Development of an animal model of chronicalcohol-induced pancreatitis in the rat

HIROSHI KONO,1 MIKIO NAKAGAMI,1 IVAN RUSYN,1,2 HENRY D. CONNOR,1

BRANKO STEFANOVIC,3 DAVID A. BRENNER,3 RONALD P. MASON,2,4

GAVIN E. ARTEEL,1 AND RONALD G. THURMAN1,2

1Laboratory of Hepatobiology and Toxicology, Department of Pharmacology, 2Curriculum inToxicology, and 3Department of Medicine, University of North Carolina at Chapel Hill,Chapel Hill 27599-7365; and 4Laboratory of Pharmacology and Chemistry, National Instituteof Environmental Health Sciences, Research Triangle Park, North Carolina 27709Received 13 September 2000; accepted in final form 7 January 2001

Kono, Hiroshi, Mikio Nakagami, Ivan Rusyn, HenryD. Connor, Branko Stefanovic, David A. Brenner,Ronald P. Mason, Gavin E. Arteel, and Ronald G. Thur-man. Development of an animal model of chronic alcohol-induced pancreatitis in the rat. Am J Physiol GastrointestLiver Physiol 280: G1178–G1186, 2001.—This study wasdesigned to develop an animal model of alcoholic pancreatitisand to test the hypothesis that the dose of ethanol and thetype of dietary fat affect free radical formation and pancre-atic pathology. Female Wistar rats were fed liquid diets richin corn oil (unsaturated fat), with or without a standard orhigh dose of ethanol, and medium-chain triglycerides (satu-rated fat) with a high dose of ethanol for 8 wk enterally. Thedose of ethanol was increased as tolerance developed, whichallowed approximately twice as much alcohol to be deliveredin the high-dose group. Serum pancreatic enzymes and his-tology were normal after 4 wk of diets rich in unsaturated fat,with or without the standard dose of ethanol. In contrast,enzyme levels were elevated significantly by the high ethanoldose. Increases were blunted significantly by dietary satu-rated fat. Fibrosis and collagen a1(I) expression in the pan-creas were not detectable after 4 wk of enteral ethanolfeeding; however, they were enhanced significantly by thehigh dose after 8 wk. Furthermore, radical adducts detectedby electron spin resonance were minimal with the standarddose; however, the high dose increased carbon-centered rad-ical adducts as well as 4-hydroxynonenal, an index of lipidperoxidation, significantly. Radical adducts were alsoblunted by ;70% by dietary saturated fat. The animal modelpresented here is the first to demonstrate chronic alcohol-induced pancreatitis in a reproducible manner. The key fac-tors responsible for pathology are the amount of ethanoladministered and the type of dietary fat.

medium-chain triglycerides; fibrosis; free radical; enteralfeeding

EPIDEMIOLOGICAL EVIDENCE INDICATES that chronic alcoholabuse, but not occasional alcohol intoxication, is amajor cause of chronic pancreatitis in adult patients(26). Therapies are limited by an inadequate under-

standing of the mechanisms of pathophysiology, andprogress has been slow because an appropriate animalmodel for alcohol-induced pancreatitis is lacking (6).The pathology of chronic alcoholic pancreatitis consistsof both intralobular fibrosis involving pancreatic aciniand interlobular fibrosis with fibrotic strictures of pan-creatic ducts (8). Tsukamoto et al. (32) demonstratedthat alcohol-induced pancreatitis can be studied in anintragastric enteral feeding model in the rat. In thatstudy, enteral ethanol along with a high-fat diet causedatrophy and apoptosis in pancreatic acinar cells; how-ever, focal necrosis and fibrosis were only present in;30% of rats given ethanol chronically for 30–160days. Thus modifications of this enteral feeding proto-col could theoretically produce a useful model.

Free radicals are involved in alcohol-induced tissueinjury (17). Indeed, a-hydroxyethyl free radical wasdetected in pancreatic secretions after alcohol expo-sure, but before increases in pancreatic enzymes andpathological changes occurred (15). The type of dietaryfat is involved in free radical formation and pathogen-esis in the liver after alcohol consumption. Indeed,medium-chain triglycerides (saturated fat) blunted in-creases in endotoxin levels in portal blood, free radicalformation in the liver, and liver injury significantly inthe Tsukamoto-French enteral model (18, 23). More-over, a high-fat liquid diet (35% of total calories) rich incorn oil increased pancreatic injury in long-term en-teral feeding (32). Thus dietary unsaturated fat hasbeen associated with alcohol-induced pancreatic in-jury. Therefore, the present study was designed todevelop an animal model and, using the intragastricalcohol feeding protocol, to test the hypothesis that thedose of ethanol and the type of dietary fat affect freeradical formation and pathology in the pancreas. In thepresent study, rats received progressively increasingdoses of ethanol through careful challenge as theirtolerance developed over the 8 wk of the experimentalperiod, because pancreatic injury was minimal in the

Address for reprint requests and other correspondence: R. G.Thurman, Dept. of Pharmacology, CB#7365, Mary Ellen Jones Bldg.,Univ. of North Carolina, Chapel Hill, NC 27599-7365 (E-mail:thurman@med.unc.edu).

