Associations Between Drug Metabolism Genotype, Chemotherapy Pharmacokinetics, and Overall Survival in Patients With Breast Cancer
http://www.100md.com
《临床肿瘤学》
the West Virginia University Health Sciences Center, Morgantown, WV
Duke University Medical Center, Durham, NC
DNA Sciences Inc, Fremont, CA
ABSTRACT
PURPOSE: To evaluate associations between patient survival, pharmacokinetics, and drug metabolism–related genetic polymorphisms in patients receiving a combination chemotherapy regimen for breast cancer.
PATIENTS AND METHODS: A genotype association study was conducted on 85 chemotherapy-na?ve patients with metastatic or inflammatory breast cancer that were evaluated for an extended period after receiving standard-dose chemotherapy followed by high-dose cyclophosphamide, cisplatin, and carmustine. Blood pharmacokinetics were evaluated, and DNA was genotyped for 29 polymorphisms in 17 drug metabolism genes.
RESULTS: Patients with cyclophosphamide plasma exposures above the median (implying slower metabolic activation) had a shorter survival than those below the median (1.8 v 3.8 years, respectively; P = .042). Patients having a variant genotype of cytochrome P450 3A4 displayed higher blood concentrations of parent (inactive) cyclophosphamide with the second and third doses (P = .024 and .028, respectively) in addition to slower cyclophosphamide activation over the three doses (P = .031). Median survival for these patients was 1.3 years compared with 2.7 years for those without the variant (P = .043). Similar results were observed for patients carrying a genetic variant of P450 3A5. Median survival for patients with deletions of glutathione-S-transferase M1 gene was 3.5 v 1.5 years for patients with one or both copies (P = .041). Patients with a polymorphism in a gene regulating metallothionein had lower platinum concentrations and shorter survival (P = .033).
CONCLUSION: These data suggest that pretreatment evaluation of drug metabolism genes may explain some interindividual differences in both anticancer drug pharmacokinetics and response. The correlations found here may have implications for other commonly used anticancer drugs.
INTRODUCTION
The clinical relevance of interpatient variability in drug disposition and effect is most evident for agents with low therapeutic indices (ie, the dose prescribed is close to the dose likely to produce toxicity in most individuals). Such characterization is typical of many clinically used anticancer drugs due to their nonspecific cytotoxicity and because the doses thought necessary for optimal eradication of malignant cells are close to those that damage normal cells. Substantial interpatient differences in both clinical response and toxicity are common with many cancer chemotherapy regimens, yet clinicians have few logistically feasible drug monitoring strategies to maximize therapeutic indices in individual patients.1
The rapidly evolving field of pharmacogenetics holds great promise for assisting selection of patient-individualized treatment regimens and dosages.2 A vast number of single nucleotide polymorphisms have been discovered in genes thought to be involved in the regulation of drug metabolism; however, relatively few studies have been conducted that establish a link between genotype and drug efficacy. It is estimated that several million polymorphisms are found in the human genome; however, a majority of these lie in regions of DNA that do not code for proteins and thus may not be of biologic importance. A major focus of current research entails clinical evaluation of polymorphisms for their impact on gene function. Such studies are difficult since nongenetic factors also can have a significant impact on the drug metabolism phenotype. Yet metabolic enzyme genetics could play a very important role in the disposition of prodrugs, such as cyclophosphamide, which require biochemical conversion in order to elicit their bioactivity.
We have recently reported an association between the pharmacokinetics of cyclophosphamide and overall survival following an intensive chemotherapy regimen for patients with high-risk, locally advanced breast cancer.3 The current study was conducted in a separate cohort of patients in order to substantiate the importance of anticancer drug pharmacokinetics with this regimen and evaluate the influence of metabolic enzyme pharmacogenetics on outcome in women with metastatic breast cancer. In order to accomplish this goal, we evaluated the pharmacokinetic and pharmacogenetic data from a subset of patients whose clinical response was reported in a large, previously published trial.
PATIENTS AND METHODS
Patients, Treatment Regimen, and Response Assessment
Consent was received from all patients, and they were enrolled between 1987 and 1995 on our previously published, institutional review board–approved, single-institution, randomized clinical trial of high-dose chemotherapy for metastatic or inflammatory breast cancer.4 Inclusion criteria for this trial mandated that either the cancer be hormone receptor–negative or the patient fail hormonal therapy (if the tumor was hormone receptor positive before enrollment). In order to limit variables that could confound our ability to evaluate relationships between systemic chemotherapy exposure, genotype, and outcome, all 85 patients who had not received chemotherapy before enrolling in the clinical trial were utilized for this study.
The patients received standard-dose induction chemotherapy consisting of doxorubicin, fluorouracil, and intermediate-dose methotrexate (AFM), followed by high-dose chemotherapy with hematopoietic support (recombinant, colony-stimulating factor–primed, autologous peripheral blood progenitor cells, and nonprimed autologous bone marrow).5
The high-dose chemotherapy consisted of cyclophosphamide, cisplatin, and carmustine as previously described.6 Cyclophosphamide was administered at a dose of 1,875 mg/m2/d (intravenously over 1 hour) on days –6 through –4. Cisplatin was delivered as a continuous infusion of 55 mg/m2/d on days –6 through –4, followed by carmustine 600 mg/m2 (intravenously over 2 to 3 hours) on day –3. Autologous hematopoietic support was provided on days –1 through +1. Patients were consolidated with radiation therapy or surgery post-transplantation if these therapies were likely to convert them to complete responders. We have previously shown that such a strategy leads to improvements in event-free, but not overall, survival.7
Following completion of high-dose chemotherapy, patients underwent disease restaging every 3 months for 1 year and then every 6 months for 5 years. Survival was measured from the date of initial restaging, approximately 21 days before transplantation.
Pharmacokinetic Analyses
The pharmacokinetics of cyclophosphamide, carmustine, and non–protein bound cisplatin were evaluated using serial blood samples collected on each of the high-dose treatment days as previously described by our laboratory.8,3 The concentrations of drugs were analyzed in all samples within several days of their collection via assays validated for clinical laboratory use. Each dose of cyclophosphamide was modeled independently due to metabolic induction over the treatment regimen. Compartmental analyses were conducted in the standard, two-stage approach using WinNonlin software (version 2.1; Pharsight Inc, Mountain View, CA).
Pharmacogenetic Analyses
Genomic DNA was extracted from each patient's cryopreserved peripheral blood lymphocytes (obtained under institutional review board–approved informed consent) using Gentra PureGene kit K-50 (Gentra, Minneapolis, MN). Each patient's DNA was then diluted to 4 ng/μL, then 5 μL (20 ng) was used for each polymerase chain reaction.
Twenty-nine previously identified polymorphisms in 17 genes potentially important in the metabolism of chemotherapy agents, or which may have an association with therapeutic response of breast cancer, were selected for evaluation in this study (Table 1).
Polymorphisms were detected by the TaqMan method.9,10 Two polymerase chain reaction primers were designed to flank the probe sequence for each variant giving an amplification product of approximately 100 bp and having melting temperatures compatible with the given probes. Allele typing was obtained by comparing the ratio of fluorescence of the two probes in comparison with controls whose genotypes had been verified by sequencing.
Statistical Analyses
Genotype results for polymorphisms were integrated with demographic and clinical outcome data for all 85 patients. Demographic data included age, ethnicity, and menopausal status. Clinical outcome data included response to therapy (complete response, partial response, no response, disease progression), progression-free survival, and overall survival.
Differences in median plasma chemotherapy systemic exposure of carmustine, cisplatin, and cyclophosphamide were compared by genotype using the Wilcoxon rank sum test.
