Biliary fibrosis associated with altered bile composition in a mouse model of erythropoietic...

Biliary Fibrosis Associated With Altered Bile Compositionin a Mouse Model of Erythropoietic Protoporphyria

LEO MEERMAN,*,‡ NYNKE R. KOOPEN,*,§ VINCENT BLOKS,*,§ HARRY VAN GOOR,*,00

RICK HAVINGA,*,§ BERT G. WOLTHERS,*,¶ WERNER KRAMER,# SIEGFRIED STENGELIN,#

MICHAEL MULLER,*,‡ FOLKERT KUIPERS,*,§ and PETER L. M. JANSEN*,‡

*Center for Liver, Digestive, and Metabolic Diseases, Groningen Institute for Drug Studies, Groningen, The Netherlands; ‡Division ofGastroenterology and Hepatology, Department of Internal Medicine, §Department of Pediatrics, 00Department of Pathology, and ¶Central ClinicalLaboratory, University Hospital Groningen, Groningen, The Netherlands; and #Hoechst Marion Roussel, Frankfurt am Main, Germany

Background & Aims: Reduced activity of ferrochelatasein erythropoietic protoporphyria (EPP) results in proto-porphyrin (PP) accumulation in erythrocytes and liver.Liver disease may occur in patients with EPP, some ofwhom develop progressive liver failure that necessi-tates transplantation. We investigated the mecha-nisms underlying EPP-associated liver disease in amouse model of EPP. Methods: Liver histology, indica-tors of lipid peroxidation, plasma parameters of liverfunction, and bile composition were studied in micehomozygous (fch/fch) for a point mutation in theferrochelatase gene and in heterozygous (fch/1) andwild-type (1/1) mice. Results: Microscopic examina-tion showed bile duct proliferation and biliary fibrosiswith portoportal bridging in fch/fch mice. PP contentwas 130-fold increased, and thiobarbituric acid–reactive substances (130%) and conjugated dienes(175%) were slightly higher in fch/fch than in fch/1and 1/1 livers. Levels of hepatic thiols (212%) andiron (252%) were reduced in fch/fch livers. Liverenzymes and plasma bilirubin were markedly increasedin the homozygotes. Plasma bile salt levels were 80times higher in fch/fch than in fch/1 and 1/1 mice,probably related to the absence of the Na1-taurocho-late cotransporting protein (Ntcp) in fch/fch liver. Para-doxically, bile flow was not impaired and biliary bile saltsecretion was 4 times higher in fch/fch mice than incontrols. Up-regulation of the intestinal Na1-depen-dent bile salt transport system in fch/fch mice mayenhance efficiency of bile salt reabsorption. The bilesalt/lipid ratio and PP content of fch/fch bile wereincreased 2-fold and 85-fold, respectively, comparedwith 1/1, whereas biliary glutathione was reduced by90%. Similar effects on bile formation were caused bygriseofulvin-induced inhibition of ferrochelatase activ-ity in control mice. Conclusions: Bile formation isstrongly affected in mice with impaired ferrochelataseactivity. Rather than peroxidative processes, formationof cytotoxic bile with high concentrations of bile saltsand PP may cause biliary fibrosis in fch/fch mice bydamaging bile duct epithelium.

Erythropoietic protoporphyria (EPP), an inheriteddisorder of heme synthesis, is caused by deficiency of

the mitochondrial enzyme ferrochelatase (EC 4.99.1.1)responsible for insertion of iron into protoporphyrin(PP).1 Reduced ferrochelatase activity in humans resultsin increased PP concentrations in erythrocytes, blood,liver, and feces.1 Because of its hydrophobic nature, PPcan only be removed from the body through secretioninto bile followed by fecal elimination. Transport acrossthe canalicular membrane appears to be rate-limiting inbiliary PP secretion,2 but the transport systems involvedhave not been elucidated. It has been shown that PPsecretion in experimental animals is stimulated by bilesalts.3,4 Recent studies from our laboratory showed thatPP associates with biliary lipids and that the presence ofcholesterol and phospholipids in bile is essential forefficient PP secretion.5 Hence, stimulation of lipidsecretion may explain the reported stimulatory effects ofbile salts on biliary PP secretion. When production of PPexceeds its utilization in heme synthesis and its maximalbiliary excretion rate, progressive accumulation of PP inliver, erythrocytes, and skin will occur.

Abnormal liver function test values are frequentlyobserved in patients with EPP, but the hepatic manifesta-tions of the disease are diverse.1,6 Approximately 40different mutations in the human ferrochelatase genehave been described so far.7,8 It seems that only severedeficiencies in ferrochelatase activity predispose individu-als to the development of liver disease for which livertransplantation is indicated.6,8 Recently, Bloomer et al.8

Abbreviations used in this paper: CYP7A, gene encoding choles-terol 7a-hydroxylase; CYP27, gene encoding sterol 27-hydroxylase;EPP, erythropoietic protoporphyria; fch, disabling mutation in murineferrochelatase gene; ibst, Na1-dependent ileal bile salt–transportingprotein; mdr2, murine gene encoding mdr2 P-glycoprotein; Ntcp,Na1-dependent taurocholate cotransporting protein; PP, protopor-phyrin.

r 1999 by the American Gastroenterological Association0016-5085/99/$10.00

GASTROENTEROLOGY 1999;117:696–705

reported genetic heterogeneity in 8 unrelated patientswith EPP who received liver transplants; all of themutations found caused a major structural alteration inthe ferrochelatase protein and drastically reduced enzymeactivity. Liver fibrosis progressing to cirrhosis develops ina considerable number of patients with EPP, e.g., in 7 of55 patients in a 20-year follow-up study by Doss andFrank.9 Biliary fibrosis has also been described in a mousemodel of EPP10 in which the homozygous presence of apoint mutation in the ferrochelatase gene ( fch/fch) leadsto a reduction of enzyme activity by 95%. Developmentof liver damage in EPP has been ascribed to toxic actionsof PP because of its peroxidative properties and to theprecipitation of (poorly soluble) PP in bile canaliculi.6

Although plasma bilirubin and liver enzyme levels areincreased in fch/fch mice,10 it is not known to what extentthe bile formation process is actually affected in theseanimals. This is important because recent studies in mdr2P-glycoprotein–deficient mice have shown that the forma-tion of cytotoxic bile per se (without cholestasis) can leadto severe biliary fibrosis. In this particular animal model,in which phospholipid/cholesterol-free bile is produced,the degree of liver pathology is related to the hydropho-bicity of the bile salt pool.11–13 Because insight in themechanism(s) leading to development of fibrosis isessential for prevention and treatment of this complica-tion of EPP, we have investigated liver morphology inrelation to indicators of hepatic lipid peroxidation as wellas bile composition in ferrochelatase-deficient fch/fchmice.

