk2 Radiometer, Copenhagen, Denmark) were performed immediately before mucosal damage and 15 and 90 min after the end of this period, corresponding to the times of blood flow determination. A midline laparotomy was performed. A plastic tube, 5 mm in diameter, was entered through the mouth and advanced to below the esophagogastric junction. The tube was kept in place, and leakage was prevented, by a ligature around the abdominal part of the esophagus, taking care to avoid damage to the vagal nerves. Another fenestrated tube of equal diameter was inserted into the stomach via the duodenum and placed along the greater curvature. Along with this second tube, a catheter, 1.02 mm in diameter, and two agar bridges were inserted. The pylorus was closed around these cannulas with a ligature not interfering with the gastric blood supply. Isotonic saline (150 mM) of pH 1 was circulated from a reservoir at a rate of 5 mUmin through the esophageal tube into the stomach and evacuated through the duodenal tube back into the reservoir by means of a constant-rate peristaltic pump [Harvard Apparatus, 1203 A, Millis, Mass.). The volume of fluid in the stomach was kept at 30 ml to avoid gastric distention as judged visually and by monitoring the intragastric pressure by a Statham P23De transducer. Intragastric pressure in this study remained below 6.5 mmHg throughout the experiments. Mucosal

Damage

After completion of the cannulation procedures, the animals were allowed to stabilize for ‘15-30 min. Baseline variables were thereafter recorded for 30 min, and in case of acid stimulation or inhibition, another 60 min were added to this period. The gastric luminal perfusion was then stopped, and the tubes to the stomach were clamped. Gastric fluid was removed, and the gastric mucosa was exposed to hypertonic NaCl by instillation and withdrawal of 30 ml of 2 M NaCl followed by another instillation of 30 ml of 2 M NaCl. After 10 min, the hypertonic solution was removed and the stomach was rinsed twice with 30 ml of isotonic saline from the reservoir. Gastric perfusion with saline at pH 1 was then reestablished and continued throughout the rest of the experiment. Net H+ Flux The acidity of the circulating fluid was kept constant at pH 1 and gain or loss of H+ from the gastric lumen was measured continuously by a pH-stat method (arranged with a PHM 62 standard pH meter connected to two TTT 80 titrators and two ABU 80 autoburettes; Radiometer,

GASTROENTEROLOGY

Vol. 93, No. 4

Copenhagen, Denmark) with the pH probe placed in the reservoir. As titrants, 0.5 M NaOH was used for determination of net gain of Hf in the fluid ‘returning from the stomach, and 0.5 M HCl was used for determination of net loss of H+. Fluid from the initial 10 min after mucosal damage was discarded because of the irrigation procedure following that period. Because of fluctuations associated with titration at pH 1, net Hf flux values were pooled for 30-, 60-, and JO-min periods and calculated as micromoles per minute [see Table 5). Potential

Difference

Transmucosal potential difference (PD) was measured by agar bridges (saturated KC1 in 3% agar; Intermedic Polyethylene Tubing, PE-200; internal diameter 1.40 mm) connected via calomel electrodes to voltmeters (Fluke 8020 B Multimeter, Tilburg, Holland). Two agar bridges were inserted via the duodenostomy into the stomach, one with the tip in the middle part of the corpus close to the greater curvature and the other in the antrum z cm above the pylorus. The reference agar bridge was placed in the right femoral vein (4,9). The potentials were recorded every 10 min. Mucosal

Blood

Flow

Regional tissue blood flow and cardiac output were determined by means of the distribution of embolized microspheres, 15 * 1 km in diameter, labeled with 4eSc, 85Sr, g5Nb, l13Sn, 14’Ce, or 153Gd [New England Nuclear, Boston, Mass.). Three measurements were performed in each animal, with the isotope sequence randomized. The spheres were suspended in 10% dextran, sonicated on the day of the experiment, and mechanically shaken immediately before injection. Microspheres were injected into the left cardiac ventricle for a total of 30 s. Starting 15 s before sphere injection and terminating 75 s after its completion, a reference blood sample was drawn from the abdominal aorta by means of a constant rate pump set at 1 mUmin (Sage Instruments 351, Cambridge, Mass.). Collected blood samples were later weighed to determine the exact extraction rate. Tissue samples used for flow determination were taken from the antrum, corpus, and fundus regions. Full-wall samples were divided into a combined mucosa/submucosa portion and a muscularis portion by stripping these layers apart. Radioactivity of blood and whole wet tissue samples was measured in an automatic y-spectrometer (CompuGamma 1282, LKB-Wallac Company, Turku, Finland) connected by an interface to a computer. Window settings were appropriate for the photopeak of each isotope energy. Tissue blood flow rate and cardiac output were calculated as previously described (10-12). As mucosal blood flow was similar in corresponding regions of the anterior and posterior walls of the stomach, values from anterior and posterior walls were averaged. Blood flow data are given for the antrum, corpus, and fundus, respectively. Microscopy Immediately after killing the animal with an intracardiac injection of saturated KC1 solution, initial

GASTRIC EPITHELIAL RESTITUTION AT LOW pH

October 1987

Figure

755

1. A. Micrograph of cat gastric fundus mucosa after gastric luminal perfusion with isotonic saline at pH 1 for 2 h. Normal surface epithelium and gastric pits are seen (X 280). B. Micrograph of gastric corpus mucosa 5 min after damage by exposure to 2 M NaCl showing detachment of epithelial cells leaving gaps in the surface epithelium. Pits and glandular cells appear intact (x320). C. Section of the corpus mucosa 5 min after exposure to 2 M NaCl showing denuding of the surface between two pits and a partly damaged pit to the right (x320).

