Abstract
Objective. The incidence of nonphysiologic neonatal hyperbilirubinemia is twice as high in East Asians as in whites. We studied whether the condition was associated with mutations in the gene for bilirubin uridine 5′-diphosphate-glucuronosyltransferase (UGT1A1), a key enzyme of bilirubin catabolism.
Design. We analyzed the UGT1A1 gene in 25 Japanese neonates who had nonphysiologic hyperbilirubinemia (serum bilirubin >257 μmol/L) with no obvious cause. They had all received phototherapy. The background control population consisted of 50 Japanese neonates whose transcutaneous jaundice index was monitored during the first week of life. We detected mutations by direct sequencing of polymerase chain reaction-amplified fragments of the gene.
Results. We found a polymorphism for UGT1A1in exon 1; a G→A transition at nucleotide 211 caused arginine to replace glycine at position 71 of corresponding protein product (G71R). The frequency of the mutated allele in the hyperbilirubinemic group (0.34) was significantly higher (χ2 = 5.56) than in the control group (0.16). In the control group the peak transcutaneous jaundice index of the carriers of G71R was significantly higher than it was in the normal infants.
Conclusions. The missense mutation causing G71R is the first reported polymorphism for UGT1A1, and the mutation is a risk factor for nonphysiologic neonatal hyperbilirubinemia. The high incidence of hyperbilirubinemia in the Japanese may be attributable to the high frequency of this missense mutation.
- neonatal hyperbilirubinemia
- bilirubin UDP-glucuronosyltransferase
- UGT1A1
- polymorphism
- Gilbert's syndrome
- UDP =
- uridine 5′-diphosphate •
- PCR =
- polymerase chain reaction
Neonatal hyperbilirubinemia is a common problem and is of concern not only to obstetricians and pediatricians but also to the parents of neonates. Traditionally, authorities have distinguished between physiologic hyperbilirubinemia, which is benign, and nonphysiologic hyperbilirubinemia, which requires further etiologic investigation. Major textbook authors agree that if the total serum bilirubin level exceeds 206 to 223 μmol/L (12 to 13 mg/dL), the bilirubinemia is nonphysiologic.1 ,2 Factors that are associated with nonphysiologic neonatal hyperbilirubinemia include maternal diabetes, East Asian ancestry, low birth weight, prematurity, breast milk feeding, and a sibling with hyperbilirubinemia.1 In half of the cases, however, no such factors are identified.2
Risk factors such as East Asian (Japanese, Korean, or Chinese) ancestry, an affected sibling, and a family history of the condition suggest that a genetic factor is involved,1 ,3 but that factor has not been fully investigated. Recently the structure of the bilirubin uridine 5′-diphosphate (UDP)-glucuronosyltransferase gene,UGT1A1, has been clarified.4 The transferase is a key enzyme for bilirubin metabolism, and mutations ofUGT1A1 cause the unconjugated hyperbilirubinemias known as Crigler-Najjar syndrome and Gilbert's syndrome.5 Patients with Crigler-Najjar syndrome are usually diagnosed just after birth because of their extremely high serum bilirubin concentrations, and they suffer from severe hyperbilirubinemia throughout life. Gilbert's syndrome, on the other hand, which is caused by defects in the same enzyme, is a mild hyperbilirubinemia that generally appears after puberty. Our genetic analysis of patients with Gilbert's syndrome suggested that nonphysiologic neonatal hyperbilirubinemia may be caused by mutations in the UGT1A1 gene. In this report, we studied the relationship between UGT1A1 mutations and nonphysiologic neonatal hyperbilirubinemia.
PATIENTS
We studied 25 Japanese neonates (13 males and 12 females) with nonphysiologic hyperbilirubinemia necessary for phototherapy (serum bilirubin >257 μmol/L) who were born from July 1996 through March 1997 at Ohmi-Hachiman Municipal Hospital in Shiga prefecture. All were full-term with a mean gestational age of 38.2 ± 1.3 weeks (range, 37–41 weeks), and birth weights greater than 2500 g (mean, 3061 ± 393 g; range, 2530–3644 g); they had no known risk factors (maternal diabetes, hemolytic anemia, polycythemia, infection, asphyxia, hypothermia, hypoglycemia, drug therapy, cephalohematoma, or liver dysfunction). The neonates received phototherapy from the second to the fifth day of life, and serum total and unconjugated bilirubin concentrations just before the start of the therapy were 307.8 ± 20.0 μmol/L (range, 259–345 μmol/L) and 295.5 ± 19.2 μmol/L (range, 249–324 μmol/L), respectively. The direct reacting bilirubin concentrations were from 3.4% to 4.6% of the total bilirubin concentrations. The infants were monitored for serum bilirubin level, blood cell count, liver function, and C-reactive protein; all measures were within normal range, except for the serum bilirubin level. Leukocytes were used for the sequencing ofUGT1A1.
