Download pdf - Mechanism of gastroprotection by bismuth subsalicylate against chemically-induced oxidative injury in human gastric mucosal cells

Mechanism of Gastroprotection by Bismuth

Subsalicylate Against Chemically Induced

Oxidative Stress in Cultured Human Gastric

Mucosal Cells

DEBASIS BAGCHI, PhD, THOMAS R. MCGINN, MD, XUMEIN YE, BS, JAYA BALMOORI, MS,

MANASHI BAGCHI, PhD, SIDNEY J. STOHS, PhD, CHARLES A. KUSZYNSKI, PhD,

OWEN R. CARRYL, PhD, and SEKHAR MITRA, PhD

Reactive oxygen species (ROS) are implicate d in the pathogenesis of chemically induced

gastric mucosal injury. We have investigated the effects of ethanol, hydrochloric acid (HCl),

and sodium hydroxide (NaOH) on: (1) enhance d production of ROS including superoxide

anion and hydroxyl radicals, (2) modulation of intrace llular oxidized states by lase r scanning

confocal microscopy, and (3) DNA fragmentation, indice s of oxidative tissue , and DNA

damage in a primary culture of normal human gastric mucosal cells (GC), which were isolated

and cultured from Helicobacter pylori-negative endoscopic biopsie s from human subje cts. The

induction of ROS and DNA damage in these cells following exposure to ethanol (15% ), HCl

(150 mM) and NaOH (150 mM) were assessed by cytochrome c reduction (superoxide anion

production) , HPLC detection for enhanced production of hydroxyl radicals, changes in

intrace llular oxidized states by lase r scanning confocal microscopy, and DNA damage by

quantitating DNA fragmentation. Furthe rmore, the protective ability of bismuth subsalicylate

(BSS) was assessed at concentrations of 25, 50, and 100 mg/liter. Incubation of GC with

ethanol, HCl, and NaOH increased superoxide anion production by approximate ly 8.0-, 6.1-

and 7.1-fold and increased hydroxyl radical production by 13.3-, 9.6-, and 8.9-fold, respec-

tively, compared to the untreated gastric cells. Incubation of GC with ethanol, HCl, and

NaOH increased DNA fragmentation by approximate ly 6.7-, 4.3-, and 4.8-fold, respective ly.

Approximate ly 20.3-, 17.5-, and 13.1-fold increases in ¯ uore scence intensitie s were obse rved

following incubation of gastric cells with ethanol, HCl, and NaOH, respective ly, demonstrat-

ing dramatic change s in the intrace llular oxidize d states of GC following exposure to these

necrotizing agents. Preincubation of GC with 25, 50, and 100 mg/lite r of BSS decreased

ethanol-induce d increases in intrace llular oxidized states in these cells by 36% , 56% , and

66% , respective ly, demonstrating a concentration-de pendent protective ability by BSS. Sim-

ilar results were obse rved with respect to BSS in terms of superoxide anion and hydroxyl

radical production, and DNA damage . The present study demonstrates that ethanol, HCl,

and NaOH induce oxidative stress and DNA damage in GC and that BSS can signi® cantly

attenuate gastric injury by scavenging these ROS.

KEY WORDS: human gastric ce lls; necrotizing age nts; reactive oxyge n species; laser scanning confocal microscopy;

DNA fragmentation.

Digestive Diseases and Sciences, Vol. 44, No. 12 (Decem ber 1999), pp. 2419 ± 2428

2419Digestive Diseases and Sciences, Vol. 44, No. 12 (Decem ber 1999)

0163-2116/99/1200-2419$16.00/0 Ñ 1999 Plenum Publishing Corporation

An effective balance exists in a healthy human stom-

ach between the potential for gastric acid and pepsin

to damage gastric mucosal cells and the ability of

these gastric cells to protect themselves from injury

(1± 4). Disruption of this balance has been attribute d

to several factors including environmental, emotional,

and chemical stress, age , overindulge nce of food,

genetics, and individual behavior, and is evidenced as

a burning, aching, or gnawing pain that may be per-

ceived as abdominal pressure or fullne ss (1, 3). Most

of the symptoms experienced by the patients unde r

such conditions result from a breakdown of the nor-

mal mucosal defense mechanisms (3). It is well doc-

umented that gastric acid and pepsin are important in

the pathogenesis of dyspepsia, stomach upse t, gastro-

esophage al re¯ ux disease, and duodenal and gastric

ulce r (1, 3, 4). Several mechanisms are be lieved to be

important in protecting gastric and duodenal mucosa

from damage by acid, pepsin, bile , pancreatic en-

zymes, as well as these external stressors/factors (1, 2,

4). These defense mechanisms include mucus, muco-

sal blood ¯ ow, cell renewal, and bicarbonate (4).

These factors acting in balance he lp in maintaining

mucosal integrity. It has been suggested that reactive

oxyge n species (ROS) are greatly involve d in the

pathoge nesis of gastric injury (1). However, the de-

tailed mechanistic pathways of these events are still

unknown.

In our previous studie s, we have demonstrated the

role and involve ment of ROS in the pathogenesis of

restraint stress, high fat diet, spicy food, and alcohol-

induced gastric lesions and gastrointe stinal mucosal

injury and that protection against the ROS induce d

by these stressors can ameliorate gastric injury (5).

We have also demonstrated that oral administration

of necrotizing agents including 0.6 M HCl, 0.2 M

NaOH, 80% ethanol, and 200 mg/kg aspirin induce

increased lipid peroxidation and membrane microvis-

cosities in gastric and inte stinal mucosa as well as

cause gastric and intestinal mucosal DNA fragmen-

tation in female Sprague -Dawley rats (6). Pretreat-

ment and posttreatment of the rats with bismuth

subsalicylate (BSS) provide d signi® cant protection

against the mucosal injury induced by these necrotiz-

ing agents (6).

To furthe r establish our hypothe sis, we have inve s-

tigated the effects of ethanol (15% ), HCl (150 mM),

and NaOH (150 mM) on the enhance d production of

ROS including superoxide anion and hydroxyl radi-

cals, and modulation of intrace llular oxidize d states

by lase r scanning confocal microscopy and DNA frag-

mentation, indice s of oxidative tissue damage , in cul-

tured human gastric mucosal cells isolate d from en-

doscopic biopsie s. Furthe rmore, the concentration-

dependent protective ability of BSS was determined

against the gastric injury induce d by these stressors.

