Haplotypes in the UGT1A1 gene and their role as genetic determinants of bilirubin concentration in healthy German volunteers

Available online at www.sciencedirect.com

Clinical Biochemistry 42 (2009) 1635 – 1641

Haplotypes in the UGT1A1 gene and their role as genetic determinants of bilirubin concentration in healthy German volunteers Katrin Borucki a , Cornelia Weikert b,d , Eva Fisher b , Sibylle Jakubiczka c , Claus Luley a , Sabine Westphal a , Jutta Dierkes a,⁎ a Institute of Clinical Chemistry, Medical Faculty University Magdeburg, Germany Department of Epidemiology, German Institute of Human Nutrition, Nuthetal, Germany c Institute of Medical Genetics, Medical Faculty University Magdeburg, Germany Institute for Social Medicine, Epidemiology, and Health Economics, Charité University Medicine Berlin, Germany b

d

Received 11 May 2009; received in revised form 20 August 2009; accepted 22 August 2009 Available online 2 September 2009

Abstract Background: Genetic variations of UDP-glucuronyltransferase 1A1 (UGT1A1) influence the concentration of serum bilirubin. We investigated the association of four common polymorphisms including UGT1A1-53(TA)n, and common haplotypes of the UGT1A1 gene with bilirubin levels in 218 Caucasian volunteers. Methods: Total bilirubin was measured in serum of 218 healthy Caucasian volunteers. Genotyping of four genetic variants was performed: UGT1A1-53(TA)6/7, UGT1A1c.-3279TNG, UGT1A1c.-3156GNA, and UGT1A1c.211GNA. The association between polymorphisms/haplotypes and bilirubin levels were determined. Results: Minor allele frequencies were 0.36 for UGT1A1-53(TA)7, 0.47 for c.-3279G, 0.33 for c.-3156A and 0.006 for c.211A. The three promoter polymorphisms were in close linkage disequilibrium. Common haplotypes were: -53(TA)6/c.-3279T/c.211G (frequency 0.530), -53 (TA)7/c.-3279G/c.211G (frequency 0.365), and -53(TA)6/c.-3279G/c.211G (frequency 0.099). Male sex, UGT1A1-53(TA)6/7 and the c.-3279GG variant were significantly associated with higher bilirubin concentrations. Conclusions: Two UGT1A1 promoter polymorphisms (-53(TA)6/7 and c.-3279TNG) and a common haplotype of the UGT1A1 gene are associated with serum bilirubin concentrations in Caucasians. © 2009 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved. Keywords: UDP-glucuronosyltransferases; Haplotypes; Bilirubin; Caucasians

Introduction UDP-glucuronosyltransferases (UGT) catalyze the conjugation of the glucuronyl group from uridine 5-diphosphoglucuronic acid with endogenous and exogenous substrates. The resulting glucuronide products are more polar, less toxic, and more easily excreted. One prominent endogenous substrate is bilirubin, derived from the catabolism of hemoglobin. The UGTs are a superfamily of phase II biotransformation enzymes, which in ⁎ Corresponding author. Institute of Clinical Chemistry, Leipziger Str. 44, D39120 Magdeburg, Germany. Fax: +49 391 6713902. E-mail address: [email protected] (J. Dierkes).

humans are divided into two families: the gene encoding UGT1A protein family is located on chromosome 2, while the gene encoding UGT2B is located on chromosome 4 [1,2]. The proteins of the UGT1A family share four common exons (exons 2–5), but differ in exon 1 and the promoter region of the gene. Several polymorphisms have been reported, especially in the promoter region and in exon 1, which occur with highly variable frequencies in different populations and regions of the world. Some of them are highly relevant as they exert a significant effect on enzyme function [1,3]. Bilirubin is a non-polar metabolite formed in the catabolism of hemoglobin, which is bound to glucuronic acid in the liver by the UGT1A1 enzyme to form so-called conjugated bilirubin.

0009-9120/$ - see front matter © 2009 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.clinbiochem.2009.08.011

