Plasma gastrin and gastric enterochromaffinlike cell activation and proliferation

GASTROENTEROLOGY

1986;90:391-9

Plasma Gastrin and Gastric Enterochromaffinlike Cell Activation and Proliferation Studies With Omeprazole Intact and Antrectomized

and Ranitidine Rats

in

HAKAN LARSSON, ENAR CARLSSON, HILLEVI MATTSSON, LARS LUNDELL, FRANK SUNDLER, GUNHILD SUNDELL, and BJbRN WALLMARK, TAKEHIKO WATANABE, ROLF HAKANSON Department of GI Pharmacology, HBssle Research Laboratory, M6lndal; Departments of Histology and Pharmacology, University of Lund, Lund; Department of Surgery II, Sahlgren’s Hospital, Gothenburg, Sweden; and Department of Pharmacology, Osaka University Medical School, Osaka, Japan

Unoperated female rats were subjected to daily oral treatment with omeprazole (10 or 400 pmollkg body wt), ranitidine (175 + 175 + 350 PmoJlkg body wt), or vehicle and antrectomized rats were treated with omeprazole (400 ,umollkg body wt) or vehicle. After 10 wk of treatment, plasma gastrin levels were high in unoperated rats treated with the high omeprazole dose and with ranitidine, and low in antrectomized controls. Plasma gastrin levels were slightly higher in the low-dose omeprazole group than in the intact controls. In antrectomized rats treated with the high dose of omeprazole, the plasma gastrin level was in the same range as in intact control rats. A close correlation (r = 0.89, p < 0.0001) wasfound between the plasma gastrin level and the oxyntic mucosal enterochromaffinlike cell density [as well as the Received May 23, 1985. Accepted July 29, 1985. Address requests for reprints to: Hhkan Larsson, Ph.D., Department of GI Pharmacology, AB HBssle, S-431 83 MBlndal, Sweden. This work was supported by the Swedish Medical Research Council (04x-1007 and 04x-4499) and A PBhlssons’ foundation. Preliminary reports of parts of this study have appeared in abstract form (Hepatogastroenterology 1985;32:48 and Gastroenterology 1985;88:1464). The authors thank Kjell Andersson, Marie-Louise Berglund, Marita Blom. John Briiutigam, D.V.M., Kerstin Carlsson, Berit Elander. Eva Engelbert, Jan Fryklund, Lena Gunnarsson, Inger Hartmark, Inger Lind, Birgitta Ryberg, Marie Strimfors, Lennart Svensson, and Eva Varai for expert technical assistance. They also thank Dr. Ika Hirsch and Dr. Bengt Orebick for help with the statistical evaluation, Elsa Svensson for preparing the figures, and Bodil Sederquist for manuscript preparation. 0 1986 by the American Gastroenterological Association 0016-5085/86/$3,50

tissue levels of histidine decarboxylase and histamine in the oxyntic mucosa) in all groups. The somatostatin cell density in the oxyntic mucosa was not altered by the various treatments. During a recovery period of 10 wk after the IO-wk treatment, the enterochromafinlike cell density and histamine concentration decreased by 30%-40% in the rats treated with the high dose of omeprazole, whereas the corresponding values increased by 50% and 40% respectively, in the control rats. The difference between the two groups, however, was still statistically significant. Plasma gastrin levels and gastric histidine decarboxylase activity returned to control values during recovery. The results suggest that the observed changes in enterochromaffinlike cell density are related to the plasma gastrin levels and that they are reversible. It is concluded that neither omeprazole nor ranitidine per se is likely to induce proliferation of enterochromaffinlike cells. Oral administration of omeprazole (H168/68) to rats over a time period of 24 mo resulted in diffuse endocrine cell hyperplasia and focal neoplasia, characterized as carcinoids, in the gastric corpus (oxyntic gland area). In the highest-dose groups (400 pmol/kg body wt), 40% of the female rats and 10% of the male rats were reported as having carcinoids (1). The hyperplastic and neoplastic cells were idenAbbreviations HDC, histidine

used in this decarboxylase.

paper:

ECL. enterochromaffinlike;

392 LARSSDNET AL.

Table

1. Experimental

GASTROENTEROLOGY Vol. 90, No. 2

Groups

Group Treatment

surgery

(oral)