756

GRiiNBECH

ET AL.

GASTROENTEROLOGY

Vol. 93. No. 4

Fig! ure 2. A. Gastric corpus mucosa exposed to 2 M NaCl for 10 min, followed by gastric luminal perfusion with 150 mM NaCl at f IH 1 for 90 min. The micrograph shows epithelial restitution with a complete surface epithelial lining, shallow pits, and pari eta1 cells close to the surface (X 320). B. Corpus mucosa 90 min after exposure to 2 M NaCl. Epithelial restitution; the cells on the surface are cuboidal and irregularly arranged but form a complete lining (~400).

fixation of the mucosa was performed by instillation of 39 ml of Bouin’s solution into the in situ stomach for 10 min. The stomach was then removed, opened, and inspected. Samples of the anterior and posterior walls of antrum, corpus, and fundus regions, respectively, were selected for histologic examination corresponding to samples for determination of tissue blood flow. Samples for morphologic examination were further fixed in Bouin’s solution and 7-pm-thick sections were prepared and stained with hematoxylin and eosin. Subsequent to completion of all experiments, the sections were coded so that the examiner was unaware of which experimental group and region of the stomach they belonged to. Conventional light microscopy was used for histologic examination. The lengths of mucosal surface of a tissue section showing normal [Figure lA), damaged [Figures 1B and lC), and restituted (Figures 2A and 2B) surface epithelium were measured with an ocular micrometer. As shown in previous studies (1,4), epithelial restitution can be separated from normal and damaged epithelium by the following criteria: shallow pits and a continuous epithelial lining consisting mostly of irregularly arranged cuboidal and flattened cells. One section from each of the six different areas of a stomach was examined, which means a total of 318 sec-

tions. From each stomach area 14.7 + 3.8 mm of mucosal surface lengths (mean * SD] were quantitated. The results are expressed as the fraction (percent) of mucosal surface showing normal, damaged, or restituted epithelium according to Lacy and Ito (3). Experimental

Groups

Four experimental groups were studied. One group served as a control without gastric mucosal damage and with nonstimulated acid secretion. The three other groups were all subjected to gastric mucosal damage by 2 M NaCl and acid secretion was either nonstimulated, pharmacologically stimulated by pentagastrin, or inhibited by cimetidine. The three last groups were each subdivided into two parts. Most of the animals were observed for 90 min after exposure to 2 M NaCl, but in each group 6 of the animals were killed 5 min after exposure to 2 M NaCl to obtain information about early histologic changes. Nondamage group. Eight cats were subjected to gastric luminal perfusion with isotonic saline at pH 1. Equivalent to the period of gastric mucosal damage in the subsequent groups, 30 ml of 160 mM NaCl replaced the 2 M NaCl instillation. Mucosal blood flow was determined before discontinuing isotonic luminal gastric perfusion,

October

and 15 and 90 min after reestablishing this perfusion. Ht flux and PD were measured as outlined above. Data were obtained for 7 cats concerning net H+ flux and for 4 cats concerning PD in the antrum; all other measurements are given for 8 cats. Mucosal damage. In 16 cats, the gastric mucosa was exposed to 2 M NaCl for 10 min while discontinuing luminal perfusion. Apart from this, the experiments were conducted in a manner similar to that for the nondamage control group. Ten of these cats were killed 90 min after gastric mucosal exposure to 2 M NaCl. H+ flux data were obtained from 7 of these 10 cats, and data on PD in the corpus and antrum from 9 and 8 cats, respectively. To obtain information about early histologic changes, 6 cats from this group were killed 5 min after luminal exposure to 2 M NaCl. Mucosal damage and pentagastrin. Fifteen cats received a continuous infusion of pentagastrin (Peptavlon, Imperial Chemical Industries, Macclesfield, U.K.) at a rate of 16 pg. kg-’ . h-l, starting 60 min before induction of gastric mucosal damage by 2 M NaCl. This dose of pentagastrin is known to produce a submaximal stimulation of acid secretion in cats (8). Nine of the cats in this group were observed for 90 min after mucosal exposure to 2 M NaCl, whereas 6 cats were killed 5 min after gastric mucosal damage by 2 M NaCl. Otherwise the experimental protocol was the same as in the preceding groups. Mucosal damage and cimetidine. Fourteen cats received a bolus of 10 mg/kg cimetidine (Tagamet, Smith, Kline & French Laboratories, Philadelphia, Pa.), followed by a continuous infusion of cimetidine at a rate of 2 mg . kg-’ . h-l throughout the rest of the experiment, which gives a near complete inhibition of the acid secretion (8). Cimetidine administration started 60 min before

induction of gastric mucosal damage by 2 M NaCl. Eight of these animals were killed 90 min after mucosal damage by 2 M were were after

GASTRIC EPITHELIAL RESTITUTION AT LOW pH

1987

NaCl. Potential difference measurements in antrum obtained from 4 of these cats; all other measurements obtained from 8 cats. Six of the cats were killed 5 min mucosal damage for morphologic evaluation.

Statistics Analysis of variance for repeated measurements was used to determine statistical probabilities for changes in PD and H+ flux throughout the experiments (13). Analysis of variance was also used to evaluate changes in blood gases within and between groups. Two-way analysis of variance was performed for testing changes in arterial HCOB- concentration, cardiac output, mucosal blood flow, and morphology within and between groups exposed to 2 M NaCl using the BMDP;P2V program (14). Because morphology, classified as normal, damaged, or restituted epithelium, was quantitated as the percentage of the surface length of each microscopic section, these data were transformed to arcsine for the performance of analysis of variance. Each of the categories of normal, damaged, and restituted mucosal surface was subjected to separate twoway analysis of variance. When justified by preceding analysis of variance, contrast tests were used to calculate statistical probabilities within or between groups (15,16).

