anovel antiulcerogenic stable radical gastric ... · anovelantiulcerogenic stable radical prevents...

Gut 1994; 35: 1181-1 188

A novel antiulcerogenic stable radical preventsgastric mucosal lesions in rats

D Rachmilewitz, F Karmeli, E Okon, A Samuni

AbstractThe pathogenesis of gastric mucosalinjury is still poorly understood. Recentreports implicate redox active metalsand reactive oxygen species as mediatorsof gastric damage induced by ethanol ornon-steroidal anti-inflammatory drugs.Attempts were made therefore to preventgastric injury using chelators and theantioxidant enzymes catalase and super-oxide dismutase. These attempts, atbest, would only detoxify extracellularreactive species, such as those producedby activated circulating granulocytes andmacrophages. This study utilises anotherstrategy by pre-emption of both intraand extraceliular reactive species usingradical-radical annihilation reactions andby detoxifying redox active transitionmetals. Nitroxide, stable free radicalswere shown to enter mucous cells, protectagainst the ethanol induced damage, andprevent gastric lesions induced by aspirin,indomethacin, 25% NaCl, or 0.6 N HCI.These findings confirm that gastricmucosal damage from the above agents ismediated by free radicals and, moreover,introduce a prototypical agent withina potential new class of gastric ulcerpreventing drugs.(Gut 1994; 35: 1181-1188)

Departments ofMedicine andPathology, HadassahUniversity Hospital,Mount Scopus,Jerusalem, IsraelD RachmilewitzF KarmeliE Okon

Department ofMolecular Biology,Hebrew University -Haddassah MedicalSchool, Jerusalem,IsraelA Samuni

Correspondence to:Dr A Samuni, Department ofMolecular Biology, HebrewUniversity- HaddassahSchool of Medicine,Jerusalem 91120, Israel.

Accepted for publication21 December 1993

Oxygen derived radicals and non-radicalspecies have been previously implicated inthe pathogenesis of gastric ulceration inducedby ethanol.'-2 Reactive oxygen species havebeen assumed to be mediators of acutegastric damage induced also by non-steroidalanti-inflammatory drugs (NSAIDs) suchas indomethacin and aspirin. 1-6 Previousattempts to attenuate gastric damage,therefore, included pre-treatment with catalaseand superoxide dismutase to detoxifypresumed reactive oxygen species.' 2 S7 Inaddition, chelating agents have been used tosequester redox active transition metals andprevent them from aggravating the damage.Protective effects of catalase and superoxidedismutase, previously seen, are anticipated ifextracellular reactive oxygen species generatedby circulating leucocytes predominantlycontribute to the damage. The use of reagents,however, which do not enter the cell limits thisapproach to extracellular reactive species.

If radicals are instrumental in mediatingbiological damage, their rapid removal shouldprovide protection. Reactions of free radicalwith diamagnetic molecules, however, give riseto secondary radicals, whereas radical-radical

reactions are rapid and often result in stablenon-toxic species. Such a strategy has beenpreviously used with nitroxides to protectcells,8 organs,9 and whole animals1o againstdiverse insults. Nitroxides are cell permeable,comparatively non-toxicl1 stable radicals,which are widely used as biophysical probesfor monitoring membrane state, cellular pH,and oxygen concentration,'2-14 intracellularredox reactions, or as contrast agents fornuclear magnetic resonance imaging (MRI). 16Being capable of undergoing one electronredox reactions, 5-membered ring nitroxidessuch as 2-ethyl-2,5,5-trimethyl-3-oxazolidi-noxyl, have been found to catalytically dismu-tase superoxide radicals:

O-x-_° N-OH

+ 02- + I -4

+ °02 + HF

0-xN-OH

V_:~

+ 02 [1]

+ H202 [2]

yielding H202 and molecular oxygen.Previous studies showed that nitroxidesprevent lipid peroxidation17 and also protectcells by selectively detoxifying other freeradicals. Studies using bacterial and culturedmammalian cells showed that nitroxidesprovide cytoprotection from toxicity inducedby H202,8 18-19 t-butyl hydroperoxide(t-BuOOH), tumour necrosis factor,20mitomycin c,21 and ionising radiation.22

This study was undertaken in the light of theclinical challenge to prevent gastric mucosalinjury and to gain better understanding of thepathogenesis of mucosal damage, this gastricmucosal damage has been induced in rats byethanol, indomethacin, aspirin, 25% NaCl,and 0-6 N HCI while the effect of 4-hydroxy-2,2,6,6-tetramethylpiperidine- 1 -oxyl (TEM-POL) on the induction of gastric mucosalinjury was investigated. The results showeffective protection by TEMPOL againstgastric mucosal injury induced by ethanol andnon-steroidal anti-inflammatory drugs. Thisprotection also provided direct evidence thatfree radicals mediate gastric mucosal damage.

