- © 2008 by American Society of Clinical Oncology
Variation in Bleomycin Hydrolase Gene Is Associated With Reduced Survival After Chemotherapy for Testicular Germ Cell Cancer
- Esther C. de Haas,
- Nynke Zwart,
- Coby Meijer,
- Janine Nuver,
- H. Marike Boezen,
- Albert J.H. Suurmeijer,
- Harald J. Hoekstra,
- Gerrit van der Steege,
- Dirk Th. Sleijfer and
- Jourik A. Gietema
- From the Departments of Medical Oncology, Epidemiology, Pathology, Surgical Oncology, Medical Biology, Medical Genetics; and the Genotyping Facility, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
- Corresponding author: Jourik A. Gietema, MD, PhD, Department of Medical Oncology, University Medical Center Groningen, University of Groningen, PO Box 30001, 9700 RB Groningen, the Netherlands; e-mail: j.a.gietema{at}int.umcg.nl
Abstract
Purpose Response to chemotherapy may be determined by gene polymorphisms involved in metabolism of cytotoxic drugs. A plausible candidate is the gene for bleomycin hydrolase (BLMH), an enzyme that inactivates bleomycin, an essential component of chemotherapy regimens for disseminated testicular germ-cell cancer (TC). We investigated whether the single nucleotide polymorphism (SNP) A1450G of the BLMH gene (rs1050565) is associated with survival.
Patients and Methods Data were collected on survival and BLMH genotype of 304 patients with TC treated with bleomycin-containing chemotherapy at the University Medical Center Groningen, the Netherlands, between 1977 and 2003. Survival according to genotype was analyzed using Kaplan-Meier curves with log-rank testing and Cox regression analysis with adjustment for confounders.
Results BLMH gene SNP A1450G has a significant effect on TC-related survival (log-rank P = .001). The homozygous variant (G/G) genotype (n = 31) is associated with decreased TC related survival compared with the heterozygous variant (A/G; n = 133) and the wild-type (A/A; n = 140). With Cox regression the G/G genotype proves to be an unfavorable prognostic factor, in addition to the commonly used International Germ Cell Consensus Classification prognosis group, with a hazard ratio of 4.97 (95% CI, 2.17 to 11.39) for TC-related death. Furthermore, the G/G genotype shows a higher prevalence of early relapses.
Conclusion The homozygous variant G/G of BLMH gene SNP A1450G is associated with reduced survival and higher prevalence of early relapses in TC patients treated with bleomycin-containing chemotherapy. This association is hypothesis generating and may eventually be of value for risk classification and selection for alternative treatment strategies in patients with disseminated TC.
INTRODUCTION
Testicular germ-cell cancer (TC), the most common malignancy among young adult men, has shown an increasing worldwide incidence over the past 30 years.1 In contrast, the mortality rate of TC, consisting of approximately 95% of testicular germ-cell tumors (seminomatous and nonseminomatous), has decreased. Long-term survival in disseminated TC exceeds 80%.2,3 This improvement in survival is related to the introduction of cisplatin for the treatment of disseminated TC in the 1970s and the further development of cisplatin-based regimens.4-6
To enable risk-based decisions about treatment of disseminated TC, the International Germ Cell Consensus Classification (IGCCC) has been developed. The IGCCC identifies three prognosis groups with good, intermediate, and poor prognosis.7 With cisplatin-containing chemotherapy, 5-year survival rate is 91% for the good prognosis group, 79% for the intermediate prognosis group, and 48% for the poor prognosis group.7
As survival of TC has improved, there is increasing interest for reducing toxicity of the current standard chemotherapy regimen, consisting of bleomycin, etoposide, and cisplatin (BEP). The last major change in this area is the confirmation that three courses of BEP are as effective as four courses in good prognosis germ-cell cancer.8 In addition to reduction of toxicity, further improvement of survival, especially in poor prognosis patients, forms a challenge.