Survival time was defined as the length of time from restaging to death. Patients who were alive were censored at date of last follow-up visit.
We estimated overall survival according to the Kaplan-Meier product limit method and applied the log-rank test to compare two distributions based on genotypes or pharmacokinetics.11
All statistical analyses were performed using SAS for Windows version 8.02 (SAS Institute, Cary, NC).12 Results that yielded a P value .05 were considered statistically significant. No adjustments were made to account for number of statistical tests being conducted.
Cox proportional hazards multivariate modeling was used to identify genotypes that significantly predict overall and progression-free survival. The Spearman correlation coefficient was used to determine pairwise associations between genotypes.
RESULTS
Patient Demographics
All 85 patients were female, with a median age of 44 years. Eighty-seven percent of the population was white, while African Americans accounted for the remaining 13%. A complete response to high-dose chemotherapy was evident in 52% of the patients. Five of the complete responders (6% of the total population) were African American and 39 (46%) were white (Table 2). No patients died due to chemotherapy-induced toxicity. The median overall survival for all patients was 2.2 years (95% CI, 1.7 to 3.3 years), with a median follow-up of 8.2 years (range, 2 to 12.5 years).
Pharmacokinetics
Pharmacokinetic parameters were fully assessable in 71 patients for cyclophosphamide, 61 patients for carmustine, and 52 patients for cisplatin (Tables 2 and 3). The reasons for lack of data include insufficient samples collected on at least one of the treatment days for cyclophosphamide and carmustine, whereas some patients did not have cisplatin data due to inaccurate sample processing.
Carmustine plasma concentration time data sets were adequately described with an open, linear, one-compartment pharmacokinetic model. The median systemic exposure was 529 μg/mL x min, but values varied over a 12-fold range among patients. Ultrafiltrable plasma platinum concentrations varied over a three to four-fold range within each sample time, and the median values increased over the course of the 3-day infusion, as anticipated (Table 3). Neither carmustine nor platinum systemic exposure was associated with overall survival.
Cyclophosphamide disposition was estimated using an open, linear, two-compartment model. The systemic exposure to parent drug varied over a three-fold range and declined within each patient by approximately one half after the first dose, compared with later doses, as anticipated. Patients who achieved overall systemic exposure value below the population median (suggesting extensive metabolic activation) experienced a significantly longer overall survival than those above the median (3.8 [95% CI, 2.4 to 10.1] v 1.8 [95% CI, 1.2 to 2.5] years, respectively; P = .042; Fig 1).
Pharmacogenetics
Genotype data for potentially important polymorphisms were available in approximately 70% of patients (Table 1). All polymorphisms could not be determined in all patients due to limited sample volume or degradation of DNA. Polymorphisms were within the probability limits of Hardy-Weinberg equilibrium stratified by race, except for CYP2B6_C-1184GT (P = .009) in white patients. Given the relatively small sample size and large number of polymorphisms assayed, one polymorphism would be expected, by chance, to be in disequilibrium. The allelic frequencies of polymorphisms associated with CYP3A4 and CYP3A5 were substantially higher in African Americans compared with white patients, as anticipated (Table 1). Approximately 26% of patients had at least one polymorphic allele for either CYP3A4 or CYP3A5. Genotypes for CYP3AP1*1 were in virtually complete concordance with those of CYP3A5*1. The GSTM1 null genotype occurred in approximately one half of the white patients, but was infrequent in the African Americans (10%).
The median pharmacokinetic parameters differed by genotype of CYP3A4, CYP3A5, GSTM1, and MET1F. Differences in the systemic exposure (area under the concentration-time curve [AUC]) of cyclophosphamide became evident by dose 2 and persisted into dose 3 when patients were segregated based on the CYP3A4*1B polymorphism (P = .024 and P = .028, respectively; Table 2). Those with the polymorphic variant displayed higher AUC of parent drug, suggesting less enzyme activity. Likewise, apparent induction of cyclophosphamide metabolism (dose 1 AUC/dose 3 AUC) was reduced significantly for patients with that polymorphism (P = .031). The AUC of cyclophosphamide also seemed to be higher by the second dose in patients with the CYP3A5*1 variant, or its pseudogene, CYP3AP1*1 (P = .002 and .049, respectively; Table 2). No differences in cyclophosphamide AUC were found when patients were segregated based on ethnicity alone, despite the higher frequency of CYP3A4/5 polymorphisms in the African American patients (Table 1).
Patients with homozygous GSTM1 null genotype had a 19% lower median carmustine AUC (504 μg/mL x min) compared with patients with either one or two functional alleles (620 μg/mL x min; P = .031). Further, the clinical response to therapy was 58% in the null group compared with 44% in those with one or two functional alleles (P = .22).
Concentrations of platinum at 72 hours after administration were significantly lower in patients with polymorphisms in MET1F G-7T (P = .003; Table 3). Segregation of the cisplatin pharmacokinetic data by genotype revealed that patients with the polymorphism did not experience accumulation of platinum in the plasma ultrafiltrate (Wilcoxon rank sum test, P = .012).
Survival time was found to be associated with genotypes from the following polymorphisms: CYP3A4*1B, CYP3A5*1, CYP3AP1*1, GSTM1 null, and MET1FG-7T. Median survival of patients with the CYP3A4 polymorphism was 1.3 years (95% CI, 0.6 to 2.1) compared with 2.7 years (95% CI, 1.8 to 4.1) for patients with both copies of the common allele (P = .043; Fig 2A). Similar differences in median survival years were observed for polymorphisms in CYP3A5 (95% CIs, 0.8 to 2.2 v 2.1 to 7.4; P = .002; Fig 2B) and for polymorphisms in the promoter of MET1F (95% CIs, 0.9 to 3.2 v 1.7 to unbounded; P = .033; Fig 2C). Survival increased for patients with the null genotype of GST M1 compared with patients with one or two functional alleles. Patients with one or two wild-type alleles of GSTM1 showed a median survival of 1.5 years (95% CI, 1.2 to 2.6), whereas patients with the null genotype of GSTM1 had a median survival of 3.5 years (95% CI, 2.2 to unbounded; P = .041; Fig 2D).
Spearman correlations were conducted between the variants found to be important to pharmacokinetics of the drugs or overall survival. Significant associations were found between the GSTM1 variant and CYP3A5*1 (r = 0.28; P = .014), in addition to one between CYP3A5*1 and CYP3A4*1B (r = 0.62; P < .0001).
A Cox proportional hazards multivariate model found that patients had risk ratios for disease progression of 2.4 and 1.7 if they had CYP3A5*1 and MET 1F G-7T genotypes, respectively (P = .0078 and .076, respectively). Similarly, a multivariate evaluation for overall survival demonstrated a 2.6 and 1.9 risk ratio for death if patients had CYP3A5*1 (P = .0035) and MET1F G-7T (P = .029) genotypes, respectively.
DISCUSSION
It is logical to assume that the observed variability in clinical response to anticancer agents is due to both tumor and host factors. In the current study, we demonstrate links between the variability in systemic drug disposition (pharmacokinetics) with specific genetic variability and outcome.
Associations between the pharmacokinetics of anticancer drugs and their clinical toxicities are well documented in the literature13; however, relatively little data are available that substantiate relationships between drug pharmacokinetics and anticancer efficacy.14 We found a significant association between the disposition of high-dose cyclophosphamide and overall survival in a chemotherapy-na?ve group of patients with metastatic breast cancer. These results corroborate our previous findings with the same high-dose regimen when it was used in earlier stage disease,3 as well as those results observed by others.15
It is not clear which of the CYP450 isoenzymes are involved with in vivo activation of cyclophosphamide in humans; however, in vitro studies have primarily implicated CYP3A4/5, CYP2C9, and CYP2B6.16-18 We found that polymorphisms in CYP3A4 and CYP3A5 are associated with a reduced systemic clearance of parent cyclophosphamide and a poorer overall clinical response. These data suggest an important clinical role for the CYP3A isoenzymes in cyclophosphamide metabolic activation.
