Superoxide Anions Produced by Inflammatory Cells

Gut 1997; 40: 175-181 175 Superoxide anions produced by inflammatory cells

play an important part in the pathogenesis of acid Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from and pepsin induced oesophagitis in rabbits

M J Naya, D Pereboom, J Ortego, J 0 Alda, A Lanas

Abstract ability of the acid extruding membrane ex- Background-Reactive oxygen metab- changers and the activity of the Na+/K+ olites have been associated with gastro- ATPase. Other potential mechanisms of intestinal injury. mucosal damage have not yet been explored. Objective-To investigate whether muco- There is substantial evidence that oxygen sal reactive oxygen metabolites are derived free radicals play an important part in involved in acid and pepsin induced oeso- the pathogenesis of the injury ofvarious tissues phagitis, and if so, which specific including the digestive system.4 5 The involve- metabolites. ment of oxygen derived free radicals, such as Methods-The effects of free radical the superoxide anion, hydrogen peroxide, and scavengers and the anti-inflammatory hydroxyl radical, have been well established in drug ketotifen on rabbit oesophagitis the pathogenesis of ischaemic injury of the induced by acidified pepsin were studied. gastrointestinal mucosa and in other models of Isolated oesophageal cells were obtained mucosal damage induced by non-steroidal before and after oesophageal injury and anti-inflammatory drugs,6 ethanol,7 haemor- the generation of superoxide anion and rhagic shock,8 feeding restriction stress,9 hydrogen peroxide was analysed by flow platelet activating factor'0 and Helicobacter cytometry. The presence of inflammatory pylori. " In all these models the presence of acid celis was determined by indirect immuno- and pepsin increased and potentiated the fluorescence with a mouse antirabbit mucosal damage.'2 13 However, it is not clear CDllb antibody. whether oxygen free radicals are involved in the Results-Of the free radical scavengers pathogenesis of oesophagitis induced by acid tested, superoxide dismutase, which reacts and pepsin. Potential sources of free radicals

with the superoxide anion, significantly include the activated inflammatory cells, the http://gut.bmj.com/ reduced oesophagitis, whereas catalase, hypoxanthine-xanthine oxidase system, the which reacts with hydrogen peroxide, had disrupted mitochondrial electron transport only a mild effect and dimethylsulphoxide system, the metabolism of arachidonate via the had no effect. Ketotifen significantly lipoxygenase pathway, and vascular endothelial reduced the inflammation and also cells.4 From a theoretical point of view, several prevented the induction of oesophagitis. of them could be involved in the pathogenesis

Isolated cells obtained from the oeso- of mucosal damage, including the lipoxygenase on September 30, 2021 by guest. Protected copyright. phageal mucosa after acidified pepsin pathway and the activated inflammatory cells. exposure generated increased amounts of Oesophageal mucosal cells metabolise ara- superoxide anions, which were mainly chidonic acid via both the cyclo-oxygenase and produced by CDllb positive cells. lipoxygenase pathways'4 '5 and the presence of Conclusion-Reactive oxygen metab- inflammatory cell infiltrates in the oesophageal olites, especially superoxide anion, pro- mucosa, lamina propria, and muscularis duced by inflammatory cells play a mucosae, usually accompanies epithelial significant part in the genesis of oeso- erosions and ulceration in reflux oesophagitis.' Gastroenterology phagitis induced by acid and pepsin in The objective of this investigation was, Services M J Naya rabbits and might be a target for future therefore, to examine the role of inflammatory A Lanas medical therapy. cells and the superoxide anion and hydrogen (Gut 1997; 40: 175-181) peroxide generation in the mechanisms of Pathology Services J Ortego oesophageal injury induced by acid and pepsin Keywords: oesophagitis, superoxide anion, free radical, and to determine whether scavenging radicals Department of inflammation, acid, pepsin. and the anti-inflammatory drug ketotifen Physiology, University protects ofZaragoza and against this damage. Finally we sought Hospital Clinico to determine the cell source of the free radical Universitario, Reflux oesophagitis results from contact of the release. Unidad Mixta de oesophageal epithelium with gastric juice, Investigaci6n, Zaragoza, Spain which contains hydrochloric acid and pepsin.' D Pereboom The intrinsic mechanisms of acid and pepsin Methods J 0 Alda induced oesophageal mucosal damage are not This investigation comprised both in vivo and Correspondence to: well known. Recent studies have shown that in vitro studies. Dr Angel Lanas, Servicio de Aparato acid damages oesophageal epithelial cells by Digestivo, Hospital Clinico inducing cell oedema, cell acidification, and Universitario, 50009 Zaragoza, Spain. necrosis2 3 after affecting the lipid bilayers of IN VIVO STUDIES Accepted for publication the apical cell membranes, their intercellular All animal studies were carried out at the 20 September 1996 junction structures, and overwhelming the Service of Biomedicine and Biomaterials of the 176 Naya, Pereboom, Ortego, Alda, Lanas