1636

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

Unconjugated hyperbilirubinemia occurs if the activity of the hepatic UGT1A1 enzyme is limited or absent. If the enzyme activity is severely affected, this results in the Crigler–Najjar syndrome type I (OMIM 218800) and type II (OMIM 606785). Gilbert's syndrome (OMIM 143500) is a benign hyperbilirubinemia caused by reduced enzyme activity. A tandem repeat in the 5′ promoter region (UGT1A1-53 (TA)n) has been identified as a main cause of Gilbert's syndrome [4]. Usually, six copies of this tandem repeat are present and are associated with normal enzyme activity, thus representing the wild type enzyme. Fewer than six copies are rare in Caucasian populations and are associated with increased enzyme activity, while more than six copies are associated with reduced enzyme activity. The frequency of an allele with seven copies is about 0.35 in Caucasians, leading to a homozygous genotype in about 10% of the population. More than seven copies of the TA repeat occur only occasionally in Caucasians. Heterozygous subjects usually have normal or slightly elevated bilirubin levels [4–6]. The frequency distribution of this polymorphism is highly variable in different ethnicities with an allele frequency of the (TA)7 allele equal to about 0.10 in Asians, and one of about 0.45 in African Americans [7,8]. The association between the (UGT1A1-53 (TA)n variant and Gilbert's syndrome was identified in 1995 by Bosma and colleagues [4]. Since then, numerous studies investigated the association between this variant and hyperbilirubinemia [9,10]. It has thus been found that this variant is not the only common genetic variant causing hyperbilirubinemia and that not all cases of hyperbilirubinemia are caused by this variant [11–13]. As a result of many investigations carried out during the last few years, it has appeared that this gene variation occurs in linkage disequilibrium with other variations in the promoter region (c.-3156GNA or c.-3279TNG). It has been suggested that the c.-3156GNA polymorphism is also involved in toxicity of irinotecan [14]. The c.-3279TNG polymorphism is located in a phenobarbital-responsive enhancer module of the promoter and it has also been suggested that it is associated with reduced enzyme activity [11] and involved into irinotecan toxicity [15]. A polymorphism in exon 1, (c.211GNA) causing an amino acid change in the protein (G71R), is a genetic cause of hyperbilirubinemia in Asian populations [16,17], but the frequency of this variant in Caucasian populations appears to be low [18]. The combined effects of these polymorphisms on bilirubin concentrations in healthy subjects have not been studied systematically. There is also a lack of information on UGT1A1 haplotypes. We therefore studied the common polymorphisms in the UGT1A1 gene in 343 healthy Caucasian Germans and their effects on bilirubin levels. Patients and methods Recruitment of patients Participants were recruited by advertisements on the campus of the School of Medicine of Otto-von-Guericke University in Magdeburg and during lectures. Participation was voluntary and

was not associated with attendance in courses, lectures or examinations. The advertisements invited students to take part in a screening procedure for a clinical trial which would be limited to subjects with a -53 (TA)7/7 genotype. Thus, it cannot be ruled out that subjects who felt anxious about their bilirubin levels (e.g. those with “yellow eyes”) are over-represented in the sample. However, it was not the aim to arrive at a representative sample but to collect as many healthy subjects as possible. The criteria for inclusion in the screening procedure were as follows: (1) written informed consent to testing of the variants of the UGT1A1 gene, (2) good health and (3) not taking any drugs or medications and reporting a low or moderate alcohol consumption. Volunteers were only included after signing informed consent for genetic testing and if they did not use any medication (except for contraceptives) and if there was no doubt on their reported alcohol consumption. Furthermore, we included only volunteers of Caucasian ancestry. For one part of the study, the screening procedure involved the collection of a venous blood sample during an arranged visit to the Lipid Clinic of the Institute in order to obtain blood cells for DNA analysis and serum for bilirubin measurements. This visit was usually in the morning between 7 and 10 a.m. but did not require that the person was fasting. To increase the effectiveness of the screening procedure a number of subjects were also screened without an appointment. These subjects gave only their written consent, provided a sample of buccal cells, and filled in a health questionnaire. This non-invasive procedure was also carried out in the Lipid Clinic but without appointments. In total, 343 healthy adult volunteers of Caucasian descent were included (mean age 25.3 ± 6.2 years, 193 women and 150 men), recruited between May 2004 and February 2008. Bilirubin was determined in a subcohort (n = 218) due to the availability of serum (see above). The study design and all procedures involving patients were first approved by the Ethics Committee of the School of Medicine of the University of Magdeburg. Each participant gave his or her written informed consent. All procedures were in accordance with the data protection laws and the declaration of Helsinki in its revised form. All data were analyzed after blinding and steps taken to ensure the subjects' anonymity. Genetic and biochemical methods DNA was isolated from white blood cells using the QIAamp DNA blood mini kit or from buccal cells using the QIAamp DNA mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The following gene variations were analyzed: within the promoter sequence: UGT1A1-53(TA)6/7, UGT1A1c.-3279TNG, UGT1A1c.-3156GNA, and within exon 1: UGT1A1c.211GNA. Our nomenclature follows the recommendations of Mackenzie [1,19]. For studying the polymorphisms DNA was amplified by polymerase chain reaction (Master Mix including Taq DNA Polymerase from Promega, Madison, USA) followed by genotype-determination by using in-house dHPLC methods

