Effective Treatment of Unconjugated Hyperbilirubinemia With Oral Bile Salts in Gunn Rats

ES

FC

*GCDW

BoiibGs(upAiktcPUotwcPa(irscBcau

UgbceIi

GASTROENTEROLOGY 2009;136:673–682

ffective Treatment of Unconjugated Hyperbilirubinemia With Oral Bilealts in Gunn Rats

RANS J. C. CUPERUS,* ANJA M. HAFKAMP,* RICK HAVINGA,* LIBOR VITEK,‡,§ JAROSLAV ZELENKA,§

LAUDIO TIRIBELLI,� J. DONALD OSTROW,¶ and HENKJAN J. VERKADE*

Pediatric Gastroenterology, Department of Pediatrics, Center for Liver, Digestive, and Metabolic Diseases, University Medical Center Groningen, University ofroningen, Groningen, The Netherlands; ‡4th Department of Internal Medicine, 1st Faculty of Medicine, Charles University, Prague, Czech Republic; §Institute oflinical Biochemistry and Laboratory Diagnostics, 1st Faculty of Medicine, Charles University, Prague, Czech Republic; �Centro Studi Fegato, AREA Science Park andepartment of Life Sciences, University of Trieste, Trieste, Italy; and ¶Gastroenterology/Hepatology Division, University of Washington School of Medicine, Seattle,

(bndgphettstidc

Ucnhfaatlpur

gtmcc

Adm

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

ackground & Aims: We tested the hypothesis thatral administration of bile salts, which are known to

ncrease the biliary excretion of unconjugated bil-rubin (UCB), decreases unconjugated hyperbiliru-inemia in the Gunn rat model. Methods: Adultunn rats were fed a standard diet or the same diet

upplemented with 0.5 weight % ursodeoxycholic acidUDCA) or cholic acid (CA) for 1 or 6 weeks. UCB androbilinoids, a family of intestinal UCB breakdownroducts, were determined in plasma, feces, or both.fter 6 weeks of treatment, tracer 3H-UCB was admin-

stered intravenously to determine steady-state UCBinetics over the next 60 hours. Results: One-weekreatment with UDCA or CA decreased plasma UCBoncentrations by 21% and 30%, respectively (each

< .01). During the first 4 days of treatment, bothDCA and CA increased the combined fecal excretionf UCB and urobilinoids (�52% and �32%, respec-ively; each P < .01). Prolongation of treatment to 6eeks caused a persistent decrease in plasma UCB

oncentrations to �40% below baseline (each bile salt< .001). 3H-UCB kinetic studies showed that UDCA

nd CA administration decreased UCB pool size�33% and �32%, respectively; each P < .05) andncreased UCB fractional turnover (�33% and �25%,espectively; each P < .05). Conclusions: Dietary bilealt administration induces a large, persistent de-rease in plasma UCB concentrations in Gunn rats.oth UDCA and CA enhance UCB turnover by in-reasing its fecal disposal. These results support thepplication of oral bile salt treatment in patients withnconjugated hyperbilirubinemia.

nconjugated hyperbilirubinemia occurs in condi-tions such as neonatal hemolytic jaundice and Cri-

ler-Najjar disease. Crigler-Najjar disease is characterizedy a genetically absent (type I) or decreased (type II)apacity to conjugate bilirubin in the liver,1 which isssential for efficient biliary excretion of the pigment.mpaired conjugation results in unconjugated hyperbil-

UCB) in the body. Severe unconjugated hyperbiliru-inemia can lead to deposition of UCB in the centralervous system, causing bilirubin-induced neurologicysfunction (BIND), kernicterus, and death.2 Unconju-ated hyperbilirubinemia is conventionally treated byhototherapy, which induces photoisomerization of theydrophobic UCB to polar isomers that can readily bexcreted into the bile.3 Although generally effective, pho-otherapy does not always decrease plasma UCB to non-oxic levels. Most importantly, long-term phototherapy,uch as needed for patients with Crigler-Najjar diseaseype I, becomes less effective with age and has a profoundmpact on social life.4,5 These considerations favor theevelopment of effective alternative treatments for un-onjugated hyperbilirubinemia.

During severe unconjugated hyperbilirubinemia, mostCB does not enter the intestinal lumen via biliary ex-

retion, but rather via direct diffusion across the intesti-al mucosa.6,7 The efficiency of this pathway is decreased,owever, by the ability of the intestine to reabsorb UCB

rom its lumen.8,9 Several experimental therapies haveimed to prevent reabsorption by oral administration ofgents that trap UCB in the intestinal lumen. However,rapping agents tested so far, including agar,10 cho-estyramine,11 charcoal,12 amorphous calcium phos-hate,13 zinc salts,14 and orlistat,15 have been clinicallynsatisfactory because of side effects and inconsistentesults.

Since bile salts can stimulate biliary excretion of or-anic anions,16 including bilirubin in rats,17 we reasonedhat bile salt administration could be relevant for treat-

ent of unconjugated hyperbilirubinemia. Ursodeoxy-holic acid (UDCA) treatment in healthy volunteers de-reased the expiration of 14CO2 from triolein, suggesting

Abbreviations used in this paper: ALT, alanine aminotransferase;ST, aspartate aminotransferase; BW, body weight; CA, cholic acid;pm, disintegrations per minute; HPLC, high-performance liquid chro-atography; UCB, unconjugated bilirubin; UDCA, ursodeoxycholic acid.

© 2009 by the AGA Institute0016-5085/09/$36.00

doi:10.1053/j.gastro.2008.10.082

tMptlUao

tbdipawsUb

2shaop

Ftdgb3

p5AUi

(d5bU

6rm

apa0woatpa

trd6tw5tbdamelgbi

tasgwt6UweAgvbdwn�fadtwU

f

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

674 CUPERUS ET AL GASTROENTEROLOGY Vol. 136, No. 2

hat UDCA also mildly decreases the absorption of fat.18

ild fat malabsorption induced by orlistat decreasedlasma bilirubin levels in a subset of Crigler-Najjar pa-ients and in homozygous Gunn rats, their well-estab-ished animal model.15,19 Finally, bile salts associate withCB in vitro, and bile salt administration could therefore

lso lower plasma UCB concentrations via enhancementf fecal excretion of UCB– bile salt complexes.20,21

In the present study we show that dietary administra-ion of UDCA indeed reduces unconjugated hyperbiliru-inemia in homozygous Gunn rats. We studied severalosages and administration periods to evaluate the clin-

cal applicability of this potential treatment. We com-ared the UDCA effects in Gunn rats with those obtainedfter administration of cholic acid (CA) to evaluatehether effects were bile salt specific. Finally, we studied

teady-state kinetics of intravenously administered 3H-CB, to gain insight into the underlying mechanisms ofoth UDCA and CA administration.