University of Zaragoza, officially inscribed as a and injected in 50 ml saline. Catalase and "Research Establishment" for the adequate DMSO were obtained from Sigma Chemical husbandry and use of all research animals Co (St Louis, MO, USA) and SOD under the Good laboratory practices norms. New (Orgoteina-Ontosein) from Tedec-Zambaletti Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from Zealand white rabbits weighing 2-5-3-5 kg (Madrid, Spain). The SOD used is a liver each were studied. The animals were bovine metaloprotein which has been widely anaesthetised with intramuscular ketamine used to evaluate the role of superoxide anion KCI (75 mg/kg) and intraperitoneal 20% in different experimental models.7 1218 Keto- urethane (40 mg/kg) before all studies. tifen (Zasten) was obtained form Sandoz- Pharma (Barcelona, Spain). Experimental model The oesophagitis model is a modification of Indicators ofdamage one described by Lillemoe et al, 6 which has Macroscopic and microscopical changes - been described elsewhere.'7 In brief, after Immediatley after removal, the oesophaus was sedation, the rabbits underwent laparotomy examined for gross changes, fixed in 10% and neck dissection. The oesophagus was formalin, and photographed. Histological cannulated in the neck at the pharyngo- examination was performed on sections taken oesophageal junction and in the abdomen at from the proximal, central, and distal portions the gastro-oesophageal junction with plastic of the oesophagus and stained with haema- tubing (internal diameter 2 mm) that was toxylin and eosin. The extent of both macro- secured in place with umbilical ligatures. The scopic and microscopical mucosal damage was oesophagus was then perfused with 50 ml of graded by two independent observers unaware the various damaging solutions at a flow rate of the treatments given. The macroscopic and of 10 ml/min via a recirculatin system using a microscopical oesophagitis indices were de- peristaltic pump (Microtube Pump, MP-3. termined following previously described Eyela, Tokyo Rikakikai, Japan). The tempera- criteria.'7 19 20 Macroscopic: O=normal appear- ture of the perfusate was maintained at 37°C ance; 1=hyperaemia involving >50% of the by a thermoregulator (P-Selecta, Barcelona, tissue; 2=non-confluent mucosal or sub- Spain). To induce oesophageal damage, mucosal haemorrhage; 3=confluent intramural acidified pepsin (normal saline acidified to pH haemorrhages or erosions. Microscopical: 2-0 with 1-0 N HC1+2000 units/ml pepsin; O=normal oesophagus; 1=submucosal oedema Sigma Chemical Co, St Louis, MO, USA) was or separation of epithelial layers or vascular perfused for 80 minutes in each animal congestion alone or in combination; 2focal

(exposed period) via a recirculating system. To areas of intramural haemorrhage or partial http://gut.bmj.com/ evaluate ion flux rates after damage, the epithelial loss or inflammation alone or in exposure period was followed by another 40 combination; 3=large areas of haemorrhage minute period in which acidified saline (normal and/or complete epithelial desquamation. The saline acidified to pH 2 0 with 1'0 N HCI) was severity of infiltration by inflammatory cells perfused (flux period). In each experiment a was assessed subjectively from the following five minute washout period with normal saline scale: O=no infiltration; 1 =mild infiltration;

was performed between the two different 2=pronounced infiltration.2' on September 30, 2021 by guest. Protected copyright. solutions perfused. Assessment of mucosal barnier function - As Aliquots ofthe flux solution were taken from previously described,'7 mucosal permeability the reservoir (50 ml) at the beginning and after was assessed by the calculation of fluxes of H+ the perfusion period for the later analysis of (,ueq/h), K+ (,ueq/h). H+ concentrations were pH, K+, and haemoglobin (see below). After determined by measuring the pH (Crison, the completion of the flux period, each animal micropH 2001, Barcelona, Spain) and K+ con- was killed with an intracardiac bolus of pento- centrations were assayed by a flame photo- barbitone and the oesophagus removed and meter (Eppendorf Genatebau, Netherler Hinz opened longitudinally for pathological exam- GMBH, Hamburg, Germany). Mucosal ination. bleeding was quantified as the total amount of haemoglobin shed into the perfusate in both the exposure and flux periods by a colorimetric Experimental protocol method (Quimica Clinica Aplicada SA Six different groups of six to eight animals each Terragona, Spain). were included in this part of the study. In all of them the same protocol of induction of mucosal damage (described earlier) was IN VITRO STUDIES performed. However, different substances were given 10 minutes before and during the ex- Cell isolation posure period, when an intraluminal acidified Isolated cells from the oesophageal mucosa pepsin solution was perfused. Unless otherwise were obtained according to a procedure pre- specified, all drugs were given intravenously. viously described22 with small modifications. In These substances were 1 mg/kg dimethyl- brief, the oesophagus from overnight fasted sulphoxide (DMSO); 1 mg/kg intraperitoneal New Zealand white rabbits (3-3 5 kg) was dimethylsulphoxide (one bolus); 90 000 units/ removed after a lethal intravenous injection of kg catalase; 6 mg/kg superoxide dismutase 3 ml sodium pentobarbitone (50 mg/kg). The (SOD); 0-05 mg/kg ketotifen; and saline as muscle layer was stripped from the submucosa placebo. All substances were finally dissolved while tissues were bathed in a 4°C Ringer's Superoxide anion and oesophagitis 177