Materials and Methods

MiceMale and female fch/fch and fch/1 mice with a BALB/c

background were generously supplied by Dr. Xavier Montagu-telli (Institut Pasteur, Paris, France) for establishing a breedingprogram at the Central Animal Laboratory, Faculty of MedicalSciences, University of Groningen. The animals were kept in atemperature- and light-controlled environment and were pro-tected from direct light. Homozygous and heterozygous off-spring were individually phenotyped by plasma PP andbilirubin analysis. Age-matched control BALB/c mice werepurchased from Harlan BV (Zeist, The Netherlands). Controlanimals were kept in the same environment and under the samead libitum dietary conditions (SRM-A; Hope Farms BV,Woerden, The Netherlands) for at least 3 weeks beforeexperiments. Animals received humane care according to localguidelines; experimental procedures were approved by the localEthical Committee for Animal Experiments.

Experimental Procedures

Only male mice aged 12–14 weeks were used. Animalswere anesthetized by intraperitoneal (IP) injection of Hypnorm

(Janssen Cilag, Beerse, Belgium) (fentanyl/fluanisone) anddiazepam, and their gallbladders were cannulated for collectionof bile as described previously.14 During surgery, livers wereprotected from UV light by means of a filter. During the1-hour bile collection period, animals were placed in ahumidified incubator to ensure maintenance of body tempera-ture. After bile collection, a large blood sample (0.8–1.0 mL)was collected by cardiac puncture and immediately transferredto an EDTA-containing tube for separation of erythrocytes andplasma. Erythrocytes were washed with phosphate-bufferedsaline and stored for porphyrin analysis at –20°C untilmeasurements. Livers were excised and weighed. Parts of thelivers were stored in paraformaldehyde or frozen in liquidisopentane for microscopic evaluation. The remainders weresnap-frozen in liquid nitrogen for biochemical analyses, isola-tion of a liver membrane fraction, and RNA isolation. Thesmall intestines of the mice were flushed with cold saline andfrozen in liquid nitrogen for preparation of mucosal scrapings.

For determination of fecal and urinary bile salt output,fch/fch and 1/1 mice (n 5 4 in both groups) were housed inindividual metabolic cages for 3 days. Feces and urine werecollected daily, and the quantities were determined by weigh-ing (the feces were weighed after freeze-drying). The 3 dailycollections for each individual mouse were then pooled andcarefully mixed before bile salt measurements (see below).

The distribution of intravenously administered radiolabeledtaurocholate over plasma, hepatic, and biliary compartmentswas determined in female 1/1 and fch/fch mice. For thispurpose, a tracer amount of [3H(G)]taurocholate (3.47 Ci/mmol; DuPont/New England Nuclear, Boston, MA) wasinjected via the left jugular vein in anesthetized mice equippedwith a gallbladder catheter. At 30 minutes after injection,blood was obtained by heart puncture and the liver was excisedand weighed. Subsequently, radioactivity was determined inplasma, liver homogenates, and bile by scintillation counting(Packard 1500 Scintillation Counter; Packard Instruments,Downers Grove, IL).

Bile, blood, and liver collection procedures were also appliedto male Swiss mice (Harlan BV) aged 12–14 weeks fed eitherthe control chow diet or the same diet to which griseofulvin(2.5%, wt/wt) was added for a period of 2.5 weeks.15

Microscopic Evaluation

Liver tissue was fixed in 4% paraformaldehyde (vol/vol)and embedded in paraffin. Sections were stained with hematoxy-lin or Masson stain before evaluation.

Analytical Procedures

Concentrations of plasma alkaline phosphatase, alanineaminotransferase (ALT), aspartate aminotransferase (AST), andbilirubin were determined by routine clinical procedures.Plasma bile salts were measured enzymatically.16 PP andcoproporphyrins I and III in erythrocytes, liver tissue, and bilewere determined by high-performance liquid chromatography,as described previously.5 Levels of conjugated dienes,17 thiobar-bituric acid–reactive substances,18 and thiols19 in liver homog-

September 1999 BILIARY FIBROSIS IN MICE WITH EPP 697

enates were measured according to published methods. Totalbiliary bile salts, cholesterol, and phospholipids were mea-sured,14 the latter 2 after Bligh and Dyer extraction.20 Biliary,fecal, and urinary bile salt composition were determined bycapillary gas chromatography14; the identity of some minormetabolites present in mouse bile was verified by gas chroma-tography and mass spectrometry on a Finnegan-MAT SSQcombination (Finnegan MAT, Bremen, Germany) and bycomparison to reported spectra. Hepatic and biliary ironcontents were determined by atomic absorption spectrometry.

Western Blotting

A crude membrane fraction and a plasma membranefraction were prepared from pooled mouse livers (n 5 5–6 pergroup) and characterized as described by Wolters et al.21 forassessment of Na1-taurocholate cotransporting protein (Ntcp),organic anion transporting polypeptide 1 (oatp-1), and Na1,K1-adenosine triphosphatase (ATPase), a-subunit, levels by West-ern analysis.22 The anti-Ntcp and anti–oatp-1 immunoglobu-lin G (IgG) K4 and K10, respectively, were kindly provided byDrs. Bruno Stieger and Peter Meier-Abt, Zurich, Switzerland.Antibodies for detection of Na1,K1-ATPase were a kind giftfrom Dr. Wilbert Peters (Nijmegen, The Netherlands.23)

Antisera (RIBMAL1) against a 53-residue carboxy-terminalportion of the rat intestinal Na1-dependent ileal bile salttransporter (ibst), fused at its amino terminus to an Escherichiacoli maltose–binding protein moiety, were raised in guineapigs. These antisera were used to assess ibst levels in homoge-nates and crude membrane fractions prepared from intestinalscrapings. For the latter purpose, the entire small intestine of1/1 and fch/fch mice was divided into 4 equal parts andopened longitudinally, and mucosae were scraped using a glassslide. Homogenates were prepared after 10-fold dilution(wt/wt) with NaHCO3 (1 mmol/L, pH 7.5) containing 17 mg/Lphenylmethylsulfonyl fluoride, and a crude membrane prepara-tion was isolated by centrifugation at 45,000 rpm and 4°C for60 minutes in an Optima TLX tabletop ultracentrifuge(Beckman Instruments, Palo Alto, CA). Pellets were resus-pended in 250 mmol/L sucrose in 10 mmol/L Tris-HCl buffer(pH 7.5) with the protease inhibitor cocktail (Complete;Boehringer Mannheim, Mannheim, Germany, 1 tablet/50 mL)by passing 20 times through a 26.6-gauge needle. Thehomogenates of intestinal mucosa and crude membranes(50–75 µg protein/lane) were loaded and run on a 4%–20%polyacrylamide gel. Subsequently, the proteins were transferredto nitrocellulose (Amersham, Little Chalfont, England) andimmunoblotting was performed with the antisera at a 1:10.000dilution. Immunodetection was carried out using a rabbitanti-guinea pig IgG antibody conjugated to horseradish peroxi-dase (Sigma Chemical Co., St. Louis, MO) at a 1:2.000dilution. Detection was by the ECL Western blotting kit(Amersham) according to the manufacturer’s instructions.