Materials and MethodsAnimalsHomozygous adult male Gunn Rats (RHA/jj,

40 –360 g), obtained from our breeding colony (Univer-ity Medical Center Groningen, The Netherlands), wereoused in an environmentally controlled facility and fedd libitum. The Ethics Committee for Animal Experimentsf the University of Groningen approved all experimentalrotocols.

MaterialsChemicals. UDCA was a generous gift from Dr

alk Pharma GmbH (Freiburg, Germany). CA and hep-adecanoic acid (C17:0) were obtained from Sigma-Al-rich (St Louis, MO). Xanthobilirubin-methyl ester was aenerous gift from Dr J. Fevery (Leuven, Belgium). Uro-ilin was obtained from Frontier Scientific (Logan, UT).H-labeled UCB (specific activity 6.02 �Ci/�mol) wasrepared by biosynthetic labeling of 2,3-3H-labeled-aminolevulinic acid (specific activity 13 mCi/mmol;mersham Biosciences, Piscataway, NJ).22–24 3H-labeledCB solution was prepared immediately before injection

nto Gunn rats as described before.23

Diets. The semisynthetic, purified control dietcode 4063.02) was produced by Hope Farms BV (Woer-en, The Netherlands) and contained 13 energy% fat and.2 weight (wt%) long-chain fatty acids. Diets containingile salts were identical except for supplementation withDCA or CA (0.05%–1.5% by chow weight).

MethodsPreliminary dose-response experiment. After a

-week run-in period on the control diet, Gunn rats wereandomly assigned to receive the control diet supple-

ented with either UDCA or CA (n � 6 per group). All a

nimals were housed and fed by dietary group for aeriod of 10 weeks, during which the dosage of UDCAnd CA was increased every 2 weeks. Used dosages were.05 wt%, 0.1 wt%, 0.5 wt%, 1 wt%, and 1.5 wt% (by choweight). Heparinized samples of tail vein blood werebtained under isoflurane anesthesia before and 2, 4, 6, 8,nd 10 weeks after dietary randomization for determina-ion of plasma UCB concentrations and, in the lastlasma sample, aspartate aminotransferase (AST) andlanine aminotransferase (ALT) concentrations.

Short-term experiment. After a 6-week period onhe control diet, individually housed Gunn rats wereandomly assigned to receive the control diet or the sameiet supplemented with UDCA or CA (each 0.5 wt%; n �per group). Food intake and animal weights were de-

ermined daily. Heparinized samples of tail vein bloodere obtained under isoflurane anesthesia at day 0, 1, 3,, and 8 for determination of plasma UCB concentra-ions. Feces were collected every 24 hours for 4 daysefore and for 4 days after dietary randomization toetermine fecal excretion of UCB, urobilinoids, and bilecids. Eight days after dietary randomization, the com-on bile duct was cannulated under pentobarbital an-

sthesia, and bile was collected for 30 minutes underight-protected conditions. Bile flow was determinedravimetrically, assuming a density of 1 g/mL. A 1-mLlood sample was then obtained by puncture of the

nferior vena cava to determine AST and ALT.Long-term experiment. After 6 weeks on the con-

rol diet, individually housed Gunn rats were randomlyssigned to receive either the control diet or the same dietupplemented with UDCA or CA (each 0.5 wt%; n � 6 perroup). Food intake and animal weights were determinedeekly. Heparinized samples of tail vein blood were ob-

ained under isoflurane anesthesia before and at 2, 4, andweeks after dietary randomization to determine plasmaCB concentrations. At 5 weeks, the rats were gavagedith 1 mL (20 mg/mL) carmine red, and stools were

xamined for red staining to assess intestinal transit time.t 6 weeks, the 3H-labeled UCB solution (�0.29 �Ci/100body weight [BW]) was administered via the penile

ein.23 Subsequently, heparinized samples of tail veinlood were collected every 12 hours for 60 hours foretermination of plasma UCB concentrations, and fecesere collected to determine fecal excretion of urobili-oids, 3H-label, bile salts, calcium, phosphate, and fat. At60 hours after the 3H-UCB-injection, bile was collected

or 30 minutes, followed by vena cava inferior punctures described above. The intestine was then removed andivided into 5 segments (3 equal parts of small intestine,he cecum, and the remaining colon) that were flushedith phosphate-buffered saline (pH 7.4) for analysis ofCB and urobilinoids.

Plasma analysis. Blood samples were protectedrom light and processed immediately. Bilirubin, AST,

ia(lhNpatp

zlttUGibmtl

aanlafgHlcda2atgofdSasmammct

3

U�r

3

siacamc

tLpafcnapwiS

Ulw2pswtTbdsasb

tCcttbwr

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

February 2009 BILE SALT THERAPY FOR UNCONJUGATED HYPERBILIRUBINEMIA 675

cal spectrophotometry on a P800 unit of a modularnalytics serum work area from Roche DiagnosticsBasel, Switzerland). Hemoglobin, hematocrit, and reticu-ocyte counts were determined on a Sysmex-XE-2100ematology analyzer (Goffin Meyvis, Etten-Leur, Theetherlands). UCB levels were confirmed by reversed-hase high-performance liquid chromatography (HPLC)fter chloroform extraction as described before.19 3H con-ent was determined by liquid scintillation as describedreviously.19

Bile analysis. Bile samples were immediately fro-en under argon and processed within 24 hours. UCBevels were determined by HPLC after chloroform extrac-ion as described above. Urobilinoid levels were de-ermined as zinc complexes of total urobilinoids on aV-2401PC spectrophotometer (Shimadzu, Duisburg,ermany).25 Bile salt concentration was determined us-

ng the 3�-hydroxysteroid dehydrogenase method,26 andile salt composition was measured by capillary gas chro-atography after conversion of bile salts to methyl ester–

rimethylsilyl derivatives.27 3H content was determined byiquid scintillation as described previously.23