The costs of publication of this article were defrayed in part by thepayment of page charges. The article must therefore be herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

Am J Physiol Gastrointest Liver Physiol280: G1178–G1186, 2001.

0193-1857/01 $5.00 Copyright © 2001 the American Physiological Society http://www.ajpgi.orgG1178

Tsukamoto-French enteral protocol with the standarddose of ethanol (15).

METHODS

Diets and animals. A liquid diet described first by Thomp-son and Reitz (27), supplemented with lipotropes as de-scribed by Morimoto et al. (20), was used in this study. Itconsisted of either corn oil (unsaturated fat) or medium-chaintriglycerides (saturated fat) as a source of fat (37% of totalcalories), protein (23%), carbohydrate (5%), minerals, andvitamins plus either ethanol (35–40% of total calories) ormaltose-dextrin (control diet), as described in detail else-where (29).

Female Wistar rats (200–225 g) were fed high-fat liquiddiets (200 kcal zkg body wt21 zday21) rich in unsaturated fat,without or with ethanol (standard dose, 8–12 g zkg21 zday21;high dose, 10–18 g zkg21 zday21) or a diet rich in saturated fatalong with a high dose of ethanol continuously for up to 8 wk,using the enteral protocol developed by Tsukamoto andFrench (13, 30). In the standard dose protocol, ethanol wasinitially delivered at 8.3 g zkg21 zday21 (35% of total calories)and was increased 0.6 g/kg every 2 days until the end of thefirst week; then it was increased 0.6 g/kg every 4 days up to11.5 g zkg21 zday21 of ethanol (37% of total calories). Thisdose was continued until the end of the experimental period.On the other hand, in the high-dose protocol, ethanol wasdelivered initially at 10.2 g zkg21 zday21 (35% of total calo-ries) and was increased 0.6 g/kg every 2 days until the end ofthe first week and then 0.6 g/kg every 4 days until the end ofweek 4. During the second 4-wk period, ethanol delivery wasincreased 0.6 0.6 g/kg each week up to 17–18 g zkg21 zday21

(40% of total calories). All animals received humane care incompliance with institutional guidelines, and alcohol intoxi-cation was assessed carefully to evaluate the development oftolerance to ethanol with the use of a 0–3 scoring system(0, normal; 1, sluggish movement; 2, loss of movement butstill moving if stimulated; 3, loss of consciousness). Theethanol dose was increased progressively to challenge ani-mals based on this assessment, allowing ;1.4-fold moreethanol to be delivered to adult Wistar rats than in previousstudies by this and other laboratories (15, 22, 28).

Clinical chemistry. Ethanol concentrations in urine, whichare representative of blood ethanol levels (3), were measureddaily. Rats were housed in metabolic cages that separatedurine from feces, and urine was collected over 24 h in bottlescontaining mineral oil to prevent evaporation. Each day at9:00 AM, the urine collection bottles were changed and a 1-mlsample was stored at 220°C for later analysis. The ethanolconcentration was determined by measuring the absorbanceat 366 nm resulting from the reduction of NAD1 to NADH byalcohol dehydrogenase (4).

After 4 or 8 wk of the enteral diet, the rats were anesthe-tized with pentobarbital sodium (75 mg/kg ip) and blood wascollected via the aorta just before euthanasia. Serum wasstored at 280°C until it was assayed for amylase, lipase, andcreatine kinase with analytical kits (Sigma, St. Louis, MO).

Pathological evaluation. Pathological changes in the pan-creas were scored as described by Tsukamoto et al. (32) asfollows: steatosis (percentage of acinar cells containing fatdroplets): ,25% 5 11, ,50% 5 21, ,75% 5 31, .75% 541; inflammation and necrosis: #1 focus per low-powerview 5 11; .1 focus 5 21. The type of infiltrating inflam-matory cells was also determined morphologically in hema-toxylin and eosin-stained sections. Fibrosis was scored asfollows: thickened perivenular collagen and a few thin colla-gen septa 5 11, thin septa with incomplete bridging between

regions 5 21, thin septa and extensive bridging 5 31,thickened septa with complete bridging of regions and nod-ular appearance 5 41.

Collection of pancreatic samples. Pancreatic tissues wereformalin fixed, embedded in paraffin, and stained with he-matoxylin and eosin or trichrome. Pathology was evaluatedin a blinded manner by one of the authors. Ethanol concen-tration in the breath was analyzed by gas chromatography toverify that the levels were similar between the groups whencollection of the pancreatic secretions was initiated (14). Therats were anesthetized with pentobarbital sodium (75 mg/kg), and the pancreatic secretions were collected with amethod described previously (15), with minor modifications.Briefly, the distal bile duct was cannulated with PE-10 tub-ing, and a ligature was placed on the distal end of thepancreatic duct near the sphincter to prevent contaminationof the pancreatic juice with bile. The spin-trapping agenta-(4-pyridyl-1-oxide)-N-t-butylnitrone (POBN; Sigma) wasadministered (1 g/kg body wt) intravenously, and the pancre-atic secretions were collected into 35 ml of 0.5 mM deferox-amine mesylate (Sigma) via retrograde flow from the ligatedbile duct for 3 h to avoid ex vivo radical formation. Sampleswere stored at 280°C until analysis of free radical adducts byelectron spin resonance (ESR) spectroscopy (17).

