INTRODUCTION

Nonsteroidal anti-inflammatory drugs (NSAIDs) such asaspirin and indomethacin are the most commonly prescribeddrugs for arthritis, inflammation, and cardiovascular protection(1). However, they cause gastrointestinal complications such asulcers and erosions (1). The pathophysiology of thesecomplications has mostly been ascribed to NSAID's action oncyclooxygenase inhibition and subsequent prostaglandin (PG)deficiency (2). In the 1970s and 1980s, under the concept ofcytoprotection, extensive research revealed the role of PG inthe gastric mucosal defense system (3). According to thesestudies, an insufficiency of gastric mucosal mucin productioncaused by the inhibition of PG synthesis by NSAIDs is mostlikely the reason for the formation of these lesions. Becausemucin is a very important protective factor, insufficient mucinproduction would be expected to cause an imbalance betweenaggressive factors (such as acid) and protective factors. In fact,some researchers have also shown that inhibition of acidsecretion with histamine-H2 receptor antagonists or protonpump inhibitors significantly inhibited the formation of lesions(4-6).

Recent clinical studies have shed some light on NSAID-induced small intestinal mucosal injury. Capsule endoscopic anddouble-balloon endoscopic examinations have revealed thatNSAID-induced mucosal damage in the small intestine,including erosion and ulceration, occurs more often than thatpreviously expected (7, 8). There are no acid-secreting cells inthe small intestine; therefore, the balance theory cannot apply inthis case. A pathophysiological concept that would also explainNSAID-induced injury in the small intestine is desired, becausethis lesion has been reported to be fatal (9). However, thepathophysiology of NSAID-induced small intestinal injury is notas well understood as that of gastric injury.

The authors have previously reported that NSAID-derivedmitochondrial reactive oxygen species (ROS) are directlyinvolved in gastrointestinal (GI) cellular injury in vitro and thattreatment with rebamipide, which is clinically prescribed forgastric injury patients, prevented this cellular injury byupregulating the expression of manganese superoxide dismutase(MnSOD) (10). To investigate the mechanisms underlyingNSAID-induced intestinal injury, in this study, we performedlive imaging of intestinal mucosa with a microscope duringintestinal injury formation.

JOURNAL OF PHYSIOLOGY AND PHARMACOLOGY 2014, 65, 3, 435-440www.jpp.krakow.pl

T. TOMITA1,3, H. SADAKATA1,3, M. TAMURA2, H. MATSUI2

INDOMETHACIN-INDUCED GENERATION OF REACTIVE OXYGEN SPECIES LEADS TO EPITHELIAL CELL INJURY BEFORE THE FORMATION

OF INTESTINAL LESIONS IN MICE

1Graduate School of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Japan; 2Division of Clinical Medicine, Faculty of Medical Sciences, University of Tsukuba, Tsukuba, Japan; 3TimeLapse Vision Inc., Asaka, Japan

Recently, with the increasing number of elderly patients who continuously take aspirin and/or nonsteroidal anti-inflammatory drugs (NSAIDs), the number of cases of severe hemorrhagic gastrointestinal (GI) bleeding is also on theincrease. Gastric acid has been reported to play the most important role in hemorrhagic gastric mucosal injury. However,the pathogenesis of NSAID-derived mucosal injury in the intestine, where there is no acidic environment, remainsunknown. We previously reported that NSAID-derived mitochondrial reactive oxygen species (ROS) are directly involvedin GI cellular injury in vitro, although an in vivo study has not yet been carried out. In this study, we investigated therelationship between NSAID-derived ROS and intestinal injury formation. For this purpose, intestinal mucosal liveimaging in mice was carried out using an ROS-indicating fluorescent probe. Treatment with indomethacin causedmacroscopic intestinal injury in mice; however, many dying cells were observed even in areas that macroscopicallyappeared to have no injury after treatment with indomethacin. A fluorescent probe revealed that mucosal cells in theapparently uninjured areas had a high concentration of ROS. Treatment with rebamipide significantly decreased both theROS concentration and the number of dying cells: this drug is prescribed clinically for gastric injury patients and has beenreported to upregulate the expression of manganese superoxide dismutase. On the basis of these results, we propose thatNSAID treatment causes a high cellular concentration of ROS in mucosae, possibly decreasing mucosal organo-protectiveefficacy. Moreover, intestinal food contents are likely to damage the mucosal structure when it is in such a fragile condition.