Analysis of feces and intestinal content. Fecesnd intestinal content were immediately frozen underrgon, freeze dried for 24 hours, mechanically homoge-ized, and promptly analyzed for UCB and urobilinoid

evels as described previously.19,25 Bile acid concentrationnd composition were determined as described be-ore.26,27 Fatty acid levels in feces were determined byas chromatography on an HP-Ultra-1 column fromewlett-Packard (Palo Alto, CA) after extraction, hydro-

ysis, and methylation of aliquots of feces.19,28 Fecal cal-ium levels were determined in duplicate 2.5-g aliquots ofried feces. Hydrogen chloride (2.2 mol/L; 25 mL) wasdded, and the mixture was refluxed for 10 minutes at00°C. After cooling, ammonia (10% weight/volume) wasdded to adjust the pH to 4, and the mixture was filteredhrough grade-1 filter paper from Whatman (Kent, En-land). The filtrate was analyzed spectrophotometricallyn a P800 unit of a modular analytics serum work arearom Roche Diagnostics. Fecal phosphate levels wereetermined in duplicate 1.0-g aliquots of dried feces.ulfuric acid (96% weight/volume; 7.5 mL) was added,nd the mixture was heated at 200°C for 10 minutes andubsequently at 330°C for 60 minutes. After cooling, 5

L of hydrogen peroxide (33% weight/weight) wasdded, and the mixture was heated at 330°C for 15inutes and filtered through grade-1 filter paper (What-an). The filtrate was analyzed as described above for

alcium. 3H content was determined by liquid scintilla-ion as described before.23

Calculation of fluxes based on steady-stateH-UCB kinetics. The natural logarithm of plasma 3H-CB specific activity (disintegrations per minute [dpm]/mol) was plotted against time, and the best-fit linear

egression curves were calculated. Fractional turnover of d

H-UCB (%/h) was obtained from the slope of the regres-ion line, and bilirubin pool size was calculated by divid-ng the specific activity at 0 hour (Y-axis intercept) by thedministered dose (dpm) of 3H-UCB. Total turnover wasalculated as the product of 3H-UCB fractional turnovernd pool size.29 Fractional biliary and fractional net trans-ucosal fluxes of UCB and of UCB derivatives were

alculated as described previously.23

Statistical analysis. Normally distributed datahat displayed homogeneity of variance (by calculation ofevene’s statistic) were expressed as mean � SD, andarametric statistical analysis was used. Analysis of vari-nce with post-hoc Bonferroni correction was performedor comparisons between groups, and Student t test foromparison of paired data within groups. If the data wereot normally distributed, nonparametric Kruskal–Wallisnd Mann–Whitney U tests (with corrected P values) wereerformed for comparison between groups, and the dataere expressed as median and range. The level of signif-

cance was set at P � .05. Analyses were performed usingPSS 14.0 for Windows (SPSS, Chicago, IL).

ResultsEffects of UDCA and CA Treatment:Dose DependencyFigure 1A shows the effect of increasing doses of

DCA or CA on plasma bilirubin levels in Gunn rats. Theowest-used UDCA or CA dose of 0.05 wt% (by choweight) resulted in a mean daily bile salt intake of 32 �mg/kg BW. This dose already effectively decreased

lasma bilirubin concentrations (�17% and �25%, re-pectively; each P � .05). At the highest dose used (1.5t%), UDCA and CA decreased plasma bilirubin concen-

rations by 42% and 50%, respectively (each P � .001).he rats in both groups did not differ significantly inody weight before the experiments, and none of theoses of UDCA or CA affected growth rate (data nothown). We selected a dose of 0.5 wt%, corresponding todaily bile salt intake of 317 � 23 mg/kg BW, for further

tudies to have a substantial effect yet minimize possibleile salt toxicity.

Effects of Short-Term Administration ofUDCA and CARapid decrease in plasma bilirubin concentra-

ions. Figure 1B shows that 8 days of dietary UDCA orA administration (each 0.5 wt%) decreased plasma UCB

oncentrations in Gunn rats by 21% and 30%, respec-ively, compared with controls (each P � .01). Adminis-ration of UDCA or CA was statistically hypobiliru-inemic within 3 and 5 days, respectively. Mean bodyeight and growth rate did not differ significantly among

ats in the control, UDCA, or CA group either before or

tf(ptmcsa(drSa

nhubctoPrbc(n(u

tCtpbao

siC�ectIimubg

Ftftsdt2mb0d317 � 23, 633 � 50, and 948 � 73.

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

676 CUPERUS ET AL GASTROENTEROLOGY Vol. 136, No. 2

Rapid increase in fecal and biliary bile salt excre-ion. Figure 2A shows that dietary UDCA or CA increasedecal bile salt excretion in the first 4 days of administration�663% and �466%, respectively; each P � .001) as com-ared with the 4-day pretreatment period. UDCA adminis-ration increased fecal excretion of UDCA, lithocholate, and

uricholates, while CA administration increased fecal ex-retion of CA and deoxycholate (Figure 2B). Figure 2Chows that the increase in fecal bile salt excretion wasccompanied by an increase in biliary bile salt secretionUDCA �127%, CA �128%; each P � .01), measured after 8ays of treatment. Changes in fecal bile salt compositioneflected changes in biliary bile salt composition (Figure 2D,upplementary Figure 1 [see supplementary material onlinet www.gastrojournal.org]).

Rapid increases in fecal, but not biliary, urobili-oid and UCB excretion. If bile salt administration en-ances bilirubin disposal, the fecal excretion of UCB androbilinoids, a family of intestinally formed bacterialreakdown products of UCB, would be expected to in-rease upon starting treatment. Figure 3A and B showhat UDCA and CA indeed increased the fecal excretionf both urobilinoids (�42% and �48%, respectively; each� .01) and UCB (�56% [P � .05] and �25% [P � .06],

espectively) in the first 4 days of treatment. The com-ined fecal excretion of urobilinoids and UCB was in-reased by 52% with UDCA and 32% with CA treatmentFigure 3C; each P � .01). In contrast, UDCA and CA didot influence the biliary excretion of urobilinoids or UCB

data not shown), or the combined biliary excretion ofrobilinoids � UCB, after 8 days of treatment (Figure 3D).

