Influence of GST Gene Polymorphisms on the Clearance of Intravenous Busulfan in Adult Patients Undergoing Hematopoietic Cell Transplantation

Influence of GST Gene Polymorphisms on the Clearance of Intravenous Busulfan in Adult Patients Undergoing Hematopoietic Cell Transplantation Sung-Doo Kim,1 Je-Hwan Lee,1 Eun-Hye Hur,1 Jung-Hee Lee,1 Dae-Young Kim,1 Sung-Nam Lim,1 Yunsuk Choi,1 Hyeong-Seok Lim,2 Kyun-Seop Bae,2 Gyu-Jeong Noh,2 Sung-Cheol Yun,3 Sang Beom Han,4 Kyoo-Hyung Lee1 Intravenous (i.v.) busulfan can produce a more consistent pharmacokinetic profile than oral formulations can, but nonetheless, significant interpatient variability is evident. We investigated the influence of polymorphisms of 3 GST isozyme genes (GSTA1, GSTM1, and GSTT1) on i.v. busulfan clearance. Fifty-eight adult patients who received 3.2 mg/kg/day of busulfan as conditioning for hematopoietic cell transplantation were included in this study. Stepwise multiple linear regression demonstrated that GSTA1 variant GSTA1*B (P 5 .004), GSTM1/GSTT1 double-null genotype (P 5 .039), and actual body weight (P 5 .001) were significantly associated with lower clearance of i.v. busulfan. A trend test analyzing the overall effect of GST genotype on busulfan pharmacokinetics, combining GSTA1 gene polymorphism and the number of GSTM1- and GSTT-null genotypes, showed a significant correlation between GST genotype and busulfan clearance (P 5 .001). The clearance of i.v. busulfan was similar between patients with GSTA1*A/*A and GSTM1/GSTT1 doublenull genotypes and those with GSTA1*A/*B and GSTM1/GSTT1 double-positive genotypes. In conclusion, a pharmacogenetic approach using GST gene polymorphisms may be valuable in optimizing the i.v. busulfan dosage scheme. Our results also highlight the importance of including polygenic analyses and addressing interactions among isozyme genes in pharmacogenetic studies. Biol Blood Marrow Transplant 17: 1222-1230 (2011) Ó 2011 American Society for Blood and Marrow Transplantation

KEY WORDS: Hematopoietic cell transplantation, Busulfan, Pharmacokinetics, Glutathione S-Transferase, Pharmacogenetics

INTRODUCTION Hematopoietic cell transplantation (HCT) is a potentially curative therapy of a range of hematologic diseases [1-3]. Busulfan, a bifunctional DNA alkylating agent, is commonly used in conditioning regimens for HCT [3-5]. However, differences in busulfan exposure attributable to variable pharmacokinetics are associated with variations in clinical outcomes: high From the 1Department of Hematology, 2Department of Clinical Pharmacology and Therapeutics, 3Department of Clinical Epidemiology and Biostatistics, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea; and 4Department of Pharmaceutical Analysis, College of Pharmacy, Chung-Ang University, Seoul, Korea. Financial disclosure: See Acknowledgments on page 1229. Correspondence and reprint requests: Je-Hwan Lee, MD, PhD, Department of Hematology, Asan Medical Center, 86, Asanbyeongwon-gil, Songpa-gu, Seoul 138-736, Korea (e-mail: [email protected]). Received October 29, 2010; accepted December 22, 2010 Ó 2011 American Society for Blood and Marrow Transplantation 1083-8791/$36.00 doi:10.1016/j.bbmt.2010.12.708

1222

busulfan exposure is associated with hepatic venoocclusive disease (VOD), whereas low busulfan levels are associated with graft rejection or relapse [6-8]. Wide interpatient variability (10-fold or more) is observed when oral busulfan is used, largely attributable to unpredictable levels of intestinal absorption and individual differences in first-pass metabolism [9,10]. An intravenous (i.v.) formulation of busulfan was developed to overcome the problems associated with oral administration. Intravenous busulfan offers more consistent dosing and pharmacokinetic profiles than do oral formulations [11]. However, interpatient variability in the pharmacokinetics of i.v. busulfan is still considerable [12,13] even after correcting the busulfan dose for body size parameters, including actual body weight (ABW), ideal body weight (IBW), and body mass index (BMI), all of which have been shown to correlate with busulfan clearance [14]. Pharmacokinetics-guided busulfan dose adjustment is an attractive method because blood concentrations are measurable, and there is interpatient variability and a relationship between busulfan exposure and outcome in a narrow therapeutic index

Biol Blood Marrow Transplant 17:1222-1230, 2011

Pharmacogenetics of Intravenous Busulfan Clearance in Adult Patients

[15]. Subsequent dose adjustments to target an area under the concentration-time curve (AUC) based on measurements of the first-dose AUC are typically required in 30% to 40% of patients [16]. However, this approach is limited by time, cost, expertise, and instrumentation. Busulfan is metabolized mainly in the liver by glutathione S-transferases (GSTs) [17,18] a family of phase II detoxification enzymes that catalyze the conjugation of glutathione to various xenobiotics [19,20]. Human GSTs are divided into 3 main families based on subcellular localization: cytosolic, mitochondrial, and microsomal. The cytosolic family is further subdivided into 7 classes: alpha, mu, omega, pi, sigma, theta, and zeta [20]. Increasing number of GST genes have been recognized to be polymorphic, and certain alleles confer impaired catalytic activity [19]. The alpha class of GSTs consists of 5 isoforms (GSTA1-5) that are mostly expressed in the liver [17,18]. The human GSTA1 gene has 3 linked single nucleotide polymorphisms (SNPs) in the promoter region, and 2 genotypes, GSTA1*A and GSTA1*B, have been defined based on the linked polymorphisms 2631 T.G, 2567 T.G, 269 C.T, and 252 G.A [21]. Homozygous GSTA1*B status results in lower transcriptional activity and about a 4-fold reduction in mean hepatic expression compared to the homozygous GSTA1*A situation [21,22]. The mu class contains 5 GST isoforms (GSTM1-5), and the theta class 2 (GSTT1 and GSTT2). GSTM1 and GSTT1 exhibit genetic polymorphisms among human populations, and a large proportion of individuals present with homozygous deletions resulting in a null genotype characterized by a complete lack of the respective protein [22]. The regulation of GST expression also appears to differ according to tissue and cell type. The liver contains high levels of the GSTA1 isozyme; expression of GSTM1 and GSTT1 isozymes in liver tissues has also been confirmed [17,23-25]. Although busulfan is mainly metabolized by GSTA1 in the liver, different GST isozymes may play overlapping roles in busulfan metabolism, and impaired activity of 1 GST isozyme may be compensated by expression of other GSTs [17,26]. GST gene polymorphism may partly explain the interpatient variability of busulfan clearance because higher hepatic GST activity is associated with elevated busulfan clearance and lower plasma busulfan concentrations [27]. Recently, attempts have been made to explain interpatient variability by examination of GST gene polymorphisms, mostly in pediatric patients [28-31]. But, current data regarding the association of busulfan pharmacokinetics and GST polymorphisms are conflicting in pediatric patients. There is only 1 small study in adult patients, regarding the importance of GSTA1 polymorphism in determining busulfan pharmacokinetics [32]. In this study, oral busulfan was administered. Until now, GSTA1, or

