Mucosal reactive oxygen metabolite production in duodenal ulcer disease

Transcription

1 Gut 1992; 33: Gastrointestinal Siene Researh Unit, Department of Histopathology, and Inflammation Researh Group, The London Hospital Medial College, London G R Davies N J Simmonds T R J Stevens A Grandison D R Blake D S Rampton Correspondene to: Dr G R Davies, Gastrointestinal Siene, Researh Unit, The London Hospital Medial College, London l 2AD. Aepted for publiation 1 1 June 1992 Muosal reative oxygen metabolite prodution in duodenal uler disease G R Davies, N J Simmonds, T R J Stevens, A Grandison, D R Blake, D S Rampton Abstrat To investigate the hypothesis that reative oxygen metabolites are important in the pathophysiology of duodenal uler disease, their prodution by duodenal muosal biopsy speimens was measured using luminol and luigenin amplified hemiluminesene. Luminol hemiluminesene, expressed as bakground orreted median photon emission/mg/min x 13 (95% onfidene intervals), was inreased in duodenal inflammation as assessed marosopially: ulers 2.3 (4.8 to 51.3), n=29; severe duodenitis 13.9 (6.6 to 75.3), n=16; mild duodenitis. (-.5 to.8), n=56; ontrols -.8 (-1.3 to -.1), n=41; p=.1, Kruskal-Wallis) and mirosopially: severe 17. (9.3 to 51.3), n=12; moderate.3 (-2.8 to 5.8), n=17; mild -.1 (-1.8 to 1.), n=17; ontrols -*8 (-1.6 to.), n=15; (p=.1). Luminol hemiluminesene was diretly related to both the marosopi and mirosopi severity of duodenal damage (Spearman's R=+.53, +.55 respetively, both p=.1), to histohemial assessment (myeloperoxidase ativity) of neutrophil infiltration (R= +.63; p=.4), and to luigenin hemiluminesene (R=+-56, p=.2). Luminol hemiluminesene was inhibited by sodium azide (-8%), atalase (-73%), and dimethyl sulphoxide (-24%). Superoxide dismutase inhibited luigenin more than luminol dependent hemiluminesene (-61% and -7% respetively, p<.5). Within disease groups, Heliobater pylon antral infetion was assoiated with inreased duodenal hemiluminesene, whereas smoking, alohol, and use ofnsaids or H2 blokers had no influene. Their disease related generation in duodenal muosa supports a role for reative oxygen metabolites in the pathogenesis of duodenitis and duodenal uler. These metabolites might inlude superoxide, hydrogen peroxide, hydroxyl, and produts of myeloperoxidase ativity. (Gut 1992; 33: ) Reative oxygen metabolite (ROM) prodution by inflammatory ells, while essential for survival, an be inappropriate or exessive, ontributing more to host destrution than defene.' Inreased ROM prodution has been linked to the pathogenesis of many inflammatory onditions, inluding those of the human gastrointestinal trat.2 Information diretly linking ROM prodution with duodenal uleration is mostly onfined to animal models.34 One study to date, however, has shown uler healing in humans using savengers or inhibitors of ROM prodution5; in addition, reent data suggest that the pathogeniity of Heliobater pylori in the upper gastrointestinal trat might be related to stimulation of neutrophil ROM prodution.6 We have now looked for evidene of disease related prodution of ROMs in duodenal biopsy speimens from patients with duodenal uler disease and duodenitis using hemiluminesene assays. We have orrelated the results with marosopi and mirosopi appearanes and myeloperoxidase ativity. To show that ROMs were the oxidants deteted by the hemiluminesene assays and in an attempt to identify the type and soure of ROM involved, we examined the effets of various enzymes, enzyme inhibitors, and ROM savengers and ompared the results of luminol and luigenin amplified hemiluminesene responses. We also examined the influene of known risk fators for duodenal uler disease suh as antral H pylori infetion, smoking, alohol, non-steroidal anti-inflammatory drugs (NSAIDs) and of urrent use of H2 blokers, on duodenal ROM prodution. Methods 1467 RAGNTS Chemials were obtained from Sigma Chemial Co, Poole, Dorset. Stok solutions of luminol (5 mg/ml in dimethyl sulphoxide (DMSO)) and luigenin (3,uM in phosphate buffered saline (PBS)) were prepared weekly and stored at 4 C. Other reagents were prepared on the day of eah experiment. PBS with added gluose (5 mm), alium, and magnesium (both 1 mm) was preoxygenated and adjusted to ph 7-4. This solution was used to dilute luminol and luigenin to a final onentration of 75 FtM, and also as a transport medium for biopsy speimens. Superoxide dismutase-cu Zn (SOD) from bovine erythroytes (final onentration 3 U/ml), atalase from bovine liver (3 U/ml), and sodium azide (. 1 mm) were made up in PBS. When used as a ROM savenger, DMSO was used at a final onentration of 5%. Inativated SOD was prepared by boiling a solution in PBS for two hours, and inativated atalase by heating the powder to 1'C for two hours and then leaving the solution in PBS at room temperature for 24 hours. Inativation of SOD and atalase was onfirmed as previously desribed.' Reagents for myeloperoxidase staining were obtained from Sigma in kit form. PATINTS Muosal biopsy speimens (2-6) were obtained from 145 patients undergoing routine endosopy at The Royal London Hospital. Patient hara- Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.