Methods

CHEMICALSDiethylenetriaminopentaacetic acid (DTPA),and 4-hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPOL), were purchased from

1 181

on April 1, 2021 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.35.9.1181 on 1 S

eptember 1994. D

ownloaded from

http://gut.bmj.com/

Rachmilewitz, Karmeli, Okon, Samuni

Aldrich Chemicals. Fatty acid free bovineserum albumin, indomethacin, superoxidedismutase ferricytochrome c were obtainedfrom Sigma Laboratories, Israel; aspirin(Aspegic, Lab Egic, Amily, France); ranitidine(Zantac, Glaxo, UK), leukotriene B4 andleukotriene C4 radioimmunoassay (Amer-sham, England). Mn(III)-desferrioxamine 1:1chelate (Mn-DF) has been preparedby dissolving an excess of MnO2 in DFovernight, followed by centrifugation toremove unreacted MnO2, as previouslydescribed.23-24 Tri(oxalato) chromiate(III)K3[Cr(C204)3] 3H20 (CrOx) was prepared aspreviously described.25

ELECTRON PARAMAGNETIC RESONANCE (EPR)MEASUREMENTS

Liquid samples - a volume of 100 ml wasdrawn into a gas permeable, Teflon capillary of0-8 mm inner diameter. The capillary wasinserted into a quartz tube open at both endsand then placed within the EPR cavity.

Tissue samples - a gastric mucosa sampleof 10-40 mg wet weight was excised fromthe rat stomach immediately after deathand placed in a special groove of a standardquartz made tissue cuvette. During theexperiment, the sample within the spectro-meter cavity was flushed with either air or N2,without disturbing the sample, and theEPR spectra were recorded on a Varian E9X-band spectrometer, operating at 9.45 GHz,100 kHz modulation frequency, 1G modula-tion amplitude, and 10-20 mW microwavepower. The concentrations of the nitroxide inthe samples studied were quantitated using aTEMPOL solution of a known concentrationas a calibration standard. For the determina-tion of the total concentration of nitroxide +reduced nitroxide, ferricyanide was used tooxidise the hydroxylamine.

ENZYME PREPARATIONSCatalase (EC 1.11.1.6) was purchased fromSigma (10 U/,ug) and from BoehringerBiochemicals (65 U/,ug). Superoxide dis-mutase (EC. 1.15.1.1) was obtained fromSigma. The superoxide dismutase activity wasassayed using hypoxanthine/xanthine oxidaseas a generator of superoxide in the presenceof 100 ,uM ferricytochrome c and after itssuperoxide dismutase inhibitable reduction,using a Uvikon dual beam spectrophoto-meter.26 Both sample and reference cuvettescontained 130 U/ml catalase to remove H202.The reference cuvette contained an excess ofsuperoxide dismutase, the reaction was startedby adding 0-01 U/ml XO to both cuvettes. Theinitial rates of superoxide dismutase increase,in the absence (V) and the presence (v) ofvarious amounts of superoxide dismutase werespectrophotometrically followed at 550 nm(E550=21 mlM'Icm'1). The superoxide dis-mutase activity was evaluated from the plots ofV/v dependence on the amount of super-oxide dismutase added. Boiling the enzymewas not enough for the preparation of

inactive-superoxide dismutase. Instead, theenzyme was incubated overnight with 10 mMH202 at pH 10, followed with dialysis againstphosphate buffer pH 7 containing catalase.Subsequently the residual activity was deter-mined as described above and was found to beless then 3%.

ANIMALS AND EXPERIMENT DESIGNMale rats (Sprague-Dawley strain), weighing150-200 g were fasted overnight and permittedfree access to water. To study the effect ofTEMPOL, rats were treated intragastricallywith 1-100 mg of TEMPOIJkg bw. Unlessotherwise stated, the standard protocol includedintragastric treatment with TEMPOL0.1 g/kg bw five minutes before induction ofdamage by intragastric administration of 1 ml96% ethanol. Control animals received equalvolumes of saline before intragastric admini-stration of ethanol. Another control group wastreated intragastrically with only 0.9% NaCl.To assess the damage 10 minutes after ethanoladministration, rats were killed by cervicaldislocation; the stomach was removed andwashed with ice cold Od15 M NaCl, and theextent ofhaemorrhagic erosions in the glandularmucosa was assessed blindly by measuring thearea of the gastric lesions. Sections wereobtained for morphological studies, the mucosawas then scraped, and samples taken fordetermination of leukotrienes.To assess the effect of scavengers, chelating

agents, and proteins on the ethanol inducedinjury, rats were treated intragastrically witheither catalase, bovine serum albumin, super-oxide dismutase, Mn-DF, or 2,2-dipyridyl5-15 minutes before intragastric administra-tion of 1 ml 96% ethanol.To attenuate gastric acid secretion and

maintain neutral pH in the stomach, in certainexperiments, rats were treated intragastricallywith ranitidine (75 mg/kg bw) 30 minutesbefore intragastric administration of TEM-POL followed five minutes later by intragastricadministration of 1 ml 96% ethanol. Controlanimals in this experiment received ranitidineand ethanol.When gastric mucosal damage was induced

by NSAIDs, rats received intragastric 0.1 gnitroxide/kb bw, five minutes before orone hour and two hours after subcutaneousinjections of indomethacin (30 mg/lkb bw)or intragastric administration of aspirin(0. 1 g/kg bw). In these experiments, ratswere killed three hours after NSAID adminis-tration.To assess the nitroxide protection from

gastric damage induced by acid or 25% NaCl,TEMPOL was given five minutes beforeintragastric administration of either 1 ml 0-6N HCI or 25% NaCl. Control rats were treatedwith the irritants only; all rats were killed onehour after damage induction.

