Induction of CYP1A1 and CYP1B1 and formation of carcinogen–DNA adducts in normal human mammary epithelial cells treated with benzo[a]pyrene

Abstract

Inter-individual variation in formation of carcinogen–DNA adducts and induction of cytochrome P450 genes was measured in 23 cultured normal human mammary epithelial cell (NHMEC) strains established from reduction mammoplasty tissue. Semi-confluent cells were exposed to 4 μM benzo[a]pyrene (BP) for 12 h and BP–DNA adduct levels were measured by chemiluminescence immunoassay using antiserum elicited against DNA modified with r7, t8-dihydroxy-t-9, 10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE). BP–DNA adduct levels for 22 of 23 different cell strains ranged from non-detectable (three samples) to about 15 adducts/10⁸ nucleotides. Increases in levels of CYP1A1 and CYP1B1 were detected using both oligonucleotide arrays and reverse transcription/quantitative real-time polymerase chain reactions (RT-PCRs). For CYP1A1 and CYP1B1, the oligonucleotide array data and RT-PCR data were highly correlated (r=0.73 and 0.70, respectively), suggesting that oligonucleotide arrays are a suitable gene discovery tool, and demonstrating that the complementary and efficient RT-PCR may be used to confirm microarray data for a specific gene in a large number of samples. As measured by RT-PCR, inter-individual variation in CYP1A1 induction was 100-fold, while the variation in CYP1B1 induction was almost 40-fold. On a per-person basis, CYP1A1 and CYP1B1 induction were well-correlated (r=0.88, P<0.001), which is to be expected as they are under the control of a common transcriptional regulation mechanism in response to BP exposure. Inter-individual variation in carcinogen–DNA adduct formation could not be explained only by variation in levels of CYP1A1 or CYP1B1 induction, as neither was well-correlated with BPDE–DNA adduct level (r=0.40 and 0.50 for CYP1A1 and CYP1B1, respectively). Evaluation of glutathione-S-transferase M1 genotype (GSTM1 positive or null) revealed an apparent correlation between positive GSTM1 genotype and BPDE–DNA adduct levels (r=0.84 and 0.77 for CYP1A1 and CYP1B1, respectively); however, after removal of the single outlier this relationship was not significant. Overall the data suggest that BPDE–DNA adduct levels in normal human breast tissue may be modulated by multiple factors that include, but are not exclusive to, CYP1A1 and CYP1B1 inducibility and the presence or absence of GSTM1.

Introduction

Human exposure to mixtures of environmental pollutants that include polycyclic aromatic hydrocarbons (PAHs) is unavoidable [1]. Because many of these compounds are human carcinogens, elevated cancer rates may occur in individuals subjected to environmental and occupational PAH exposures [2]. Carcinogenic PAHs induce multiple metabolic pathways in human tissues and cells, as evidenced by DNA-microarray studies [3], [4]. In humans, carcinogenic PAHs are biotransformed through induction of multiple enzyme systems, particularly the cytochrome P450s [5], [6], and wide inter-individual variation in response to PAH exposures has been documented [4]. As breast cancer incidences appear to be increasing in recent years, several molecular epidemiologic studies have focused on normal breast tissue and breast cancer cells, and established associations between patterns of carcinogen–DNA adduct formation and cytochrome P450 polymorphisms [7], [8], [9]. These studies suggest that both environmental and genetic factors may contribute to the burden of human female breast cancer.

While CYP1A1 was cloned in 1985 and its expression had been studied long before that, CYP1B1 was not cloned until almost a decade later and has been less intensively investigated [10], [11]. Both CYP1A1 and CYP1B1 enzymes may be involved in the metabolic activation of exogenous carcinogens that reach human mammary tissues [12], [13], [14], [15], [16]. In addition, CYP1B1, known to metabolize steroids, may be involved in the metabolic activation of estrogens to carcinogenic intermediates; consequently, it has been identified as a target for inhibition in anticancer strategies [6], [17]. Both CYP1A1 and CYP1B1 are activating enzymes that can metabolize xenobiotics in concert with detoxifying enzymes. Glutathione-S-transferase M1 (GSTM1), a detoxication enzyme, contributes to the disposition of PAHs by effectively lowering the concentration of activated xenobiotic available for DNA-binding.

In these experiments, we have used normal human mammary epithelial cells (NHMECs), cultured from normal mammary tissues obtained at reduction mammoplasty from 22 individuals to study gene expression of the carcinogen-metabolizing enzymes CYP1A1 and CYP1B1. The data have been stratified by GSTM1 genotype. NHMEC strains were exposed to the carcinogenic PAH benzo[a]pyrene (BP) to explore inter-individual variability in BP–DNA adduct formation and to compare BP–DNA adduct formation with expression of the metabolic enzymes. An earlier study has documented the metabolism of PAHs in cells of similar derivation [18], but advanced methods to compare gene expression of the metabolic activation enzymes with DNA adduct formation (the toxicogenomic aspects of PAH metabolism) were not available at that time. In a previous study [4], we used oligonucleotide arrays to investigate the complete pattern of gene expression changes in BP-exposed cultured NHMECs obtained from four different donors, and those studies showed that some of the most consistently and extensively induced genes were CYP1A1 and CYP1B1. Here we compare the formation of BP–DNA adducts with the extent of CYP1A1 and CYP1B1 induction in cell strains from 22 individuals exposed to BP. Values for CYP1A1 and CYP1B1 induction determined by oligonucleotide arrays were confirmed by reverse transcription-polymerase chain reaction (RT-PCR) and stratified by GSTM1 status for comparison with BP–DNA adduct levels.