No.

n

1 2

20 10

None None

0.25% Methocel” Omeprazole 1 x 10 pmolikg

3

20

None

Omeprazole

body body 4

10

None

5 6

11 11

Antrectomy Antrectomy

wt

1 x 400 pmolikg

wt

Ranitidine 2 x 175 +l x 350 pmolikg body wt 0.25% Methocel Omeprazole 1 x 400 wmolikg body wt

Details on the mode of administration of the drugs are given in Methods. ” 90 HG 15000, Dow Corning Corp., Midland, Mich.

tified as enterochromaffinlike (ECL) cells by silver staining [Grimelius (+), Sevier-Munger (+), and Masson (-)] and by electron microscopy (1, cf. 2). The ECL cells are endocrine cells known to contain large amounts of histamine and histidine decarboxylase (HDC) in the rat and to be functionally and trophically controlled by gastrin (2). They have a characteristic ultrastructure with numerous vesicular-type cytoplasmic granules that have a small electron-dense core separated by a large electron-lucent halo from the limiting membrane. Prolonged elevation of the circulating gastrin concentration causes proliferation of the ECL cells within 6-10 wk (2). The secretion of gastrin from the antrum, which is the major source of gastrin in the body, is stimulated by nervous activity, and by peptides and amino acids in food, and is inhibited by a low antral pH. Food in the stomach in combination with high antral pH will stimulate secretion of gastrin into the bloodstream. Therefore, it could be anticipated that a long-lasting inhibition of gastric acid secretion would cause a sustained high plasma level of gastrin, leading in turn to proliferation of ECL cells in the oxyntic mucosa. The aim of the present study was to investigate whether ECL cell proliferation occurred after prolonged treatment with antisecretagogues and to elucidate the mechanism behind such proliferation and its reversibility. The study included antrectomized rats, in which the major source of gastrin had been eliminated.

rat food pellets and water. The light cycle was 12 h light to 12 h darkness and the room temperature was kept between 21” and 24°C. Antrectomy type Billroth I [gastroduodenostomy end-to-end (groups 5 and S)] was performed (3,4) under Rompun (Bayer, Leverkusen, F.R.G.)/Ketalar (Parke-Davis, Morris Plains, N.J.) anesthesia (0.12 ml/l.0 ml per kg body wt, i.p.) after 2 days of starvation [free access to water and Ringer’s glucose (ACO, Sweden)]. Immediately after surgery the animals received 10 ml of Ringer’s glucose subcutaneously and Doctacilline (Astra, Sweden) 0.1 ml intramuscularly (125 mgiml). If necessary, Ringer’s glucose was given subcutaneously during the first 3 days after surgery. Food was given approximately 4 days after surgery. Before assignment to groups, blood samples for determination of plasma gastrin were taken from all the antrectomized animals to verify the antrectomy. Treatment was started 10 days after surgery. The rats were randomly assigned to groups according to Table 1.All animals were treated for 10 wk and all, except for those in group 4, received vehicle or suspension orally by gavage once daily, at 10 AM to 12 noon. Group 4 received ranitidine (Zantac, Glaxo Laboratories, Greenford, England) orally by gavage three times daily, at 7 AM (175 pmolikg body wt), 12 noon (175 pmollkg body wt), and at 5 PM (350 pmolikg body wt). The animals were given 5 ml of suspension per kilogram body weight, which was individually determined once a week. After 10 wk of treatment the animals, except for 10 animals each in groups 1 and 3, were killed 2 h after the last dose by bleeding under COz anesthesia. Blood w’as collected from the neck for determination of plasma gastrin. The stomach was opened along the greater curvature, washed in saline, weighed, and pinned to a cork plate. Specimens from the corpus (oxyntic gland area) were collected. Each specimen was fixed for histologic examination and immunocytochemical staining. Sections from the oxyntic gland area were stained with antiserum against HDC to visualize ECL cells and with antiserum against somatostatin to visualize the somatostatin cells. Mucosal thickness was measured in sections stained with hematoxylin and eosin. The remainder of the oxyntic mucosa was scraped off and homogenized for determination of histamine levels and HDC activity. Recovery groups (10 animals each in groups 1 and 3) were killed 10 wk after treatment was discontinued. Blood samples for plasma gastrin determination were drawn from the tip of the tail 2 h after administration of the suspensions during the 2nd and 10th week of treatment and on days 3,6, and 10 after discontinuation of treatment in these two groups. Determination