Immunohistochemistry

Localization of ibst was studied by immunohistochem-istry on frozen sections (10 µm) of 1/1 and fch/fch livers.

Sections were fixed with 100% acetone for 10 minutes andair-dried. Blocking was performed with 5% rabbit serum, andthe sections were then incubated with a 1:50 dilution of theantibodies described above. Immunodetection was carried outusing a rabbit anti-guinea pig antibody conjugated to horserad-ish peroxidase (1:400 dilution).

Northern Blotting

Total RNA was isolated from livers according to themethod of Chomczynski and Sacchi.24 Steady-state messengerRNA (mRNA) levels for CYP7A and CYP27 were determinedby using Northern blotting, with hybridization conditionsexactly as described.25 28S ribosomal RNA was used as aninternal standard to correct for differences in amounts of totalRNA. The mRNA levels were quantified by PhosphoImager(Fuji Fujix Bas 1000; Fuji, Tokyo, Japan) using the programTINA, version 2.08c.

Statistics

Comparison between data from 1/1, fch/1, and fch/fchmice was done by analysis of variance (ANOVA) and post hocNewman–Keuls t test. Differences between control and griseo-fulvin-treated animals were tested for significance by theMann–Whitney U test. A P value of ,0.05 was consideredsignificant.

Results

Liver Characteristics and Morphology

The liver weight of fch/fch mice was greater thanthat of fch/1 and 1/1 mice, which, in combination withan ,30% lower body weight, resulted in a significantincrease of liver to body weight ratio in fch/fch comparedwith fch/1 and 1/1 mice: 0.118 6 0.021, 0.063 60.010, and 0.046 6 0.004, respectively.

Microscopic examination of liver sections (Figure 1)showed massive bile duct proliferation and biliary fibrosisin fch/fch mice. Portoportal bridging was frequentlyobserved in these animals. PP deposits were present insmall bile ducts of fch/fch livers. Except for the presence ofabundant fat droplets, mainly in perivenous hepatocytes,no morphological abnormalities were detected in livers offch/1 mice (not shown).

Hepatic PP content was 130-fold increased in fch/fchmice compared with fch/1 and 1/1 animals, as de-scribed previously,10 whereas coproporphyrin I and IIIcontents were 17-fold and 25-fold higher than controlvalues, respectively (Table 1). No significant differencesbetween fch/1 and 1/1 mice in this respect were noted.Contents of hepatic conjugated dienes and thiobarbituricacid–reactive substances were slightly but significantlyhigher in fch/fch than in fch/1 and 1/1 mice, whereastotal thiol content was somewhat lower. Hepatic iron

698 MEERMAN ET AL. GASTROENTEROLOGY Vol. 117, No. 3

content was reduced in both fch/fch and fch/1 micecompared with 1/1 mice.

Blood Analyses

PP contents of washed red blood cells were 21.0 68.7, 2.0 6 0.2, and 1.7 6 0.1 µmol/L for fch/fch, fch/1,and 1/1 animals, respectively. The difference betweenfch/fch mice and fch/1 and 1/1 mice is highly signifi-cant. These values are comparable to those reportedpreviously.10

Table 2 shows that alkaline phosphatase, ALT, andAST levels in plasma of fch/fch mice were increased incomparison to fch/1 and 1/1 mice, as were conjugatedbilirubin and bile salt levels. Plasma bile salt concentra-tions up to 4700 µmol/L were measured in the fch/fchmice, that is, values exceeding those observed after bileduct ligation in mice (900–1400 µmol/L after 14 days; R.Ottenhof, Amsterdam, personal communication, July1998).

Bile Formation

Table 3 summarizes data on bile formation in the3 strains of mice. In contrast to what was expected on thebasis of plasma analysis, bile formation was not impairedin fch/fch mice. When expressed relative to body weight,

Figure 2. Western blots. (A) Ntcp levels in total liver membranefractions and plasma membrane fractions of wild-type mice (lanes 1and 4, respectively), fch/1 mice (lanes 2 and 5), and fch/fch mice(lanes 3 and 6). Note the absence of Ntcp in fch/fch mice. (B) Oatp-1levels in total liver membrane fractions and plasma membranefractions of wild-type mice (lanes 1 and 4, respectively), heterozygousfch/1 mice (lanes 2 and 5), and homozygous fch/fch mice (lanes 3and 6). Note the absence of oatp-1 in fch/fch mice. (C) Levels ofNa1,K1-ATPase, a-subunit, in total liver membrane fractions of wild-type mice (lanes 1 and 4, respectively), heterozygous fch/1 mice(lanes 2 and 5), and homozygous fch/fch mice (lanes 3 and 6). Fifty(lanes 1–3) or 75 (lanes 4–6) micrograms of protein were applied.Na1,K1-ATPase is clearly present in membrane fractions of fch/fchlivers, showing that the absence of Ntcp (A) and oatp-1 (B) inmembrane fractions of fch/fch mice is not the result of an isolationartifact.

Figure 3. Immunohistochemical detection of ibst on cholangiocytesof frozen liver sections of a fch/fch mouse. A clear staining of theapical membranes of cholangiocytes was observed, as indicated byclosed arrowheads. Open arrowhead indicates precipitated PP local-ized in the dilated bile duct. When tissue was processed withoutaddition of the primary antibody as a negative control, no staining wasobserved (original magnification 603).

Figure 1. Representative sections of liver from wild-type (1/1) andhomozygous (fch/fch) mouse. (A) Liver of 1/1 mouse showing normalmorphology. (B) Severe bile duct proliferation and fibrosis in fch/fchmice, with PP precipitation in bile ducts (open arrowheads; Massonstain; original magnification 403).


bile flow was even higher in fch/fch and fch/1 mice than incontrols. The biliary secretion of PP was markedly higherin fch/fch mice than in the other 2 groups, as was that ofconjugated bilirubin. Biliary glutathione secretion, onthe other hand, was reduced by 90% in fch/fch micecompared with 1/1 mice. Surprisingly, biliary bile saltsecretion was markedly stimulated in the fch/fch mice.Increased bile salt output was accompanied by an in-creased output of cholesterol and phospholipids, but theratio of bile salts to lipids was clearly higher in fch/fchmice (mean bile salt to lipid ratio, 24.0) than in fch/1(ratio, 10:1) and in 1/1 mice (ratio, 11:8). Finally,biliary iron concentrations were below 3 µmol/L in all 3groups.

Table 4 summarizes biliary bile salt composition in the

3 strains. The secondary bile salt deoxycholate could notbe detected in the bile of fch/fch mice, and the relativecontribution of its precursor (of cholate) was significantlyreduced. Contribution of b-muricholate, on the otherhand, was significantly higher in fch/fch bile. In addition,D22-muricholate was detectable in the bile of fch/fch micebut not in that of the other 2 groups.