2 1468 Davies, Simmonds, Stevens, Grandison, Blake, Rampton TABL I Patient harateristis teristis are shown in Table I. These studies were approved by the Tower Hamlets Health Authority thial Committee. Control Mild duodenit's Severe duodenitis Duodenal uler No Age (range) (yr) 52 (23-83) 49 (41-82) 45 (26-65) 57 (22-8) H pylori+() Smokers (%) NSAID use (%) H2 antagonist use (%) rlign; 14iah aionoi -aitrhni use-t :k-{ n /o) -1 1o I1 7/ R *(>2 U/week females; >3 U/week males) X 15 X,o m 1 i U). C) L -5 i I 1 p = -1 liil * i illill p =.1 r P =.1 1 rp = O.1 - Control Mild Severe duodenitis duodenitis Figure 1: Prodution ofluminol amplified hemiluminesene by duodenal m speimens aording to the endosopi grade. Results are shown as photonslmgilmnute after subtration of bakground. Points represent the mean oftwo or more biopsy s same patient. Medians and Mann-Whitney U test p values are shown. m * TABL II Histologial (haematoxvlin and eosin) grading sheme for duodenal biopsies Normal: Mild: Moderate: Severe: STUDY PROTOCOL No inrease in inflammatorv ell infiltrate or disruption of ellular arhiteture. Santy inflammatory ells with preservation of glandular and villous arhiteture. Presene of moderate to severe inflammators infiltrate but with preservation of glandular and villous arhiteture. Presene of a severe inflammators infiltrate with one or more of the following: branhing rvpts, loss of villous struture, vasular ongestion, uleration, glandular disruption. Drug, alohol, and igarette onsumption, were reorded. Duodenal inflammation was graded endosopially aording to a modified Lanza sale7 as follows: muosal injetion, haemorrhagi and erosive lesions were eah sored on a sale of -3, and the three sores summated to give the following groups: ontrol sore=, mild duodenitis= 1-3, severe duodenitis= >4. Duodenal ulers, onsidered separately, were defined as a breah in the muosa greater than.5 m diameter, with depth. videne for H pylon infetion was based on results of antral biopsy ulture and/or ommerial urease (CLO) test. Biopsy speimens (2-6) were taken from the first part of the duodenum for hemiluminesene assay, and, in the ase of abnormal muosa, either from areas of maximum inflammation or the edge of uleration. CHMILUMINSCNC ASSAY FOR ROMS Light emission aompanying the generation of ROMs an be amplified by hemiluminogeni probes suh as luminol or luigenin: these moleules form exited states when exposed to * oxidants, relaxing to ground state with the emission of light, detetable using a photomultiplier devie.8 Muosal biopsy speimens were washed and transported in preoxygenated PBS; hemiluminesene assays were arried out within three hours using a Pakard Tri Carb 16 CA liquid sintillation analyser operated in the 'out of oinidene' mode. Sintillation tubes ontaining 1 ml of either 75 FtM luminol or 75 FtM luigenin were preounted for bakground hemiluminesene. After addition of the biopsy tissue, sintillation tubes were reounted for five minutes. The baseline hemiluminesene for,. eah biopsy is expressed as photons/minute/mg - * * - wet weight x 13 after subtration of vial bakground, and the results for eah patient were averaged. In paired biopsy speimens from the same site with high hemiluminesene (arbitrarily defined as >4 photons/minute after subtration of bakground), inhibitors enzymes, or savengers affeting ROM prodution or ontrols (PBS or heat inativated atalase or SOD) were added to the vial, and samples were reounted for a further five minutes. ffets of inhibitors and savengers are Duodenal expressed as the % hange in hemiluminesene uler zluosal biopsy ompared with the ontrol value. Immediately after hemiluminesene assays, biopsy peimens fromthe speimens were dry blotted, weighed, and preserved for histologial analysis (see below). Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.