Materials

and Methods

General Female Sprague-Dawley rats (MDllegaards Breeding Centre, Skensved, Denmark), 160-250 g in body weight, were used. The animals were kept in Macrolon cages, 2 or 3 animals in each, with free access to standard

of Plasma

Levels

of Gastrin

As already indicated, gastrin was determined in plasma obtained from blood sampled at death (all groups) and also during and after the treatment period in the two recovery groups. The animals were freely fed during all sampling occasions. At death about 2 ml of blood was collected in Venoject (ethylenediaminetetraacetic acid) test tubes and centrifuged. The plasma was stored frozen until analysis. In the recovery groups (during and after

February 1986

GASTRIN AND ECL CELL PROLIFERATION IN THE RAT 393

treatment) about 150 ~1 of blood was collected from the tip of the tail in small preweighed Eppendorf tubes containing 100 ~1 of ethylenediaminetetraacetic acid (3.8 mM). After centrifugation the plasma was separated and stored frozen until analysis. Gastrin was determined using a doubleantibody, liquid-phase, lzsI radioimmunoassay (Diagnostic Products Corporation, Los Angeles, Calif.). The concentration was expressed in terms of equivalents of the standard, synthetic human gastrin I. The gastrin levels in blood samples drawn from the tail were 10%~20% lower than in blood samples collected at death. The reason for this discrepancy is not known.

supernatant, essentially according to the enzymatic isotopic assay described by Snyder (9). Aliquots of 100 ~1 of the supernatant were added to polypropylene tubes and mixed with [‘4C]S-adenosylmethionine (40 nCi and 0.7 nmol, New England Nuclear, Boston, Mass.), [“Hlhistamine (200 nCi and 0.025 nmol, New England Nuclear), 50 ~1 of histamine-N-methyltransferase [prepared as described by Snyder (911, and 0.01 M sodium phosphate buffer (pH 7.4) to a total volume of 1.0 ml. The tubes were incubated at 37°C for 60 min and the reaction was stopped by addition of 0.5 ml of 4 M NaOH. The samples were saturated with NaCl and the [‘“C-“Hlmethylhistamine formed was extracted with 4.0 ml of chloroform. The organic phase was washed once with 1 ml of 2 M NaOH, and 2.0 ml of the chloroform extract was transferred to glass scintillation vials and evaporated to dryness under air. Scintillation liquid (Ready solve EP, Beckman Industrial Corp., Cedar Grove, N.J.) was added and the ratio between 14C and “H was determined in a Mark III liquid scintillation spectrometer. In each series of assays, both blanks, containing all reagents and buffer instead of homogenate, and histamine standards (25-200 pg histamine/ml) were included. Duplicate tests were run on all samples and the results were expressed as micrograms of histamine per gram tissue wet weight.

Determination of Enterochromafinlike and Somatostatin Cell Density

Cell

Specimens of the oxyntic gland area (diameter 4 mm) were immersed in 4% buffered formaldehyde (pH 7.2), fixed overnight, washed repeatedly in sucroseenriched buffer, frozen, and sectioned in a cryostat; the sections (8 pm) were cut perpendicular to the mucosal surface and processed for immunocytochemical demonstration of HDC (5,6) using an antiserum raised against fetal rat liver HDC (5). The antiserum was applied in a dilution of 1:1600 at 4°C for about 20 h. Alternatively, the sections were exposed to an antiserum against somatostatin (K 18,Milab, Malmo, Sweden) in a dilution of 1:800 (7).In both cases, the site of the antigen-antibody reaction was visualized by the indirect immunofluorescence method of Coons et al. (8). Antiserum, inactivated by the addition of excess amounts of antigen (50-100 pgiml diluted antiserum), was used for controls. Immunofluorescent cells were counted in 2-5 randomly selected visual fields (entire thickness of mucosa visible) from each section. At least three sections from each specimen were examined. Cell counts were expressed as the number of cells per visual field. Examination was performed with eyepiece 12.5~ and objective 10X (visual field diameter 1.2 mm). Determination Thickness

