K. KANATANI*, M. EBATA*, M. MURAKAMI**, S. OKABE*

EFFECTS OF INDOMETHACIN AND ROFECOXIB ON GASTRICMUCOSAL DAMAGE IN NORMAL AND HELICOBACTER PYLORI-

INFECTED MONGOLIAN GERBILS

*Department of Applied Pharmacology, Kyoto Pharmaceutical University, Misasagi, Yamashina, Kyoto 607-8414, Japan

**Department of Medical Welfare, Nippon Bunri University, Ooita 870-0397, Japan

This study examined the effects of indomethacin and rofecoxib on normal andHelicobacter pylori (H. pylori)-infected gastric mucosa of Mongolian (M.) gerbils.M. gerbils (6-wk-old) were orally administered H. pylori (ATCC43504, 2×108

CFU/ml) after fasting for 24 hours. Beginning 3 mo after inoculation, indomethacin(2 mg/kg, s.c) or rofecoxib (10 mg/kg, p.o.) was administered once daily for 2 wk tothe gerbils. At autopsy, gastric mucosal ulcer area, myeroperoxidase (MPO) activity,prostaglandin (PG) E2 synthesis, and H. pylori viability were determined. Histamine-stimulated gastric acid secretion was measured with the acute gastric fistula method.Histological study was performed with H&E staining. H. pylori infection causedsevere mucosal damage and production of lymphoid follicles in the gastricsubmucosa. In H. pylori-infected gerbils, indomethacin aggravated the gastricmucosal damage induced by H. pylori infection. Furthermore, indomethacin by itselfinduced gastric ulcers at an incidence of 6/10. In contrast, rofecoxib did notaggravate the H. pylori-induced mucosal damage. Indomethacin and rofeocoxibsignificantly reduced H. pylori viability. MPO activity was significantly increased inH. pylori-infected gerbils compared with H. pylori-uninfected gerbils. Indomethacinand rofecoxib reduced MPO activity in H. pylori-infected gerbils. PGE2 synthesiswas markedly increased in H. pylori-infected gerbils (approximately 3-times)compared with the normal gerbils. Indomethacin significantly inhibited PGE2

synthesis in the gastric mucosa, both in normal and H. pylori-infected gerbils.Rofecoxib did not reduce PGE2 synthesis in normal gerbils, however, PGE2 synthesiswas reduced to normal levels in H. pylori-infected gerbils. In H. pylori-infectedgerbils, histamine-stimulated gastric acid secretion was reduced compared withnormal gerbils. Indomethacin significantly increased histamine-stimulated gastricacid secretion and rofecoxib tended to increase secretion in H. pylori-infectedgerbils. It was concluded that indomethacin enhances development of gastricmucosal damage in normal gerbils and aggravates H. pylori-induced gastric damage,resulting in gastric ulcers. Rofecoxib did not induce gastric damage in normal gerbils

JOURNAL OF PHYSIOLOGY AND PHARMACOLOGY 2004, 55, 1, 207�222

www.jpp.krakow.pl

and did not aggravate damage in H. pylori-infected gerbils, suggesting that rofecoxibis less damaging to the stomach than indomethacin.

K e y w o r d s : H. pylori, M. gerbil, COX-1, COX-2.

INTRODUCTION

Our group has already reported that a single inoculation of Helicobacter pylori(H. pylori) could induce atrophic gastritis, as well as ulcer and cancer-likemucosal changes in the stomachs of Mongolian gerbils (M. gerbils) (1, 2). Inaddition, we have reported that various non-steroidal anti-inflammatory drugs(NSAIDs) invariably induce acute gastric lesions and delay ulcer healing inanimals. It had been generally believed that the mechanism of action for NSAIDsis inhibition of both cyclooxygenase-1 (COX-1) and COX-2 isoenzymes in thearachdonic acid cascade. Both COX enzymes produce prostaglandin (PG), butCOX-1 is considered to be responsible for maintenance of gastric mucosalintegrity. In contrast, COX-2 is induced by such processes as inflammation. It iswell known that PGs regulate gastric acid secretion, gastric mucus andbicarbonate secretion, gastric mucosal blood flow, and gastric motility (3 - 6).Accordingly, NSAIDs cause gastric mucosal damage by inhibiting both COX-1and COX-2, resulting in reduced PG synthesis. Consequently, NSAID-inducedgastric mucosal damages are invariably inhibited by PG supplementation (7). Inorder to reduce such gastric damage, COX-2 selective inhibitors, such ascelecoxib and rofecoxib were developed and clinically used (8, 9). Indeed, theincidence of gastric ulcers in patients taking COX-2-selective NSAIDs issignificantly reduced compared with the incidence for those taking non-selectiveNSAIDs (10, 11). Such data indicates that PG production induced by COX-2plays an important role in gastric mucosal repair. Consequently, H. pyloriinfection and NSAIDs are considered to represent the most important risk factorsfor gastric mucosal damage.