3 Muosal reative oxygen metabolite prodution in duodenal uler disease 1469 x._ J Cn.,o m.._ Un ) -_ ~ % * 4 %... ', * 4 Spearman's R = +-56 p = Photon emission minus bakground/min/mg x 13 (luigenin) Figure 2: Comparison ofhemiluminesene obtainedfrom duodenal biopsy speimens from same patient using either luminol or luigenin amplifiation. Reproduibility of the method (expressed for eah marosopi group as the median 95% onfidene interval (CI) perentage differenes between the seond reading and the first, in patients who had two biopsies taken from the same area at the same time) was as follows: ontrols 22 (-6 to 66) n=3 pairs; mild duodenitis 47 (-38 to 122), n=32; severe duodenitis 23 (-128 to 99), n=24; ulers 18 (-12 to 84), n=28. MICROSCOPIC ASSSSMNTS In 66 patients, haematoxylin and eosin stained setions of formalin preserved biopsy tissues were assessed by two independent pathologists: the semiquantitative grading system used is summarised in Table II. In 11 biopsies, the hemiluminesene result was ompared with the myeloperoxidase ativity, an index of neutrophil infiltration.9 For this assessment biopsies were snap frozen, and later setioned, fixed for 1 minutes in glutaraldehyde-aetone, and stained aording to the Sigma proedure 391 TABL iii ffet ofaddition ofenzymes, inhibitors, and savengers on reative oxygen metabolite prodution by duodenal biopsies. Results are median % inhibition of hemiluminesene (95% onfidene intervals) overfive minutes ompared with addition ofphosphate buffered saline (or inativated enzyme in the ase ofsuperoxide dismutase (SOD) and atalase) Chemiluminesene Agent Amplifier No Inhibition (95% CI) Azide Luminol 6-8 (-77 to -9)* DMSO Luminol 6-24 (-1 to -54) Catalase Luminol 7-73 (-61 to -77)* SOD Luminol 8-7 (-11 to -56)* SOD Luigenin 5-61 (-56 to -9)*t *p< 5 ompared with ontrol; tp< 5 ompared with SOD effets with luminol amplifier; DMSO=dimethyl sulphoxide. (but using double the stated reation times), whih is based on myeloperoxidase-mediated oxidation of diaminobenzidine. Setions were graded in arbitrary units aording to staining intensity. For either method, mirosopists were blinded to the other details of the biopsy speimen. STATISTICS Two tailed statistis were used throughout. Chemiluminesene prodution between the different marosopi and mirosopi groups was ompared initially using the Kruskal-Wallis non-parametri test for multiple groups, and subsequent Mann-Whitney U test to look for differenes between paired groups. The influene of risk fators within eah group was assessed using the Mann-Whitney U test, with orretions for multiple omparisons. Relationships between ROM prodution and mirosopi and marosopi sores were assessed by Spearman's 6 rank orrelation. The effet of inhibitors was assessed using the Wiloxon signed rank test for the paired variables. Results CHMILUMINSCNC ACCORDING TO MACROSCOPIC GRAD When duodenal appearanes were graded aording to the endosopi appearanes, luminol-amplified hemiluminesene (expressed as median photon emission/min/mg wet weight, after subtration of vial bakground (95% CI)) was inreased in duodenal uleration and inflammation: ulers 2.3 x 13 (4-8 to 51-3) n=29; severe duodenitis 13.9 (6.6 to 75.3) n= 16; mild duodenitis. (--5 to -8) n=56; ontrols --8 (-1-3 to --1) n=41; (p=-1 Kruskal- Wallis) (Fig 1). Luminol amplified hemiluminesene was positively orrelated with the Lanza grading of duodenal muosa (Spearman's R=+ 53, p=-1). Luminol and luigenin hemiluminesene were ompared in biopsy speimens from areas of similar marosopi damage in 41 patients. The results from the two methods were diretly related (Spearman's R=+-56, p=.2), however, luminol amplifiation tended to give greater hemiluminesene (Fig 2). In the smaller number of biopsy speimens assessed for hemi- Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.