of Oxyntic

Mucosal

Formalin-fixed, paraffin-embedded specimens were sectioned and the sections were stained with hematoxylin and eosin. At least 10 determinations of mucosal thickness (expressed in micrometers) were made on at least two sections from each animal. For examination we used eyepiece 8X and objective 6.3~ (visual field diameter 2.5 mm). Determination Histamine

of Oxyntic

A4ucosal

Histidine

Eighty microliters of the oxyntic mucosal homogenate was incubated with [l-*4C]L-histidine (20 nCi, New England Nuclear), 5 X 10m4 M t.-histidine, and lo-” M pyridoxal-5-phosphate in a total volume of 160 ~1 at 37°C for 60 min according to the microprocedure described by Beaven et al. (10). The 14C02 formed was trapped in 50 ~1 of Protosol (New England Nuclear) after addition of 80 ~1 of 2 M perchloric acid and reincubation for 30 min. Radioactivity was measured by liquid scintillation counting, and the HDC activity was expressed as picomoles of CO2 formed per milligram of wet weight per hour. In each series of assays, blanks were included containing all reagents and 0.01 M sodium phosphate buffer, pH 7.4, instead of tissue homogenate. Duplicate tests were run on all samples and blank values were subtracted before calculation of HDC activity. Statistics Statistical evaluation of the results was made by using the nonparametric Mann-Whitney U-test.

Genera]

was determined

of Oxyntic Activity

Results

Mucosa]

The oxyntic mucosa was scraped off, weighed, and homogenized in ice-cold 0.01 M sodium phosphate buffer, pH 7.4, to a concentration of 100 mg wet wt/ml. The homogenate was diluted 1: 10 in the same phosphate buffer and heated in boiling water for 10 min to release bound histamine. The homogenate was then centrifuged at 6000 g for 20 min and histamine

Determination Decarboxylase

in the

Comments

In groups 5 and 6, animals of slightly different ages were included. Half of the animals in each of these groups were about 3 mo older than the other half. At the start of treatment the older rats weighed 274 ? 4.5 g and 272 t 11.2 g and the younger rats weighed 182 ? 9.7 g and 185 t 3.7 g in groups 5 and 6, respectively. At the end of the treatment period

394

Table

GASTROENTEROLOGY Vol. 90, No. 2

LARSSON ET AL.

2. Body Weight, Stomach Weight, Mucosal Thickness at Death

Group No.

Stomach weight (9)

Body weight (g)

1 276 T 2 287 + 3 278 " 4 276 2 5 All 290 + Young 282 t (II = 6) Old 299 ?

6 6 7 4 4 4

P

and

effects were noticed any of the groups.

Oxyntic

Mucosal thickness (pm1

P

1.68r!z 0.02 500 I?13.7 1.812 0.05 NS 513 -r10.5 NS 1.99f 0.07 CO.01 611 f 28.9
6

6 Al!*= 5, 302" 7 Young 288 2 9 (n = 6]

'

487 5 16

NS

Old 320 2 6 (n = 5) Values are mean f SEM (n = 10-11).Mann-Whitney U-test. Statistical significance of difference versus the intact controls (group 1)indicated by p values. L1The stomach weights of antrectomized animals were not recorded.

the differences in body weights between the groups were small, but a tendency still remained toward higher body weights in the two groups of antrectomized animals (Table 2). In the following, groups 5 and 6 will be regarded as homogeneous. The drugs were well tolerated, as reflected by the normal body weight in treated groups. No adverse

Ezlantrectomized

during the treatment

period

in

Plasma Gastrin Levels The plasma gastrin levels determined at death are shown in Figure 1. Low-dose omeprazole produced a slight but significant increase, whereas greater increases were observed in the high-dose omeprazole group and in the ranitidine group. Antrectomized controls [group 5) had significantly lower gastrin levels than intact controls. Gastrin levels in antrectomized animals treated with highdose omeprazole (group 6) did not differ from those in intact controls, but were higher than in the antrectomized controls. Plasma gastrin levels during treatment with highdose omeprazole and after treatment was discontinued are shown in Figure 2. Untreated animals had fairly constant gastrin levels of around 150 pg/ml. During treatment with omeprazole, gastrin increased to >2000 pg/ml in week 10. After treatment was discontinued the gastrin levels rapidly decreased, already reaching control levels after 610 days. Gross Morphology