1223

GSTM1 has been reported to be associated with busulfan pharmacokinetics [29,32-34]. GSTT1 has been reported to have combined pharmacologic effects with GSTM1 [35-38]. Thus, we investigated the influence of polymorphisms in 3 GST genes, GSTA1, GSTM1, and GSTT1, on the pharmacokinetics of i.v. busulfan in adult patients undergoing HCT. We also analyzed the combined effects of polymorphisms in these 3 GST genes. The ultimate goal of the study was to explore the utility of a pharmacogenetic approach toward optimizing the dosage scheme of i.v. busulfan.

PATIENTS AND METHODS Patients Sixty hematopoietic cell transplant recipients enrolled in our previous study on the pharmacokinetics of i.v. busulfan were included in this pharmacogenetic study [39]. All patients received 3.2 mg/kg/day i.v. busulfan on the first day of conditioning therapy for hematopoietic cell transplantation. This trial was approved by the institutional review board of the Asan Medical Center, Seoul, Korea, and all the patients gave written informed consent. Transplantation Outline Patients were randomized to receive i.v. busulfan (3.2 mg/kg/day) either as a 4-times-daily regimen (0.8 mg/kg over 2 hours, 4 times a day) or a oncedaily regimen (3.2 mg/kg over 3 hours, once a day). All doses of busulfan were calculated using: (1) ABW, if less than or equal to IBW; (2) IBW, if ABW was greater than IBW, but within 120% of IBW; or (3) IBW 1 (0.25
[ABW – IBW]), if ABW exceeded IBW by .120%. Busulfan doses were not adjusted on subsequent days. All patients received 1 of the following conditioning regimens: i.v. busulfan (3.2 mg/kg/day on days 27 to 24) plus i.v. cyclophosphamide (60 mg/kg/day on days 23 and 22; n 5 15); i.v. busulfan (3.2 mg/kg/ day on days 27 and 26) and i.v. fludarabine (30 mg/ kg/day on days 27 to 22) plus varying doses of i.v. anti-thymocyte globulin (ATG) (daily on days 24 to 22 or 21; n 5 33); or i.v. busulfan only (3.2 mg/kg/ day daily on days 26 to 23; n 5 2). No patient received total-body irradiation (TBI). Phenytoin was given for prevention of seizure from the day before the first dose of busulfan to the day after the last dose of busulfan. Nonchemotherapeutic concomitant medications including antiemetics and hydration were used per institutional guidelines in all patients. Pharmacokinetic Studies Blood samples (5 mL) were obtained from all patents at 5 time points for pharmacokinetic studies. In

1224

S.-D. Kim et al.

Biol Blood Marrow Transplant 17:1222-1230, 2011

patients receiving 4 times-daily-busulfan, blood samples were drawn at 2.5, 3, 4, 5, and 6 hours after the start of infusion; in those receiving once-daily busulfan, samples were drawn 3.5, 5, 6, 7, and 22 hours after the start of the infusion. Blood samples, collected in vacutainer tubes (BD Vacutainer, Franklin Lakes, NJ) containing sodium heparin, were centrifuged (2000
g) for 10 minutes at 4 C, and the separated plasma samples were stored at 240 C until assay. Plasma busulfan concentrations were quantified using high-performance liquid chromatography (HPLC) with tandem mass spectrometry (LC-MS/ MS) on an API 3000 triple-quadruple mass spectrometer equipped with an electrospray ion source (MDS SCIEX, South San Francisco, CA). In brief, an aliquot of the sample (20 mL) was delivered into the electrospray ion source via an HPLC system (Agilent 1100 series, Agilent Technologies Inc., Santa Clara, CA) with a C18 Capcell Pak MG column (2.0
50 mm, 3.0-mm particle size). Linear calibration curves were established over a range of busulfan concentrations from 30 to 6000 ng/mL; correlation coefficients were .0.998 in all cases. The limit of quantification for busulfan was 30 ng/mL. The precision of intra- and interday measurements for all analytes was within \12.5%. Busulfan clearance was determined based on a 1-compartment model using WinNonlin v5.0.1 (Pharsight Corporation, Mountain View, CA). Busulfan AUC per dose was calculated by dividing the drug dose by the busulfan plasma clearance estimate.