Effects of Long-Term Administration ofUDCA and CASustained decrease in plasma bilirubin concen-

rations. Figure 1C shows that 6 weeks of dietary UDCA orA (each 0.5 wt%) decreased plasma bilirubin concentra-

ions by �40% in Gunn rats from week 2 onwards, com-ared with stable values in controls (each P � .001). Meanody weight and growth rate did not differ significantlymong rats in the control, UDCA, or CA group either beforer during the 6-week period (data not shown).

Changes in biliary and fecal excretion of bilealts, urobilinoids, and UCB. Table 1 shows that, mim-cking the short-term experiment, 6-weeks of UDCA orA administration increased fecal bile salt (�566% and652%, respectively; each P � .001) and fecal urobilinoid

xcretion (�98% and �103%, respectively; each P � .01),ompared with controls. UDCA, but not CA, increasedhe fecal excretion of UCB (�256%, P � .001; Table 1).nterestingly, only administration of CA, but not UDCA,ncreased biliary bile salt secretion after 6 weeks of treat-

ent (�306%, P � .001; Table 1). Changes in fecalrobilinoid and UCB excretion were not reflected in theile, since their biliary excretion was similar among all 3

igure 1. UDCA or CA administration decreases plasma UCB concen-ration in Gunn rats. Gunn rats (n � 6 per group) were fed the control dietor 6 weeks, followed by: dietary UDCA or CA supplementation in doseshat were increased every 2 weeks for 10 weeks (A); dietary UDCA or CAupplementation (0.5 wt% each), or no supplementation for 8 days (B); orietary UDCA or CA supplementation (0.5 wt% each), or no supplemen-ation for 6 weeks (C). Plasma UCB values at T0 (�mol/L) in B: controls,41 � 22; UDCA, 260 � 25; CA, 251 � 26 (not significant). Data representean � SD. *P � .05; **P � .01; ***P � .001, compared with plasmailirubin at T0 (A), or compared with controls (B and C). The used doses of.05 wt%, 0.1 wt%, 0.5 wt%, 1.0 wt%, and 1.5 wt% corresponded with aietary bile salt intake (mg/24 h/kg BW) of, respectively, 32 � 2, 62 � 5,

roups (Table 1).

cfwrwm.te

tpita

ttimecpwm

dnafila(cwcfictp�(nbr

tgHqs

FCc(b(cptgfU(tlfarcr*amv(asMccdc

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

February 2009 BILE SALT THERAPY FOR UNCONJUGATED HYPERBILIRUBINEMIA 677

No increases in fecal excretion of fatty acids,alcium, or phosphate. An increased fecal excretion ofatty acids, calcium, or phosphate has been associatedith decreased plasma UCB concentrations in Gunn

ats.13,19 Table 1 shows that UDCA administration for 6eeks did not affect fecal fat excretion, whereas CA ad-inistration even decreased fecal fat excretion (–75%, P �

001). Table 1 also shows that UDCA and CA adminis-ration did not influence fecal calcium or phosphatexcretion.

CA administration decreases intestinal transitime. A decrease in intestinal transit time may decreaselasma UCB concentrations in Gunn rats.30 CA admin-

stration for 6 weeks moderately decreased the intestinalransit time (�18%, P � .001; Table 1), whereas UDCAdministration at the same dosage showed no effect.

3H-Bilirubin Turnover StudiesDecreased pool size and increased fractional

urnover of bilirubin. To obtain more detailed mechanis-ic insights, we determined steady-state 3H-UCB kineticsn the Gunn rats over 60 hours, after 6 weeks of treat-

ent with control, UDCA, or CA diet. During the kineticxperiment, the hematocrit, hemoglobin level, reticulo-yte count, and AST and ALT levels were unaltered (Sup-lementary Table 1 [see supplementary material online atww.gastrojournal.org]). Also, plasma bilirubin levels re-

igure 2. Short-term UDCA orA administration to Gunn rats in-reases fecal bile salt excretionA); changes the composition ofile salts excreted via the feces

B); increases biliary bile salt ex-retion (C); and changes the com-osition of bile salts excreted viahe bile (D). Gunn rats (n � 6 perroup) were fed the control diet

or 6 weeks, followed by dietaryDCA or CA supplementation

0.5 wt% each), or no supplemen-ation for 8 days. Feces were col-ected during a 4-day period be-ore (pretreatment period) andfter (treatment period) dietaryandomization. At 8 days, bile wasollected during 30 minutes. Dataepresent mean � SD. *P � .01;*P � .001; †P � .05. Statisticalnalysis in feces: 4-day pretreat-ent period (area under the curve)

ersus 4-day treatment periodarea under the curve). Statisticalnalysis in bile: UDCA or CA ver-us controls. LC, lithocholic acid;, muricholic acid; DC, deoxy-

holic acid; C, cholic acid; CDC,henodeoxycholic acid; HDC, hyo-eoxycholic acid; UDC, ursodeoxy-holic acid; HC, hyocholic acid.

ained stable, and the plasma 3H-UCB specific activity o

eclined in a semilogarithmic manner in all groups (dataot shown). These findings are in accordance with thebsence of significant hemolysis and with the presence ofrst-order steady-state conditions. Analysis of the semi-

ogarithmic specific activity curves showed that UDCAnd CA treatment decreased bilirubin pool sizes by 33%P � .01) and 32% (P � .05), respectively, compared withontrols (Table 2). As shown in Figure 4, the pool sizesere strongly, positively correlated with plasma UCB

oncentrations (y � 50x � 40; r � 0.91; P � .001). Theractional turnover of 3H-UCB increased by 33% and 25%n UDCA- and CA-treated animals, respectively, whenompared with controls (each P � .05; Table 2). Frac-ional turnover was negatively correlated with bothlasma UCB concentrations (y � �177x � 506; r �0.89; P � .001) and the calculated bilirubin pool sizes

y � �3.2*x � 8.8; r � �0.92; P � .001). Neither UDCAor CA administration significantly affected total biliru-in turnover, in accordance with similar UCB productionates in the 3 groups (Table 2).