Data concerning the effect of NSAIDs on H. pylori-infected patients remaincontroversial, i.e., NSAIDs were found to aggravate gastric mucosal damageinduced by H. pylori infection (12), as well as exert no effect (13). In addition, itwas reported that H. pylori eradication achieved prior to administration ofNSAIDs could reduce the incidence of gastric ulcers (14). Another studyconcluded that NSAIDs had no effect on the gastric ulcer recurrence and delay ofgastric ulcer healing (15). Accordingly, the interplay between H. pylori infection,NSAIDs, and the timing of H. pylori eradication remains uncertain.

The present study examined the effects of rofecoxib and indomethacin on thegastric mucosa of normal and H. pylori-infected M. gerbils.

208

MATERIALS AND METHODS

Animals

Male Mongolian gerbils (6-wk-old, 40-50 g) were purchased from Seac Yoshitomi (Fukuoka,Japan). The gerbils were kept in an isolated clean room with regulated temperature (approximately20-22°C), humidity (approximately 55%), and light/dark cycle (12/12 h). The gerbils were deprivedof food for 24 h before and 4h after H. pylori inoculation, but were otherwise afforded free accessto food and tap water. A total of 5 to 10 animals were used for each study. The maintenance of theanimals and the experimental procedures were carried out in accordance with the guidelines of theEthics Committee of Kyoto Pharmaceutical University.

H. pylori preparation and inoculation

Cag A- and Vac A-positive standard strains of H. pylori (ATCC43504; American Type CultureCollection, Rockville, MD) were used for this study. The bacteria were incubated overnight in abrain-heart infusion broth (Difco Laboratories, Detroit, MI) containing 10% fetal bovine serum(Gibco BRL, Grand Island, NY) at 37°C under a microaerophilic atmosphere; the bacteria wereallowed to grow to a concentration of approximately 2.0×108 colony-forming units (CFU)/mL. H.pylori (2.0×108 CFU/mL) were orally inoculated into each animal at a dose of 1.0 mL/animal.

Drugs

In H. pylori-infected animals, drug treatment was started 3 mo after inoculation. Indomethacin(Sigma), suspended in a trace of Tween 80 and saline, was subcutaneously administered at a doseof 2 mg/kg once daily for 2 wk to normal and H. pylori-infected gerbils beginning 3 mo post-infection. Rofecoxib (Nippon Chemiphar Co., Tokyo) was suspended in 0.5%hydoxypropylcellulose and orally administered at a dose of 10 mg/kg daily for 2 wk to normal andH. pylori-infected gerbils beginning 3 mo post-infection. Control animals received vehicle alone.To measure histamine-stimulated gastric acid secretion, histamine (Sigma) was suspended in salineand subcutaneously administered at a dose of 10 mg/kg to normal and H. pylori-infected gerbilsbeginning 3 mo post-infection. Indomethacin and rofecoxib were also subcutaneously administeredto normal and H. pylori-infected gerbils 1 h before stimulation with histamine.

Macroscopical Studies

The animals were killed with an overdose of ether. The stomach of each animal was removed,opened along the greater curvature, and then spread upon a corkboard. The lesion area (mm2) waspromptly measured under a dissecting microscope (×10; Olympus, Tokyo, Japan). The author (S.O.)who determined the ulcer size was blinded with regards to the treatment that any given animalreceived.

Quantification of viable H. pylori

After sacrificing the animals with ether anesthesia, the stomach of each animal washomogenized in 10 ml of phosphate-buffered saline with a Polytron (Kinetica, Steinhofhalde,Sweden) and then diluted with the same buffer. Aliquots (0.1ml) of the dilutions were applied toBrucella agar plates (Gibco BRL) containing 10% horse blood (Nippon Bio-Test Laboratories,Tokyo), 2.5 µg/ml amphotericin B (Sigma, St. Louis, MO), 9 µg/ml vancomycin (Sigma), 0.32

209

µg/ml polymyxin B (Sigma), 5 µg/ml trimethoprim (Sigma), and 50 µg/ml 2, 3, 5-triphenyltetrazolium chloride (Wako Pure Chemicals, Osaka). The plates were incubated at 37°Cunder a microaerophilic atmosphere for 7 days. All colonies were either black or gold and wereidentified as H. pylori by appreciating the characteristic spiral shape under a microscope (×2,000;Olympus). The number of colonies was determined and the viable H. pylori count was expressed asthe ratio CFU/stomach.

Measurement of MPO activity

We previously reported that MPO activity was negligible in normal gerbils. To elucidatechanges in MPO activity induced by the test drugs, activity was measured by the method of Krawiszet al. (16). Gastric tissue (approximately 50 mg) was extracted from each stomach, homogenizedwith a Polytron in 1.0 mL of 50 mM phosphate buffer (pH 6.0) containing 0.5%hexadecyltrimethyl-ammonium bromide (Sigma), and then subjected to freeze-thawing three times.The homogenates were centrifuged at 1,600 g for 10 min at 4 °C. After an aliquot (5 µL) of eachsupernatant had been mixed with 145 µL of phosphate buffer (pH 6.0), containing 0.167 mg/mL o-dianisidine dihydrochloride (Sigma) and 0.0005% H2O2, the change in the rate of absorbance at 450nm was measured with a microplate reader (Thermo Max; Molecular Devices, Sunnyvale, CA).MPO activity was expressed as the degradation of H2O2 µmol/min/g tissue.