K e y w o r d s : nonsteroidal anti-inflammatory drugs, intestinal injury, reactive oxygen species, manganese superoxide dismutase,intestinal villi, prostaglandins, rebamipide

MATERIALS AND METHODS

Reagents

Indomethacin was purchased from Wako Pure ChemicalIndustries, Ltd. (Osaka, Japan); aminophenyl fluorescein (APF),

from Sekisui Medical Co. Ltd. (Japan); propidium iodide (PI),from Sigma-Aldrich; and the Guava ViaCount reagent, fromEMD Millipore Co. (MA, USA). Rebamipide was kindlyprovided by Otsuka Pharmaceutical Co., Ltd. (Tokyo, Japan);the molecular structure of rebamipide is shown in Fig. 1.

Ethics

For the animal experiments, all procedures and animal carewere approved by the Committee on the Ethics of AnimalExperimentation of the University of Tsukuba and wereconducted according to the Guidelines for AnimalExperimentation of the University of Tsukuba.

Live imaging of indomethacin-induced small intestinal injury inmice

Seven-week-old male ICR mice were purchased from SLCCo., Ltd. (Hamamatsu, Japan). Indomethacin or rebamipide wasresuspended in 0.5% carboxyl methyl cellulose. Indomethacin(20 mg/kg) was administered p.o. to mice. In the rebamipidepretreatment group, rebamipide (30 mg/kg) was administeredp.o. 10 min before the administration of indomethacin. Thetreated mice were kept for 12 hours without fasting and thensubjected to live imaging. For live imaging, mice wereanesthetized with urethane (1.75 g/kg, i.s.), and the jejunum waspicked out through a small incision of the peritoneum. Themucosa of jejunum was visualized by making an incisionapproximately 2 cm long along the long axis for microscopicobservation. Twenty microliters of the fluorescent ROS indicatoraminophenyl fluorescein (APF) (11) (50 µmol/L) was poured onthe mucosa after wiping away the intraluminal contents with apainting brush to remove detached cells. Fifteen minutes afterthe addition of APF, the formation of ROS on the mucosa wasobserved by fluorescence microscopy (Optiphoto Nikon, Tokyo,Japan). Then, 90 min after adding APF, propidium iodide (PI)(10 µL; 5 µg/mL) was applied to the mucosal surface tovisualize dead cells. The fluorescence intensity of both APF andPI images was measured at randomly chosen observation fieldsby using NIH ImageJ software, and averages and standard errorwere calculated.

Flow cytometric analysis of detaching cells

Following live imaging of indomethacin-induced smallintestinal injury as described above, cells that were detached orin the process of detaching from the mucosa were obtained usinga No. 2 round-type painting brush (Sakura Color Products Co.,Japan). The mucosa was brushed ten times in the same direction,and then the brush was washed in 1 mL of saline to collect thedetached cells. This procedure was performed three times. The

436

���������������� �������������������������������� ���������������� ������������� ����

����� ��!�"

Fig. 1. Chemical structure of rebamipide.

�

�

�

Fig. 2. Images of small intestinal mucosa 12 hours afterindomethacin administration. (a) The ulcer in the lower intestinecommonly developed in association with intestinal contents.Arrows indicate the perforations confirmed outside the intestinaltract. (b) Perforations were observed in the intestinal innercavity. (c) Image of staining with Evans blue.

number of the detached cells was counted using a flowcytometer after staining them with the Guava ViaCount Reagent.Flow cytometric analysis was carried out using the GuavaEasyCyte Mini system (Guava Technologies, Hayward, CA).