Increased efficiency of biliary and intestinalransmucosal UCB excretion. In each experimentalroup, the excretion of UCB into the bile, measured byPLC, comprised �20% of its total turnover. Because the

uantitative excretion of bilirubin occurs almost exclu-ively via the feces, this implies that the remaining �80%

tocsUpifi

ttt(wsu

ipom

Cpkascfbpmstoctar

etBb

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

678 CUPERUS ET AL GASTROENTEROLOGY Vol. 136, No. 2

ion. The treated groups thus excreted similar amountsf UCB via either excretory pathway, compared withontrols. However, because of smaller bilirubin poolizes, all fractional fluxes (flux per hour as fraction of theCB pool size) are increased in the treated groups, com-ared with controls. Figure 5 shows that UDCA mainly

ncreased the fractional transmucosal UCB flux and theractional biliary flux of derivatives, whereas CA predom-nantly increased the fractional biliary flux of UCB.

UDCA, but not CA, administration increases in-estinal UCB and urobilinoid content. Table 3 shows thathroughout the bowel, UCB and urobilinoid contentended to be higher in the UDCA-treated animals�143% and �106%, respectively; P � .01), comparedith controls. By contrast, CA administration did not

ignificantly alter total intestinal content of UCB androbilinoids.

DiscussionIn this study, we demonstrate that dietary admin-

stration of either UDCA or CA significantly decreaseslasma UCB concentrations in Gunn rats. The decreaseccurs within 3 days after starting administration, is

The conclusion that the administration of UDCA orA enhances fecal bilirubin disposal is based on 2 inde-endent analytic moieties, namely, biochemical and 3Hinetic measurements. First, the fecal excretion of UCBnd urobilinoids increased promptly upon starting bilealt treatment. Previous studies have shown that bio-hemical measurements of UCB and urobilinoids in theeces only account for 25%–50% of the expected fecalilirubin disposal.31 To address this and to investigateossible alterations in UCB metabolism by the treat-ents, we performed a 3H-UCB kinetic experiment in a

teady-state condition.23 This experiment demonstratedhat either bile salt decreased the bilirubin pool size byne third, mirroring similar decreases in plasma UCBoncentrations. Combined, these methodologies indicatehat UDCA or CA enhances UCB disposal from the bodynd that the hypobilirubinemic effects are not due toedistribution among different body compartments.32

During steady-state conditions, bilirubin productionquals its excretion, and both are synonymous with theotal turnover of bilirubin (expressed as nmol/h/100 gW; Table 2). If bile salt treatment has no effect onilirubin production, the total excretory flux of bilirubin

Figure 3. Short-term UDCA orCA administration to Gunn ratsincreases fecal urobilinoid excre-tion (A); increases fecal UCB ex-cretion (B); increases fecal urobi-linoid � UCB excretion (C); anddoes not affect biliary urobilin-oid � UCB excretion (D). For ex-perimental setup, please refer toFigure 2. Data represent mean �SD. †P � .06; *P � .05; **P �.01. Statistical analysis in feces:4-day pretreatment period (areaunder the curve) versus 4-daytreatment period (area under thecurve). Statistical analysis in bile:UDCA or CA versus controls.

s thus similar in the treated groups and the control

gebbel

dtci

3

socutbTtttowrth

fltaetfa

FrctomSt

T

F

B

Noma

b

T

P

3

B

T

B

F

Nsa

b

c

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

February 2009 BILE SALT THERAPY FOR UNCONJUGATED HYPERBILIRUBINEMIA 679

roup during steady-state. To describe differences in thefficiency of bilirubin disposal, the fractional turnover ofilirubin—that is, the total turnover as percent of theilirubin pool size (expressed as %/h; Table 2)—is prefer-ntially used. The kinetic study showed that the hypobi-irubinemic effect of bile salt treatment was not due to a

able 1. Steady-State Excretion of Several Fecal and BiliaryComponents and Intestinal Transit Time 6 WeeksAfter Dietary Randomization

ControlsUDCA0.5% CA 0.5%

ecesBile salts (�mol/h/100

g BW)0.3 � 0.0 2.1 � 0.1a 2.4 � 0.4a

UCB (nmol/h/100 g BW) 1.6 � 0.5 5.8 � 3.8a 2.0 � 0.6Urobilinoids (nmol/h/

100 g BW)1.8 � 0.2 3.6 � 0.7b 3.7 � 0.5b

UCB � urobilinoids(nmol/h/100 g BW)

3.4 � 0.5 9.4 � 3.5a 5.7 � 0.9b

Fat (�mol/h/100 g BW) 1.2 � 0.5 1.1 � 0.4 0.3 � 0.0a

Calcium (mmol/h/100 gBW)

33 � 5 34 � 9 33 � 4

Phosphate (mmol/h/100 g BW)

26 � 5 24 � 7 20 � 4

Transit time (h) 11 � 1 11 � 1 9.0 � 1a

ileBile salts (�mol/h/100

g BW)7.1 � 1 10 � 4 29 � 5a

UCB (nmol/h/100 g BW) 11 � 3 8.2 � 3 11 � 2Urobilinoids (nmol/h/

100 g BW)5.9 � 4 8.7 � 8 5.1 � 3

UCB � urobilinoids(nmol/h/100 g BW)

17 � 2 17 � 10 17 � 4

OTE. For experimental setup, please refer to Figure 4. After 5 weeksf treatment, intestinal transit time was determined. Data representean � SD.

P � .001, compared with controls.P � .01, compared with controls.

able 2. Steady-State 3H-Bilirubin Kinetics 6 Weeks AfterDietary Randomization

Controls UDCA 0.5% CA 0.5%

lasma bilirubin at T0h(�mol/L)

287 � 40 187 � 25a 181 � 14a

H-UCB fractionalturnover (%/h)

1.4 � 0.2 1.9 � 0.2b 1.8 � 0.1c

ilirubin pool size(�mol/100 g BW)

4.2 � 0.9 2.8 � 0.7c 2.9 � 0.0c

otal bilirubin turnover(nmol/h/100 g BW)

59 � 4 52 � 8 51 � 2

iliary 3H excretion T60h(10^3 dpm/h)

10 � 1 12 � 4 11 � 3

ecal 3H excretion T0–60h(10^3 dpm/h)

5.0 � 0.5 5.4 � 0.5 5.8 � 0.4

OTE. For experimental setup, please refer to Figure 4. Data repre-ent mean � SD.P � .001, compared with controls.P � .01, compared with controls.

iP � .05, compared with controls.

ecrease in UCB production, as reflected by the similarotal bilirubin turnover, but rather to an increased effi-iency of UCB disposal from the body, as reflected by thencreased fractional turnover compared with controls.