Determination of PGE2 synthesis

PGE2 production in the gastric mucosa of the gerbils was determined according to the methodof Lee and Feldman (17, 18). Gastric specimens were extracted from the fundic-antral area. Eachgastric specimen was placed in 50 mM Tris-HCl (pH 8.4) buffer and then minced with scissors.After the tissue samples had been washed and re-suspended in 1 ml of buffer, each sample wassubjected to vortex mixing at room temperature for 1 min to stimulate PGE2 production, followedby centrifugation at 10,000 g for 15 sec. The PGE2 levels in the resulting supernatants weredetermined by means of an enzyme immunoassay (PGE2 EIA kit; Cayman Chemicals, Ann Arbor,MI). PGE2 production was expressed as pg of PGE2/min/mg tissue.

Measurement of histamine-stimulated gastric acid secretion

The effect of indomethacin and rofecoxib on histamine-stimulated gastric acid secretion wasdetermined in normal and H. pylori-infected gerbils with the acute gastric fistula method. Thegerbils were fasted for 23 h, then anesthetized with an intraperitoneal injection of urethane at a doseof 1.25 g/kg. After tracheotomy, a polyethylene tube was inserted into the trachea to ensure a patentairway, the abdomen was incised, and the stomach and duodenum were exposed. An acute fisutlawas achieved by insertion of a polyethylene tube towards the direction of the antrum through anincision in the duodenum. A total of 1 milliliter of saline was injected and collected every 15minutes; acid output (µmol H+/15 min) was determined by titration with 10 mmol/L NaOH.Animals were subcutaneously injected with 2 mg/kg of indomethacin or 10 mg/kg of rofecoxib 1 hbefore stimulation with histamine (10 mg/kg subcutaneous dose).

Histological observation

After gross observation, several specimens were extracted from each stomach, fixed withbuffered 10% formalin, and embedded in paraffin. Subsequently, 4 µm paraffin samples wereprepared and stained with hematoxylin and eosin.

210

Statistical Analysis

Data are presented as means ± SEM. Statistical differences were evaluated using the Student'st-test or Dunnett's multiple comparison test, with a P value of < 0.05 regarded as significant.

RESULTS

Effects of indomethacin and rofecoxib on the gastric mucosa of normal and H.pylori-infected M. gerbils

In untreated normal gerbils, no gastric damage was observed. Indomethacininduced a few erosions in the region of the corpus, while rofecoxib caused nogastric erosion in the gastric mucosa. The erosion area in indomethacin-treatedanimals was 1.8 ± 0.3 mm2. Three mo after H. pylori infection, some superficialdamage and erosions were observed, but no ulcers were noted. The areas forsuperficial damage and erosions were 33.3± 24.4 mm2 and 2.8± 0.6 mm2,respectively. Indomethacin aggravated gastric damage induced by H. pylori andresulted in gastric ulcers in the fundic-antral area at an incidence of 6/10. Theareas for superficial damage, erosions, and ulcers were 12.2± 2.6 mm2, 6.1± 2.0mm2, and 7.1± 3.4 mm2, respectively. Moreover, in indomethacin-treated infectedgerbils, largest ulcer area was 31 mm2. Rofecoxib did not affect H. pylori-inducedgastric mucosal damages and did not precipitate development of gastric ulcers.The areas for superficial damage and erosions were similar to the H. pylori-infected control group, i.e., 33.6± 14.2 mm2 and 2.2± 1.2 mm2 (Figs 1-2).

211

Fig.1. Effects of indomethacinand rofecoxib on the gastricmucosa of normal and H.pylori-infected M. gerbils.Indomethacin (2 mg/kg, s.c.)and rofecoxib (10 mg/kg, p.o.)were administered once dailyfor 2 wk to M. gerbils with andwithout H. pylori infection.Drug treatment was started 3mo after inoculation. Note thatindomethacin clearly inducedgastric ulcers in H. pylori-infected animals, whilerofecoxib did not. Data arepresented as means±SEM for5-10 animals.

Effects of indomethacin and rofecoxib on histological changes in normal and H.pylori-infected M. gerbils

In normal gerbils, no gastric damage was observed, hence the mucosalstructure was normal. In the indomethacin-treated animals, the mucosa epithelialcells were observed to peel away from the underlying cells. In contrast,rofecoxib-treated animals exhibited normal mucosa. In H. pylori-infectedmucosa, superficial damage and erosions were observed to a small extent.Moreover, lymphoid follicles, neutrophil infiltration, gastric hyperplasia were

212

Fig. 2. Gross appearances of the gastric mucosa in normal (A) and H. pylori-infected M. gerbils (B),with and without indomethacin (2 mg/kg, s.c.) or robecoxib (10 mg/kg, p.o.). Note that gastriculcers were observed only after indomethacin treatment was administered to H. pylori-infectedgerbils.

A: Normal

B: H. pylori infection

also observed. Indomethacin treatment aggravated such mucosal damageobserved in H. pylori-infected mucosa and H. pylori-induced gastric ulcers,which were found to penetrate the muscularis and serosa and adhere to the liver.In rofecoxib-treated animals, mucosa superficial damage and erosions wereobserved to a small extent, however, gastric ulcers were not induced as in theindomethacin-treated gerbil group (Fig. 3).