solution previously gassed with 5% CO2 and the other area of greater granularity the cell 95% 02. The standard Ringer solution con- population was composed of smaller cells from tained 82-2 mM NaCl, 4 0 mM KCI, 18 mM the oesophageal basal layers. Moreover, no CaCl2, 0-8 mM MgCl2, 0-8 mM NaSO4, 17-8 mast cells could be identified either in isolated Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from mM NaHC03, 08 mM NaH2PO4, 115 mM cells or in histological specimens (toluidine glucose, and 0-2% bovine serum albumin. The blue stain). medium was equilibrated with 95% 02-5% The presence of monocytes, macrophages, CO2 and pH was adjusted to 7-4. The oeso- and neutrophils in isolated oesophageal phagus was then transferred to a low calcium mucosal cells was determined by indirect Ringer solution (Ca++ low medium contained immunofluorescence. The cells were firstly 01 mM CaCl2 and no MgCl2), containing 1 labelled with a mouse antirabbit CD1 lb mg/ml trypsin (Sigma Chemical Co) and 0 I% (Labgen, Barcelona, Spain) antibody for 45 collagenase (Worthington, New Jersey, NJ, minutes at 4°C and then with a goat antimouse USA; digestion medium). After 30 minutes, Ig Gl-FITC (Southern Biotechnology Asso- the mucosa was cut into small pieces and ciates Inc, Birmingham, AL, USA) for 45 transferred to a new digestion medium in a minutes at 4°C. FITC fluorescence was deter- shaking water bath at 37°C for another 30 mined at 525±+ 15 nm with a band pass filter. minute period. The digestion medium was continuously oxygenated with 95%0 2 and 5% CO2. During this period the tissue was In vitro experimental protocol subjected to repetitive pipetting for several Superoxide anion, hydrogen peroxide, and cell minutes. Finally, the tissue fragments and cells mortality were measured in isolated cells were collected by centrifugation and re- obtained from the oesophageal mucosa of suspended in Ringer's solution allowing the rabbits with and without previous 80 minutes tissue fragments to settle and leaving isolated of oesophageal acidified pepsin perfusion. In mucosal cells suspended in solution. This both types of isolates simultaneous measure- suspension was collected and then filtered ments of superoxide anion and CD 1 lb positive through 100 ,um Teflon mesh. A small aliquot cells were performed. Simultaneous measure- ofcells was removed for a cell count. Cells were ments ofhydrogen peroxide in CD1 lb positive >90% viable as determined by flow cytometry cells could not be undertaken because the using the fluorochromes rhodamine'23 and fluorochromes used to determine the presence propidium iodide to measure live and dead of hydrogen peroxide and CD 1 lb receptors cells respectively.23 have the same wavelength. The cytometric measurements were performed in isolated

mucosal cells as a group and individually in http://gut.bmj.com/ Cytometric determinations each population identified by flow cytometry. The cellular content of superoxide anion and hydrogen peroxide, cell mortality, and population studies were carried out by flow cyto- DATA ANALYSIS metry with an EPICS Elite (Coulter, Hialeah, Results are expressed as mean (SEM). The FL, USA) equipped with an Argon Laser (ILT differences between means were evaluated for

550, Salt Lake City, UT, USA). All deter- significance (p<0 05). After logarithmic Box- on September 30, 2021 by guest. Protected copyright. minations were carried out with 10 000 cells. Cox transformation, continuous data were Intracellular superoxide anion - *°2- was analysed by analysis ofvariance (ANOVA) with measured by intracellular ethidium bromide post hoc comparison where appropriate. Two- formation from hydroethidine as previously group comparisons were conducted with described.24 Briefly, the separated cells were Student's unpaired t test (two tailed). Cate- resuspended at 1 X 1 o6 cell/ml in Hank's gorical data were analysed by Fisher's exact balanced salt solution medium with 20 ,uM test. Cell mortality and CD1 lb labelling hydroethidine and incubated for 20 minutes; studies were analysed by Kruskal Wallis test the transformation of ethidium bromide by and Mann-Whitney U test. *O2 was measured with a DL filter of 625 nm. Intracellular hydrogen peroxide - Hydrogen peroxide was determined as rhodamine'23 Results formation from the oxidation of dihydrorho- damine'23.25 Oesophageal isolated mucosal IN VIVO STUDIES cells were resuspended at 1 X 106 cell/ml in Hank's balanced salt solution and incubated Effects ofradical scavengers on acid andpepsin for 20 minutes with 2 ,uM dihydrorho- induced oesophageal mucosal damage damine123 after which the rhodamine123 Exposure of the oesophageal mucosa to acidi- fluorescence was recorded at 510-540 nm with fied pepsin for 80 minutes induced severe gross a band pass filter. and microscopical mucosal damage (Fig 1). Determination of the cell populations of the Mucosal barrier function was also severely oesophageal mucosa - The oesophageal isolated affected, showing loss of hydrogen ions from mucosal cells were cytometrically studied by the lumen and a positive flux of potassium and size and granularity and two populations were haemoglobin into the oesophageal lumen (Figs clearly separated. The histological analysis of 2 and 3). The parenteral administration of these populations obtained by cytometric either intraperitoneal or intravascular DMSO, sorting showed that in one of them there was a scavenger of hydroxyl radicals, during the a predominance of squamous cells and that in exposure period failed to modify this mucosal 178 Naya, Pereboom, Ortego, Alda, Lanas