at 72 C for 1 minute, followed by a final extension at 72 C for 6 minutes. After amplification, PCR products were digested with 10 U of the restriction endonuclease EarI by incubating at 37 C for 12 hours. Homozygous deletions of GSTM1 and GSTT1 were detected using a PCR technique, as described previously [42]. b-Actin was used as an internal control. PCR was carried out in a 50-mL reaction mixture containing 30 ng of DNA template, 5
premix buffer (GeneChem Inc.), and 0.1 mg of the following primer pairs: GSTM1, 50 -GAA CTC CCT GAA AAG CTA AAG C-30 (forward) and 50 -GTT GGG CTC AAA TAT ACG GTG G-30 (reverse), which yield a 215bp fragment; GSTT1, 50 -TCA CCG GAT CAT GGC CAG CA-30 (forward) and 50 -TTC CTT ACT GGT CCT CAC ATC TC-30 (reverse), which yield a 480-bp fragment; and b-actin, 50 -GCC CTC TGC TAA CAA GTC CTA C-30 (forward) and 50 -GCC CTA AAA AGA AAA TCC CCA ATC-30 (reverse), which yield a 345-bp fragment. Amplification conditions consisted of initial denaturation at 95 C for 12 minutes, 35 cycles of denaturation at 94 C for 1 minute, annealing at 62 C for 1 minute, and extension at 72 C for 1 minute, followed by a final extension at 72 C for 10 minutes. Genotype assignments were made based on electrophoretic analyses of amplified products in 3.5% agarose gels. Cases where appropriate PCR products for GSTM1 or GSTT1 were absent and b-actin was present were considered a null genotype (ie, homozygous deletion of GSTM1 or GSTT1).

Identification of GSTA1, GSTM1, and GSTT1 Genotypes

Statistical Analysis

For pretransplant genotype identification, genomic DNA was isolated from peripheral blood lymphocytes using a genomic DNA extraction kit (Qiagen, Chatsworth, CA). For genotyping of GSTA1 (GSTA1*A and GSTA1*B), the 269 C.T variation in the promoter region of GSTA1 was analyzed using the polymerase chain reaction (PCR) followed by restriction fragment-length polymorphism analysis, as previously described [40,41]. It was previously shown that this SNP is in complete linkage disequilibrium with 2631 T.G, 2567 T.G, and 252 G.A [21]. The promoter region of the GSTA1 gene was amplified with the forward primer GSTA1F (50 -TGT TGA TTG TTT GCC TGA AAT T-30 ) and the reverse primers GSTA1-R1 (50 -GTT AAA CGC TGT CAG CCG TCC T-30 ) and GSTA1-R2 (50 -TTT GTT AAA TGC TGT CAC CTT TGT30 ). The PCR mixture contained 30 ng of DNA template, 0.1 mg of each primer (100 pmol), and 5
premix buffer (GeneChem Inc., Daejeon, Korea). PCR conditions were as follows: initial denaturation at 95 C for 10 minutes, 35 cycles of denaturation at 96 C for 1 minute, annealing at 55 C for 1 minute, and extension

Descriptive statistics, including means and standard deviations, and minimum, median, and maximum values, were applied as appropriate. The chi-square test was used to assess deviation of GSTA1 allele frequency from Hardy-Weinberg equilibrium. Categoric variables were compared using chi-square or Fisher’s exact tests, and continuous variables were compared using Student’s t-test. We used Spearman’s correlation coefficient to test relationships between the number of GSTM1- and GSTT1-null genotypes and busulfan pharmacokinetics. A logarithmic transformation of the pharmacokinetic variables (clearance and AUC) was performed to stabilize the models. Busulfan clearance was selected as representative of interpatient variability of the busulfan pharmacokinetics rather than AUC because 2 different dosage regimens (once daily and 4 times daily) were used in our study. AUC was evaluated as a secondary parameter. To select candidate variables for the multiple regression model of busulfan pharmacokinetics (clearance and AUC), we initially examined the individual effects of demographic data (age, gender, ABW, BMI, and height), preconditioning laboratory data (serum total bilirubin, aspartate aminotransferase, alanine

Biol Blood Marrow Transplant 17:1222-1230, 2011

1225

Pharmacogenetics of Intravenous Busulfan Clearance in Adult Patients

aminotransferase, and creatinine), genetic variables (presence of GSTA1*B genotype or number of GSTM1- and GSTT1-null genotypes), and dosing regimen (a 4-times-daily dosing versus a once-daily dosing). Variables that were associated (P \ .10) with busulfan pharmacokinetics and genetic variables were considered further in a stepwise multiple linear regression model. Linear regression analysis was used to test the correlation of combinations of the 3 GST genotypes with busulfan pharmacokinetics. A trend test for allele dose-effect relationships of the GST genotypes was performed using the presence of GSTA1*B genotype and GSTM1/GSTT1-null genotype (the number of GSTM1- and GSTT1-null genotypes), which were entered as ordered categoric variables. P values \.05 were considered statistically significant. Statistical analyses were carried out using the SPSS package (version 14.0, SPSS, Chicago, IL).

RESULTS Patient Characteristics GST polymorphisms were determined in 58 of the 60 hematopoietic cell transplant recipients enrolled in our previous study [39]. Lack of sufficient DNA prevented genotype analysis of 2 patients. Patient demographics and preconditioning laboratory data are shown in Table 1. Acute leukemia, observed in 56.9% of patients, was the most common underlying disease indication for HCT. All but 2 patients received allogeneic HCT. The hematopoietic cell graft donor was an HLA-matched sibling in half of patients. Bone marrow was used in 34 patients (58.6%), and granulocyte colony-stimulating factor (G-CSF)–mobilized peripheral blood mononuclear cells in 24 (41.4%). Frequencies of GST Genotypes Fourteen of the 58 patients (24.1%) were heterozygous for GSTA1*B (GSTA1*A/*B) (Table 2). No

Table 1. Patients’ Demographics and Preconditioning Laboratory Data Characteristic

Number of Patients (Total: 58)

Sex, male versus female Age, year, median (range) Diagnosis AML ALL CML MDS MPD PNH Type of graft donor HLA-matched sibling HLA haploidentical family Unrelated volunteer Autologous Conditioning regimen Busulfan-cyclophosphamide Busulfan-fludarabine-ATG Busulfan ABW, kg; median (range) Height, cm; median (range) Hemoglobin, g/dL; median (range) Total leukocytes,
103/mL; median (range) Platelets,
103/mL; median (range) Aspartate aminotransferase, IU/L; median (range) Alanine aminotransferase, IU/L; median (range) Total bilirubin, mg/dL; median (range) Serum creatinine, mg/dL; median (range)