CYP3A4 and CYP3A5 are thought to be the most important CYP 450 isoenzymes pertaining to drug metabolism, due to the diversity of drugs that they act on, as well as their relative abundance in humans.19 Several polymorphisms have been identified in CYP3A4; however, controversy exists as to their functional significance.20 We found that patients with a genetic variant in the 5' promoter region (CYP3A4_A292G; CYP3A4*1B allele) had a higher AUC and shorter overall survival, which could result from a reduced metabolic activation of cyclophosphamide. This polymorphism has been previously associated with a higher clinical stage/grade of prostate cancer and a higher risk of treatment-related leukemia.21,22 However, altered constitutive protein expression in subjects with the variant has not been shown.23,24 Our data support this result since we did not see differences in the pharmacokinetic parameters until the metabolic induction process was evident on the second and third doses per day of cyclophosphamide. It will thus be important to expand this research to investigate polymorphisms in other genes that control transcription of CYP3A in such a setting.25
Multiple drugs used in clinical practice are inhibitors of CYP3A, including some antihistamines, antibiotics, antivirals, and importantly, antiemetics.26 Our data would suggest that concomitant use of such drugs with cyclophosphamide has potential to reduce its clinical effectiveness. We recently reported that patients who receive this chemotherapy regimen display a 43% induction in CYP3A metabolic activity after three doses of cyclophosphamide when compared with the level measured just before therapy,27 consistent with data from studies conducted in vitro.28 Thus, patients who receive cyclophosphamide concurrently with other drugs metabolized by CYP3A should also be monitored closely for accelerated clearance of the latter agents.
We found a similar association of a variant in CYP3A5 (CYP3A5_A22893G; CYP3A5*1), with both survival and cyclophosphamide pharmacokinetics. It has been shown that this polymorphism is in linkage disequilibrium with a pseudogene found upstream of the CYP3A5 promoter (CYP3A5_5UTR_A-44G; CYP3AP1*1), which is in agreement with our data.29,30 A recently published study by Floyd et al using midazolam as a probe and rifampin as a metabolic inducer in 57 healthy subjects suggests that the CYP3A5*1 polymorphism is particularly important in the setting of CYP3A enzyme induction.31 This study and ours both show that CYP3A is less inducible for individuals with CYP3A5*1 or CYP3A4*1B. One possible explanation for the reduced activation of cyclophosphamide in women carrying the CYP3A5*1 allele is that they fail to induce total CYP3A levels to the same extent as those homozygous for *3, since the 3A5 gene does not appear to be inducible.32,33
The results of this study pertaining to CYP3A have potentially dramatic implications for the pharmacogenetic evaluations of many drugs used in clinical care today. This isoenzyme is known to be involved in the metabolism of approximately one half of US Food and Drug Administration–approved drugs, spanning all therapeutic classes.34 Some anticancer drugs that are metabolized via CYP3A include: docetaxel, paclitaxel, irinotecan, tamoxifen, and imatinib. In addition, it is an important enzyme involved in the metabolic conversion of endogenous estrogens to potential carcinogens.35 The variant CYP3A alleles are relatively uncommon in whites but occur in approximately 50% of African Americans.36 These variants may be partially responsible for decreased cyclophosphamide efficacy previously observed in African Americans.37
Cytosolic glutathione-S-transferase (GST) enzymes catalyze reactions between lipophilic compounds with electrophilic centers and glutathione. Their primary role is believed to be in facilitating the detoxification of reactive molecules such as environmental carcinogens and xenobiotics (eg, cisplatin). Some polymorphisms in GSTs result in complete deletion of the gene, and thus no functional protein. These occur in up to 50% of white subjects and approximately 28% of African Americans. The GST null genotype occurred in only 10% of the patients on our study; however, this may be due to our relatively small number of patients in this ethic group. We found that patients with the GSTM1 null genotype had a prolonged overall survival. This is consistent with data from other studies that have demonstrated high expression of some GSTs resulted in lower response to chemotherapy,38-40 whereas patients with null alleles may exhibit higher degrees of toxicity.41
Metallothioneins are thought to play a role in neutralization of toxic heavy metals and are highly inducible (both intracellularly and in plasma) by a variety of xenobiotics, including cisplatin.42,43 Upregulation of metallothionein in cancer cells can render them resistant to cisplatin.44,45 The size of metallothionein (molecular weight, 6,000 to 7,000 Da) when bound to cisplatin is well below the molecular weight cutoff of the ultrafiltration apparatus used to prepare our clinical samples. Thus, we suspect that cisplatin induces plasma metallothionein, and this accounts for the accumulation of the ultrafiltrable platinum over the 72-hour infusion seen here and previously.3
Our data suggest that evaluation of independent patient pharmacogenetics may provide a logistically feasible mechanism for tailoring the dosage of chemotherapy in hopes of optimizing efficacy and minimizing toxicity. The confounding factor of higher frequencies of some polymorphisms based on ethnicity, suggests that confirmatory studies should be conducted to attempt establishment of similar relationships within ethnic groups. Such would then allow for formal evaluation of dose individualization based on a patient's genetic profile. When coupled with the rapidly evolving field of tumor genomics,46 these strategies could provide a substantial impact on improving the therapeutic index of anticancer regimens.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported in part by NIH 5-P01-CA47741 and NIH P50-CA68438; Specialized Program in Research Excellence in Breast Cancer, NIH 5-P30-CA14236-26, NIH UO1 CA84955; and the Mylan Chair of Pharmacology.
