Pergamon 0306-3623(95)00037-2

Gen. Pharmac. Vol. 26, No. 7, pp. 1455-1459, 1995 Copyright © 1995 Elsevier Science Inc.

Printed in Great Britain. All rights reserved. 0306-3623/95 $29.00 + 0.00

REVIEW

Non-Steroidal Anti-Inflammatory Drug-Induced Gastric Damage in Experimental Animals:

Underlying Pathological Mechanisms M I C H A E L A. T R E V E T H I C K * I A N OAKLEY, N I C H O L A S M. CL A Y T O N and P E T E R S T R O N G

Biology Division, Glaxo Research and Development Ltd., Park Road, Ware, Herts, SG12 ODP, U.K. [Tel: 01920-882796; Fa~" 01920-882263]

(Received 12 December 1994)

INTRODUCTION

The effects of chronic use of non-steroidal anti-inflam- matory drugs (NSAIDs) on the gastrointestinal tract continue to pose a major challenge for rheumatolo- gists (Ehsanullah et al., 1988; Roth and Bennett, 1987). NSAIDs have been reported to induce a wide spectrum of mucosal damage ranging from erosions (damage to the mucosa) to ulcers (damage penetrating through to the muscularis mucosa). Erosions are mainly found in the fundus, and both develop and heal rapidly with little overall clinical significance (McCarthy, 1989; Soil et al.,

1991). Ulcers, in contrast, occur principally in the an- trum (Roth and Bennett, 1987), have a more protracted development and can persist for long periods of time. The major risk, however, is that of the ulcer damaging important blood vessels, generating a possibility of fa- tal haemorrhage (McCarthy, 1989; Soil etaL, 1991). This analysis suggests that in terms specifically of gastric

damage, it is ulceration of the antrum that is of real significance rather than damage to the fundus (McCar- thy, 1989; Graham, 1990; Soil et al., 1991).

However, experimental research has concentrated on investigating NSAID-induced damage to the fundus that is typically superficial, with penetrating ulcers sel- dom occurring (McCarthy, 1990). If these two pathol- ogies have different underlying mechanisms (Soil et al., 1991), research on fundic damage may have limited rel- evance to the clinical setting.

Some years ago, Satoh et al. (1981) described antral ulceration in re-fed rats caused by a single dose of in- domethacin. Table 1 compares the characteristics of

*To whom all correspondence should be addressed.

clinical ulceration induced by NSAIDs with those of the pathology in the re-fed rat and demonstrates marked similarity utilising a variety of criteria encompassing biochemical, histological and pharmacological criteria. We have therefore used the re-fed rat to investigate the mechanisms responsible for the pathogenesis of NSAID-induced antral ulceration in the hope that this

may be predictive for the clinical setting.

ROLE OF ACID/PEPSIN

Clinical studies demonstrated that standard doses of histamine H2 antagonists do not inhibit the devel- opment of NSAID-induced antral ulceration (Ehsanul- lah et al., 1988; Robinson et al., 1991). However, NSAID-induced duodenal ulceration was inhibited. Animal studies produce similar results, with cimetidine being a more effective inhibitor of fundic damage than of antral ulceration (Satoh et al,, 1981; Bunce et al.,

1981). Taken together, these results suggest that there may be no role for acid/pepsin in the genesis of antral ulceration.

We have re-examined the potential role of acid in NSAID-induced antral ulceration by comparing the effects of the long-acting H2 receptor antagonist lox- tidine (Boyd and Wormsley, 1984) and the shorter act- ing H2 receptor antagonist ranitidine (Daly et aL, 1981) on indomethacin-induced ulceration of the gastric an- trum in the re-fed rat. Pretreatment with either loxti- dine (0.1-10.0 mg/kg p.o.) or ranitidine (10-100 mg/kg p.o.) produced a dose-dependent inhibition of the area of macroscopic damage to the antrum induced by ino domethacin with near maximal inhibition at 10 and 100

1455

1456 Michael A. Trevethick et al.

Table 1. Comparison of NSAID-induced ulceration of the gastric antrum in rat and human

Human Rat

NSAID damage Spectrum of lesions

Role by food Infiltration by inflammatory cells Eicosanoid Synthesis Pharmacological Manipulation

Primarily antral* Erosions and deep

penetrating ulcers* Enhances ulceration§ Predominantly neutrophils[I ~PGtLTB, U PGs potent inhibitors. H2

antagonists not effective*

Primarily antralt Erosions and deep $

penetrating ulcers Enhances ulcerationt Predominantly neutrophils¶ ~,PG tLTB4~ PGs potent inhibitors.