It seems conceivable that toxicity, as well as tumor response, is influenced by polymorphisms of genes involved in metabolism or target pathways of cytotoxic drugs. The discovery of an association with such genetic polymorphisms may contribute to tailoring of chemotherapy to reduce toxicity and improve survival.9,10
A candidate gene might be the gene for bleomycin hydrolase (BLMH), an enzyme that can inactivate bleomycin. Bleomycin has proven to be an essential component of the cisplatin-based chemotherapy regimens, cisplatin, vinblastin, and bleomycin (PVB) and BEP, used in the treatment of TC.11-13 However, the use of bleomycin is limited because of bleomycin-induced pneumonitis, a complication occurring in approximately 10% of patients treated with bleomycin and fatal in approximately 10% of the instances.14-17
In mice, lack of enzymatic activity of BLMH is associated with increased bleomycin-induced pulmonary toxicity.18,19 In contrast, increased enzymatic activity has been found to be associated with human tumor resistance to bleomycin20 and elevated expression of BLMH has been observed in human tumor cell lines.21,22
In the BLMH gene, an A1450G polymorphic site (rs1050565) has been identified. This single nucleotide polymorphism (SNP) leads to the presence of either isoleucine or valine as amino acid residue 443.21 The SNP is located in the C-terminal region that is thought to be involved in controlling the enzymatic activity.23,24 Although the effect of this SNP on the enzymatic activity of BLMH and the consequences for the metabolism of bleomycin are unknown, the SNP appears to influence the level of bleomycin-induced DNA damage.25
We hypothesize that the BLMH gene SNP A1450G influences BLMH activity and may, consequently, by an altered metabolism of bleomycin, be associated with differences in TC-related survival. Therefore, we investigated whether the BLMH gene SNP A1450G is associated with differences in survival in patients with nonseminomatous TC treated with bleomycin-containing chemotherapy.
PATIENTS AND METHODS
Patients
Patients with nonseminomatous TC who have been treated with bleomycin- and platinum-containing chemotherapy at the University Medical Center Groningen, between January 1977 and January 2003, were eligible for this study. Patients with a primary mediastinal nonseminomatous tumor were excluded. In total, this cohort comprised 372 patients.
For these patients, data were collected on baseline characteristics (age at start of treatment, creatinine clearance according to Cockcroft-Gault formula), clinical stage (Royal Marsden Hospital staging), IGCCC prognosis group, combination of chemotherapy received, and disease outcome.
DNA Collection and Genotyping
After informed consent, genomic DNA was isolated from peripheral blood samples collected into EDTA tubes at the general practitioner's or at the outpatient clinic. For deceased patients, genomic DNA was isolated preferentially from routinely stored serum and if no serum was present from paraffin-embedded histological material. In case no healthy tissue was available, tumor samples were used. DNA isolation was performed as previously described.17
BLMH genotype was determined by polymerase chain reaction (PCR) and a restriction fragment length polymorphism technique, according to the previously described procedure.17 In summary, a reverse primer and a forward mismatch primer with a universal M13 primer extension 5′-CGA CGT TGT AAA ACG ACG GCC AGT AGT GCT GTG TTA GAG CAG GAA CCA ATT-3′ (Invitrogen, Merelbeke, Belgium) were used for amplification to create a 150 basepair (bp) DNA fragment which contains in case of A1450G transition a MunI (Roche Diagnostics GmbH, Mannheim, Germany) restriction site. After digestion with MunI and electrophoresis on a 2.5% agarose gel, the wild-type (A/A) genotype was identified by a 150 bp fragment, the heterozygous variant (A/G) genotype by 150, 111, and 39 bp fragments, and the homozygous variant (G/G) genotype by 111 and 39 bp fragments. When the BLMH genotype could not be accurately determined by this procedure, the DNA sample was sequenced.
Survival Measurements
Last follow-up date, vital status at last follow-up date, and, if applicable, date and cause of death were collected from the medical record or available via the general practitioner's files. Survival time was calculated from start of chemotherapy to death or last follow-up. Overall survival was distinguished from TC-related survival, defined as death due to TC and not to chemotherapy-induced toxicity.
With respect to disease, outcome distinction was made between refractory disease, defined as absence of normalization of tumor markers lactate dehydrogenase, α-fetoprotein, and β-human chorionic gonadotropin, or renewed elevation of tumor marker levels within 4 weeks after completion of chemotherapy, early relapses (occurring within 2 years after start of treatment after an initial complete response), and late relapses (occurring > 2 years after start of treatment).
Statistical Methods
Distribution of the genotypes was tested for Hardy-Weinberg equilibrium. The three BLMH genotype groups were compared for patient characteristics using the χ2 test for categoric variables and the Kruskal-Wallis test for continuous variables. Differences in overall and TC-related survival between the genotype groups were analyzed using Kaplan-Meier curves and tested formally using the log-rank test. To analyze the independent effects of BLMH genotype on survival, Cox proportional hazards regression was performed with adjustment for potential confounders. All statistical analyses were performed with SPSS for Windows version 14.0 (SPSS Inc, Chicago, IL).