Effects of indomethacin and rofecoxib on H. pylori viability in H. pylori-infected M. gerbils

H. pylori was detected in all the gerbil stomachs that were examinedimmediately after treatment with vehicle, indomethacin, or rofecoxib. H. pyloriviability was 3.1x105 CFU/stomach in indomethacin control animals.Indomethacin significantly decreased H. pylori viability to 1.8x105 CFU/stomach.In rofecoxib control animals, H. pylori viability was 3.6x105 CFU/stomach,

213

Fig. 3. Histological appearances of the gastric mucosa of normal (A) and H. pylori-infected M.gerbils (B) with or without indomethacin (2 mg/kg, s.c.) or rofecoxib (10 mg/kg, p.o.). (a) Control,(b) Indomethacin (2 mg/kg, s.c.), (c) Rofecoxib (10 mg/kg, p.o.). (H&E staining; x50).

A: Normal B: H. pylori infection

which was also significantly reduced to 2.0x105 CFU/stomach with rofecoxibtreatment (Fig. 4).

Effects of indomethacin and rofecoxib on the MPO activity of normal and H.pylori-infected M. gerbils

In the indomethacin and rofecoxib normal control groups, MPO activity in thegastric mucosa was negligible (7.2±3.0 µmol H2O2/min/g tissue and 6.4±1.8µmolH2O2/min/g tissue, respectively). In indomethacin-treated gerbils, MPO activity inthe gastric mucosa was significantly increased, however, rofecoxib did not affectMPO activity (26.7±6.3 µmol H2O2/min/g tissue and 10.0±3.4 µmol H2O2/min/gtissue). MPO activity in H. pylori-infected gastric mucosa was markedlyincreased compared with normal gerbils. In indomethacin-treated H. pylori-infected gerbils, MPO activity in the gastric mucosa was significantly higher thanthat of normal gerbils, however, the level was reduced compared with the H.pylori-infected control group (907.2±218.5 µmol H2O2/min/g tissue vs.1690.7±128.1 µmol H2O2/min/g tissue). In the rofecoxib-treated H. pylori-infected gerbils, MPO activity in the gastric mucosa was also significantly higherthan that of normal gerbils, but less than the H. pylori-infected control group(1255.1±236.0 µmol H2O2/min/g tissue vs. 1906.4±47.2 µmol H2O2/min/g tissue)(Fig. 5).

214

Fig. 4. Effects of indomethacinand rofecoxib on H. pyloriviability in H. pylori-infectedM. gerbils. Indomethacin (2mg/kg, s.c.) or rofecoxib (10mg/kg, p.o.) was administeredonce daily for 2 wk to M.gerbils 3 mo after inoculation.Data are presented as means ±SEM for 4-5 animals. *, #Significantly different from thecontrol (H. pylori infected)group, P<0.05.

Effects of indomethacin and rofecoxib on PGE2 synthesis in normal and H.pylori-infected M. gerbils

PGE2 synthesis in the indomethacin-treated gastric mucosa was significantlyreduced compared with normal gerbils (8.0 ± 2.4 pg/min/mg tissue vs. 49.8 ± 25.6pg/min/mg tissue). PGE2 synthesis in the rofecoxib-treated mucosa was similar tothe normal control mucosa (37.8 ± 7.0 pg/min/mg tissue vs. 42.0 ± 14.5pg/min/mg tissue). In H. pylori-infected gastric mucosa, PGE2 synthesis wasmarkedly increased to approximately 3 times that of the control gastric mucosa.Indomethacin significantly inhibited PGE2 synthesis less than the normal controlgroup. In rofecoxib-treated gastric mucosa, PGE2 synthesis was significantlyinhibited, resulting in a value similar to the control gastric mucosal PGE2

synthesis (Fig. 6).

Effects of indomethacin and rofecoxib on histamine-stimulated gastric acidsecretion in normal and H. pylori-infected M. gerbils

In normal gerbils, histamine was found to significantly increase gastric acidsecretion. The maximum level of gastric acid secretion was 75 min afterstimulation with histamine. After indomethacin treatment 1 h before histaminestimulation, the maximum gastric acid secretion tended to increase comparedwith that of vehicle group (42.5±31.0 µEq/15 min vs. 17.6±4.9 µEq/15 min) 90min after stimulation with histamine. In rofecoxib-pretreated group, the

215

Fig. 5. Effects of indomethacinand rofecoxib onmyeloperoxidase (MPO)activity in the gastric mucosaof normal and H. pylori-infected M. gerbils.Indomethacin (2 mg/kg, s.c.) orrofecoxib (10 mg/kg, p.o.) wasadministered once daily for 2wk to M. gerbils with orwithout H. pylori infection.Drug treatment was started 3mo after inoculation. Data arepresented as means ± SEM for4-5 animals. *, # Significantlydifferent from the normal andH. pylori-infected groupsrespectively, P<0.05.