Presented in part at the American Association for Cancer Research Annual Meetings, New Orleans, LA, 2001 (abstr 1435) and Orlando, FL, 2004 (abstr 2106); however, no other aspects of the data have been published.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Evans WE, Hon YY, Bomgaars L, et al: Preponderance of thiopurine S-methyltransferase deficiency and heterozygosity among patients intolerant to mercaptopurine or azathioprine. J Clin Oncol 19:2293-2301, 2001
Nagasubramanian R, Innocenti F, Ratain MJ: Pharmacogenetics in cancer treatment. Annu Rev Med 54:437-452, 2003
Petros WP, Broadwater G, Berry D, et al: Association of high-dose cyclophosphamide, cisplatin and carmustine pharmacokinetics with response, toxicity and dosing weight in patients with primary breast cancer. Clin Cancer Res 8:698-705, 2002
Rizzieri DA, Vredenburgh JJ, Jones R, et al: Prognostic and predictive factors for patients with metastatic breast cancer undergoing aggressive induction therapy followed by high-dose chemotherapy with autologous stem-cell support. J Clin Oncol 17:3064-3074, 1999
Jones RB, Shpall EJ, Shogan J, et al: The Duke AFM Program: Intensive induction chemotherapy for metastatic breast cancer. Cancer 66:431-436, 1990
Peters WP, Ross M, Vredenburgh JJ, et al: High-dose chemotherapy and autologous bone marrow support as consolidation after standard-dose adjuvant therapy for high-risk primary breast cancer. J Clin Oncol 11:1132-1143, 1993
Carter DL, Marks LB, Bean JM, et al: Impact of consolidation radiotherapy in patients with advanced breast cancer treated with high-dose chemotherapy and autologous bone marrow rescue. J Clin Oncol 17:887-893, 1999
Gilbert CJ, Petros WP, Vredenburgh J, et al: Pharmacokinetic interaction between ondansetron and cyclophosphamide during high-dose chemotherapy for the treatment of breast cancer. Cancer Chemother Pharmacol 42:497-503, 1998
Abdel-Rahman SZ, el-Zein RA, Anwar WA, et al: A multiplex PCR procedure for polymorphic analysis of GSTM1 and GSTT1 genes in population studies. Cancer Lett 107:229-233, 1996
Livak KJ: Allelic discrimination using fluorogenic probes and the 5[prime] nuclease assay. Genet Anal 14:143-149, 1999
Cox DR: Regression models and life-tables. J R Stat Soc B 34:187-220, 1972
SAS/Stat user guide, version 8. SAS Institute Inc. SAS Campus Drive, Cary, NC
Ratain MJ, Plunkett W: Pharmacology, in Kufe DW, Pollock RE, Weichselbaum RR, et al (eds): Cancer Medicine (ed 6). London, UK, BC Decker, 2003
Evans WE, Relling MV, Rodman JH, et al: Conventional compared with individualized chemotherapy for childhood acute lymphoblastic leukemia. N Engl J Med 338:499-505, 1998
Ayash LJ, Wright JE, Tretyakov O, et al: Cyclophosphamide pharmacokinetics: Correlation with cardiac toxicity and tumor response. J Clin Oncol 10:995-1000, 1992
Ren S, Yang J-S, Kalhorn TF, et al: Oxidation of cyclophosphamide to 4-hydroxycyclophosphamide and deschloroethylcyclophosphamide in human liver microsomes. Cancer Res 57:4229-4235, 1997
Roy P, Yu LJ, Crespi CL, et al: Development of a substrate-activity based approach to identify the major human liver P-450 catalysts of cyclophosphamide and ifosfamide activation based on cDNA-expressed activities and liver microsomal P-450 profiles. Drug Metab Dispos 27:655-666, 1999
Xie H-J, Yasar U, Lundgren S, et al: Role of polymorphic human CYP2B6 in cyclophosphamide bioactivation. Pharmacogenomics J 3:53-61, 2003
Gellner K, Eiselt R, Hustert E, et al: Genomic organization of the human CYP3A locus: Identification of a new, inducible CYP3A gene. Pharmacogenetics 11:111-121, 2001
Flockhart DA, Rae JM: Cytochrome P450 3A pharmacogenetics: The road that needs traveled. Pharmacogenomics J 3:3-7, 2003
Rebbeck TR, Jaffe JM, Walker AH, et al: Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4. J Natl Cancer Inst 90:1225-1229, 1998
Felix CA, Walker AH, Lange BJ, et al: Association of CYP3A4 genotype with treatment-related leukemia. Proc Natl Acad Sci U S A 95:13176-13181, 1998
Ball SE, Scatina J, Kao J, et al: Population distribution and effects on drug metabolism of a genetic variant in the 5[prime] promoter region of CYP3A4. Clin Pharmacol Ther 66:288-294, 1999
Westlind A, Lofberg L, Tindberg N, et al: Interindividual differences in hepatic expression of CYP3A4: relationship to genetic polymorphism in the 5[prime]-upstream regulatory region. Biochem Biophys Res Commun 259:201-205, 1999
Zhang J, Kuehl P, Green ED, et al: The human pregnane X receptor: Genomic structure and identification and functional characterization of natural allelic variants. Pharmacogenetics 11:555-572, 2001
Majumdar AK, McCrea JB, Panebianco DL, et al: Effects of aprepitant on cytochrome P450 3A4 activity using midazolam as a probe. Clin Pharmacol Ther 74:150-156, 2003
Petros W, Vredenburgh J, Fehdrau R, et al: Evaluation of CYP3A4 metabolic phenotype and cyclophosphamide pharmacokinetics in patients receiving high-dose therapy. Proc Am Assoc Cancer Res 43:274, 2002 (abstr 318)
Lindley C, Hamilton G, McCune JS, et al: The effect of cyclophosphamide with and without dexamethasone on cytochrome P450 3A4 and 2B6 in human hepatocytes. Drug Metab Dispos 30:814-821, 2002
Paulussen A, Lavrijsen K, Bohets H, et al: Two linked mutations in transcriptional regulatory elements of the CYP3A5 gene constitute the major genetic determinant of polymorphic activity in humans. Pharmacogenetics 10:415-424, 2000
Finta C, Zaphiropoulos PG: The human cytochrome P450 3A locus: Gene evolution by capture of downstream exons. Gene 260:13-23, 2000
Floyd MD, Gervasini G, Masica AL, et al: Genotype-phenotype associations for common CYP3A4 and CYP3A5 variants in the basal and induced metabolism of midazolam in European- and African-American men and women. Pharmacogenetics 13:595-606, 2003
Kuehl P, Zhang J, Lin Y, et al: Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet 27:383-391, 2001
Goh B-C, Lee S-C, Wang L-Z, et al: Explaining interindividual variability of docetaxel pharmacokinetics and pharmacodynamics in Asians through phenotyping and genotyping strategies. J Clin Oncol 20:3683-3690, 2002
de Wildt SN, Kearns GL, Leeder JS, et al: Cytochrome P450 3A: Ontogeny and drug disposition. Clin Pharmacokinet 37:485-505, 1999
Huang Z, Guengerich FP, Kaminsky LS: 16Alpha-hydroxylation of estrone by human cytochrome P4503A4/5. Carcinogenesis 19:867-872, 1998
Walker AH, Jaffe JM, Gunasegaram S, et al: Characterization of an allelic variant in the nifedipine-specific element of CYP3A4: ethnic distribution and implications for prostate cancer risk. Hum Mutat 12:289, 1998
Dooley MA, Hogan S, Jennette C, et al: Cyclophosphamide therapy for lupus nephritis: Poor renal survival in black Americans—Glomerular Disease Collaborative Network. Kidney Int 51:1188-1195, 1997
Ambrosone CB, Sweeney C, Coles BF, et al: Polymorphisms in glutathione S-transferases (GSTM1 and GSTT1) and survival after treatment for breast cancer. Cancer Res 61:7130-7135, 2001
Cheng X, Kigawa J, Minagawa Y, et al: Glutathione S-transferase-pi expression and glutathione concentration in ovarian carcinoma before and after chemotherapy. Cancer 79:521-527, 1997
Bai F, Nakanishi Y, Kawasaki M, et al: Immunohistochemical expression of glutathione S-transferase-Pi can predict chemotherapy response in patients with nonsmall cell lung carcinoma. Cancer 78:416-421, 1996
Peters U, Preisler-Adams S, Hebeisen A, et al: Glutathione S-transferase genetic polymorphisms and individual sensitivity to the ototoxic effect of cisplatin. Anticancer Drugs 11:639-643, 2000
Vandier D, Calvez V, Massade L, et al: Transactivation of the metallothionein promoter in cisplatin-resistant cancer cells: A specific gene therapy strategy. J Natl Cancer Inst 92:642-647, 2000
Tohyama C, Shaikh ZA: Metallothionein in plasma and urine of cadmium-exposed rats determined by a single-antibody radioimmunoassay. Fundam Appl Toxicol 1:1-7, 1981
Kelley SL, Basu A, Teicher BA, et al: Overexpression of metallothionein confers resistance to anticancer drugs. Science 241:1813-1815, 1988
Toyoda H, Mizushima T, Satoh M, et al: HeLa cell transformants overproducing mouse metallothionein show in vivo resistance to cis-platinum in nude mice. Jpn J Cancer Res 91:91-98, 2000
Evans WE, McLeod HL: Pharmacogenomics: Drug disposition, drug targets, and side effects. N Engl J Med 348:538-549, 2003(William P. Petros, Penelo)
Duke University Medical Center, Durham, NC
DNA Sciences Inc, Fremont, CA
ABSTRACT
PURPOSE: To evaluate associations between patient survival, pharmacokinetics, and drug metabolism–related genetic polymorphisms in patients receiving a combination chemotherapy regimen for breast cancer.