37 (63.8%) versus 21 (36.2%) 35.5 (16-58) 33 (56.9%) 6 (10.3%) 8 (13.8%) 8 (13.8%) 2 (3.4%) 1 (1.7%) 29 (50.0%) 6 (10.3%) 21 (36.2%) 2 (3.4%) 23 (39.7%) 33 (56.9%) 2 (3.4%) 64.0 (53.0-116.0) 168.1 (146.0-183.0) 10.7 (6.5-14.8) 4.0 (0.6-9.6) 127.5 (11.0-347.0) 21.5 (12-77) 23.5 (7-231) 0.8 (0.4-3.1) 0.8 (0.4-1.5)

AML indicates acute myeloid leukemia or acute mixed lineage leukemia; ALL, acute lymphoblastic leukemia; CML, chronic myeloid leukemia; MPD, myeloproliferative disease; PNH, paroxysmal nocturnal hemoglobinuria; ATG, anti-thymocyte globulin.

patient was homozygous for GSTA1*B (GSTA1*B/*B). The allele frequency of the GSTA1 genotype was in Hardy-Weinberg equilibrium (c2 5 1.09, P 5 .30). The GSTM1-null genotype was detected in 25 patients (43%) and the GSTT1-null genotype in 34 (59%). The baseline characteristics of patients were not significantly different among GST genotypes, except for female predominance in patients with the GSTA1*A/*B genotype (P 5 .023).

Table 2. Genotype Distribution of GSTA1, GSTM1, and GSTT1, and Patient Characteristics of 58 Hematopoietic Cell Transplant Recipients Parameter

n (%)

GSTA1 genotype *A/*A 44 (75.9) *A/*B 14 (24.1) GSTM1 genotype Present 33 (56.9) Null 25 (43.1) GSTT1 genotype Present 24 (41.4) Null 34 (58.6) GSTM1/GSTT1-null genotype Neither 12 (20.7) Either 33 (56.9) Both 13 (22.4)

Age, years

Female, n (%)

Weight, kg

Cr, mg/dL

37.1 ± 10.9 35.2 ± 11.7

12 (27.3)† 9 (64.3)†

66.8 ± 12.0 66.7 ± 9.5

0.86 ± 0.21 0.74 ± 0.24

37.4 ± 10.0 35.7 ± 12.4

13 (39.4) 8 (32.0)

66.7 ± 12.6 66.8 ± 9.8

0.88 ± 0.24 0.78 ± 0.19

35.9 ± 9.3 37.2 ± 12.2

8 (33.3) 13 (38.2)

65.3 ± 8.9 67.8 ± 12.9

0.78 ± 0.16 0.87 ± 0.25

39.3 ± 8.9 34.9 ± 10.1 38.7 ± 14.7

4 (33.3) 13 (39.4) 4 (30.8)

66.1 ± 9.5 66.1 ± 12.4 69.0 ± 10.7

0.83 ± 0.20 0.85 ± 0.23 0.82 ± 0.24

GSTM1/GSTT1-null genotype indicates number of GSTM1- and GSTT1-null genotypes; SD, standard deviation; Cr, serum creatinine. Data are presented as number and percent, or mean 6 SD, as appropriate (†P 5 .023).

1226

S.-D. Kim et al.

Biol Blood Marrow Transplant 17:1222-1230, 2011

Table 3. Association between GST Gene Polymorphisms and Pharmacokinetic Parameters of i.v. Busulfan Parameter

n (%)

GSTA1 genotype *A/*A 44 (75.9) *A/*B 14 (24.1) GSTM1 genotype Present 33 (56.9) Null 25 (43.1) GSTT1 genotype Present 24 (41.4) Null 34 (58.6) GSTM1/GSTT1-null genotype Neither 12 (20.7) Either 33 (56.9) Both 13 (22.4)

Clearance mL/min/kg†

P

AUC, mM$min/mg

P

2.03 ± 0.35 1.79 ± 0.30

.015

31.53 ± 5.56 35.42 ± 6.59

.039

2.01 ± 0.39 1.92 ± 0.28

.382

31.93 ± 6.68 33.18 ± 5.01

.340

2.05 ± 0.29 1.92 ± 0.38

.086

31.49 ± 5.12 33.16 ± 6.54

.355

2.08 ± 0.30 1.99 ± .038 1.82 ± 0.25

.048

30.37 ± 4.30 32.75 ± 6.96 33.70 ± 4.34

.097

AUC indicates dose-normalized area under the plasma concentration-time curve; GSTM1/GSTT1-null genotype, number of GSTM1- and GSTT1-null genotypes; SD, standard deviation. Data are presented as number and percent, or mean 6 SD, as appropriate. †Total plasma clearance.

the number of GSTM1- and GSTT1-null genotypes) was significantly associated with busulfan clearance (P 5 .048). Among demographic and preconditioning laboratory variables, ABW (R 5 2.420; P 5 .001) and BMI (R 5 2.394, P 5 .002) were statistically significant factors with respect to busulfan clearance (data not shown). Because ABW and BMI were highly intercorrelated (R 5 .805, P \ .001), only ABW was included as a body size measure in the multivariate models. Gender was not significantly associated with busulfan clearance (P 5 .950). Thus, the GSTA1 genotype, the GSTM1/GSTT1 genotype combination, and ABW were included in

Factors Associated with Busulfan Clearance

2.5

2.0

1.5

1.0

3.0

2.5

2.0

1.5

1.0

Busulfan clearanc e (m L/ min/ kg)

3.0

Busulfan clearanc e (m L/ min/ kg)

Busulfan clearanc e (m L/ min/ kg)

Busulfan clearanc e (m L/ min/ kg)

Despite adjusting busulfan dosing with respect to body weight, busulfan clearance varied by about 2fold among patients (range: 1.41-2.78 mL/min; median, 1.89 mL/min). Associations between GST gene polymorphisms and the pharmacokinetic parameters of busulfan are listed in Table 3 and depicted in Figure 1. Carriers of GSTA1*B showed significantly lower busulfan clearance than did GSTA1*A/*A carriers (P 5 .015). Although neither GSTM1- nor GSTT1-null genotypes alone were significantly associated with busulfan clearance (P 5 .382 and P 5 .086, respectively), the GSTM1/GSTT1-null genotype (ie,

3.0

2.5

2.0

1.5

1.0

3.0

2.5

2.0

1.5

1.0

Figure 1. Comparison of busulfan clearance distribution according to GST genotypes. Horizontal lines indicate the means of each group. (A) GSTA1, (B) GSTM1, (C) GSTT1, and (D) GSTM1/GSTT1-null genotype (ie, the number of GSTM1- and GSTT1-null genotypes).