757

Probabilities of <0.05 were regarded as significant,

except for interaction effects that were regarded as significant at the 0.1 level. If not otherwise indicated, data are expressed as mean f SEM.

Results No

Mucosal Damage

In control cats without exposure to hypertonic saline, pharmacologic stimulation, or inhibition of acid secretion, the heart rate, mean aortic pressure, and cardiac output averaged 183 -+ 8 beats/min, 136 2

7 mmHg,

These

remained tential

and

343

variables, unaltered

difference

?

together

26

ml/min,

with

throughout

respectively.

blood

gas

values,

the experiments.

in the corpus

and antrum,

Po-

averag-

ing -49.4 ? 2.6 mV and -25.2 + 2.4 mV, respectively, also remained unaltered throughout the experiments as did H+ loss of 62 + 19 pmol/min

from the gastric outflow fluid. Gastric mucosal blood flow did not change significantly throughout the experiments (Table 1). Microscopy largely revealed

normal epithelium, apart from some small areas with epithelium that was characterized as damaged or restituted (Figure 1A and Figure 3, upper panel). Gastric Mucosal Damage Central hemodynamics. Heart rate and mean aortic pressure did not change significantly upon, or following, exposure to 2 M NaCl within or between groups, except that there was a small but significant increase (p < 0.001) in mean aortic pressure at the end of the experiments (Table 2). Cardiac output showed a small increase 15 min after damage (p < O.OOl), and then reached predamage levels 90 min after gastric damage. There were no differences between groups in these respects. Heart rate, mean aortic pressure, or cardiac output did not differ between groups in animals killed 5 min after exposure to 2 M NaCl. Blood gases. Table

Arterial

1. Gastric Mucosal Control Groupa

acidity

(pH transformed

Blood Flow in Nondamaged Afterb

Gastric

region

Beforeb

15 min

90 min

Antrum Corpus Fundus

0.71 k 0.13 0.74 k 0.12 0.52 If:0.12

0.81 "_ 0.12 0.71 * 0.10 0.61 2 0.13

0.89 2 0.13 0.69 2 0.09 0.58 T 0.11

n Mean 2 SEM in milliliters per minute per gram (n = 8). Blood flow was determined immediately before discontinued gastric luminal perfusion with saline at pH 1, and 15 and 90 min after reestablishing luminal perfusion. b Refers to measurements before and after exposure to 150 mM NaCl, which corresponds exposure to 2 M NaCl in the other groups.

to

758

GRijNBECH

ET AL.

GASTROENTEROLOGY

Vol. 93, No. 4

Table 2. Central Hemodynamic Variables in Cats With Gastric Mucosal Damage Caused by Exposure to 2 M NaCl

Groups

After damage

Before damage

15 min

90 min

ANOVA statistics

n

Baseline

10

173 + 7 180 ? 17 193 2 6

182 k 6 181 ? 14 198 ? 4

191 k 5 177 2 16 186 f 7

189 2 6 174 k 15 196 2 4

pb = 0.53 pw = 0.63 pi = 0.16

130 2 9 133 2 7 138 t 6

133 + 8 125 + 7 138 t 6

135 5 8 124 k 8 131 2 6

143 ? 6 137 f 7 144 2 8

pb = 0.73 pw < 0.01 pi = 0.27

298 f 30 307 t 24 332 L 37

312 ? 27 373 ? 35 393 2 32

300 f 21 339 2 24 331 2 52

pb = 0.43 pw = 0.01 pi = 0.50

Heart rate (beats per min) Damage alone Pentagastrin Cimetidine Mean aortic pressure Damage alone Pentagastrin Cimetidine

9

8 (mmHg) 10 9

8

Cardiac output (mJ/min) Damage alone Pentagastrin Cimetidine

10 9

8

Mean values ? SEM. Pentagastrin (16 pg * kg-’ * h-‘) and cimetidine (IO-mg bolus followed by 2 mg . kg-’ . h-‘) were given intravenously, starting 60 min before exposure to 2 M NaCl. With exception of the gastric damage period, the gastric lumen was perfused with saline at pH 1. pb, p,,,, and pi are the probabilities for differences between groups, within groups, and for interaction by analysis of variance (ANOVA), respectively.

to the concentration of Hf), Pacoz, and Pao2 did not show any significant changes over time or differences between groups. Bicarbonate concentrations, however, were elevated in pentagastrin-stimulated animals (p < 0.0005) compared with the other two groups exposed to 2 M NaCl. There was also a small but significant drop (p < 0.0005) in bicarbonate concentration in all groups after exposure to 2 M NaCl (Table 3). The cats killed 5 min after mucosal damage showed similar changes in bicarbonate levels under pentagastrin stimulation, and there was no difference between groups in other blood gas values in these cats. Potential difference. The PD was similar in the three exposed groups before the application of 2 M NaCl. The average corpus PD was approximately -45 mV and the antrum PD was about -25 mV. Shortly after the application of 2 M NaCl, the corpus and antrum PD dropped to about zero and +5 mV, respectively (Table 4). During the next 90 min, PD recovered and reached about half the predamage