MATERIALS AND METHODS

Cell Culture and Treatment. A primary culture of normalhuman gastric mucosal cells was deve loped using keratino-cyte growth medium supplemented with pituitary extract(KGM) (Clonetics, La Jolla, California) following a modi-® ed procedure of Wagner et al (7). Helicobacter pylori-negative endoscopic biopsies (four to six biopsies) fromhuman subjects were predigested in 2 ml medium A (70mmol/liter NaCl, 20 mmol/liter NaHCO 3, 1.5 mmol/literNa2HPO4, 5 mmol/liter KCl, 1.5 mmol/liter MgCl2, 1 mmol/liter CaCl2, 11 mmol/liter glucose and 50 mmol/literHEPES) containing 3 mg pronase. The presence of Helico-bacter pylori in these biopsies was determined using a com-mercial CLO-Test strip (Tri-Med Specialities Inc., Lenexa,Kansas). Afte r predigestion for 15 min at 37°C, the tissuewas washed with medium A and later digested for 45 minwith 2 mg collagenase in 2 ml medium A containing 0.2%bovine serum albumin. The cell suspensions were then® ltered through 100- and 60- m M nylon ® lters and sus-pended in KGM growth medium. The gastric cells were® nally incubated at 37°C in collagen-coated (200 m g rat tailcollagen/well) microplates in KGM growth medium supple-mented with 10% heat-inactivated fetal calf serum in anatmosphere of 5% CO2 in 100% humidity, and the culturemedium was changed daily. The parietal and chief cells didnot attach to the collagen-coated culture plates and wereremoved during change of medium. Periodic acid± Schiff(PAS) staining was used to determine the integrity of mu-cosal cells. Figure 1 represents the progressive growth ofhuman gastric cells. Primary cultures of these cells growsigni® cantly rapidly, and 96 ± 100% con¯ uence can be ob-tained in the culture ¯ asks in approximate ly four to ® vedays. An approval (IRB # 94-11121) was obtained from theCreighton University Institutional Review Board for thisproject. Cultured cells were preincubated with 25, 50, or 100mg BSS/liter for 12 hr prior to the exposure to the followingnecrotizing agents including ethanol (15% ), hydrochloricacid (150 mM), and sodium hydroxide (150 mM).

Ethanol. In this study, we have demonstrated the effect of15% ethanol for 1 hr on human gastric mucosal cells,isolated from Helicobacter pylori-negative endoscopic biop-sies from human subjects. Previous workers have demon-strated (8 ± 10) that incubation of gastric cells with 15%ethanol for 1 hr can signi® cantly elicit cell damage in gastric

Manuscript receive d December 28, 1998; revised manuscriptrece ived April 7, 1999; acce pted April 20, 1999.

From the Creighton University Health Science s Center, Omaha,Nebraska; University of Nebraska Medical Center, Omaha, Ne-braska; and Health Care Research Center, The Procter & GambleCompany, Mason, Ohio.

Address for reprint requests: Debasis Bagchi, Department ofPharmace utical and Administrative Sciences, Creighton UniversitySchool of Pharmacy and Allied Health Professions, 2500 CaliforniaPlaza, Omaha, Nebraska 68178.

BAGCHI ET AL

2420 Digestive Diseases and Sciences, Vol. 44, No. 12 (Decem ber 1999)

Fig 2. Changes in intracellular oxidized states following treatment of cultured normal human gastric mucosal ce lls with15% ethanol and/or selected concentrations of bismuth subsalicylate (BSS) (a) Control; (b) BSS 100 mg/lite r; (c)

ethanol (15% ); (d) BSS 25 mg/liter 1 ethanol (15% ); (e) BSS 50 mg/lite r 1 ethanol (15% ); (f) BSS 100 mg/lite r 1ethanol (15% ). (Reduced at 85% for reproduction.)

Fig 1. Progre ssive growth of normal human gastric mucosal ce lls. 100 3 phase-contrast photographs under light microscope. (A)

day 2; (B) day 3; (C) day 4; (D) day 5, 96 ± 100% con¯ uent. (Reduced at 85% for reproduction.)

GASTROPROTECTION BY BISMUTH SUBSALICYLATE


mucous cells. Human gastric mucosal cells were exposed to

a premixed medium containing 15% ethanol.

Hydrochloric Acid. The effect of 150 mM hydrochloricacid on cultured human gastric mucosal cells for 1 hr was

investigated in this project. This concentration of hydro-

chloric acid was used by previous investigators (11, 12).

Hydrochloric acid was initially added to the cell culturemedium and later the human gastric mucosal cells were

exposed to this medium.

Sodium Hydroxide. Cultured human gastric mucosal cellswere exposed to 150 mM NaOH for 1 hr to assess the effectof sodium hydroxide. A sterilized stock solution of sodiumhydroxide was added to the culture medium and later thegastric mucosal cells were exposed to this medium. Thisconcentration was selected from the earlier study of Ogiu(13) .

Chemicals. Bismuth subsalicylate (BSS; material #4524;batch # 4; Trash date 11-15-97; moisture content 53.6% ]was obtained from The Procter & Gamble Company (Cin-cinnati, Ohio). With the exception of the HPLC suppliesand selected cell culture items, all other chemicals used inthis study were obtained from Sigma Chemical Co. (St.Louis, Missouri) and were of analytical grade or the highestgrade commercially available . Sodium acetate trihydrateand citric acid used for HPLC were purchased from Fluka(Buchs, Switzerland). 2,3-Dihydroxybenzoic acid, 2,5-dihydroxybenzoic acid, salicylic acid, and HPLC-grade wa-ter were purchased from Aldrich Chemical Company (Mil-waukee, Wisconsin).