Microscopic dramatic improvement of all indicators of M Macroscopic mucosal damage. The effect was similar to that n = 6-8 observed with SOD treatments (Figs 1-3). The ~ * p <0.05 effects of ketotifen on acid and pepsin induced Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from **p < 0.001 oesophageal mucosal damage were associated C with a significant decrease in the presence of inflammatory cells in the mucosa and lamina propia of the oesophagus (Fig 5).

c/io 0 IN VITRO STUDIES Control DMSO Catalase SOD Ketotifen Isolated cells from the oesophageal mucosa Figure 1: Effects ofdifferentfree radical scavengers and ketotifen on the oesophageal mucosal damage induced by obtained after the in vivo rabbit oesophageal acidified pepsin in rabbits (DMSO=dimethylsulphoxide; exposure to acidified pepsin for 80 minutes SOD=superoxide dismutase). induced a significant increase (300%) in generation of superoxide anions compared with control M K+ H+ experiments (isolated cells without previous in n = 6-8 vivo acidified pepsin exposure; Table I). The cell 1100 p0Q05 < populations identified by flow cytometry showed of superoxide **p<0.01 similar increases in the production 850 anions, but the high correlation obtained az 600 | * ** between the superoxide anion content and the immune CD1 lb staining suggest that most xi 350° superoxide anions were produced by the inflammatory cells that were equally distributed in the epithelial populations identified by flow o 1001501 Ii _ cytometry (Fig 6). The percentage of CD1 lb 400 positive cells isolated from the oesophageal 650- mucosa increased after the rabbit oesophageal Control DMS0 Catalase SOD Ketotifen exposure in vivo to acidified pepsin from 2 97 to 6 06 (p<001). Hydrogen Figure 2: Effects of differentifree radical scavengers and (1 08) (0 97)% ketotifen on the function of rabbit oesophageal mucosal peroxide production by isolated cells from this barrier damaged by the intraluminal perfusion of acidified preparation did not differ from production by pepsin in rabbits (DMSO=dimethylsulphoxide; control cells (Table II). SOD=superoxide dismutase). http://gut.bmj.com/

150 Discussion n = 6-8 - The data obtained in this study show that 125- ** p < 0.01 generation of superoxide anions is involved in E 100- the pathogenesis of acid and pepsin induced c .0 oesophageal damage in rabbits. The admin- 0 75 istration of SOD, a scavenger of this radical, on September 30, 2021 by guest. Protected copyright. o 50 significantly reduced the oesophageal damage E a) induced by acidified pepsin. Other free radical 1:m 25 scavengers were either weaker or did not offer Involve- 0 any significant mucosal protection. Control DMSO Catalase SOD Ketotifen ment of oxygen free radicals in mucosal injury Figure 3: Effects ofdifferentfree radical scavengers and has usually been assessed indirectly, but we ketotifen on the oesophageal mucosal bleeding induced by also directly evaluated and confirmed the acidified pepsin in rabbits (DMSO=dimethylsulphoxide; SOD=superoxide disniutase). presence of superoxide anion in cells isolated from the oesophageal mucosa after acid exposure by measuring ethidium bromide, damage. However, catalase, a scavenger of which is intracellularly formed by the oxidation hydrogen peroxide radicals, induced a weak of hydroethidine in the presence of the reduction in mucosal injury induced by superoxide anion. This study also suggests that acidified pepsin. All the indicators of damage the presence of inflammation is associated with were reduced but the differences were only oesophageal damage as the anti-inflammatory statistically significant for macroscopic damage drug ketotifen, which is not a free radical (Figs 1-3). By contrast, treatment with SOD, scavenger, was as effective as SOD in the a scavenger of superoxide anion radicals, prevention of oesophageal damage induced by greatly reduced the mucosal injury induced by acid and pepsin, and that activated anti-inflam- acidified pepsin, and all indicators of damage matory cells (neutrophils, or monocytes, or were significantly improved (Figs 1-4). macrophages) are one of the main sources of this free radical, although other potential, but probably minor, sources were not excluded. Effects of ketotifen on acid and pepsin induced Oxygen free radicals are detrimental to the oesophageal mucosal damage integrity of biological tissues and mediate their The parenteral administration of ketotifen injury. The mechanisms of damage involve during the exposure period of the oesophageal lipid peroxidation, which destroys cell mem- mucosa to acidified pepsin also induced a branes with the release of intracellular Siuperoxi'de anion anid oesopliagiti1sI17i9

