Unconjugated Bilirubin and an Increased Proportion ofBilirubin Monoconjugates in the Bile of Patients withGilbert's Syndrome and Crigler-Najjar Disease

JOHANFEVERY, NORBERTBLANCKAERT,and KAREL P. M. HEIRWEGH,Laboratory for Liver Physiopathology, Department of Medical Research,University of Leuven, 3000 Leuven, Belgium

ANNE-MARIE PREAUXand PIERRE BERTHELOT, Unite de Recherches U-99,Institut National de la Sante et de la Recherche Medicale, Hopital HenriMondor, 94010 Creteil, France

A B S T RA C T Bilirubin pigments were studied in thebile of 20 normal adults, 25 patients with Gilbert'ssyndrome, 9 children with Crigler-Najjar disease, and 6patients with hemolysis, to determine how a deficiencyof hepatic bilirubin UDP-glucuronosyltransferasewould affect the end products of bilirubin biotrans-formation.

In the bile from patients with Gilbert's syndrome, astriking increase was found in the proportion ofbilirubin monoconjugates (48.6±9.8% of total con-jugates) relative to that in normal bile (27.2±7.8%).This increase was even more pronounced in childrenwith Crigler-Najjar disease, in whom, even in the mostsevere cases, glucuronide could always be demon-strated in the bile. Furthermore, unconjugatedbilirubin-IXa was unquestionably present in the bileof these children and amounted to 30-57% of theirtotal bilirubin pigments (<1% in the controls). It wasnot possible to predict from the biliary bilirubin com-position whether a child would respond to pheno-barbital therapy or not. Bile composition was normal inpatients with hemolysis, except when there was as-sociated deficiency of hepatic glucuronosyltransferase.Therefore, the observed alterations were not a simpleconsequence of unconjugated hyperbilirubinemia.

The present findings suggest that Crigler-Najiardisease represents a more pronounced expression thanGilbert's syndrome of a common biochemical defect.Hepatic bilirubin UDP-glucuronosyltransferase de-ficiency leads to decreased formation of diconjugates

This work was presented in part at the 8th Annual Meetingof the European Association for the Study of the Liver, 7-8September 1973, Vittel, France (1).

Received for publication 16 July 1976 and in revised form17 May 1977.

with an ensuing increase in the proportion of biliru-bin monoconjugates in bilei in the most severe cases,an elevated content of biliary unconjugated bilirubinis also found.

INTRODUCTION

In both Gilbert's syndrome and Crigler-Najjar disease,hepatic bilirubin UDP-glucuronosyltransferasel ac-tivity (UDP-GTA)2 is significantly decreased (2-7), butdoes not correlate well with the serum bilirubin levelsin the individual cases (2, 3, 6, 7). Indeed, otherprocesses also influence the serum levels as illustrated,for example, by the fasting hyperbilirubinemia (7, 8),whereas the in vitro enzyme assay does not take ac-count of endogenous factors which may regulate thetransferase activity in vivo (9, 10).

An interesting approach to both disorders is based onthe delayed plasma disappearance rate of injectedbilirubin (11-15). Such analysis also permitted the de-tection and quantification of the overproduction ofbilirubin (13, 16, 17) which is often associated withGilbert's syndrome (6, 18). However, the precise mean-ing of the kinetic parameters is not yet known, as"the models are merely mathematical conceptualiza-tions of the sites and rates of bilirubin elimina-tion. They provide no information about the biochemi-cal mechanisms involved" (cited from Bloomeret al. [14]).

1 UDP-glucuronosyltransferase, or UDP-glucuronate 1-glucuronosyltransferase (acceptor-unspecific), EC 2.4.1.17.

2Abbreviations used in this paper: EA, ethyl anthranilate;PIA, p-iodoaniline; TLC, thin-layer chromatography; UDP-GTA, bilirubin UDP-glucuronosyltransferase activity.

The Journal of Clinical Investigation Volume 60 November 1977- 970-979970

Decreased hepatic bilirubin UDP-GTA anid de-creased hepatic clearance of labeled bilirubin canl bothbe used for diagnosis as they yield indices of theconjugation rate of bilirubin-IXa (by far the predomi-nant isomer in man). In the present work, another ap-proach has been investigated. It is indeed reasonableto assume that the relative amounts of bilirubin-IXaand of its mono- and diconjugates in bile reflect theconjugating processes operating in the hepatocyte,provided that the secreted pigments are inot signifi-cantly altered in the biliary system. For these reasons,we have analyzed the nature and amounts of coinjuga-ted and unconjugated bilirubins that are present inbile. Bile analysis was compared with the results ofhepatic bilirubin UDP-GTA. Recent developmenits inmethods for bile pigment analysis have made this ap-proach possible (19).

METHODSBile collection and patients. Bile was obtained by duo-

clenal intubation in 20 normal controls, 25 adults with

Gilbert's syncdrome, 8 children with Crigler-Najjar disease,and 6 patients with hemolytic jaundice who awaitedsplenectomiiy. The cause of the hemolysis was conigenitalspherocytosis in five, and autoimmune hemolysis in thesixth patient. In one additional child with Crigler-Najjardisease (child S), the bile sample was obtained by gallbladderpunicture at time of surgery. The conitrols were informiie(dpatients examinied for psychosomatic complaints or memiibersof the laboratory. Gilbert's syndromiie was dlefinecl as chronicunconijugated hyperbilirubinemia (>1.5 mg/100 ml) in theabsence of overt hemolysis and of any other disease; thesepatients included 14 males and 11 females; their agesranged from 15 to 64 yr.

The ninle children with Crigler-Najjar disease were un-related to each other; there were five males and four females,ranging from 1 mo to 3 yr of age; clinical data on1 sonme ofthem have been published (20, 21). The nine children withCrigler-Najjar disease (Table I) showedl serumil concenitrationisof unconjugated bilirubin ranginig from 20 to 35 mg/100 ml.During the survey period, kernicterus developed in sixchildren. In these six patients, adminiistrationi of pheno-barbital failed to lower serum bilirubin, whereas it clearlyrdecreased the hyperbilirubinemiiia in two of the remiiainingthree children (20, 21). Exact data on the effect of pheno-barbital in the third child (S) could not be obtainied. If onetakes the response to phenobarbital and the developmenit of

TABLE ICharacteristics of the Nine Children with Crigler-Najjar Disease

Bile analysisBiliary

Unconjugated Responise bilirubin (a,-PIA-bilirubin Kernic- to pheno- Hepatic ao-EA" a0-EA) &-EA1

Child in serum terus* barbital* UDP-GTAt PIA EA as-PIA§ (pH 2.7) (pH 2.7) (pH 2.7)

mgIlOO ml lAglh/g mg/1O0 ml %of total azopigment

K 25.0 - + 145 2.6 1.9 68 51 17 27B 23.5 - + 0; 0; 0; 252 2.9 1.2 73 52 21 24S** - NDtt 93.0 62.0 68 56 12 23M,§§ 24.0 + - 0 0.9 0.4 72 60 12 17M2 2.9 1.2 73 52 21 15D,§§ 19.6 + - ND 2.7 1.1 81 47 34 40D25§§2.4 0.8 56 41 15 48D3 1.3 0.6 84 61 23 25E 23.1 + - ND 3.2 1.3 99 77 22 traceVI 5.2 3.0 87 69 18 3V2 24.0 + - 0 0.6 0.3 97 58 39 traceV3A § 1.1 0.7 90 80 10 traceC 35.0 + - 0 4.8 1.6 86 60 26 23G, 25.0 + - ND 7.1 4.0 87 70 17 9G255§ 8.5 3.2 86 66 20 18GM55 5.3 3.4 83 68 15 30

* Absence (-) or presence (+) of kernicterus or positive response to treatment.Bilirubin UDP-glucuronosyltransferase activity.