DETERMINATION OF LEUKOTRIENESSamples (150 mg) of mucosa were placed inpre-weighed tubes containing 1 ml of 50 mM

182



j.com/


eptember 1994. D

ownloaded from

http://gut.bmj.com/

A novel antiulcerogenic stable radical prevents gastric mucosal lesions in rats

phosphate buffer, pH 7.4 minced with scissors,and centrifuged in an Eppendorf centrifuge for10 seconds. The pellet was resuspended in1 ml of the above buffer, incubated for oneminute in a vortex mixer, 10 ,ug indomethacinwas added, and the tubes were centrifugedfor 60 seconds. The supernatants were keptat -20°C until radioimmunoassay(s) wereperformed. The capability of the mucosa togenerate leukotriene C4 and leukotriene B4 wasexpressed as ng/g wet tissue weight.

MEASUREMENT OF LEUKOTRIENE B4Immunoreactivity of leukotriene B4 was deter-mined by radioimmunoassay kit (Amersham,TRK 940). The assay combines the use of ahigh specific activity leukotriene B4 tracer, anantiserum specific for leukotriene B4 (crossreactivity 100%), and leukotriene standard(range 1.6-200 pg/tube). The specific bindingof tracer is 42.5%, non-specific binding 2.4%.Fifty per cent B/BO displacement is obtainedwith 15 pg/tube and 90% B/BO displacementwith 2.2 pg/tube of leukotriene B4.

buffer were pipetted to assay tubes and 100 jilantibody was added to all tubes except for totaland blank tubes. The tubes were incubated at4°C for 30 minutes. 100 ,ul tritiated leukotrienewere added to all tubes and incubated at 4°Covernight. 200 gl buffer was added, followedby 200 pLI dextran coated charcoal solution,excluding the total tube to which 200 pI bufferwas added. The tubes were vortexed, incu-bated at 4°C for 10 minutes, and centrifuged ina refrigerated Beckman centrifuge at 1600 g for10 minutes. The supernatant was transferredto scintillation vials to which 7 ml scintillationfluid was added and counted in a Kontronliquid scintillation counter.

MORPHOLOGICAL STUDIESSections of the gastric mucosa from fiverepresentative rats of each treatment groupwere fixed in phosphate buffered formalde-hyde, embedded in paraffin wax, and routine5 ,um sections were prepared. Tissues wereroutinely stained with haematoxylin and eosinand blindly evaluated for severity of damage bylight microscopy.

MEASUREMENT OF LEUKOTRIENE C4Immunoreactivity of leukotriene C4 was deter-mined by a radioimmunoassay kit (Amersham,TRK 905). The assay combined the use of ahigh specific activity leukotriene C4 tritiatedtracer with a monoclonal antibody specific forleukotriene C4 and leukotriene C4 standard.The standard curve covers the range 8-500pg/tube and was performed in serial dilution.The assay uses highly specific leukotriene C4antiserum (cross reactivity 100%) and has lowcross reactivity with leukotriene D4 (


Figure 2: Histological sections (haematoxylin and eosin, original magnification Xl 73) ofgastric mucosa of rats treated with 1 ml 96% ethanol only (A); or after pre-treatment withTEMPOL 0.1 g/kg bw (B). (A) shows (on top right side) an erosion that is sharplydemarcatedfrom the surrounding mucosa with considerable congestion, oedema, and acuteinflammatory infiltrate underneath. (B) a section obtainedfrom rats pre-treated withTEMPOL 0. 1 g/kg bw, shows normal gastric mucosa.

be distinguished from that in control rats. Thesurface epithelium of the entire mucosa wasexamined and found to be totally protectedand normal looking (Fig 2B).

TIME DEPENDENCE OF NITROXIDE EFFECTTo examine the duration of the nitroxideeffect, rats were treated with intragastricnitroxide at various periods before induction ofdamage. TEMPOL 0. 1 g kg, when given 30 or60 minutes before administration of ethanol,prevented lesion formation (Fig 3). With alower dose of 20 mg nitroxide/kg bw, given 30minutes and 60 minutes before ethanol, theprotection was partial and lesion areadecreased by 76% and 450/o respectively.When TEMPOL (0. 1 g/kg bw) was given5 minutes after ethanol administration, no

60

'U-40

EE

CD

.2 20

00 5 30 60

Addition of TEMPOL before ethanol (min)Figure 3: Protection by TEMPOL given at various timeintervals before ethanol administration. Rats were treated,intragastrically with TEMPOL 0.1 g/kg bw at varioustime intervals before intragastric administration of 1 ml96% ethanol. Rats were killed after 10 minutes and thearea ofmacroscopically visible gastric lesions was measured.Results are mean (SEM).

protection was seen (61 (9) and 56 (4) mm2/ratrespectively).