PATIENTS AND METHODS: A genotype association study was conducted on 85 chemotherapy-na?ve patients with metastatic or inflammatory breast cancer that were evaluated for an extended period after receiving standard-dose chemotherapy followed by high-dose cyclophosphamide, cisplatin, and carmustine. Blood pharmacokinetics were evaluated, and DNA was genotyped for 29 polymorphisms in 17 drug metabolism genes.
RESULTS: Patients with cyclophosphamide plasma exposures above the median (implying slower metabolic activation) had a shorter survival than those below the median (1.8 v 3.8 years, respectively; P = .042). Patients having a variant genotype of cytochrome P450 3A4 displayed higher blood concentrations of parent (inactive) cyclophosphamide with the second and third doses (P = .024 and .028, respectively) in addition to slower cyclophosphamide activation over the three doses (P = .031). Median survival for these patients was 1.3 years compared with 2.7 years for those without the variant (P = .043). Similar results were observed for patients carrying a genetic variant of P450 3A5. Median survival for patients with deletions of glutathione-S-transferase M1 gene was 3.5 v 1.5 years for patients with one or both copies (P = .041). Patients with a polymorphism in a gene regulating metallothionein had lower platinum concentrations and shorter survival (P = .033).
CONCLUSION: These data suggest that pretreatment evaluation of drug metabolism genes may explain some interindividual differences in both anticancer drug pharmacokinetics and response. The correlations found here may have implications for other commonly used anticancer drugs.
INTRODUCTION
The clinical relevance of interpatient variability in drug disposition and effect is most evident for agents with low therapeutic indices (ie, the dose prescribed is close to the dose likely to produce toxicity in most individuals). Such characterization is typical of many clinically used anticancer drugs due to their nonspecific cytotoxicity and because the doses thought necessary for optimal eradication of malignant cells are close to those that damage normal cells. Substantial interpatient differences in both clinical response and toxicity are common with many cancer chemotherapy regimens, yet clinicians have few logistically feasible drug monitoring strategies to maximize therapeutic indices in individual patients.1
The rapidly evolving field of pharmacogenetics holds great promise for assisting selection of patient-individualized treatment regimens and dosages.2 A vast number of single nucleotide polymorphisms have been discovered in genes thought to be involved in the regulation of drug metabolism; however, relatively few studies have been conducted that establish a link between genotype and drug efficacy. It is estimated that several million polymorphisms are found in the human genome; however, a majority of these lie in regions of DNA that do not code for proteins and thus may not be of biologic importance. A major focus of current research entails clinical evaluation of polymorphisms for their impact on gene function. Such studies are difficult since nongenetic factors also can have a significant impact on the drug metabolism phenotype. Yet metabolic enzyme genetics could play a very important role in the disposition of prodrugs, such as cyclophosphamide, which require biochemical conversion in order to elicit their bioactivity.
We have recently reported an association between the pharmacokinetics of cyclophosphamide and overall survival following an intensive chemotherapy regimen for patients with high-risk, locally advanced breast cancer.3 The current study was conducted in a separate cohort of patients in order to substantiate the importance of anticancer drug pharmacokinetics with this regimen and evaluate the influence of metabolic enzyme pharmacogenetics on outcome in women with metastatic breast cancer. In order to accomplish this goal, we evaluated the pharmacokinetic and pharmacogenetic data from a subset of patients whose clinical response was reported in a large, previously published trial.
PATIENTS AND METHODS
Patients, Treatment Regimen, and Response Assessment
Consent was received from all patients, and they were enrolled between 1987 and 1995 on our previously published, institutional review board–approved, single-institution, randomized clinical trial of high-dose chemotherapy for metastatic or inflammatory breast cancer.4 Inclusion criteria for this trial mandated that either the cancer be hormone receptor–negative or the patient fail hormonal therapy (if the tumor was hormone receptor positive before enrollment). In order to limit variables that could confound our ability to evaluate relationships between systemic chemotherapy exposure, genotype, and outcome, all 85 patients who had not received chemotherapy before enrolling in the clinical trial were utilized for this study.
The patients received standard-dose induction chemotherapy consisting of doxorubicin, fluorouracil, and intermediate-dose methotrexate (AFM), followed by high-dose chemotherapy with hematopoietic support (recombinant, colony-stimulating factor–primed, autologous peripheral blood progenitor cells, and nonprimed autologous bone marrow).5
The high-dose chemotherapy consisted of cyclophosphamide, cisplatin, and carmustine as previously described.6 Cyclophosphamide was administered at a dose of 1,875 mg/m2/d (intravenously over 1 hour) on days –6 through –4. Cisplatin was delivered as a continuous infusion of 55 mg/m2/d on days –6 through –4, followed by carmustine 600 mg/m2 (intravenously over 2 to 3 hours) on day –3. Autologous hematopoietic support was provided on days –1 through +1. Patients were consolidated with radiation therapy or surgery post-transplantation if these therapies were likely to convert them to complete responders. We have previously shown that such a strategy leads to improvements in event-free, but not overall, survival.7
Following completion of high-dose chemotherapy, patients underwent disease restaging every 3 months for 1 year and then every 6 months for 5 years. Survival was measured from the date of initial restaging, approximately 21 days before transplantation.
Pharmacokinetic Analyses
The pharmacokinetics of cyclophosphamide, carmustine, and non–protein bound cisplatin were evaluated using serial blood samples collected on each of the high-dose treatment days as previously described by our laboratory.8,3 The concentrations of drugs were analyzed in all samples within several days of their collection via assays validated for clinical laboratory use. Each dose of cyclophosphamide was modeled independently due to metabolic induction over the treatment regimen. Compartmental analyses were conducted in the standard, two-stage approach using WinNonlin software (version 2.1; Pharsight Inc, Mountain View, CA).
Pharmacogenetic Analyses
Genomic DNA was extracted from each patient's cryopreserved peripheral blood lymphocytes (obtained under institutional review board–approved informed consent) using Gentra PureGene kit K-50 (Gentra, Minneapolis, MN). Each patient's DNA was then diluted to 4 ng/μL, then 5 μL (20 ng) was used for each polymerase chain reaction.
Twenty-nine previously identified polymorphisms in 17 genes potentially important in the metabolism of chemotherapy agents, or which may have an association with therapeutic response of breast cancer, were selected for evaluation in this study (Table 1).
Polymorphisms were detected by the TaqMan method.9,10 Two polymerase chain reaction primers were designed to flank the probe sequence for each variant giving an amplification product of approximately 100 bp and having melting temperatures compatible with the given probes. Allele typing was obtained by comparing the ratio of fluorescence of the two probes in comparison with controls whose genotypes had been verified by sequencing.
Statistical Analyses
Genotype results for polymorphisms were integrated with demographic and clinical outcome data for all 85 patients. Demographic data included age, ethnicity, and menopausal status. Clinical outcome data included response to therapy (complete response, partial response, no response, disease progression), progression-free survival, and overall survival.
Differences in median plasma chemotherapy systemic exposure of carmustine, cisplatin, and cyclophosphamide were compared by genotype using the Wilcoxon rank sum test.
Survival time was defined as the length of time from restaging to death. Patients who were alive were censored at date of last follow-up visit.
