Digestive and Liver Disease 39 (2007) 351355
Liver, Pancreas and Biliary Tract
Inheritance of hyperbilirubinemia major autosomal recess
. FabL. O
ediatriciences,, Cagliaestive D
Room 233, Yale University, New Haven, USA
Received 6 September 2006; accepted 18 December 2006Available online 7 March 2007
Backgrouimpaired bilof the UGT1
Subjectsconsisting ofrom hyperbexcluded poand liver dis
Results.in males. Comajor recess
Conclusiomental facto 2007 Edit
Unconjutive disoraccumulatijugated hyobservation
Corresponical Sciences,35100 Padua,
1590-8658/$3doi:10.1016/jnd and aim. To clarify the precise mode of inheritance of Gilbert syndrome, an unconjugated familial hyperbilirubinemia, whereirubin conjugation is caused by reduced UGT1A1 activity determined by a defective function of the A(TA)6TAA promoter regionA1 gene.and methods. Serum bilirubin levels were measured in a large, homogeneous resident population from North-Eastern Italy,f 1.639 males (age 44.5 13.9, range 1889 years), and 1.420 females (age 45.1 15.0, range 1885). In 112 nuclear familiesilirubinemic probands living in the same area a complex segregation analysis was then performed. In both samples we carefullytentially confounding factors of bilirubin levels (alcohol abuse, excessive cigarette smoking, drug consumption, overt haemolysisease).Mean serum bilirubin concentrations are higher in males than in females, showing fluctuations through the different age periodsmplex segregation results demonstrate that unconjugated hyperbilirubinemia exhibits a precise mode of inheritance in which aive gene with a frequency of 0.45 is responsible for higher serum bilirubin values.ns. This major recessive gene accounts only for a part of the serum bilirubin concentration, thus implying additional, environ-
rs for the clinical appearance of GS.rice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
ilirubin; Complex segregation analysis; Gilbert syndrome; Inheritance
gated hyperbilirubinemia syndromes are selec-ders of bilirubin metabolism, resulting inon of pigment in the blood . Isolated uncon-perbilirubinemia can be defined as the repeateds of total serum bilirubin concentration clearly in
ding author at: Department of Surgical and Gastroenterolog-University of Padua, Ospedale Busonera, Via Gattamelata,Italy. Tel.: +39 049 821 5617; fax: +39 049 821 5619.dress: email@example.com (L. Okolicsanyi).
excess of the values found in normal individuals of the sameage and sex, with direct-reacting bilirubin comprising lessthan 1520% of the total. Clinical forms vary widely fromthe mild, asymptomatic hyperbilirubinemia of Gilbert syn-drome (GS) to the severe CriglerNajjar syndrome (type I)where kernicterus is frequent.
In GS unconjugated hyperbilirubinemia occurs withoutovert signs of haemolysis and other abnormalities of liverfunction. Increased bilirubin levels are caused by a reducedactivity of the liver microsomal enzyme uridine-diphospho-glucuronate glucuronosyl-transferase (UGT), which has dif-ferent isoforms, one of which, the UGT1A1, catalyses the
0 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved..dld.2006.12.019M. Clementi a, E. Di Gianantonio a, LM. Strazzabosco d, R. Tenconi a,
a Clinical Genetics and Epidemiology, Department of Pb Department of Surgical and Gastroenterological Sc
c Centro Nazionale Ricerche (CNR)d Department of Internal Medicine, Section of Diga: Evidence forive generis b, P. Forabosco c,kolicsanyi b,
s, University of Padua, ItalyUniversity of Padua, Italyri, Italyiseases, TAC Building,
352 M. Clementi et al. / Digestive and Liver Disease 39 (2007) 351355
transformation of the lipid-soluble bilirubin IX- to the morepolar, water soluble bilirubin, through its conjugation withtwo molecules of glucuronic acid. The gene encoding forUGT1A1 iarm) and ismRNA trantranscriptioleft in the pinheritancepast an autoas reported
More rewith GS arehaving twoTATAA elegested thathyperbiliruthe longerhas been eabnormal psubjects. Hof homozybilirubin lehyperbilirution, withobilirubin [4ies on assodisease shogene withbilirubin lebilirubin cober of heamay influedrug consuusing apprfore needeGS [4,5].