To gain insight in mechanisms potentially underlying

Table 1. Contents of Porphyrins, Conjugated Dienes,Thiobarbituric Acid–Reactive Substances, Thiols,and Iron in Liver Homogenates of Wild-Type (1/1)Mice and in Heterozygous (fch/1) and Homozygous(fch/fch) Ferrochelatase-Deficient Mice

1/1 fch/1 fch/fch

PP ( pmol/g liver ) 721 6 593 506 6 142 93081 6 43696a

Coproporphyrin I( pmol/g liver ) 4 6 4 ND 71 6 31a

Coproporhyrin III( pmol/g liver ) 26 6 26 16 6 25 654 6 214a

Conjugated dienes(absorbance of233/mmol lipid) 0.74 6 0.07 0.68 6 0.36 1.30 6 0.09a

TBARS (nmol/mgprotein) 1.05 6 0.13 1.21 6 0.12a 1.36 6 0.07a,b

Thiols (nmol/mgprotein) 78.6 6 5.1 82.3 6 7.8 69.1 6 11.5b

Iron (nmol/mg liver ) 7.55 6 1.31 5.49 6 1.05a 3.60 6 0.44a,b

NOTE. Mean values 6 SD are shown for 6–7 mice/group.TBARS, thiobarbituric acid–reactive substances; ND, nondetectable.aSignificant difference from 1/1 by ANOVA and post hoc Newman–Keuls t test (P , 0.05).bSignificant difference between fch/fch and fch/1 by ANOVA and posthoc Newman–Keuls t test (P , 0.05).

Table 2. Plasma Liver Function Parameters inWild-Type (1/1) Mice and in Heterozygous(fch/1) and Homozygous (fch/fch)Ferrochelatase-Deficient Mice

1/1 fch/1 fch/fch

AP (U/L) 21 6 15 20 6 10 327 6 114a

AST (U/L) 80 6 36 57 6 19 1263 6 777a

ALT (U/L) 53 6 9 77 6 37 1757 6 876a

Bilirubin (mmol/L)Total 9 6 5 8 6 3 152 6 100a

Direct 8 6 5 7 6 3 150 6 98a

Bile salts (mmol/L) 25 6 8 36 6 12 1819 6 1610a

NOTE. Data given as means 6 SD (n 5 6 in all groups).AP, alkaline phosphatase.aSignificant difference between fch/fch and fch/1 and 1/1 mice byANOVA (P , 0.05).

Table 3. Bile Flow and Biliary Secretion Rates inWild-Type (1/1) Mice and in Heterozygous(fch/1) and Homozygous (fch/fch)Ferrochelatase-Deficient Mice

1/1 fch/1 fch/fch

Bile flow(mL · min21 · 100g body wt21) 7.0 6 1.3 9.4 6 1.6a 9.7 6 0.9a

PP(pmol · min21 · 100g body wt21) 2.8 6 1.2 7.0 6 1.9 243.3 6 85a

Coproporhyrin I(pmol · min21 · 100g body wt21) 1.11 6 0.24 1.28 6 0.18 0.43 6 0.52a

Coproporhyrin III(pmol · min21 · 100g body wt21) 3.38 6 4.53 4.05 6 2.53 1.97 6 0.98

Bilirubin(nmol · min21 · 100g body wt21) 0.35 6 0.05 0.41 6 0.13 0.76 6 0.25a

Glutathione(nmol · min21 · 100g body wt21) 23.9 6 7.3 53.1 6 14.7a 3.7 6 0.8a,b

Bile salts(nmol · min21 · 100g body wt21) 185 6 55 250 6 63 910 6 463a,b

Phospholipids(nmol · min21 · 100g body wt21) 13.2 6 3.2 19.0 6 4.7 27.7 6 10.8a

Cholesterol(nmol · min21 · 100g body wt21) 2.5 6 0.9 5.7 6 1.7 10.3 6 4.5a

NOTE. Data given as means 6 SD (n 5 6/group).aSignificant difference from 1/1 values (P , 0.05).bSignificant difference between fch/fch and fch/1 mice (P , 0.05).

Table 4. Biliary Bile Salt Concentrations and Compositionsin Wild-Type (1/1) Mice and in Heterozygous(fch/1) and Homozygous (fch/fch)Ferrochelatase-Deficient Mice

Bile salt 1/1 fch/1 fch/fch

Total (mmol/L) 26.3 6 4.4 27.4 6 7.9 86.4 6 34.7a

Deoxycholate (%) 6.5 6 0.8 6.7 6 1.5 NDCholate (%) 71.4 6 2.7 62.7 6 2.5a 45.2 6 7.9a,b

Chenodeoxycholate (%) 1.7 6 1.0 5.2 6 1.7a 0.4 6 0.2a,b

b-Muricholate (%) 13.0 6 2.6 14.6 6 1.0 41.4 6 8.2a

D22-Muricholate (%) ND ND 1.8 6 0.5a

v-Muricholate (%) 7.5 6 1.1 10.9 6 1.0a 11.2 6 3.2a

NOTE. Data given as mean 6 SD (n 5 5–6 in each group).aSignificant difference from control values (P , 0.05).bSignificant difference between fch/fch and fch/1 (P , 0.05).


altered bile composition in ferrochelatase-deficient mice,we determined steady-state mRNA levels of cholesterol7a-hydroxylase (CYP7A) and sterol 27-hydroxylase(CYP27), the enzymes catalyzing the initial steps in bilesalt formation via the acidic and the neutral pathway,respectively. The mRNA levels of CYP7A and/or CYP27tended to be higher in fch/fch (1126% and 1170%,respectively) and fch/1 (no change for CYP7A, and1115% for CYP27) mouse livers than in the controlgroups. Because of the large variations, however, onlyCYP7A mRNA levels in fch/fch mice differed signifi-cantly (P , 0.05) from those in fch/1 and 1/1 mice.However, subsequent studies revealed that these elevatedhepatic CYP7A mRNA levels are not associated withincreased bile salt synthesis in this particular animalmodel. Fecal bile salt excretion, which is equal to hepaticsynthesis rate under steady-state conditions, was similarin fch/fch mice (4.0 6 2.5 µmol/day) and in 1/1 controls(5.5 6 3.3 µmol/day) (P . 0.05). In addition, urinarybile salt loss was significantly higher in fch/fch mice(174 6 87 nmol/day) than in 1/1 mice (18 6 2nmol/day), but more than an order of magnitude lowerthan the fecal loss.

Bile Salt Uptake Systems in Liver andIntestine

To explain the cause of the exceedingly highplasma and biliary bile salt concentrations in fch/fch mice,membrane fractions were isolated for detection of Ntcpand oatp-1, involved in Na1-dependent and Na1-independent uptake of bile salts by the liver, respectively.The Western blot in Figure 2A shows that Ntcp is belowdetectable levels in the crude membranes and plasmamembrane fractions isolated from fch/fch livers, in con-trast to the situation in fch/1 and 1/1 membranefractions. In addition, oatp-1 could not be detected infch/fch membranes (Figure 2B). In contrast, the a-subunitof Na1,K1-ATPase was clearly present, although insomewhat reduced amounts, in membrane fractions fromfch/fch livers (Figure 2C), indicating that the absence ofNtcp and oatp-1 is not the result of an isolation artifact.