ITABLE II Aleas urenment of hvdro-gen pe ox-.ide production by cellflow cvtomietrvin i'solated cells obtained from the ocsophageal m)lucosa before a)ld after thc perfusion of acidified pepsin in vi. o Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from C-ll popudlationl Ititltoutr dantag. 18 itlr daiiagc 'A hB:t.:. - -As, - ,>; -...... ~~.... Total 5.59 A1 74' 6 36 '3 03) Population A 7 51 2 70' 6 95 4 15) Population B 4 25 18 4 (54 1 99)

The generation of hydrogen peroxide is expressed as the mean (SEM) of arbitrars units of fluorescence of rhodamine n in four experiments. In population A there x-as a predominance of squamous cells and in population B a predominance of basal cells.

generation of the superoxide anion as a mechanism of damage is well established in different models of acute and chronic injury in the stomach and intestine,4 13 but it is not well established whether this radical is involved in the pathogenesis of acid and pepsin induced Figuire 4: M1icrophotographs ofoesophageal speciniens in control experinients (A and B) oesophageal damage. In a recent report by anzd ani'mals treated .it/i superoxide dismnziutase (C). Sniperoxide disniutase redutced thzc WVetscher et al,I oesophagitis induced bv Mnicroscopical mnucosal daniiage and inflammatioon inldniced by acidified pepsiml. duodenogastro-oesophageal reflux after duo- (A) Oesophageal epitthelial ulceration and oedemna, vascular congestionI, haemnorrhage, and polvmnorphonuclear cell infiltrationz; (B) Detail of the polvmnorph/onuclcar cell infiltration; denojejunal ligation in rats was prevented or (C) Smnall suipeificial epitlielial loss with oedeniia and z ascuilar congestion. No significantly reduced bv the blockade of the polvmnomphonuclear cell infiltrationm .as Seemi in ttins specinMein. superoxide anion with SOD, but it was not reduced with catalase, dimethvlthiourea, allo- purinol, or vitamin E succinate. Treatment None Mild M Marked with catalase reduced lipid peroxidation by 506% but was not sufficient to inhibit oeso- n = 6-8 *p < 0.05 phagitis on gross appearance. In a different model of oesophagitis, induced by perfusing 100 o< a) A http://gut.bmj.com/ 400 n = 6 co as 50 * p < 0.01 coC:'1 300 as-a)

CZ Y- 200 as0~ o o on September 30, 2021 by guest. Protected copyright. Control DMSO Catalase SOD Ketotifen Q ,) 100 Figure 5: Effects ofdiffcerent frec radical scaven)ges and -E ketotifecn oni tlhe presence of inflammatory cells in the mnnztcosa and lamiina propn7a of t/ic oesophags after the intraluminal 0 pefiusion of acidified pepsin in rabbits (DAISO= CD1lb (+) cells CD1lb ( cells dimethlv sulpihoxide; SOD=supcwxtide dismnutas).

TIABLE I .lcasutreinent ofs'npcoxide anion prodnction by, B 4 cell flow- cvtomnetrltin i'solated cells obtained front tuze oeSopliageal muicosa before and after tiie pctfisnion (if 3 acidified pepsi i1n vico - 02-(CDl 1 b-) C'.ll popultati() lf'oitulrli dawaclfadt'itilr dawiia- 2 ~~~~~~~HE Total 45-48 (25 43' 152 50 (207 3 * HE 02-(CD11 b+) Population A 84 46 (4070) 2'55TO0875)* 1 \ ~~~~~HE-

Population B 4427 1 20 04) 121 13 (4 83' C {1,A HE 02- tota The generation of superoxide anion is expressed as the mean Total (SEMI) of arbitrarx units of fluorescence of ethidium bromide in four expenrments. In population A there wi as a LgFITC predominance of squamous cells and in population B a Ftgniire 6: (A) Snperooxide anlioni production as predominance of basal cells. (meaasmured *p<0 0O. arbittrarr iunti offluorc sicnce osf/tvdroethiditn=HE) bv iSolatcd CDI lb (+) anid CD] lb (-) cells obtaincd fronlt rabbit oesop/lageal mnuiicosa after t/zc in vivo intraluminm al pcr-fnion (f acidified pcpsin. (B) ExamnplIc of onc (7f t/hc components, such as lvsosomal enzymes, c.xpcrimntcts pcfrfcsncd: (1) Total flutorcscentcc obtatnic'd in1 all cell populations aftic t/te addition of t/ic labellecd (FITC) leading to further tissue damage. The radicals CDJ lb antibody. (2) Fluiorescccncc correiponding to also promote mucosal damage bv causing in7tracellutlar superoxide anion (HE) produiccd by the w/tolc degradation of the epithelial basement mem- ceil population. (3) Fluiorcescenicc of ittraccillular sipLc)ox dic brane components, complete alteration of the anlionz produiecd bs' CD/ lb (+) clls. (4) No significant fl uosrescence corresponding to inttraccllular- sipc10ox'Sic/c is cell metabolism, and DNA damage.) The generated tint CD] lb (-) cells. 180 Naya, Pereboom, Ortego, Alda, Lanas