of the Stomach

Stomach weights were determined in groups l-4. In intact controls (group 1) the weight of the stomach was 1.68 ? 0.02 g. An increase to 1.99 + 0.07 g was seen in group 3 (omeprazole 400 pmol/kg body wt, p < 0.01). The thickness of the oxyntic mucosa (Table 2) was found to be 500 t 13.7 pm in the intact controls. Statistically significant differences from this value were found in the high-dose omeprazole group (group 3, 611 * 28.9 pm, p < 0.01) and in the antrectomized controls (group 5,453 + 14.3 wrn, p < 0.05).

***

I,

I Treatment Figure

1 Plasma gastrin levels after 10 wk omeprazole, ranitidine, or vehicle. mined in blood samples collected SEM, n = 10-11.**p i 0.01; ***p

of treatment with Gastrin was deterat death. Mean * < 0.001.

period

1 Recovery

period

Figure 2. Plasma gastrin levels during the 2nd and 10th week of treatment with high-dose omeprazole and on the 3rd, 6th, and 10th day after treatment was discontinued. Gastrin was determined in blood samples taken from the tail (see Methods]. Mean t SEM; n = 10.

February

1986

GASTRIN

Effects on Histidine Decarboxylase and Histamine Content in Oxyntic Mucosa

Pm01 CO2. mg wet weight-’

The HDC activity (Figure 3A) and histamine concentration (Figure 3B) in the oxyntic mucosa were found to change roughly in parallel with the pattern observed for plasma gastrin levels, i.e., not differently from intact controls in group 6 (antrectomized, omeprazole 400 pmolikg body wt), a small increase in group 2 (omeprazole 10 pmolikg body wt) and group 4 (ranitidine), a larger increase in group 3 (omeprazole 400 pmol/kg body wt), and a decrease in group 5 (antrectomized controls). After a lo-wk recovery period the HDC activity was back to the control level in the animals treated with the high-dose omeprazole (Table 3). During the recovery period the histamine concentration was reduced by 36% in the omeprazole-treated animals, whereas it was increased by 38% in the corresponding controls. Nonetheless, the difference between the two groups was statistically significant. Effects on Enterochromafinlike Somatostatin Cell Density

AND ECL CELL PROLIFERATION

IN THE RAT

395

h-l

antrectomized

pg/g wet weight 200

1

***

l-4-l n

intact

a

antrectomized

and

The effects on the ECL cell density are illustrated in Figures 3C, 4, and 5. In intact controls the number of ECL cells per visual field amounted to -190. Low-dose omeprazole (group 2) did not significantly affect the ECL cell density, whereas treatment with high-dose omeprazole (group 3) and ranitidine (group 4) increased the ECL cell density to 560 and 340, respectively. Antrectomy (group 5) lowered the ECL cell density. Antrectomized animals treated with omeprazole (group 6) showed roughly the same ECL cell density as intact controls (Figures 3C and 5) and thus had significantly higher ECL cell density than antrectomized controls. The density of ECL cells in the oxyntic mucosa was closely correlated to the plasma gastrin concentration. This is shown in Figure 6, where individual values from all six groups, as well as group mean values, have been plotted. The immunofluorescence intensity of the ECL cells was high in the high-dose omeprazole group and in the ranitidine group and low in antrectomized controls (Figures 4 and 5). The density of somatostatin cells (number per visual field) did not differ significantly among control rats (group 1, 45 2 5.0) and rats treated with high-dose omeprazole (group 3, 45 2 5.0), rats treated with ranitidine (group 4, 45 + 4.0), antrectomized controls (group 5, 39 ? 3.O), and omeprazoletreated antrectomized rats (group 6, 45 2 2.0). After a lo-wk recovery period the ECL cell density had increased from -190 to almost 300 (p < 0.05) in the control group (Figure 7, Table 3), whereas a

Number/visual 6001

Figure

3

field *** 1

Oxyntic mucosal histidine decarboxylase (HDC) activity (A). histamine concentration (B), and ECL cell density (C) after 10 wk of treatment with omeprazole, ranitidine, or vehicle. The cells were visualized by immunofluorescence using antibodies against HDC. Mean + SEM; n = 9-11. *+p C 0.01 and ***p < 0.001 according to Mann-Whitney U-test.