Biol Blood Marrow Transplant 17:1222-1230, 2011

Pharmacogenetics of Intravenous Busulfan Clearance in Adult Patients

1227

Table 4. Multiple Linear Regression Model for Prediction of Clearance of i.v. Busulfan Variable

Regression Coefficient

95% CI

Partial R2 (%)

P

GSTA1 genotype* GSTM1/GSTT1 genotype† Actual body weight (in kg)

20.058 20.027 20.003

20.097 to 20.020 20.052 to 20.001 20.004 to 20.001

10.1 5.5 17.6

.004 .039 .001

95% CI indicates 95% confidence interval for regression coefficient. *Coded as 0 if *A/*A and 1 if *A/*B. †Number of GSTM1- and GSTT1-null genotypes.

Impact of the Combination of GST Polymorphisms on Busulfan Clearance We analyzed the overall effect of GST gene polymorphisms on busulfan pharmacokinetics by combining GSTA1 gene polymorphism and the number of GSTM1- and GSTT1-null genotypes (Table 5). A trend test showed that even after adjusting for ABW, GST genotypes were significantly correlated with busulfan clearance (Figure 2, P 5 .001). A similar trend was also observed in the busulfan AUC after inclusion of ABW in the regression model (P 5 .002). Table 5. Clearance of i.v. Busulfan According to Combined GST Gene Polymorphism Clearance (mL/min/kg)† GSTA1

GSTM1/GSTT1-null

No.

mean ± SD

95% CI

*A/*A

Neither Either Both Neither Either Both

7 28 9 5 5 4

2.17 ± 0.31 2.03 ± 0.37 1.91 ± 0.26 1.95 ± 0.25 1.75 ± 0.40 1.63 ± 0.06

1.62-2.89 1.50-2.67 1.42-2.53 1.33-2.38 1.35-2.44 1.21-2.20

*A/*B

GSTM1/GSTT1-null indicates number of GSTM1- and GSTT1-null genotypes; SD, standard deviation; 95% CI, 95% confidence interval for regression coefficient. †P 5 .001.

DISCUSSION The pharmacogenetic evaluation of genes involved in drug pharmacokinetics has been suggested to be an efficient method for individualizing drug therapy, which may lead to improved efficacy or reduced drug toxicity [43]. The variability in busulfan pharmacokinetics has hampered the effectiveness of busulfanbased high-dose conditioning therapy for HCT. Use of an i.v. busulfan formulation reduces, but does not eliminate, interpatient variability in busulfan pharmacokinetics. Recently, GST gene polymorphisms have been recognized to have a significant role as a determinant of interpatient variability in busulfan pharmacokinetics, mostly in pediatric patients. Three such studies analyzed the impact of polymorphisms in GST genes, including GSTA1, GSTM1, GSTP1, and/or GSTT1, on the pharmacokinetics of i.v. busulfan [28-30]. Two of the cited studies found a significant correlation between GSTA1 polymorphic genotype and the clearance of i.v. busulfan [28,29], whereas the other reported no correlation between any GST gene polymorphism and the pharmacokinetics of i.v. busulfan [30]. A study of 12 Japanese adult patients who had received oral busulfan also reported that the

3.0

Busulfan clearance (mL/min/kg)

a multiple regression model of busulfan clearance (Table 4). After stepwise regression, the GSTA1 genotype, the GSTM1/GSTT1 genotype combination, and ABW were all retained in the final model (R2 5 .333, P \ .001), which explained 33.3% of total variability. Notably, the predictive power of GST genotypes (15.6%) was comparable to that of ABW (17.6%). With respect to busulfan AUC, the GSTA1 genotype (P 5 .039), the GSTM1/GSTT1 genotype combination (P 5 .097), ABW (R 5 2.355, P 5 .006), BMI (R 5 2.245, P 5 .064), gender (P 5 .001), serum creatinine (R 5 2.303, P 5 .021), and bilirubin level (R 5 2.218, P 5 .099) were significantly or marginally correlated. A multivariate model that included GST genotypes, ABW, gender, creatinine, and bilirubin showed that the GSTA1 genotype (P 5 .021), the GSTM1/GSTT1 genotype combination (P 5 .036), and ABW (P \ .001) were independent predictors of the busulfan AUC (R2 5 .351, P \ .001), as was the case with busulfan clearance.

2.5

2.0

1.5

1.0

Figure 2. Clearance of i.v. busulfan according to GST gene polymorphism. Horizontal lines indicate the means of each group. Busulfan clearance was grouped by GSTA1 genotype and GSTM1/GSTT1-null genotype (ie, the number of GSTM1- and GSTT1-null genotypes).

1228

S.-D. Kim et al.