4 147 l x * Uw e._ n.._ (). 5) - ~ base _ I p = p1 p = OpOO **..: I * _.* _ r p = 2 -l -5 Control Mild Moderate Severe Histologial grade Figure 3: Prodution ofluminol amplified hemiluminesene by duodenal muosal biopsy speimens aording to mirosopi (haematoxylin and eosin) assessment ofinflammation. Results are shown as photons/mglminute after subtration ofbakground. Medians and Mann- Whitney U test p values are shown. luminesene with luigenin, a trend to disease related inreases in hemiluminesene did not ahieve statistial signifiane (ulers and severe duodenitis -6 (-2.3 to 81) n=12; mild duodenitis -16 (-3-5 to.) n=22; ontrol -1[ (-14. to 2.1) n=7 (p=2 Kruskal- Wallis). CHMILUMINSCNC ACCORDING TO MICROSCOPIC GRAD AND MYLOPROXIDAS SCOR As for marosopi grading, the luminol hemiluminesene result was related to duodenal disease when inflammation was graded mirosopially: severe 17. (9.3 to 51[3) n= 12; moderate 3 (-2-8 to 5.8) n=17; mild --1 l Davies, Simmonds, Stevens, Grandison, Blake, Rampton (-18 to 1) n= 17; normal --8 (- 16 to ) n= 15; (p=1, Kruskal-Wallis) (Spearman's R=+ 55, p=-1) (Fig 3). The myeloperoxidase based assay of neutrophil infiltration in duodenal biopsy speimens was also positively orrelated with the hemiluminesene result (Spearman's R= +-63, p= 4, n= 11) (Fig 4). FFCT OF INHIBITORS (TABL III) Luminol amplified hemiluminesene was onsiderably inhibited by the addition of sodium azide, whih inhibits neutrophil myeloperoxidase. There were smaller but signifiant dereases with atalase (whih breaks down hydrogen peroxide) and DMSO (a hydroxyl radial savenger). SOD (whih onverts superoxide to hydrogen peroxide and oxygen) produed a greater redution of luigenin than luminol amplified hemiluminesene (p<5). FFCT OF H PYLORI AND OTHR RISK FACTORS Antral H pylori infetion was assessed in 89 patients (Table IV). The inreased hemiluminesene in the H pylon positive patients was not simply related to the inreased likelihood of more severe duodenal damage than in those who were H pylori negative. Biopsy samples from areas of similar inflammation produed signifiantly more hemiluminesene if antral H pylon infetion oexisted: onversely, in the absene of duodenal damage antral H pylon status had no effet on the hemiluminesene result. Luminol amplified hemiluminesene was not signifiantly different when the whole study group or disease subgroups were subdivided aording to presene of the following risk fators for duodenal uler disease: smoking, high alohol onsumption (>3 U/week in men, >2 U/week in women) or use of NSAIDs: similarly, use of H2 blokers had no influene on the hemiluminesene result in any group (data not shown). Disussion ROMs have been impliated in the pathogenesis of human disease haraterised by an aute inflammatory ell response, inluding, in gastroenterology, inflammatory bowel disease, panreatitis, and hepatitis.' Surprisingly little is known of their role in duodenal uler disease. We believe our present results provide the first diret evidene in man for disease related prodution of ROMs at the site of duodenal uleration and duodenitis. There is no natural animal model for duodenal uler disease: studies linking ROM prodution and duodenal injury relate mostly to aute and hroni seretagogue indued experimental duodenal uieration, where allopurinol, DMSO, and other ROM savengers have had dose dependent effiay in reduing muosal injury.3 ` In humans, indiret evidene for a role of ROMs in duodenal uler disease has been shown in two studies. A randomised study of 22 patients with healed duodenal uler showed lower relapse rates in those taking allopurinol 5 Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.