RESULTS
Patient Characteristics and BLMH Genotype
For 321 patients of the total cohort of 372 patients, DNA was available for genotyping. BLMH genotype could be accurately assessed in 304 of the 321 DNA samples (94.7%). The analyzable group of 304 patients had a median age of 28 years at start of chemotherapy (range, 16 to 64 years) and a median survival time of 10 years (range, 0 to 27 years). Age at start of treatment, survival time, and prognosis group distribution (according to IGCCC) do not differ from the group of 68 nonanalyzable patients (data not shown).
Of the 304 analyzed patients, genotype distribution was as follows: 140 A/A (46%), 133 A/G (44%), and 31 G/G (10%). Allele frequencies are in Hardy-Weinberg equilibrium. Baseline characteristics and chemotherapy data for the three genotype groups are presented in Table 1. The genotypes do not differ significantly with respect to age, creatinine clearance, initial prognosis (according to IGCCC), and received cumulative dose of bleomycin and cisplatin.
Patient Baseline Characteristics and Chemotherapy Data
BLMH Genotype and Survival
Overall, 59 (19.4%) of the 304 evaluated patients have died. With 41 patients (13.5%) dying of TC, it was the most common cause of death. Other most common causes of death were cardiovascular disease (n = 5; four fatal myocardial infarctions and one ruptured abdominal aortic aneurysm), bleomycin-induced pulmonary toxicity (n = 3), and second non-TC malignancy (n = 3; Table 2).
Number of Deceased Patients With Causes of Death According to BLMH Genotype
Kaplan-Meier curves for overall survival and TC-related survival according to BLMH genotype are shown in Figure 1. Log-rank testing shows a significant difference for overall survival according to BLMH genotype (P = .006). Overall survival of the homozygous variant (G/G) genotype group is decreased compared with the wild-type (A/A) and heterozygous variant (A/G) genotype groups (Fig 1A).
Kaplan-Meier curves for (A) overall survival, (B) testicular cancer (TC)–related survival, and (C) progression-free survival according to bleomycin hydrolase (BLMH) genotype. A/A, wild-type; A/G, heterozygous variant; G/G, homozygous variant.
Looking specifically at TC-related survival, the homozygous variant (G/G) genotype group shows worse TC-related survival than the wild-type (A/A) and heterozygous variant (A/G) genotype groups, mainly within the first years after treatment (Fig 1B). TC-related survival for the heterozygous variant (A/G) genotype seems merely to follow the wild-type (A/A) group, suggesting a recessive instead of a dominant effect of the analyzed SNP. Because still curative options exist for refractory disease and relapses, progression free survival (date of progression for refractory patients [reflected by an increase in tumor marker levels, occurrence of new lesions, or growing lesions other than growing teratoma] or date of first relapse for initially responding patients) was also analyzed. The Kaplan-Meier curve for progression-free survival showed a similar pattern as the curve for TC-related survival with reduced progression-free survival for the G/G genotype (log-rank test P = .013; Fig 1C).
Because of the observed difference in TC-related survival and progression-free survival, the three genotypes were compared for disease outcome after completion of chemotherapy (Table 3). The G/G genotype group shows a higher prevalence of early TC relapses than the A/G and A/A group (P = .019). The prevalence of refractory disease also seems to be higher in the G/G genotype group. These results suggest a worse tumor response to chemotherapy. The prevalence of late relapses does not differ.
Disease Outcome and Median Duration of Follow-Up to Diagnosis of Testicular Cancer Relapse After Completion of Chemotherapy (N = 300)
Cox Regression Model
In order to analyze the independent effect of BLMH genotype on TC-related survival, a multivariate Cox regression analysis was performed with adjustment for the following disease- and chemotherapy-related confounders: age at start of chemotherapy, stage, prognosis according to IGCCC, separate levels of the tumor markers lactate dehydrogenase, α-fetoprotein, and β-human chorionic gonadotropin, cumulative dose of the cytotoxic agents bleomycin and cisplatin, and creatinine clearance. Prognosis group according to IGCCC and BLMH genotype appeared to be independent prognostic factors for TC-related survival (Table 4). Adjusted for IGCCC, patients with the G/G genotype showed a significantly increased risk for TC-related death with a hazard ratio (HR) of 4.97 (95% CI, 2.17 to 11.39; P = .000) compared with the A/A group. The presence of the A/G genotype does not affect the risk for TC-related death when adjusted for prognosis group. Based on this Cox regression model with adjustment for prognosis group, predicted TC-related survival for each BLMH genotype group is shown in Figure 2.
Predicted testicular cancer (TC)–related survival according to bleomycin hydrolase (BLMH) genotype after adjustment for prognosis group (International Germ Cell Consensus Classification). For the homozygous variant (G/G) genotype the hazard ratio for TC-related death is 4.97 (95% CI, 2.17 to 11.39). A/A, wild-type; A/G, heterozygous variant.