§ Unconjugated azodipyrrole formed by coupling with PIA.11 Unconjugated azodipyrrole formed by coupling with diazotized EA, a system in which, at pH 2.7, only the conjugatedbilirubin reacts. Therefore the difference (ao-PIA minus ao-EA) gives an estimate of the amount of unconjugated bilirubinpresent.¶ Glucuronidated azodipyrrole formed by coupling with diazotized EA.$* Bile obtained by gallbladder puncture at surgery.

4 Not determined.§§ Samples taken under phenobarbital therapy. Subscripts after the letters refer to successive bile samples from the samepatient.

Unconijugated and Alonoconjugated Bilirubin in Humant Bile 971

UCBUCB J'N=~~~NN-R

MONOGLUCURONIDE

' s N=N-R

Di GLUCURONIDE

FIGURE 1 Schematic representation of diazo-cleavage of bili-rubin, leading to dipyrrolic azopigments. Diazotized p-iodoaniline (PIA) reacts with both unconjugated (UCB) andconjugated bilirubin, whereas only the conjugates react withdiazotized ethyl anthranilate (EA) at pH 2.7. In the PIA sys-tem, one molecule of UCB leads to the formation of twomolecules of unconjugated azodipyrrole (ao). In both PIAand EA systems: one molecule of bilirubin monoglucuronideproduces one molecule of a0 and one of conjugated 8-azopig-ment; diglucuronide yields only 8-azopigment.

kernicterus as discriminative features, then the former sixchildren would belong to the so-called group I (22). In fourof these patients, two consecutive bile samples were ob-tained, one before and one during treatment with pheno-barbital, for periods varying from 1 mo to 1 yr.

Gallbladder bile was obtained after stimulation with intra-duodenal magnesium sulfate (15 g) or intravenous (i.v.) in-jection of cholecystokinin (1 Ivy dog U/kg; CCK-PZ, GIHUnit, Karolinska Institute, Stockholm, Sweden). In all cases,a first sample (5-15 ml) taken before gallbladder stimula-tion was discarded because gastric juice present as a contami-nant may induce artificial changes in the pattern of biliarybilirubin (23). The bile samples were analyzed either freshor after storage at -20°C in the dark. Storage for at least 3wk under these conditions did not alter the bilirubin com-position, provided the samples were deep-frozen immediatelyafter collection.

In the patients, but not in the controls, liver tissue wasobtained by needle biopsy or at the time of splenectomy inthe hemolytic patients. Specimens were assayed for bilirubinUDP-GTAimmediately or after storage of the biopsy as suchat -20°C. The activity remains unaltered for periods up to 6wk at least under these conditions (24, 25).

Chemical methods. Bilirubin UDP-GTAwas determinedin digitonin-activated liver homogenates as described byBlack et al. (25). In two children (C and B), transferaseactivity was measured by a slight modification of thismethod yielding essentially comparable results (26). /3-glucuronidase activity was assayed with phenolphthalein asthe substrate (27).

Bile pigments were assayed with various diazo-couplingtechniques (Fig. 1). Before the assays, bile samples from thenormal controls and adult patients were diluted 21 to 51-foldwith distilled water. To obtain sufficient sensitivity in theassay, bile of patients with Crigler-Najjar disease was dilutedonly 3 to 11-fold, except in the child S (101-fold). Conjugatedbilirubin was determined by treatment of diluted bile withdiazotized ethyl anthranilate (EA) at pH 2.7 (28, 29). Totalbile pigment was determined by coupling with diazotizedp-iodoaniline (PIA) in the presence of a reaction ac-celerator (30). To detect any acid-labile conjugating bonds,bile was treated with diazotized EA at pH 6.0 (31); thisalso promotes reaction of unconjugated bilirubin-IXa ifpresent.

The azopigments formed were separated by thin-layerchromatography (TLC). The silica gel plates were firstdeveloped with benzene:ethyl acetate, 85:15 (vol/vol), to re-move lipids and excess of diazoreagent. This washingprocedure was followed by successive developments withchloroform:methanol:water, 65:25:3 (vol/vol/vol), for 10 cmand with chloroform:methanol, 85:15 (vol/vol), for 16 cm. Theseparated azopigments were quantitated either by densitome-try or by photometric reading of methanol eluates (19).Preparations of azodipyrrole (ao; from bilirubin-IXa), ofazodipyrrole f3-D-monoxyloside (a2), 8-D-monoglucoside(a3), and 8-D-monoglucuronide (8; from normal rat and dogbile), whose structures have all been previously established,were used as chromatographic references (19, 32-34). Theratio of the a,,-EA (pH 2.7) over total azopigment wascalculated, thus permitting the determination of the relativeamounts of mono- and diconjugated bilirubin-IXa (seebelow, Validity of methods).

The spectral purity of the separated azopigments was moni-tored by comparing the "characteristic spectra," obtained inmethanol, with those of pure reference compounds (35).Azopigment a0 was identified as azodipyrrole by TLC separa-tion into a mixture of the vinyl and isovinyl isomers ofazodipyrrole, both as the free acids and as their methylesters (19). Separation of the methyl esters allows easy dif-ferentiation from mesoazodipyrrole (35). The 8-azopigmentobtained from bile of three patients with Crigler-Najjardisease was further subjected to TLC, after formation ofthe methyl esters (35), and of the fully acetylated methylesters (19). The corresponding derivatives of azodipyrrole,/-D-glucopyranuronoside obtained from normal rat bile (32,34) were used as chromatographic references.

Unconjugated bilirubin-IXa was estimated by the use ofchloroform extraction and spectrophotometry at 454 nmas de-scribed for serum by Brodersen and Vind (36). 17 bilesamples were assayed immediately after collection; an ad-ditional sample had been stored at -20°C. To allow quantita-tion of any coextracted conjugated bilirubin, the chloroformextracts were reacted with diazotized PIA and the azopig-ments analyzed by TLC (23). In some instances, the chloro-form extracts were applied directly to thin-layer plates andthe bile pigments developed with chloroform:acetic acid, 99:1(vol/vol) (37). Purified bilirubin-IXa served as a referencecompound.

Validity of methodsCOUPLINGOF BILE PIGMENTSWITHDIAZOTIZED ETHYL ANTHRANILATEATpH 2.7; SIGNIFICANCE OF THE AZOPIGMENTFRACTION a0-EA

Each mole of monoconjugated biirubin-IXa gives rise to theformation of 1 mol of unconjugated and 1 mol of con-jugated azodipyrrole (Fig. 1). Provided certain conditions aremet, the azopigment fraction a0-EA (pH 2.7) multiplied by2 equals the amount of monoconjugates expressed as a per-centage of the total conjugated bilirubin-IXa (38, 39). Thisao-EA (pH 2.7) fraction can assume the extreme values of 0(no monoconjugate) and 50% (only monoconjugates), buttheoretically should never exceed 50%. The analytical re-quirements are as follows:

(a) Conjugated bilirubin-IXa must react completely withthe diazo-reagent. Previous work established that this condi-tion is fulfilled for the reaction at pH 2.7 with diazotizedEA (28) provided that bilirubin concentrations are <5 mg/100 ml as shown in the present work.