Samples were thawed and transferred to a quartz flat cell,and ESR spectra were obtained with a Varian E-109 spec-trometer equipped with a TM110 cavity. Instrument condi-tions were as follows: 20-mW microwave power, 1.0-G mod-ulation amplitude, 80-G scan width, 16-min scan, and 1-stime constant. Spectral data were stored on an IBM-compat-ible computer and were analyzed for ESR hyperfine couplingconstants by computer simulation (10). ESR signal intensitywas determined from the amplitude of the high-field memberof the low-field doublet (second line from the left) of the ESRspectra and expressed in arbitrary units (1 unit 5 1 cm chartpaper).

Immunohistochemical detection of 4-hydroxynonenal-mod-ified proteins. Paraffin-embedded sections of pancreatic tis-sue were deparaffinized, rehydrated, and stained immuno-histochemically for the presence of an in vivo marker of lipidperoxidation, 4-hydroxynonenal protein adducts, by sequen-tial incubation with a polyclonal antibody (Alpha DiagnosticInternational, San Antonio, TX) in PBS (pH 7.4) containing1% Tween 20 and 1% bovine serum albumin. Peroxidase-linked secondary antibody and diaminobenzidine (Peroxi-dase Envision kit, DAKO, Carpinteria, CA) were used todetect specific binding. The slides were rinsed two times withPBS-0.1% Tween 20 between all incubations, and sectionswere counterstained with hematoxylin as described else-where (12). To control for nonspecific binding of the second-ary antibody, sections from the same animals were processedwithout the primary antibody, followed by the proceduredetailed above. No positive staining was observed in thiscontrol experiment (data not shown).

RNase protection assays for detection of collagen I mRNAlevels. Briefly, the template for the cRNA probe for rat colla-gen a1(I) was generated by subcloning the AvaI/PstI frag-ment of its cDNA into the in vitro transcription plasmidpGME3 (5). The antisense collagen a1(I) probe was gener-ated by digestion of this plasmid with HindIII restrictionendonuclease and transcribed with the T7 RNA polymerasein the presence of [a-32P]UTP. The antisense rat L-32 probewas generated according to the manufacturer’s instructions(Ambion, Austin, TX). Both radiolabeled probes were simul-taneously hybridized to total RNA and digested with RNaseA2 and RNase T1. Protected fragments were resolved byelectrophoresis on a 6% polyacrylamide sequencing gel and

G1179ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

quantified by PhosphoImage analysis (Molecular Dynamics,Sunnyvale, CA).

Statistics. A multiple-comparison ANOVA with Bonfer-roni’s post hoc analysis was used for the determination ofstatistical significance, as appropriate. For comparison ofpathology scores, the Mann-Whitney rank sum test was used.Before the study, a P value ,0.05 was selected as the level ofsignificance.

RESULTS

Physiological parameters. The diets were initiated 1wk after surgery to allow for full recovery, and steadygrowth rates were observed during enteral feeding.Weight gain in rats fed high-fat liquid diets with orwithout ethanol was similar to that in rats fed thechow diet [high-fat control (without ethanol), 2.0 6 0.2g/day; with standard dose of ethanol, 1.9 6 0.1; withhigh dose of ethanol, 1.9 6 0.2], indicating that the ratswere adequately nourished. Including medium-chaintriglycerides as the source of dietary fat did not affectweight gain. As previously reported (1, 22, 28), forunknown reasons, daily urine alcohol concentrationsfluctuated between 0 and 600 mg/dl in all groups stud-ied (Fig. 1). In rats fed the standard dose of ethanol,mean urine alcohol concentration was 185 6 11 mg/dl(Fig. 1A); however, the high dose of ethanol increasedthese values significantly, by ;50% (274 6 15 mg/dl;Fig. 1B). Furthermore, in the second 4-wk period, themean urine alcohol concentrations were 319 mg/dl inthe high-dose group and 246 mg/dl in the standard-dose group. There were no differences in mean urinealcohol concentrations between rats fed the high doseof ethanol with unsaturated fat and those fed the highdose with saturated fat, as expected (271 6 17 mg/dl).

In the first 4 wk, all animals were healthy andsurvived during enteral feeding; however, ;30% mor-tality was observed in the high-dose ethanol groupduring the second 4 wk. Because analysis of theseanimals was postmortem, the cause of mortality wasnot clear in all cases, although it was observed thatsome animals died because of convulsions during alco-hol withdrawal when urine alcohol levels were low.Serum creatine kinase values were not significantlyelevated as would be expected if circulatory shock oc-

curred. Furthermore, there was no effect of alcohol onaverage systemic blood pressure.