PARTITION OF NITROXIDE INTO CELL,STo ascertain that TEMPOL does indeed enterthe mucous cells, the effect of an extracellularparamagnetic line broadening agent on theEPR signal of the nitroxide was examined. Inprinciple, a distinction between the EPRsignals arising from intra and extracellularspecies can be achieved with selective linebroadening agents that cannot readily enter thecells. CrOx, which was previously found to besuperior to ferricyanide or Ni(II),27 was usedin our study. After mixing 40 ,ul of300 mM CrOx with 40 pl of 10 mMTEMPOL, the EPR signal of the nitroxidedisappeared (data not shown). Later, immedi-ately after the addition of a tissue sample ofgastric mucosa to the solution, the EPR signalof TEMPOL has been instantaneouslyrestored. Figure 4 shows the results of a typicalexperiment that compares the EPR signal seenupon scanning a sample containing 38 mg wetweight gastric mucosa soaked with 40 KId of5 mM TEMPOL in phosphate buffered saline,before (upper spectrum) and immediately afterthe addition of 40 ,ul of 300 mM CrOx (lowerspectrum). As Fig 4 shows, the intensity of therestored signal, because of the intracellularTEMPOL species inaccessible to the linebroadening agent, represented only a smallfraction of the total signal, reflectingthe volume fraction occupied by the cells.The results showed that TEMPOL readilyenters the mucous cells, as previously foundfor isolated cells in tissue culture.28 TheTEMPOL distribution between the intra andextracellular compartments is readily evaluatedfrom the respective spectra. The volumefraction occupied by the cells and inaccessiblefor CrOx is less than 15% of the total samplevolume (38 mg of gastric mucosa containingabout 15 mg of mucous cells immersed in a

Gain: 12 500

+ CrOx

20 GAUSSFigure 4: Nitroxide partition into mucous cells. Electronparamagnetic resonance spectra recordedfor tissue sampleofgastric mucosal soaked with 5 mM TEMPOL in 0 1 MpH7, phosphate buffered saline before (upper spectrum,gain =400) and immediately after the addiiton ofCrOx(lower spectrum, gain 5=12 500). Experimental condi-tions: room temperature, aerobic conditions, 9.45 GHz,3360 Gauss centre field, 10 mWincident microwavepower, and IG modulation amplitude.

1184



j.com/


eptember 1994. D

ownloaded from

http://gut.bmj.com/


TABLE I Effect of catalase, superoxide dismutase, bovine serum albumin (BSA),superoxide dismutase-mimic and chelating agents on gastric mucosal lesions induced byethanol

Time before administration Lesion areaAdditive Rats (n) of ethanol (min) mm2/rat

Control ethanol only 48 56 (4)Catalase (Sigma), 50 000 U/lkg bw 5 5 8-4 (2.6)*Catalase (Sigma), 50 000 U/kg bw 10 15 40 (14)Catalase (Boeheringer), 50 000 U/kg bw 5 5 42 (8)BSA, 15 mg/kgbw 5 5 18 (7.6)*BSA, 15 mg/kgbw 5 10 33 (107)BSA, 15 mg/kg bw 5 15 90 (13)Superoxide dismutase 12 000 U/kg bw 10 5 56 (12)Superoxide dismutase 60 000 U/kg bw 10 5 19 (6.3)Superoxide dismutase 60 000 U/kg bw 9 15 24 (78)*Inactive superoxide dismutase

60 000U/kgbw 5 5 37 (7.8)Mn-DF, 0-1 g/lkgbw 8 5 37 (7.8)DF, 0-1 glkgbw 10 5 38 (9.2)2,2dipyridyl, 0-1 g/kgbw 5 5 63 (19)

The additives were given intragastrically before treatment of the rats with 1 ml 96% ethanol.The rats were killed 10 minutes later and mucosal lesions measured. Results are mean (SEM)for rats in each group. *Significantly different from rats treated with ethanol alone.

total volume of 118 ,ul). Taking into accountthe respective 2:1 ratio in signal heights shownin Fig 4 and a 1:31 ratio of spectrometer gainyields a ratio of 1:2 for nitroxide concentrationinside and outside the cells.

PERSISTENCE OF NITROXIDE IN THE MUCOSALCOMPARTMENTTo evaluate the persistence of the nitroxidein the mucosal tissue, rats were treated intra-gastrically with 0.1 glkg bw TEMPOL, eitherfive minutes or 60 minutes before administra-tion of 1 ml saline or ethanol 96% (n=4). Fiveminutes later, the rats were killed 20 mg wetweight samples were excised from the gastricmucosal tissue, blended in 1 ml phosphatebuffered saline, and frozen. To estimate theresidual nitroxide concentration in the tissue,the frozen suspension were thawed, samples of100 pul were drawn into a Teflon capillary,placed in the EPR cavity, and the EPRspectrum of the TEMPOL was recorded (seemethods). Because TEMPOL predominantlydecays within cells by reduction to the hydroxy-

:.!,.,-.''.': ...'S. ; 0. j i. 0 _

0° s_ . :s .'X , ,,f g J ; '..

