Most cancers can be attributed to environmental factors that act in conjunction with both genetic and acquired susceptibility.1 It has been suggested that individuals possessing a modified ability to metabolize carcinogens are at increased risk for cancer.1 In fact, germ-line polymorphisms of genes encoding carcinogen-metabolizing enzymes, namely, phase I cytochromes P-450 (CYPs) and phase II glutathione S-transferases (GSTs), have been shown to influence the risk of a variety of disorders thought to be caused by environmental exposure to toxic agents, including Parkinson disease2 and cancers of the gastrointestinal tract, skin, bladder, cervix, and lung.1,3 In contrast to these solid tumors, the etiologic role of chemical carcinogens is less obvious for many hematologic malignancies. Nevertheless, Krajinovic et al4 reported theGSTM1 null and CYP1A1*2A genotypes to be risk factors for childhood acute lymphoblastic leukemia (ALL), and myelodysplastic syndromes (MDSs) seem to be associated with theGSTT1 null genotype.5
For chronic myelogenous leukemia (CML), a malignancy of the hematopoietic stem cell characterized by the Philadelphia chromosome and the hybrid BCR-ABL fusion gene, epidemiologic studies have failed to find any reproducible, significant association with either genetic or environmental factors except for ionizing radiation and benzene. In order to find new evidence for the hypothesis that CML is a combined result of environmental exposure and genetic susceptibility, we used the polymerase chain reaction (PCR)–based genotyping approach described by Krajinovic et al4 to examine the relationship between BCR-ABL+ CML and genetic polymorphisms in the CYP1A1, GSTM1, andGSTT1 gene loci.
Genomic DNA of 141 BCR-ABL+ CML cases (78 males, 63 females) was randomly selected from a large institutional DNA bank where DNA of German CML Study Group patients is collected. Diagnosis ofBCR-ABL+ CML was confirmed according to the criteria of the German CML Study Group. Informed consent was obtained from all participating individuals. For a general population control group, peripheral blood samples of 150 randomly selected, anonymous, healthy blood donors (97 males, 53 females) were collected. After isolation of purified genomic DNA, PCR-based genotyping was performed as described.4
Whereas no significant differences between cases and controls were found in the frequencies of heterozygous or homozygous presence of the mutant alleles CYP1A1*2B and CYP1A1*4 and of homozygous deletions of GSTM1 and GSTT1, the mutant allele CYP1A1*2A was significantly underrepresented among CML cases (6.4% versus 16.0%, P < .01). The significance was calculated using the χ2 test and confirmed by using the Bonferroni correction for multiple testing (k = 5). As an estimate for the relative risk, the crude odds ratio for CYP1A1*2A was 0.36 with a 95% confidence interval of 0.16-0.80 (Table 1). Two individuals in the control group were not successfully genotyped for GSTM1,thus explaining the different number of control samples for this locus. The observed prevalences in the general population are in accordance with data from other studies of groups of whites.4
These data indicate a reduced risk for CML in individuals carrying the mutant allele CYP1A1*2A. This is, to our knowledge, the first report of a protective role of this allele, which is a risk factor for childhood ALL according to Krajinovic et al.4The latter finding is explained by the elevated metabolizing activity associated with CYP1A1*2A, which results in enrichment of reactive intermediates of some carcinogens, for example, polycyclic aromatic hydrocarbons (PAHs), in phase I of metabolism.1 These intermediates must be detoxified by the phase II enzymes such as GSTs. Accordingly, homozygous GSTM1 or GSTT1 deletions are risk factors for childhood ALL and several other neoplasias.1,4 5 In contrast, the relevant carcinogens for CML seem to be detoxified by CYP1A1. This should mean that the phase II metabolism is not needed for detoxification of these carcinogens, which is in accordance with our observation that there is no association between GSTM1 or GSTT1 deletions and CML risk.
Our result that CYP1A1*2A is a protective factor against CML means (1) that genetic susceptibility may be relevant for CML risk, (2) that environmental carcinogens seem to play a role in the etiology of CML, and (3) that the carcinogens relevant for CML risk might differ from carcinogens relevant for other malignancies, for example, PAHs. Moreover, even different hematologic malignancies seem to be preferentially attributed to different chemical carcinogens. Taken together with other results, the available knowledge of inherited genetic and environmentally acquired susceptibility might be relevant for predicting individual risk patterns for hematologic and other malignancies.
Acknowledgments. We thank Ms Susanne Brendel for excellent technical assistance, Dr H. P. Altenburg for statistical analysis, and the DRK-Blutspendezentrale Mannheim for collecting samples of healthy donors.
Supported by the Deutsche Krebshilfe (grant no. 10-1179-Kr1), the Deutsche José Carreras Leukämie-Stiftung eV, and the Forschungsfonds, Fakultät für Klinische Medizin Mannheim, Universität Heidelberg, Germany (grant no. 0022/97)