972 Fevery, Blanckaert, Preaux, Heirwegh, and Berthelot

(b) Foreign color must not be present in the quantitatedazopigment spots. Gross spectral impurity can easily bemonitored by comparing "characteristic spectra" with thoseof pure reference compounds. "Characteristic spectra" areobtained by plotting the logarithm of the extinction (opticaldensity) against the wavelength, thereby rendering the shapeof the spectra independent of the concentration and opticalpath length (35). This test is especially indicated for azo-pigment a, which moves near the solvent front. In general,TLC with respect to known reference compounds allowsunequivocal localization of conjugated azodipyrroles derivedfrom conjugated bilirubin-IXa. By TLC of the methyl esterof azopigment a, any contamination with mesoazodipyrrolecan easily be detected. (35).

(c) Acid-labile conjugating linkages could be present insome bilirubin-IXa diconjugates. If split during diazo-coupling at pH 2.7, then some monoconjugated bilirubin-IXa could be formed. Inasmuch as such cleavage wouldbe slower at less acidic pH, bile samples were treated inparallel with diazo-reagent at pH 2.7 (standard system) and atpH 6.0, and the a,-fractions quantitated (Table II). Identicalvalues were found when bile samples from normal adultsand from Gilbert's patients were analyzed.

(d) Formation of unconjugated azodipyrrole from othersources:

Normal controls and patients with Gilbert's syndrome. Ingeneral, minute amounts of unconjugated bilirubin-IXa arefound in human bile (23). In confirmation of previous data(28) and assays of urine (40) and serum (41), uncon-jugated bilirubin-IXa (10-20 mg/100 ml) added to dilutedbile (5 to 51-fold) reacted only to a negligible extent. Inparticular the a,-fractions were unchanged. It is likely,therefore, that reaction of endogenous uncoinjugated pigmentcan be disregarded. Even if it reacted completely its conceni-tration in fresh samples of both these types of bile is solow (Table III) (23) that the ratios would be barely af-fected. This is borne out by the similar amounts of ao-frac-tions obtained in the presence (PIA) or the absence (EA, pH2.7) of accelerating substances (Table II).

Children with Crigler-Najjar disease. Treatment of bilewith diazotized EA at pH 2.7 yielded percentages ofao-azopigment equal to, or higher than, 50% (TablesI and II), thus indicating that pigments other than mono-conjugated bilirubin-IXa contributed to the formationof the ao-azodipyrrole. In these bile samples, uncon-jugated bilirubin-IXa could be a major source of ao-azopigment, because it was a relatively important componentof diazo-positive material (Table III). Owing to the rather lowdilutions of bile that could be applied, bile salts could havepromoted its diazo-coupling at pH 2.7 (42). Another source ofazodipyrrole is suggested by a recent study of bile fromhomozygous Gunn rats, in which at least 24% of total diazo-positive bile pigment was shown to be composed ofdiazo-positive unconjugated tetrapyrroles related to bilirubin-IXa but containing only one of both dipyrrole halves in un-modified form (43). In addition, a number of ill-defined pig-ments which also yielded some azodipyrrole were observed.It is suggested that similar pigments occur in bile fromCrigler-Naijar patients, with an ensuing increase in the ao-EA (pH 2.7) fraction.

COUPLINGWITH DIAZOTIZED PIA

In this procedure both unconjugated and conjugatedbilirubin-IXa are transformed into azo-derivatives (Fig.1). The ao-PIA can thus derive from unconjugated and

TABLE IIUnconjugated Azopigment a0* Obtained with Various Diazo-Procedures from Bile of Controls and Patients with Either

Gilbert's Syndrome or Crigler-Najjar Disease

a,-EA a(,-EA(pH 2.7) (pH 6.0) a0-PIA

ControlsV 13 13 14Mo 7 8 9L 12 10 11Le 14 12 7M 16 16 13

Adults with Gilbert's syndromeBb 24 23 21VdR 32 29 25VR 26 25 26Wm 20 22 19Sw 26 28 30

Children with Crigler-NajjarVI 69 77 87V2 58 82 97M2 52 71 73D2 42 48 56

* The percentages of the ao azopigment are expressedrelative to the total azopigments resulting from treatmentwith diazotized EA at pH 2.7 or 6, and with diazotized PIA.

monoconjugated bilirubin-IXa. With bile from Crigler-Najjar patients, additional unconjugated azodipyrrole couldIoriginate from unconjugated, tetrapyrrolic breakdown prod-ucts of bilirubin-IXa retaining an unmodified dipyrrolemoiety (43).

DETERMINATIONOF UNCONJUGATEDBILIRUBIN-IXa

The procedure is based on chloroform extraction. Correctionfor any coextraction of conjugated pigment was performed bytreatment of the chloroform extracts with diazotized PIA andsubsequent analysis of the azopigment distribution (23). As theuncorrected concentrations for niormal bile anid bile fromnGilbert's patients are already very low (in general <1% oftotal bile pigment), even a rough estimate of the correctionterm is adequate. Chloroform extracts from bile of Crigler-Najjar patients will not onily conitain uniconjugated bilirubini-IXa, but possibly also some conijugated pigment, anid yellowbreakdown products of bilirubin-IXa. Furthermore, if unlconl-jugated bilirubin-IXa is relatively importanit, theni both thebile pigment concen-trationi anid the a0-fraction obtainied withPIA will exceed the values obtainecl with EA at pH 2.7.

RESULTS

Hepatic bilirubin UDP-GTAA liver biopsy was obtained from 18 of the 25

adults with Gilbert's syndrome. UDP-GTA rangedbetween 42 and 494 Ag of bilirubin-IXa coIn-jugated per hour and per gram of wet weight ofliver. This is far below the values found in normalindividuals by the same method: 1,100+280 (1 SD, [2])

Unconjugated and Monoconjugated Bilirubin in Humlan Bile 973

TABLE IIIUnconjugated Bilirubin-IXa (UCB) in Bile of Normal Individuals and Patients with Either

Gilbert's Syndrome or Hemolysis or Crigler-Najjar Disease

DifferenceTotal bilirubini Concentration UCBas %of total (aO-PIA*-ao-EAt)

concentration in bile of UCBin bile diazopositive material (pH 2.7)

,ng1100 mlmg/i 0 nil

Normals(10 individuals) 3.3-60.9 0-0.35 0-0.8 0-2.1

(but in one: 2.05)Gilbert's syndrome

B 29.5 0.02 0.07 -2.5Be 10.2 0.30 3.00 0.5

HemolysisVdP 45.0 traces 0 -2.2H 402.9 0.002 traces -3.0

Crigler-Najjar diseaseVI 5.2 2.45 47 18V2 0.6 0.34 57 39V3 1.1 0.60 54 10G3 5.3 1.59 30 15

* Unconjugated azodipyrrole formed by coupling with diazotized PIA.Unconjugated azodipyrrole formed by coupling with diazotized EA at pH 2.7.

or 1,330+390 (6). In Crigler-Najjar disease, the activityof the enizyme was 0 in two chil(dreni (NI ain(i V) andl145 in aniother (K; Table I). In two other patienits,tranisferase activity miieasuredI bv a imio(lified versioni ofthe saimie metho(d (26) was zero (C), anid at stuccessiveassays, 0, 0, 0, anid 252 (B).