GSTA1 genotype was important in busulfan clearance [32]. Another pediatric study that analyzed polymorphisms of GSTA1, GSTM1, and GSTP1 genes found that only the GSTM1 polymorphism was associated with the pharmacokinetics of i.v. busulfan [34]. The reason for these discrepancies in the reported impact of GST gene polymorphisms on the pharmacokinetics of busulfan is not clear, but may include difficulties in interpreting these relationships, differences in study population, sample size, and GST isozymes tested. Pharmacokinetics of busulfan seems to be influenced by a particular combination of GST genotypes. In fact, we observed a significant correlation between pharmacokinetics of busulfan and the GSTM1/ GSTT1 genotype combination, although neither GSTM1 nor GSTT1 alone were significantly correlated with i.v. busulfan clearance (Tables 3 and 4). The biochemical basis for this combined gene effect is not clear, but most drug effects are determined by interplay among several gene products that influence pharmacokinetics and pharmacodynamics [44,45]. Thus, pharmacogenetics and pharmacogenomics are increasingly being applied to identify polygenic determinants of drug effects. Understanding the role of different combinations of genotypes on pharmacokinetics of busulfan is relevant in particular when the complex network of metabolic pathway is considered. Until now, this polygenic approach has also been employed in epidemiologic studies to study the role of GSTM1 and GSTT1 polymorphisms in cancer. These studies have shown that GSTM1 and GSTT1 double-null genotypes tend to have negative prognostic values in various cancers, including acute myeloid leukemia, childhood B-precursor acute lymphoblastic leukemia, and lymphoma [35-37]. Furthermore, in 1 pharmacogenetic study designed to predict drug-induced liver injury, only GSTM1 and GSTT1 double-null genotypes were found to play a role in determining such susceptibility, regardless of the drug involved [38]. In addition, ethnic differences in the frequencies of GST polymorphic genotypes have been reported and may contribute to the variation seen. The frequency of GSTM1- and GSTT1-null genotypes was similar to those found in previously studied Asian population. In the present study, the GSTM1-null genotype was found in 57% of the study population compared with frequencies of 48% to 56% in Asians, 34% to 58% in Caucasians, and 17% to 47% in Africans [46-49]. The frequency of the GSTT1-null genotype was 41% in our study, compared with 45% to 64% in Asians, 15% to 22% in Caucasians, and 20% to 38% in Africans [47,49,50]. The minor allele frequency of GSTA1 in our study (12%) was also similar to that found in previously studied Asian population (12%-14%), compared with that in Caucasians (34-36%) [28,29,32,51].

Biol Blood Marrow Transplant 17:1222-1230, 2011

In our patients, the GSTA1*A/*B genotype was more frequent in women than in men. As such gender differences have not been observed in previous studies, this finding may be insignificant and relate to the limited sample size [51-53]. Busulfan clearance was independent of gender in univariate analysis, and the multiple regression model to predict busulfan clearance was not influenced by the gender factor. Thus, we presented the final multiple regression model without including gender as a covariate, which coincides with other population pharmacokinetic studies [28-30]. A shortcoming of most studies on the biological effects of GSTM1 and GSTT1 genotypes, including our study, is that individuals who are heterozygous or homozygous for the functional alleles are not distinguished, rather being grouped together. As a consequence, the significance of heterozygosity for wild-type GSTM1 or GSTT1 has seldom been addressed [19]. Our investigation of the relationship between GSTA1, GSTM1, and GSTT1 genotypes, and i.v. busulfan pharmacokinetics in adult HCT recipients, showed that GST gene polymorphisms significantly affected the pharmacokinetics of i.v. busulfan and were independent factors underlying interpatient variation in pharmacokinetic parameters. These polymorphic GST enzymes would have reduced excretion of busulfan (and hence, decreased clearance) because reduced GST activities cause lesser degrees of water solubility of busulfan. Of the GST gene polymorphisms tested, the GSTA1 genotype was the most important in predicting busulfan clearance and doseadjusted AUC. Patients of GSTA1*B genotype had significantly reduced busulfan clearance, which was consistent with previous reports [28,29,32] and is consistent with the fact that GSTA1 is the main enzyme involved in busulfan metabolism [17,18]. In the present study, linear regression models demonstrated that both the GSTA1 genotype and the combination of GSTM1/GSTT1 genotypes were independently associated with the clearance and dose-normalized AUC of i.v. busulfan. These results suggest overlapping roles for different GST genes in busulfan metabolism, and prompted us to analyze the overall effect of different GST genes (Table 5). The clearance of i.v. busulfan was similar between patients with GSTA1*A/*A and GSTM1/GSTT1 double-null genotypes and those with GSTA1*A/*B and GSTM1/GSTT1 double-positive genotypes. Our study showed the importance of interactions among isozyme genes, suggesting that expression of GSTM1/GSTT1 may compensate for impaired activity of GSTA1 in the liver metabolism of busulfan. To the best of our knowledge, this is the first comprehensive GST pharmacogenetic analysis to assess the pharmacokinetics of i.v. busulfan in adult patients and to demonstrate the combined effect of GST isozyme genes on the pharmacokinetics of the drug.

Biol Blood Marrow Transplant 17:1222-1230, 2011

Pharmacogenetics of Intravenous Busulfan Clearance in Adult Patients

The goal of pharmacogenetics is to identify genetic differences among patients that influence treatment responses, and devise a revised treatment strategy on the basis of the genotype of the individual. In our study, busufan genotyping explained 15.6% of total variability of busulfan pharmacokinetics, which was comparable to ABW (17.6%). Randomized, controlled trials will be needed to determine whether this genotyping is valuable. In conclusion, our results suggest that a pharmacogenetic approach using GST gene polymorphisms may be useful for optimizing an i.v. busulfan dosage scheme. Our results also highlight the importance of including polygenic analyses and addressing interactions among isozyme genes in pharmacogenetic studies.

ACKNOWLEDGMENTS The authors are grateful to the patients who participated in this study. The authors also thank the nursing and medical staffs who contributed in completing this study.

AUTHORSHIP STATEMENT S.D.K. analyzed the data, performed statistical analyses, and wrote the manuscript; J.H.L. designed and performed experiments, analyzed data, performed statistical analyses, and contributed to the writing of the manuscript; K.S.B. performed pharmacokinetic analyses and analyzed data; H.S.L. and S.B.H. performed pharmacokinetic analyses; S.C.Y. performed statistical analyses; E.H.H. genotyped target genes; and J.H.L., D.Y.K., S.N.L, Y.C., G.J.N., and K.H.L. contributed to the analyses of data. Financial disclosure: This study was supported by a grant of the Korea Health 21 R&D Project, Ministry of Health, Welfare and Family Affairs, Republic of Korea (A062254).