Table 3. Bicarbonate Concentrations in Arterial

Blood

level at the end of the experiment. The initial level of PD and the subsequent drop after exposure to 2 M NaCl were similar in animals killed 5 min after mucosal injury. (Data not shown in Table 4.) Net H+ flux. Under nonstimulated conditions, net loss of H+ from the gastric fluid was increased (p < 0.005) during the first 40-min period after exposure to 2 M NaCl and later decreased (p < 0.05) during the last 40-min period (Table 5). The average net loss of H+ from gastric outflow fluid shifted after pentagastrin stimulation to a net gain of H+ (p < 0.001). After exposure to 2 M NaCl, the H’ flux changed (p < 0.001) to a mean net loss of H+, followed by a decrease (p < 0.05) in Hf loss at the end of the experiment (Table 5). The mean loss of Hf from gastric luminal perfusate in cimetidine-treated animals also increased after damage. The increase, however, was not statistically significant (Table 5). In cats killed 5 min after gastric mucosal injury, the predamage rate of H+ flux and the change after

of Cats With Gastric

Mucosal

Damage

Caused

by Exposure

to

2 M NaCl After damage Groups

n

damage

15 min

90 min

Damage alone Pentagastrin Cimetidine

10 9 8

15.7 2 0.4 18.5 t 0.7 15.1 + 0.5

14.2 2 0.5 16.6 + 0.5 13.8 ? 0.6

14.2 k 0.7 16.5 k 0.5 12.7 T 0.5

ANOVA statistics pb = 0.0007 PW = 0.0001 pi = 0.65

Mean values ? SEM in millimoles per liter. Pentagastrin (16 pg. kg-’ * h-‘) and cimetidine (lo-mg bolus followed by 2 mg kg-’ h-l) were given intravenously, starting 60 min before exposure to 2 M NaCl. With the exception of the gastric damage period, the gastric lumen was perfused with saline at pH 1. pb, p,,,, and pi are the probabilities by analysis of variance (ANOVA), respectively.

for differences

between

groups, within

groups, and for interaction

October

Table

GASTRIC

1987

4. Effect of Gastric Mucosal

Exposure

EPITHELIAL

to 2 M NaCl on Transmucosal

Potential

RESTITUTION

AT LOW pH

759

Difierence in Cats

After damage

Before damage

0 min

30 min

60 min

90 min

Groups

n

Before drugs

Antrum Damage alone Pentagastrin Cimetidine

8 9 4

-25.7 + 1.4 -23.5 f 1.9

-25.6 k 1.7 -29.9 * 2.2 -20.2 f 1.0

5.4 2 1.6 7.0 + 1.4 1.5 r 2.0

-1.7 2 1.6 -4.4 t 1.9 -5.8 k 2.4

-10.6 +- 2.3 -9.2 k 1.8 -12.6 + 3.6

-16.9 k 1.7 -14.2 k 2.1 -17.0 +- 4.3

Corpus Damage alone Pentagastrin Cimetidine

9 9 8

-44.8 IL 2.3 -42.7 2 2.4

-48.3 2 3.6 -48.8 f 2.8 -44.2 f 1.9

-0.1 + 1.5 -5.2 rfr1.4 0.5 2 2.1

-5.1 2 2.8 -12.4 * 1.4 -9.2 + 3.6

-17.0 * 3.4 -19.7 k 2.3 -17.1 k 3.5

-26.8 2 2.8 -27.9 f. 3.2 -23.8 f 3.4

Mean values + SEM in millivolts. Pentagastrin (16 pg. kg-’ * h-‘) and cimetidine (IO-mg bolus followed by 2 mg - kg-’ * h-‘) were given intravenously, starting 60 min before exposure to 2 M NaCl. The gastric lumen was perfused with saline at pH 1, except for a IO-min period of exposure to 2 M N&.

pentagastrin stimulation were similar to that described for the animals observed for 90 min. (Data not shown in Table 5.) Mucosal blood flow. A summary of the results is given in Table 6. In the antrum, mucosal blood flow was markedly increased 15 min after exposure to 2 M NaCl (p < 0.001). Blood flow subsided but did not completely reach predamage values 75 min later. The flow changes were not significantly influenced by pentagastrin or cimetidine treatment. In the corpus mucosa of both nonstimulated and cimetidine-treated animals, blood flow was significantly increased 15 min after application of 2 M NaCl (p < 0.001) and approached the predamage level after 90 min. Corpus mucosal blood flow of pentagastrin-treated animals was high even before damage was induced, and flow did not change significantly in this region after mucosal exposure to 2 M NaCl. Similarly, in the fundus region, mucosal blood flow was significantly increased (p < 0.001) 15 min after exposure to 2 M NaCl, and thereafter subsided but did not reach predamage level at the end of the experiment. Furthermore, flow was not significantly influenced by cimetidine or pentagastrin in the fundus region. In the cats killed 5 min after mucosal injury, there was no difference between groups in mucosal blood Table

5.