Pathophysiology. In rats fed a high-fat control dietfor 4 wk, the values of the serum pancreatic enzymesa-amylase and lipase were normal (Fig. 2). Increases inthese values were minimal in rats given the standarddose of ethanol; however, values increased signifi-cantly, approximately threefold, in rats given higherdoses of ethanol (Fig. 2). In contrast, the amylase andlipase values were blunted significantly (;50%) bydietary saturated fat in the rats given the higher dosesof ethanol (Fig. 2). After 8 wk, the amylase and lipasevalues in the animals given higher doses of ethanolwere also significantly elevated compared with those inanimals that received the standard dose of ethanol(amylase, 1,877 6 188 IU/l; lipase, 82 6 38 IU/l) or thehigh-fat control animals (amylase, 1,477 6 173 IU/l;lipase, 136 6 26 IU/l). However, amylase levels at 8 wkin animals given the higher doses of ethanol weresignificantly lower compared with the values at 4 wk inthe same group (Fig. 2A); this effect is similar to thatobserved in humans. Specifically, it is known thatserum amylase levels decrease as pancreatic exocrinefunction declines during the progression of pancreaticdamage (9).

Pancreatic histology data from adult female Wistarrats on the chow diet are shown in Fig. 3A. In rats fedthe high-fat diet with or without the standard dose ofethanol, pathological changes were minimal after 4 wk(Fig. 3, B and C). In contrast, the high dose of enteralethanol given with unsaturated fat for 4 wk causedacinar cell atrophy, fatty infiltration in acinar and isletcells (Fig. 3D), inflammatory cell infiltration (Fig. 3E),and focal necrosis (Fig. 3F) in the pancreas, resultingin the total pathology score of 4.4 6 0.5 (Table 1). The

Fig. 1. Representative plots of daily urine alcohol concentrations.Typical urine alcohol concentrations in rats fed the high-fat liquiddiet with the standard dose of ethanol (A) and the high dose ofethanol (B) and rats fed the high dose of ethanol plus medium-chaintriglycerides (MCT; C) are shown.

Fig. 2. Effect of the dose of ethanol and lipid type on levels of seruma-amylase and lipase. Blood samples were collected via the aorta at4 or 8 wk (W), and a-amylase and lipase were measured as describedin detail in METHODS. Data are means 6 SE; n 5 4–6 rats. CO, cornoil; CON, high-fat control diet; ETH, high-fat ethanol-containingdiet; STD, standard dose of ethanol; High, high dose of ethanol. *P ,0.05 compared with rats fed a high dose of ethanol with unsaturatedfat by ANOVA and Bonferroni’s post hoc test; #P , 0.05 comparedwith rats fed the standard ethanol dose.

G1180 ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

infiltrating cells were predominantly lymphocytes. Di-etary medium-chain triglycerides prevented thesepathological changes nearly completely (total pathol-ogy score, 1.0 6 0.4). Furthermore, after 8 wk ofthe high dose of ethanol, pathological changes weregreater than after 4 wk (total pathology score, 5.0 60.4; Fig. 3E).

Fibrogenesis in the pancreas after chronic enteralethanol feeding. Trichrome staining of the pancreasfrom adult female Wistar rats on the chow diet is

shown in Fig. 4A. No significant fibrotic changes weredetected after 4 wk of enteral feeding as reflected bythe fibrosis score (Table 2), confirming a previous studyfrom this laboratory (15). However, 8 wk of enteralfeeding of the high dose of ethanol significantly in-creased the fibrotic score compared with the fibroticscore of animals receiving the standard dose of alcoholor the high-fat control diet (Fig. 4E, Table 2). Underconditions in which fibrosis was detected, thesechanges tended to be focal. In addition to the scoring of

Fig. 3. Representative photomicrographs ofpancreata from rats fed different diets afterethanol treatment. A: chow. B: 4-wk high-fatcontrol diet. C: 4-wk high-fat standard doseethanol-containing diet. D: 4-wk high-fat,high ethanol-containing diet. E: 8-wk high-fat, high ethanol-containing diet. Originalmagnification, 3200. F: higher magnification(3400) view of inflammation and necrosis inrats fed the high-fat liquid diet with the highdose of ethanol for 4 wk.

Table 1. Pathological scores in the pancreas

No. of Weeks Steatosis Inflammation Necrosis Total

Control 4 0 0 0 0Standard dose 4 1.060.4 0.360.2 0.360.2 1.560.8High dose 4 2.860.5* 1.060.3* 1.060.3* 4.460.5*High dose 8 2.260.5* 1.360.3* 1.860.3* 5.060.4*High dose1MCT 4 0.560.2† 0.560.2† 0.460.1† 1.060.4†

Pathological changes in the pancreas (means6SE) were scored as described in METHODS. Control, high-fat control diet; standard dose,high-fat diet 1 standard ethanol dose; high dose, high-fat diet 1 high ethanol dose; high dose 1 MCT, high-fat [(medium-chain triglycerides(MCT)] diet 1 high ethanol dose. *P , 0.05 compared with standard-dose group; †P , 0.05 compared with high-dose group byMann-Whitney rank sum test.