controlled amounts of acid and pepsin for fixed phageal biopsy specimens showed an increased periods, we have shown both directly and production of these compounds in patients with indirectly that superoxide anions were directly oesophagitis that was partially reduced by the in involved in the oesophageal damage induced vitro treatment with the myeloperoxidase Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from by acid and pepsin in rabbits. In our model, inhibitor azide and catalase, suggesting that catalase was much less effective than SOD in neutrophils were not the only source of mucosal preventing or reducing the mucosal damage, luminol chemiluminescence.2" and generation of hydrogen peroxide was not In our model, ketotifen also significantly and detected in mucosal cells isolated after dramatically reduced the oesophageal damage damage, but we cannot exclude that small induced by acidified pepsin. The effect was amounts of this radical might be produced in similar to that obtained with the perfusion of vivo. The involvement of these two radicals the superoxide anion scavenger SOD. Keto- seems plausible because SOD efficiently and tifen is an anti-inflammatory drug, usually specifically scavenge the superoxide radical by known as a "mast cell stabiliser" with other catalysing its dismutation to hydrogen important effects that include the inhibition of peroxide and oxygen.27 Hydrogen peroxide, neutrophilic migration,29 30 the reduction of which is a relatively non-toxic substance can be eosinophil viability,3' and inhibition of cell scavenged by catalase to form water and NADPH oxidase,32 and the inhibition ofleuko- molecular oxygen.27 trienes, platelet activating factor, prostaglandin Hydroxyl radicals are formed by the reaction E2, and tromboxane B2 concentrations in in- of superoxide anions with hydrogen superoxide flamed tissues.3335 Previous studies have also via the Haber-Weiss reaction. This reaction shown that treatment with ketotifen either proceeds at too slow a rate to be of physio- reduced or totally prevented mucosal damage logical relevance. However, this reaction is in different models of experimental colitis,33 accelerated by the presence oftransition metals Clostridium difficile toxin A induced enteritis in such as iron. In these conditions the generation rat ileum and ethanol and gastric mucosal of hydroxyl radicals, which are highly reactive damage induced by non-steroidal anti- and cytotoxic, is particularly harmful. In our inflammatory drugs in rats and humans.29 33 3 model, parenteral DMSO, a scavenger of The presence of increased amounts of neutro- hydroxyl radicals, did not have a significant phils and macrophages in the oesophageal effect on oesophageal damage, suggesting that mucosa after damage suggests the involvement this radical is not involved in the pathogenesis of soluble chemotactic mediators causing mi- of acid and pepsin induced oesophagitis. gration of inflammatory cells into the mucosa. Similar results were reported by Westcher et In other models of damage, mast cells are

al,'5 who did not find any significant effect with involved in both the induction and amplifi- http://gut.bmj.com/ dimethylthiourea alone, another hydroxyl cation of the inflammatory proceSS33-36 by radical scavenger, on rat oesophagitis. releasing soluble mediators. Therefore, the However, the hydroxyl radical is extremely effects of ketotifen on this model of acid and reactive and it is unlikely that DMSO in vivo pepsin induced oesophagitis could be due to an experiments will reach the site of generation of effect on resident oesophageal mast cells. How- this radical in sufficient concentration to pre- ever, mast cells, which are present in the adult

vent hydroxyl radical induced local damage and North American opossum and humans, are on September 30, 2021 by guest. Protected copyright. therefore, despite our findings and Westcher's very rare in rabbit oesophageal mucosa,37 38 data, a role for hydroxyl radicals in the patho- suggesting that most effects of ketotifen in this genesis of acid and pepsin induced oesophagitis model of oesophagitis might be mast cell inde- cannot be definitely excluded. pendent and probably related to other drug An important question related to this investi- actions (for example, inhibition ofneutrophilic gation concerns the source of the oxygen migration). Our histological studies have radicals produced during oesophageal injury shown a significant reduction in the presence induced by acid and pepsin. Our in vitro ofinflammatory cells only in those experiments studies suggest that the inflammatory infiltrate in which rabbits were treated with ketotifen. is the main source of generation of superoxide The results of this study demonstrate that anions. As measured by flow cytometry, iso- oxygen derived free radicals, especially the lated oesophageal cells obtained after the per- superoxide anion, are involved in the patho- fusion of acidified pepsin showed increased genesis of acid and pepsin induced oeso- generation of superoxide anions. Pretreatment phagitis in rabbits, and that the inflammatory with a monoclonal antibody directed against infiltrate - namely, neutrophils, macrophages, subunits of the granulocyte adherence glyco- and monocytes - are the main source of this protein CD 1 lb showed that most of the super- radical and the drugs that modulate this in- oxide anion was generated in CDl lb positive flammatory response or radical scavengers cells, which include neutrophils, monocytes, dramatically reduce the acid and pepsin and macrophages in rabbits. Eosinophils are induced injury. Because oesophagitis in not labelled by the CD1 lb antibody in rabbits humans is often characterised by epithelial and, therefore, we cannot exclude the possi- erosion, ulceration, and accompanying inflam- bility that this type of cell, usually present as matory infiltrate,' these results may be useful a small percentage of the inflammatory infil- in the devleopment of future treatment of trate, may contribute to the generation of super- reflux oesophagitis. oxide anions. In a recent study, the indirect We thank Dr Tomas Casuella for his help in the cytological measurement of reactive oxygen species by analysis and Clara Tapia for technical assistance. This study was luminol chemiluminescence from human oeso- supported by grants 92/0670 and 91/0613 from the Fondo de Superoxide anion andoesophagitis 181