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

either in the sizing mode at 50 °C (TA-repeat) or SNPspecific dHPLC protocols (c.-3279TNG, c.-3156GNA, c.211GNA). (dHPLC, Transgenomic Ltd, Omaho, USA). Information on primers and details of the dHPLC protocol are listed in Table 1. For each SNP at least 10 subjects with different chromatograms from the dHPLC analysis were sequenced as quality controls. Sequencing was done with an ALFexpress (GE Healthcare Europe, Freiburg, Germany), using the DYEnamic ET terminator kit (GE Healthcare). The concordance rate was 100% for each variant. Bilirubin was determined by the Total Bilirubin liquid test (Roche Diagnostics GmbH, Mannheim, Germany) using a random access analyser in a subsample of n = 218 due to the availability of serum. Total bilirubin was determined by the 2,5dichlorophenyldiazonium (DPD) method [20]. The upper limit of the normal range is 17 μmol/L.

1637

type frequencies and the effect of haplotypes on bilirubin concentrations were calculated using SAS Genetics software (Release 9.1, SAS Institute Inc., Cary, North Carolina, USA). Linkage disequilibrium is expressed both as linkage correlation coefficient r2 and Lewontin's D'. A value of D' = 1 indicates complete linkage disequilibrium in which case one haplotype out of four possible haplotypes deduced from two biallelic variants is not observed, while r2 = 1 indicates a perfect linkage disequilibrium in which case two haplotypes out of four possible haplotypes in two biallelic variants are not observed. Individual haplotypes were calculated by the method described by Zaykin et al. [21]. The haplotypic effect determined by their probabilities (0 = no copy of the haplotype, 0.5 = one copy of the haplotype, and 1 = two copies of the haplotype) on bilirubin concentration was then tested by linear regression with sex as covariate. In all analyses p = 0.05 was taken as the threshold of significance.

Statistical analyses Results Age and bilirubin are treated as continuous variables. The distribution of bilirubin was right-skewed and was normalized by logarithmic transformation. The logarithmically transformed bilirubin (log bilirubin) was then used in all analyses. Univariate analyses whether a genetic variant was associated with log bilirubin concentrations was first done by analysis of variance (ANOVA) with the number of mutated alleles (0, 1, 2) as factor and the log bilirubin as dependent variable. Sex and age-adjusted analyses of the effects of the genetic variants on bilirubin concentrations were tested by linear regression. When the ANOVA revealed an allele-dose effect of a certain variant on the log bilirubin concentration, a co-dominant model was assumed. When the ANOVA revealed similar bilirubin concentrations for those harbouring two major alleles and those who are heterozygous, a recessive model was assumed. In case the ANOVA revealed similar bilirubin concentrations for those being heterozygous and those harbouring two alleles of the minor allele, a dominant model was assumed. These analyses were carried out using SPSS version 15 (SPSS Inc., Chicago, USA). The Hardy–Weinberg equilibrium for single polymorphisms, linkage disequilibrium between genetic variants, haplo-

Allele frequencies of genetic variants and their effect on serum bilirubin concentrations. All polymorphisms were in compliance with the Hardy– Weinberg equilibrium (p N 0.05). The genotype and minor allele frequencies (MAF) and the bilirubin concentrations per genotype are presented in Table 2. While the promoter polymorphisms were common (MAF range 0.332–0.477), the UGT1A1c.211GNA variant was very rare (MAF 0.007). Promoter polymorphisms were found to have a significant effect on bilirubin concentrations, the latter being significantly higher in heterozygous or homozygous carriers of UGT1A1-53 (TA)7 and UGT1A1c.-3156A. The calculated age- and sexadjusted increase in serum bilirubin per allele was 7.4 μmol/L (p = 0.001) for UGT1A1-53(TA)7 and 6.6 μmol/L (p = 0.001) for UGT1A1c.-3156A. For UGT1A1c.-3279TNG, only subjects homozygous for the minor allele (n = 48) had higher bilirubin concentrations than heterozygous (n = 94) or wild type subjects (n = 57, p = 0.001, Table 2). The effect of the UGT1A1 -3279TNG variant was apparent in those who were heterozygous for the UGT1A1 -53(TA)6/7 variant (n = 78). In 67 of these subjects, the UGT1A1-3279TG

Table 1 Protocols for genetic analysis of four gene variants in the UGT1A1 gene by dHLPC. Gene variant