Cox Regression Hazard Analysis for Risk of Testicular Cancer–Related Death According to BLMH Genotype (adjusted for prognosis group according to IGCCC)
With multivariate Cox regression analysis, a comparable relationship was found between BLMH genotype and progression-free survival (data not shown). Adjusted for IGCCC, patients with the G/G genotype showed an increased risk for disease progression (either in refractory disease or as a relapse) with a HR of 3.10 (95% CI, 1.50 to 6.44; P = .002) compared with the A/A group.
DISCUSSION
Bleomycin is an essential component of the platinum-based chemotherapy regimens for disseminated testicular germ-cell cancer. Because metabolic inactivation of bleomycin by BLMH appears to be associated with human tumor resistance to bleomycin20, we investigated whether the BLMH gene SNP A1450G influences survival in patients with nonseminomatous TC treated with bleomycin-containing chemotherapy. The presented results suggest that presence of the G/G genotype is associated with decreased overall survival, compared with the A/G and A/A genotypes. This decreased overall survival is caused by a higher prevalence of deaths due to TC in the homozygous variant G/G genotype group.
Cox regression analysis shows that, after correction for the commonly used IGCCC prognosis group, the G/G genotype is an independent, unfavorable prognostic factor with an approximately five times increased risk (HR, 4.97; 95% CI, 2.17 to 11.39) for TC-related death.
Furthermore, the G/G genotype is associated with reduced progression-free survival and a higher prevalence of early relapses (relapses occurring within 2 years), and appears also, although not significantly, to be associated with a higher prevalence of refractory disease. This suggests that tumor response to chemotherapy in this group is worse compared with the A/G and A/A genotype group.
Patients in the G/G genotype group have received a comparable cumulative dose of bleomycin during first-line chemotherapy. With the exception of one patient in the G/G genotype group, none of the patients with refractory disease or relapse received additional bleomycin-containing chemotherapy. In addition, the genotype groups do not differ with respect to renal function (as estimated by Cockcroft-Gault formula), suggesting that they have been exposed to a similar amount of bleomycin.
The aforementioned data show that the observed difference in TC-related survival and disease outcome cannot be explained by differences in initial IGCCC prognosis or differences in cumulative bleomycin dose. Because the effect of the analyzed SNP on bleomycin pharmacokinetics is unknown, it is unclear whether the associations with disease outcome and TC-related survival are due to an altered metabolism of bleomycin. Based on its location, the BLMH gene SNP A1450G is a plausible candidate for influencing the enzymatic activity of BLMH. The SNP is located in the C-terminal region that intrudes the active-site cleft of BLMH and might influence by its position the substrate specificity as observed for the yeast BLMH homolog Gal6.26,27 Although in the variant genotypes substitution of isoleucine by valine would unlikely lead to change in the protein conformation of BLMH, it has been suggested that this region forms a likely candidate for interaction with a protein regulating the positioning of the C-terminal arm.26
Moreover, data have been published that suggest that this SNP is functional. Tuimala et al25 found an association between the BLMH gene SNP A1450G and the level of bleomycin-induced DNA damage, that appears to be lower in smoking individuals with the A/G and G/G genotype.
Aforementioned associations suggest that presence of a variant BLMH genotype may lead to a change in BLMH activity. Theoretically, increased BLMH activity would lead to increased inactivation of bleomycin, contributing to bleomycin resistance. In this study the prevalence of refractory disease appears, although not significant, to be higher in the G/G genotype group. However, the analysis of an association between BLMH genotype and refractory disease may be limited by the relatively small number of patients with initial refractory disease.
A contribution of the G/G BLMH genotype to bleomycin resistance would also suggest a protective effect of this genotype against bleomycin-induced toxicity to healthy tissues. In an earlier report of our group on the effect of BLMH genotype on bleomycin-induced pulmonary toxicity, in a largely overlapping cohort, we did not find differences in the development of bleomycin-induced pulmonary toxicity according to BLMH genotype.17 A possible explanation for this lack of relationship between BLMH genotype and bleomycin-induced pulmonary toxicity is that other factors than BLMH genotype may play a more important role in the development of pulmonary toxicity.15
An alternative explanation for the observed difference in disease outcome and TC-related survival is the involvement of BLMH in intracellular response pathways to bleomycin and/or to other cytotoxic agents, although other substrates of BLMH, besides bleomycin, homocysteine28 and probably amyloid precursor protein in Alzheimer's disease29,30, are currently unknown. BLMH is supposed to have a conserved cellular function, because of its wide distribution in nature and its importance for neonatal survival, as shown in BLMH-knockout mice.19,21,26 It has been found to have a structure resembling the proteasome31 and is able of binding ubiquitin-conjugating enzyme 9, suggesting a role in intracellular protein processing and degradation.32 Besides, BLMH binds ribosomal proteins and may consequently have a protective role against cytotoxic proteins that bind to RNA.33
In addition to the involvement of BLMH in intracellular response pathways to cytotoxic agents, a linkage disequilibrium to a yet unknown gene that is located near the locus of the BLMH gene (chromosome 17q11.1 to q11.2) and that may influence TC-related survival, cannot be excluded since the present study is an association study.