Cimetidine not effective.t

References: *Soil et al., 1991; #Satoh et al., 1981; :~Trevethick et al., 1994a; §Hudson et aL, 1992; [[Hudson et al., 1993; ¶Trevethick et al., 1993.

PGs, prostaglandins; LT, leukotriene.

m g / k g p.o., respectively. Compar ison of the doses re-

quired to reduce the area o f ulceration by 50~/0 with

those required to effect a similar reduction in acid secre-

tion revealed that ranitidine was more potent as an in-

hibitor of acid secretion than as an anti-ulcerant in this

model (Table 2). In marked contrast, however, loxti-

dine had almost identical potency in both systems.

These results suggest that acid/pepsin may have a role

in the pathology of indomethacin-induced antral ul-

ceration, but this can only be uncovered with long-

lasting inhibitors o f acid secretion.

In subsequent studies, we examined the histological

nature o f indomethacin-induced antral ulceration and

that of the protection afforded by loxtidine (3 m g / k g

p.o.) and ranitidine (100 m g / k g p.o.). Histologically,

indomethacin-induced lesions in the gastric antrum

are primarily ulcers (lesions extending down to the

musclaris mucosa) with a few superficial and deep ero-

sions (Table 3). Both loxtidine and ranitidine substan-

tially inhibited the area of ulceration detected macro-

scopically. In contrast, these pre-treatments did not

significantly reduce the area of damage detected on

histological examination; however, they did alter the

profile o f lesions detected. These "microscopic" lesions

consisted almost entirely o f superficial and deep ero-

sions, the proport ions of which were significantly

greater than with indomethacin alone (Table 3). The

hypothesis that acid/pepsin was important in the con-

version o f erosions to ulcers was investigated further

by re-introducing acid into the gastric lumen of animals

pre-treated with loxtidine or ranitidine. Hour ly dosing

with 160 m M HC1 completely reversed the protective

effect o f both histamine H2 antagonists when assessed

by both macroscopic and histological means of analy-

sis. Importantly, acid dosing itself did not affect the

ability of indomethacin to cause ulceration nor did it

influence the distribution o f lesions detected histolog-

ically (Table 3).

Two conclusions were drawn from these studies. First,

acid/pepsin is important in the pathogenesis of

indt~methacin-induced ulceration of the gastric antrum,

but [he role o f acid can only be uncovered if secretion

is almost completely suppressed for the duration of the

experiment. This situation can be revealed if rats are

dosed with low doses o f loxtidine (a long lasting in-

hibitor of acid secretion) or high doses o f shorter act-

ing compounds such as ranitidine (or cimetidine). Sec-

ond, these data are consistent with the idea that

ulceration is a two-stage process: an initial "mucosal

lesion" is apparently formed that is then converted into

a penetrating ulcer by the action of acid/pepsin. Treat-

ment with a long-lasting inhibitor o f the secretion o f

acid/pepsin would therefore prevent the progression of

the initial mucosal lesion into a frank ulcer.

Table 2. Comparison of anti-ulcerant and acid inhibitory potencies of ranitidine and loxtidine

Inhibition of antral ulceration* Inhibition of acid secretion'~

(mg/kg p.o.) (mg/kd i.d.)

Ranitidine 26.0 1.6 (2.2-50.0) (0.9-2.9)

Loxtidine 0.3 0.4 (0.2-0,4) (0.3-0.5)

Results are presented as IDso values, with 95°70 confidence intervals in paren- theses.

* Data obtained using the protocol described by Trevethick et al. (1993). t Data obtained using the rat perfused stomach preparation as described by Par-

sons (1969). In brief, acid secretion was induced by intravenous infusion of histamine (0.1 mg/kg/min) until the response had plateaued. The antagonist under study was then administered intraduodenally and the decrease in acid secretion determined.