maximum gastric acid secretion was the same as that of the vehicle group(17.8±3.6 µEq/15 min vs. 17.6±4.9 µEq/15 min) 75 min after stimulation withhistamine. In H. pylori-infected gerbils, the maximum gastric acid secretion wasreduced (5.5±3.0 µEq/15 min vs. 17.6±4.9 µEq/15 min) 60 min after stimulationwith histamine. Moreover, total gastric acid secretion 2 h after stimulation withhistamine was also significantly reduced compared with that of normal gerbils(31.3±0.6 µEq/2 hr vs. 97.6±2.1 µEq/2 hr). In contrast, in the indomethacinpretreatment group, maximum gastric acid secretion was enhanced 75 min afterstimulation with histamine (11.1±4.1 µEq/15 min vs. 5.5±3.0 µEq/15 min).Furthermore, total gastric acid secretion 2 h after stimulation with histamine wassignificantly increased compared with that of H. pylori-infected gerbils (62.8±1.3µEq/2 hr vs. 31.3±0.6 µEq/2 hr). In the rofecoxib pretreatment group, themaximum gastric acid secretion was enhanced 75 min after stimulation withhistamine (7.4±2.8 µEq/15 min vs. 5.5±3.0 µEq/15 min) (Fig.7).

DISCUSSION

This study demonstrated that indomethacin aggravated gastric mucosaldamage induced by H. pylori, resulting in gastric ulcers. Moreover, the study alsoconfirmed that rofecoxib, a COX-2 selective inhibitor, had an effect on not onlynormal mucosa, but also H. pylori-induced mucosal damage. In general, themechanisms of NSAID-induced gastric damage and H. pylori-induced mucosal

216

Fig. 6. Effects of indomethacinand rofecoxib on PGE2

synthesis in gastric mucosa ofnormal and H. pylori-infectedM. gerbils. Indomethacin (2mg/kg, s.c.) or rofecoxib (10mg/kg, p.o.) was administeredonce daily for 2 wk to M.gerbils with or without H.pylori infection. Drugtreatment began 3 mo afterinoculation. Note that bothindomethacin and rofecoxibsignificantly reduced PGE2synthesis, which was markedlyincreased with H. pyloriinfection. Data are presented asmeans ± SEM for 4-5 animals.*, # Significantly differentfrom the normal and H. pylori-infected groups respectively,P<0.05.

damage are thought to be independent, however, factors such as neutrophilactivation and gastric acid secretion are thought to be involved in thepathogenesis of gastric damage.

Gastric acid is recognized as a caustic agent for the stomach. In the presentstudy, although gastric acid secretion was increased by stimulation with histaminein H. pylori-infected gerbils, the increase was low compared with normal gerbils.Nonetheless, pretreatment with indomethacin increased histamine-stimulatedgastric acid secretion more than the vehicle in H. pylori-infected M. gerbils. It isknown that gastric acid secretion is decreased by H. pylori infection and that sucha phenomenon might be caused by a component of H. pylori, inflammatorycytokines induced by H. pylori infection, or a decrease in the number of parietalcells caused by atrophic gastritis (19 - 21). It has been reported that expression ofseveral inflammatory cytokines, such as Interleukin (IL)-1ß and tumor necrosisfactor (TNF)-α resulting from gastric ulcers, inhibits gastric acid secretion (22)and elevates expression of COX-2 mRNA in gastric mucosa (23, 24). It has alsobeen reported that PG inhibits gastric acid secretion by binding to EP3 receptors(25). In our study, PGE2 synthesis in H. pylori-infected gerbils was increased 3

217

Fig. 7. Effects of indomethacinand rofecoxib on histamine-stimulated gastric acidsecretion in normal and H.pylori-infected M. gerbils.Indomethacin (2 mg/kg, s.c.) orrofecoxib (10 mg/kg, p.o.) wasadministered 60 min beforehistamine (10 mg/kg, s.c.)treatment. It was of note that H.pylori infection significantlyreduced histamine-stimulatedgastric acid secretion.Indomethacin, but notrofecoxib, significantlysuppressed the reduced acidsecretion seen in infectedanimals. Data are presented asmeans ± SEM for 6-7 animals.*, # Significantly differentfrom the normal and H. pylori-infected groups respectively,P<0.05.

times greater than that of normal gerbils. In H. pylori-infected gerbils, treatmentwith indomethacin significantly decreased PGE2 synthesis in the gastric mucosa;the level was lower than that of normal gerbils. Such findings indicate that PGinduced by COX-2 in the inflammatory process is associated with a decrease ingastric acid secretion seen in H. pylori infection. Indomethacin increased gastricacid secretion as a result of inhibition of PG synthesis. The present studydemonstrated that in normal gerbils, rofecoxib decreased PGE2 synthesis to thesame level as that of normal gerbils; pretreatment with rofecoxib tended toincrease gastric acid. Xiao (26) has also reported that etodolac, a COX-2 selectiveinhibitor, increased gastric acid secretion in H. pylori-infected mice. Accordingly,PG production induced by COX-2 is involved in the inhibition of gastric acidsecretion observed in H. pylori infection.