Primer/length of PCR product

Genotyping by dHPLC

UGT1A1⁎28 -53 (TA)6/7 repeat

forward 5′GTCACGTGACACAGTCAAAC3′ reverse 5′TTTGCTCCTGCCAGAGGTT3′ 98 or 96 bp forward 5′ CTAGCCATTCTGGATCCCTTG 3′ reverse 5′TTTTGAGATCTGAGTTCTCTTCACCTC 3′ 367 bp forward 5′AAAGGCGGCATGGCTGTGGA 3′ Reverse 5′GAGGCCATGAGCTCCTTGTTGTGCAGTAAG 3′ 425 bp forward 5′TTCAGGTTATGTAACTAGAGGT3′ reverse 5′CACCTCCTCCTTATTCTCTT3′ 211 bp

Sizing mode, 50 °C

UGT1A1⁎60 promoter c.-3279 TNG

UGT1A1⁎6 Exon 1 211 GNA (Protein: G71R)

UGT1A1⁎93 promoter c.-3156 GNA

Mutation analysis at 59.3 °C

Mutation analysis at 63.0 °C

Mutation analysis at 60.8 °C

1638

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

Table 2 Bilirubin concentrations according to genotype of four UGT1A1 polymorphisms in n = 218 healthy Caucasian volunteers. Variant

Homozygous genotype, major allele

Heterozygous genotype

Homozygous genotype minor allele

UGT1A1-53 (TA)6/7 (n = 218) N (%) Bilirubin, μmol/L UGT1A1 c.-3279 TNG (n = 199) N (%) Bilirubin, μmol/L UGT1A1 c.-3156 GNA (n = 217) N (%) Bilirubin, μmol/L UGT1A1 c.211 GNA (n = 213) N (%) Bilirubin, μmol/L

6/6 99 (45.4) 8.2 (7.5–9.0)a TT 57 (28.7) 8.2 (7.4–9.3)a GG 104 (47.9) 8.3 (7.6–9.3)a GG 210 (98.6) 10.7 (9.9–11.6)a

6/7 80 (36.7) 10.4 (9.4–11.4)b TG 94 (47.2) 9.4 (8.6–10.3)a GA 82 (37.8) 11.2 (10.0–12.6)b GA 3 (1.4) 17.8 (10.0–32.1)a

7/7 39 (17.9) 22.7 (19.4–26.6)c GG 48 (24.1) 19.0 (16.0–22.6)b AA 31 (14.3) 21.5 (18.0–25.6)c AA 0 –

Minor allele frequency

P for trend ⁎ (Model)

0.362 0.001 (co-dominance) 0.477 0.001 (recessive model) 0.332 0.001 (co-dominance) 0.007 0.148 (dominant model)

The exact no. of subjects differs for genotypes and is given for each genotype separately. Bilirubin per genotype is presented as geometric means with 95% confidence intervals. a, b, c Superscript letters: same letters mean no difference at p = 0.05 in serum bilirubin concentration according to one-way ANOVA using log-normalized bilirubin data. ⁎ p value of age and sex-adjusted linear regression analysis using log-normalized bilirubin data.

genotype was present and the associated age-and-sex adjusted bilirubin concentration was 9.9 μmol/L (95% confidence interval 9.0–10.9 μmol/L). In 11 of these subjects, the UGT1A1-3279GG genotype was present and the age- and sex-adjusted bilirubin concentration was 12.9 μmol/L (95% confidence interval 10.2–16.4 μmol/L). The general linear model revealed a significant difference between the UGT1A13279 genotypes (p = 0.045). The linear regression analysis revealed a significant effect of sex and UGT1A1-53(TA)6/7 genotype on log bilirubin concentrations. Sex contributed 8%, and the UGT1A1-53 (TA)6/7 genotype (assuming co-dominance) 35% to the explained variance. When UGT1A1c.-3279TNG, assuming a recessive effect (c.-3279TT combined with GT versus c.3279GG), was included in the model, the variant contributed 15% to the explained variance and reduced the explained variance of UGT1A1-53(TA)6/7 to 23%. The explained variance due to sex was not influenced by UGT1A1c.3279TNG. In total, sex, UGT1A1-53(TA)6/7 and UGT1A1c.3279TNG explained 50% of the variance in log bilirubin.

c.3279G/c.211G with a frequency of 0.099 (These frequencies are based on 343 observations, while the frequencies presented in Table 4 are based on 218 observations with bilirubin measurements). Haplotypes were assigned to individuals with a certain degree of probability (0 = no copy of the haplotype present, 0.5 = one copy of the haplotype present, 1 = two copies of the haplotype present). Table 4 shows the least square means of bilirubin with their 95% confidence limits for all three haplotype probabilities in the 218 subjects with bilirubin measurements and linear trends across those groups. All three haplotypes were found to have a significant effect on the bilirubin concentrations.

Linkage disequilibrium

Table 3 Linkage disequilibrium between four UGT1A1 polymorphisms in the promoter region and in exon 1, in terms of D' and r2.