Concentration of bile pigments in bileCollection of duodenal fluid enriched with bile by

means of intraduodenal magnesium sulfate or intravenouscholecystokinin inevitably leads to great individualvariation in pigment concentration as a result of dilu-tion by other secretionis. Total bile pigment concentra-tion in the samples was 4-194 mg/100 ml in the nor-mals, 3-132 mgl100 ml in the adult Gilbert's patients,and 0.6-8.5 mg/100 ml in eight of the Crigler-Najjar childreni. Bile obtainedl from child S by directgallbladder puncture yielded 93 mg/100 ml. It is ofiniterest that the conicentrations calculated from thethe method using diazotized PIA (total bilirubini) didniot dliffer signiificanitly from the values obtained withEA at pH 2.7 (conijugate(d bilirubini) in the normiialindividuals and in the adult patienits with Gilbert'ssvindromiie. In conitrast, in the bile samples of all thechildreni suffering from Crigler-Naijar disease, totalbilirubini conicentrationis greatly exceeded the valuesfound with EA at pH 2.7, thus demonstratinig thepresence of unconjugated bilirubin in these samples(Table I).

Unconjugated azodipyrrole (bilirubin-IXamonoconjugates) obtained from bileChromatographic analysis of EA azopigments formed

at pH 2.7 in the bile of the control group, showed thata0-azopigment amounited to 13.6+3.9% (n = 20) of totalazopigment. This proportion was significantly increasedin patients with Gilbert's syndrome, representing 24.3+4.9% (n = 25), P < 0.001. Markedly increased values(60.5±10.7) were obtained in the 16 bile samples from thenine children with Crigler-Najjar disease (Fig. 2).Determination of the a0-fraction with the two otherdiazo-methods (PIA and EA at pH 6.0) gave very simli-lar values to those obtainied with EA at pH 2.7 in thecontrols anid in the adlults with Gilbert's syndrome; thedifferenices niever exceeded 4, except in two patients:one in the former group (Le) and one in the latter (VdR)in whom it was 7 (Table II). In contrast, in the 16samiiples obtained from the nine children with Crigler-Najjar disease, ao-PIA on the average exceeded ao-EA(pH 2.7) by 20.13+7.78 (Table I). This again shows thepresence of significant amounits of uniconjugated biliru-bini in the bile of childreni with Crigler-Najjar disease.

Bile of niormal controls anid of patients withGilbert's syndrome. Spectral and TLC analysis ofazopigment a0, both as the free acid and as its methyl ester,showed that the azopigment corresponded exclusively tounconjugated azodipyrrole. A critical analysis of compar-ative assays with three diazo-methods (Table II) excluded

974 Fevery, Blanckaert, Preaux, Heirwegh, and Berthelot

the formation of azodipyrrole from dicon-jugates con-taining acid-labile conjugating bonds or from sourcesother than monoconjugates (see Validity of methods).Therefore, the amount of monoconjugates expressed asa percentage of total conjugated bilirubin-IXa (i.e.ao-EA [pH 2.7] multiplied by 2) was 27.2+7.8% (it= 20) and 48.6+9.8% (it = 25) in bile of normal adultsan-d in that of patients with Gilbert's syndrome, re-spectively (P < 0.001). It is likely that these values arevalid for freshly secreted bile. Thus, the presence ofspecific deconjugating factors in the bile of the pa-tients could be excluded on the following basis: equalvolumes of normal bile and patient's bile (four samplestested) were mixed and incubated at 37°C for 30 min.After diazo-treatment at pH 2.7 and TLC analysis of theproducts, the concentrations of azopigment a, did notexceed the mean values calculated from assays per-formed on the individual bile samples. Furthermore,,8-glucuronidase activity could not be detected in thebile from two normal controls nor in that from twopatients with Gilbert's syndrome. This observation isin agreemiient with the studies of Felsher et al. (3) inliver tissue and of Boonyapisit et al. (44) in bile fromadult patienits. Wetherefore conclude that the propor-tion of moniocon-jugates is significantly increased in thebile of patients with Gilbert's syndrome or, conversely,that the excretion of (licoinjugated bilirtubin-IXa is de-pressed.

Bile of patietnts with Crigler-Najjar disease. Aswith bile from normal adults and from patientswith Gilbert's syndrome, azopigmenit a0 correspondedto uncon-jugated azodipyrrole, but its chemical signifi-cance is unclear (see Validity of mizethods). Obviously,all conjugated bilirubin-IXa could consist of monocon-jugates as more than the required equivalent of un-conjugated azodipyrrole is formed at pH 2.7 (Table I).However, uncon-jugated azodipyrrole must also havebeen formed from sources other than monoconjugatesin those samples where ao-EA (pH 2.7) significantlyexceeded the 50% value. Owinig to lack of sufficientbile it has not been possible to analyze bile pigmentsdirectly by TLC in ordler to quantitate the dicon-jugates present. However, a reasonable assumnption isthat the levels of dicon-jugates in the bile of thesechildreni are verv low, or even absent.

Unconjugated bilirubin-IXa

The proportion of unconjugated bilirubin-IXain bile, determined by PIA azopigment analysisof chloroform extracts, never exceeded 3%, andwas usually below 0.8% in 10 normal subjects,two patients with Gilbert's syndrome and two withhemolvsis (Table III). In contrast, in four bile

soF

c- EAO/o

60-

20F

0

S

*-o

0O*4z

6*

*0

asJ.

00

N GS CNTIIFIGURE 2 The percentage of a,, unconjugated azodipyrroleformed with diazotized EA at pH 2.7 in bile from normaladults (N), from patients with Gilbert's syndrome (GS), andfrom patients with Crigler-Naijar (CN) disease of type I (0)or type 11 (0).

samples from two children with Crigler-Naijar disease,unconjugated bilirubin-IXa, estimated by choloroformextraction followed by PIA azopigment analysis,amounted to 30-57% of total diazo-positive material.These values may slightly overestimate the uncon-jugated azodipyrrole fraction (see Validity of methods).However, in one case (G3), sufficient material wasavailable for direct analysis of the chloroform extractby TLC, thereby permitting unequivocal determina-tion of a component which migrated as unconjugatedbilirubin-IXa. In comparison to ao-EA (pH 2.7), theincrease seen in unconjugated azodipyrrole a,0 whenbile samples from all the children were treated withdiazo-reagent, under conditions promoting reaction- ofall bile pigments (EA pH 6.0 and PIA; Tables I and II),supports the contentioni that unconjugated bilirubin-IXa is of importance in the bile of Crigler-Najjarpatients.