Detection of intestinal injury by Evans blue staining

Small intestinal injury was induced by indomethacintreatment without prior rebamipide administration. Intestinalinjury detection was performed as previously described (12, 13).Briefly, 0.5 mL of 1% Evans blue was bolus-injected into the tailvain. After 15 min, the mouse was killed with an pentobarbitaloverdose. The small intestine of the mouse was removed from thebody and then incised longitudinally along the antimesentericborder. Lesions in the small intestinal mucosa as revealed byleakage of Evans blue were observed macroscopically.

Statistical analysis

The statistical significance of the data was evaluated usinganalysis of variance (ANOVA) followed by Tukey method. A P-value of <0.05 was considered significant.

RESULTS

Indomethacin-induced small intestinal injury in the mouse

At 12 huors after oral administration of indomethacin, micehad serious ulcers in the ileum (the lower part of the intestine;Fig. 2a), and the ulcers could be visualized with Evans blue (Fig.2b). However, Evans blue was not detected in the jejunum (theupper part of the small intestine), which is usually empty. Theseobservations suggest that the ulcers were caused by mechanicaldamage from the contents in the intestine because the area of the

ulcer was almost the same size as the area that had been coveredby intestinal contents.

Live imaging of the jejunum

Injury prior to ulcer development was detected in thejejunum. Twelve hours after oral administration of indomethacin,mucosal damage in the jejunum could not be seen in bright-fieldimages; however, fluorescence from APF, an ROS indicator,could be observed in the mucosal epithelial cells (representativeimages are shown in Fig. 3a, 3d, 3g). APF fluorescence in thecontrol mice, which had been administered the vehicle instead ofindomethacin, was almost undetectable and was weaker inrebamipide-pretreated mice than in indomethacin-treated mice(representative images are shown in Fig. 3b, 3e, 3h).

Ninety minutes after removing intestinal contents such asdigested foods, we treated villi with 10 µL of PI (5 µg/mL).Because the compromised cellular membranes of dead cellsallow the membrane-impermeant PI dye to enter and stainnuclei, dead cells can be easily detected as PI-positive cells (14).In the intestines of indomethacin-treated mice, many PI-positivecells could be seen: they covered almost all the intestinalmucosa. In contrast, only a few PI-positive cells could bedetected on the summit of each villus in the control mice.Moreover, there were fewer PI-positive cells in the intestines ofrebamipide-treated mice than in those of indomethacin-treatedmice. Fluorescence intensity was the highest in indomethacin-treated mice, followed by that in rebamipide and indomethacin-treated mice, and that in the control mice (representative imagesare shown in Fig. 3c, 3f, 3i).

In additional, when cells that had detached from the mucosawere analyzed by flow cytometry, the same trend in the intensityof APF fluorescence was observed (Fig. 5). These resultsindicate that indomethacin induces cellular injury in the smallintestinal mucosa through oxidative stress.

437

������� ������� ��������� �� ��������

�

�

�

�

�

&

$

�

�

�''�(�

Fig. 3. Representative images oflive imaging of mouse smallintestine (jejunum). (a), (d), (g):Bright-field images 12 hoursafter administration p.o. ofindomethacin or vehicle. (b), (e),(h): Fluorescence images afterAPF application on the mucosafor 15 min. ROS producing cellswere observed. (c), (f), (i): PIfluorescence images taken 90min after the initial mucosalclean up. Fluorescent dots aredead cells. Scale bar: 100 µm;magnification of each image issame.

DISCUSSION

In this study, we found for the first time that indomethacintreatment causes the death of intestinal cells by inducing a highconcentration of ROS prior to the formation of lesions.

Intestinal cells are produced at the area close to the crypt(located at the base of the villus). The cells age as they move tothe summit of the villus, where they are sloughed off after theirdeath. In our study, representative of this cell-kineticphenomenon, dead cells, which were stained with PI, could beobserved at the summit of each villus in control mice. Incontrast, a larger number of dead cells were observed in theintestines from indomethacin-treated mice. An obvious increasein dead cells was also observed in areas that macroscopicallyappeared to be free of lesions. Therefore, we conclude that theformation of macroscopic lesions such as ulcers and erosions

are involved physical force such as that caused by themovement of foods.