W .ffi,. .,.3 *.,;e ^> ~

.50C0'{r,

-.. ...^ ... ... ..i : .trifotha cinFigure 5: Effect of nitroxides on NSAID induced gastricmucosal lesion. Rats, fasted overnight, were treatedintragastrically with 041 g TEMPOIJkg bw, five minutesbefore as well as one hour and two hours after subcutaneousadministration of indomethacin (30 mg/kg bw) orintragastric administration of aspirin (0.1 g/kg bw). Therats were killed three hours after NSAID administrationand the area of macroscopically visible gastric lesions wasmeasured.

lamine 4-hydroxy-2,2,6,6-tetramethyl-piperi-dine-loxyl (TEMPOL-H), the total concentra-tion of TEMPOL+TEMPOL-H was alsoevaluated. To oxidise the TEMPOL-H that wasaccumulated 2 ,lI of 0.1 M K3Fe(CN)6 wasadded to each sample and the EPR spectrumwas recorded again. From the intensity of theEPR signal, the total concentrationof TEMPOL+TEMPOL-H was calculated,taking into account the respective dilutions.When TEMPOL was given five minutesbefore treatment with ethanol, the residual(TEMPOL) was 1.3 (0.2) mM while(TEMPOL)+(TEMPOL-H) was 3.65 (1. 1)mM. When TEMPOL was given 60 minutesbefore treatment with ethanol the respectivevalues were: 0.32 (0.24) mM and 1-7 (0.65)mM. This shows that after one hour, the totalconcentration of TEMPOL decreased byhalf, from which, over 80% was reduced toTEMPOL-H. This explains the decreasewith time in the protective effect of nitroxide(see Fig 3).

EFFECT OF PHTo find out if acidic pH was required for theprotective activity of TEMPOL, its effect ongastric mucosal injury in ranitidine treated ratswas studied. Under such conditions, gastricacid secretion was inhibited, the pH remainedneutral, and no free gastric acid wasdetectable. Administration of ethanol, in theabsence of TEMPOL, 30 minutes aftertreatment with ranitidine resulted in lesionareas of 55 (10) mm2/rat (n= 7). Conversely, inranitidine treated rats, a pre-treatment withTEMPOL totally prevented ethanol inducedgastric damage, showing that neither theextent of damage nor the protective effect ofTEMPOL differ from these obtained undernormal gastric pH1 7.

EFFECT OF EXOGENOUS CATALASE,SUPEROXIDE DISMUTASE, AND BOVINESERUM ALBUMINAs the deleterious effect of ethanol is exerted infasted, but not fed, animals, the intragastricadministration of enzymic protein might havesome non-specific effect resembling that ofrat feeding. To distinguish between specificand non-specific protective effects, the effectsof intragastric administration of superoxidedismutase, inactive superoxide dismutase,catalase, and bovine serum albumin wasstudied. In addition, the time intervalbetween the administration of protein andethanol was varied. When given five minutesbefore ethanol, even a foreign protein suchas bovine serum albumin or inactive super-oxide dismutase, diminished the extent ofgastric mucosal lesions. Similarly, theprotection afforded by catalase of Boheringerwas smaller than that exerted by catalase ofSigma, which contained sixfold more protein(Table I). When bovine serum albuminwas given 10-15 minutes before ethanol,however, no protection against ulceration wasevident.

1185



j.com/


eptember 1994. D

ownloaded from

http://gut.bmj.com/


TABLE ii Effect of acid, 25% NaCI, and TEMPOL on gastric mucosal lesions andeicosanoids

Rats Leukotriene C4 Leukotriene B4 LesionsTreatment (n) (ng/g wet weight) (mm2/rat)

0.6NHCI 9 8-2 (0-9) 6-4 (1-0) 106 (18)0.6 N HCI + TEMPOL 10 6-8 (0.6) 3-6 (0.5)* 28 (7)*25%/NaCI 14 6-1 (1-4) 1-0 (0.3) 34 (10)TEMPOL + 25%/NaC1 9 8-9 (1-0) 0-2 (0-03) 0*

TEMPOL (0-1 g/kg/bw) was given intragastrically 5 minutes before intragastric administrationof saline or irritant. Rats were killed 60 minutes later, mucosal lesions were measured andeicosanoids were determined as described in methods. Results are mean (SEM) for rats in eachgroup. *Significantly different from rats treated with 0-6 N HC1 or 25% NaCl alone.

EFFECT OF MN-DF SUPEROXIDEDISMUTASE-MIMICMn-DF, a low molecular weight, cellpermeable, superoxide dismutase-mimic, isreportedly effective in protecting variouscells against superoxide induced damage.23Pre-treatment with Mn-DF before ethanoladministration, however, did not decrease theextent of gastric mucosal injury (Table I) toany statistically significant extent.

PROTECTION BY TEMPOL AGAINST DAMAGEINDUCED BY NSAIDS AND OTHER IRRITANTSThe protection provided by nitroxide was alsoevaluated with respect to damage induced byNSAIDs and other irritants. TEMPOL (threedoses, each of 0O 1 g/kg bw) prevented gastriculceration induced by either indomethacin oraspirin (Fig 5). Similarly, pre-treatment withTEMPOL (0 1 g/kg bw) prevented gastriclesions induced by 25% NaCl and partiallyprotected from 0-6NN HC1 (Table II).