Conjugated azo-derivativesAll bile samples from the patients with Gil-

bert's syndrome yielded abundant 6-azopigmenton treatment with diazotized EA at pH 2.7. This

Ulicotnjugated and(c AlonoconijuggatedI Bilirubini in Hiumiian Bile 975

pigment amounted to 54.0±+4.9% of the total azo-color in bile samples from 12 patients, a valuewhich is lower than that previously found for normaladults, 75.4±5.7% (23). 8-Azopigment was also formedupoIn diazo-treatment of bile from almost all patientswith Crigler-NajiJar disease (Table I). In three caseswith severe type I disease (patients D, V, and G), theisolatedI 8-azopigments were further analyzed. OnTLC, the methyl esters and the fully acetylatedmethyl esters of the 8-azopigment moved in eachcase as the corresponding derivatives of authenticazodipyrrole ,8-D-glucopyranuronoside. The identity ofthe unknowns and reference material was furthersupported by the parallel separation into their vinyland isovinyl isomers. From the presently applieddouble-derivative formation technique, it may be con-cluded unequivocally that bilirubin-IXa glucuronidewas present in the bile of patients D, V, and G. Thisobservation was further supported by mass spectro-metric analysis of the 6-azopigment found in the bilesamples of children D and G. In view of the importantimplicatioins of these results to our concepts of Crigler-Najjar disease, an extensive organochemical and massspectrometric study of bilirubin derivatives is in pro-gress (45).

Azopigments a2 and a3 were found in all bilesamples. In dog bile, they have been identified as thef3-D-xylopyranoside and the f8-D-glucopyranoside ofazodipyrrole, respectively (31, 33). The same con-clusion has been reached for human bile (Fevery,unpublished work). In the bile of patients with Gil-bert's syndrome, azopigments a2 and a3 amounted to1.6±1.0 and 3.7±1.2% respectively, values which didnot differ significantly from those found previously (23)for normal human bile, i.e. 1.2±0.4 and 3.5±0.8%.Somewhat higher amounts of azopigment a3, rangingfrom 3-17%, were found in bile from children withCrigler-Najjar disease.

DISCUSSION

Gilbert's syndrome or "constitutional hepatic dysfunc-tion," is characterized by chronic, often mild, uncon-jugated hyperbilirubinemia in the absence of overthemolysis (for a review, see reference 46). It has longremained a diagnosis made by exclusion of otherdiseases, and therefore depended on the extent of theinvestigations that had been performed. The recent(lemonstration of markedly decreased bilirubin UDP-GTA, both in unactivated (5) and in digitonin-activatedliver homogenates (2-4, 6, 7), complemented theprevious observations of decreased transferase activityin some patients with more pronounced unconjugatedhyperbilirubinemia (47, 48). In 1969, Arias et al. (22)described 16 cases of severe chronic nonhemolytic un-conjugated hyperbilirubinemia which they proposed to

subdivide into two groups. Group I was composed ofthe most severely affected children who developedkernicterus and whose condition did not improve upontreatment with phenobarbital. Their bile was "virtuallycolorless and contained only a trace of unconjugatedbilirubin" (cited from Arias et al. [22]). In contrast,conjugates were detected in the bile of patients ingroup II, and their bilirubinemia decreased on treat-ment with phenobarbital. The activity of hepaticbilirubin UDP-GTAwas near to zero in both groups.Further aspects of the disease have been discussedrecently by Blaschke et al. (49).

In the present study, several parameters of bilirubinmetabolismii have been investigated in both disorders.In our patients with Gilbert's syndrome, transferaseactivity was 4-45% of the control values. Zero ornear-zero activities were found in the children withCrigler-Najjar disease, in agreement with other work(22, 49, 50). This, however, does not offer anabsolute diagnostic criterion, as values approachingzero were occasionally found in patients with Gil-bert's syndrome (50, 51) and in neonates withoutliver disease (52). However, for the present, transferaseassays allow differentiation of Gilbert's syndrome andCrigler-Najjar disease from other types of uncon-jugated hyperbilirubinemia.

The analysis of bile showed that conjugated biliru-bin-IXa was present in all samples examined, in-cluding those obtained from the children with Crigler-Najjar disease. The conjugating groups detected wereglucuronic acid, glucose, and xylose. In seven of ninechildren with Crigler-Najjar disease, 9-48% of the EA-azopigments contained glucuronic acid (8-azopig-ment); in two children (E and V), only trace amountswere found (Table I). Glucose residues (3-17%) wereusually present in slightly higher relative proportionsthan in the bile from the controls. The present resultsindicate that the detectability or nondetectability ofconjugated bilirubin-IXa (22) is not a reliable criterionfor the differentiation of patients with Crigler-Najjardisease, although it should be noted that children Eand V, whose bile contained only traces of glucuronide,belonged to the more severely affected group. The use-fulness of differentiation based on response to pheno-barbital (group II) or lack of response and develop-ment of kernicterus (group I) was confirmed (22). How-ever, colorless bile was never observed in any of ourchildren with Crigler-Najjar disease, even in the mostsevere cases.

The most striking feature of the present study wasthe highly significant increase in azopigment a0 (un-conjugated azodipyrrole) in the EA (pH 2.7)-treatedbile from all patients with congenital nonhemolyticunconjugated hyperbilirubinemia. Minor increases inthis pigment have occasionally been detected in pa-tients with liver disease and cholestasis, but were

976 Fevery, Blanckaert, Pr6aux, Heirwegh, and Berthelot

accomnpanied by other specific chaniges in biliarybilirubini compositioni (23). The a0-fraction was signifi-canitly higher in Crigler-Najjar disease than in Gilbert'ssyndlromiie. A similar increase has been recently coIn-firmned in four adults with Crigler-Naijar (lisease typeII (53, 54). As the a0-fraction was not increase(d in pa-tients with hemolysis in the absence of associatedGilbert's syndrome (Table IV), determination of the a0-fraction imay allow an easy, safe, and rapid diagnosisof bilirubin-IXa UDP-GTAdeficiency (Fig. 2). In pa-tients with Gilbert's syndrome, the deficiency is clearlyexpressed in decreased formation and biliary secretionof dliconijugated bilirubin-IXa

In patients with Crigler-Najjar disease, our mainifindinig was that conjugates were always present tosome extent in bile, whatever the severity of thedisease.

Evidence was found that these conjugates weremainly, and perhaps exclusively, bilirubin-IXamonioglucuroniide, an observation which is in ac-cordanice with a study by Gordon et al. (54), whoshowed in addition that there was a definite decreasein the total biliary bilirubin output in these patients.This clearly indicates that the maini abnormality inCrigler-Najjar disease is the decreased excretion ofdliconijugates, reflecting the transferase deficienicywhich is miiore pronounced in this condition than inGilbert's syndrome. In addition, the percentage ofunlconijugated bilirubin-IXa that we found in bile wastypically increased to 30-57% of the total bile pigment.Such an increase has also been found by isotopicalmlethods (55, 56). However, the concentrations of un-conjugated bilirubin-IXa in the bile of patients withCrigler-Najjar disease were only one- to sevenfold thevalues foundl in normal adults oir patients with Gilbert'ssyndrome. This pigment therefore constitutes only aminor fraction of total heme breakdown. A comparablesituation is found in Gunn rat bile where 31-40% ofthe biliary bile pigments is unconjugated bilirubin-IXa, although the daily excretion of this unconjugatedpigment is barely higher than that of control Wistarrats and approaches only 3-4% of total heme turnover(43).