G1181ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

pancreata for fibrotic changes, levels of collagen a1(I)mRNA expression in the pancreas were also deter-mined (Fig. 5). There were no detectable signals inanimals fed the chow or high-fat diets, regardless of thelength of treatment (Fig. 5). Enteral feeding with eth-anol for 4 wk also did not affect collagen a1(I) mRNAexpression (Fig. 5), as observed previously (15). How-ever, high-dose ethanol given for 8 wk caused extensiveincreases in collagen a1(I) mRNA expression.

Table 2. Fibrosis in the pancreas

Enteral Feeding 4 Wk 8 Wk

High-fat control 0 0Standard dose 0.260.1 0.560.2High dose 0.460.2 1.360.3*High dose1MCT 0.260.1 N.D.

Pancreatic fibrosis (means6SE) was scored as described inMETHODS. *P , 0.05 compared with rats fed the standard dose ofethanol by Mann-Whitney rank sum test. N.D., not determined.

Fig. 4. Trichrome staining in the pancreasafter chronic enteral ethanol feeding. Originalmagnification, 3200. Representative photo-micrographs show histology of the pancreasfrom rats given high-fat control (chow; A) andhigh-fat ethanol-containing (B–E) diets.

Fig. 5. Effect of chronic ethanol and lipid type on collagen a1(I)mRNA expression in pancreatic tissue. Pancreatic tissues were as-sayed for fibrillar collagen a1(I) mRNA using an RNase protectionassay with L32 as the housekeeping gene. Representative gels areshown. EtOH, ethanol.

G1182 ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

Radical adducts in pancreatic secretions. Radicaladducts were barely detectable in pancreatic secretionsfrom rats fed either the chow or the high-fat controldiet without ethanol (data not shown). Moreover,POBN radical adducts in pancreatic secretions fromrats given the standard dose of ethanol were minimalafter 4 wk (Fig. 6). In contrast, the high dose of ethanolincreased radical adducts approximately twofold. Fur-thermore, dietary saturated fat in the form of medium-chain triglycerides blunted this increase significantly.Computer simulation of these spectra identified thePOBN-a-hydroxyethyl radical adduct (17). The aver-age radical intensity was approximately fourfoldgreater in pancreatic secretions from rats fed the highdose than those fed the standard dose of ethanol (Fig.7). Furthermore, this increase was blunted signifi-cantly (;70%) by dietary saturated fat.

Immunohistochemical detection of 4-hydroxynonenalmodified proteins. To test whether ethanol administra-tion causes lipid peroxidation in the pancreas, 4-hy-droxynonenal-modified proteins were detected by im-munohistochemistry (24). Despite increases in freeradical formation (Fig. 7), no detectable increases in 4-hydroxynonenal protein adducts were detected after

4 wk of enteral feeding in any group (data not shown).After 8 wk of enteral feeding, significant accumulationof 4-hydroxynonenal (brown staining) was observed inthe pancreas of rats fed high-dose ethanol but not inrats fed the chow or high-fat control diets (Fig. 8) or inrats given the standard dose of ethanol (data notshown).

DISCUSSION

Development of a new animal model to study alcohol-induced pancreatic injury. The mechanism of chronicalcohol-induced pancreatitis remains unclear, and amajor problem has been the lack of an appropriateanimal model for study. Tsukamoto et al. (32) previ-ously reported that feeding a high-fat diet with thestandard dose of ethanol caused atrophy and apoptosisin pancreatic acinar cells; however, focal necrosis andfibrosis were only present in ;30% of the rats fedenteral ethanol for 30–160 days. Because the severityof clinical alcohol-induced pancreatitis and fibrosis isrelated to dose- and time-dependent alcohol consump-tion in humans (11), here, the chronic intragastricenteral feeding protocol was modified to increase eth-anol delivery. This goal was achieved by increasing thealcohol levels in the diet based on behavioral assess-ment of the development of tolerance to ethanol withthe use of the 0–3 score system. In this manner, alcoholdelivery in female Wistar rats in the 200- to 225-gweight range could be increased to 17–18 g zkg21 zday21. Alcohol delivery was increased ;1.4-fold overprevious studies from this and other laboratories withrats in the same weight range (Refs. 13, 15, 31; Fig. 2).As a result, pancreatic enzymes (amylase and lipase)and pathology scores were increased after 8 wk of thehigh dose of ethanol (Table 1, Figs. 2 and 3). Further-more, fibrosis and collagen a1(I) mRNA expressionincreased significantly after 8 wk of the high dose ofethanol in nearly all animals (Table 2, Figs. 4 and 5).