ROS are known to be involved in GI mucosal injury.Kwiecien et al. reported that peroxidation of gastric mucosaltissue is involved in gastric lesion formations in rats with water-immersion restraint stress (WRS), a model of gastric mucosalinjury due to stress (15). In their model, tissue SOD and GSHproduction are reduced by stress. Moreover, the calcitonin gene-related peptide (CGRP)-sensitive sensory nerves are involved inthe formation of ROS scavengers. The two known primarysources of cellular ROS generation are NADPH oxidase (NOX)in cell membranes and the mitochondrial respiratory chain.Proinflammatory cytokines induce ROS formation via activationof NOX, which triggers prostaglandin production (16).Meanwhile, ROS leakage to the cytoplasm owing tomitochondrial dysfunction triggers cellular events such as

438

Fig. 4. The amount of ROS production and number ofdead epithelial cells in areas of the small intestinemucosa (jejunum) that macroscopically appeared to beuninjured. Fluorescence intensities were measuredfrom images taken with the same settings as Fig. 3. (A)ROS production calculated from the APF fluorescenceimages. (B) The number of dead cells calculated fromPI fluorescence images. n=6, mean ± S.E. ** p<0.05,ANOVA followed by the Tukey method.

apoptosis. We have previously reported that cellular stress byexposure to hypertonic salt directly generates ROS from themitochondria of gastric mucosal cells (17).

ROS from mitochondria are known to play an importantrole in NSAIDs-induced GI injury (18, 19). The mitochondrialrespiratory chain is the major source of ROS, which aremainly generated at Complex I and III of the respiratory chain(20). The mitochondrial respiratory chain is also an importanttarget for the damaging effects of ROS. ROS frommitochondria play an important role in the release ofcytochrome c and other pro-apoptotic proteins, which cantrigger caspase activation and apoptosis (21). NSAIDs inhibitoxidative phosphorylation to involve both ROS productionand the liberation of cytochrome c from mitochondrialintermembranous space into cytosol, phenomena which causecaspase 9 and caspase 3 activation and cellular lipidperoxidation, leading to apoptosis (22-25). The uncoupling ofmitochondria also decreases the concentration of intracellularATP and causes Ca2+ to leak out of the mitochondria, leadingto cellular osmotic imbalance and a loss of control overintracellular junctions and resulting in increased permeabilityand subsequent mucosal damage (26).

In the intestines of indomethacin-treated mice, distinctlyfluorescent villi were observed after application of APF. Incontrast, little fluorescence was observed in the villi of controlmice. Moreover, the intestinal villi of mice that were pretreatedwith rebamipide before the indomethacin treatment were lessfluorescent, and a smaller number of dead cells were observed inthis group than in the indomethacin-treated mice. We havepreviously reported that rebamipide treatment upregulates theexpression of MnSOD (10). MnSOD has been reported toscavenge mitochondria-specific O2

-, thus reducing ROSinvolved in tissue injury, and it has also been reported that cellsoverexpressing MnSOD are resistant to oxidative stresses (27-31). Therefore, we propose that the APF fluorescence observedin our mouse model is likely to represent the presence ofmitochondrial ROS after indomethacin treatment. Consequently,treatment with a mitochondrial ROS scavenger such asrebamipide should be considered as a preventive therapy forNSAID-induced intestinal injury.

In conclusion, indomethacin treatment causes intestinalcellular injury from mitochondrial ROS in areas both with and

without macroscopic lesions. The MnSOD-inducing reagentrebamipide was shown to be effective at preventing this injury.

Conflict of interests: None declared.

REFERENCES

1 Arora G, Singh G, Triadafilopoulos G. Proton pumpinhibitors for gastroduodenal damage related to nonsteroidalanti-inflammatory drugs or aspirin: twelve importantquestions for clinical practice. Clin Gastroenterol Hepatol2009; 7: 725-735.