In H. pylori infection, IL-1, as well as TNF-α produced by monocytes andmacrophages, induced IL-8 production (27). H. pylori stimulate IL-8 productionin gastric epithelial cells (28). It has also been reported that IL-8-enhancedneutrotaxis, as well as expression of CD11b/CD18 and free radicals, resulted incontinuous infiltration of inflammatory cells, in addition to cell injury (29). Onthe other hand, NSAIDs decrease PG synthesis through COX inhibition. PGI2

produced in endothelial cells and PGE2 produced in gastric mucosal andendothelial cells inhibit neutrophil activation, production of cytokines, andexpression of adhesion molecules (30). Moreover, PGE2 also inhibits neutrophilsfrom producing elastase and free radicals (31). The decrease of PG synthesisresulting from NSAIDs caused neutrophil activation, leading to gastric mucosalinjury. In this study, gastric mucosal MPO activity was very low level in normalgerbils treated with vehicle or rofecoxib. In contrast, MPO activity was increasedin normal gerbils treated with indomethacin. H. pylori-infected gerbils hadsignificantly higher MPO activity and exhibited severe neutrophil infiltration. Aswas noted in a previous study, consumption of NSAIDs during H. pylori infectionresulted in more severe injury due to combination of neutrophil activationinduced by NSAIDs with neutrophil infiltration induced by H. pylori infection.Nonetheless, indomethacin and rofecoxib decreased MPO activity in gerbilgastric mucosa. Two possible mechanisms are involved. One possibility is thatindomethacin induces lower IL-8 levels than H. pylori infection alone. Moreover,activatived neutrophils caused endothelial cell injury in only blood vessels, not inthe gastric mucosa. Another possibility is that rofecoxib inhibits only COX-2,therefore, PGE2 and PGI2 produced by COX-1 inhibits neutrophil activation,resulting in decreased gastric mucosal MPO activity.

Gastric mucous is one gastric mucosal protective factor. The mechanism ofaction involves binding to EP4 receptors (3). As seen in macroscopic views ofstomachs, gastric mucous is significantly increased by H. pylori infection.Nonetheless, indomethacin and rofecoxib decreased the volume of gastricmucous in H. pylori infection. Moreover indomethacin and rofecoxibsignificantly decreased H. pylori viability. Accordingly, the decrease in H. pylori

218

viability might be caused by the decrease in gastric mucous, an element essentialfor H. pylori survival, as a result of decreasing PG synthesis.

The other gastric protective actions of PG include PGE2 binding to EP1,resulting in inhibition of gastric motility, and to EP2/EP3/EP4, resulting in anincrease of gastric blood flow (5, 30). It has been reported that the gastric mucosalinjury induced by non-selective inhibitors, such as indomethacin, results inacceleration of gastric motility and a decrease in gastric mucosal blood flow (32).On the other hand, it is known that rofecoxib has no effect on gastric motility (33)and celecoxib, a COX-2 selective inhibitor, does not decrease gastric mucosalblood flow (34). Consequently, COX selective inhibition allows reduction of theincidence of gastric mucosal injury.

In a clinical study, H. pylori and non-selective NSAIDs were found to resultin severe gastric damage compared with COX-2 selective NSAIDs, due toinhibition of both COX-1 and COX-2. Nonetheless, it is not clear that NSAIDsaggravate gastric mucosal damage induced by H. pylori. If such a possibilityholds true, then the timing of NSAID ingestion and H. pylori infection, as well asbasal gastric acid secretion and the degree of gastric mucosal injury, becomeimportant. In particular, the timing of H. pylori infection is thought to be veryimportant for induction of a gastric mucosal injury. We previously performedsuch a study 4 wk after induction of H. pylori infection, determining erosion area,MPO activity, PGE2 synthesis, and H. pylori viability in the gastric mucosa ofgerbils. The results were as follows. At baseline, 4 wk after H. pylori infectioninduction, some mucosal erosions were observed. Indomethacin aggravated thegastric mucosal erosion induced by H. pylori infection, resulting in gastric ulcerproduction (1/6) 4 wk after inoculation (35). In this study, treatment withindomethacin beginning 3 mo after H. pylori infection inoculation resulted ingastric ulcer production at an incidence of 6/10. Accordingly, the longer theperiod of H. pylori infection, the more severe the gastric mucosal injury is due topersistent inflammatory reaction.

It was expected that COX-2 selective inhibitors would exhibit a strong anti-inflammatory effect with less gastric mucosal injury than non-selective NSAIDs.Nonetheless, COX-2 selective NSAIDs still possess problems. One studyreported that COX-2 selective NSAIDs induced gastric mucosal injury withoutinhibition of PG in normal gastric mucosa (36). Another study reported thatCOX-2 selective NSAIDs delay the healing of gastric ulcers. Such reportsindicate that COX-2 might be needed to maintain normal conditions in the gastricmucosa. Consequently, prescribing COX-2 selective NSAIDs to H. pyloripositive patients warrants prudence.