Fig. 6. Effect of the dose of ethanol and lipid type on electron spinresonance (ESR) spectra in pancreatic secretions. After 4 wk ofenteral feeding, the diet was discontinued immediately before exper-iments. After injection of the spin-trapping agent a-(4-pyridyl-1-oxide)-N-t-butylnitrone (POBN; 1 g/kg body wt iv), pancreatic secre-tions were collected as described in detail in METHODS. Computersimulation identified the species as the a-hydroxyethyl radical ad-duct (coupling constants: aN 5 15.8 G, ab

H 5 2.3 G). RepresentativeESR spectra are shown.

Fig. 7. Effect of the dose of ethanol and lipid type on average radicaladduct signal intensity from pancreatic secretions. Conditions werethe same as those depicted in Fig. 6. ESR signal intensity wasquantitated as the double integral of peaks from pancreatic secre-tions. Data are means 6 SE; n 5 6 rats. *P , 0.01 compared withrats fed CO with EtOH, #P , 0.05 compared with the high-dosegroup by ANOVA and Bonferroni’s post hoc test.

G1183ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

Thus it is concluded that the total amount of alcoholconsumed is a critical factor in producing chronic alco-hol-induced pancreatic injury. Furthermore, the ani-mal model presented here is the first reproducibledemonstration of chronic alcohol-induced pancreatitisin the rat.

Gender is also an important factor in alcohol-inducedtissue injury. Indeed, alcohol-induced liver injury isgreater in females than in males in the enteral alcoholmodel (16). The present study differs from previouswork (32) in that female rats were used, and gendermay therefore play a role in the observed pancreaticdamage; however, the damage resulting from the use of

the alcohol dosing regimen described in the METHODS

section also occurs in males (Kono et al., unpublishedobservations), consistent with the idea that gendercannot completely explain the greater pathology ob-served in this study.

Dietary fat and alcohol-induced pancreatic injury.Dietary fat is also an important factor in alcohol-in-duced tissue injury (21). Indeed, dietary medium-chaintriglycerides reduced lipid peroxidation in the enteralmodel (23). Furthermore, dietary medium-chain tri-glycerides prevented free radical formation and earlyalcohol-induced liver injury in the Tsukamoto-Frenchenteral model (18). In the present study, dietary me-dium-chain triglycerides also blunted free radical for-mation and tissue injury in the pancreas (Figs. 6 and7). Dietary medium-chain triglycerides significantlyblunted the increase in plasma endotoxin levels afterenteral ethanol, the response of Kupffer cells to endo-toxin, and liver injury (18). Therefore, medium-chaintriglycerides could affect activation of tissue macro-phages or other critical cell types in the pancreas. Thusdietary medium-chain triglycerides prevented pancre-atic injury, most likely by inhibition of free radicalformation. Together, these data indicate that dietaryfat is also an important factor in the pathogenesis ofchronic alcohol-induced pancreatic injury.

Fig. 9. Working hypothesis. Ethanol changes gut permeability andmicroflora during enteral feeding, and the high dose of ethanol couldincrease endotoxin levels. Increased endotoxin most likely increasesinfiltration of leukocytes such as macrophages and neutrophils,which produce free radicals in the pancreas. One key factor incausing pathology seems to be the total amount of ethanol consumedchronically.

Fig. 8. Chronic ethanol treatment increases 4-hydroxynonenal inthe pancreas. Sections of pancreatic tissue from untreated rats (A)and rats treated for 8 wk with high-fat (B) or high-dose ethanol-containing (C) liquid diets were stained immunohistochemically for4-hydroxynonenal-modified proteins (brown staining) as detailed inMETHODS. C: pathology caused by alcohol, including acinar atrophy,necrotic foci, and increased stroma around acini. Original magnifi-cation, 3200. Representative photomicrographs from 4 animals/group.

G1184 ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

What is the mechanism of damage caused by alcoholin the pancreas ? Figure 9 is a schematic depiction ofour working hypothesis on the mechanism by whichalcohol damages the pancreas. In the present study,enteral feeding with a higher delivery of ethanol in-creased formation of both carbon-centered free radicalsand 4-hydroxynonenal-modified proteins in the pan-creas (Figs. 6–8). These data are consistent with thehypothesis that oxidative stress is involved in thepathogenesis of early alcohol-induced pancreatic injury(see Fig. 9).