Fractional turnover, pool size, and total turnover ofH-UCB of the control animals in this study were reas-uringly similar to the values obtained in previous studiesf radiolabeled UCB kinetics in Gunn rats.3,23,32 We cal-ulated steady-state fluxes of UCB and its derivativessing a previously used mathematical model, based onhe assumption that the quantitative disposal of biliru-in and its derivatives occurs exclusively via the feces.23

his assumption seems reasonable since only �6% ofotal bilirubin turnover in Gunn rats is disposed of viahe urine.7 Our calculations showed that only �20% ofhe total bilirubin turnover in each group was disposedf via the bile, and �80% of the total bilirubin turnoveras disposed of via net transmucosal excretion. This

esult underlines the previously described importance ofhe transmucosal excretory route during unconjugatedyperbilirubinemia.6,7,23

Analysis of the fractional excretory fluxes (Figure 5), orux as a proportion of the bilirubin pool size, suggestshat the principal mechanisms differ by which UDCAnd CA lower plasma UCB levels. The hypobilirubinemicffect of UDCA is mainly due to an increased fractionalransmucosal excretion of UCB, whereas CA increases theractional biliary excretion of UCB. Unlike the controlnd CA groups, UDCA-fed animals showed a marked

igure 4. Bilirubin pool sizes and plasma UCB concentrations in Gunnats were positively correlated. Gunn rats (n � 6 per group) were fed theontrol diet for 6 weeks, followed by dietary UDCA or CA supplemen-ation (0.5 wt% each), or no supplementation for 6 weeks. After 6 weeksf treatment, tracer 3H-UCB was intravenously administered to deter-ine UCB kinetics. Plasma was collected every 12 hours over 60 hours.pecific activity of plasma bilirubin declined semilogarithmically with

ime in each group.

ncrease in UCB content in the colon—that is, rather

dnmtibetsbtrteoaUpftItfuf

tabdcp

pda

pawpowHcbs

mttnfrUmddccfai

Fatdti n � S

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

680 CUPERUS ET AL GASTROENTEROLOGY Vol. 136, No. 2

istal from the biliary excretion of UCB into the duode-um. Based on this finding, we hypothesize that UDCAainly exerts its hypobilirubinemic effect via intestinal

rapping of UCB. This hypothesis is supported by thencreased colonic content and the increased fractionaliliary flux of urobilinoids in the UDCA group, reflectingnhanced formation and enterohepatic circulation ofhese bacterial metabolites, probably due to the increasedupply of UCB in the intestinal lumen. The mechanismy which CA treatment lowers plasma bilirubin concen-rations in our Gunn rats is less clear. Intravenous tau-ocholic acid administration enhanced biliary UCB excre-ion in Gunn rats.17 We observed no increase in biliaryxcretion of UCB and urobilinoids after 8 days or 6 weeksf CA treatment, although we did not measure directlyfter starting bile salt treatment. The steady-state 3H-CB kinetic study and the analysis of intestinal bileigment content favor greater enhancement by CA of theractional biliary excretion of UCB, but less effectiverapping of UCB in the intestine compared with UDCA.n a binary system, UCB binds less avidly to UDCA thano CA, but in the presence of lecithin, UDCA favors theormation of vesicles over micelles, which enhances sol-bilization of cholesterol and might conceivably do so

or UCB.33

We acknowledge that other mechanisms may also con-ribute to the increase in UCB disposal during bile saltdministration. The intestinal transit time was decreasedy 18% in the 0.5 wt% CA-treated group, which wouldecrease contact time for mucosal reabsorption.30 In-reased amounts of fat and amorphous calcium phos-

igure 5. Fractional biliary and transmucosal fluxes of UCB and UCB dre expressed as %, and fractional UCB-derivative fluxes are expressedurnover of UCB; [b] fractional biliary UCB excretion; [c] fractional biliaerivatives; [e] estimated net transmucosal flux of UCB from the blood inhat of [a] in a steady-state, assuming that UCB turnover equals the fecan upper left of each panel show bilirubin pool size (�mol/100 g BW; mea

hate could lower plasma bilirubin via intestinal trap- d

ing.13,19 However, their fecal excretion was not increaseduring bile salt treatment. Also, UDCA treatment couldffect the expression of relevant transporters.34

Among all experimental groups, we observed a strong,ositive, linear correlation between bilirubin pool sizend plasma UCB concentrations, similar to our findingshen treating Gunn rats with orlistat and photothera-y.23 During treatment of Gunn rats with phototherapy,rlistat, or bile salts, changes in UCB pool sizes are thusell predicted by changes in plasma UCB concentrations.owever, 3H-UCB kinetic studies, using our mathemati-

al model (Figure 5), remain necessary to estimate theiliary and transmucosal UCB fluxes, and their relation-hip to UCB pool size.

Mendez-Sanchez et al35 showed that in nonjaundicedice and rats, dietary UDCA supplementation increased

he enterohepatic UCB circulation. This opposite findingo our study, in which UDCA decreased net UCB intesti-al reabsorption, might result from strain-induced dif-

erences in bilirubin metabolism. Homozygous Gunnats cannot conjugate bilirubin, and the accumulatedCB diffuses from the plasma into the intestinal lu-en.6,7 This diffusion is inversely directed in nonjaun-

iced rodents, as used in the study of Mendez-Sanchez,ue to lower plasma UCB levels.8,9 UDCA-inducedhanges in microfloral hydrolysis of bilirubin conjugatesould further enhance this diffusion. The observed dif-erences between rat strains illustrate that therapeuticpplication of UDCA in patients with an etiology that isncompatible to our animal model, such as chronic liver

tives in Gunn rats, 6 weeks after randomization. Fractional UCB fluxesuivalent% of the bilirubin pool size that is excreted per hour. [a] fractionalCB-derivative excretion; [d] fractional fecal excretion of UCB � UCBintestinal lumen, calculated as [a] – [b]. The magnitude of flux [d] equalsetion of UCB � UCB derivatives. EHC, enterohepatic circulation. OvalsD). For calculation of fractional fluxes, please refer to Methods section.