In view of recent reports describing the presence of ibstand Na1-dependent taurocholate transport in bile ductepithelial cells,26,27 immunohistochemical studies wereperformed on frozen sections of 1/1 and fch/fch livers.Figure 3 shows that ibst is present at the apical mem-brane of the epithelial cells lining the proliferated bileducts in fch/fch livers. No clear signal was detected by thisapproach in the small ducts of 1/1 livers. Westernanalysis of crude membrane fractions indicated thepresence of ibst both in 1/1 and fch/fch livers, with ahigher content in the latter (data not shown).

Figure 4 shows that the levels of ibst are clearlyincreased in the small intestine of fch/fch mice, as detectedby antisera raised against the rat protein. At all 4 levels ofthe intestine, immunoreactive material with molecularmasses of about 46 kilodaltons (indicated by arrow) and93 kilodaltons was present, considered to represent themonomeric and dimeric forms of ibst, respectively.28,29

To evaluate the consequences of altered expression ofbile salt uptake systems in fch/fch mice, the distribution ofintravenously administered [3H]taurocholate over plasma,hepatic, and biliary compartments was determined 30minutes after injection. Figure 5 shows that virtually all(.95%) recovered radioactivity is present in bile of 1/1mice after 30 minutes. In contrast, a significant amountof radioactivity (,8%) is present in plasma of fch/fch miceat this time point. Yet, despite the absence of Ntcp andoatp-1, about 40% and 50% of recovered radioactivity ispresent in liver and bile, respectively, indicating thathepatic taurocholate processing from plasma to bile isdelayed but still functioning in fch/fch mice.

Bile Formation in Griseofulvin-Induced EPPin Mice

To ascertain whether the changes observed in bileformation are related to impaired ferrochelatase activityrather than to aspecific effects associated with the proce-dure used to generate the mutation, a limited experimentwas performed in which mice treated for 2.5 weeks withgriseofulvin, a well-known inhibitor of hepatic ferroche-latase activity.30,31 Of note, in this model the liver is theonly source of PP, whereas in fch/fch mice the bonemarrow is a major contributor.31 Data summarized inTable 5 clearly show that griseofulvin treatment leads to

Figure 4. Western blot analysis of ibst levels in the small intestines ofwild-type and fch/fch mice, using the anti-rat ibat antiserum (RIBMAL1).Small intestines of both groups of mice were divided into 4 equalparts, mucosae were collected, and crude membrane preparationswere made. Lanes 1–4 represent subsequent sections of wild-typemice, from proximal (lane 1) to distal (lane 4). Lanes 5–8 representsubsequent sections of fch/fch mice, from proximal (lane 5) to distal(lane 8). Lane 9 shows a crude membrane fraction of rat ileum forcomparison. Arrow indicates localization of monomeric ibst (,46kilodaltons).


PP accumulation in red blood cells and in liver as well asto increased disposition of PP into bile. Griseofulvin-treated mice also had an increase in liver weight and liverto body weight ratio, similar to the situation observed inferrochelatase-deficient mice. Most importantly, plasmabile salt levels were only slightly elevated, whereas bileflow and biliary bile salt secretion were markedly stimu-lated by griseofulvin. The bile salt to lipid ratio wasincreased from 11.9 to 25.9. The similarity between bothindependent models of impaired ferrochelatase activitysupports the view that altered bile composition is a directconsequence of this condition.

Discussion

Liver fibrosis and cirrhosis as well as PP gallstoneshave been reported in humans with EPP. Acute liverfailure requiring liver transplantation is a relatively rarecomplication of the disease.6,32 All the mutations in theferrochelatase gene found in patients with liver failurelead to major alterations in the structure of the protein.7,8

However, there appears to be no clear-cut relationbetween the genetic defect and PP accumulation, residualenzyme activity, or severity of the disease. Evidently,other factors contribute to the variable disease expression;Rufenacht et al.7 recently described clinical manifesta-tions ranging from mild skin photosensitivity to terminalliver failure in 4 unrelated subjects with the same pointmutation.

In this study, we evaluated liver pathology in relationto bile composition in a mouse model of EPP generatedduring the course of a chemical mutagenesis experimentwith ethylnitrosourea.10 Analysis of complementary DNAclones obtained from these mice revealed a T to Atransition at nucleotide 293, leading to a methionine tolysine substitution at position 98 of the protein.33

Homozygosity for this mutation ( fch/fch) reduces ferroche-latase activity in liver, spleen, and kidney to 2.7%, 6.3%,and 3.3% of control values, respectively.10 Reportedenzyme activities in heterozygotes in these organs are65%, 45%, and 54% of control, respectively. Examina-tion of livers obtained from 3-month-old fch/fch miceshowed portal inflammation and severe bile duct prolifera-tion and biliary fibrosis with portoportal bridging,associated with a marked enlargement of the liver. Thehepatic PP content was dramatically increased in homozy-gous mutants, and PP deposits were seen in bile ductulesand in sinusoidal cells, presumably representing Kupffercells. In contrast, no pathology was seen in livers of fch/1mice, except for the presence of macrovesicular fat inperivenous areas. This indicates that 50% of normalenzyme activity, sufficient to maintain normal erythro-cyte and hepatic PP levels, is also sufficient to preventdevelopment of liver disease in these mice.

It has been suggested that so-called ‘‘dark effects’’ of

Figure 5. Distribution of radioactivity over plasma (j), hepatic (N),and biliary (h) compartments in female wild-type (1/1) and fch/fchmice 30 minutes after intravenous injection of a tracer amount of[3H]taurocholate via the left jugular vein. To account for small butvariable losses of radioactivity during the injection procedure, data areexpressed as percentage of recovered radioactivity (sum of plasma,liver, and biliary radioactivity). Average values are shown (n 5 4).

Table 5. Effects of Griseofulvin Treatment on Liver Weight,PP Levels, and Bile Formation in Swiss Mice

Control Griseofulvin

LiverLiver wt 1.6 6 0.2 (19) 2.9 6 0.8a (8)Liver/body wt ( g ) 0.05 6 0.005 (19) 0.09 6 0.01a (8)PP ( pmol/g ) 398 6 101 (3) 143,000 (1)

BloodPP ery (nmol/L) 1569 6 510 (6) 8714 6 4847a (5)Bile salt plasma

(mmol/L) 46 6 32 (6) 112 6 98 (5)Bile

Bile flow(mL · min21 · 100 gbody wt21) 5.28 6 2.69 (19) 9.13 6 1.31a (3)

Bile salts(nmol · min21 · 100 gbody wt21) 112 6 46 (19) 244 6 32a (3)

Cholesterol(nmol · min21 · 100 gbody wt21) 1.67 6 0.80 (19) 2.95 6 0.06a (3)

Phospholipids(nmol · min21 · 100 gbody wt21) 13.8 6 6.2 (19) 8.3 6 5.1 (3)

PP (pmol · min21 · 100 gbody wt21) 3.9 6 3.2 (15) 1208 6 181a (3)

NOTE. Data given as means 6 SD of indicated number of animals (inparentheses).aSignificant difference between control and griseofulvin-treated mice(P , 0.05).