Gastric damage in experimental animals 1457

Table 3. Reversal of the protective effect of ranitidine (100 mg/kg p.o.) and loxtidine (3 mg/kg p.o.) by hourly dosing with 160 mM HCI

Treatment

% Area of damage to antrum % Distribution of microscopic damage

Macroscopic Microscopic Ulceration Deep erosions Superficial erosions

Hourly dosing with 0.5% methyl cellulose Indomethacin 16.08 ± 1.12 Indomethacin + ranitidine 2.50 ± 1.06" Indomethacin + loxtidine 3.83 ± 1.33"

Hourly dosing with 160 mM HCI Indomethacin 14.75 _+ 2.02 Indomethacin + ranitidine 17.00 _+ 2.43 Indomethacin + loxtidine 17.00 ± 2.13

19.77 ± 1.76 80.64 ± 5.68 18.27 ± 5.93 1.09 ± 0.78 9.81 + 2.29 19.33 ± 14.47" 48.50 ± 11.66 32.17 ± 7.91" 9.72 ± 3.35 34.67 + 12.25 46.17 _+ 11.70 18.83 _+ 9.36*

11.54 + 3.00 69.33 + 9.04 17.33 + 7.25 13.25 _+ 5.45 21.78 + 5.14 85.17 _+ 8.03 11.00 ± 5.96 3.83 + 3.45 15.37 ± 3.13 78.67 _ 8.06 16.33 +_ 7.83 5.0 _+ 1.26

Values are means _+ SEM. *P < 0.05 vs. indomethacin (Student's unpaired t-test).

I N F L A M M A T I O N A N D A N T R A L ULCERATION

Recent reports suggested that the pathogenesis o f

NSAID-induced fundic lesions in rats and rabbits is

a neutrophil-dependent process (Wallace and Granger,

1992) and that leukotriene (LT)B4 generation may be

pivotal in the pathology. Stimulated by these reports,

we decided to investigate the role of neutrophils and

LTB, in the development o f indomethacin- induced ul-

ceration o f the rat gastric antrum. In particular, we in-

vestigated whether these cells initiate the primary

mucosal lesion predisposing the s tomach wall to de-

velop ulceration after exposure to acid/pepsin.

Our results revealed that indomethacin- induced ul-

ceration of the gastric antrum was associated with blood

neutrophilia, neutrophil infiltration into the antrum and

enhanced synthesis of LTB4 by both blood and antrum

(Trevethick et aL, 1993). Importantly, however, these

processes preceded the development of antral ulcera-

tion, suggesting that neutrophil infiltration may be a

pivotal event in the ulcerative process. This hypothesis

was strengthened by our observation that the anti-

inflammatory corticosteroid, dexamethasone, inhibited

indomethacin- induced antral ulceration and the as-

sociated neutrophil infiltration but did not affect

indomethacin- induced b lood neutrophilia. As dexa-

methasone also inhibited LTB4 synthesis f rom the

blood, we hypothesised that dexamethasone inhibited

antral ulceration by reducing cell infiltration as a con-

sequence o f inhibit ion o f the potent chemoat t ractant

LTB4 (Trevethick et al., 1993). However, pre- treatment

with the potent and selective leukotiene synthesis in-

hibitor, MK886, al though totally inhibiting LTB, syn-

thesis from both blood and antrum, did not alter the

ulcerogenic activity of indomethacin in the an t rum or

the associated neutrophil infiltration (Trevethick et al.,

1994a). Thus, although controversial, these studies sug-

gest that whereas neutrophil infiltration may be a prime

event in the process of antral ulceration, leukotrienes

are not major chemoattractants in this pathology.

To further define the role of neutrophil infiltration

in antral ulceration, we developed an ant iserum that

selectively depleted neutrophil numbers in rat blood.

Acute (single dose) or chronic (multiple doses over

2 days) pre-treatment with this antisera reduced the

number of circulating neutrophils to near zero and to-

tally abolished the increase in blood neutrophilia and

infiltration into the an t rum after administrat ion o f in-

domethacin (Trevethick et al., 1994b). These treatments

did not, however, affect the ability of indomethacin to

induce ulceration of the gastric an t rum both in terms

o f the total area o f damage or in the distr ibution of

Table 4. Comparison of single and multiple doses of anti-neutrophil serum on indomethacin-induced ulceration of the rat gastric antrum

Single Dose Multiple Doses

NIS ANS NIS ANS

% Area damage Macroscopic 26.1 _+ 6.3 22.3 _+ 6.7 21.1 ± 4.3 24.3 _+ 6.8 Microscopic 22.4 _+ 4.7 22.5 _+ 4.7 24.6 ± 5.5 28.3 _+ 6.9

Blood neutrophils (× 103/~.1) 3.26 _+ 0.19 0.13 _+ 0.10" 3.36 _+ 0.45 0.03 ± 0.01"

Antral LTB4 synthesis (ng/g/30 rain) 31.5 + 15.1 5.9 __. 2.3" 33.2 _+ 13.1 8.8 _+ 2.6*

Blood LTB, synthesis (ng/g/30 rain) 24.3 +_ 3.6 5,2 _+ 0.3* 27.1 + 2.9 6.8 _+ 2.5*

Values are means ± SEM, n = 5. NIS and ANS dosed at 0.4 ml/100 g. Multiple doses were administerd at 36, 18 and l hr before indomethacin, whereas the single dose was given 18 hr before indomethacin. All parameters were assessed 6 hr after dosing with indomethacin.