Investigaciones Sanitarias de la Seguridad Social (FIS). Part of prevent reflux esophagitis in rats. Dig Dis Sci 1995; 40: this work was presented at the 1995 meeting of the American 1292-6. Gastroenterological Association. 19 Schweitzer EJ, Bass BJ, Johnson LF, HarmonJW. Sucralfate prevents experimental peptic esophagitis in rabbits.

Gastroenterology 1985; 88: 611-9. Gut: first published as 10.1136/gut.40.2.175 on 1 February 1997. Downloaded from 20 Tay HP, Chaparala RC, Harmon JW, HueskenJ, Saini N, 1 Kahrilas PJ, Hogan WJ. Gastroesophageal reflux diseases. Hakki FZ, et al. Bismuth subsalicylate reduces peptic In: Sleisenger MH, Fordtran JS, eds. Gastrointestinal injury of the oesophagus in rabbits. Gut 1990; 31: 11-6. diseases. London:WB Saunders, 1993: 378-401. 21 Trevethick MA, Clayton NM, Strong P, HarmanIW. Do 2 Orlando RC. Esophageal epithelial defences against acid infiltrating neutrophils contribute to the pathogenesis of injury. AmJ' Gastroenterol 1994; 89: 48-52. indomethacin induced ulceration of the rat gastric 3 KhalbussWE, Marousis CG, SubramanyamM, Orlando RC. antrum? Gut 1993; 34: 156-60. Effect of HCI on transmembrane potentials and intra- 22 Layden TJ, Agnone LM, Schmidt LN, Hakim B, cellular pH in rabbit esophageal epithelium. Gastro- Goldstein JL. Rabbit esophageal cells possess an Na+, H+ enterology 1995; 108: 662-72. antiport. Gastroenterology 1990; 99: 909-17. 4 Grisham MB, Hernandez LA, Granger DN. Xanthine oxi- 23 Lanas A, Nerin J, Esteva F, Sainz R. Nonsteroidal anti- dase and neutrophil in filtration in intestinal ischemia. Am inflammatory drugs and prostaglandin effects on pep- JPhysiol 1986; 251: G567-74. sinogen secretion by dispersed human peptic cells. Gut 5 Tarnasky PR,Livingston EH, Jacobs KM, Zimmerman BJ, 1995; 36: 657-63. Guth PH, Garrick TR. Role of oxyradicals in cold water 24 Rothe G, Valet G. Flow cytometric analysis of respiratory immersion restraint-induced gastric mucosal injury in the burst activity in phagocytes with hydroethidine and 2',7' rat.DigDis Sci 1990; 35: 173-7. dichlorofluorescein. JLeukoc Biol 1990; 47: 440-8. 6 Wallace JL, Arfors KE, Mcknight GW. A monoclonal 25 Rothe G, Oser A, Valet G. Didydrorhodamine 123: a new antibody against the CD18 leukocyte adhesion molecule flow cytometric indicator for respiratory burst activity in prevents indomethacin-induced gastric damage in the neutrophil granulocytes. Naturwissenchaften 1988; 75: rabbit. Gastroenterology 1991; 100: 878-83. 354-5. 7 Terano A, Hiraishi H, Ota S, Shiga J, Sugimoto T. Role of 26 Halliwell B. Reactive oxygen species in living systems: superoxide and hydroxyl radicals in rat gastric mucosal source, biochemistry and role in human disease. Am J7 injury induced by ethanol. Gastroenterol Jrpn 1989; 24: Med 1991; 91: 14-22. 488-93. 27 Mutoh H, OtaS, Hiraishi H, Ivey KJ, Terano A, Sugimoto T. 8 Yasue N, Guth PH. Role of exogenous acid and re- Reduced glutathione protects cultured gastric mucosal transfusion in hemorrhagic shock induced gastric lesions cells from suckling rats against acid. Am J Physiol 1991; in the rat. Gastroenterology 1988; 94: 1135-43. 261: 65-70. 9 Nakamura K, Aoike A, Rokutan K, Hosokawa T, Koyama K, 28 Olayaee M, SontagS, Salman W, Schnell T, Mobarhan S, Kawai K. The role of oxygen radicals in the pathogenesis Eiznhamer D, et al. Mucosal reactive oxygen species of gastric mucosal lesions induced in mice by feeding- production in oesophagitis and Barrett's oesophagus. Gut restriction stress. Scand J Gastroenterol 1989; 24 (suppl 1995; 37: 168-73. 162): 47-50. 