EFFECT ON GASTRIC MUCOSAL MEDIATORSThe generation of mucosal leukotriene B4 andC4, was increased two to threefold 10 minutesafter damage induction by ethanol comparedwith their generation in saline treated rats(Table III). TEMPOL at 0O 1 g/kg bwprevented the ethanol induced increase inleukotriene C4 and B4 generation (p


nitroxides, which act both intra and extra-cellulary are fully protected. The dramaticdifference between the effects of superoxidedismutase and nitroxide clearly shows theimportant part played by intracellular radicals.

In this study, Mn-DF did not show anyprotection. This synthetic superoxide dismu-tase-mimic, which reportedly23 can providecytoprotection by removing Oy is prone todissociate in biological systems and lose itscatalytic activity.24 Therefore, the failure ofMn-DF to prevent gastric mucosal injury doesnot necessarily exclude the role of intracellularO2- toward mucosal damage.A similar conclusion can be drawn from the

failure of DF to provide protection. DF is achelating agent, which avidly binds transitionmetal ions40 particularly iron (III); however,because it enters the cell very slowly8 thecytoprotective effect of DF requires about onehour of preincubation. As Table I shows, DFdid not provide protection against gastricmucosal injury.

MECHANISM OF NITROXIDE ACTIONReactions of free radicals with diamagneticmolecules yield secondary radicals, whichmight propagate injurious process. Conversely,nitroxides, which are free radicals themselves,terminate radical chain reactions. Severalmechanisms might underly the nitroxideseffect on the biological damage, all of whichentail reaction with paramagnetic species.Firstly, five membered ring nitroxides havebeen found capable of catalytically dismutatingO°-. Both the nitroxide and its respectivehydroxylamine can react with Or yielding 02and H202, whereas the concentration ratio ofthe redox couple [nitroxide]/[hydroxylamine]remains time invariant and independent of °2-flux.38 Secondly, more recently, 6 memberedring nitroxide derivatives were found toremove °2- by a different, yet catalytic, mode.Both the nitroxide and its respective oxoam-monium cation, acting as an superoxide dis-mutase mimic, rapidly react with °2-, yieldingH202 and 0241:

Note that generally the concentration ofcellular enzymes is dwarfed by that of theirrespective substrates. The exception for thisrule are cellular superoxide dismutase andnitroxide superoxide dismutase-mimics whoseconcentrations exceeds that of Oy by severalorders of magnitude. Thirdly, both 5 and 6

HOf- N + 2H*

HOfC N-0 + 02

HO- N=O + H202 [6]

HOf1N O + 02 [7]

membered ring nitroxides have been found toprotect cells through reaction with reducedmetal ions. 'Normally', in aerated systems,reduced redox active metal ions react with

H202 yielding a peroxo complex (reaction 4).As cellular transition metals do not exist as'free' ions, but rather coordinated to cellularcomponents, deleterious species, such asauthentic secondary *OH (reaction 5a) andhypervalent metal (reaction 5b), are formed'site specifically' in a close proximity to criticalbiological targets. This site specific formationis responsible for damage potentiation andaccounts for the frequent failures of radicalscavengers to provide protection.42 43 Thereactions of nitroxides with transition metalshave been previously studied and reported.8 19These results suggested that nitroxides cancompete with H202 and oxidise the reducedmetal ions19:

L-M(n-l)++H+ +nitroxide -4L-Mn+ +hydroxylamine

thus inhibiting reaction 4. Finally, nitroxideshave been previously shown to protect hypoxiccells also where reactive oxygen species arepractically absent. The hypoxic cytotoxicity ofeither t-BuOOH8 or mitomycin c21 to Chinesehamster cells could be totally eliminatedby nitroxides. While t-BuOOH inducedtoxicity is catalysed by transition metals,that of mitomycin c is not. The cytoprotectiveeffect of TEMPOL, in the last case, wasattributed to the rapid removal of the deleteri-ous semiquinone radicals of mitomycinc through reaction with nitroxide.21The findings of this study confirmed that in

ethanol treated rats, the generation of gastricmucosal leukotriene B4 and leukotriene C4increases, suggesting that they participatein some fashion in the pathogenesis of thedamage.44 TEMPOL effectively preventedgastric mucosal damage (Figs 1 and 2) andattenuated the increase in leukotriene B4 andleukotriene C4 generation induced by ethanol(Table III). Yet, gastric mucosal generation ofleukotrienes in rats treated with indomethacinor aspirin, did not change in the presence ofnitroxide (Table III). These findings suggestthat leukotrienes and platelet activating factordo not contribute toward gastric mucosalinjury but are formed after damage inductionafter free radical generation. This may partlyexplain why specific receptor antagonists orsynthesis blockers of leukotrienes do notprovide total protection.

In conclusion, our results suggest thatTEMPOL prevents gastric mucosal injuryby breaking some chain reaction through atermination reaction with free radicals. Thus,the strategy of using selective radical-radicalreactions to block biological damage, seemsapplicable also for the prevention of gastricinjury. While the part played by free radicals inthe mediation of gastric mucosal injury is wellestablished and the potential of nitroxidesfor intervention is evidenced, further researchis still needed to select better nitroxides andoptimise their protective effect.