2 Graham DY, Agrawal NM, Roth SH. Prevention of NSAID-induced gastric ulcer with misoprostol: multicentre, double-blind, placebo-controlled trial. Lancet 1988; 332(8623):1277-1280.

3 Laine L, Takeuchi K, Tarnawski A. Gastric mucosal defenseand cytoprotection: bench to bedside. Gastroenterology2008; 135: 41-60.

4 Guth PH, Aures D, Paulsen G. Topical aspirin plus HClgastric lesions in the rat. Cytoprotective effect ofprostaglandin, cimetidine, and probanthine.Gastroenterology 1979; 76: 88-93.

5 Laine L, Takeuchi K, Tarnawski A. Gastric mucosal defenseand cytoprotection: bench to bedside. Gastroenterology2008; 135: 41-60.

6 Higuchi K, Yoda Y, Amagase K, et al. Prevention of NSAID-induced small intestinal mucosal injury: prophylactic potentialof lansoprazole. J Clin Biochem Nutr 2009; 45: 125-130.

7 Tajima A. Non-steroidal anti-inflammatory drug (NSAID) -induced small intestinal injury. Pharm Anal Acta 2014; 5: 282

8 Higuchi K, Umegaki E, Watanabe T, et al. Present status andstrategy of NSAIDs-induced small bowel injury. J Gastroenterol 2009; 44: 879-888.

9 Lanas A, Garcia-Rodríguez LA, Polo-Tomas M, et al. Timetrends and impact of upper and lower gastrointestinalbleeding and perforation in clinical practice. Am JGastroenterol 2009; 104: 1633-1641.

10 Nagano Y, Matsui H, Shimokawa O, et al. Rebamipideattenuates nonsteroidal anti-inflammatory drugs (NSAID)induced lipid peroxidation by the manganese superoxide

439

Fig. 5. Determination of the number of detaching cells.Indomethacin or vehicle was administered p.o. to mice,and mice were kept for 12 hours. The intestine wasexposed through a small incision of the peritoneum.Detaching cells were obtained by wiping the smallintestinal mucosa with a painting brush. Numbers ofthe detaching cells were counted by flow cytometryafter staining with the Guava ViaCount Reagent. n=4,mean ± S.E. ** p<0.01, ANOVA followed by theTukey method.

dismutase (MnSOD) overexpression in gastrointestinalepithelial cells. J Physiol Pharmacol 2012; 63: 137-142.

11 Setsukinai K, Urano Y, Kakinuma K, Majima HJ, Nagano T.Development of novel fluorescence probes that can reliablydetect reactive oxygen species and distinguish specificspecies. J Biol Chem 2003; 278: 3170-3175.

12 Takagi T, Naito Y, Yoshikawa T, et al. Identification ofdihalogenated proteins in rat intestinal mucosa injured byindomethacin. J Clin Biochem Nutr 2011; 48: 178-182.

13 Amagase K, Yoshida Y, Hara D, Murakami T, Takeuchi K.Prophylactic effect of egualen sodium, a stable azulenederivative, on gastrointestinal damage induced byischemia/reperfusion, double antiplatelet therapy andloxoprofen in rats. J Physiol Pharmacol 2013; 64: 65-75.

14 Moore A, Donahue CJ, Bauer KD, Mather JP, et al.Simultaneous measurement of cell cycle and apoptotic celldeath. Methods Cell Biol 1998; 57: 265-278.

15 Kwiecien S, Konturek PC, Sliwowski T, et al. Interactionbetween selectiove cyclooxygenase inhibitors and capsaicin-sensitive afferent sensory nerves in pathogenesis of stress-induced gastric lesions. Role of oxidative stress. J PhysiolPharmacol 2012; 63: 143-151.

16 Korbecki J, Baranowska-Bosiacka I, Gutowska I, ChlubekD. The effect of reactiove oxygen species on the synthesis ofprostanoids from arachidonic acid. J Physiol Pharmacol2013; 64: 409-421.