We estimated overall survival according to the Kaplan-Meier product limit method and applied the log-rank test to compare two distributions based on genotypes or pharmacokinetics.11
All statistical analyses were performed using SAS for Windows version 8.02 (SAS Institute, Cary, NC).12 Results that yielded a P value .05 were considered statistically significant. No adjustments were made to account for number of statistical tests being conducted.
Cox proportional hazards multivariate modeling was used to identify genotypes that significantly predict overall and progression-free survival. The Spearman correlation coefficient was used to determine pairwise associations between genotypes.
RESULTS
Patient Demographics
All 85 patients were female, with a median age of 44 years. Eighty-seven percent of the population was white, while African Americans accounted for the remaining 13%. A complete response to high-dose chemotherapy was evident in 52% of the patients. Five of the complete responders (6% of the total population) were African American and 39 (46%) were white (Table 2). No patients died due to chemotherapy-induced toxicity. The median overall survival for all patients was 2.2 years (95% CI, 1.7 to 3.3 years), with a median follow-up of 8.2 years (range, 2 to 12.5 years).
Pharmacokinetics
Pharmacokinetic parameters were fully assessable in 71 patients for cyclophosphamide, 61 patients for carmustine, and 52 patients for cisplatin (Tables 2 and 3). The reasons for lack of data include insufficient samples collected on at least one of the treatment days for cyclophosphamide and carmustine, whereas some patients did not have cisplatin data due to inaccurate sample processing.
Carmustine plasma concentration time data sets were adequately described with an open, linear, one-compartment pharmacokinetic model. The median systemic exposure was 529 μg/mL x min, but values varied over a 12-fold range among patients. Ultrafiltrable plasma platinum concentrations varied over a three to four-fold range within each sample time, and the median values increased over the course of the 3-day infusion, as anticipated (Table 3). Neither carmustine nor platinum systemic exposure was associated with overall survival.
Cyclophosphamide disposition was estimated using an open, linear, two-compartment model. The systemic exposure to parent drug varied over a three-fold range and declined within each patient by approximately one half after the first dose, compared with later doses, as anticipated. Patients who achieved overall systemic exposure value below the population median (suggesting extensive metabolic activation) experienced a significantly longer overall survival than those above the median (3.8 [95% CI, 2.4 to 10.1] v 1.8 [95% CI, 1.2 to 2.5] years, respectively; P = .042; Fig 1).
Pharmacogenetics
Genotype data for potentially important polymorphisms were available in approximately 70% of patients (Table 1). All polymorphisms could not be determined in all patients due to limited sample volume or degradation of DNA. Polymorphisms were within the probability limits of Hardy-Weinberg equilibrium stratified by race, except for CYP2B6_C-1184GT (P = .009) in white patients. Given the relatively small sample size and large number of polymorphisms assayed, one polymorphism would be expected, by chance, to be in disequilibrium. The allelic frequencies of polymorphisms associated with CYP3A4 and CYP3A5 were substantially higher in African Americans compared with white patients, as anticipated (Table 1). Approximately 26% of patients had at least one polymorphic allele for either CYP3A4 or CYP3A5. Genotypes for CYP3AP1*1 were in virtually complete concordance with those of CYP3A5*1. The GSTM1 null genotype occurred in approximately one half of the white patients, but was infrequent in the African Americans (10%).
The median pharmacokinetic parameters differed by genotype of CYP3A4, CYP3A5, GSTM1, and MET1F. Differences in the systemic exposure (area under the concentration-time curve [AUC]) of cyclophosphamide became evident by dose 2 and persisted into dose 3 when patients were segregated based on the CYP3A4*1B polymorphism (P = .024 and P = .028, respectively; Table 2). Those with the polymorphic variant displayed higher AUC of parent drug, suggesting less enzyme activity. Likewise, apparent induction of cyclophosphamide metabolism (dose 1 AUC/dose 3 AUC) was reduced significantly for patients with that polymorphism (P = .031). The AUC of cyclophosphamide also seemed to be higher by the second dose in patients with the CYP3A5*1 variant, or its pseudogene, CYP3AP1*1 (P = .002 and .049, respectively; Table 2). No differences in cyclophosphamide AUC were found when patients were segregated based on ethnicity alone, despite the higher frequency of CYP3A4/5 polymorphisms in the African American patients (Table 1).
Patients with homozygous GSTM1 null genotype had a 19% lower median carmustine AUC (504 μg/mL x min) compared with patients with either one or two functional alleles (620 μg/mL x min; P = .031). Further, the clinical response to therapy was 58% in the null group compared with 44% in those with one or two functional alleles (P = .22).
Concentrations of platinum at 72 hours after administration were significantly lower in patients with polymorphisms in MET1F G-7T (P = .003; Table 3). Segregation of the cisplatin pharmacokinetic data by genotype revealed that patients with the polymorphism did not experience accumulation of platinum in the plasma ultrafiltrate (Wilcoxon rank sum test, P = .012).
Survival time was found to be associated with genotypes from the following polymorphisms: CYP3A4*1B, CYP3A5*1, CYP3AP1*1, GSTM1 null, and MET1FG-7T. Median survival of patients with the CYP3A4 polymorphism was 1.3 years (95% CI, 0.6 to 2.1) compared with 2.7 years (95% CI, 1.8 to 4.1) for patients with both copies of the common allele (P = .043; Fig 2A). Similar differences in median survival years were observed for polymorphisms in CYP3A5 (95% CIs, 0.8 to 2.2 v 2.1 to 7.4; P = .002; Fig 2B) and for polymorphisms in the promoter of MET1F (95% CIs, 0.9 to 3.2 v 1.7 to unbounded; P = .033; Fig 2C). Survival increased for patients with the null genotype of GST M1 compared with patients with one or two functional alleles. Patients with one or two wild-type alleles of GSTM1 showed a median survival of 1.5 years (95% CI, 1.2 to 2.6), whereas patients with the null genotype of GSTM1 had a median survival of 3.5 years (95% CI, 2.2 to unbounded; P = .041; Fig 2D).
Spearman correlations were conducted between the variants found to be important to pharmacokinetics of the drugs or overall survival. Significant associations were found between the GSTM1 variant and CYP3A5*1 (r = 0.28; P = .014), in addition to one between CYP3A5*1 and CYP3A4*1B (r = 0.62; P < .0001).
A Cox proportional hazards multivariate model found that patients had risk ratios for disease progression of 2.4 and 1.7 if they had CYP3A5*1 and MET 1F G-7T genotypes, respectively (P = .0078 and .076, respectively). Similarly, a multivariate evaluation for overall survival demonstrated a 2.6 and 1.9 risk ratio for death if patients had CYP3A5*1 (P = .0035) and MET1F G-7T (P = .029) genotypes, respectively.
DISCUSSION
It is logical to assume that the observed variability in clinical response to anticancer agents is due to both tumor and host factors. In the current study, we demonstrate links between the variability in systemic drug disposition (pharmacokinetics) with specific genetic variability and outcome.
Associations between the pharmacokinetics of anticancer drugs and their clinical toxicities are well documented in the literature13; however, relatively little data are available that substantiate relationships between drug pharmacokinetics and anticancer efficacy.14 We found a significant association between the disposition of high-dose cyclophosphamide and overall survival in a chemotherapy-na?ve group of patients with metastatic breast cancer. These results corroborate our previous findings with the same high-dose regimen when it was used in earlier stage disease,3 as well as those results observed by others.15
It is not clear which of the CYP450 isoenzymes are involved with in vivo activation of cyclophosphamide in humans; however, in vitro studies have primarily implicated CYP3A4/5, CYP2C9, and CYP2B6.16-18 We found that polymorphisms in CYP3A4 and CYP3A5 are associated with a reduced systemic clearance of parent cyclophosphamide and a poorer overall clinical response. These data suggest an important clinical role for the CYP3A isoenzymes in cyclophosphamide metabolic activation.