The three promoter polymorphisms were in close linkage disequilibrium, while UGT1A1c.211GNA was not in linkage disequilibrium with the promoter polymorphisms (Table 3). Since UGT1A1c.-3156GNA was in nearly perfect linkage disequilibrium (r2 = 0.86) with UGT1A1-53(TA)6/7, this variant was omitted from the haplotype analysis.

Discussion This study is a cross-sectional study of the frequency of common polymorphisms in the UGT1A1 gene and of their combined effects on serum bilirubin levels, an important endogenous substrate of the UGT1A1 enzyme. The study was conducted in young and healthy Caucasians living in

Haplotype analysis Three haplotypes with frequencies N 2% were identified: -53 (TA)6/c.-3279T/c.211G with a frequency of 0.530, -53(TA)7/c.3279G/c.211G with a frequency of 0.365, and -53(TA)6/

Upper part of the table: D'. Lower part of the table: correlation coefficient r2. The calculation of linkage disequilibrium was based on genotypes of 343 healthy Caucasian volunteers.

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

1639

Table 4 Bilirubin concentrations according to UGT1A1 haplotypes. Haplotype -53(TA)6/c.-3279T/c.211G n (%) Bilirubin, μmol/L -53(TA)6/c.-3279G/c.211G n (%) Bilirubin, μmol/L -53(TA)7/c.-3279G/c.211G n (%) Bilirubin, μmol/L

Probability = 0

Probability = 0.5

Probability = 1

Haplotype frequency

p (linear trend)

n = 53 (26.8) 22.2 (20.3–24.1)

n = 89 (44.9) 10.3 (8.9–11.8)

n = 56 (28.3) 9.0 (7.2–10.8)

0.526

b0.001

n = 162 (80.5) 13.6 (12.3–15.0)

n = 36 (17.9) 10.8 (7.9–13.7)

n = 3 (1.5) 7.6 (− 2.3 to 17.5)

0.111

0.045

n = 99 (47.1) 9.1 (7.9–10.3)

n = 75 (35.7) 11.1 (9.7–12.5)

n = 36 (17.1) 24.7 (22.7–26.7)

0.360

b0.001

Bilirubin is presented as sex-adjusted log-normalized means with 95% confidence intervals. The calculations are based on haplotypes in 218 healthy Caucasian volunteers with information on genotypes and bilirubin determinations.

Germany. The major findings are that UGT1A1-53(TA)6/7 occurs in complete and nearly perfect linkage disequilibrium with UGT1A1c.-3156GNA. Furthermore, UGT1A1c.211GNA occurs with an allele frequency of less than 1% in this population, thus not fulfilling the definition of a polymorphism. The major determinants of serum bilirubin concentrations in these healthy subjects are sex, UGT1A1-53(TA)6/7, and UGT1A1c.3279TNG. Three haplotypes exhibiting a significant influence on the serum bilirubin concentration were identified. Occurrence of polymorphisms Polymorphisms in the UGT1A1 gene are of interest as genetic determinants of both benign hyperbilirubinemia and of toxicity of the anti-cancer drug irinotecan. Both UGT1A1-53 (TA)6/7 and c.-3156GNA have been associated with hyperbilirubinemia and irinotecan toxicity [22]. Due to the observed nearly complete linkage disequilibrium of UGT1A1-53(TA)6/7 with UGT1A1c.-3156GNA, the relevance of the UGT1A1c.3156GNA can be questioned. UGT1A1-53(TA)7 leads to decreased affinity of TATAA binding protein and consequently the UGT1A1 expression declines [5]. This was not shown for UGT1A1c.-3156GNA. Glucuronidation of SN-38, the active metabolite of irinotecan, did not differ in liver microsomes differing in the UGT1A1c.-3156GNA genotype [8]. Obviously, in Caucasians the UGT1A1c.-3156GNA variant does not occur independently of the UGT1A1-53(TA)6/7 variant. In Caucasian populations, the variant UGT1A1c.211GNA obviously occurs as mutation rather than a polymorphism [18]. In the Japanese population this genetic variant occurs with a minor allele frequency of about 0.15- 0.20, and is associated with neonatal hyperbilirubinemia (OMIM 237900) [17]. Furthermore, it was significantly associated with severe neutropenia in cancer patients receiving irinotecan [23]. In German populations, however, this genetic variant is evidently of only minor importance. Association with serum bilirubin concentrations In awareness with the results of other studies on Caucasians [6,24], we observed a strong and significant association between both sexes and bilirubin and between the UGT1A1-