Cytochrome c Reduction. Superoxide anion productionwas measured by the assay method of Babior et al (14) andas described previously by us (15), which is based on cyto-chrome c reduction. In 1.0 ml, the reaction mixtures con-tained 1 mg protein and 0.05 mM cytochrome c in theincubation buffer. The incubation mixtures were incubatedfor 15 min at 37°C. The reactions were terminated byplacing the reaction mixtures on ice. The mixtures werecentrifuged for 10 min, and the supernatant fractions weretransferred to clean tubes for spectrophotometric measure-ment at 550 nm. Absorbance values were converted tonanomoles of cytochrome c reduced per 15 min per milli-gram protein using the extinction coef® cient 2.1 3 10

4/M/

cm.Hydroxyl Radical Detection by HPLC. The detection of

hydroxyl radicals based on the formation of 2,3-dihydroxy-benzoic acid and 2,5-dihydroxybenzoic acid was determinedin an HPLC equipped with a Waters 460 electrochemicaldetector as previously described by us (16, 17). The culturedcells were incubated in the presence of 100 m M sodiumsalicylate solution 1 hr prior to the completion of theexperiment. The samples were ® ltered through a Rainin0.22- m m pore size Nylon-66 sample ® lter, and 25- m l vol-umes of the samples were injected onto an Altex Ultras-phere (3 m m ODS, 75 3 4.6 mm) column (Rainin) pro-tected by a Rainin Brownlee Spheri-5 polyfunctional C18

ODS-GU precolumn (30 3 4.6 mm). The hydroxylatedproducts of salicylic acid afte r interaction with hydroxylradicals, 2,3-dihydroxybenzoic acid, and 2,5-dihydroxyben-zoic acid, were eluted with a mobile phase containing 0.03mol/liter sodium acetate and 0.03 mol/liter of citric acid (pH3.6) at a ¯ ow rate of 1 ml/min. The detection potential wasmaintained at 1 0.6 V, employing a glassy carbon working

electrode and an Ag± AgCl reference electrode. Retentiontimes for the peaks of 2,3-dihydroxybenzoic acid and 2,5-dihydroxybenzoic acid were veri ® ed by injecting authenticstandards.

Laser Scanning Confocal Microscopy. The overall intra-cellular oxidized state s of cells before and after exposure toBSS and/or the necrotizing agents were measured at anexcitation wave length of 513 nm using a Meridian ACS 570Scanning Confocal Laser Microscope (Okemos, Missouri)and 2,7-dichloro¯ uorescein diacetate (DCFD) as the ¯ uo-rescent probe (Molecular Probes, Inc., Eugene, Oregon).Concentration- and time-dependent changes in ¯ uores-cence intensity were assessed according to the methodpreviously described by us (18, 19) . Signals were quantitatedby integrating ¯ uorescence intensity over a user de ® nedarea cell number. Relative ¯ uorescence intensity of eachcell was calculated relative to untreated control cells. Ap-proximately 50 cells were used for each time point deter-mination.

DNA Fragm entation . Human gastric mucosal cells wereexposed to BSS and/or the necrotizing agents at 37°C. Thecontrol and treated cells were removed from the platesusing 1.5 ml of a lysis buffer consisting of 5 mM Tris HCl, 20mM EDTA, and 0.5% Triton X-100 at pH 8.0, and allowedto lyse for 15 min on ice. The samples were then centrifugedat 27,000g for 20 min to separate intact chromatin in thepellets from fragmented DNA in the supernatant fractions.Pellets were resuspended in 0.5 N perchloric acid and 5.5 Nperchloric acid was added to supernatant samples (100

m l/ml) to reach a concentration of 0.5 N. Supernatant andresuspended pellet samples were heated at 90°C for 15 minand centrifuged at 1500g for 10 min to precipitate thedebris. Resulting supernatant fractions were assayed forDNA content using diphenylamine, a reagent that excludesRNA by reacting preferentially with 2-deoxy sugars for16 ± 20 hr at room temperature, and absorbances were mea-sured at 600 nm. DNA fragmentation in control samples is

TABLE 1. NECROTIZING AGENT-INDUCED PRODUCTION OF

SUPEROXIDE ANION (BASED ON CYTOCHROME C REDUCTION) IN

CULTURED HUMAN GASTRIC MUCOSAL CELLS AND PROTECTIVE

ABILITY OF BISMUTH SUBSALICYLATE (BSS)*

Cytochrom e c (nmol reduced/15

min/26 3 104

cells)

Control group 1.50 6 0.21a

BSS 50 mg/liter 1.68 6 0.35a

BSS 100 mg/liter 1.90 6 0.35b

EtOH 15% 11.97 6 1.54c

BSS 25 mg/liter 1 EtOH 9.69 6 2.34d

BSS 50 mg/liter 1 EtOH 8.74 6 1.26e

BSS 100 mg/liter 1 EtOH 8.12 6 1.24f

HCl 150 mM 9.19 6 1.60d,e

BSS 25 mg/liter 1 HCl 7.00 6 0.98g

BSS 50 mg/liter 1 HCl 6.43 6 1.08h

BSS 100 mg/liter 1 HCl 5.81 6 1.02 i

NaOH 150 mM 10.62 6 1.65c,d

BSS 50 mg/liter 1 NaOH 8.95 6 1.63e

BSS 100 mg/liter 1 NaOH 8.21 6 1.27e

* Cultured human gastric mucosal ce lls were incubated with BSSand/or the necrotizing agent( s). Each value is the mean 6 SD of

4 ± 6 individual expe riments. Values with different superscriptsare signi® cantly different (P , 0.05) .

BAGCHI ET AL


expressed as percent of total DNA appearing in the super-natant fraction. Treatment effects are reported as percentof control fragmentation as described earlier by us (5, 6) .

Statis tical Methods. The presence of signi® cant differ-ences between mean values was determined using Student’st test or by analysis of variance (ANOVA) followed byScheffe’ s S method as the post hoc test. Each value is themean 6 SD from at least four to six experiments. The leve lof statistical signi® cance in all cases was P , 0.05.

RESULTS

Superoxide Anion and Hydroxyl Radical Produc-

tion in Cultured Human Gastric Mucosal Cells Fol-

lowin g Incubation with Ethanol (15%), Hydroch loric

Acid (150 mM), and Sodium Hydroxide (150 mM).

The production of reactive oxyge n species, including

superoxide anion and hydroxyl radicals, in cultured

human gastric mucosal cells following incubation with

these necrotizing agents was assessed by cytochrome c

reduction and hydroxyl radical production . Cyto-

chrome c reduction is a relative ly speci® c test for

superoxide anion production. Production of hydroxyl

radicals was determined based on the formation of

hydroxylate d products of salicylic acid (2,3-dihydroxy-

benzoic acid and 2,5-dihydroxybe nzoic acid) follow-

ing interaction with hydroxyl radicals. These hydroxy-

lated products were de te rmined using a HPLC

equippe d with an electrochemical detector.