CYP3A4 and CYP3A5 are thought to be the most important CYP 450 isoenzymes pertaining to drug metabolism, due to the diversity of drugs that they act on, as well as their relative abundance in humans.19 Several polymorphisms have been identified in CYP3A4; however, controversy exists as to their functional significance.20 We found that patients with a genetic variant in the 5' promoter region (CYP3A4_A292G; CYP3A4*1B allele) had a higher AUC and shorter overall survival, which could result from a reduced metabolic activation of cyclophosphamide. This polymorphism has been previously associated with a higher clinical stage/grade of prostate cancer and a higher risk of treatment-related leukemia.21,22 However, altered constitutive protein expression in subjects with the variant has not been shown.23,24 Our data support this result since we did not see differences in the pharmacokinetic parameters until the metabolic induction process was evident on the second and third doses per day of cyclophosphamide. It will thus be important to expand this research to investigate polymorphisms in other genes that control transcription of CYP3A in such a setting.25
Multiple drugs used in clinical practice are inhibitors of CYP3A, including some antihistamines, antibiotics, antivirals, and importantly, antiemetics.26 Our data would suggest that concomitant use of such drugs with cyclophosphamide has potential to reduce its clinical effectiveness. We recently reported that patients who receive this chemotherapy regimen display a 43% induction in CYP3A metabolic activity after three doses of cyclophosphamide when compared with the level measured just before therapy,27 consistent with data from studies conducted in vitro.28 Thus, patients who receive cyclophosphamide concurrently with other drugs metabolized by CYP3A should also be monitored closely for accelerated clearance of the latter agents.
We found a similar association of a variant in CYP3A5 (CYP3A5_A22893G; CYP3A5*1), with both survival and cyclophosphamide pharmacokinetics. It has been shown that this polymorphism is in linkage disequilibrium with a pseudogene found upstream of the CYP3A5 promoter (CYP3A5_5UTR_A-44G; CYP3AP1*1), which is in agreement with our data.29,30 A recently published study by Floyd et al using midazolam as a probe and rifampin as a metabolic inducer in 57 healthy subjects suggests that the CYP3A5*1 polymorphism is particularly important in the setting of CYP3A enzyme induction.31 This study and ours both show that CYP3A is less inducible for individuals with CYP3A5*1 or CYP3A4*1B. One possible explanation for the reduced activation of cyclophosphamide in women carrying the CYP3A5*1 allele is that they fail to induce total CYP3A levels to the same extent as those homozygous for *3, since the 3A5 gene does not appear to be inducible.32,33
The results of this study pertaining to CYP3A have potentially dramatic implications for the pharmacogenetic evaluations of many drugs used in clinical care today. This isoenzyme is known to be involved in the metabolism of approximately one half of US Food and Drug Administration–approved drugs, spanning all therapeutic classes.34 Some anticancer drugs that are metabolized via CYP3A include: docetaxel, paclitaxel, irinotecan, tamoxifen, and imatinib. In addition, it is an important enzyme involved in the metabolic conversion of endogenous estrogens to potential carcinogens.35 The variant CYP3A alleles are relatively uncommon in whites but occur in approximately 50% of African Americans.36 These variants may be partially responsible for decreased cyclophosphamide efficacy previously observed in African Americans.37
Cytosolic glutathione-S-transferase (GST) enzymes catalyze reactions between lipophilic compounds with electrophilic centers and glutathione. Their primary role is believed to be in facilitating the detoxification of reactive molecules such as environmental carcinogens and xenobiotics (eg, cisplatin). Some polymorphisms in GSTs result in complete deletion of the gene, and thus no functional protein. These occur in up to 50% of white subjects and approximately 28% of African Americans. The GST null genotype occurred in only 10% of the patients on our study; however, this may be due to our relatively small number of patients in this ethic group. We found that patients with the GSTM1 null genotype had a prolonged overall survival. This is consistent with data from other studies that have demonstrated high expression of some GSTs resulted in lower response to chemotherapy,38-40 whereas patients with null alleles may exhibit higher degrees of toxicity.41
Metallothioneins are thought to play a role in neutralization of toxic heavy metals and are highly inducible (both intracellularly and in plasma) by a variety of xenobiotics, including cisplatin.42,43 Upregulation of metallothionein in cancer cells can render them resistant to cisplatin.44,45 The size of metallothionein (molecular weight, 6,000 to 7,000 Da) when bound to cisplatin is well below the molecular weight cutoff of the ultrafiltration apparatus used to prepare our clinical samples. Thus, we suspect that cisplatin induces plasma metallothionein, and this accounts for the accumulation of the ultrafiltrable platinum over the 72-hour infusion seen here and previously.3
Our data suggest that evaluation of independent patient pharmacogenetics may provide a logistically feasible mechanism for tailoring the dosage of chemotherapy in hopes of optimizing efficacy and minimizing toxicity. The confounding factor of higher frequencies of some polymorphisms based on ethnicity, suggests that confirmatory studies should be conducted to attempt establishment of similar relationships within ethnic groups. Such would then allow for formal evaluation of dose individualization based on a patient's genetic profile. When coupled with the rapidly evolving field of tumor genomics,46 these strategies could provide a substantial impact on improving the therapeutic index of anticancer regimens.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported in part by NIH 5-P01-CA47741 and NIH P50-CA68438; Specialized Program in Research Excellence in Breast Cancer, NIH 5-P30-CA14236-26, NIH UO1 CA84955; and the Mylan Chair of Pharmacology.