The present study found that indomethacin aggravated gastric mucosal injuryinduced by H. pylori infection through inhibition of PG synthesis, resulting in adecrease of gastric mucosal protective factors and an increase in gastric acidsecretion. Moreover, rofecoxib had no effect on gastric mucosa of normal and H.pylori-infected gerbils. Such a result would be expected with preservation of

219

normal PG synthesis levels and inhibition of neutrophil activation. Therefore,these results indicate that non-selective NSAIDs enhance gastric mucosal injuryinduced by H. pylori infection to a greater extent than COX-2 selective NSAIDs.Accordingly, our study further elucidates the effects of H. pylori infection andCOX-2 selective NSAIDs on gastric mucosal injury.

In conclusion, it was found that indomethacin, a non-selective NSAID,aggravated gastric mucosal damage induced by H. pylori infection, leading todevelopment of gastric ulcers. In contrast, rofecoxib, a COX-2 selective inhibitor,had no effect on normal and H. pylori-infected gastric mucosa. This differencemay result from different inhibition levels of PG synthesis that is the differencebetween inhibition of both COX-1 and COX-2, versus only COX-2. Rofecoxibthus appears to be safer to use than indomethacin in regards to the stomach.

REFERENCES

1. Keto Y, Ebata M, Okabe S. Gastric mucosal changes induced by long term infection withHelicobacter pylori in Mongolian gerbils: effects of bacteria eradication. J Physiol (Paris)2001; 95: 429-436.

2. Ebata M, Keto Y, Okabe S. Long-term studies of Helicobacter pylori infection in Mongoliangerbils; effects of eradication timing on prevention of malignant gastric changes. In Gut-BrainPeptides in the New Millennium. Y. Tache, Y. Goto, G. Ohning, T. Yamada (eds), Los Angeles,CURE Foundation, 2002, pp. 339-352.

3. Takahashi S, Takeuchi K, Okabe S. EP4 receptor mediation of prostaglandin E2-stimulatedmucus secretion by rabbit gastric epithelial cells. Biochem Pharmacol 1999; 58: 1997-2002.

4. Takeuchi K, Ukawa H, Furukawa O et al. Prostaglandin E receptor subtypes involved instimulation of gastroduodenal bicarbonate secretion in rats and mice. J Physiol Pharmacol1999; 50: 155-167.

5. Araki H, Ukawa H, Sugawa Y, Yagi K, Suzuki K, Takeuchi K. The roles of prostaglandin Ereceptor subtypes in the cytoprotective action of prostaglandin E2 in rat stomach. AlimentPharmacol Ther 2000; 14: 116-124.

6. Sanders KM. Role of prostaglandins in regulating gastric motility. Am J Physiol 1984; 247:G117-126.

7. Johansson C, Kollberg B, Nordemar R, Samuelson K, Bergstrom S. Protective effect ofprostaglandin E2 in the gastrointestinal tract during indomethacin treatment of rheumaticdiseases. Gastroenterology 1980; 78: 479-483.

8. Bensen WG, Fiechtner JJ, McMillen JI et al. Treatment of osteoarthritis with celecoxib, acyclooxygenase-2 inhibitor: a randomized controlled trial. Mayo Clin Proc 1999; 74: 1095-1105.

9. Bombardier C, Laine L, Reicin A et al. VIGOR Study Group. Comparison of uppergastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. VIGORStudy Group. N Engl J Med 2000; 343: 1520-1528.

10. Silverstein FE, Faich G, Goldstein JL et al. Gastrointestinal toxicity with celecoxib vsnonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASSstudy: A randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA 2000;284: 1247-1255.

220

11. Watson DJ, Harper SE, Zhao PL, Quan H, Bolognese JA, Simon TJ. Gastrointestinal tolerabilityof the selective cyclooxygenase-2 (COX-2) inhibitor rofecoxib compared with nonselectiveCOX-1 and COX-2 inhibitors in osteoarthritis. Arch Intern Med 2000; 160: 2998-3003.

12. Aalykke C, Lauritsen JM, Hallas J, Reinholdt S, Krogfelt K, Lauritsen K. Helicobacter pyloriand risk of ulcer bleeding among users of nonsteroidal anti-inflammatory drugs: a case-controlstudy. Gastroenterology 1999; 116: 1305-1309.

13. Malfertheiner P, Labenz J. Does Helicobacter pylori status affect nonsteroidal anti-inflammatory drug-associated gastroduodenal pathology? Am J Med 1998; 104: 35S-40S.

14. Chan FK, Sung JJ, Chung SC et al. Randomized trial of eradication of Helicobacter pyloribefore non-steroidal anti-inflammatory drug therapy to prevent peptic ulcers. Lancet 1997; 350:975-979.

15. Hawkey CJ, Tulassay Z, Szczepanski L et al. Randomized controlled trial of Helicobacterpylori eradication in patients on non-steroidal anti-inflammatory drugs: HELP NSAIDs study.Helicobacter Eradication for Lesion Prevention. Lancet 1998; 352: 1016-1021.

16. Krawisz JE, Sharon P, Stenson WF. Quantitative assay for acute intestinal inflammation basedon myeloperoxidase activity. Assessment of inflammation in rat and hamster models.Gastroenterology 1984; 87: 1344-1350.

17. Lee M, Feldman M. Nonessential role of leukotrienes as mediators of acute gastric mucosalinjury induced by aspirin in rats. Dig Dis Sci 1992; 37: 1282-1287.