The results of cytochrome c reduction (an index of

superoxide anion production) are presented in Table

1. These data demonstrate the enhanced production

of superoxide anion following incubation with these

necrotizing agents. Approximate ly 8.0-, 6.1-, and 7.1-

fold increases in cytochrome c reduction were ob-

served in gastric mucosal cells following incubation

with ethanol (15% ), hydrochloric acid (150 mM), and

sodium hydroxide (150 mM), respective ly, compared

to the control cells (Table 1). Pretreatment of gastric

cells with 50 and 100 mg/lite r concentrations of BSS

alone increased cytochrome c reduction by 1.1- and

1.3-fold, respectively, compared to control cells (Ta-

ble 1).

Pretreatment of the gastric mucosal cells with 25,

50, and 100 mg/liter concentrations of BSS decreased

ethanol-induce d cytochrome c reduction by approxi-

mately 19% , 27% , and 32% , respective ly, while unde r

these same conditions and concentrations with BSS,

HCl-induce d cytochrome c reduction was decreased

by 24% , 30% , and 37% , respectively, compared to the

corresponding control sample s (Table 1). Approxi-

mately 16% and 23% decreases in DNA fragmenta-

tion were obse rved in NaOH-induce d gastric injury

following coincubation of the cultured gastric muco-

sal cells with 50 and 100 mg/liter of BSS, respective ly

(Table 1).

The results of hydroxyl radical production in the


BSS and/or the necrotizing agent(s) are shown in

Table 2. These data illustrate the enhanced produc-

tion of hydroxyl radicals following incubation of hu-

man gastric mucosal cells with these necrotizing

agents. Approximate ly 13.3-, 9.6-, and 8.9-fold in-

creases in hydroxyl radical production were observed

in gastric mucosal cells following incubation with eth-

anol (15% ), hydrochloric acid (150 mM), and sodium

hydroxide (150 mM), respectively, compared to the

control cells. Pretreatment of these gastric cells with

50 and 100 mg/lite r concentrations of BSS alone

increased hydroxyl radical production by 1.4- and

1.6-fold, respective ly, compared to the control cells

(Table 2).



ethanol-induce d hydroxyl radical production in hu-

man gastric cells by approximate ly 17% , 27% , and

51% , respectively, while unde r these same conditions

TABLE 2. NECROTIZING AGENT-INDUCED PRODUCTION OF

HYDROXYL RADICALS BASED ON FORMATION OF 2,3- AND

2,5-DIHYDROXYBENZOIC ACID (DHBA) IN CULTURED HUMAN

GASTRIC MUCOSAL CELLS AND PROTECTIVE ABILITY OF BISMUTH

SUBSALICYLATE (BSS)

Form ation of 2,3- and2,5-DHBA (nm oles/m l)

2.5-DHBA 2.3-DHBA Total

Control group 0.16 6 0.03a

0.29 6 0.08a

0.45BSS 50 mg/liter 0.22 6 0.02b 0.43 6 0.08b 0.65

BSS 100 mg/liter 0.21 6 0.03b

0.50 6 0.07b

0.71EtOH 15% 1.59 6 0.18c 4.40 6 0.76c 5.99

BSS 25 mg/liter

1 EtOH 1.74 6 0.18c 3.22 6 0.62d 4.95

BSS 50 mg/liter

1 EtOH 1.42 6 0.19d 2.95 6 0.96e 4.37

BSS 100 mg/liter

1 EtOH 1.10 6 0.30e 1.81 6 0.41f 2.91

HCl 150 mM 1.44 6 0.22d

2.88 6 0.24e

4.33BSS 25 mg/liter

1 HCl 1.29 6 0.33d

2.14 6 0.51f

3.44BSS 50 mg/liter

1 HCl 1.24 6 0.20d,e

1.79 6 0.52f

3.03BSS 100 mg/liter

1 HCl 1.02 6 0.32e

1.18 6 0.24g

2.20NaOH 150 mM 1.33 6 0.08d 2.66 6 0.50e 3.99

BSS 50 mg/liter

1 NaOH 1.19 6 0.15e 2.44 6 0.80e ,f 3.63

BSS 100 mg/liter

1 NaOH 1.04 6 0.33e 2.25 6 0.69f 3.29





and concentrations with BSS, HCl-induce d hydroxyl

radical production was decreased by 21% , 30% , and

49% , respective ly, compared to the corresponding

control sample s. Approximate ly 9% and 18% de-

creases in hydroxyl radical production were observed

in NaOH-induced gastric injury following coincuba-

tion of the cultured gastric mucosal cells with 50 and

100 mg/liter of BSS, respective ly, compared to the

control cells (Table 2).

Modulation of Intracellu lar Oxid ized States in

Cultu red Human Gastric Mucosal Cells Followin g

Exposure to These Necrotizin g Agen ts. Laser scan-

ning confocal microscopy is an exce llent tool to de-

termine the modulation of intrace llular oxidize d

states in cultured cells following exposure to BSS

and/or dive rse necrotizing agents. Change s in ¯ uore s-

cence intensity were determined following exposure

to these necrotizing agents and/or BSS. Signals were

quantitate d by integrating ¯ uorescence intensity over

a user de ® ned area cell number. Relative ¯ uore s-

cence intensity of each cell was calculated relative to

untreated control cells. Approximate ly 50 cells were

used for individual experiment. The concentration-

dependent prote ctive ability of BSS on e thanol

(15% )-induce d modulation of intrace llular oxidize d

states in cultured human gastric mucosal cells is

shown in Figure 2.

The modulation of intrace llar oxidize d states in

cultured human gastric mucosal cells following incu-

bation with BSS and/or the necrotizing agent(s) is

shown in Table 3. These data demonstrate the en-

hanced production of ROS following incubation with

these necrotizing agents and the protection by BSS.

Approximate ly 20.3-, 17.5-, and 13.1-fold increases in

¯ uorescent intensity in these cells were obse rved fol-

lowing incubation with ethanol (15% ), hydrochloric

acid (150 mM), and sodium hydroxide (150 mM),

respective ly, compared to the control cells (Table 3).