17 Tamura M, Matsui H, Nagano YN, et al. Salt is an oxidativestressor for gastric epithelial cells. J Physiol Pharmacol2013; 64: 89-94.

18 Somasundaram S, Sigthorsson G, Simpson RJ, et al.Uncoupling of intestinal mitochondrial oxidativephosphorylation and inhibition of cyclooxygenase arerequired for the development of NSAID-enteropathy in therat. Aliment Pharmacol Ther 2000; 14: 639-650.

19 Matsui H, Shimokawa O, Kaneko T, Nagano Y, Rai K, HyodoI. The pathophysiology of non-steroidal anti-inflammatorydrug (NSAID)-induced mucosal injuries in stomach and smallintestine. J Clin Biochem Nutr 2011; 48: 107-111.

20 Nishikawa T, Edelstein D, Du XL, et al. Normalizingmitochondrial superoxide production blocks three pathwaysof hyperglycaemic damage. Nature 2000; 404: 787-790.

21 Ott M, Gogvadze V, Orrenius S, Zhivotovsky B.Mitochondria, oxidative stress and cell death. Apoptosis2007; 12: 913-922.

22 Brand MD, Affourtit C, Esteves TC, et al. Mitochondrialsuperoxide: production, biological effects, and activation ofuncoupling proteins. Free Radic Biol Med 2004; 37: 755-767.

23 Brzozowski T, Konturek PC, Konturek SJ, et al. Role ofgastric acid secretion in progression of acute gastric erosionsinduced by ischemia-reperfusion into gastric ulcers. Eur JPharmacol 2000; 398: 147-158.

24 Tsutsumi S, Tomisato W, Takano T, Rokutan K, Tsuchiya T,Mizushima T. Gastric irritant-induced apoptosis in guineapig gastric mucosal cells in primary culture. BiochimBiophys Acta 2002; 1589: 168-180.

25 Nagano Y, Matsui H, Muramatsu M, et al. Rebamipidesignificantly inhibits indomethacin-induced mitochondrialdamage, lipid peroxidation, and apoptosis in gastric epithelialRGM-1 cells. Dig Dis Sci 2005; 50 (Suppl. 1): S76-S83.

26 Somasundaram S, Hayllar H, Rafi S, Wrigglesworth JM,Macpherson AJ, Bjarnason I. The biochemical basis of non-steroidal anti-inflammatory drug-induced damage to thegastrointestinal tract: a review and a hypothesis. Scand JGastroenterol 1995; 30: 289-299.

27 Indo HP, Inanami O, Koumura T, et al. Roles ofmitochondria-generated reactive oxygen species on X-ray-induced apoptosis in a human hepatocellular carcinoma cellline, HLE. Free Radic Res 2012; 46: 1029-1043.

28 Majima HJ, Matsui H, Ozawa T, et al. Mitochondria assource of free radicals. Free Radic Biol 2011; 29: 12-22.

29 Indo HP, Davidson M, Yen HC, et al. Evidence of ROSgeneration by mitochondria in cells with impaired electrontransport chain and mitochondrial DNA damage.Mitochondrion 2007; 7: 106-118.

30 Majima HJ, Oberley TD, Furukawa K, et al. Prevention ofmitochondrial injury by manganese superoxide dismutasereveals a primary mechanism for alkaline-induced cell death.J Biol Chem 1998; 273: 8217-8224.

31 Majima HJ, Matsui H, Suenaga S, Matsui H, Yen HC,Ozawa T. Mitochondria as possible pharmaceutical targetsfor the effects of vitamin E and its homologues in oxidativestress-related diseases. Curr Pharm Des 2011; 17: 2190-2195.

R e c e i v e d : November 30, 2013A c c e p t e d : April 18, 2014

Author's address: Dr. Hirofumi Matsui, Division of ClinicalMedicine, Faculty of Medical Sciences, University of Tsukuba,1-1-1, Ten-nohdai, Tsukuba, JapanE-mail: hmatsui@md.tsukuba.ac.jp

440