The observations reported in the present work raiseseveral problems regarding the biochemical defin i-tion of the enzyme deficiencies involved in congeni-tal, nonhemolytic, unconjugated hyperbilirubinemia.Firstly, a disproportion is apparent between trans-ferase activities and the rates of formation of con-jugated bilirubin-IXa in vivo. The near-zero GTAlevels found occasionally in the liver of Gilbert'spatients in previous studies (50, 51), in the presentwork, and in neonatal liver (52), suggest that thestandard enzyme assay which we have employed andwhich is optimized for normal liver tissue, may notalways be adequate. The simplest explanation would

TABLE IVComparison of the Percenttage of Unconjugated Azodipyjrrole,

UDP-GTA in Hepatic Biopsies and Serulm BiliruibinLevels in Six Patients with Hemolysis

and in Normal Individuals

Totalao-EA bilirubin

(pH 2.7)* UDP-GTA in serum Diagniosis

%of total mg/h/g mg/il0 0ilazopigment liver

PatientsXre 13.1 1.242 3.8 SpherocytosisVm 11.3 1.065 2.8 SpherocytosisP 10.6 2.260 2.6 SpherocytosisG 23.2 0.316 3.5 Autoimmune

hemiioly sisVdP 26.3 0.418 2.3 SpherocytosisS 25.7 0.234 4.5 Spherocvtosis

Normals(20 individuals) 13.6+3.9 1.100±0.280 <1.0

The liver biopsies were taken in the patients at time of splenectomy.Unconjugated azodipyrrole formed by treatment of bile with diazotized EA at

pH 2.7.

appear to be that in Gilbert's syndrome anl abniormiialtransferase is present in the liver.

A situation of this sort seems to exist in Gunn ratswhose liver has a normal amounit of UDP-GTAactinigon p-nitrophenol with an abnormally low affinity forUDP-glucuronic acid (57). The near-zero tranisferaseactivities typically found in the liver of Crigler-Najjar patients (Table 1) (22, 49, 50) are consistentwith the low biliary output of bilirubin con-jugates inthis condition (54, 55). The difference in responsetowards phenobarbital found in type I and type II pa-tients may also point to a fundamental differenice inthe type of transferase responsible for conjugatinigbilirubin-IXa.

How can the decreased excretion of disconjugatedbilirubin-IXa be related to the dlocumented bilirubin-IXa UDP-GTAdeficiency? Too little is known aboutthe biochemical mechanisms underlying the con-jugation and excretion of bilirubin to warrant anvthorough discussion. It may suffice to conisider brieflya few of many possibilities which couild explain ourbasic observations. If bilirubin-IXa were converted toits monoglucuronide and subsequently to its di-glucuronide by a single enzymic site, then trans-ferase deficiency would of course result in an increasedbilirubin concentration in the cytosol. If one now as-sunmes that unconjugated bilirubin-IXa binds imiorestrongly to the enzyme than its monocon-jugate, thena smnaller fraction of the enzymic sites would befree to bind the monoglucuronide under conditionswhen a decreased amount of enzyme is available. Asa consequence, a less rapid conversion to the di-conjugate would ensue, together perhaps with a moreefficient direct biliary elimination of the monoglucuro-

Unconjugated andl Monoconjugated Bilirubin in Human Bile 977

nide. In the case of a two-enzyme system, an as-sociated deficiency of the enzyme catalyzing con-version of mono- to diconjugated bilirubin-IXa couldexplain the present data. However, even if the activityof the second enzyme is normal in vitro, as proposedby Jansen et al. (58) in studies of two patients withCrigler-Najiar type IL, other mechaniism,is, such as ac-cumulation of bilirubin-IXa within the hepatocyte,may inhibit the second step in vivo. Clearly, furtherinvestigation is necessary to elucidate the underlyingmechanisms of glucuronide formiation.

ACKNOWLEDGNIENTSWe thank Doctors Brun, Dodion, Eggermiiont, G6itze, Klinie,Levy, Nussbaumer, Rozenkranz, Sinaasappel, and Thalerwho generously providecl us with bile samiiples of childreniwith Crigler-Najjar (lisease. Our thanks are also (lue to Doc-tors De Groote, Dhumeaux, Hanounie, NI1treau, Vand(leni-broucke, anid Verwilghen for support an (l discussionis, and( toDr. Comperniolle for mass spectromiietric anialysis. We aregreatly indebted to the nurses and to P. Leroy, D. Serpeau,anl XI. Vandevijver for their skillful help in collecting ani(analyzing the samples. Wethank Dr. John Chapmani for care-ful assistance in reviewing the mnaniuscript and Mrs. NI. Tassierfor the preparation of the maniuscript.

This work was performed with grants from InstitutNational de la Sante et de la Recherche Medicale, France, andfrom Nationaal Fonds voor Wetenschappelyk Onderzoek,Belgium.

REFERENCES1. Fevery, J., N. Blanckaert, A. NI. Preaux, K. P. MI.

Heirwegh, anid P. Berthelot. 1973. Unconijugated and(l coni-jugated bilirubin pigments in bile in Gilbert's and(Crigler-Najjar's syndromes. Digestion. 8: 445-446A.(Abstr.)

2. Black, NI., andl B. H. Billing. 1969. Hepatic bilirubinUDP-glucuronvl transferase activity in liver disease anidGilbert's syndromiie. N. Engl. J. AMed1. 280: 1266-1271.

3. Felsher, B. F., J. R. Craig, and(l N. Carpio. 1973. Hepaticbilirubini glucuroniidationi in Gilbert's syn(lromie. J. Lab.Clin. Medl. 81: 829-837.

4. Fevery, J., J. De Groote, and(l K. P. NI. Heirwegh. 1973.Quantitationi of hepatic bilirubini conijugation. In TheLiver: Quanititative Aspects of Structure and( Funiction. G.Paumgartner and R. Preisig, editors. S. Karger, Basel.203-209.

5. Bellet, H., and A. Raynauid. 1974. Anl assay of bilirubiniUDP-glucuronyl tran sferase onl needlle-biopsies appliedto Gilbert's syndromiie. CliG. Chinii. Acta. 53: 51-55.

6. N1ktreau, J. NI., F. Roudot, A. NI. Preaux, D. Dhumileaux,anid P. Berthelot. 1974. D)vserythropoiesis, a necessarycondition- for (lisclosinig Gilbert's (lisease. Digestion. 10:315A. (Abstr.)

7. Felsher, B. F., aind N. NI. Carpio. 1975. Caloric intakeani( unicon-jugated hvperbiliruhbinieml ia. Gastroen terology.69: 42-47.

8. Kirshenbaumn, G., D. NI. Shamiies, and(I R. Schmiii(l. 1976i.An expand(ledl imodel of bilirubini kinietics: effect of feed-ing, fasting, an(d pheniobarbital in Gilbert's syndrome. J.Pharinacokinet. Biopharmn. 4: 115-155.

9. Zakimii, 1)., and( D. A. Vessey. 1974. NMembrane (le-penidenice of uri(dinie (liphosphate glucuronvltransferase:effect of the imiemiibranie oni kinietic properties. Biocheoil.Soc. Trans. 2: 1165-1167.