Pretreatment of these gastric cells with 50 and 100

mg/liter concentrations of BSS alone resulted in only

1.6- and 2.7-fold increases in ¯ uore scent intensity,

respective ly, compared to the control cells (Table 3).



ethanol-induce d increases in ¯ uorescent intensitie s by

approxim ate ly 36% , 56% , and 69% , respective ly,

while unde r these same conditions and concentra-

tions with BSS, HCl-induce d change s in ¯ uorescent

intensitie s decreased by 39% , 58% , and 70% , respec-

tive ly, compared to the corresponding control sam-

ples (Table 3). Approximate ly 11% and 46% de-

creases in ¯ uore scent intensity were observed in

NaOH-induced gastric injury following coincubation

of the cultured gastric mucosal cells with 50 and 100

mg/lit of BSS, respective ly (Table 3).

Necrotizing Agen t-Induced Changes in DNA Frag-

mentation in Gastr ic Mucosal Cells . Necrotizing

agent-induce d DNA fragmentation was determined

as an index of oxidative DNA damage in cultured

human gastric mucosal cells, and the results are

shown in Table 4. Furthe rmore , the concentration-

dependent protective ability of BSS was assessed.

Approximate ly 6.7-, 4.3-, and 4.8-fold increases in

DNA fragmentation were obse rved in gastric mucosal

cells following incubation with ethanol (15% ), hydro-

chloric acid (150 mM), and sodium hydroxide (150

mM), respectively, compared to the control cells (Ta-

ble 4). Pretreatment of these gastric cells with 50 and

100 mg/lite r concentrations of BSS alone increased

DNA fragmentation by 1.3- and 1.6-fold, respectively,

compared to the control cells (Table 4).



e thanol-induced DNA fragmentation by approxi-

mately 18% , 28% , and 38% , respective ly, while unde r

these same conditions and concentrations with BSS,

HCl-induce d DNA fragmentation was decreased by

14% , 28% , and 36% , respective ly, compared to the

corresponding control sample s (Table 4). Approxi-

mately 11% and 18% decreases in DNA fragmenta-

TABLE 3. MODULATION OF INTRACELLULAR OXIDIZED STATES IN

CULTURED HUMAN GASTRIC MUCOSAL CELLS FOLLOWING

INCUBATION WITH NECROTIZING AGENTS AND PROTECTIVE

ABILITY OF BISMUTH SUBSALICYLATE (BSS)*

Fluorescent intensity

Control group 85 6 15a

BSS (50 mg/liter) 136 6 38b

BSS (100 mg/lite r) 211 6 103c

EtOH 15% 1725 6 511d

BSS 25 mg/liter 1EtOH 1105 6 247e

BSS 50 mg/liter 1EtOH 759 6 172 f

BSS 100 mg/liter 1EtOH 535 6 85g

HCl 150 mM 1485 6 331h

BSS 25 mg/liter 1 HCl 906 6 197e

BSS 50 mg/liter 1 HCl 624 6 204g

BSS 100 mg/liter 1 HCl 446 6 93g

NaOH 150 mM 1115 6 301e

BSS 50 mg/liter 1NaOH 992 6 178

e

BSS 100 mg/liter 1NaOH 602 6 119

g



BAGCHI ET AL


tion were obse rved in NaOH-induce d gastric injury

following coincubation of the cultured gastric muco-

sal cells with 50 and 100 mg/liter of BSS, respective ly

(Table 4).

DISCUSSION

The main obje ctive of this proje ct was to establish

a primary culture of human gastric mucosal cells from

Helicobacter pylori-negative endoscopic biopsie s and

to assess the relationship between gastrointe stinal

oxidative stress induce d by chemical stressors includ-

ing ethanol (15% ), hydrochloric acid (150 mM), and

sodium hydroxide (150 mM) and to determine the

gastroprotective ability of bismuth subsalicylate

(BSS). The primary culture of normal human gastric

mucosal cells was deve lope d using ke ratinocyte

growth medium supplemented with pituitary extract

(KGM) following a modi® ed procedure of Wagner et

al (7). The KGM growth medium provided improve d

structural integrity and viability of these primary cul-

tures. These cells were exposed to the chemical stres-

sors, and the extent of injury in these gastric mucosal

cells was determined by measuring cytochrome c re-

duction (an index of superoxide anion production) ,

production of hydroxyl radicals and DNA fragmenta-

tion. Modulation of intrace llular oxidized states fol-

lowing exposure to these chemical stressors was also

determined. Furthermore, the chemoprote ctive abil-

ity of BSS was also assessed. Bismuth salts are com-

monly used to treat a varie ty of gastrointe stinal dis-

orders, including peptic ulce rs, dyspepsia, infectious

diarrhea, and parasitic infections (20, 21) . Agents that

provide protection against gastric and intestinal mu-

cosal injury by a mechanism other than inhibition or

neutralization of gastric acid have been described as

gastric cytoprote ctants (4). Several mechanisms of

gastric cytoprote ction have been propose d (22) . How-

ever, the mechanism associated with cytoprote ction

by BSS in gastrointe stinal ecology is not clearly un-

derstood.

The involve ment of ROS and its metabolite s has

been demonstrated in several gastrointe stinal pathol-

ogies, including inte stinal ischemia and reperfusion,

gastrointe stinal in¯ ammatory diseases, in¯ ammatory

bowe l diseases, acute and chronic pancreatitis, during

gastrointe stinal metabolism of xenobiotics, and after

hemorrhagic shock (23) . It has been demonstrated

that ethanol may serve as an antioxidant at low con-

centrations and protect from coronary heart diseases

by preventing athe rosclerosis through the action of

high-de nsity-lipoprote in chole sterol or through a he-

mostatic mechanism (24, 25) . However, at moderate

and high concentrations, ethanol functions as a po-

tent prooxidant and causes signi® cant gastrointe stinal

mucosa injury (24, 25) . Previous workers have dem-

onstrated that a concentration of 15% ethanol can

induce signi® cant mucosal injury (8 ± 10) , and accord-

ingly this concentration was selected. Previous studie s

in our laboratory (6) and others (26, 27) have dem-

onstrated that concentrated acid and base can cause

signi® cant gastrointe stinal injury mediated by ROS

especially hydroxyl radicals, and selected free radical

scavengers including dimethylthioure a (DMTU) can

protect against these gastrointe stinal mucosal inju-

ries. These studie s suggest that ROS may be one of

the pathways of inducing oxidative gastrointe stinal

injury by hydrochloric acid or sodium hydroxide . Our

present study was primarily focused on determining

the individual effects of ethanol, hydrochloric acid, or

sodium hydroxide on a primary culture of human

gastric mucosal cells because similar chemical insults

are like ly to be induce d in the gastric mucosa follow-

ing overindulge nce of food or fried, spicy, or high-fat

diet or alcoholic beverages.