The control group, representing a background Japanese population sample, comprised 50 full-term neonates (24 males and 26 females) born in the same hospital from April 1997 to May 1997. Transcutaneous jaundice index of the control group was monitored during the first week of life. Their mean gestational age was 39.1 ± 1.4 weeks (range, from 37–41 weeks) and their birth weights were greater than 2500 g (mean, 3071 ± 428 g; range, 2530–3950 g). Four of the 50 infants received phototherapy. All infants in both groups received a combination of breastfeeding and formula milk.
METHODS
Sequence Analysis of UGT1A1
Genomic DNA was isolated from the leukocytes of patients and the controls by DNAQUICK (Dai-Nippon Pharm, Osaka, Japan), with the informed consent of the relevant parties. Amplification of exons and the promoter region of UGT1A1 by polymerase chain reaction (PCR) from genomic DNA has been described elsewhere.6 ,7 In brief, ∼100 ng of total genomic DNA was amplified by three pairs of oligonucleotide primers. Exons 2, 3, and 4, and their intervening introns were simultaneously amplified as a single DNA fragment by a primer pair of 5′-CTCTATCTCAAACACGCATGCC-3′/TTTTATCATGAATGCCATGACC-3′. The 5′ region of UGT1A1 including the TATA box to exon 1 and exon 5 were amplified separately by primer pairs of 5′-AAGTGAACTCCCTGCTACCTT-3′/5′-GCTTGCTCAGCATATATCTGGG-3′ and 5′GAGGATTGTTCATACCACAGG-3′/5′-GCACTCTGGGGCTGATTAAT-3′, respectively. Conditions for the PCR reaction were as follows: an initial denaturation for 2 minutes at 94°C followed by 1 minute at 94°C, 1 minute at 58°C, and 2 minutes at 72°C for 30 cycles with a Minicyler (MJ Research, Inc, Watertown, MA). A final extension for 8 minutes at 72°C was performed to ensure complete extension of PCR products. Direct sequencing of DNA was performed by the method of Yamada et al.8 Sequence primers used for the determination of TATA box and exon 1 position 211 mutations were 5′-AAGTGAACTCCCTGCTACCTT-3′ and 5′-TTGTTGTGCAGTAAGTGGGA, respectively. Other sequencing primers have been described elsewhere.6 ,7
The PCR products for the control group with both heterozygous exon 1 position 211 and TATA box mutations were subcloned to pCR II vectors using a TA-cloning kit (Invitrogen, San Diego, CA) to examine cis- or trans-arrangement of the two mutations on homologous chromosomes. Thirty ng of PCR fragments, including the regions of exon 1 and the TATA box, were ligated with 50 ng of the vector, and transformation by the ligated products was performed using Competent High JM109 (Toyobo, Osaka, Japan).
Serum bilirubin concentration of infants in the hyperbilirubinemic group was measured by the azo-bilirubin method. Transcutaneous jaundice index of infants in the control group was measured with a Minolta Jaundice Meter (Minolta, Osaka, Japan) during the first week of life. Transcutaneous jaundice index has been shown to be correlated with serum bilirubin concentrations.9
Statistics
Genotype results were analyzed as the hyperbilirubinemic group versus the control group. χ2 analysis was conducted on the raw frequencies. The peak transcutaneous jaundice indexes for the different genotypes in the control group were analyzed by analysis of variance and the Scheffé's test for pairwise comparisons.
RESULTS
Analysis of UGT1A1 revealed that 14 of the 25 neonates (56%) with nonphysiologic hyperbilirubinemia had the identical transition mutation at nucleotide number 211 in exon 1 (Table 1). The mutation changed the codon from GGA to AGA, causing arginine to replace glycine at position 71 (G71R) of the corresponding protein (Fig 1). Three of the 14 cases were homozygous and the other 11 were heterozygous. None of the patients with G71R had the TATA box mutation (two-base insertion mutation: A(TA)7TAA instead of wild type A(TA)6TAA), which is the most common mutation found in white patients with Gilbert's syndrome.7 ,10 No additional mutations were detected in exons 2 through 5 of UGT1A1 in the 14 cases with the missense mutation. Two of the remaining 11 patients without the substitution were heterozygous for A(TA)7TAA (Table 2), and 1 also had a different type of heterozygous mutation in the promoter region: the CAT box changed from the wild type sequence of CAAT to CAGT (not shown). The other 8 cases had no mutation in either the promoter or the coding regions of UGT1A1.