Nevertheless, our results suggest that presence of the G/G genotype of the BLMH gene SNP A1450G is an independent, unfavorable prognostic factor for survival in patients with TC treated with chemotherapy. Furthermore, the homozygous variant seems to be associated with worse tumor response to chemotherapy. To our knowledge, this is the first study describing an unfavorable influence of the G/G of the BLMH gene SNP A1450G on survival in patients with TC and as a result confirmation in other cohorts of patients with is needed.
The results of this study are hypothesis generating and not yet applicable to clinical practice. Studies in patients with other forms of cancer treated with bleomycin-containing chemotherapy or studies in patients with TC not treated with bleomycin may give insight into whether the influence of BLMH genotype on survival is specific for TC and/or treatment with bleomycin. In case presence of the G/G of the BLMH gene SNP A1450G appears to be associated with decreased bleomycin efficacy, substitution of bleomycin by ifosfamide may be warranted because no differences in response rate and survival have been observed.34
In conclusion, although it is currently unknown what the underlying mechanism is of the worse TC-related survival in patients carrying the G/G of the BLMH gene SNP A1450G, confirmation of the observed association may have consequences for risk classification in patients with disseminated TC and may be of use to select patients who will benefit from alternative treatment strategies.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Esther C. de Haas, H. Marike Boezen, Gerrit van der Steege, Jourik A. Gietema
Financial support: Jourik A. Gietema
Provision of study materials or patients: Coby Meijer, Albert J.H. Suurmeijer, Harald J. Hoekstra
Collection and assembly of data: Esther C. de Haas, Nynke Zwart, Coby Meijer, Janine Nuver, Gerrit van der Steege
Data analysis and interpretation: Esther C. de Haas, Coby Meijer, H. Marike Boezen, Jourik A. Gietema
Manuscript writing: Esther C. de Haas, Coby Meijer, H. Marike Boezen, Harald J. Hoekstra, Dirk Th. Sleijfer, Jourik A. Gietema
Final approval of manuscript: Esther C. de Haas, Nynke Zwart, Coby Meijer, Janine Nuver, H. Marike Boezen, Albert J.H. Suurmeijer, Harald J. Hoekstra, Gerrit van der Steege, Dirk Th. Sleijfer, Jourik A. Gietema
Glossary Terms
- SNP (single nucleotide polymorphism):
- Genetic polymorphisms are natural variations in the genomic DNA sequence present in greater than 1% of the population, with SNP representing DNA variations in a single nucleotide. SNPs are being widely used to better understand disease processes, thereby paving the way for genetic-based diagnostics and therapeutics.
- PCR (polymerase chain reaction):
- PCR is a method that allows logarithmic amplification of short DNA sequences within a longer DNA molecule.
- Hardy-Weinberg equilibrium:
- A state in which genotype frequencies and ratios remain constant from generation to generation and in which genotype frequencies are a product of allele frequencies. A randomly mating population tends toward a Hardy-Weinberg equilibrium state if there are no mutations, migrations, or environmental factors favoring particular genotypes.
- Cox proportional hazards regression model:
- The Cox proportional hazards regression model is a statistical model for regression analysis of censored survival data. It examines the relationship of censored survival distribution to one or more covariates. It produces a baseline survival curve, covariate coefficient estimates with their standard errors, risk ratios, 95% CIs, and significance levels.
- Linkage disequilibrium:
- Nonrandom association of linked genes. This is the tendency of the alleles of two separate but already linked loci to be found together more frequently than would be expected by chance alone.
Footnotes
-
Supported by Grants No. CVZ 01-211 from the Health Care Insurance Board/Dutch Association of University Medical Centers, the Netherlands, and RUG2000-2177 from the Dutch Cancer Society, the Netherlands.
-
Presented in part as an oral presentation at the 43rd Annual Meeting of the American Association of Clinical Oncology, Chicago, IL, June 1-5, 2007.
-
Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org.
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received September 7, 2007.
- Accepted December 6, 2007.