Since oxidative tissue damage occurs in response to

ROS, the in vitro effects of ethanol (15% ), hydrochlo-

ric acid (150 mM), and sodium hydroxide (150 mM)

were individually assessed on a primary culture of

gastric mucosal cells, isolated and cultured from Hel-

icobacter pylori-negative endoscopic biopsie s from hu-

TABLE 4. NECROTIZING AGENT-INDUCED DNA FRAGMENTATION

IN CULTURED HUMAN GASTRIC MUCOSAL CELLS AND

PROTECTIVE ABILITY OF BISMUTH SUBSALICYLATE (BSS)

DNA fragmentation (% control)

Control group 3.25 6 0.21a (100)

BSS (50 mg/liter) 4.36 6 0.32b

(134)BSS (100 mg/lite r) 5.02 6 0.29c (155)

EtOH 15% 21.77 6 1.75d

(670)BSS 25 mg/liter 1

EtOH 17.84 6 2.13e

(549)BSS 50 mg/liter 1

EtOH 15.59 6 1.79f(480)

BSS 100 mg/liter 1EtOH 13.44 6 1.64

g(414)

HCl 150 mM 14.07 6 2.21f,g (433)

BSS 25 mg/liter 1 HCl 12.14 6 1.74g

(374)BSS 50 mg/liter 1 HCl 10.13 6 1.47g,h (312)

BSS 100 mg/liter 1 HCl 8.95 6 1.51h

(275)NaOH 150 mM 15.59 6 2.22f (480)

BSS 50 mg/liter 1NaOH 13.84 6 2.10g (426)

BSS 100 mg/liter 1NaOH 12.81 6 2.91g (394)

* Cultured human gastric mucosal ce lls were incubated with BSS

and/or the necrotizing agent( s). Each value is the mean 6 SD of4 ± 6 individual expe riments. Values with different superscripts

are signi® cantly different (P , 0.05) .



man subje cts. Cytochrome c reduction was assessed in

these gastric cells as an index of superoxide anion

production following incubation with these chemical

stressors. Cytochrome c reduction is a relative ly spe-

ci® c test for superoxide anion production (14, 28) .

Necrotizing agent-induce d production of hydroxyl

radicals in these gastric cells was determined based on

the formation of hydroxylate d products of salicylic

acid (2,3-dihydroxybe nzoic acid and 2,5-dihydroxy-

benzoic acid) following inte raction with hydroxyl rad-

icals (16, 17) . These hydroxylate d products were de-

te rmine d using an HPLC e quipped with an

electrochemical detector. A novel, state -of-the -art la-

ser scanning confocal microscopic technique was used

to assess free radical-induce d modulation of intrace l-

lular oxidized states in the primary culture of normal


these necrotizing agents as an index of oxidative dam-

age to these primary cultures at the single cell leve l.

This technique provide s the oxidative damage caused

by the dive rse necrotizing agent(s) at the single cell

leve l, the concentration-de pendent protective ability

of BSS, and can be implemented as a tool to assess

the cytoprote ctive ability of drugs. Programme d cell

death (apoptosis) is a selective process of physiolog-

ical cell deletion (29, 30) . It plays a major role in

developmental biology and in the maintenance of

homeostasis in vertebrate s. Apoptosis is accompanie d

by condensation of cytoplasm, loss of plasma mem-

brane microvilli, condensation and fragmentation of

nucle i, and extensive degradation of chromosomal

DNA (29, 30) . Since DNA fragmentation is very

closely associate d with apoptosis, DNA fragmenta-

tion was quantitate d and corre lated with oxidative

stress. The results of this study sugge st that a strong

corre lation exists between the formation of ROS and

oxidative tissue damage in gastric mucosal cells.

These results were independent of the stressor result-

ing in the production of free radicals, as well as the

presence or absence of BSS as an inhibitor of ROS.

Naganuma et al (31) demonstrated that bismuth

subnitrate signi® cantly decreased lethal toxicity, car-

diotoxicity, and bone marrow toxicity of adriamycin.

In a more recent study, Nakagawa et al (32) have

demonstrated that pretreatment of mice with bismuth

nitrate signi® cantly prevent the clastoge nicity of

adriamycin, cyclophosphamide , cisplatin, and L-

phenylalanine mustard. Since a signi® cant increase in

tissue concentrations of metallothione in is observed

following treatment with bismuth subnitrate , the au-

thors conclude d that the ability of bismuth subnitrate

to reduce toxicity of adriamycin and clastogenicity of

various antineoplastic drugs may be ascribed to the

induction of metallothione in. From these studies we

concluded that the protective effect of metallothio-

ne in might be due to its ability to scavenge reactive

oxyge n species or inhibit their formation. However,

recent studie s by DiSilve stro et al (33) have demon-

strated that transgenic mice , which overexpre ssed car-

diac metallothione in and also had moderately high

glutathione concentrations, were not resistant to ad-

riamycin-induce d cardiotoxicity. Cardiotoxicity was

assessed by survival, ¯ uid accumulation, and lipid

peroxidation. Furthermore, the effect of bismuth on

adriamycin toxicity did not differ between these ge-

netically alte red mice and control animals. Thus, it

appears that metallothione in may not be involve d in

the chemoprote ctive effects of bismuth salts.

Forse ll (3) propose d that bismuth acts as a mild

irritant, inducing liberation of endoge nous mucosal

prostaglandins that help to maintain blood ¯ ow and

prevent vascular injury caused by irritants such as

ethanol. This hypothe sis has not been con® rmed with

respect to the chemoprote ctive abilitie s of bismuth

salts. In addition, the mild irritant properties of BSS

may explain the small but signi® cant increases pro-

duced by BSS in enhanced production of ROS (Ta-

bles 1 and 2), modulation of intrace llular oxidize d

states (Table 3), and DNA fragmentation (Table 4) in

cultured normal human gastric mucosal cells. Since

bismuth does not unde rgo redox cycling, it might

produce oxidative stress by binding to sulfhydryl

groups of prote ins. The salicylic acid moiety in BSS

also may play a major role in trapping hydroxyl

groups in biological systems.