396

Table

GASTROENTEROLOGY

LARSSON ET AL.

3. Oxyntic

Mucosal

Enterochromafinlike

Cell Density,

Histamine

Concentration,

and Histidine

Vol. 90, No. 2

Decarboxylase

Activitv’ 10 wk of treatment + 10 wk of recovery

10 wk of treatment

Parameter ECL cell density (cells/visual field) Histamine concentration (b&g wet W HDC activity (pmol CO, formedimg wet wt * h)

Control

Omeprazole

Control

Omeprazole

189 +- 25.6

564 + 50.Zb

286 2 22.5

410 2 42.1"

57.0 + 4.9

183 4 12.5"

78.8 2 7.0

117 t 4.1d

39.8 t 4.3

394 2 41.5b

51.6 f 11.4

36.5 +- 6.5"

Values are mean 2 SEM (n = 10).a In rats treated for 10 wk with vehicle or omeprazole (400 pmol/kg body wt) and in rats treated in the same way but followed by a lo-wk recovery period. b p < 0.001; (-p < 0.05;d p < 0.01;' NS by Mann-Whitney U-test.

decrease from 560 to almost 400 (p < 0.05) was seen in the group previously treated with omeprazole. The difference between the groups was statistically significant. The results obtained with the HDC immunofluorescence technique were verified in sections stained for argyrophilic cells by the methods of Grimelius (11) or Sevier-Munger (12), or by using an antiserum against histamine (not shown). Generally, the argyrophilic reaction in the ECL cells was weaker after high-dose omeprazole treatment than in control rats.

Discussion The present study demonstrates a direct relationship between the plasma gastrin concentration and the density of ECL cells in the rat oxyntic mucosa. Interestingly, the effect of gastrin was not an unspecific one on all gastric endocrine cells, as the somatostatin cells were not affected by the widely differing plasma gastrin concentrations. Permanently altered levels of plasma gastrin were induced in two ways: (a) increases were obtained by inhibition of acid secretion using either a high daily dose of omeprazole or three daily doses of the

Figure 4. Enterochromaffinlike cells in oxyntic mucosa visualized by immunofluorescence using antibodies against HDC. A. Control rat (group 1).B. Ranitidine treatment (group 4). C. Omeprazole treatment (high dose, group 3).Blockade of acid secretion (B and C) caused a marked proliferation of the ECL cells.

February

Figure

1986

GASTRIN

AND ECL CELL PROLIFERATION

IN THE RAT

397

by immunofluorescence using 5. Enterochromaffinlike cells in the oxyntic mucosa of antrectomized rats, demonstrated antibodies against HDC. A. Antrectomized controls (group 5). B. Omeprazole treatment (high dose, group 6). Note that the number of ECL cells and the immunofluorescence intensity were reduced after antrectomy and that the reduction was reversed by high-dose omeprazole

histamine Hz-receptor antagonist ranitidine (Zantat), or (b) a decrease was obtained by antrectomy. Omeprazole at 400 pmolikg body wt (p.0.) once daily is known to inhibit gastric acid secretion over a period of almost 24 h (13,14), whereas ranitidine has a shorter duration of action. When the dose regimen used for ranitidine in the present study was tested in gastric fistula rats, acid secretion was inhibited for -5 h after each dose and secretory rates above control level were attained before the morning dose (data not shown). These differences in potency and duration of action between the two drugs probably explain why ranitidine induced a comparatively lower increase in plasma gastrin and a lower increase in HDC activity, histamine concentration, and ECL cell density than high-dose omeprazole. Also the difference between omeprazole and ranitidine in general trophic effects, as reflected by stomach weight and oxyntic mucosal thickness, can be explained in similar terms. A higher dose of ranitidine (5 mmolikg body weight daily) given to gastric fistula rats as a continuous subcutaneous infusion during 1 wk was found to produce a sustained 80% inhibition of maximally stimulated acid secretion and an increase in plasma