Distribution of UGT1A1 Genotypes and Frequencies of Allele With Mutated Codon 71 (G71R) in the Control and Hyperbilirubinemic Groups
Nucleotide sequences of the mutated section of exon 1 of theUGT1A1 gene amplified from the genomic DNA of normal, heterozygous, and homozygous patients. The transition of G (▿) to A (▾) at position 211 in UGT1A1 cDNA, changes codon 71 from a glycine to an arginine codon (G71R).
Distribution of Genotypes and Allele Frequencies of TATA Box Mutation in the Control and Hyperbilirubinemic Group
In the control group, 13 of 50 (26%) neonates had the transition mutation (1 homozygous and 12 heterozygous) at nucleotide number 211 in exon 1 (causing G71R) (Table 1) and no TATA box mutation. Eleven of the 50 (22%) had the TATA box mutation, A(TA)7TAA (2 homozygous and 9 heterozygous) (Table 2) and no exon 1 mutation. Two of the 50 (4%) were heterozygous for both A(TA)7TAA and the exon 1 mutation (G71R); both types of mutation were in trans-arrangement on homologous chromosomes (not shown).
The frequency of the allele with the codon 71 mutation (G71R) in the hyperbilirubinemic group (0.34) was significantly higher (χ2 = 5.56, P < .025) than it was in the control group (0.16) (Table 1). The frequency of the A(TA)7TAA mutation (0.04) in the hyperbilirubinemic group, however, was significantly lower (χ2 = 5.92, P < .025) than it was in the control group (0.15) (Table 2).
The relationship between genotype and peak transcutaneous jaundice index in the control group is shown in Fig 2. The peak transcutaneous jaundice indexes of normals, heterozygotes of A(TA)7TAA, and heterozygotes of G71R were 20.4 ± 2.48 (n = 24), 19.6 ± 1.99 (n = 9), and 22.8 ± 1.99 (n = 12), respectively. Analysis of variance revealed significant differences among the three genotypes (degrees of freedom: 2/42,F = 6.452, P = .0036). (We did not analyze homozygous infants with G71R [n = 1] and A(TA)7TAA [n = 2] or compound heterozygotes with G71R and A(TA)7TAA [n = 2] because the sample size was small.) The mean peak transcutaneous jaundice index of the heterozygotes of G71R was significantly higher than that of the normals (P = .0162) and the heterozygotes of A(TA)7TAA (P = .0088). However, the mean peak transcutaneous jaundice index of the heterozygotes of A(TA)7TAA was not significantly different (P = .6430) from that of the normals.
Relationship between UGT1A1 mutations and peak transcutaneous jaundice indexes during the first week of life in the control group. Squares and circles represent male and female newborns, respectively. Closed symbols represent the infants who received phototherapy. Mean peak transcutaneous jaundice indexes are indicated by horizontal lines. N* represents no mutation. +P values were obtained by using Scheffé's test.
DISCUSSION
In newborn infants, bilirubin metabolism is still in transition from the fetal to the adult stage. The physiologic unconjugated hyperbilirubinemia that we see in more than half of full-term infants during the first week of life is usually attributable to prematurity of bilirubin UDP-glucuronosyltransferase activity11 and overloading of bilirubin, which is caused by the degradation of fetal hemoglobin.12 Moreover, in the early neonatal period newborns are often exposed to risk factors that cause elevation of serum bilirubin concentration, as described earlier. Those risk factors, however, are present in only approximately half of the neonates with nonphysiologic hyperbilirubinemia.2
As shown in Table 1, the frequency of the missense mutation causing G71R in the control group was 0.16. That suggests an enzyme polymorphism in the Japanese population. In our in vitro expression study, G71R in the homozygous and heterozygous genetic states decreased the enzyme activities to 32% and 60% of normal, respectively.13 That was the first reported evidence of polymorphism for the primary structure of bilirubin UDP-glucuronosyltransferase.
Gene analysis of UGT1A1 showed a statistically significant increase in the frequency of the mutation coding for G71R (0.34) in the hyperbilirubinemic group compared with the control group (0.16) (Table 1). In the control group peak transcutaneous jaundice indexes were higher in neonates heterozygous for the codon 71 mutation (G71R) than in those with the normal allele (Fig 2). Furthermore, we detected the identical UGT1A1 mutation in both the heterozygous and homozygous state in previous studies of Gilbert's syndrome,14 ,15 and that same substitution was the one we found most frequently in Japanese patients with Gilbert's syndrome. Our present results, our in vitro expression studies,13and our accumulated data for Gilbert's syndrome14indicate that G71R in both the heterozygous and homozygous states is an important risk factor for nonphysiologic neonatal hyperbilirubinemia.