This research was partly supported by grant 89-00124 from theUnited States-Israel Binational Science Foundation (BSF),Jerusalem, Israel. The technical assistance ofMs R Carmeli andMs L Rabi is greatly acknowledged.

1187

[8]



j.com/


eptember 1994. D

ownloaded from

http://gut.bmj.com/

1188 Rachmilewitz, Karmeli, Okon, Samuni

1 Pihan G, Regilo C, Szabo S. Free radicals and lipid peroxi-dation in ethanol-or aspirin-induced gastric mucosalinjury. Dig Dis Sci 1987; 32: 1395-401.

2 Szelenyi I, Brune K. Possible role of oxygen free radicals inethanol-induced gastric mucosal damage in rats. Dig DisSci 1988; 33: 865-71.

3 Terano A, Hiraishi H, Ota S, Shiga J, Sugimoto T. Role ofsuperoxide and hydroxyl radicals in rat gastric mucosalinjury induced by ethanol. Gastroenterol Jpn 1989; 24:488-93.

4 Salim AS. Removing oxygen-derived free radicals stimulateshealing of ethanol-induced erosive gastritis in the rat.Digestion 1990; 47: 24-8.

5 Kvietys PR, Twohig B, Danzell J, Specian RD. Ethanol-induced injury to the rat gastric mucosa. Role of neutro-phils and xanthine oxidase-derived radicals.Gastroenterology 1990; 98: 909-20.

6 Vaananen PM, Meddings JB, Wallace JL. Role of oxygen-derived free radicals in indomethacin-induced gastricinjury. Am Jf Physiol 199 1; 261: G470-5.

7 Takeuchi K, Ushima K, Hironaka Y, Fujioka Y,Matsumoto J, Okabe S. Oxygen free radicals and lipidperoxidation in the pathogenesis of gastric mucosal lesionsinduced by indomethacin in rats. Digestion 1991; 49:175-84.

8 Mitchell JB, Samuni A, Krishna MC, DeGraff WG, AhnMS, Samuni U, Russo A. Biologically active metal-independent superoxide dismutase mimics. Biochemistry1990; 29: 2802-7.

9 Gelvan D, Saltman P, Powell SR. Cardiac reperfusion dam-age prevented by a nitroxide free radical. Proc Natal AcadSci USA. 1991; 88: 4680-4.

10 Hahn SM, Tochner Z, Krishna MC, Glass J, Wilson L,Samuni A, et al. Tempol, a stable free radical, is a novelmurine radiation protector. Cancer Res 1992; 52: 1750-3.

11 Ankel EG, Lai CS, Hopwood LE, Zivkovic Z. Cytotoxicityof commonly used nitroxide radical spin probes. Life Sci1987; 40: 495-8.

12 Froncisz W, Lai CS, Hyde JS. Spin-label oximetry: kineticstudy of cell respiration using a rapid-passage TI-sensitiveelectron spin resonance display. Proc Natl Acad Sci USA1985; 82: 411-5.

13 Belkin S, Mehlhorn RJ, Packer L. Electron spin resonanceoximetry. Methods Enzymol 1988; 167: 670-7.

14 Glockner JF, Chan HC, Swartzz HM. In vivo oximetryusing a nitroxide-liposome system. Magn Reson Med1991; 20: 123-33.

15 Bacic G, Nilges MJ, Magin RL, Walczak T, Swartz HM. Invivo localized ESR spectroscopy reflecting metabolism.Magn Reson Med 1989; 10: 266-72.

16 Keana JF, Pou S, Rosen GM. Nitroxides as potential con-trast enhancing agents for MRI application: influence ofstructure on the rate of reduction by rat hepatocytes,whole liver homogenate, subcellular fractions, and ascor-bate. Magn Reson Med 1987; 5: 525-36.

17 Nilsson US, Olsson LI, Carlin G, Bylund FA. Inhibitionof lipid peroxidation by spin labels. Relationshipsbetween structure and function. Jf Biol Chem 1989; 264:11131-5.

18 Samuni A, Winkelsberg D, Pinson A, Hahn SM, MitchellJB, Russo A. Nitroxide stable radicals protect beatingcardiomyocytes against oxidative damage. J Clin Invest1991; 87: 1526-30.

19 Samuni A, Godinger D, Aronovitch J, Russo A, Mitchell JB.Nitroxides block DNA scission and protect cells fromoxidative damage. Biochemistry 199 1; 30: 555-61.

20 Pogrebniak H, Matthews W, Mitchell J, Russo A, SamuniA, Pass H. Spin trap protection from tumor necrosisfactor cytotoxicity. J Surg Res 199 1; 50: 469-74.

21 Krishna MC, DeGraff W, Tamura S, Gonzalez FJ, SamuniA, Russo A, et al. Mechanisms of hypoxic and aerobiccytotoxicity of mitomycin C in Chinese hamster V79 cells.Cancer Res 1991; 51: 6622-8.