5 Muosal reative oxygen metabolite prodution in duodenal uler disease X "' 175 -C 15. = ' Q v) 75 5 ) - CL Spearman's R = + 63 p = 4 by standard histologial methods or by (neutrophil speifi9) myeloperoxidase staining. Neutrophil ativation, with release of ROMs and myeloperoxidase, is the most likely explanation linking their presene and the hemiluminesene response. Chemiluminesene responses of the same tissue were greater with luminol than luigenin in almost all ases. This may be explained by presene of peroxidases in the tissue (for example myeloperoxidase from ativated neutrophils or lipid peroxidases generated through ROM-mediated damage) as they atalyse the interation of ROMs with 25 luminol but not luigenin. 4 The pattern ofresponse to inhibitors, enzymes, o * * * * and savengers speifially affeting ROM -25 *prodution provides the most diret evidene -25 that ROMs were responsible for the hemi- the partiular ROMs present. The most notie- able inhibition of hemiluminesene was found shown as photonslmglminute after subtration of bakground. after addition of sodium azide, whih is an inhibitor of myeloperoxidase, and thus inhibits both prodution ofhypohlorite (HOCl1) and the Myeloperoxidase sore (AU) Figure 4: Luminol amplified hemiluminesene aording to myeloperoxidase ativity. Results are luminesene seen, and gives some indiation of mg four times daily (12%) or DMSO 5 mg four atalysed reation of superoxide and hydrogen times daily (13%) ompared with those taking peroxide with luminol. Although sodium azide is either plaebo (65%) or imetidine 4 mg one a relatively non-speifi agent, inhibiting a wide daily (3%).5 The effets of allopurinol and range of ellular proesses in addition to neutro- DMS wereattributedtotheir abiltyto savenge phil myeloperoxidase, a diret ytotoxi effet is hydroxyl radials, and in the ase of allopurinol, unlkely at the onentrations used.i SOD also to inhibition of xanthine oxidase, an enzyme produed only a small derease in hemi- lumiesene from our biopsy speimens in the entral to ROM prodution during reperfusion after ishaemia. Jankowski et al found redued luminol system, but larger effets with luigenin, levels of the antioxidant glutathione and a pattern repeated when SOD was used to inhibit inreased levels of a marker of ROM mediated luminol and luigenin hemiluminesene from damage (malondialdehyde like material) in 23 stimulated neutrophils.'5 The explanation may patients with ative duodenal ulers. In a separate be that SOD plays a relatively minor part in group. glutathione and malondialdehyde like luminol-amplified hemiluminesene ommaterial levels returned towards normal after pared with other peroxidase atalysed ROM uler healing.2 Indiret evidene for inreased interations, partiularly with hydrogen peroxide ROM prodution has also been found in human whose levels would inrease in the presene of gastri uleration, with redued amounts of the SOD. The moderate inhibitory effets of atalase endogenous antioxidant SOD in biopsy spei- and DMSO suggest that hydrogen peroxide and mens from the edges of ative ulers, and the hydroxyl radial also ontribute to the inreased amounts in biopsy speimens of hemiluminesene response. neaiing 1nn-;,,1- uiers... r 13 In duodenal uler disease, the relationships The hemiluminesene assay we used to demonstrate the presene of muosal ROMs is sensitive but non-speifi. Although any oxidant may at as a substrate for the luminol or luigenin hemiluminesene reations, our results support the onlusion that hemiluminesene was produed by the interation of ROMs with the hemiluminogeni probes. Firstly hemiluminesene orrelates strongly with the degree of inflammatory ell infiltrate, whether assessed TABL IV Influene ofantral Heliobater pylori infetion on hemiluminesene (CL) result, shown as median (photon emission/min/mg - bakground) x 1JO (95% onfidene intervals), for the study group as a whole, andfor the vanrous endosopi gradings H pylori negative H pylori positive Population No CL No CL p* All biopsies (-1-6 to 5) (-2 to 37.6).1 Control 11-4(-14to14) 1-5(-8to1 1) 89 Mild duodenitis (-1 7 to.) 2 1 ( 1 to 14.6).2 Duodenal uler disease (- 1.6 to 3.9) (6O to 812).3 * Compared with H pylori negative ases in same ategory between hemiluminesene and muosal damage, and the effets of inhibitors, resemble those we have reported previously in biopsy speimens from patients with inflammatory bowel disease,2 in the pathogenesis of whih a larger body of evidene impliates neutrophil derived ROMs. The above arguments indiate that the hemiluminesene responses seen in the present study represent duodenal muosal ROM prodution. The relationship between ROM prodution and the severity of muosal injury, taken together with the established pathogeni role of ROMs in experimental models of duodenal uleration and muosal damage, and healing ofhuman duodenal muosal damage with ROM savengers, strongly, if indiretly, supports a pathogeni role for ROMs in duodenal uleration and duodenitis. There are two likely soures of ROMs in the biopsy speimens we studied: (i) ativated neutrophils, whih produe primarily superoxide and hydrogen peroxide via the ell membrane enzyme NADPH oxidase, and hypohlorite via myeloperoxidase,'6 and (ii) xanthine Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.