*P < 0.05 compared with NIS. NIS, non-immune serum; ANS, anti-neutrophil serum. From Trevethick et al. (1995).

1458 Michael A. Trevethick et al.

lesions detected microscopically (Table 4) (Trevethick et al., 1995). Thus, our data suggest that in contrast to NSAID-induced fundic damage, neutrophil infiltra- tion does not contribute to the ulcerogenic effects of indomethacin in the gastric antrum. Overall, our anal- ysis suggests that the mechanisms underlying the gastric-damaging effects of indomethacin in the fundus and antrum of the rat are clearly different.

ALTERATIONS IN MUCOSAL BLOOD FLOW: THE INITIATING FACTOR IN

NSAID-INDUCED ULCERATION?

Several publications suggested that gastric ulceration is more common in those areas of the stomach where there is little or no collateral blood flow (Piasecki, 1974; Prokopiw et al., 1991). The fundic region of the stom- ach possesses numerous horizontal interconnections be- tween capillaries, whereas such connections rarely oc- cur in the antrum (Prokopiw et al., 1991). In the fundus, therefore, the multiple interconnections between capil- laries may provide several lateral interconnections that would maintain mucosal blood flow and limit progres- sion of superficial ischaemia and hence ulcer forma- tion. In the antrum, however, such interconnecting anastomoses are rare, and thus reductions in blood sup- ply may render areas of the mucosa ischaemic and sus- ceptible to ulcer formation by luminal aggressors such as acid.

Although there are no reports comparing the effect

of NSAIDs on blood flow in the antrum and fundus, it is clear these agents can reduce mucosal blood flow in a variety of animal species, including humans (see Vigneri et al., 1992). In addition, ingestion of NSAIDs is associated with an increased luminal uptake of su- crose (Meddings et aL, 1994) suggesting that a loss of the barrier function of the gastric mucosa may occur as a consequence of ischaemia. Three mechanisms have been proposed to explain the reduction in blood flow. First, is the direct inhibition of vasodilator prostaglan- din synthesis (Vigneri et al., 1992). Second, NSAIDs have been shown to induce spasm of gastric smooth muscle (e.g., Piasecki and Thrasivoulou, 1993) and in- duce ischaemia by direct compression of the vascula- ture. The third, and more speculative, suggestion is that one of the major functions of mucosal prostaglandin synthesis is to inhibit mediator release from resident inflammatory cells such as mast cells (Hogaboam et

aL, 1993). Removal of this negative feedback by NSAIDs would allow increased mediator release with the subsequent induction of ischaemia.

Whatever the mechanism(s) of NSAID-induced re- ductions in blood flow, literature reports examining gas- tric vascular anatomy suggest that this is likely to lead

to ulceration in the antrum and not the fundus (Pia- secki, 1974; Prokopiw et al., 1991). Thus, alterations in blood flow may be the key initiating factor in the process of NSAID-induced antral ulceration.

The presence of food, however, is also critical for NSAID-induced antral ulceration in the rat and also

enhances antral damage in humans (Table 1). Food in the gastric lumen will stimulate acid secretion, provide an antigenic environment and increase antral motility. Thus, we envisage that the following two stages con- tribute to the mechanism of NSAID-induced antral ul- ceration. Stage 1, the initiation of mucosal lesions, oc- curs as a consequence of ischaemia induced after administration of an NSAID. Stage 2, the conversion of erosions to ulcers, occurs in the presence of a "weakened" mucosa after NSAID administration. The presence of luminal aggressors (food, acid, pepsin) will convert mucosal lesions into frank ulceration. This hy- pothesis probably explains why long-, but not short- acting, inhibitors of acid secretion suppress the forma- tion of ulcers but not lesions, because a prolonged and sustained inhibition of acid secretion would be required while food is present in the gastric lumen. In addition, our results also suggest that neutrophil infiltration oc- curs as a consequence of tissue damage and does not initate the process.