Although bismuth salts are wide ly used to prevent

gastric injury caused by dive rse stressors, and ROS

are believed to participate in many gastrointe stinal

disorde rs (1± 4), no studies have been conducte d to

determine if the cytoprote ction by bismuth salts is

associate d with their ability to scavenge ROS. Previ-

ous studie s in our laboratory (6) have demonstrated

that BSS can induce a concentration-de pendent inhi-

bition in the biochemically generated production of

superoxide anion, hydroxyl radical, and hypochlorite

radical plus hypochlorous acid in vitro. In these stud-

ies, the free radical scavengers SOD plus catalase ,

mannitol, and allopurinol inhibite d the production of

superoxide anion, hydroxyl radical, and hypochlorite

radical plus hypochlorous acid, respective ly. Thus,

these results suggest that BSS exhibite d the ability to

scavenge or quench various ROS or prevent their

formation.

In order to determine whether BSS could prevent

BAGCHI ET AL


oxidative tissue damage to gastric and intestinal mu-

cosa, mucosal tissues were incubate d with oxyge n free

radical generating systems in the presence and ab-

sence of the BSS (6). The results clearly demonstrate

that BSS induce s a concentration-de pendent inhibi-

tion of lipid peroxidation in gastric and inte stinal

mucosa that was produce d by using a superoxide

anion generating system, a hydroxyl radical generat-

ing system, and a hypochlorite radical plus hypochlo-

rous acid generating system (6). Thus, our previous

studie s demonstrated that BSS can provide protection

against free radical-induce d tissue injury in gastroin-

testinal mucosa. Furthermore , we have demonstrated

that pre- and postadministration of BSS to rats can

partially protect the gastric and intestinal mucosa

against increases induce d by 0.2 M NaOH, 0.6 M

HCl, 80% ethanol, and aspirin (200 mg/kg) in lipid

peroxidation, membrane microviscositie s, and DNA

fragmentation (6). The results of these in vitro and in

vivo studies support the hypothe sis that the gastro-

protective ability of BSS at least in part involve s the

ability of BSS to scavenge or prevent the formation of

ROS and prevent oxidative tissue damage by short-

lived reactive forms of oxygen (6). However, we have

not demonstrated the enhanced production of super-

oxide anion and hydroxyl radicals in the gastric and

inte stinal mucosa following exposure to these necro-

tizing agents (6). In a separate set of experiments, we

have demonstrated that ROS, including superoxide

anion (as determined by cytochrome c reduction) and

hydroxyl radicals, are involve d in the pathoge nesis of

the stress and gastrointe stinal injury produced by

restraint, a spicy food die t, a high-fat diet, and 40%

alcohol, as exempli® ed by the production of lipid

peroxidation, DNA damage , and altered membrane

¯ uidity (5). Exce llent corre lation coef® cients (r val-

ues) were shown to exist between ROS formation and

oxidative tissue damage . Furthe rmore, BSS-provide d

potential protection against restraint stress, spicy

food, high-fat die t, and 40% alcohol-induce d gastro-

inte stinal injury, with less protection being observed

against restraint stress (5).

The present studie s have assessed the protective

ability of bismuth subsalicylate (BSS) against necro-

tizing agent-induced oxidative injury in a primary

culture of normal human gastric mucosal cells. These

studie s demonstrated enhance d production of ROS

including superoxide anion and hydroxyl radicals in

the primary culture of normal human gastric mucosal

cells following exposure to ethanol (15% ), hydrochlo-

ric acid (150 mM), and sodium hydroxide (150 mM).

Laser scanning confocal microscopy has also demon-

strated the modulation of intrace llular oxidized states

in these gastric cells following incubation with these

necrotizing agents. Programme d cell death (apopto-

sis) is a selective process of physiological cell deletion

and is accompanie d by condensation of cytoplasm,

loss of plasma membrane microvilli, condensation

and fragmentation of nucle i, and extensive degrada-

tion of chromosomal DNA. Thus, DNA fragmenta-

tion was quantitate d and corre lated with oxidative

stress.

In summary, a nove l method was developed for

establishing a primary culture of normal gastric mu-

cosal cells from Helicobacter pylori-negative endo-

scopic biopsie s from human subjects. The results of

the present study have provide d extensive informa-

tion on the possible mechanism of BSS as a gastro-

protectant, as well as the role of ROS including

superoxide anion and hydroxyl radicals, in necrotizing

agent-induce d mucosal injury in normal human gas-

tric cells.

ACKNOWLEDGMENTS

These studies were supported in part by a grant from theNational Cancer Institute (NIH CA70976-01A 2) and agrant from the Procter & Gamble Company (Mason, Ohio).The authors thank Ms. Sheri Elliott for technical assistance.