53(TA)6/7 genotype (assuming co-dominance) and bilirubin, but no interaction between sex and UGT1A1 genotype. UGT1A1c.-3279TNG was entered in the model assuming a recessive mode of action. In fact, a univariate ANOVA did not reveal different bilirubin levels between subjects harbouring either two copies of the major allele or the heterozygous genotype. Our results suggest that the UGT1A1c.-3279TNG genotype is relevant for serum bilirubin levels in addition to the UGT1A1-53(TA)6/7 variant and independently of it. The functional relevance of UGT1A1c.-3279TNG for UGT1A1 activity, neonatal hyperbilirubinemia and irinotecan toxicity has not been investigated in detail. Experimental and clinical observations are hampered by the linkage disequilibrium with UGT1A1-53(TA)6/7. Sugatani et al. [11] suggested that this variant contributes to mild hyperbilirubinemia, showing that the transcriptional activity of the UGT1A1 gene carrying this variant, was about 62–64% of the norm. Costa et al. [25] observed a significantly higher frequency of c.3279G alleles in Caucasian subjects with Gilbert's syndrome compared with control subjects, but did not compare bilirubin levels in the controls. Maruo et al. [10] reported that both variants (UGT1A1-53(TA)6/7 and UGT1A1c.-3279TNG) were in linkage disequilibrium in both Japanese and Caucasian patients with Gilbert's syndrome and in controls, but did not report bilirubin levels. Haplotype analysis Even though there have been a few studies on haplotypes of the UGT1A1 promoter and exon 1 variants, it is evident that the haplotype structure of the promoter and of exon 1 is highly variable between different ethnic groups. In 343 healthy Caucasians we observed three haplotypes with a frequency of N2%. Ki et al. [9] investigated the same genetic variants (c.-3279TNG, -53(TA)n, c.211GNA) in 324 healthy Koreans and observed four haplotypes. The main differences from our study were the high frequency of the c.211A allele (21.3%), which was included in two haplotypes, and the lower frequencies of the c.-3279G allele (26.7%) and the -53(TA)7 allele (12.7%). However, the r2 between the c.-3279TNG locus and the -53(TA)6/7 locus was 0.399, somewhat lower than the r2 of 0.67 in our study. In Caucasians, linkage disequilibrium between the c.-3279TNG and the -53(TA)6/7 loci has been

1640

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

reported by Jirsa et al. [26] and by Innocenti et al. [27]. Jirsa investigated 25 Caucasians with Gilbert's syndrome and 179 subjects with normal bilirubin levels and obtained an r2 of 0.66, which is similar to that found in our study. However, bilirubin levels were not reported. A similar degree of linkage disequilibrium between the -53(TA)6/7 locus and the c.-3279TNG locus in 133 healthy unrelated Caucasians was also reported by Innocenti et al. [27] when the r2 was 0.65. The linkage disequilibrium was much weaker in Asians. Again, no data on bilirubin were reported. The ethnic differences in the UGT1A1 gene were also studied by Kaniwa et al. [7], who investigated the DNA from each of 150 Japanese, Caucasians and African-Americans and confirmed very large ethnic differences in the frequencies of the promoter and exon 1 variants. In addition, the latter authors tested variants in exons 2–5, which occur with relatively low frequency in Caucasians, and reported extremely large ethnic differences. Variants in exons 2–5 were not included in the present study due to their low frequency in Caucasians. Our study is the first report on bilirubin concentrations for different haplotypes in Caucasians. We observed a significant effect of haplotypes on bilirubin concentrations. However, it must be kept in mind, that the biomarker concentration of a particular haplotype can only be calculated with a degree of uncertainty, since haplotypes in heterozygous subjects have to be inferred. Ki et al. [9] determined bilirubin concentrations for various haplotypes in Koreans and reported highest bilirubin values in subjects with the c.-3279G/-53(TA)7/c.211G haplotype (which was rare in Koreans, but made up 36% of our population). In our study, subjects with this haplotype also had the highest bilirubin levels (Table 4). Limitations of the present study It is important to consider potential limitations of our study. In the first phase, this is a cross-sectional analysis with only one serum bilirubin determination. Although the participants gave the blood samples during the morning, we had no control of their food intake, which may influence bilirubin concentrations somewhat. We selected four common gene variants, of which three are frequent in Caucasians. Other variants, especially within the coding region of exons 2–5, have not been selected, though other studies indicated that most of these variants occur rarely in Caucasians [7]. We included healthy subjects with a mean age of 25 years. In such study population of young adults chronic diseases are rare, and the use of medications is also infrequent. Thus, a subjective health assessment was believed to be sensitive enough to exclude subjects with diseases, although other, unidentified sources of hyperbilirubinemia may not be completely excluded by such recruitment procedure. Conclusions The study has shown that in addition to UGT1A1-53(TA)6/7, the variant UGT1A1c.-3279TNG, which is in linkage disequilibrium with UGT1A1-53(TA)6/7, contributes to high bilirubin