SUMMARY AND PERSPECTIVES

A major impetus to experimental studies examining the pathogenesis of NSAID-induced gastric damage is the hope that key mechanisms can be identified that may lead to the design of "safer" NSAIDs or to the development of novel cytoprotectives to co-administer with current NSAIDs. Virtually every hypothesis pro- posed to explain the pathogenesis of NSAID-induced gastric damage has arisen from studies examining fun- dic lesions. Our results suggest that not only is the pathogenesis of gastric damage in the fundus and an- trum of the rat potentially very different but that it is NSAID-induced damage to the rat gastric antrum (and not the fundus) that more closely resembles that of hu- mans. Thus, hypotheses constructed from experimen- tal studies examining fundic damage may not be predic- tive of the clinical setting. We would suggest, therefore, that future studies should concentrate on studying the pathogenesis of antral ulceration induced by NSAIDs. In particular, we believe that future research should as- sess whether NSAIDs do indeed reduce blood flow in the gastric antrum and define the mechanism(s) in- volved. Identification of these processes should sig- nificantly advance our understanding of antral ulcera- tion and may suggest novel approaches in the design of cytoprotective agents.

Gastric damage in experimental animals 1459

Recent work has es tabl ished tha t two dist inct forms

o f the enzyme cyclooxygenase (COX) can catalyse the

metabol ism of arachidonic acid and initiate prostaglan-

din synthesis. It is hypothes ised (De Wi t t et al., 1993)

tha t the ana lges ic /an t i - in f lammatory effect of current

NSAIDs is achieved t h r o u g h inhib i t ion of C O X 2,

whereas their side effects (such as ant ra l ulcerat ion) re-

sult as a consequence of inh ib i t ion of gastric C O X 1.

The recently described selective inh ib i to r of C O X 2,

NS398, has been shown to be analgesic and anti-inflam-

matory wi thou t causing gastric u lcera t ion (Futaki et

aL, 1993; Masferrer et al., 1994). I f key mechanisms

(such as a l terat ions in b lood flow) of NSAID- induced

ant ra l damage can be identified, then it would be

hypothesised tha t selective inhibi tors of C O X 1 would

be ulcerogenic and reduce ant ra l b lood flow, whereas

inhibi tors of COX 2 would not share ei ther o f these

properties. I f this hypothesis can be substantiated, then

inhibitors of COX 2 may be the next generation of"gas-

tric safe" NSAIDs.

REFERENCES

Boyd Eo J. S. and Wormsley K. G. (1984) Effects of loxtidine, a new histamine H: antagonist on 24 hour gastric secre- tion in man. Eur J. clin. Pharmac. 26, 443-447.

Bunce K. T., Daly M. J., Humphray J. M. and Stables R. (1981) H2 receptor antagonists protect against aspirin-induced gastric lesions in the rat. Agents Actions 11, 167-170.

Daly M. J., Humphray J. M. and Stables R. (1981) Some in vitro and in vivo actions of the new histamine H2 recep- tor antagonist, Ranitidine. Br. J. Pharmac. 72, 49-52.

De Witt D. L., Meade E. A. and Smith W. L. (1993) PGH synthase selectivity: the potential for safer non-steroidal anti-inflammatory drugs. Am. J. Med. 95, 455-465.

Ehsanullah R. S. B., Page M. C., Tildesley G. and Wood J. R. (1988) Prevention of gastroduodenal damage induced by non-steroidal anti-inflammatory drugs: controlled trial of ranitidine. Br. Med. J. 297, 1017-1021.

Futaki N., Yoshikawa K., Hamasai Y., Arai I., Higuchi S., Izuka H. and Otomo S. (1993) NS398, a novel non-steroidal anti- inflammatory drug with potent analgesic and anti-pyretic effects which causes minimal stomach lesions. Gen. Phar- mac. 24, 105-110.

Graham D. Y. (1990) The relationship between non-steroidal anti-inflammatory drug use and peptic ulcer disease. Gas- troenterol. Clin. North Am. 19, 171-182.

Hogaboam C. M., Bissonnette E. Y., Chin B. C., Befus A. D. and Wallace J. L. (1993) Prostaglandins inhibit inflamma- tory mediator release from rat mast cells. Gastroenterol- ogy 104, 122-129.