TakingconcentratiNorth-Eastin which aorder to de
This wa. A ransubjects liva rural areaasked to pasent. Thereyears), andData from
normal population values (distribution by sex and age) to beused for the segregation analysis.
nal inwas twas
day ining (l
fere wod saAfterand a
niconand dase, aultrasese suatologditioninatio
iary ancreasas theia :total
cular,bin nerobanted 11the s
able inFathe, moth12 to 7and/orenzym
digresegregrmulaPOINn runl a mand amajors proded no
r-locus located in region q37 of chromosome 2 (longregulated by the promoter TATAA box that directsscription by controlling RNA polymerase II andn factor 2D . Population and family studiesast a degree of uncertainty regarding the mode ofof GS, although it has been hypothesized in thesomal dominant pattern of transmission (#143500in the OMIM catalog) .cently, Bosma et al. observed  that patientshomozygous for the TATAA box polymorphism,extra bases of thymidineadenine, instead of a
ment with six repeats. The same authors also sug-this abnormality is not sufficient by itself to revealbinemia and subicterus. The allele frequency ofA(TA)7TAA sequence in the normal populationstimated to be 0.40, allowing a frequency of theromoter element homozygosity of about 16% ofowever, less than half of the expected proportiongotes actually have increased serum unconjugatedvels [4,6], so that the estimated prevalence ofbinemia is only around 5% in the adult popula-ut taking into account the sex differences in serum]. Furthermore, during the last years, three stud-ciation of serum bilirubin with coronary heartwed that serum bilirubin segregates as a majorthe rarer genotype associated with high serumvels . Extensive family studies of serumncentrations, including a sufficiently large num-
lthy subjects following exclusion of factors thatnce serum bilirubin concentration (alcohol andmption, manifest haemolysis or liver disease) andopriate procedures of genetic analysis, are there-d in order to clarify the genetic heterogeneity of
this approach, we measured serum bilirubinons in a large, homogeneous population fromern Italy and then selected 112 nuclear familiescomplex segregation analysis was performed in
fine the mode of inheritance of GS.
ls and methods
s an ancillary investigation of the M.I.COL studydom control sample of 3.059 unrelated healthying in the district of Montegrotto Terme (Padua),of the Veneto Region, North-Eastern Italy, were
rticipate in this study, and gave their informed con-were 1.639 males (age 44.5 13.9, range 18891.420 females (age 45.1 15.0, range 1885).
this sample were used in this study to obtain the
AwentpersoNonethere40 g/smokinterof blotion.takenTechtotaldatabliverof thhaemin adexam
tobilfor inGS wcriterwhenpartibilirumic pselecfromualsparenavailable.57.9)fromease
majoal data recording
participants (controls and family samples) under-ne clinical and laboratory examination, and aterview using the M.I. COL questionnaire .aking any drug, including over-the-counter drugs,no excessive alcohol consumption (less thanmales and 20 g/day in females) and cigarette
ess than 10 cigarettes/day) potentially capable toith liver function and bilirubin metabolism at timempling. All were healthy at the physical examina-an overnight fasting, venous blood samples werenalysed by an autoanalyzer (SMAC II Analyzer,Instrument Corp.). For each individual, serum
irect-reacting bilirubin values were recorded in and serum indirect bilirubin obtained by difference;ound was also performed in each subject. Nonebjects evidenced overt haemolysis on standardic tests or routine liver enzyme abnormalities;
, none had hepatobiliary alterations at ultrasoundn. Taking these data together, concurrent hepa-d haematologic disorders potentially responsible
ed bilirubin levels could be reasonably ruled out.n defined according to previously well-establishedhyperbilirubinemia in both sexes was consideredserum bilirubin was greater than 17.5mol/l. Inin all of these selected individuals direct reactingver exceeded 30% of total. From hyperbilirubine-ds, observed in the gastroenterology unit, we then2 nuclear families (parents and siblings) derived
ame geographical area, resulting in 427 individ-l, which agreed to participate in this study. Bothre available in 17 families, only one parent was49 families and 46 families had no parents avail-
rs ranged in age from 46 to 77 years (mean ageers from 44 to 82 (mean age 61.9), and siblings6 (mean age 44.7). Relatives evidencing liver dis-anomaly in standard haematologic tests or routinee were excluded from the study.
tical methods of complex segregation analysis
es of nuclear families were analysed by com-ation analysis (CSA) using the mixed model ted and implemented in the POINTER programTER was used on a SUN Sparc Classic X work-ning the UNIX operating system. In the mixedajor gene, a transmissible multifactorial compo-residual random source of variation are assumed.locus (ML) component is determined by twoucing three possible genotypes. Phenotypes arermally distributed around the mean values of eachs genotype and have equal variances. Complex
M. Clementi et al. / Digestive and Liver Disease 39 (2007) 351355 353
Table 1Population sample
Age groups 1830 3140 4150 5160 6189 1889
Males 14.5 8.3* 12.8 5.2* 12.9 5.4* 13.4 6.0* 12.1 4.8* 13.2 6.1*N 294 347 433 359 206 1639Females 10.6 4.8 10.4 5.1 10.5 4.7 10.1 5.1 10.3 4.2 10.4 4.8N 289 284 299 310 238 1420
Distribution of total serum bilirubin levels (mol/l) by sex and age groups (years). *p < 0.01.