10. Abou-el Nlakaremi, NI. NI., anid K. WV. Bock. 1976. UDP-glucuroniyltransferase in perfuse(d rat liver ancl in micro-somes: glucuronidlationi of bilirubini. Eur. J. Biocheii.62: 411-416.

11. Billinig, B. H., R. Williamiis, an(d T. G. Richards. 1964.Defects in hepatic transport of bilirubini in conigeniitalhyperbilirubinaemiiia: an analysis of plasmla bilirubin (lis-appearanice curves. CliGu. Sci. (Oxf.). 27: 245-257.

12. Barrett, P. V. D., P. D. Berk, NI. NMenken, and(l N. I. Berlini.1968. Bilirubin turniover studies in niormal and(l pathologicstates using bilirubin-14C. Ann. Initertn. Med. 68: 355-377.

13. Berk, P. D., J. R. Bloomiier, R. B. Howe, and(l N. I.Berlin. 1970. Coiistitutionial hepatic dylsfunction (Gil-bert's syndrome): a niew dlefinition cbased oni kinieticstudies wvith uniconijuigate(d radiobilirubin. Am.J. Med. 49:296-305.

14. Bloomiier, J. R., P. 1). Berk, R. B. Howe, and(I N. I. Ber-lini. 1971. Bilirubin metabolism in conigeniitatl nonlhemoly-tic jaund(lice. Pediatr. Res. 5: 256-264.

15. Crigler, J. F., and(l N. I. Goldl. 1969. Effect of so(dium11phenobarbital onl bilirubini metabolism in ani inifaint withconigeniital, nonhemolvtic, uniconijtugate(d hyperbilirui-biniemiiia, and(l kernicterus.J. CliGu. Iivest. 48: 42-55.

16. Berk, P. D., J. R. Bloomer, R. B. Howe, T. F. Blaschke,anid N. I. Berlin. 1972. Bilirubini pro(luctioni as a miieasureof red cell life spani. J. Lab. Cliii. Med. 79: 364-378.

17. Frezza, M., G. Perona, R. Corrocher, R. Cellerinio, NI. A.Bassetto, anid G. D)e Sand(lre. 1973. Bilirubiin-H3 kinieticstud(lies: patterni of niormal Gilbert's syndrome and(lhemolytic states. Acta lie pato-GCastroen terol. 20: 363-371.

18. Auclair, C., G. Feldmnann, J. Hakiini, P. Boivini, J.Boucherot, and(I H. Troube. 1976. Bilirubin and(l para-i itropheniol glucuronyl tranisferase activities and(I uiltra-structural aspect of the liver in patienits with chroniichemolvtic aniemicias. Biomedicine (Paris). 25: 61-65.

19. Heirwegh, K. P. NI., J. Fevery, J. A. T. P. Mlewvissen, J.De Groote, F. Comiiperniolle, V. Desimiet, and(I F. P.Vcain Rov. 1974. Recenit advances in the separation an(Ianalysis of' diazo-positive bile pigimienits. MletliodsBiochemii. AnIal. 22: 205-250.

20. Klinie, WV., and(I A. Rosenikraiiz. 1971. Kliinische tndbiochemilische Sttuldienl zUmIl Crigler-Najjar-Syndrom.Monatsschr. Kin derheilk. 119: 555-558.

21. G6tze, H., 1). Sidiropoulos, F. A. Hess, and( P. Berthe-lot. 1972. Das Crigler-Najjar-Syndrom: klinisclhe, bio-chemiiische, imiorphologische u1nd1 therapeutisclhe Aspekte.Hel. Paedliatr. Acta. 27: 335-351.

22. Arias, I. NI., L. NI. Gartnier, NI. Coheni, J. Beni Ezzer,and(l A. J. Levi. 1969. Chroniic nonhemolvtic unconi-jugate(d hvperbilirubiniemia with glucuronyl tran sferasedeficiency: cliniical, biochemical, phlarmiaceologic andlgenietic evi(lenice for heterogeneity. A in. J. .Mled. 47:395-409.

23. Fevery,, J., B. Vani Damme, R. NMechiel, J. De Groote,anid K. P. NI. Heirwegh. 1972. Bilirtibin conijugates inbile of mani and(l rat in the niormnal state and(l in liver(lisease.j. Cliii. Intvest. 51: 2482-2492.

24. Black, NI., B. H. Billinig, and(l K. P. NI. Heirwegh. 1970.Determiinlationi of bilirubin UDP-glucuronvl tranisferiaseactivit\ in eedle-biopsy specimens of hiuima lier.Cliii. Chimi Acta. 29: 27 -35.

25. Duvaldlestini, P., J. L. Malhu, an(d P. Berthlelot. 1975. Eftfect of fastinig on substrate specificity of rat liver UDP-glticuironios\ ltranisfer-ase. Biochini. Bioplniqs. Acta. 384:81-86.

26. Frei, J. 1970. Nlultiplicity and specificity of UD)P-gluctironyltransferase. I. Effect of (li-alet catiomis d

978 Fereril, Blanickaert, Preaux, Heirwegh, and(c Ber-thelot

EDTAon the activity of UDP-glucuronvl-transferase as-saved with bilirubin, 4-m-nethvlumbelliferone and( p-nitropheniol. Etnzi iol. Biol. Clint. 11: 385-401.

27. Rodriguez Garay, E. A., MI. Del Rosario Spetale, andL: S. Morisoli. 1972. Relationship between bilirubin and,8-glucuronidase activity in rats with experimenital ob-structive jaundice. Clini. Chimii. Acta. 37: 171-177.

28. Vani Roy, F. P., aind K. P. MI. Heirwegh. 1968. De-termiiniationi of bilirubini glucuronidle and( assay ofglucuronvlx-tranisferase with bilirubin as acceptor. Bio-chemii. J. 107: 507-518.

29. Heirwegh, K. P. MI., G. P. Vani Hees, P. Leroy, F. P.Vani Roy, and F. H. Janiseni. 1970. Heterogeneity of bilepigmlenit conijugiates as reveale(d by chromatography oftheir ethvl anthraniilate azopigmlenits. Bioche in. J. 120:877-890.

30. Vani Roy, F. P., J. A. T. P. leuiwissen, F. l)e NMeiter,and(l K. P. MI. Heirwegh. 1971. Determiiniationi of bilirubinin liver homiiogeniates anl( serumil with diazotize(l p-iodloainiliine. Clini. Chimii. Acta. 31: 109-118.

31. Fevery, J., G. P. Van Hees, P. Leroy, F. Comlperniolle,and K. P. MI. Heirwegh. 1971. Excretion in dog bile ofglucose and( xvlose conijugates of bilirtubin. Biochemrl. J.125: 803-810.

32. Comlperniolle, F., F. H. Jansein, and K. P. NI. Heirwegh.1970. Mlass-spectrometric studys of the azopigments ob-tained froml bile pigimenits with (liazotizedl ethyl anthranii-late. Biochem. J. 120: 891-894.

33. Compernolle, F., G. P. Van Hees, J. Fevery, and K. P. M.Heirwegh. 1971. Nlass-spectrometric structure elucida-tioni of dog bile azopigmiienits as the acyl glycosidlesof glucopyranose anid xylopyranose. Biochemii. J. 125:811-819.

