Chemotherapy-induced Unconjugated Hyperbilirubinemia Caused by a Mutation of the Bilirubin Uridine-5[acute]-Diphosphate-Glucuronosyltransferase Gene

Abstract

Chemotherapy for malignant neoplasms sometimes causes unconjugated hyperbilirubinemia in the absence of liver dysfunction. We analyzed the association of chemotherapy-induced hyperbilirubinemia with mutations of the bilirubin uridine-5[acute]-diphosphate (UDP)-glucuronosyltransferase gene (UGT1A1) from two leukemic patients in whom chemotherapy resulted in a hyperbilirubinemic response. We isolated genomic DNA from peripheral blood samples and amplified UGT1A1 by polymerase chain reaction. The amplified DNA fragments were analyzed by direct sequencing. The genes of the two patients revealed an identical heterozygous missense mutation in exon 1 (211G[rarr]A[colon] G71R). This UGT1A1 mutation may be the basis of chemotherapy-induced unconjugated hyperbilirubinemia.

Chemotherapy for malignant neoplasms sometimes induces unconjugated hyperbilirubinemia without liver dysfunction. This hyperbilirubinemia is mild and transient and is thought to be the result of therapy-induced liver damage. Whereas we do not know the cause of the hyperbilirubinemia, we do know that the inherited moderate unconjugated hyperbilirubinemia of Gilbert syndrome, which generally manifests after puberty ⁽¹⁾, is caused by mutations of the bilirubin uridine-5[acute]-diphosphate (UDP)-glucuronosyltransferase gene (UGT1A1) ⁽²⁾. To explore whether there may be a similar genetic basis for chemotherapy-induced unconjugated hyperbilirubinemia, we analyzed the UGT1A1 genes of two patients with cancer (one with acute myeloblastic leukemia and the other with acute lymphoblastic leukemia) in whom chemotherapy had induced unconjugated hyperbilirubinemia with no liver dysfunction.

CASE REPORTS

The patient in our first case report was a 7-year-old Japanese boy with acute myeloblastic leukemia (French-American-British classification M2). After the onset of chemotherapy he had mild, transient, predominantly unconjugated hyperbilirubinemia develop (Fig. 1A) without any evidence of liver dysfunction. Although he was treated with an anti-emetic drug (granisetron hydrochloride) during chemotherapy, he experienced mild emesis. Before starting chemotherapy, his daily caloric intake was approximately 1,500 kcal and during therapy, it was 1,000 to 1,300 kcal. Serum liver enzyme levels remained normal whereas total bilirubin values peaked at 27.3 [mgr]mol/L (1.6 mg/dL). More than 99[percnt] of the serum bilirubin was unconjugated according to high-performance liquid chromatography analysis. The patient was treated with several antileukemic agents, including mitoxantrone hydrochloride, cytarabine, etoposide, pirarubicin hydrochloride, 6-mercaptopurine, and prednisolone. During the chemotherapy, he did not receive any blood transfusions.

FIG. 1.: Serum total bilirubin concentration during chemotherapy. A[colon] Case 1. The patient received pirarubicin hydrochloride on day 1 (35 mg/m²; open arrow), etoposide (100 mg/m²; open triangle), and cytarabine (160 mg/m²; dotted box) from day 2 to day 5, and 6-mercaptopurine (175 mg/m²; solid box) from day 1 to day 5. Before chemotherapy, the total serum bilirubin concentration was normal (5.1 [mgr]mol/L). After chemotherapy was started, it increased (23.9 [mgr]mol/L ) and returned to normal within 1 week after therapy was stopped. Dotted line represents bilirubin concentration at 17 [mgr]mol/L (1 mg/dL). B. Case 2. The patient received pirarubicin hydrochloride (20 mg; open arrow) and vincristine (2 mg; solid arrow) on day 1, 6-mercaptopurine (250 mg; solid box) and prednisolone (120 mg; open box) from day 1 to day 5. Before chemotherapy, the total serum bilirubin concentration was normal (10.2 [mgr]mol/L). After chemotherapy was started, it increased (39.3 [mgr]mol/L ) and returned to normal within 1 week after therapy was stopped.

The patient in our second case report was a 14-year-old Japanese girl with acute lymphoblastic leukemia (French-American-British classification L2). After the onset of chemotherapy, she had mild, transient, unconjugated hyperbilirubinemia develop (Fig. 1B) without any evidence of liver dysfunction. During chemotherapy, she experienced mild emesis despite treatment with an anti-emetic drug (granisetron hydrochloride). Before starting chemotherapy, her daily caloric intake was approximately 1,500 to 1,600 kcal and during therapy, it was 1,000 to 1,350 kcal. Serum liver enzyme levels remained normal whereas total bilirubin values peaked at 39.3 [mgr]mol/L (2.3 mg/dL) (Table 1). More than 99[percnt] of serum bilirubin was unconjugated, based on high-performance liquid chromatography analysis. Serum haptoglobin concentration was 113 mg/dL. The patient was treated with vincristine, cytarabine, cyclophosphamide, pirarubicin hydrochloride, 6-mercaptopurine, and prednisolone. During the chemotherapy, she did not receive blood transfusion.