segregation analysis is a general method for evaluating thetransmission of a trait/disease within pedigrees. It proceedsby testing models of varying degrees of generality, both todetermine whether a Mendelian locus is likely to exert a largeeffect on the phenotype (trait or disease) and to estimate themagnitude of genetic sources of variation in the trait. Thesedata are useful to start linkage analysis, and to provide a modelon which to base parametric linkage methods. The major lim-itations of Cof data is gguish betwand the effesimilar transive loci arthey woulddiseasealldisease allewould resudetecting aof inheritancontributinstudies havcomponent
The anaamong agefor these eby POINTcarried outtribution wis consider
a major gene  although restricts power to detect majorgene segregation. Data were then corrected performing lin-ear regression of serum bilirubin values on age, separately formales and females. Residual values were computed for eachindividual as the difference between the observed phenotypeand the predicted value on the basis of the regression model.Residuals were then standardized using the program Skumix that estimates an appropriate value of p for standardized
ble 1s inbin vsignifi
ved bales a
and obtion f
Table 2Segregation a
No familiar e
Tests on ML dHigh valueHigh valueHigh valueHigh value
Tests on the trHigh value
U: The overahomozygousvariance attribSA are that a large amount of a very specific typeenerally needed and that CSA is unable to distin-een the effect of a single locus that underlies a traitcts of two or more independently acting loci withsmission patterns. For example, if many reces-
e involved in determining a complex trait/disease,be detected as if they were a single locus with a
ele frequency equalling that of the sum of severalles. The resulting overestimate of allele frequencylt both in an overestimate of the probability oflocus, as well as in misspecification of the modelce. This error will reduce the power to detect eachg locus. Notwithstanding these limitations, manye used CSA methodology to identify the geneticand the inheritance model in complex diseases.
alization of data
lysed data showed differences between sexes andgroups (Table 1), thus they have been adjustedffects using the control sample. As requested
ER for a quantitative trait, a transformation wasto eliminate skewness and to obtain a normal dis-ith mean = 0 and variance = 1. This transformationed appropriate to guard against false inference of
Regrage ino imand gremo
nalysis of log-transformed standardized residuals (family sample)U V Q
ffect 0 1 (0)ominance:
s: ML recessive .03 1.02 .4488s: ML dominant .03 1.08 .0085
s: ML additive .02 1.01 .0544s: ML general .03 1.02 .4488
ansmission probability 2:s: ML general .01 .98 .4400ll phenotypic mean; V: variance; Q: the gene frequency at major locus which fameans; D: the degree of dominance which corresponds to the relative displacement outable to the multifactorial component; : the transmission probabilities.removes skewness in the presence of up to threed distributions.
ics and data adjustment
shows the distribution of total serum bilirubinmales and females in the population sample.alues in all age classes considered in this studycantly lower (p < 0.01) in females than in males.analysis showed a slight inverse correlation with
es (BILmales = 0.655 + 4.111/age), while age hadon bilirubin levels in females. Effects of ageon observed phenotype (serum bilirubin) were
y performing linear regression on age separatelynd females. Skumix  was used to transformtain a distribution which did not show significant
tion analysis of serum bilirubin log transformedd residuals with the program POINTER is shown
T D 2 2lnL+C1209.55
1.74 (0) 1190.213.10 (1) 1204.04
2.47 (.5) 1206.521.74 0 1190.21
1.70 0 .44 1189.16
vours high values of the trait; T: the displacement betweenf the heterozygous mean; H: the proportion of the phenotypic
354 M. Clementi et al. / Digestive and Liver Disease 39 (2007) 351355
in Table 2. The hypothesis of no familial effect and thehypothesis of no major locus (q = 0) were rejected when com-pared with the major locus model (2 = 19.34, p < .0001). Thehypotheseswere testedwere also rmodel (2quent majoof SB gaveFor this mo45% of the
GS is cof bilirubinoccurrence
its precise mological stuincluding 1ysis we conmode of inhcontrollingmodel.
Duringthat the premoter regireduced UGHowever, agenotype wabout 5%,the UGT1Ais not suffi[6,15]. Thecondition rronmental)concentratithe promotpreviouslymay have alular uptakof organicor indocyain GS subjtransport sycient gluculess than haabnormalitcharacterisals with GSindicatinghaemolyticglucuronylconcurrentactivity [17
Sex and age are additional factors that may be involvedin the regulation of bilirubin metabolism. In our study, datawere obtained in healthy subjects randomly c...