It is also proposed that the infiltration of inflamma-tory cells plays a key role in the progression of pancre-atic damage (Fig. 9). Infiltrating leukocytes such aslymphocytes and macrophages can be activated by gut-derived endotoxin and/or other stimuli to produce oxi-dants after alcohol consumption (Fig. 9). Indeed, 4 wkof high-dose ethanol increased endotoxin levels in por-tal blood (e.g., 149 6 17 pg/ml), and free radical forma-tion (Figs. 6 and 7) was ;2.5-fold higher than in aprevious study (16) performed with the standard eth-anol protocol. Furthermore, inflammatory foci in-creased significantly in the high-dose group comparedwith those in the standard group in this study (Table1). There are already links in the literature betweenacute pancreatitis, endotoxins, and inflammatory cells.For example, endotoxemia predicts outcome in acutepancreatitis in humans (34), and gut sterilization withantibiotics protects against acute pancreatitis in rats(19). Inflammatory cell infiltration is also known tocorrelate well with the prognosis of pancreatitis inhumans and in animal models (2, 7, 25, 33). The ques-tion remains whether these mechanisms are also im-portant during the chronic pancreatitis caused by al-cohol, and this question will be the focus of futureresearch.

Clinical implications. The Tsukamoto-French en-teral alcohol model has a strong nutritional compo-nent, with liver injury dependent on unsaturated fat inthe diet (21). In this study, pancreatic injury caused bythe high dose of ethanol was prevented by medium-chain triglycerides (Table 1). Importantly, in thisstudy, increases in fibrosis were observed only after 8wk (Figs. 4 and 5), leading to the conclusion thatchronic alcohol-induced pancreatic injury is dependenton the total amount of alcohol consumed and the typeof dietary fat. Furthermore, this is the first demonstra-tion of reproducible chronic alcohol-induced pancreati-tis in a relatively short experimental period. Thus thisanimal model may be useful for the study of mecha-nisms of chronic alcohol-induced pancreatitis and fordeveloping useful therapeutic strategies.

This study was supported, in part, by grants from the NationalInstitute on Alcohol Abuse and Alcoholism.

REFERENCES

1. Adachi Y, Bradford BU, Gao W, Bojes HK, and ThurmanRG. Inactivation of Kupffer cells prevents early alcohol-inducedliver injury. Hepatology 20: 453–460, 1994.

2. Andersson R, Deng XM, and Wang XD. Role of macrophageoveractivation in the development of acute pancreatic injury inrats. Br J Surg 84: 775–780, 1997.

3. Badger TM, Crouch J, Irby D, and Shahare M. Episodicexcretion of ethanol during chronic intragastric ethanol infusionin the male rat: continuous vs. cyclic ethanol and nutrient infu-sions. J Pharmacol Exp Therapeut 264: 938–943, 1993.

4. Bergmeyer HU. Methods of Enzymatic Analysis. New York:Academic, 1988.

5. Brenner DA, Veloz L, Jaenisch R, and Alcorn JM. Stimu-lation of the collagen a1(I) endogenous gene and transgene incarbon tetrachloride-induced hepatic fibrosis. Hepatology 17:287–292, 1993.

6. Buckelew DQ and Schenker S. Pathogenesis of alcohol pan-creatitis—a peek into a black box. Alcohol Clin Exp Res 22:550–552, 1998.

7. Burkart V and Kolb H. Macrophages in islet destruction inautoimmune diabetes mellitus. Immunobiology 195: 601–613,1996.

8. De Angelis C, Valente G, Spaccapietra M, Angonese C,DelFavero G, Naccarato R, and Andriulli A. Histologicalstudy of alcoholic, nonalcoholic and obstructive chronic pancre-atitis. Pancreas 7: 193–196, 1992.

9. Dominguez-Munoz JE, Pieramico O, Buchler M, and Mal-fertheiner P. Ratios of different serum pancreatic enzymes inthe diagnosis and staging of chronic pancreatitis. Digestion 54:231–236, 1993.

10. Duling DR. Simulation on multiple isotropic spin-trap EPRspectra. J Magnetic Resonance 104: 105–110, 1994.

11. Durkec JP and Sarles H. Multicenter survey of the etiology ofpancreatic disease: the relationship between the relative risk ofdeveloping chronic panceratitis and alcohol, protein and lipidconsumption. Digestion 18: 337–350, 1978.

12. Eldridge SR, Tilbury LF, Goldsworthy TL, and Butter-worth BE. Measurement of chemically induced cell prolifera-tion in rodent liver and kidney: comparison of 5-bromo-29-de-oxyuridine and [3H]thymidine adminstered by injection orosmotic pump. Carcinogenesis 12: 1557–1561, 1991.

13. French SW, Miyamoto K, and Tsukamoto H. Ethanol-in-duced hepatic fibrosis in the rat: role of the amount of dietary fat.Alcohol Clin Exp Res 10: 13S–19S, 1986.

14. Glassman EB, McLaughlin GA, Forman DT, Felder MR,and Thurman RG. Role of alcohol dehydrogenase in the swiftincrease in alcohol metabolism (SIAM). Studies with deer micedeficient in alcohol dehydrogenase. Biochem Pharmacol 34:3523–3526, 1989.

15. Iimuro Y, Bradford BU, Gao W, Kadiiska M, Mason RP,Stefanovic B, Brenner DA, and Thurman RG. Detection ofa-hydroxyethyl free radical adducts in the pancreas after chronicexposure to alcohol in the rat. Mol Pharmacol 50: 656–661, 1996.