PP—free-radical mediated damage initiated by PP in theabsence of light—contribute to development of liverdisease in EPP.34,35 In the present study, we foundindications for peroxidation in livers of fch/fch livers:increased concentrations of conjugated dienes in thehepatic lipid fraction and of thiobarbituric acid–reactivesubstances.Yet the differences between fch/fch mice on theone hand and fch/1 and 1/1 on the other hand weremodest and relatively small compared with those ob-served in established models of liver fibrosis and cirrhosisrelated to oxidant stress, such as CCl4 and ethanol.36,37

The unexplained low iron content of fch/fch livers may bebeneficial in this respect because iron has been shown toexacerbate oxidant stress–related fibrogenesis.38 In ouropinion, therefore, it is unlikely that oxidant stress aloneis responsible for the development of liver disease infch/fch mice.

Persisting cholestasis per se can lead to irreversibleliver damage. The bile duct–ligated rat is a widely usedmodel of biliary fibrosis progressing to cirrhosis.39 Devel-opment of cirrhosis in this model has been claimed to bepreceded by a complex sequence of events involving toxicactions of bile salts on bile duct epithelium, chemokineand mitogen production by activated neutrophils, andgeneration of reactive oxygen species. Based on plasmaparameters, the fch/fch mice at first sight appear to beextremely cholestatic, with high liver enzyme levels,conjugated hyperbilirubinemia, and particularly highbile salt levels. Surprisingly, bile formation appeared notto be compromised. On the contrary, bile flow and biliarybile salt secretion are higher in fch/fch mice than incontrols, although the amount of fluid per mole of bilesalt is reduced in fch/fch mice. The latter may be a result ofimpaired bile salt–independent bile flow caused by thelow biliary glutathione secretion.

It could be argued that cannulation of the gallbladderfor collection of bile alleviates obstruction of the bileducts by PP and thereby causes an enhanced flow of bile.However, there are a number of arguments against thisoption. First, obstruction of bile ducts is most likely tooccur in smaller units inside the liver; this would not besolved by creating an outlet at the level of the gallbladder.Second, urinary bile salt loss is only 10-fold higher infch/fch mice than in controls, whereas bile duct ligation inrats leads to a persistent 4000-fold increase of urinary bilesalts.40 Third, fecal bile salt output is not affected infch/fch mice, but a strong decrease would be expected incase of an obstruction.40 In addition, induction of ibst inthe small intestine is consistent with an increased bile saltload in the intestine,41 although, obviously, other factorsmay also be involved. Fourth, release of bile ductobstruction is associated with an isolated hypersecretion

of phospholipids into bile, at least in rats.42 This is notseen in fch/fch mice. Finally and most importantly,short-term treatment of control mice with the ferroche-latase inhibitor griseofulvin stimulates bile flow andbiliary bile salt secretion and leads to an increased bilesalt/lipid ratio while plasma bile salts are not signifi-cantly elevated. This indicates that impaired ferroche-latase activity leading to elevated hepatic PP levels per seaffects bile composition. Together these observationsstrongly indicate that frank biliary obstruction is not thecause of fibrosis, although it cannot be excluded thatobstruction of some small bile ductules by PP precipi-tates does occur.

We propose that bile duct proliferation and fibrosis arenot a result of the absence of bile formation but theformation of bile with proliferation-inducing properties.Indeed, the bile of fch/fch mice has a very high concentra-tion of detergent bile salts that, on a molar basis, isaccompanied by less cholesterol and phospholipid than infch/1 and in 1/1 mice. Formation of ‘‘cytotoxic’’lipid-free bile has recently been shown to underlie thedevelopment of bile duct proliferation and biliary fibrosisin mdr2 Pgp-deficient mice,12,13 a pathology similar tothat observed in fch/fch mice. It is evident that in fch/fchmice the bile salt/lipid ratio is affected to a lesser extentthan in mdr2 Pgp-deficient mice. In addition to straindifferences that may influence susceptibility of the biliaryepithelium, the presence of other bile components such asPP and coproporphyrins in high concentrations, and theabsence of others such as glutathione, may contribute tothe overall cytotoxic potential of fch/fch bile toward theepithelial cells.

Presently we can only speculate about the causes ofabnormalities in bile formation induced by ferrochelatasedeficiency. Low biliary glutathione may be an indicator ofthe compromised thiol status of liver cells and/or reducedmrp2/cmoat activity.43 The latter may also contribute toconjugated hyperbilirubinemia in fch/fch mice. Of particu-lar interest, however, are the overt changes in bile saltmetabolism. Bile salt secretion is increased in fch/fch micewith proportionally more muricholates and no secondarybile salts. Fecal and urinary bile salt analysis showed thathepatic bile salt synthesis is not increased in fch/fch micebut actually tends to be decreased. The finding thathepatic mRNA levels of CYP7A are increased in fch/fchmice therefore most likely has no physiological meaning.

Why then is biliary bile salt output increased in fch/fchmice and why are plasma bile salt levels extremelyelevated at the same time? The latter is probably relatedto the absence of Ntcp and oatp-1 in the liver, leading toan impaired capacity for Na1-dependent and-indepen-dent bile salt uptake. Whether this is a cause or a


consequence, as proposed by Gartung et al.,44 of highplasma bile salts cannot be concluded at this point.Experiments with radiolabeled taurocholate suggest im-paired hepatic bile salt uptake in fch/fch mice. Yet, themajority of the recovered dose was present in liver andbile at 30 minutes after injection, indicative for taurocho-late uptake by the liver via alternative mechanisms.

We speculate that, in states of reduced ferrochelataseactivity, high biliary bile salt concentrations not accompa-nied by sufficient lipids and high PP concentrationsdamage cholangiocytes and/or obstruct small bile duct-ules. This will induce bile duct proliferation, leading toan increased overall ibst activity in the bile ducts and toincreased reabsorption of bile salts at this level. The bilesalts are thus partly retained in the liver and return to thehepatocytes for resecretion into bile. Such an ‘‘intrahe-patic bile salt circulation’’ may lead to increased bile saltconcentrations in the hepatic circulation and insidehepatocytes and, by spillover, to increased concentrationsin the systemic circulation. Increased intracellular bilesalt concentrations, perhaps in conjunction with inflam-mation-related factors, may down-regulate Ntcp andoatp-1, thereby aggravating bile salt accumulation in thesystemic circulation. Accumulation of radiolabeled tauro-cholate in fch/fch liver after its intravenous administrationmay be a reflection of this proposed sequence of events.The increased concentration of bile salts in the bile offch/fch mice is also associated with up-regulation of ibst inthe intestine.