In the hyperbilirubinemic group, we found a CAT box mutation that had not previously been reported for UGT1A1. It is uncertain whether the mutation would cause elevation of serum bilirubin levels, but our in vitro expression study demonstrated that the complete change of nucleotide sequence of the CAT box from CCAAT to GTGCT decreased transcriptional activity of UGT1A1 to 62% of normal (unpublished data).
The frequency of the TATA box mutation, A(TA)7TAA, in Japanese is quite low (0.15) (Table 2) compared with that of whites (0.33–0.40).7 ,10 Values nearly matching ours have been reported by Doyama et al16 (0.17) and Ueyama et al17 (0.105). We found no direct relationship between the A(TA)7TAA mutation and neonatal hyperbilirubinemia (Table 2). On the contrary, the frequency of A(TA)7TAA (0.04) was significantly lower in the hyperbilirubinemic group than it was in the control group (0.15). Furthermore, there was no significant increase in the mean transcutaneous jaundice index in control group infants who carried A(TA)7TAA (Fig 2). Recently, Bancroft et al18 also reported that there was no significant difference in the peak transcutaneous jaundice index during the first week of life among the heterozygous and homozygous carriers of A(TA)7TAA and noncarriers, although the A(TA)7TAA homozygotes had a greater increase in jaundice index during the first 2 days of life compared with heterozygotes and noncarriers. Those results and this report indicate that A(TA)7TAA may not be the principal cause of nonphysiologic neonatal hyperbilirubinemia. Gene analysis of heterozygotes carrying both A(TA)7TAA and the codon 71 mutation (G71R) revealed that 2 cases were compound heterozygotes. The lower frequency of A(TA)7TAA in the hyperbilirubinemic group might have been the result of a nonlinked arrangement between the TATA box and codon 71 mutations.
The major mutations found in UGT1A1 differ between Japanese and whites: the predominant mutation in Japanese is the heterozygous codon 71 mutation (G71R),14 and in whites it is the homozygous A(TA)7TAA.7 ,10 Interestingly, the codon 71 mutation has not been reported in whites.10 In the present study, we demonstrated that G71R is a risk factor for nonphysiologic neonatal hyperbilirubinemia both in the heterozygous and homozygous state, but we did not find a significant relationship between the A(TA)7TAA mutation and hyperbilirubinemia. Different genetic backgrounds may contribute to the incidence of nonphysiologic neonatal hyperbilirubinemia being twice as high in Japanese as in whites.
CONCLUSIONS
In summary, we conclude that there is a polymorphism for bilirubin UDP-glucuronosyltransferase: a codon 71 mutation leads to G71R, and G71R is a risk factor for nonphysiologic neonatal hyperbilirubinemia. The high incidence of hyperbilirubinemia in the Japanese may be attributable to the high frequency of G71R.
ACKNOWLEDGMENTS
This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan (Grant 08557030 and Grant 08670878).
We thank R. Nishikawa, A. Sato, and T. Narita of the Department of Pediatrics, and other medical staff at the Department of Gynecology, Ohmi-Hachiman Municipal Hospital, for their clinical treatment and examination of the patients. We also thank T. Tachibana at Institute for Developmental Research for help with statistical analysis.
Footnotes
- Received July 6, 1998.
- Accepted November 17, 1998.
Correspondence to: Dr. Hiroshi Sato, Dept. of Biology, Shiga University of Medical Science, Seta Tsukinowa, Otsu, Shiga 520-2192, Japan. E-mail: satoh{at}bellebsd.shiga-med.ac.jp
REFERENCES
- Copyright © 1999 American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- International Perspectives: Hyperbilirubinemia and Kernicterus in Neonates in China
- Genetic Variants of Human UGT1A3: Functional Characterization and Frequency Distribution in a Chinese Han Population
- Vigintiphobia Revisited
- UDP-GLUCURONOSYLTRANSFERASE 1A4 POLYMORPHISMS IN A JAPANESE POPULATION AND KINETICS OF CLOZAPINE GLUCURONIDATION
- GLUCURONIDATION AND THE UDP-GLUCURONOSYLTRANSFERASES IN HEALTH AND DISEASE
- Bilirubin Genetics for the Nongeneticist: Hereditary Defects of Neonatal Bilirubin Conjugation
- Polymorphisms of UDP-Glucuronosyltransferase Gene and Irinotecan Toxicity: A Pharmacogenetic Analysis
- Prolonged Unconjugated Hyperbilirubinemia Associated With Breast Milk and Mutations of the Bilirubin Uridine Diphosphate- Glucuronosyltransferase Gene