Efiect

of Gastric

Mucosal

Exposure

Groups

n

Before drugs

Damage alone Pentagastrin Cimetidine

7 9 8

-44 t 14 -26 + 22

to 2 M NaCI

flow in corresponding regions of the stomach. However, in these animals the level of blood flow in corpus mucosa was significantly higher in pentagastrin-stimulated compared to nonstimulated or cimetidine-treated animals. Microscopy. EARLY HISTOLOGY. Microscopy of specimens from the cats killed 5 min after gastric mucosal exposure to 2 M NaCl revealed damage of 49%-W% of the surface length (range of means of all examined stomach areas). Damage varied from complete denuding of the surface between pits with columnar cells lying in lumen clearly separated from the basal lamina, to areas with focal detachment of cells leaving gaps in the epithelial lining with apparently normal adjacent cells. In most areas, the epithelial changes included the superficial parts of the pits, whereas the glands appeared essentially normal. Little or no evidence of vascular damage, such as thrombosed vessels, hemorrhage, or capillary erythrostasis, was detected (Figure 1). Table 7 shows the distribution of damaged epithelium in stomachs of cats killed 5 min after exposure to 2 M NaCl. Analysis of variance did not reveal any difference in mucosal damage between groups in any part of the stomach. LATE HISTOLOGY. Ninety minutes after application of 2 M NaCl, 79%--99% (range of means) of the surface

lengths

on Net H+ Flux

of examined

From

preparations

the Stomach

had a com-

of Cats

After damage

Before damage

lo-50 min

50-90

-55 LII24 28 k 7 -17 2 19

-192 k 33 -59k 7 -46 2 10

-114 z!z38 -22 2 10 -52 rt 18

min

ANOVA statistics pw < 0.01 pw < 0.0005 D, > 0.25

Mean values 2 SEM in micromoles per minute. Pentagastrin (16 pg. kg-’ . h-‘1 and cimetidine (lo-mg bolus followed by 2 mg * kg-’ . h-‘) were given intravenously, starting 60 min before exposure to 2 M NaCl. The gastric lumen was perfused with saline at pH 1, except for a IO-min period of exposure to 2 M NaCl. Positive values represent net gain of H+ in the perfusing gastric luminal fluid; negative values represent net loss of H+ from this fluid. pw is the probability for differences within group by analysis of variance (ANOVA).

760

GRONBECH ET AL.

Table

GASTROENTEROLOGY Vol. 93, No. 4

6. Effect of Gastric Mucosal Exposure to 2 M NaCl on Regional Gastric Mucosal Blood Flow in Cats

n

Groups Antrum Damage alone Pentagastrin Cimetidine corpus Damage alone Pentagastrin Cimetidine Fundus Damage alone Pentagastrin Cimetidine

After damage

Before damage

15 min

ANOVA statistics

90 min

10

0.43

2 0.07

1.11

2 0.17

0.59

k 0.10

pb = 0.461

9

0.53

? 0.09

1.33

2 0.20

0.86

f

pw < 0.0001

8

0.58

* 0.09

1.11

2 0.16

0.81

-c 0.22

p, = 0.757

10

0.61

f

0.06

1.18

2 0.18

0.67

+ 0.14

pb = 0.283

9

1.04

IT 0.11

1.11

k 0.13

0.96

+ 0.14

pw < 0.0001

8

0.42

2 0.07

1.13

* 0.27

0.75

2 0.15

pi = 0.044

10

0.47

k 0.04

0.75

+ 0.09

0.54

IT 0.06

pb = 0.052

9

0.71

+ 0.08

0.89

+ 0.12

0.87

‘- 0.10

pw < 0.0001

8

0.61

2 0.07

0.93

+ 0.11

0.78

‘- 0.10

p, = 0.584

0.20

Mean values ? SEM in milliliters per minute per gram. Pentagastrin (16 pg. kg-’ . h-‘) and cimetidine (IO-mg bolus followed by 2 mg - kg-’ * h-‘) were given intravenously, starting 60 min before exposure to 2 M NaCl. The gastric lumen was perfused with saline at pH 1, except for a IO-min period of exposure to 2 M NaCl. Mucosal blood flow was determined immediately before instillation of 2 M NaCl, and 15 and 90 min after removal of 2 M NaCl. pb, p,,., and pi are the probabilities for differences between groups, within groups, and for interaction by analysis of variance (ANOVA), respectively.

thelium (black columns) than the pentagastrin group (lower middle panel) and the cimetidine group (bottom panel); p < 0.001 and p < 0.025, respectively. The amount of normal and restituted epithelium differed between areas of the stomach, but no difference was found between corresponding areas of the three groups exposed to 2 M NaCl.

plete epithelial lining; in 3%58% of the surface length the cells appeared normal, whereas the remaining surface with intact cellular lining showed shallow pits and cells between pits clearly different from normal surface cells. In some areas cylindrical cells were seen, but most cells were cuboidal or flattened with irregularly arranged nuclei. According to previously described criteria, these changes are taken as evidence of epithelial restitution and constituted 31%96% of the surface (Figure 2). Areas which, 90 min after hyperosmolar injury, still lacked a complete epithelial lining between the pits were considered still damaged and constituted 1%~21% of the surface. Figure 3 shows the distribution of surface epithelium according to the categories of normal, damaged, and restituted epithelium. The categories add up to 100% for each area. In this blind examination, the control group, not exposed to 2 M NaCl (top panel), showed small portions of abnormal epithelium. This quantity thus expresses nonspecific damage due to manipulation of the stomach as well as erroneous morphologic quantitation. At 90 min after exposure to 2 M NaCl, the nonstimulated group (upper middle panel) contained more damaged epi-

Table

of Damaged

7. Distribution

Gastric

Surface

Discussion This study evaluates the restitution process of damaged gastric mucosa in the presence of a strongly acidic milieu in the stomach. The most important finding is that epithelial restitution did occur to a large extent over a 1.5-h period after damage. Exposure of the gastric mucosa to solutions with a low pH leads to back-diffusion of H+ into the mucosa, and it is well established that such backdiffusion is enhanced when the gastric mucosal barrier is disrupted (17,18). Hypertonic solutions are able to break the mucosal barrier (19). We have recently shown that this can be achieved using 2 M NaCl for 10 min in the in vivo cat model (4). The present experiments showed a substantial net loss of Hf from the stomach during nonstimulation and

Epithelium

in Cats 5 min After Exposure to 2 M NaCl

Anterior wall CorDus

Fundus

62 + 14

76 + 13

63 ? 11

98 t

1

88 2 7

59 t

87 + 10

80 t

56 + 9

99 r

1

68 ” 11

68 k 15

49 2 12

99 k 1

84 k 8

54 + 13

Groups

n

Antrum

Damage alone Pentagastrin Cimetidine

6 6

6

66 +

Mean

values

f

SEM in percentage

Posterior wall

99 + 1

14

of surface

12

length

of examined

sections.