levels. On the other hand, it should be noted that the bilirubin elevation associated with these variants is benign in nature and therefore should not be regarded as pathological. However, as UGT1A1-53(TA)6/7 has been associated with severe toxicity of the anticancer drug irinotecan, it would be of great interest to see whether UGT1A1c.-3279TNG is also associated with irinotecan toxicity. Their implication for chronic diseases (e.g. coronary heart disease [28–30]) has to be shown in large epidemiological studies. However, such studies would require a huge sample size. Conflicts of interest The authors declare that none of them has any conflict of interest.

Acknowledgments We are indebted to Elke Hinze for the genetic analyses, and to the technical staff of the Institute of Clinical Chemistry for the bilirubin determinations. This work was made possible through a grant to Katrin Borucki from the Foundation of the German Society of Clinical Chemistry, which is gratefully acknowledged. References [1] Mackenzie PI, Bock KW, Burchell B, Guillemette C, Ikushiro S, Iyanagi T, et al. Nomenclature update for the mammalian UDP glycosyltransferase (UGT) gene superfamily. Pharmacogenet Genomics 2005;15(10):677–85. [2] Gong QH, Cho JW, Huang T, Potter C, Gholami N, Basu NK, et al. Thirteen UDPglucuronosyltransferase genes are encoded at the human UGT1 gene complex locus. Pharmacogenetics 2001;11(4):357–68. [3] Miners JO, McKinnon RA, Mackenzie PI. Genetic polymorphisms of UDPglucuronosyltransferases and their functional significance. Toxicology 2002;181-182:453–6. [4] Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, Oostra BA, et al. The genetic basis of the reduced expression of bilirubin UDPglucuronosyltransferase 1 in Gilbert's syndrome. N Engl J Med 1995 Nov 2;333(18):1171–5. [5] Beutler E, Gelbart T, Bemina A. Racial variability in the UDPglucuronosyltransferase 1 (UGT1A1) promoter: a balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci U S A 1998 Jul 7;95(14):8170–4. [6] Borlak J, Thum T, Landt O, Erb K, Hermann R. Molecular diagnosis of a familial nonhemolytic hyperbilirubinemia (Gilbert's syndrome) in healthy subjects. Hepatology 2000;32(4 Pt 1):792–5. [7] Kaniwa N, Kurose K, Jinno H, Tanaka-Kagawa T, Saito Y, Saeki M, et al. Racial variability in haplotype frequencies of UGT1A1 and glucuronidation activity of a novel single nucleotide polymorphism 686C N T (P229L) found in an African-American. Drug Metab Dispos 2005;33(3):458–65. [8] Innocenti F, Grimsley C, Das S, Ramírez J, Cheng C, Kuttab-Boulos H, et al. Haplotype structure of the UDP-glucuronosyltransferase 1A1 promoter in different ethnic groups. Pharmacogenetics 2002 Dec;12(9): 725–33 Erratum in: Pharmacogenetics. 2003 Mar;13(3):183. [9] Ki CS, Lee KA, Lee SY, Kim HJ, Cho SS, Park JH, et al. Haplotype structure of the UDP-glucuronosyltransferase 1A1 (UGT1A1) gene and its relationship to serum total bilirubin concentration in a male Korean population. Clin Chem 2003;49(12):2078–81. [10] Maruo Y, D'Addario C, Mori A, Iwai M, Takahashi H, Sato H, Takeuchi Y. Two linked polymorphic mutations (A(TA)7TAA and T-3279G) of UGT1A1 as the principal cause of Gilbert syndrome. Hum Genet 2004;115 (6):525–6. [11] Sugatani J, Yamakawa K, Yoshinari K, Machida T, Takagi H, Mori M, et al. Identification of a defect in the UGT1A1 gene promoter and its