gastrin to between 1000 and 2000 pgiml (Carlsson et al., unpublished data). A low dose of omeprazole (10 pmol/kg body wt, p.o.) that exerted a submaximal effect on acid secretion (-X1%-70% acute inhibition, not shown) as well as on plasma gastrin levels had no statistically significant effect on the ECL cell density. These results indicate that gastrin is secreted in direct relation to the degree of acid inhibition. A rapid return (within 1 wk) to control plasma gastrin levels occurred upon discontinuation of treatment with omeprazole. One week might seem to be a surprisingly short time, as long-term treatment with high doses of omeprazole induces antral gastrin cell hyperplasia (15,16). However, the return of acid secretion probably evokes a powerful feedback inhibition of gastrin release. After 10 wk of recovery, the ECL cell density remained higher in the omeprazole group than in the control group, although plasma gastrin was already back to normal after 1 wk. Inasmuch as the ECL cell density seemed to rise spontaneously in the control animals during the same period, the return toward control levels in the omeprazole-treated animals might be masked by an age-dependent trophic

398

LARSSON ET AL.

Nuryber/visual

GASTROENTEROLOGY Vol. 90, No. 2

field

7.0 A

Log plasma

Number/visual

gastrin

concentration

field

600

0 Intact.

omeprawle

10 urn&kg pmovkg 175+175+350 pmol/kg

0 Intact,omeprazola 400

‘---r’ , ,

a Intact,

ranitidine

.

?? A’tytomiyed,conl’*

A Antrectomized,

2.0 B

2.2

2.4

2.6

2.8

3.0

omeprazole

3.2

, 400

3.4

pmol/kg

3.6

Log plasma gastrin concentration

Figure 6. Correlation between plasma gastrin levels and ECL density in oxyntic mucosa after 10 wk of treatment with omeprazole, ranitidine, or vehicle. A. Individual data from all groups, r = 0.86; significance level = O.OOO~.B. Group mean values 2 SEM, r = 0.99; significance level = 0.001.

“drive” on the ECL cells. Nonetheless, the normal half-life of the ECL cells might be relatively long, i.e., in the same range as the half-life of, e.g., parietal cells (17). In another study by Stockmann et al. (15), the number of argyrophilic cells (mainly ECL cells) in the rat oxyntic mucosa was back to control values 42 days after stopping treatment with omeprazole (40 pmol/kg body wt twice daily for 8 wk). The reason for the discrepancy might be related to the difference in dose, the period of treatment, and the fact that male rats were used in the study by Stockmann et al. (15). There was a close correlation between oxyntic mucosal HDC activity and plasma gastrin, as well as between histamine concentration and plasma gastrin. There was, however, a difference between the two parameters in that a more direct relation to the ECL cell density was found for the histamine level, Regardless of the treatment to which the rats had been subjected, the histamine content calculated per ECL cell seemed to be relatively constant, whereas

the activity of HDC per ECL cell increased with increasing plasma gastrin levels (not shown). This latter finding was reflected in the fact that the HDC immunofluorescence intensity was conspicuously high in omeprazole-treated animals (high plasma gastrin) and low in antrectomized animals (low plasma gastrin). As expected, the plasma gastrin level and all gastrin-dependent parameters (oxyntic mucosal thickness, HDC activity, histamine concentration, and ECL cell density) were lowered by antrectomy. However, increases in all parameters to the levels seen in intact control rats were induced by inhibition of gastric acid secretion by omeprazole. Again, the changes can be explained by the plasma gastrin concentration, as there was a close correlation of ECL cell density to plasma gastrin within that group too. The increased plasma gastrin levels and ECL cell density found in antrectomized animals upon omeprazole treatment might be explained by gastrin secretion from an extra-antral source or from gastrin cells in an unresected “borderline zone” between the corpus and antrum. Indeed, histologic examination revealed minute remnants of antral mucosa in the “borderline zone” in some of the antrectomized rats. It can be concluded that pronounced inhibition of acid secretion causes hypergastrinemia, which-if sustained-causes proliferation of ECL cells. The gastric ECL cell density reflects the circulating gastrin concentration, and in this study no evidence has been found of an additional, direct effect of omeprazole or ranitidine. The close correlation between the logarithm of the plasma gastrin level and the ECL cell density in the antrectomized groups, in the intact control group, and in the groups treated Number/visual

field

10 weekd’ treatment

Figure

10 weeks’ treatment 10 weeks’ recovery

7. Oxyntic mucosal ECL cell density in rats treated for 10 wk with omeprazole (400 Fmolikg body wt) and in rats 10 wk after stopping treatment with the same dose and for the same period of time. Mean values it SEM, n = 10. *For p < 0.05, ***for p < 0.001 according to the Mann-Whitney U-test.