The remaining

Antrum

CorDus

surface

was judged

as normal.

Fundus 10

GASTRIC EPITHELIAL RESTITUTION AT LOW pH

October 1987

ANTERIOR

WALL

, ,CORPUS, ,ANTRUM

,FUNDUS

I ,

150 mM 100 r

4 9

k

60

THERMAL

EPITHELIUM

1

DAMAGED

, PDSTERIOR WALL ,ANTRUM , ,CORPUS,

2M 100 -

F

60.

EPlTHELlUM

d RESTITUTED

EPITHELIUM

Figure 3. Distribution of normal, damaged, and restituted gastric surface epithelium in control cats (top panel; n = 8), 90 min after exposure to 2 M NaCl (upper middle panel; n= lo), during pentagastrin stimulation (lower middle panel; n = g), and in cimetidinetreated cats (bottom panel; n = 8). The sum of normal, damaged, and restituted epithelium equals 100% for each area. Bars represent SEM.

NaCI

80 .

LO

(

NaCl

.

5 ti

,FUNDUS

761

.

20 . 2M NaCI and

PENTAGASTRIN

LO 20 1 2 M Na Cl and CIMETIDINE

inhibition of acid secretion, whereas stimulation with pentagastrin evoked a net secretion of H+. Substantial evidence has been provided that the intact gastric mucosa secretes small amounts of HC03- (20,211. Svanes et al. (1)also showed that there is a considerable passive flux of HC03- from damaged mucosa into the luminal solution of the in vitro frog gastric mucosa. A magnitude of luminal alkalinization as large as in our experiments, however, indicates that back-diffusion must have been the main source of H+ loss (Table 5).The marked increase of back-diffusion after gastric damage in nonstimulated cats, and a corresponding change from net secretion to net back-diffusion in stimulated cats fits well with earlier observations (22-24). The magnitude of H+ back-diffusion obtained in the present experiments is somewhat higher than in earlier reports (18,19,23,24). This discrepancy can be explained by the constant pH of 1 obtained by continuous titration in our study, whereas the reported results were obtained by intermittent instil-

lation and withdrawal of fluid from the stomach which lead to an increase in luminal pH during intervals between the instillations. The H+ concentration gradient between the luminal fluid and the blood will, therefore, on average, be higher in our experiments than in the reported studies. Many previous studies have shown that damage to the gastric surface epithelium is associated with a decrease in PD (9,25). Evidence has also been provided that there is a connection between recovery of PD and functional and structural recovery of gastric mucosa after damage (1,4,24,26). In the present study, the transmural PD was almost abolished shortly after the application of 2 M NaCl in all three exposed groups of animals. Concomitant and abrupt reduction in PD and increase in back-diffusion, followed by gradual recovery toward the end of the experiments, strongly suggest functional recovery of the gastric mucosa after mucosal damage by 2 M NaCl. Mucosal blood flow increased in all regions of the

762

GRijNBECH

ET AL.

stomach after exposure to 2 M NaCl, in all experimental groups, except for the corpus mucosa in pentagastrin-stimulated animals where blood flow was high already before damage and did not increase further (Table 6). The increase in mucosal blood flow after damage is in the agreement with previous observations demonstrating increased mucosal blood flow after gastric mucosal damage when the stomach contains fluid with high H+ concentration (22,23,27-32). The effect of pentagastrin on corpus mucosal blood flow is in agreement with previous observations obtained at pH 7 (48). Antral and fundus blood flow rates also increased after damage in pentagastrin-stimulated animals, probably because these regions contain few or no parietal cells and therefore are less influenced by pentagastrin than the corpus region. Although our experiments confirm previous findings concerning mucosal blood flow after mucosal damage, they also provide new information by showing that the cimetidine-treated mucosa has the same ability to increase blood flow as the nonstimulated one. It should also be emphasized that a high and similar level of mucosal blood flow was reached after application of 2 M NaCl in all the experimental groups. This indicates that the mucosal blood flow approached a near maximal level after damage, independent of pharmacologic H+ stimulation or inhibition. Considerable evidence has been provided that vasoactive substances like histamine and prostaglandins are produced in damaged gastric mucosa (17,18,33-39). Increased mucosal blood flow after gastric damage in our experiments could be explained by the release of vasoactive mediators, but could also be due to a direct response of the vessels to increased tissue concentration of H+ during backdiffusion after mucosal damage. Our data do not indicate which mechanism was the most important one. The increase in cardiac output after mucosal damage suggests, however, that a release of vasoactive substances had occurred. In the present study arterial blood HCOB- concentration was elevated in pentagastrin-treated animals. This is consistent with earlier observations (40) and is probably a result of an increased alkaline tide when gastric acid secretion is stimulated. According to previous reports, the stimulated gastric mucosa is better protected against back-diffusing H+ and erosions than the nonstimulated and inhibited mucosa (41,42). Furthermore, it has also been shown that elevated plasma levels of HC03protect against shock-induced mucosal erosions (43,~). On the other hand, previous reports have noted the ability of cimetidine both in antisecretory and nonantisecretory doses to protect against erosions induced by