REFERENCES

1. van der Vliet A, Bast A: Role of reactive oxygen species in

intestinal diseases. Free Radic Biol Med 12:499 ± 513, 1992

2. Rogers C, Brown A, Szabo S: Gastric mucosal protection by

new aryl sulfhydryl drugs. Dig Dis Sci 33:324 ± 329, 1988

3. Forse ll H: Gastric mucosal defense mechanisms: A brief re -

view. Scand J Gastroenterol Suppl 155:23± 28, 1988

4. D’Souza RS, Dhume VG: Gastric Cytoprotection. Indian

J Physiol Pharmacol 35:88 ± 98, 1991

5. Bagchi D, Carryl OR, Tran MX, Krohn RL, Bagchi DJ, Garg

A, Bagchi M, Mitra S, Stohs SJ: Stress, diet and alcohol-

induced oxidative gastrointestinal mucosal injury in rats and

protection by bismuth subsalicylate. J Appl Toxicol 18:3± 13,

1998

6. Bagchi D, Carryl OR, Tran MX, Bagchi M, Vuchetich PJ,

Krohn RL, Ray SD, Stohs SJ: Protection against chemically-

induced oxidative gastrointestinal tissue injury in rats by bis-

muth salts. Dig Dis Sci 42:1890 ± 1900, 1997

7. Wagner S, Be il W, Mai UEH, Bokemeye r C, Meyer HJ, Manns

MP: Interaction between Helicobacter pylori and human gastric

epithelial cells in culture: Effects of antiulcer drugs. Pharma-

cology 49:226 ± 237, 1994

8. Tabuchi Y, Sugiyama N, Horiuchi T, Furusawa M, Furuhama

K: Ebselen, a se leno-organic compound, protects against eth-

anol-induced murine gastric mucosal injury in both in vivo and

in vitro systems. Eur J Pharmacol 272:195± 201, 1995

9. Mutoh H, Hiraishi H, Ota S, Terano A, Ogura K, Ivey KJ,

Sugimoto T: Relationships between metal ions and oxyge n free



radicals in ethanol-induced damage to cultured rat gastric

mucosal ce lls. Dig Dis Sci 40:2704 ± 2711, 1995

10. Tabuchi Y, Sugiyama N, Horiuchi T, Furuhama K, Furusawa

M: Biological characterizat ion of gastric surface mucous ce ll

line GSM06 from transgenic mice harboring temperature -

sensitive simian virus 40 large T-antigen gene. Digestion

57:141± 148, 1996

11. Wallace JL, Cirino G, De Nucci G, McKnight W, MacNaugh-

ton WK: Endothelin has potent ulcerogenic and vasoconstric-

tor actions in the stomach. Amer J Physiol 256(4 Pt 1) :G661±

G666, 1989

12. Mizui T, Shimono N, Doteuchi M: A possible mechanism of

protection by polyamines against gastric damage induced by

acidi® ed ethanol in rats: Polyamine protection may depend on

its antiperoxidative properties. Jpn J Pharmacol 44:43± 50, 1987

13. Ogiu T: DNA strand breaks and death of thymocytes induced

by N-methyl-N-nitrosourea. J Cancer Res Clin Oncol 118:23±

29, 1992

14. Babior BM, Kipnes RS, Curnutte JT: The production by leu-

kocytes of superoxide : A potential bactericidal agent. J Clin

Invest 52:741± 744, 1973

15. Bagchi M, Hassoun EA, Bagchi D, Stohs SJ: Production of

reactive oxygen specie s by peritoneal macrophages and hepatic

mitochondria and microsome s from endrin-treated rats. Free

Radic Biol Med 14:149 ± 155, 1993

16. Bagchi D, Das DK, Engelman RM, Prasad MR, Subramanian

R: Polymorphonuclear leucocytes as potential source of free

radicals in the ischaemic-reperfused myocardium. Eur Heart J

11:800 ± 813, 1990

17. Tosaki A, Bagchi D, Hellegouarch A, Pali T, Cordis GA, Das

DK: Comparison of ESR and HPLC methods for the detection

of hydroxyl radicals in ischemic/repe rfused hearts. A re lation-

ship between the genesis of free radicals and reperfusion

arrhythmias. Biochem Pharmacol 45:961± 969, 1993

18. Bagchi M, Kuszynski CA, Patterson EB, Tang L, Bagchi D,

Saini T, Stohs SJ: In vitro free radical production in human oral

keratinocytes induced by smoke less tobacco extract. In Vitro

Toxicol 10:263± 274, 1997

19. Bagchi D, Tran MX, Newton S, Bagchi M, Ray SD, Kuszynski

CA, Stohs SJ: Chromium-and cadmium-induced oxidative

stress and apoptosis in cultured J774A. 1 macrophage ce lls. In

Vitro Mol Toxicol 11:171± 181, 1998

20. Tillman LA, Drake FM, Dixon JS, Wood JR: Review article:

Safe ty of bismuth in the treatment of gastrointestinal disease s.

Aliment Pharmacol Ther 10:459 ± 467, 1996

21. DuPont HL, Sullivan P, Pickering LK, Haynes G, Ackerman

PB: Symptomatic treatment of diarrhea with bismuth subsa-

licylate among students attending a Mexican university. Gas-

troenterology 73:715± 718, 1977

22. Watson AJM: Manipulation of cell deathÐ the development of

nove l strategies for the treatment of gastrointestinal disease .

Aliment Pharmacol Ther 9:215± 226, 1995

23. Bagchi D, Bhattacharya G, Stohs SJ: Production of reactive

oxygen species by gastric cells in association with Helicobacter

pylori. Free Radic Res 24:439 ± 450, 1996

24. Nordmann R: Alcohol and antioxidant systems. Alcohol Alco-

hol 29:513± 522, 1994

25. German JB: Nutritional studies of ¯ avonoids in wine. In Fla-

vonoids in Health and Disease . CA Rice-Evans, L Packe r

(eds) . New York, Marce l Dekke r, 1997, pp 343± 358

26. Rachmilewitz D, Karmeli F, Okon E, Samuni A: A novel

antiulcerogenic stable radical prevents gastric mucosal lesions

in rats. Gut 35:1181± 1188, 1994

27. Smith GS, Barre to JC, Schmidt KL, Tornwall MS, Miller TA:

Protective effect of dimethylthiourea against mucosal injury in

rat stomach. Implications for hydroxyl radical mechanism. Dig

Dis Sci 37:1345± 1355, 1992

28. Ritchey EE, Wallin JD, Shah SV: Chemiluminescence and

superoxide anion production by leukocytes from chronic he-

modialysis patients. Kidney Int 19:349 ± 358, 1981

29. Ray SD, Kame ndulis LM, Gurule MW, Yorkin RD, Corcoran

GB: Ca2 1 antagonists inhibit DNA fragme ntation and toxic

ce ll death induced by acetaminophen. FASEB J 7:453± 463,

1993

30. Shen W, Kamendulis LM, Ray SD, Corcoran GB: Acetamin-

ophen-induced cytotoxicity in cultured mouse hepatocytes: Ef-

fects of Ca2 1

-endonuclease , DNA repair, and glutathione de-

pletion inhibitors on DNA fragme ntation and cell death.

Toxicol Appl Pharmacol 112:32± 40, 1992

31. Naganuma A, Satoh M, Imura N: Speci® c reduction of toxic

side effects of adriamycin by induction of metallothionein in

mice. Jpn J Cancer Res 79:406 ± 411, 1988

32. Nakagawa I, Nishi E, Naganuma A, Imura N: Effect of prein-

duction of metallothionein synthesis on clastogenicity of anti-

cancer drugs in mice . Mutat Res 348:37± 43, 1995

33. DiSilvestro RA, Liu J, Klaassen CD: Transge nic mice overex-

pressing metallothionein are not resistant to adriamycin car-

diotoxicity. Res Commun Mol Pathol Pharmacol 93:163± 170,

1996

BAGCHI ET AL