29 Pathoulakis C, Karmeli F, Kelly CP, Eliakim R, Joshi MA, 10 Binnaka T, Yamaguchi T, Hirohara J, Hiramatsu A, O'Keane CJ, et al. Ketotifen inhibits Clostridium difficile Mizuno T, Sameshima Y. Gastric mucosal damage toxin A induced enteritis in rat ileum. Gastroenterology induced in rats by intravenous administration of platelet- 1993; 105: 701-7. activating factor. Scand Jf Gastroenterol 1989; 24 (suppl 30 Shiratori Y, Takada H, Hai K, Kiriyama H, Mawet E, 162): 67-70. Komatsu Y, et al. Effect of antiallergic agents on chemo- 11 Perminder SP, Colin JG, Meron RJ. A radical view of the taxis of neutrophils by stimulation of chemotactic factor stomach: the role of oxygen-derived free radicals and anti- released from hepatocytes exposed to ethanol. DigDis Sci oxidants in gastroduodenal disease. Eur J Gastroenterol 1994; 39: 1569-75. Hepatol 1995; 7: 265-71. 31 Hossain M, Okubo Y, Sekiguchi M. Effects of various drugs 12 Stein HJ, Espluges J, Whittle BJ, Bauerfeind P, Hinder RA, (staurosporine, herbimycin A, Ketotifen, theophylline, Blum AL. Direct cytotoxic effect of oxygen radicals on the FK506 and cyclosporin A) on eosinophil viability. Arerugi gastric mucosa. Surgery 1989; 106: 318-24. 1994; 43: 711-7. 13 Matsumoto T, Moriguchi R, Yamada H. Role of poly- 32 Hojo M, Hamasaki Y, Fujita I, Koga H, Matsumoto S, Miyazaki S. Effects of antiallergy drugs on morphonuclear leucocytes and oxygen-derived free fMet-Leu-Phe- http://gut.bmj.com/ radicals in the formation of gastric lesions induced by stimulated superoxide generation in human neutrophils. HCl/ethanol, and a possible mechanism of protection by Ann Allergy 1994; 73: 21-6. anti-ulcer polysaccharide. J Pharm Pharmacol 1993; 45: 33 Eliakim R, Karmeli F, Okon E, Rachmilewitz D. Ketotifen 535-9. effectively prevents mucosal damage in experimental 14 Stein BE, Schwartzman ML, Carroll MA, Stahl RE, colitis. Gut 1992; 33: 1498-503. Rosenthal WS. Role of arachidonic acid metabolites in 34 Eliakim R, Karmeli F, Chorev M, Okon E, Rachmilewitz D. acid-pepsin injury to rabbit esophagus. Gastroenterology Effect of drugs on colonic eicosanoid accumulation in 1989; 97: 278-83. active ulcerative colitis. Scand J Gastroenterol 1992; 27: 15 Jimenez P, Esteva F, Piazuelo E, Garcia MA, Lanas A. 968-72. Prostaglandins and arachidonic acid metabolites pro- 35 Eliakim R, Karmeli F, Rachmilewitz D. Ketotifen old drug, duced by esophageal mucosal cells: implications for new indication: reduction of gastric mucosal injury. Scand on September 30, 2021 by guest. Protected copyright. mucosal defense. Eur Surg Res 1994; 26: 35-6. Jf Gastroenterol 1993; 28: 202-4. 16 Lillemoe KD, Johnson LF, Harmon JW. Alkaline eso- 36 Kubes P, Kanwar S, Niu XF, Gaboury JP. Nitric oxide phagitis: a comparison of the ability of components of synthesis inhibition induces leukocyte adhesion via gastroduodenal contents to injure the rabbit esophagus. superoxide and mast cells. FASEBJ7 1993; 7: 1293-9. Gastroenterology 1983; 85: 621-8. 37 Barclay RL, Dinda PK, Morris GP, Paterson WG. Morpho- 17 Lanas A, Sousa FL, Ortego J, Soria J, Esteva F, Blas JM, logical evidence of mast cell degranulation in an animal et al. Aspirin renders the esophageal mucosa more per- model of acid-induced esophageal mucosal injury. Dig Dis meable to acid and pepsin in rabbits by different mechan- Sci 1995; 40: 1651-8. isms. EurJ3 Gastroenterol Hepatol 1995; 7: 1065-72. 38 Norris HT, Zamcheck N, Gottlieb LS. The presence and 18 Wetscher GJ, Hunder PR, Bagchi D, Perdikis G, distribution of mast cells in the human gastrointestinal Redmond EJ, Glaser K, et aL Free radical scavengers tract at autopsy. Gastroenterology 1963; 44: 448-55.