REFERENCES 1. Brochstein JA, Kernan NA, Groshen S, et al. Allogeneic bone marrow transplantation after hyperfractionated total-body irradiation and cyclophosphamide in children with acute leukemia. N Engl J Med. 1987;317:1618-1624. 2. Michel G, Gluckman E, Esperou-Bourdeau H, et al. Allogeneic bone marrow transplantation for children with acute myeloblastic leukemia in first complete remission: impact of conditioning regimen without total-body irradiation—a report from the Societe Francaise de Greffe de Moelle. J Clin Oncol. 1994;12: 1217-1222. 3. Santos GW, Tutschka PJ, Brookmeyer R, et al. Marrow transplantation for acute nonlymphocytic leukemia after treatment with busulfan and cyclophosphamide. N Engl J Med. 1983;309: 1347-1353.

1229

4. Tutschka PJ, Copelan EA, Klein JP. Bone marrow transplantation for leukemia following a new busulfan and cyclophosphamide regimen. Blood. 1987;70:1382-1388. 5. Copelan EA, Deeg HJ. Conditioning for allogeneic marrow transplantation in patients with lymphohematopoietic malignancies without the use of total body irradiation. Blood. 1992; 80:1648-1658. 6. Slattery JT, Sanders JE, Buckner CD, et al. Graft-rejection and toxicity following bone marrow transplantation in relation to busulfan pharmacokinetics. Bone Marrow Transplant. 1995;16: 31-42. 7. Slattery JT, Clift RA, Buckner CD, et al. Marrow transplantation for chronic myeloid leukemia: the influence of plasma busulfan levels on the outcome of transplantation. Blood. 1997; 89:3055-3060. 8. Dix SP, Wingard JR, Mullins RE, et al. Association of busulfan area under the curve with veno-occlusive disease following BMT. Bone Marrow Transplant. 1996;17:225-230. 9. Vassal G. Pharmacologically-guided dose adjustment of busulfan in high-dose chemotherapy regimens: rationale and pitfalls (review). Anticancer Res. 1994;14:2363-2370. 10. Andersson BS, Madden T, Tran HT, et al. Acute safety and pharmacokinetics of intravenous busulfan when used with oral busulfan and cyclophosphamide as pretransplantation conditioning therapy: a phase I study. Biol Blood Marrow Transplant. 2000;6:548-554. 11. Andersson BS, Thall PF, Madden T, et al. Busulfan systemic exposure relative to regimen-related toxicity and acute graftversus-host disease: defining a therapeutic window for i.v. BuCy2 in chronic myelogenous leukemia. Biol Blood Marrow Transplant. 2002;8:477-485. 12. Nath CE, Earl JW, Pati N, Stephen K, Shaw PJ. Variability in the pharmacokinetics of intravenous busulphan given as a single daily dose to paediatric blood or marrow transplant recipients. Br J Clin Pharmacol. 2008;66:50-59. 13. Leger F, Nguyen L, Puozzo C. Exposure equivalence between IV (0.8 mg/kg) and oral (1 mg/kg) busulfan in adult patients. Eur J Clin Pharmacol. 2009;65:903-911. 14. Gibbs JP, Gooley T, Corneau B, et al. The impact of obesity and disease on busulfan oral clearance in adults. Blood. 1999;93: 4436-4440. 15. McCune JS, Holmberg LA. Busulfan in hematopoietic stem cell transplant setting. Expert Opin Drug Metab Toxicol. 2009;5: 957-969. 16. Petros WP, Evans WE. Anticancer agents. In: Burton ME, Shaw LM, Schentag JJ, Evans WE, editors. Applied Pharmacokinetics & Pharmacodynamics Principles of Therapeutic Drug Monitoring. Baltimore, MD: Lippincott Williams & Wilkins; 2006. 17. Czerwinski M, Gibbs JP, Slattery JT. Busulfan conjugation by glutathione S-transferases alpha, mu, and pi. Drug Metab Dispos. 1996;24:1015-1019. 18. Gibbs JP, Czerwinski M, Slattery JT. Busulfan-glutathione conjugation catalyzed by human liver cytosolic glutathione S-transferases. Cancer Res. 1996;56:3678-3681. 19. Hayes JD, Flanagan JU, Jowsey IR. Glutathione transferases. Annu Rev Pharmacol Toxicol. 2005;45:51-88. 20. Strange RC, Spiteri MA, Ramachandran S, Fryer AA. Glutathione-S-transferase family of enzymes. Mutat Res. 2001;482:21-26. 21. Matsuno K, Kubota T, Matsukura Y, Ishikawa H, Iga T. Genetic analysis of glutathione S-transferase A1 and T1 polymorphisms in a Japanese population. Clin Chem Lab Med. 2004;42:560-562. 22. Hayes JD, Strange RC. Glutathione S-transferase polymorphisms and their biological consequences. Pharmacology. 2000; 61:154-166. 23. Abei M, Harada S, Tanaka N, McNeil M, Osuga T. Immunohistochemical localization of human liver glutathione S-transferase (GST) isozymes with special reference to polymorphic GST1. Biochim Biophys Acta. 1989;995:279-284. 24. van Bladeren PJ. Glutathione conjugation as a bioactivation reaction. Chem Biol Interact. 2000;129:61-76.

1230

S.-D. Kim et al.