erivaas eqry U

to thel excr

isease, should be approached with caution.

btutwapdtUdsl

CUvPtb

aG1

1

1

1

1

1

1

1

1

1

1

2

T

S

S

S

C

L

T

N6ipwUa

U

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

February 2009 BILE SALT THERAPY FOR UNCONJUGATED HYPERBILIRUBINEMIA 681

The rapid and persistent hypobilirubinemic effect ofile salt administration, however, does support the po-ential clinical applicability of oral UDCA therapy fornconjugated hyperbilirubinemia in humans. The majorherapeutic effect was already present in a low dose (0.05t%), corresponding to a clinically applicable dosage ofpproximately 30 mg/kg/day. Bile salt therapy decreasedlasma UCB concentrations in Gunn rats as effectively asid phototherapy in our previous experiments.15,23 Allreated Gunn rats ingested comparable amounts of eitherDCA or CA, and treatment with neither bile salt in-uced diarrhea or impaired growth rates. This corre-ponds with the fact that UDCA treatment is well estab-ished and well tolerated in pediatric patients.36

In conclusion, dietary administration with UDCA orA induces a rapid and sustained decrease in plasmaCB concentrations in Gunn rats. The mechanism in-

olves stimulation of UCB turnover and its fecal disposal.resent results support the feasibility of oral bile saltreatment in patients with unconjugated hyperbiliru-

able 3. Intestinal UCB and Urobilinoid Content 6 WeeksAfter Dietary Randomization

Controls UDCA 0.5% CA 0.5%

mall intestine(proximal)

UCB (nmol) 4 (1–11) 2 (1–8) 3 (1–5)Urobilinoids

(nmol)2 (1–3) 2 (1–5) 3 (2–5)

mall intestine(medial)

UCB (nmol) 18 (13–22) 16 � (4–19) 14 � (1–19)Urobilinoids

(nmol)8 (5–9) 9 (3–14) 9 (3–12)

mall intestine(distal)

UCB (nmol) 19 (2–50) 90 (23–230)a 29 (4–127)Urobilinoids

(nmol)11 (5–17) 28 (12–43)a 17 (4–21)

ecumUCB (nmol) 10 (7–13) 28 (11–421)a 19 (13–61)a

Urobilinoids(nmol)

47 (25–100) 117 (65–134)a 106 (64–144)

arge intestineUCB (nmol) 14 (7–23) 116 (50–163)b 8 (4–13)c

Urobilinoids(nmol)

34 (23–46) 132 (50–169)b 38 (16–101)c

otal intestineUCB (nmol) 70 (32–92) 227 (108–790)a 72 (41–166)c

Urobilinoids(nmol)

95 (84–158) 294 (138–340)a 154 (106–254)

OTE. For the experimental setup, please refer to Figure 4. After the0-hour period of the 3H-UCB kinetic study, animals were killed. The

ntestine was then removed and divided into 5 segments (3 equalarts of small intestine, the cecum, and the remaining colon) thatere flushed with phosphate-buffered saline (pH 7.4) for analysis ofCB and urobilinoids. Data represent median and range.P � .05; bP � .01, compared with controls. cP � .05, compared withDCA-treated animals. Nonparametric tests were used.

inemia.

Supplementary Data

Note: To access the supplementary materialccompanying this article, visit the online version ofastroenterology at www.gastrojournal.org, and at doi:0.1053/j.gastro.2008.10.082.

References

1. Crigler JF Jr, Najjar VA. Congenital familial nonhemolytic jaundicewith kernicterus. Pediatrics 1952;10:169–180.

2. Shapiro SM. Definition of the clinical spectrum of kernicterus andbilirubin-induced neurologic dysfunction (BIND). J Perinatol 2005;25:54.

3. Ostrow JD. Photocatabolism of labeled bilirubin in the congeni-tally jaundiced (Gunn) rat. J Clin Invest 1971;50:707–718.

4. Van der Veere CN, Sinaasappel M, McDonagh AF, et al. Currenttherapy for Crigler-Najjar syndrome type 1: report of a worldregistry. Hepatology 1996;24:311.

5. Yohannan MD, Terry HJ, Littlewood JM. Long term phototherapy inCrigler-Najjar syndrome. Arch Dis Child 1983;58:460.

6. Kotal P, Van der Veere CN, Sinaasappel M, et al. Intestinalexcretion of unconjugated bilirubin in man and rats with inheritedunconjugated hyperbilirubinemia. Pediatr Res 1997;42:195.

7. Schmid R, Hammaker L. Metabolism and disposition of C14-bilirubin in congenital nonhemolytic jaundice. J Clin Invest 1963;42:1720–1734.

8. Lester R, Schmid R. Intestinal absorption of bile pigments. I. Theenterohepatic circulation of bilirubin in the rat. J Clin Invest1963;42:736.

9. Lester R, Schmid R. Intestinal absorption of bile pigments. II.Bilirubin absorption in man. N Engl J Med 1963;269:178.

0. Odell GB, Gutcher GR, Whitington PF, et al. Enteral administrationof agar as an effective adjunct to phototherapy of neonatalhyperbilirubinemia. Pediatr Res 1983;17:810.

1. Tan KL, Jacob E, Liew DS, et al. Cholestyramine and photother-apy for neonatal jaundice. J Pediatr 1984;104:284–286.

2. Ulstrom RA, Eisenklam E. The enterohepatic shunting of bilirubinin the newborn infant. I. Use of oral activated charcoal to reducenormal serum bilirubin values. J Pediatr 1964;65:27–37.

3. Van der Veere CN, Jansen PL, Sinaasappel M, et al. Oral calciumphosphate: a new therapy for Crigler-Najjar disease? Gastroen-terology 1997;112:455.

4. Mendez-Sanchez N, Martinez M, Gonzalez V, et al. Zinc sulfateinhibits the enterohepatic cycling of unconjugated bilirubin insubjects with Gilbert’s syndrome. Ann Hepatol 2002;1:40–43.