To the best of our knowledge, the fch/fch mouseprovides the first ‘‘pathological’’ animal model displayingincreased ibst levels in the intestine. Induction of the ibstsystem accelerates enterohepatic cycling of the bile salts.The presence of an intrahepatic circulation and a short-circuit of the enterohepatic circulation, the latter sup-ported by the absence of deoxycholate in the pool, causesa very high biliary bile salt secretion rate and animbalance between bile salt and lipid output,45 leading tothe formation of cytotoxic bile and further proliferation ofbile duct cells that progresses to biliary fibrosis. Clearly,this proposed sequence of events is highly speculative butthe various aspects of this hypothetical scheme can beapproached experimentally.

In conclusion, ferrochelatase-deficient mice developsevere biliary fibrosis progressing to cirrhosis within 3months. Development of liver disease is associated withthe formation of bile that has a high concentration of PPand bile salts and a relatively low concentration of lipids,which may, in analogy to the situation described mdr2Pgp-deficient mice, cause cytotoxicity to biliary epithe-lium. It may well be that changes in bile composition ingeneral are of importance in initiation of biliary liver

disease. Specifically, interindividual differences in thebile formation process, for example, in factors thatregulate bile salt metabolism or biliary lipid secretion,may contribute to the variability in the severity inEPP-related liver disease in humans.

References1. Kappas A, Sassa S, Galbraith RA, Nordmann Y. The porphyrias. In:

Scriver CR, Beadet AL, Sly WS, Valle D, eds. The metabolic basisof inherited disease. New York: McGraw-Hill, 1989:1305–1365.

2. Avner DL, Berenson MM. Hepatic clearance and biliary secretionof protoporphyrin in isolated, in situ-perfused rat liver. J Lab ClinMed 1982;99:885–894.

3. Berenson MM, Garcia-Marin JJ, Gunther C. Effect of bile acidhydroxylation on biliary protoporphyrin excretion in rat liver. Am JPhysiol 1988;255:G382–G388.

4. Berenson MM, el-Mir MY, Zhang LK. Mechanism of bile acidfacilitation of biliary protoporphyrin excretion in rat liver. Am JPhysiol 1995;268:G754–G763.

5. Beukeveld GJJ, In’t Veld G, Havinga R, Groen AK, Wolthers BG,Kuipers F. Relationship between biliary lipid and protoporphyrinsecretion: potential role of mdr2 P-glycoprotein in hepatobiliaryorganic anion transport. J Hepatol 1996;24:343–352.

6. Cox TM. Erythropoietic protoporphyria. J Inherit Metab Dis 1997;20:258–269.

7. Rufenacht UB, Gouya L, Schneider-Yin X, Puy H, Schafer BW,Aquaron R, Nordmann Y, Minder EI, Deybach JC. Systematicanalysis of molecular defects in the ferrochelatase gene frompatients with erythropoietic protoporphyria. Am J Hum Genet1998;62:1341–1352.

8. Bloomer J, Bruzzone C, Zhu L, Scarlett Y, Magness S, Brenner D.Molecular defects in ferrochelatase in patients with protopor-phyria requiring liver transplantation. J Clin Invest 1998;102:107–114.

9. Doss MO, Frank M. Hepatobiliary implications and complicationsin protoporphyria: a 20-year study. Clin Biochem 1989;22:223–229.

10. Tutois S, Montagutelli X, Da Silva V, Jouault H, Rouyer-Fessard P,Leroy-Viard K, Guenet JL, Nordmann Y, Beuzard Y, Deybach JC.Erythropoietic protoporphyria in the house mouse. J Clin Invest1991;88:1730–1736.

11. Smit JJM, Schinkel AH, Oude Elferink RPJ, Groen AK, Wagenaar E,van Deemter L, Mol CAAM, Ottenhof R, van der Lugt NMT, vanRoon MA, van der Valk MA, Offerhaus GJA, Berns AJM, Borst P.Homozygous disruption of the murine mdr2 P-glycoprotein geneleads to a complete absence of phospholipid from bile and liverdisease. Cell 1993;75:451–462.

12. Van Nieuwkerk CMJ, Oude Elferink RPJ, Groen AK, Ottenhof R,Tytgat GNJ, Dingemans KP, van den Bergh Weerman MA, Offer-haus GJA. Effects of ursocholate and cholate feeding on liverdisease in FVB mice with a disrupted mdr2 P-glycoprotein gene.Gastroenterology 1996;111:165–171.

13. Van Nieuwkerk CMJ, Groen AK, Ottenhof R, van Wijland M, vanden Bergh Weerman MA, Tytgat GNJ, Offerhaus GJA, OudeElferink RPJ. The role of bile salt composition in liver pathology ofmdr2 (–/–) mice: differences between males and females. JHepatol 1997;26:138–145.

14. Voshol PJ, Havinga R, Wolters H, Ottenhoff R, Princen HMG, OudeElferink RPJ, Groen AK, Kuipers F. Reduced plasma cholesteroland increased fecal sterol loss in multidrug resistance gene 2P-glycoprotein–deficient mice. Gastroenterology 1998;114:1024–1034.

15. Preisegger KH, Stumptner C, Riegelnegg D, Brown PC, SilvermanJA, Thorgeirsson SS, Denk H. Experimental Mallory body forma-tion is accompanied by modulation of the expression of multidrug-


resistance genes and their products. Hepatology 1996;24:248–252.

16. Kuipers F, Havinga R, Bosschieter H, Toorop GP, Hindriks FR, VonkRJ. Enterohepatic circulation in the rat. Gastroenterology 1985;88:403–411.

17. Recknagel PO, Glende EA. Spectrophotometric method of lipidconjugated dienes. Methods Enzymol 1984;105:331–337.

18. Ohkawa H, Ohishi N, Yagi K. Assay for lipid peroxidation in animaltissues by thiobarbituric acid reaction. Anal Biochem 1979;95:351–358.

19. Griffith OW. Determination of glutathione and glutathione disul-fide using glutathione reductase and 2-vinyl pyridine. Anal Bio-chem 1980;106:207–212.

20. Bligh EG, Dyer WJ. A rapid method of total lipid extraction andpurification. Can J Biochem Biophys 1959;37:911–917.

21. Wolters H, Spiering M, Gerding A, Slooff MJH, Kuipers F, HardonkM, Vonk RJ. Isolation and characterization of canalicular andbasolateral plasma membrane fractions from human liver. Bio-chim Biophys Acta 1991;1069:61–69.

22. Koopen NR, Wolters H, Muller M, Schippers IJ, Havinga R,Roelofsen H, Vonk RJ, Stieger B, Meier PJ, Kuipers F. Hepatic bilesalt flux does not modulate levels and activity of the sinusoidalNa1-taurocholate cotransporter (Ntcp) in rats. J Hepatol 1997;27:699–706.