25. Luckey SW, Petersen DR. Metabolism of 4-hydroxynonenal by rat Kupffer cells. Arch Biochem Biophys. 2001;389:77-83. 26. Landi S. Mammalian class theta GST and differential susceptibility to carcinogens: a review. Mutat Res. 2000;463:247-283. 27. Poonkuzhali B, Chandy M, Srivastava A, Dennison D, Krishnamoorthy R. Glutathione S-transferase activity influences busulfan pharmacokinetics in patients with beta thalassemia major undergoing bone marrow transplantation. Drug Metab Dispos. 2001;29:264-267. 28. Gaziev J, Nguyen L, Puozzo C, et al. Novel pharmacokinetic behavior of intravenous busulfan in children with thalassemia undergoing hematopoietic stem cell transplantation: a prospective evaluation of pharmacokinetic and pharmacodynamic profile with therapeutic drug monitoring. Blood. 2010;115:4597-4604. 29. Johnson L, Orchard PJ, Baker KS, et al. Glutathione S-transferase A1 genetic variants reduce busulfan clearance in children undergoing hematopoietic cell transplantation. J Clin Pharmacol. 2008;48:1052-1062. 30. Zwaveling J, Press RR, Bredius RG, et al. Glutathione S-transferase polymorphisms are not associated with population pharmacokinetic parameters of busulfan in pediatric patients. Ther Drug Monit. 2008;30:504-510. 31. Ansari M, Lauzon-Joset JF, Vachon MF, et al. Influence of GST gene polymorphisms on busulfan pharmacokinetics in children. Bone Marrow Transplant;45:261-267. 32. Kusama M, Kubota T, Matsukura Y, et al. Influence of glutathione S-transferase A1 polymorphism on the pharmacokinetics of busulfan. Clin Chim Acta. 2006;368:93-98. 33. Gaziev J, Nguyen L, Puozzo C, et al. Novel pharmacokinetic behavior of intravenous busulfan in children with thalassemia undergoing hematopoietic stem cell transplantation: a prospective evaluation of pharmacokinetic and pharmacodynamic profile with therapeutic drug monitoring. Blood. 2010;115:4597-4604. 34. Ansari M, Lauzon-Joset JF, Vachon MF, et al. Influence of GST gene polymorphisms on busulfan pharmacokinetics in children. Bone Marrow Transplant. 2010;45:261-267. 35. Voso MT, D’Alo F, Putzulu R, et al. Negative prognostic value of glutathione S-transferase (GSTM1 and GSTT1) deletions in adult acute myeloid leukemia. Blood. 2002;100:2703-2707. 36. Takanashi M, Morimoto A, Yagi T, et al. Impact of glutathione S-transferase gene deletion on early relapse in childhood B-precursor acute lymphoblastic leukemia. Haematologica. 2003; 88:1238-1244. 37. Howells RE, Redman CW, Dhar KK, et al. Association of glutathione S-transferase GSTM1 and GSTT1 null genotypes with clinical outcome in epithelial ovarian cancer. Clin Cancer Res. 1998;4:2439-2445. 38. Lucena MI, Andrade RJ, Martinez C, et al. Glutathione S-transferase m1 and t1 null genotypes increase susceptibility to idiosyncratic drug-induced liver injury. Hepatology. 2008;48:588-596. 39. Ryu SG, Lee JH, Choi SJ, et al. Randomized comparison of four-times-daily versus once-daily intravenous busulfan in con-

Biol Blood Marrow Transplant 17:1222-1230, 2011

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

ditioning therapy for hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2007;13:1095-1105. Coles BF, Morel F, Rauch C, et al. Effect of polymorphism in the human glutathione S-transferase A1 promoter on hepatic GSTA1 and GSTA2 expression. Pharmacogenetics. 2001;11:663-669. Ping J, Wang H, Huang M, Liu ZS. Genetic analysis of glutathione S-transferase A1 polymorphism in the Chinese population and the influence of genotype on enzymatic properties. Toxicol Sci. 2006;89:438-443. Lee HS, Lee JH, Hur EH, et al. Clinical significance of GSTM1 and GSTT1 polymorphisms in younger patients with acute myeloid leukemia of intermediate-risk cytogenetics. Leuk Res. 2009; 33:426-433. Evans WE, McLeod HL. Pharmacogenomics—drug disposition, drug targets, and side effects. N Engl J Med. 2003;348: 538-549. Evans WE, Johnson JA. Pharmacogenomics: the inherited basis for interindividual differences in drug response. Annu Rev Genomics Hum Genet. 2001;2:9-39. Ufer M, Dilger K, Leschhorn L, et al. Influence of CYP3A4, CYP3A5, and ABCB1 genotype and expression on budesonide pharmacokinetics: a possible role of intestinal CYP3A4 expression. Clin Pharmacol Ther. 2008;84:43-46. Garte S, Gaspari L, Alexandrie AK, et al. Metabolic gene polymorphism frequencies in control populations. Cancer Epidemiol Biomarkers Prev. 2001;10:1239-1248. Cho HJ, Lee SY, Ki CS, Kim JW. GSTM1, GSTT1 and GSTP1 polymorphisms in the Korean population. J Korean Med Sci. 2005;20:1089-1092. Lin HJ, Han CY, Bernstein DA, Hsiao W, Lin BK, Hardy S. Ethnic distribution of the glutathione transferase Mu 1-1 (GSTM1) null genotype in 1473 individuals and application to bladder cancer susceptibility. Carcinogenesis. 1994;15: 1077-1081. Rebbeck TR. Molecular epidemiology of the human glutathione S-transferase genotypes GSTM1 and GSTT1 in cancer susceptibility. Cancer Epidemiol Biomarkers Prev. 1997;6:733-743. Nelson HH, Wiencke JK, Christiani DC, et al. Ethnic differences in the prevalence of the homozygous deleted genotype of glutathione S-transferase theta. Carcinogenesis. 1995;16: 1243-1245. Kim I, Keam B, Lee KH, et al. Glutathione S-transferase A1 polymorphisms and acute graft-vs.-host disease in HLAmatched sibling allogeneic hematopoietic stem cell transplantation. Clin Transplant. 2007;21:207-213. Romero RZ, Morales R, Garcia F, et al. Potential application of GSTT1-null genotype in predicting toxicity associated to 5fluouracil irinotecan and leucovorin regimen in advanced stage colorectal cancer patients. Oncol Rep. 2006;16:497-503. Palma S, Cornetta T, Padua L, et al. Influence of glutathione Stransferase polymorphisms on genotoxic effects induced by tobacco smoke. Mutat Res. 2007;633:1-12.