Presented in part at the American Association for Cancer Research Annual Meetings, New Orleans, LA, 2001 (abstr 1435) and Orlando, FL, 2004 (abstr 2106); however, no other aspects of the data have been published.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Evans WE, Hon YY, Bomgaars L, et al: Preponderance of thiopurine S-methyltransferase deficiency and heterozygosity among patients intolerant to mercaptopurine or azathioprine. J Clin Oncol 19:2293-2301, 2001
Nagasubramanian R, Innocenti F, Ratain MJ: Pharmacogenetics in cancer treatment. Annu Rev Med 54:437-452, 2003
Petros WP, Broadwater G, Berry D, et al: Association of high-dose cyclophosphamide, cisplatin and carmustine pharmacokinetics with response, toxicity and dosing weight in patients with primary breast cancer. Clin Cancer Res 8:698-705, 2002
Rizzieri DA, Vredenburgh JJ, Jones R, et al: Prognostic and predictive factors for patients with metastatic breast cancer undergoing aggressive induction therapy followed by high-dose chemotherapy with autologous stem-cell support. J Clin Oncol 17:3064-3074, 1999
Jones RB, Shpall EJ, Shogan J, et al: The Duke AFM Program: Intensive induction chemotherapy for metastatic breast cancer. Cancer 66:431-436, 1990
Peters WP, Ross M, Vredenburgh JJ, et al: High-dose chemotherapy and autologous bone marrow support as consolidation after standard-dose adjuvant therapy for high-risk primary breast cancer. J Clin Oncol 11:1132-1143, 1993
Carter DL, Marks LB, Bean JM, et al: Impact of consolidation radiotherapy in patients with advanced breast cancer treated with high-dose chemotherapy and autologous bone marrow rescue. J Clin Oncol 17:887-893, 1999
Gilbert CJ, Petros WP, Vredenburgh J, et al: Pharmacokinetic interaction between ondansetron and cyclophosphamide during high-dose chemotherapy for the treatment of breast cancer. Cancer Chemother Pharmacol 42:497-503, 1998
Abdel-Rahman SZ, el-Zein RA, Anwar WA, et al: A multiplex PCR procedure for polymorphic analysis of GSTM1 and GSTT1 genes in population studies. Cancer Lett 107:229-233, 1996
Livak KJ: Allelic discrimination using fluorogenic probes and the 5[prime] nuclease assay. Genet Anal 14:143-149, 1999
Cox DR: Regression models and life-tables. J R Stat Soc B 34:187-220, 1972
SAS/Stat user guide, version 8. SAS Institute Inc. SAS Campus Drive, Cary, NC
Ratain MJ, Plunkett W: Pharmacology, in Kufe DW, Pollock RE, Weichselbaum RR, et al (eds): Cancer Medicine (ed 6). London, UK, BC Decker, 2003
Evans WE, Relling MV, Rodman JH, et al: Conventional compared with individualized chemotherapy for childhood acute lymphoblastic leukemia. N Engl J Med 338:499-505, 1998
Ayash LJ, Wright JE, Tretyakov O, et al: Cyclophosphamide pharmacokinetics: Correlation with cardiac toxicity and tumor response. J Clin Oncol 10:995-1000, 1992
Ren S, Yang J-S, Kalhorn TF, et al: Oxidation of cyclophosphamide to 4-hydroxycyclophosphamide and deschloroethylcyclophosphamide in human liver microsomes. Cancer Res 57:4229-4235, 1997
Roy P, Yu LJ, Crespi CL, et al: Development of a substrate-activity based approach to identify the major human liver P-450 catalysts of cyclophosphamide and ifosfamide activation based on cDNA-expressed activities and liver microsomal P-450 profiles. Drug Metab Dispos 27:655-666, 1999
Xie H-J, Yasar U, Lundgren S, et al: Role of polymorphic human CYP2B6 in cyclophosphamide bioactivation. Pharmacogenomics J 3:53-61, 2003
Gellner K, Eiselt R, Hustert E, et al: Genomic organization of the human CYP3A locus: Identification of a new, inducible CYP3A gene. Pharmacogenetics 11:111-121, 2001
Flockhart DA, Rae JM: Cytochrome P450 3A pharmacogenetics: The road that needs traveled. Pharmacogenomics J 3:3-7, 2003
Rebbeck TR, Jaffe JM, Walker AH, et al: Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4. J Natl Cancer Inst 90:1225-1229, 1998
Felix CA, Walker AH, Lange BJ, et al: Association of CYP3A4 genotype with treatment-related leukemia. Proc Natl Acad Sci U S A 95:13176-13181, 1998
Ball SE, Scatina J, Kao J, et al: Population distribution and effects on drug metabolism of a genetic variant in the 5[prime] promoter region of CYP3A4. Clin Pharmacol Ther 66:288-294, 1999
Westlind A, Lofberg L, Tindberg N, et al: Interindividual differences in hepatic expression of CYP3A4: relationship to genetic polymorphism in the 5[prime]-upstream regulatory region. Biochem Biophys Res Commun 259:201-205, 1999
Zhang J, Kuehl P, Green ED, et al: The human pregnane X receptor: Genomic structure and identification and functional characterization of natural allelic variants. Pharmacogenetics 11:555-572, 2001
Majumdar AK, McCrea JB, Panebianco DL, et al: Effects of aprepitant on cytochrome P450 3A4 activity using midazolam as a probe. Clin Pharmacol Ther 74:150-156, 2003
Petros W, Vredenburgh J, Fehdrau R, et al: Evaluation of CYP3A4 metabolic phenotype and cyclophosphamide pharmacokinetics in patients receiving high-dose therapy. Proc Am Assoc Cancer Res 43:274, 2002 (abstr 318)
Lindley C, Hamilton G, McCune JS, et al: The effect of cyclophosphamide with and without dexamethasone on cytochrome P450 3A4 and 2B6 in human hepatocytes. Drug Metab Dispos 30:814-821, 2002
Paulussen A, Lavrijsen K, Bohets H, et al: Two linked mutations in transcriptional regulatory elements of the CYP3A5 gene constitute the major genetic determinant of polymorphic activity in humans. Pharmacogenetics 10:415-424, 2000
Finta C, Zaphiropoulos PG: The human cytochrome P450 3A locus: Gene evolution by capture of downstream exons. Gene 260:13-23, 2000
Floyd MD, Gervasini G, Masica AL, et al: Genotype-phenotype associations for common CYP3A4 and CYP3A5 variants in the basal and induced metabolism of midazolam in European- and African-American men and women. Pharmacogenetics 13:595-606, 2003
Kuehl P, Zhang J, Lin Y, et al: Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet 27:383-391, 2001
Goh B-C, Lee S-C, Wang L-Z, et al: Explaining interindividual variability of docetaxel pharmacokinetics and pharmacodynamics in Asians through phenotyping and genotyping strategies. J Clin Oncol 20:3683-3690, 2002
de Wildt SN, Kearns GL, Leeder JS, et al: Cytochrome P450 3A: Ontogeny and drug disposition. Clin Pharmacokinet 37:485-505, 1999
Huang Z, Guengerich FP, Kaminsky LS: 16Alpha-hydroxylation of estrone by human cytochrome P4503A4/5. Carcinogenesis 19:867-872, 1998
Walker AH, Jaffe JM, Gunasegaram S, et al: Characterization of an allelic variant in the nifedipine-specific element of CYP3A4: ethnic distribution and implications for prostate cancer risk. Hum Mutat 12:289, 1998
Dooley MA, Hogan S, Jennette C, et al: Cyclophosphamide therapy for lupus nephritis: Poor renal survival in black Americans—Glomerular Disease Collaborative Network. Kidney Int 51:1188-1195, 1997
Ambrosone CB, Sweeney C, Coles BF, et al: Polymorphisms in glutathione S-transferases (GSTM1 and GSTT1) and survival after treatment for breast cancer. Cancer Res 61:7130-7135, 2001
Cheng X, Kigawa J, Minagawa Y, et al: Glutathione S-transferase-pi expression and glutathione concentration in ovarian carcinoma before and after chemotherapy. Cancer 79:521-527, 1997
Bai F, Nakanishi Y, Kawasaki M, et al: Immunohistochemical expression of glutathione S-transferase-Pi can predict chemotherapy response in patients with nonsmall cell lung carcinoma. Cancer 78:416-421, 1996
Peters U, Preisler-Adams S, Hebeisen A, et al: Glutathione S-transferase genetic polymorphisms and individual sensitivity to the ototoxic effect of cisplatin. Anticancer Drugs 11:639-643, 2000
Vandier D, Calvez V, Massade L, et al: Transactivation of the metallothionein promoter in cisplatin-resistant cancer cells: A specific gene therapy strategy. J Natl Cancer Inst 92:642-647, 2000
Tohyama C, Shaikh ZA: Metallothionein in plasma and urine of cadmium-exposed rats determined by a single-antibody radioimmunoassay. Fundam Appl Toxicol 1:1-7, 1981
Kelley SL, Basu A, Teicher BA, et al: Overexpression of metallothionein confers resistance to anticancer drugs. Science 241:1813-1815, 1988
Toyoda H, Mizushima T, Satoh M, et al: HeLa cell transformants overproducing mouse metallothionein show in vivo resistance to cis-platinum in nude mice. Jpn J Cancer Res 91:91-98, 2000
Evans WE, McLeod HL: Pharmacogenomics: Drug disposition, drug targets, and side effects. N Engl J Med 348:538-549, 2003(William P. Petros, Penelo)