February 1986

with omeprazole and ranitidine indicates that neither of the compounds per se is responsible for the proliferation of ECL cells. The results of the present study favor the view that the diffuse and focal ECL cell hyperplasia and focal neoplasia (ECL cell carcinoids) observed in the rat stomach after 2 yr of treatment with omeprazole are the result of sustained pronounced hypergastrinemia induced in turn by long-lasting inhibition of acid secretion.

GASTRIN AND ECL CELL PROLIFERATION

7.

8.

9.

10.

References L, Hansson E, Havu N, Carlsson E, Lundberg C. Toxicological studies on omeprazole. In: Borg KO, Olbe L, eds. Omeprazole book one. Stand J Gastroenterol 1985; 2O(Supp1 108):53-69. 2. Hdkanson R, Ekelund M, Sundler F. Activation and proliferation of gastric endocrine cells. In: Falkmer S, Hlkanson R, Sundler F, eds. Evolution and tumour pathology of the neuroendocrine system. New York: Elsevier, 1984:371-98. 3. Hlkanson R, Liedberg G. The role of endogenous gastrin in the activation of gastric histidine decarboxylase in the rat. Effect of antrectomy and vagal denervation. Eur J Pharmacol

11.

1970;12:94-103. 4. bundell L. Histamine

15.

1. Ekman

metabolism of the gastric mucosa following antrectomy. J Physiol 1974;241:437-51. 5. Watanabe T, Taguchi Y, Hayashi H, et al. Evidence for the presence of a histaminergic neuron system in the rat brain: an immunohistochemical analysis. Neurosci Lett 1983;39: 249-54. 6. Kubota H, Taguchi Y, Tohyama M, et al. Electron microscopic identification of histidine decarboxylase-containing endo-

12. 13.

14.

16.

IN THE RAT

399

crine cells of the rat gastric mucosa: an immunohistochemical analysis. Gastroenterology 1984:87:496-502. Ekblad E, Ekelund M, Graffner H, HBkanson R, Sundler F. Peptide-containing nerve fibers in the stomach wail of rat and mouse. Gastroenterology 1985;89:73-85. Coons AH, Leduc EH, Conolly JM. Studies on antibody production. I. A method for the histochemical demonstration of specific antibody and its application to a study of the hyperimmune rabbit. J Exp Med 1955;102:49-60. Snyder SH. Determination of tissue histamine by an enzymatic isotopic assay. In: Colowich SP, Kaplan NO, eds. Methods of enzymology. Vol. XVII B. New York: Academic, 1971:838-41. Beaven MA, Wilcox B, Terpstra GK. A microprocedure for the measurement of 14C0, release from “C carboxyl-labelled amino-acids. Anal Biochem 1978;84:638-41. Grimelius L. A silver nitrate stain for u cells in human pancreatic islets. Acta Sot Med Ups 1968;73:243-76. Sevier AC, Munger BL. A silver method for paraffin sections of neural tissue. J Neuropathol Exp Neurol 1965;24:130-5. Larsson H, Wallmark B. Inhibition of rat gastric H’,K’ATPase in vivo. Relation to reduction of acid secretion. Proceedings of the First International Symposium on Omeprazole. Stand J Gastroenterol (in press). Wallmark B, Larsson H, Humble L. Relation between gastric acid secretion and gastric H+,K+-ATPase activity. J Biol Chem (in press). Stijckmann F, FBlsch UR, Bonatz G, Wiilfrath M, Creutzfeld W. Der Einfluss von Omeprazol auf die endokrinen Zellen in Fundus und Antrum des Rattenmagens. Z Gastroenterol 1984; 22:449. Allen J, Bishop A, Daly M. et al. Studies on the effect of inhibition of acid secretion on the regulatory peptides of the rat stomach. Gastroenterology (in press).

17. Helander HF. The cells of the gastric mucosa. 1981;70:217-89.

Int Rev Cytol