6 1472 Davies, Simmonds, Stevens, Grandison, Blake, Rampton oxidase, abundant in the upper gastrointestinal muosa, espeially during ishaemia,' whih auses generation of superoxide and hydrogen peroxide during subsequent reperfusion.'6 The two systems are interlinked, as superoxide and other ROM formed during ishaemia/ reperfusion will ause issue damage and inreased levels of neutrophil ativators and hemoattratants. 16 Superoxide derived from either neutrophils or xanthine oxidase, spontaneously, and under the influene of SOD, is onverted to water and hydrogen peroxide. Hydrogen peroxide is a stable and diffusible radial, but relatively harmless; in the presene of neutrophil derived myeloperoxidase, however, it ombines with hloride to form hypohlorous aid, one of the most damaging radials known. While our results do not exlude the presene in our biopsy speimens of ROMs produed seondary to ishaemia and reperfusion, animal work suggests that neutrophils are the most important soure of ytotoxi ROMs in either situation: neutrophil depletion and inhibition of neutrophil adherene effetively prevents intestinal damage after ishaemia and reperfusion. 1 ROM mediated ytotoxiity is thought to be mainly seondary to lipid peroxidation of ell membranes.'8 ROMs, inluding lipid peroxides, may also ause tissue injury by inreasing synthesis of ylo-oxygenase'9 and lipoxygenasell pathway produts of arahidoni aid (results, however, onfliting in different animal models), by stimulating the release of platelet ativating fator2' and ativating the NF-kappaB transription fator, whih auses inreased expression of genes oding for a range of proinflammatory mediators, inluding interleukin 6 and tumour nerosis fator a.22 There was no evidene that smoking, alohol, or use of NSAIDs affeted muosal ROM prodution within any disease group or m ontrols. Thus, while promoting duodenal muosal injury and inflammation, these aetiologial fators do not seem to stimulate neutrophils diretly. Conversely, duodenal muosal ROM prodution was onsiderably inreased in patients positive for H pylon antral infetion ompared with those with similar marosopi disease who were H pylori negative. H pylori have been shown to serete hemotati fators23 and to stimulate ROM prodution from human neutrophils in vitro." Our observations might be explained if H pylon derived hemoattratant and neutrophil stimulating fators had effets distant from the site of infetion, or if they promoted neutrophil ativation after infetion of duodenal etopi gastri muosa. In onlusion, we have shown evidene that duodenal muosal ROM prodution is inreased in duodenal uleration and inflammation, partiularly in the presene of antral H pylon infetion. Our results suggest that antioxidants may ontribute to the treatment and prophylaxis of duodenal uler disease. We are grateful for finanial support from Searle UK (GRD); the Hilden Charitable Fund (NJS), the National Assoiation for Colitis and Crohn's disease (TRJS), and the Arthritis and Rheumatism Counil (DRB). Some of these data have previously been published in abstrat form: Gut 199; 31: Al 183 and Gut 1992; 33 (Suppl 1): S4 and S41. 