Hudson N., Hawthorene A. B., Cole A. T., Jones P. D. E. and Hawkey C. J. (1992) Mechanisms of gastric and duo- denal damage and protection. Hepato. Gastroenterol. 39, 31-36.

Hudson N., Balsitis M., Everitt S. and Hawkey C. J. (1993) Enhanced gastric mucosal leukotriene B, synthesis in pa-

tients taking non-steropidal anti-inflammatory drugs. Gut 34, 742-747.

Masferrer J. L., Zweifel B. S., Manning P. T., Hauser S. D., Leahy K. M., Smith W. G., Isakson P. C. and Seibert K. (1994) Selective inhibition of inducible cyclo-oxygenase 2 in vivo is anti-inflammatory and non ulcerogenic. Proc. natn. Acad. Sci. U.S.A. 91, 3228-3232.

McCarthy D. M. (1989) Non-steroidal anti-inflammatory drug- induced ulcers: management by traditional therapies. Gas- troenterology 96, 662-674.

McCarthy D. M. (1990) NSAID-induced gastro-inetestinal d a m a g e - a critical review of prophylaxis and therapy. J. Clin. Gastroenterol. 12 (Suppl 2), S13-$20.

Meddings J. B., Sutherland L. R., Byles N. 1. and Wallace J. L. (1993) Sucrose: a novel permeability marker for gas- troduodenal disease. Gastroenterology 104, 1619-1626.

Parsons M. E. (1969) Quantitative studies of drug-induced gastric secretion. Ph.D. thesis, University of London.

Piasecki C. (1974) Blood supply to the human gastroduodenal mucosa with special reference to the ulcer bearing areas. J. Anat. 118, 295-335.

Piasecki C. K. and Thrasivoulou C. (1993) Spasm of the gas- tric muscularis mucosae might play a key role in causing focal mucosal ischaemia and ulceration. Dig. Dis. Sci. 38, 1183-1189.

Prokopiw I., Hyanna-Liepert T. T., Dinda P. K., Prentice R. S. A. and Beck I. T. (1991) The microvascular anatomy of the canine stomach. Gastroenterology 100, 638-647.

Robinson M. G., Griffin J. J., Bowers J., Kogan E J., Kogut D. G., et al (1991) Effect of ranitidine on gastroduodenal mucosal damage induced by non-steroidal anti-inflamma- tory drugs. Dig. Dis. Sci. 34, 424-428.

Roth S. H. and Bennett R. E. (1987) Non-steroidal anti-inflam- matory drug gastropathy: recognition and response. Arch. lntern. Med. 147, 2093-2100.

Satoh H., lnada I., Hirate T. and Maki Y. (1981) Indometha- cin produces gastric antral ulcers in the refed rat. Gastroen- terology 81, 719-725.

Soil A. H., Weinstein W. M., Kurata J. and McCarthy D. (1991) Non-steroidal anti-inflammatory drugs and peptic ulcer dis- ease. Ann. Intern. Med. 114, 307-319.

Trevethick M. A., Clayton N. M., Strong P. and Harman I. W. (1993) Do infiltrating neutrophils contribute to the patho- genesis of indomethacin induced ulceration of the rat gas- tric antrum? Gut 34, 156-160.

Trevethick M. A., Clayton N. M., Bahl A. K., Strong P. and Harman I. W. (1994a) Leukotrienes do not contribute to the pathogenesis of indomethacin-induced ulceration of the gastric antrum in the re-fed rat. Agents Actions 41, 179-183.

Trevethick M. A., Clayton N. M., Bahl A. K., Strong P. and Harman 1. W. (1994b) Neutrophil depletion does not pre- vent indomethacin induced ulceration of the rat gastric an- trum. Agents Actions 41, C226-C227.

Trevethick M. A., Bahl A. K., Clayton N. M., Strong P., San- jar S. and Harman I. W. (1995) Neutrophil infiltration does not contribute to the ulcerogenic effects of indomethacin in the rat gastric antrum. Agents Actions 43, 39-43.

Vigneri S., Scialabba A., Termini R., Germana B., Vianello F., Grassi S. A., Plebani M. and Di Mario G. (1992) Patho- physiology of the gastric microcirculation. Ital. J. Gastroen- terol. 24, 22-30.

Wallace J. L. and Granger D. N. (1992) Pathogenesis of NSA1D-gastropathy-are neutrophils the culprits? Trends Pharmac. Sci. 13, 129-132.

GP 26:7-B