16. Iimuro Y, Frankenberg Mv, Arteel GE, Bradford BU, WallCA, and Thurman RG. Female rats exhibit greater suscepti-bility to early alcohol-induced injury than males. Am J PhysiolGastrointest Liver Physiol 272: G1186–G1194, 1997.

17. Knecht KT, Adachi Y, Bradford BU, Iimuro Y, Kadiiska M,Qun-hui X, and Thurman RG. Free radical adducts in the bileof rats treated chronically with intragastric alcohol: inhibition bydestruction of Kupffer cells. Mol Pharmacol 47: 1028–1034,1995.

18. Kono H, Enomoto N, Connor HD, Wheeler, MD, BradfordBU, Rivera CA, Kadiiska M, Mason RP, and Thurman RG.Medium-chain triglycerides inhibit free radical formation andTNF-a production in rats given enteral ethanol. Am J PhysiolGastrointest Liver Physiol 278: G467–G476, 2000.

19. Marotta F, Geng TC, Wu CC, and Barbi G. Bacterial trans-location in the course of acute pancreatitis: beneficial role ofnonabsorbable antibiotics and lactitol enemas. Digestion 57:446–452, 1996.

20. Morimoto M, Zern MA, Hagbjork AL, Ingelman-SundbergM, and French SW. Fish oil, alcohol, and liver pathology: role ofcytochrome P4502E1. Proc Soc Exp Biol Med 207: 197–205,1994.

G1185ALCOHOL-INDUCED PANCREATIC INJURY IN RATS

21. Nanji AA, Mendenhall CL, and French SW. Beef fat preventsalcoholic liver disease in the rat. Alcohol Clin Exp Res 13: 15–19,1989.

22. Nanji AA, Rahemtulla A, Maio L, Khwaja S, Zhao S, TahanSR, and Thomas P. Alterations in thromboxane synthase andthromboxane A2 receptors in experimental alcoholic liver dis-ease. J Pharmacol Exp Therapeut 282: 1037–1043, 1997.

23. Nanji AA, Zakim D, Rahemtulla A, Daly T, Miao L, ZhaoS, Khwaja S, Tahan SR, and Dannenberg AJ. Dietarysaturated fatty acids down-regulate cyclooxygenase-2 and tu-mor necrosis factor-alpha and reverse fibrosis in alcohol-in-duced liver disease in the rat. Hepatology 26: 1538–1545,1997.

24. Paradis V, Mathurin P, Kollinger M, Imbert-Bismut F,Charlotte F, Piton A, Opolon P, Holstege A, Poynard T,and Bedossa P. In situ detection of lipid peroxidation in chronichepatitis C: correlation with pathological features. J Clin Pathol50: 401–406, 1997.

25. Poch B, Gansauge F, Rau B, Wittel U, Gansauge S,Nussler AK, Schoenberg M, and Beger HG. The role ofpolymorphonuclear leukocytes and oxygen-derived free radi-cals in experimental acute pancreatitis: mediators of localdestruction and activators of inflammation. FEBS Lett 461:268–272, 1999.

26. Singh M and Simsek H. Ethanol and the pancreas: currentstatus. Gastroenterology 98: 151–162, 1990.

27. Thompson JA and Reitz RC. Effects of ethanol ingestion anddietary fat levels on mitochondrial lipids in male and femalerats. Lipid 13: 540–550, 1978.

28. Tsukamoto H, French SW, Reidelberger RD, and LargmanC. Cyclical pattern of blood alcohol levels during continuousintragastric ethanol infusion in rats. Alcohol Clin Exp Res 9:31–37, 1985.

29. Tsukamoto H, Gaal K, and French SW. Insights into thepathogenesis of alcoholic liver necrosis and fibrosis: status re-port. Hepatology 12: 599–608, 1990.

30. Tsukamoto H, Reiderberger RD, French SW, and Larg-man C. Long-term cannulation model for blood sampling andintragastric infusion in the rat. Am J Physiol Regulatory Inte-grative Comp Physiol 247: R595–R599, 1984.

31. Tsukamoto H, Towner SJ, Ciofalo LM, and French SW.Ethanol-induced liver fibrosis in rats fed high fat diet. Hepatol-ogy 6: 814–822, 1986.

32. Tsukamoto H, Towner SJ, Yu GSM, and French SW. Poten-tiation of ethanol-induced pancreatic injury by dietary fat. Am JPathol 131: 246–257, 1988.

33. Wang X, Andersson R, Soltesz V, Leveau P, and Ihse I. Gutorigin sepsis, macrophage function, and oxygen extraction asso-ciated with acute pancreatitis in the rat. World J Surg 20:299–307, 1996.

34. Wig JD, Kochhar R, Ray JD, Krishna Rao DV, Gupta NM,and Ganguly NK. Endotoxemia predicts outcome in acute pan-creatitis. J Clin Gastroenterol 26: 121–124, 1998.

G1186 ALCOHOL-INDUCED PANCREATIC INJURY IN RATS