1 Parks DA, Bulkley GB, Granger DN. Role of oxygen-derived free radials in digestive trat diseases. Surgery 1983; 94: Simmonds NJ, Allen R, Stevens TRJ, vansomeren RNM, Blake DR, Rampton DS. Chemiluminesene assay of muosal reative oxygen metabolites in inflammatory bowel disease. Gastroenterology 1992; 13: Salim AS. Role of.oxygen derived free radials in mehanism of aute and hroni duodenal uleration in the rat. Dig Dis Si 1989; 35: Takamasu M, Fuse Y, Kawamoto K, Ohishi T. Possible mehanisms of diethyldithio-arbamate-indued gastroduodenal muosal damage in rats. Sand J Gastroenterol Suppl 1989; 162: Salim AS. Oxygen-derived free radials and the prevention of duodenal uler relapse A new approah. AmJ Med Si 199; 3: Mooney C, Keenan J, Munster D, Wilson I, Allardye R, Bagshaw P, et al. Neutrophil ativation by Heliobater pylori. Gut 1991; 32: Lanza FL. ndosopi studies of gastri and duodenal injury after the use of ibuprofen, aspirin and other non-steroid antiinflamiatory agents. AmJ Med 1984; 77: Allen RC. Phagoyte oxygenation ativities: quantitative analysis based on hemiluminesene. In: Sholmerih J, Andreeson R, Kapp A, rnst M, Wood WG, eds. Bioluminesene and hemiluminesene. New perspetives. Chihester: John Wiley 1986: Krawisz J, Sharon P, Stenson WF. Quantitative assay for aute intestinal inflammation based on myeloperoxidase ativity. Gastroenterology 1984; 87: Cross C, Halliwell B, Borish T, Pryor WA, Ames BN, Saul RL, et al. Oxygen radials and human disease. Ann Intern Med 1987; 17: Salim AS. Role of oxygen-derived free radials in the mehanism of hroni gastri uleration in the rat: impliations for ytoprotetion. Digestion 1989; 43: Jankowski J, Bridges AB, Sott N, Wormsley KG, Belh J. Cirulating free-radial markers and pepti uler disease. urj Gastroenterol Hepatol 1991; 3: Kishi A, Yoshikawa T, Naito Y, Ando T, Yasuda M, Tsujigiwa M, et al. valuation of superoxide dismutase ativity in the gastri muosa of hroni pepti uler patients. Gastroenterology 199; 98: A Allen RC. Biohemiexitation: hemiluminesene and the study of biologial oxygenation reations. In: Adam W, Cilento G eds. Chemial and biologial generation of exited states. New York: Aademi Press, 1982: Allen RC. Phagoyti leuoyte oxygenation ativities and hemiluminesene: a kineti approah to analysis. Meth nzymol 1986; 133: Grisham MB, Granger DN. Neutrophil-mediated muosal injury. Role of reative oxygen metabolites. Dig Dis Si 1988; 33: 6-SS. 17 Hernandez LA, Grisham MB, Twohig B, Arfors K, Harlan JM, Granger DN. Role of neutrophils in ishaemiareperfusion-indued mirovasular injury. Am J Physiol 1987; 253: H Yoshikawa T, Naito Y, Ueda S, Oyamada H, Takemura T, Yoshida N, et al. Role of oxygen-derived free radials in the pathogenesis of gastri muosal lesions in rats. J Clin Gastroenterol 199; 12 (suppl 1): Chakaborti S, Gurtner GH, Mihael JR. Oxidant-mediated ativation of phospholipase A2 in pulmonary epithelium. AmJ Physiol 1989;-257: L Arakawa T, Sakuma H, Kobayashi K. ffets of ative oxygen speieson prostaglandin and leuotriene synthesis in ultured rat gastri ells. Gastroenterology 1991; 1: A Lewis MS, Whatley R, Cain P, MIntyre TM, Presott SM, Zimmerman GA. Hydrogen peroxide stimulates the synthesis of platelet-ativation fator by endothelium and indues endothelial ell-dependent neutrophil adhesion. J Clin Invest 1988; 82: Shrek R, Rieber P, Baeuerle PA. Reative oxygen intermediates as apparently widely used messengers in the ativation of the NF-kappaB transription fator and HIV-1. MBOJ 1991; 1: Kozol R, MCurdy B, Czanko R. A neutrophil hemotati fator present in Heliobater pylori but absent in H Mustelae. Gastroenterology 1991; 1: AIOI. Gut: first published as /gut on 1 November Downloaded from on 7 November 218 by guest. Proteted by opyright.