GASTROENTEROLOGY

Vol. 93, No. 4

aspirin in acidic solutions (45,46), although the mechanisms involved are not established. Our experiments in which the animals were killed 5 min after gastric mucosal exposure to 2 M NaCl did not reveal any protective effect of pentagastrin or cimetidine against hyperosmolar injury (Table 7). Ninety minutes after exposure to 2 M NaCl, however, the epithelial lining was nearly complete and very similar in pentagastrin- and cimetidine-treated cats, whereas about 15% of the mucosa lacked epithelial lining at 90 min after damage in nonstimulated animals (Figure 3). It is known that the process from damaged through restituted and finally normal epithelium is a rapid one (2,3), and an unknown amount of epithelium that initially was damaged may appear normal 90 min later. The amount of mucosal surface uncovered by epithelium (i.e., damaged) 90 min after injury is therefore the best parameter for comparison of morphologic mucosal recovery between groups in the present study. As the amount of mucosal damage at 5 min after hyperosmolar injury was similar in all groups, we conclude that both pentagastrin and cimetidine supported epithelial restitution. Our data do not allow conclusions about the mechanisms by which the drugs exert their effect on mucosal recovery. However, the very similar levels of mucosal blood flow after mucosal injury show that this support of epithelial restitution cannot be accounted for by differences in mucosal blood flow (Table 6). The main finding in the present study, however, was that epithelial restitution clearly occurred to a high degree at luminal pH 1 and during a substantial back-diffusion of H+, irrespective of the secretory state of mucosa. According to previous reports, epithelial restitution after mucosal damage was inhibited in the in vitro frog gastric mucosa at luminal pH 3, and in the in vitro guinea pig mucosa at pH 1 (6,7). In our previous work with the same animal model, and with a gastric luminal pH at 7.4, we found that 98% of the mucosa had a complete epithelial covering 90 min after damage by 2 M NaCl. A reasonable interpretation is therefore that luminal pH cannot be the only decisive factor for epithelial restitution after gastric mucosal damage under in vivo conditions. In conclusion, this in vivo study shows that gastric damage by hypertonic saline is followed by functional and structural recovery of the mucosa within 90 min, even in the presence of a strong acidic luminal solution. Both pentagastrin and cimetidine seemed to support epithelial restitution.

References 1. Svanes surface

K, Ito S, Takeuchi K, Silen W. Restitution of the epithelium of the in vitro frog gastric mucosa after

October 1967

damage with hyperosmolar sodium chloride. Gastroenterol0gy 1982;82:1409-26. 2. Rutten MJ, Ito S. Morphology and electrophysiology of guinea pig gastric mucosal repair in vitro. Am J Physiol 1983; 244:G171-82. 3. Lacy ER, Ito S. Ethanol-induced insult to the superficial rat gastric epithelium: a study of damage and rapid repair. In: Allen A, Flemstrom G, Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1984;49-55. 4. Dzienis H, Gronbech JE, Varhaug JE, Lekven J, Svanes K. Regional blood flow and acid secretion associated with damage and restitution of the gastric surface epithelium in cats. Eur Surg Res 1987;19:9a-112. 5. Critchlow J, Magee D, Ito S, Takeuchi K, Silen W. Requirements for restitution of the surface epithelium of frog stomach after mucosal injury. Gastroenterology 1985;88:237-49. 6. Svanes K, Takeuchi K, Ito S, Silen W. Effect of luminal pH and nutrient bicarbonate concentration on restitution after gastric surface cell injury. Surgery 1983;94:494-500. 7. Rutten MJ, Ito S. Luminal acid effects on reconstitution of damaged guinea pig gastric mucosa in vitro. In: Allen A, Flemstrom G, Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1964;41-7. a. Varhaug JE, Svanes K, Lekven J. Regional gastric blood flow and acid secretion during Hz receptor blockade in cats. Stand J Gastroenterol 1983;18:977-86. 9. Geall MG, Phillips SF, Summerskill WHJ. Profile of gastric potential difference in man. Gastroenterology 197o;sa: 437-43. 10. Heyman MA, Payne BD, Hoffman JI, Rudolph AM. Blood flow measurements with radionuclide-labeled particles. Prog Cardiovasc Dis 1977;29:55-79. 11. Segadal L, Svanes K. Evaluation of the microsphere-method for determination of cardiac output. Stand J Clin Lab Invest 1979;39:415-21. 12. Varhaug JE, Svanes K, Svanes C, Lekven J. Gastric blood flow determination: intramural distribution and arteriovenous shunting of microspheres. Am J Physiol 1964;247:G466-79. 13. Wallenstein S, Zucker CI, Fleiss JL. Some statistical methods useful in circulation research. Circ Res 1980;47:1-9. 14. Dixon WJ, Brown MB, Engelman L, et al. BMDP statistical software. Berkeley, Calif.: University of California, 1983. 15. Zar JH. Biostatistical analysis. 2nd ed. Englewood Cliffs, N.J.: Prentice-Hall, 1964. 16. Winer BJ. Statistical principles in experimental design. 2nd ed. Tokyo: McGraw-Hill Kogakusha, 1971:514-603. 17. Davenport HW. Fluid produced by the gastric mucosa during damage by acetic and salicylic acids. Gastroenterology 1966; 50:487-99. la. Davenport HW. Protein-losing gastropathy produced by sulfhydryl reagents. Gastroenterology 1971;69:679-9. 19. Harper AA, Reed JD, Smy JR. Effects of intragastric hyperosmolar solutions on gastric function. J Physiol 1970;299: 453-72. 20. Allen A, Garner A. Mucus and bicarbonate secretion in the stomach and their possible role in mucosal protection. Gut 1980;21:249-62. 21. Flemstrom G. Gastric secretion of bicarbonate. In: Johnson LR, Christensen J, Grossman M, Jacobson ED, Schultz SG, eds. Physiology of the digestive tract. New York: Raven, 1981: 603-16. 22. Ritchie WP. Acute gastric mucosal damage induced by bile salts, acid, and ischemia. Gastroenterology 1975;66:699-707. 23. Bruggeman TM, Wood JG, Davenport HW. Local control of