K. Borucki et al. / Clinical Biochemistry 42 (2009) 1635–1641

[12] [13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

association with hyperbilirubinemia. Biochem Biophys Res Commun 2002 Mar 29;292(2):492–7. Strassburg CP. Pharmacogenetics of Gilbert's syndrome. Pharmacogenomics 2008;9(6):703–15. Ferraris A, D'Amato G, Nobili V, Torres B, Marcellini M. Dallapiccola B. Combined test for UGT1A1-3279T N G and A(TA)nTAA polymorphisms best predicts Gilbert's syndrome in Italian pediatric patients. Genet Test 2006;10(2):121–5. Innocenti F, Undevia SD, Iyer L, Chen PX, Das S, Kocherginsky M, et al. Genetic variants in the UDP-glucuronosyltransferase 1A1 gene predict the risk of severe neutropenia of irinotecan. J Clin Oncol 2004 Apr 15;22(8): 1382–8. Kitagawa C, Ando M, Ando Y, Sekido Y, Wakai K, Imaizumi K, et al. Genetic polymorphism in the phenobarbital-responsive enhancer module of the UDP-glucuronosyltransferase 1A1 gene and irinotecan toxicity. Pharmacogenet Genomics 2005;15(1):35–41. Akaba K, Kimura T, Sasaki A, Tanabe S, Ikegami T, Hashimoto M, et al. Neonatal hyperbilirubinemia and mutation of the bilirubin uridine diphosphate-glucuronosyltransferase gene: a common missense mutation among Japanese, Koreans and Chinese. Biochem Mol Biol Int 1998;46(1): 21–6. Akaba K, Kimura T, Sasaki A, Tanabe S, Wakabayashi T, Hiroi M, et al. Neonatal hyperbilirubinemia and a common mutation of the bilirubin uridine diphosphate-glucuronosyltransferase gene in Japanese. J Hum Genet 1999;44(1):22–5. Sava M, Kraemer DM. Heterozygous G71R-mutation causing Gilbert's syndrome in one of 103 random persons of German descent. J Hepatol 2005;42(5):778–9. Mackenzie PI, Owens IS, Burchell B, Bock KW, Bairoch A, Bélanger A, et al. The UDP glycosyltransferase gene superfamily: recommended nomenclature update based on evolutionary divergence. Pharmacogenetics 1997;7(4):255–69. Wahlefeld AW, Herz G, Bernt E. Modification of the Malloy–Evelyn method for a simple, reliable determination of total bilirubin in serum [Abstract]. Scand J Clin Lab Invest 1972;29(Suppl 126):11–2.

1641

[21] Zaykin DV, Westfall PH, Young SS, Karnoub MA, Wagner MJ, Ehm MG. Testing association of statistically inferred haplotypes with discrete and continuous traits in samples of unrelated individuals. Hum Hered 2002;53: 79–91. [22] Coté JF, Kirzin S, Kramar A, Mosnier JF, Diebold MD, Soubeyran I, et al. UGT1A1 polymorphism can predict hematologic toxicity in patients treated with irinotecan. Clin Cancer Res 2007;13:3269–75. [23] Minami H, Sai K, Saeki M, Saito Y, Ozawa S, Suzuki K, et al. Irinotecan pharmacokinetics/pharmacodynamics and UGT1A genetic polymorphisms in Japanese: roles of UGT1A1⁎6 and ⁎28. Pharmacogenet Genomics 2007;17(7):497–504. [24] Rauchschwalbe SK, Zühlsdorf MT, Schühly U, Kuhlmann J. Predicting the risk of sporadic elevated bilirubin levels and diagnosing Gilbert's syndrome by genotyping UGT1A1⁎28 promoter polymorphism. Int J Clin Pharmacol Ther 2002;40(6):233–40. [25] Costa E, Vieira E, Dos Santos R. The polymorphism c.-3279T N G in the phenobarbital-responsive enhancer module of the bilirubin UDPglucuronosyltransferase gene is associated with Gilbert syndrome. Clin Chem 2005;51(11):2204–6. [26] Jirsa M, Petrasek J, Vitek L. Linkage between A(TA)7TAA and -3279T N G mutations in UGT1A1 is not essential for pathogenesis of Gilbert syndrome. Liver Int 2006;26(10):1302–3. [27] Innocenti F, Liu W, Chen P, Desai AA, Das S, Ratain MJ. Haplotypes of variants in the UDP-glucuronosyltransferase1A9 and 1A1 genes. Pharmacogenet Genomics 2005 May;15(5):295–301. [28] Lin JP, O'Donnell CJ, Schwaiger JP, Cupples LA, Lingenhel A, Hunt SC, et al. Association between the UGT1A1⁎28 allele, bilirubin levels, and coronary heart disease in the Framingham Heart Study. Circulation 2006 Oct;114(14):1476–81. [29] Gajdos V, Petit FM, Perret C, Mollet-Boudjemline A, Colin P, Capel L, et al. Further evidence that the UGT1A1⁎28 allele is not associated with coronary heart disease: the ECTIM Study. Clin Chem 2006;52(12):2313–4. [30] Bosma PJ, van der Meer IM, Bakker CT, Hofman A, Paul-Abrahamse M, Witteman JC. UGT1A1⁎28 allele and coronary heart disease: the Rotterdam Study. Clin Chem 2003;49(7):1180–1.