18. Lee M, Feldman M. Age-related reductions in gastric mucosal prostaglandin levels increasesusceptibility to aspirin-induced injury in rats. Gastroenterology 1994; 107: 1746-1750.

19. Jablonowski H, Hengels KJ, Kraemer N, Geis G, Opferkuch W, Strohmeyer G. Effects ofHelicobacter pylori on histamine and carbachol stimulated acid secretion by human parietalcells. Gut 1994; 35: 755-757.

20. Taguchi Y, Kaito M, Gabazza EC et al. Helicobacter pylori inhibits the secretory activity ofgastric parietal cells in patients with chronic gastritis. An ultrastructural study. Scand JGastroenterol 1997; 32: 656-663.

21. Crabtree JE, Shallcross TM, Heatley RV, Wyatt JI. Mucosal tumour necrosis factor alpha andinterleukin-6 in patients with Helicobacter pylori associated gastritis. Gut 1991; 32: 1473-1477.

22. Kondo S, Shinomura Y, Kanayama S et al. Interleukin-1 beta inhibits gastric histamine secretionand synthesis in the rat. Am J Physiol 1994; 267: G966-971.

23. Juttner S, Cramer T, Wessler S et al. Helicobacter pylori stimulates host cyclooxygenase-2 genetranscription: critical importance of MEK/ERK-dependent activation of USF1/-2 and CREBtranscription factors. Cell Microbiol 2003; 5: 821-834.

24. Kim JS, Kim JM, Jung HC, Song IS. Expression of cyclooxygenase-2 in human neutrophilsactivated by Helicobacter pylori water-soluble proteins: possible involvement of NF-kappaBand MAP kinase signaling pathway. Dig Dis Sci 2001; 46: 2277-2284.

25. Yokotani K, Okuma Y, Osumi Y. Inhibition of vagally mediated gastric acid secretion byactivation of central prostanoid EP3 receptors in urethane-anaesthetized rats. Br J Pharmacol1996; 117: 653-656.

26. Xiao F, Furuta T, Takashima M, Shirai N, Hanai H. Effects of cyclooxygenase-2 inhibitor ongastric acid secretion in Helicobacter pylori-infected C57BL/6 mice. Scand J Gastroenterol2001; 36: 577-583.

27. Utsunomiya I, Ito M, Watanabe K, Tsurufuji S, Matsushima K, Oh-ishi S. Infiltration ofneutrophils by intrapleural injection of tumour necrosis factor, interleukin-1, and interleukin-8in rats, and its modification by actinomycin D. Br J Pharmacol 1996; 117: 611-614.

28. Yamada H, Aihara T, Okabe S. Mechanism for Helicobacter pylori stimulation of interleukin-8 production in a gastric epithelial cell line (MKN 28): roles of mitogen-activated protein kinaseand interleukin-1beta. Biochem Pharmacol 2001; 61: 1595-1604.

221

29. Crabtree JE. Role of cytokines in pathogenesis of Helicobacter pylori-induced mucosaldamage. Dig Dis Sci 1998 Sep; 43: 46S-55S.

30. Suzuki K, Araki H, Mizoguchi H, Furukawa O, Takeuchi K. Prostaglandin E inhibitsindomethacin-induced gastric lesions through EP-1 receptors. Digestion 2001; 63: 92-101.

31. Harada N, Okajima K, Liu W, Uchiba M. Activated neutrophils impair gastric cytoprotectionrole of neutrophil elastase. Dig Dis Sci 2000; 45: 1210-1216.

32. Takeuchi K, Nishiwaki H, Okada M, Niida H, Okabe S. Bilateral adrenalectomy worsens gastricmucosal lesions induced by indomethacin in the rat. Role of enhanced gastric motility.Gastroenterology 1989; 97: 284-293.

33. Tanaka A, Araki H, Komoike Y, Hase S, Takeuchi K. Inhibition of both COX-1 and COX-2 isrequired for development of gastric damage in response to nonsteroidal antiinflammatory drugs.J Physiol Paris 2001; 95: 21-27.

34. Stika CS, Gross GA, Leguizamon G et al. A prospective randomized safety trial of celecoxibfor treatment of preterm labor. Am J Obstet Gynecol 2002; 187: 653-660.

35. Ebata M, Keto Y, Okabe S. Effects of indomethacin on the development of ulcers inHelicobacter pylori-infected Mongolian gerbils. Ulcer Res 2001; 28: 197.

36. Takeuchi K, Ueshima K, Hironaka Y, Fujioka Y, Matsumoto J, Okabe S. Oxygen free radicalsand lipid peroxidation in the pathogenesis of gastric mucosal lesions induced by indomethacinin rats. Relation to gastric hypermotility. Digestion 1991; 49: 175-184.

R e c e i v e d : November 30, 2003A c c e p t e d : February 20, 2004

Author�s address: Susumu Okabe, PhD, Department of Applied Pharmacology, KyotoPharmaceutical University, Misasagi, Yamashina, Kyoto, Japan, 607-8414, Tel.+81-75-595-4651;Fax. +81-75-595-4757E-mail: okabe@mb.kyoto-phu.ac.jp

222