GASTRIC EPITHELIAL RESTITUTION AT LOW pH

763

blood flow in the dog’s stomach: vasodilatation caused by acid back-diffusion following topical application of salicylic acid. Gastroenterology 1979;77:736-44. 24. Kelly DG, Code CF, Lechago J, Bugajski J, Schlegel JF. Physiological and morphological characteristics of progressive disruption of the canine gastric mucosal barrier. Dig Dis Sci 1979;24:424-41. 25. Rehm WS. Positive injury potentials of the stomach. Am J Physiol 1944;149:729-5. 26. Stern AI, Hogan DL, Isenberg JI. A new method for quantitation of ion fluxes across in vivo human gastric mucosa: effect of aspirin, acetaminophen, ethanol, and hyperosmolar solutions. Gastroenterology 1964;66:60-79. 27. Augur NA. Gastric mucosal blood flow following damage by 197o;sa: ethanol, acetic acid, or aspirin. Gastroenterology 311-20. 28. Cheung LY, Moody FG, Reese RS. Effect of aspirin, bile salt, and ethanol on canine gastric mucosal blood flow. Surgery 1975;77:786-92. 29. Cheung LY, Chang N. The role of gastric mucosal blood flow and H+ back-diffusion in the pathogenesis of acute gastric erosions. J Surg Res 1977;22:357-61. 30. Puurunen J. Gastric mucosal blood flow in ethanol-induced mucosal damage in the rat. Eur J Pharmacol 1980;63:275-80. 31. Starlinger M, Schiessel R, Hung CR, Silen W. H+ back diffusion stimulating gastric mucosal blood flow in the rabbit fundus. Surgery 19ai;a9:232-6. 32. McGreevy JM, Moody FG. Focal microcirculatory changes during the production of aspirin-induced gastric mucosal erosions. Surgery i981;89:337~i. 33. Johnson LR. Histamine liberation by gastric mucosa of pylorus-ligated rats damaged by acetic or salicylic acids. Proc Sot Exp Biol Med 1966;121;364-6. 34. Johnson LR, Overholt BF. Release of histamine into gastric venous blood following injury by acetic or salicylic acid. Gastroenterology 1967;52:505-9. 35. Assouline G, Leibson V, Danon A. Stimulation of prostaglandin output from rat stomach by hypertonic solutions. Eur J Pharmacol 1977;44:271-3. 36. Whittle BJR. Mechanisms underlying gastric mucosal damage induced by indomethacin and bile-salts, and the actions of prostaglandins. Br J Pharmacol 1977;60:455-60. 37. Knapp HR, Oelz 0, Sweetman BJ, Oates JA. Synthesis and metabolism of prostaglandins Ez, Fza and Dz by the rat gastrointestinal tract. Stimulation by a hypertonic environment in vitro. Prostaglandins 1978;15:751-7. 38. Soper BD, Tepperman BL. 16,16 Dimethyl prostaglandin E, reduces histamine release from the stomach and protects the gastric mucosal barrier altered by ethanol in neutral solution. Prostaglandins 1979;18:331--9. 39. Kauffman GL, Whittle BJR. Gastric vascular actions of prostanoids and the dual effect of arachidonic acid. Am J Physiol 1982;242:G582-7. 40. Albinus M, Sewing KF. Acid base balance in relation to gastric acid secretion stimulated by histamine and pentagastrin in cats. Naunyn Schmiedebergs Arch Pharmacol 1974; 285:325-35. 41. O’Brien P, Silen W. Influence of acid secretory state on the gastric mucosal tolerance to back diffusion of H+. Gastroenterology 1976;71:760-5. 42. Smith P, O’Brien P, Fromm D, Silen W. Secretory state of gastric mucosa and resistance to injury by exogenous acid. Am J Surg 1977;133:61-5. 43. Starlinger M, Jakesz R, Matthews JB, Yoon C, Schiessel R. The relative importance of HCOBm and blood flow in the protec-

764

GRijNBECH ET AL.

tion of rat gastric mucosa

GASTROENTEROLOGY Vol. 93. No. 4

during

shock. Gastroenterology

1981;81:732-5. 44. Kivilaakso

E. High plasma HCO,- protects gastric mucosa against acute ulceration in the rat. Gastroenterology 1981;81: 921-7. 48. Ku&man GL, Grossman MI. Prostaglandin and cimetidine

inhibit the formation of ulcers produced by parenteral salicylates. Gastroenterology 1978;75:1099-102. 46. Guth PH, Ames II, Paulsen G. Topical aspirin plus HCl gastric lesions in the rat. Cytoprotective effect of prostaglandin, cimetidine, and probanthine. Gastroenterology 1979; 76: 8893.