22 Mitchell JB, DeGraffW, Kaufman D, Krishna MC, SamuniA, Finkelstein E, et al. Inhibition of oxygen-dependentradiation-induced damage by the nitroxide superoxidedismutase mimic, tempol. Arch Biochem Biophys 1991;289: 62-70.

23 Rabinowitch HD, Rosen GM, Fridovich I. A mimic ofsuperoxide dismutase activity protects Chlorella soro-kiniana against the toxicity of sulfite. Free Radic Biol Med1989; 6: 45-8.

24 Hahn SM, Krishna CM, Samuni A, Mitchell JB, Russo A.Mn(III)-desferrioxamine superoxide dismutase-mimic:alternative modes of action. Arch Biochem Biophys 1991;288: 215-9.

25 Bailar JC, Young EM. Trioxalato salts. Inorg Synth 1939; 1:35-8.

26 Fridovich I. Quantitative aspects of the production of super-oxide anion radical by milk xanthine oxidase. Jf Biol Chem1970; 245: 4053-7.

27 Lai CS, Froncise W, Hopewood LE. An evaluationof paramagnetic broadening agents for spin probestudies of intact mammalian cells. Biophys J 1987; 52:625-8.

28 Samuni A, Carmichael AJ, Russo A, Mitchell JB, Riesz P.On the spin trapping and ESR detection of oxygen-derived radicals generated inside cells. Proc Natl Acad SciUSA 1986; 83: 7593-7.

29 Grisham MB, Granger DN. Neutrophil-mediated mucosalinjury. Role of reactive oxygen metabolites. Dig Dis Sci1988; 33: 6-15S.

30 Szabo S. Gastroduodenal mucosal injury-acute andchronic: pathways, mediators, and mechanisms. J ClinGastroenterol 1991; 13: S 1-8.

31 Salim AS. Gastric mucosal cytoprotection in the rat by scav-enging oxygen-derived free radicals. Am J Med Sci 1991;302: 287-91.

32 Otamiri T, Sjodahl R. Oxygen radicals: their role in selectedgastrointestinal disorders. Dig Dis 1991; 9: 133-41.

33 Mutoh H, Hiraishi H, Ota S, Ivey KJ, Terano A, SugimotoT. Role of oxygen radicals in ethanol-induced damage tocultured gastric mucosal cells. Am J Physiol 1990; 258:G603-9.

34 Hiraishi H, Terano A, Ota S, Mutoh H, Razandi M,Sugimoto T, et al. Role for iron in reactive oxygen species-mediated cytotoxicity to cultured rat gastric mucosal cells.AmJPhysiol 1991; 260: G556-63.

35 Yoshikawa T, Naito Y, Ueda S, Oyamada H, Takemura T,Yoshida N, et al. Role of oxygen-derived free radicals inthe pathogenesis of gastric mucosal lesions in rats. Jf ClinGastroenterol 1990; 12 (suppl I) S65-71.

36 Simmonds NJ, Allen RE, Stevens TR, Van SR, Blake DR,Rampton DS. Chemiluminescence assay of mucosal reac-tive oxygen metabolites in inflammatory bowel disease.Gastroenterology 1992; 103: 186-96.

37 Nordmann R, Ribiere C, Rouach H. Implication of freeradical mechanisms in ethanol-induced cellular injury.Free Radic Biol Med 1992; 12: 219-40.

38 Samuni A, Krishna CM, Riesz P, Finkelstein E, Russo A. Anovel metal-free low molecular weight superoxidedismutase mimic. _J Biol Chem 1988; 263: 17921-4.

39 Yoshikawa T, Naito Y, Kishi A, Tomii T, Kaneko T,Iinuma S, et al. Role of active oxygen, lipid, peroxidation,and antioxidants in the pathogenesis of gastric mucosalinjury induced by indomethacin in rats. Gut 1993; 34:732-7.

40 Halliwell B. Protection against tissue damage in vivo bydesferrioxamine: what is its mechanism of action? FreeRadic Biol Med 1989; 7: 645-5 1.

41 Krishna MC, Graham DA, Samuni A, Mitchell JB, MRusso A. Oxoammonium cation intermediate in thenitroxide catalyzed dismutation of superoxide. Proc NatlAcad Sci USA 1992; 89: 5537-41.

42 Samuni A, Chevion M, Czapski G. Unusual copper-induced sensitization of the biological damage dueto superoxide radicals. J Biol Chem 1981; 256:12632-5.

43 Samuni A, Aronovitch J, Godinger D, Chevion M, CzapskiG. On the cytotoxicity ofvitamin C and metal ions. A site-specific Fenton mechanism. Eur J Biochem 1983; 137:119-24.

44 Karmeli F, Eliakim R, Okon E, Rachmilewitz D. Gastricmucosal damage by ethanol is mediated by substance Pand prevented by ketotifen, a mast cell stabilizer.GastroenteroloV 1991; 100: 1206-16.



j.com/


eptember 1994. D

ownloaded from
http://gut.bmj.com/

anovel antiulcerogenic stable radical gastric ... · anovelantiulcerogenic stable radical prevents...

Documents