5. Hafkamp AM, Nelisse-Haak R, Sinaasappel M, et al. Orlistattreatment of unconjugated hyperbilirubinemia in Crigler-Najjar dis-ease: a randomized controlled trial. Pediatr Res 2007;62:725–730.

6. Vonk RJ, Jekel P, Meijer DK. Choleresis and hepatic transportmechanisms. II. Influence of bile salt choleresis and biliary mi-celle binding on biliary excretion of various organic anions. Nau-nyn Schmiedebergs Arch Pharmacol 1975;290:375–387.

7. Ostrow JD. Regulation by bile salts of the excretion of conjugatedand unconjugated bilirubin in bile. In: Fromm H, Leuschner U,eds. Proceedings of the Falk Symposium No. 84, Berlin, Ger-many, June 9–10, 1995. Lancaster, UK: Kluwer Academic Pub-lishers, 1996:78–97.

8. Einarsson K, Bjorkhem I, Eklof R, et al. Effect of ursodeoxycholicacid treatment on intestinal absorption of triglycerides in man.Scand J Gastroenterol 1984;19:283.

9. Hafkamp AM, Havinga R, Sinaasappel M, et al. Effective oraltreatment of unconjugated hyperbilirubinemia in Gunn rats. Hepa-tology 2005;41:526.

0. Ostrow JD, Celic L, Mukerjee P. Molecular and micellar associa-

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

URe

o

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

682 CUPERUS ET AL GASTROENTEROLOGY Vol. 136, No. 2

unconjugated bilirubin by bile salts. J Lipid Res 1988;29:335–348.

1. Rege RV, Webster CC, Ostrow JD. Interactions of unconjugatedbilirubin with bile salts. J Lipid Res 1988;29:1289–1296.

2. Bayon JE, Pascolo L, Gonzalo-Orden JM, et al. Pitfalls in prepa-ration of (3)H-unconjugated bilirubin by biosynthetic labeling fromprecursor (3)H-5-aminolevulinic acid in the dog. J Lab Clin Med2001;138:313–321.

3. Hafkamp AM, Havinga R, Ostrow JD, et al. Novel kinetic insightsinto treatment of unconjugated hyperbilirubinemia: phototherapyand orlistat treatment in Gunn rats. Pediatr Res 2006;59:506.

4. Webster C, Tiribelli C, Ostrow JD. An improved method for isola-tion of unconjugated bilirubin from rat and dog bile. J Lab ClinMed 2001;137:370–373.

5. Kotal P, Fevery J. Quantitation of urobilinogen in feces, urine, bileand serum by direct spectrophotometry of zinc complex. ClinChim Acta 1991;202:1.

6. Murphy GM, Billing BH, Baron DN. A fluorimetric and enzymaticmethod for the estimation of serum total bile acids. J Clin Pathol1970;23:594–598.

7. Hulzebos CV, Renfurm L, Bandsma RH, et al. Measurement ofparameters of cholic acid kinetics in plasma using a microscalestable isotope dilution technique: application to rodents andhumans. J Lipid Res 2001;42:1923–1929.

8. Lepage G, Roy CC. Direct transesterification of all classes oflipids in a one-step reaction. J Lipid Res 1986;27:114–120.

9. Zilversmit DB. The design and analysis of isotope experiments.Am J Med 1960;29:832.

0. Kotal P, Vitek L, Fevery J. Fasting-related hyperbilirubinemia inrats: the effect of decreased intestinal motility. Gastroenterology

1. Tiribelli C, Ostrow JD. Intestinal flora and bilirubin. J Hepatol2005;42:170.

2. Cohen AN, Kapitulnik J, Ostrow JD, et al. Effects of phenobarbitalon bilirubin metabolism and its response to phototherapy in thejaundiced Gunn rat. Hepatology 1985;5:310.

3. Salvioli G, Igimi H, Carey MC. Cholesterol gallstone dissolution inbile. Dissolution kinetics of crystalline cholesterol monohydrateby conjugated chenodeoxycholate-lecithin and conjugated ur-sodeoxycholate-lecithin mixtures: dissimilar phase equilibria anddissolution mechanisms. J Lipid Res 1983;24:701–720.

4. He YJ, Zhang W, Tu JH, et al. Hepatic nuclear factor 1alphainhibitor ursodeoxycholic acid influences pharmacokinetics of theorganic anion transporting polypeptide 1B1 substrate rosuvasta-tin and bilirubin. Drug Metab Dispos 2008;36:1453–1456.

5. Mendez-Sanchez N, Brink MA, Paigen B, et al. Ursodeoxycholicacid and cholesterol induce enterohepatic cycling of bilirubin inrodents. Gastroenterology 1998;115:722.

6. Balistreri WF. Bile acid therapy in pediatric hepatobiliary disease:the role of ursodeoxycholic acid. J Pediatr Gastroenterol Nutr1997;24:573–589.

Received June 5, 2008. Accepted October 30, 2008.Address requests for reprints to: Henkjan J. Verkade, MD, PhD,

niversity Medical Center Groningen, Pediatric Gastroenterology,oom Y4.107A, PO Box 30.001, 9700 RB Groningen, The Netherlands.-mail: [email protected]; fax: 0031�503611671.The authors thank Milan Jirsa (Laboratory of Experimental Hepatol-

gy, Institute for Clinical and Experimental Medicine, Prague, Czech

SGpP

February 2009 BILE SALT THERAPY FOR UNCONJUGATED HYPERBILIRUBINEMIA 682.e1

upplemental Figure 1. Short-term UDCA or CA administration tounn rats changes biliary bile salt composition. For experimental setup,lease refer to Figure 2. Data represent mean � SD. †P�0.05–

S

HHRAA

Ns

�0.001, compared with controls.

upplementary Table 1. Hematological and liver functionparameters 6 weeks after dietaryrandomisation.

controls UDCA 0.5% CA 0.5%

emoglobin (mmol/l) 7.2 � 0.7 7.3 � 0.6 7.7 � 0.4ematocrit (v/v) 0.36 � 0.03 0.37 � 0.03 0.38 � 0.03eticulocytes (‰) 58 � 17 49 � 15 42 � 8ST (U/l) 90 � 23 81 � 20 75 � 14LT (U/l) 70 � 22 49 � 6 50 � 7

OTE. For experimental setup, please refer to Figure 4. Data repre-