TABLE 1: Clinical characteristics of two patients with chemotherapy-induced hyperbilirubinemia at peak values of serum bilirubin concentration

METHODS

With the informed consent of the relevant parties, we isolated genomic DNA from peripheral blood samples of patients by DNAQUICK (Dainippon Pharmaceuticals, Osaka, Japan). We analyzed the nucleotides sequence of UGT1A1 by amplification with the polymerase chain reaction ⁽³⁾ and direct sequencing by a method previously described ^(4,5). Serum total bilirubin concentrations were measured by the azo-bilirubin method. We separated the bilirubin fraction by the method of Yamashita et al. ⁽⁶⁾, using a Shimazu high-performance liquid chromatography system (Shimazu, Kyoto, Japan) equipped with a STR ODS-II column (Shinwa Chemical Industrial, Kyoto, Japan).

RESULTS

Both patients were heterozygous for the identical transition mutation at nucleotide number 211 in exon 1. The substitution of adenine for guanine changed the codon from GGA to AGA, causing arginine to replace glycine at position 71 (G71R) of the corresponding protein (Fig. 2). The nucleotides were numbered with reference to the A nucleotide in the initiation codon as [plus]1. The two patients were homozygous for a normal TATA box [lsqb]A(TA)6TAA[rsqb] on the promoter region ⁽³⁾ (TATA box mutation is the most common genetic cause of Gilbert syndrome in whites ⁽¹⁰⁾). We found no mutation in exons 2 through 5.

FIG. 2.: Nucleotide sequences of the mutated section of exon 1 of the UGT1A1 gene amplified from the genomic DNA of a normal subject and two patients (case 1, case 2). Note the mutation, transition of G ([dtri]) to A ([dtrif]) at position 211 in UGT1A1 cDNA, change codon 71 from glycine to arginine (G71R).

DISCUSSION

Patients who receive chemotherapy for malignant neoplasms sometimes have mild and transient unconjugated hyperbilirubinemia develop with no evidence of liver dysfunction. To test whether the chemotherapy-induced hyperbilirubinemia may be caused by a mutant form of bilirubin UDP-glucuronosyltransferase, we investigated the UGT1A1 gene of two such patients. They were children who received a variety of chemotherapeutic drugs for leukemia.

We demonstrated that both patients had the identical heterozygous missense mutation in exon 1 of UGT1A1 in which mutation causes the substitution of arginine for glycine as amino acid 71 on the corresponding enzyme (G71R). In a recent in vitro expression study, we showed that the G71R enzyme had 60[percnt] of normal activity when the mutation was present in the heterozygous state ⁽¹¹⁾. The same mutation was the most common one in our study of Japanese patients with Gilbert syndrome ⁽⁷⁾, an inherited condition characterized by mild unconjugated hyperbilirubinemia that usually becomes manifest after puberty ⁽¹⁾. In recent studies, we showed that G71R was a risk factor for neonatal hyperbilirubinemia ⁽³⁾ and a genetic basis of fasting hyperbilirubinemia ⁽⁹⁾. We showed the frequency of the mutation to be 0.16 in Japanese patients ⁽³⁾.

Serum bilirubin concentrations of patients with Gilbert syndrome are known to fluctuate, and a considerable number of patients with the syndrome do not show persistent hyperbilirubinemia. In those patients, hyperbilirubinemia appears with caloric loads, and fasting is one of the strongest factors in hyperbilirubinemia induction ^(8,9,12,13). Chemotherapeutic agents for malignant neoplasms usually cause emesis and loss of appetite, and patients enter a low-caloric state. During chemotherapy, our two patients experienced mild emesis despite granisetron hydrochloride use, and their daily caloric intakes were reduced to 60[percnt] to 70[percnt] of the pretreatment level. Our recent study of the relationship between fasting hyperbilirubinemia and UGT1A1 mutations showed that reduction of intake to 400 kcal/d was necessary for the elevation of serum bilirubin concentration in men with a heterozygous G71R mutation ⁽⁹⁾. The caloric intakes of the two patients in this study, however, were far more than the value, and the caloric reduction may not have been the cause of the elevation of the serum bilirubin concentration that was diagnosed as fasting hyperbilirubinemia. Based on that fact and the present result, we suggest that chemotherapy can induce hyperbilirubinemia in patients with latent Gilbert syndrome, even in the nonfasting state.

Chemotherapeutic agents (such as 6-mercaptopurine, vincristine, mitoxantrone hydrochloride, and pirarubicin hydrochloride) may induce hyperbilirubinemia as do the agents that inhibit DNA and RNA synthesis ^{(14[ndash]16)} and that reduce the amount of bilirubin UDP-glucuronoyltransferase synthesized in hepatocytes. Patients with UGT1A1 mutations produce a less efficient form of the enzyme, and they may show a transient increase of unconjugated bilirubin.

The present results, data from the genetic analysis of Gilbert syndrome ⁽⁷⁾, our study for the genetic basis of fasting hyperbilirubinemia ⁽⁹⁾, and in vitro expression studies ⁽¹¹⁾ indicate that a mutation of UGT1A1 that expresses the G71R form of bilirubin UDP-glucuronosyltransferase may be one of the causes of chemotherapy-induced unconjugated hyperbilirubinemia. Thus, when patients receiving chemotherapy show unconjugated hyperbilirubinemia without any signs of liver dysfunction, the possibility of a UGT1A1 mutation should be considered. Once such a mutation is confirmed, the therapy can be continued with minimal concern that the chemotherapy is inducing liver damage.

Acknowledgments[colon]

We thank N. Urushiyama of Central Research Laboratory at Shiga University of Medical Science for his technical assistance. This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Science and Culture of Japan (11670494), and Hepatic Disease Research Foundation (Shiga, Japan).