23. Peters WH, de Pont JJ, Koppers A, Bonting SL. Studies on(Na1,K1)-activated ATPase. XLVII. Chemical composition, molecu-lar weight and molar ratio of the subunits of the enzyme fromrabbit kidney outer medulla. Biochim Biophys Acta 1981;641:55–70.

24. Chomczynski P, Sacchi N. Single step method of RNA isolation byacid guadinium thiocyanate-phenol-chloroform extraction. AnalBiochem 1987;162:156–158.

25. Post SM, de Wit ECM, Princen HMG. Cafestol, the cholesterol-raising factor in boiled coffee, suppresses bile acid synthesis bydown-regulation of cholesterol 7a-hydroxylase and sterol 27-hydroxylase in rat hepatocytes. Arterioscler Thromb Vasc Biol1997;17:3064–3070.

26. Lazarides KN, Pham L, Tietz P, Marinelli RA, de Groen PC, LevineS, Dawson PA, LaRusso NF. Rat cholangiocytes absorb bile acidsat their apical domain via the ileal sodium-dependent bile acidtransporter. J Clin Invest 1997;100:2714–2712.

27. Alpini G, Glazer SS, Rodgers R, Phinizy JL, Robertson WE, LasaterJ, Caligiuri A, Tretjak Z, Le Sage GD. Functional expression of theapical Na1-dependent bile acid transporter in large but not insmall rat cholangiocytes. Gastroenterology 1997;113:1734–1740.

28. Kramer W, Wess G, Bewersdorf U, Corsiero D, Girbig F, WeylandC, Stengelin S, Ehnsen A, Bock K, Kleine H, Le Dreau MA, SchaferHL. Topological photoaffinity labeling of the rabbit ileal Na1/bilesalt-cotransporting system. Eur J Biochem 1997;249:456–464.

29. Kramer W, Corsiero D, Friedrich M, Girbig F, Stengelin S, WeylandC. Intestinal absorption of bile acids: paradoxical behaviour of the14 kDa ileal lipid binding protein in differential photoaffinitylabeling. Biochem J 1998;333:335–341.

30. Yokoo H, Graig R, Haywood TR, Cochrane C. Griseofulvin-inducedcholestasis in Swiss albino mice. Gastroenterology 1979;77:1082–1087.

31. Holley A, King LJ, Gibbs AH, De Matteis, F. Strain and sexdifferences in the response of mice to drugs that induceprotoporphyria: role of porphyrin biosynthesis and removal. JBiochem Toxicol 1990:5:175–182.

32. Sarkany RP, Alexander GJ, Cox TM. Recessive inheritance oferythropoietic protoporphyria. Lancet 1994;343:1394–1396.

33. Boulechfar S, Lamoril J, Montagutelli X, Guenet J, Deybach JC,

Nordmann Y, Dailey H. Ferrochelatase structural mutant(Fechm1Pas) in the house mouse. Genomics 1993;16:645–648.

34. Koningsberger JC, Rademakers LH, van Hattum J, de la Faille HB,Wiegman LJ, Italiaander E, van Berge Henegouwen GP, Marx JJ.Exogenous protoporphyrin inhibits HepG2 cell proliferation, in-creases the intracellular hydrogen peroxide concentration andcauses ultrastructural alterations. J Hepatol 1995;22:57–65.

35. Koningsberger JC, van Asbeck BS, van Hattum J, Wiegman LJ, vanBerge Henegouwen GP, Marx JJ. The effects of porphyrins oncellular redox system: a study on the dark effects of porphyrins onphagocytes. Eur J Clin Invest 1993;23:716–723.

36. Tsukamoto H. Oxidative stress, antioxidants, and alcoholic liverfibrosis. Alcohol 1993;10:465–467.

37. Polavarapu R, Spitz DR, Sim JE, Follansbee MH, Oberley LW,Rahemtulla A, Nanji AA. Increased lipid peroxidation and impairedantioxidant enzyme function is associated with pathological liverinjury in experimental alcoholic liver disease in rats fed diets highin corn oil and fish oil. Hepatology 1998;27:1317–1323.

38. Tsukamoto H, Horne W, Kamimura S, Niemela O, Parkkila S,Yla-Hertula S, Brittenham GM. Experimental liver cirrhosis in-duced by alcohol and iron. J Clin Invest 1995;96:620–630.

39. Gross JB, Reichen J, Zeltner TB, Zimmermann A. The evolution ofchanges in quantitative liver function tests in a rat model of biliarycirrhosis: correlation with morphometric measurement of hepato-cyte mass. Hepatology 1987;7:457–463.

40. Kunigasa T, Uchida K, Kadowaki M, Takase H, Nomura Y, Saito Y.Effect of bile duct ligation on bile acid metabolism in rats. J LipidRes 1981;22:201–207.

41. Stravitz RT, Sanyal AJ, Pandak WM, Vlahcevic ZR, Beets JW,Dawson PA. Induction of sodium-dependent bile acid transportermessenger RNA, protein, and activity in rat ileum by cholic acid.Gastroenterology 1997;113:1599–1608.

42. Hardison WGM, Evans WH. Postobstructive subcellular organelleand biliary lipid composition in the rat. A selective increase inbiliary lecithins is not reflected by changes in organelle composi-tion. J Hepatol 1986;3:318–327.

43. Muller M, Jansen PLM. Molecular aspects of hepatobiliary trans-port. Am J Physiol 1997;272:G1285–G1303.

44. Gartung C, Schuele S, Schlosser SF, Boyer JL. Expression of therat liver Na1/taurocholate cotransporter is regulated in vivo byretention of biliary constituents but not by their depletion.Hepatology 1997;25:284–290.

45. Groen AK, Oude Elferink RPJ, Tager JM. Control analysis of biliarylipid secretion. J Theor Biol 1996;182:427–436.

Received November 24, 1998. Accepted June 1, 1999.Address requests for reprints to: Folkert Kuipers, Ph.D., Center for

Liver, Digestive and Metabolic Diseases, CMC IV, Room Y2115,University Hospital Groningen, P.O. Box 30.001, 9700 RB Groningen,The Netherlands. e-mail: [email protected]; fax: (31) 50-361-1746.

Supported by grant WS 95-41 from the Netherlands Foundationfor Digestive Diseases and by grants 900-523-133 and902-23-191 from the Netherlands Organization for ScientificResearch.

The authors thank Dr. X. Montagutelli and Prof. Dr. Y. Nordmann(Institut Pasteur, Paris, France) for the generous gift of fchmice; Renze Boverhof, Dr. Henk Wolters, Rob Lohr, EngeVenekamp-Hoolsma, and Alie de Jager for excellent technicalassistance; Peter van der Syde for photography; and E. Zwarteveenand Jan Louwes (Central Animal Laboratory, University ofGroningen) for breeding of fch mice.


Biliary fibrosis associated with altered bile composition in a mouse model of erythropoietic...

Documents

Transcript of Biliary fibrosis associated with altered bile composition in a mouse model of erythropoietic...