34. Heirwegh, K. P. 'I., G. P. Vlani Hees, N. Blaniekaert, J.Fevery, and(I F. Comiiperniolle. 1976. Comparative studiesoni the structures of conijugatedl bilirubin-IXa annd chanigesin cholestasis. In The Hepatobiliary System. W. Taylor,e(litor. Plenium Press, New York. 339-358.

35. Blanickaert, N., K. P. NI. Heirwegh, anid F. Comper-niolle. 1976. Synithesis anid separation bhy thin-layerchromatography of bilirubini-IX isomiiers. Biochleml. J.155: 405-417.

36. Bro(lerseni, R., and( I. V'id(l. 1963. A specific imietho(d for(Iuantitative (leterininationi of uniiconijtugatedl biliruibini andtmesobilirutbin. Scandl. J. CliGu. Lab. Incvest. 15: 225-232.

37. MlcDonagh, A. F., and(I F. Assisi. 1971. Commiiiiercialbilirtuinl: a triniity of isomiiers. FEBS (Fed. Eur. Biochlemti.Soc.) Lett. 18: 315-317.

38. Fevery, J., P. Leroy, NI. Van de Vijver, and K. P. NI.Heirwegh. 1972. Structures of bilirubin conijugatessvnthetized in vitro fromii bilirubin anid uridinie diphos-phate glucuronic acid, uridine diphosphate glucose oruridine diphosphate xylose by preparations from ratliver. Biochem. J. 129: 635-644.

39. Heirwegh, K. P. NI., J. A. T. P. MIeuwissen, anid J.Feverv. 1973. Critique of the assay and(I signiificanice ofbilirubin conijugationi. Adc. CliGu. Chemii. 16: 239-289.

40. Feverv, J., F. H. Jansein, J. A. T. P. Nleuwissen, andK. P. NI. Heirwegh. 1968. Biliruhbin-like mnaterial in urinieof niormal a(lults. CliGu. Chi in. Acta. 21: 401-410.

41. Franico, D., A. NI. Pr6aux, H. Bismiiuth, anid P. Berthelot.1972. Extra hepatic formationi of bilirubin gluicuroniides inthe rat. Biochimi Biophys. Acta. 286: 55-61.

42. With, T. K. 1968. Bile Pigmenits. Academiiic Press, Inlc.,New York. 324-328.

43. Blaniekaert, N., J. Fevery, K. P. NI. Heirweghi, and(I F.Comiiperniolle. 1977. Characterizationi of the major diazo-

positive pigmiienits in bile of homozygous Gunnii rats as thefour isomers of bilirubini-IX and(l as onie dih cdroxyl and(ltwo monohvdroxvl derivatives of bilirubin-IXa. Biocliem.

J. 164: 237-249.44. Booniyapisit, S. T., B. W. Trotmani, aind J. D. Ostrow. 1974.

Unconijugated bilirubini (UCB) anid hydrolysis of conljU-gate(l bilirubini (CB) in gallbladdler bile of patienits withpigmiienit or cholesterol gallstones. GastroenterologyJ. 67:781A. (Abstr).

45. Blanickaert, N., F. Comperniolle, J. Dodion, J. Fevery, anidK. P. NI. Heirwegh. 1976. Characterizationi of the bilirubi-nioi(d pigmiienits in bile of childIreni with Crigler-Najjardisease (CN) andcl of homozygous Gunnii rats: the effect ofphototherapy. Gastroen terologyj. 71: 897.

46. Foulk, WV. T., H. R. Butt, C. A. Owen, F. F. Whitcomb, andiH. L. Nlason. 1959. Conistitutionial hepatic dysfunction(Gilbert's disease): its niatural history anid related svn-dromiies. Me(icine (Baltimore). 38: 25-46.

47. Arias, I. NI., anid I. NI. Lond(lon. 1957. Bilirubini glucuronii(lefonniationi in vitro: ielmlonistrationi of a defect in Gilbert's(lisease. Scienice (W'ash. D.C.). 126: 563-564.

48. Arias, I. NI. 1962. Chronic uniconjugated hyperbili-rubiniemia without overt signis of hemolysis in adoles-cents anid adults.J. CliGu. Incvest. 41: 2233-2245.

49. Blaschke, T. F., P. D. Berk, B. F. Scharschmlidit, J. R.Guyther, J. NI. Vergalla, anid J. G. Waggoner. 1974.Crigler-Najjar syndrome: ani unuiisual couirse with de-velopmiient of nieurologic (lamiiage at age eighteeni. Pedliatr.Res. 8: 573-590.

50. Okolicsanyi, L., J. Frei, P. NMagnenat, and( R. Naccarato.1971. Mlultiplicitv of UDP-glucuronyl tranisferase. III.UDP-glucuronvl tran sferase anld B-glucuron idase ac-tivities assayed with differenit substrates in iniherited anidacquired humilani liver (lisease. Entzltime (Basel). 12:658-673.

51. Black, NI., R. D. Perrett, anid A. E. Carter. 1973. Hepaticbilirubini UDP-glucuronyl tranisferase activity anid cyto-chromiie P 450 conitenit in a surgical populationi, anid theeffects of preoperative drug therapy.J. Lab. CliGu. AMedl. 81:704-712.

52. Felsher, B. F., N. NI. Carpio, NI. NI. WVooley, and NI. J.Aseh. 1974. Hepatic bilirubin glucuronidlationi in nieoniateswitlh unaconijugatedl hvperbilirubinemia andcl conigeniitalgastroinitestinial obstruction. J. Lab. Clin. AMe(d. 83:90-96.

53. Gollani J. L., S. N. Huanig, B. Billinig, ain(1 S. Sherlock, 1975.Prolonige(d survival in three brotlhers witlh severe type 2Crigler-Najjar syndrome: ultrastructural and(l metabolicstuclies. Gastroen terologyI. 68: 1543-1555.

54. Gor(loni, E. R., E. A. Shaffer, and(l A. Sass-Kortsak. 1976.Bilirubini secretioni anid conijugationi in the Crigler-Najjartype II. Gastr oenterology. 70: 761-765.

55. Schmiiid, R., anid L. Hammllaker. 1963. NMetabolismn and(ldispositioni of C'4-bilirubin in conigeniital nonhemolyticjaun(lice.J. CliG. Incvest. 42: 1720-1734.

56. Thaler, NI. NI., N. H. Dawber, J. Krasnier, S. J. Yaf'fe, andclL. NMosovich. 1973. Effects of phototherapv on1 bili-rubini(B) metabolismii and(l sulfobromiiophthaleini (BSP) ex-cretioni in unlconijugatecl hyperbilirubinemia. Pediatr.Res. 7: 334A. (Abstr.)

57. Nakata, D., D. Zakimii, and(l 1). A. Vessey. 1976. Defectivefunction of a microsomiial UDP-glucuronvltran sferase inGunnii rats. Proc. Natl. Aca(l. Sci. U.S.A. 73: 289-292.

58. Janiseni, P. L. NI., J. R. Chowdhurv, E. Fischberg, and( I. NI.Arias. 1976. Ani enzyme in liver plasmiia membrane whichconverts bilirubini miionioglucuronii(le (BNIG) to bilirubin(liglucturoini(le (BDG). CliGu. Res. 24: 433A. (Abstr.)

Uniconljugateci anid AlMonocouijiugated Bilirubint in Human Bile 979