Correlation between CYP1A1 transcript, protein level, enzyme activity and DNA adduct formation in normal human mammary epithelial cell strains exposed to benzo[a]pyrene

Abstract

The polycyclic aromatic hydrocarbon (PAH) benzo(a)pyrene (BP) is thought to bind covalently to DNA, through metabolism by cytochrome P450 1A1 (CYP1A1) and CYP1B1, and other enzymes, to form r7, t8, t9-trihydroxy-c-10-(N ²-deoxyguanosyl)-7,8,9,10-tetrahydro-benzo[a]-pyrene (BPdG). Evaluation of RNA expression data, to understand the contribution of different metabolic enzymes to BPdG formation, is typically presented as fold-change observed upon BP exposure, leaving the actual number of RNA transcripts unknown. Here, we have quantified RNA copies/ng cDNA (RNA cpn) for CYP1A1 and CYP1B1, as well as NAD(P)H:quinone oxidoreductase 1 (NQO1), which may reduce formation of BPdG adducts, using primary normal human mammary epithelial cell (NHMEC) strains, and the MCF-7 breast cancer cell line. In unexposed NHMECs, basal RNA cpn values were 58–836 for CYP1A1, 336–5587 for CYP1B1 and 5943–40112 for NQO1. In cells exposed to 4.0 µM BP for 12h, RNA cpn values were 251–13234 for CYP1A1, 4133–57078 for CYP1B1 and 4456–55887 for NQO1. There were 3.5 (mean, range 0.2–15.8) BPdG adducts/10⁸ nucleotides in the NHMECs (n = 16), and 790 in the MCF-7s. In the NHMECs, BP-induced CYP1A1 RNA cpn was highly associated with BPdG (P = 0.002), but CYP1B1 and NQO1 were not. Western blots of four NHMEC strains, chosen for different levels of BPdG adducts, showed a linear correlation between BPdG and CYP1A1, but not CYP1B1 or NQO1. Ethoxyresorufin-O-deethylase (EROD) activity, which measures CYP1A1 and CYP1B1 together, correlated with BPdG, but NQO1 activity did not. Despite more numerous levels of CYP1B1 and NQO1 RNA cpn in unexposed and BP-exposed NHMECs and MCF-7cells, BPdG formation was only correlated with induction of CYP1A1 RNA cpn. The higher level of BPdG in MCF-7 cells, compared to NHMECs, may have been due to a much increased induction of CYP1A1 and EROD. Overall, BPdG correlation was observed with CYP1A1 protein and CYP1A1/1B1 enzyme activity, but not with CYP1B1 or NQO1 protein, or NQO1 enzyme activity.

Introduction

Polycyclic aromatic hydrocarbons (PAHs), partial combustion products of organic material, are ubiquitous in the environment (1), with human exposures coming primarily through inhalation and dietary routes (2). PAHs are human and animal carcinogens, and over the years animal models have provided critical insights into PAH carcinogenic mechanisms (3–5). For example, comparative dose/tumorigenicity studies of dibenzo[a,l]pyrene, 7,12-dimethyl-benz[a]anthracene, benzo[a]pyrene (BP) and two dibenzo[a,l]pyrene dihydrodiols have shown strong associations between dose level and tumour incidence in mice and rats (6). Similarly long-term studies using multiple levels of dosing have shown correlations between PAH–DNA adduct formation and tumorigenesis in trout (7) and mice (8,9). Epidemiologic studies, focused on relationships between PAH exposures, PAH–DNA adduct formation and cancer incidence (10), have shown increased risk of colon adenocarcinoma (11), and lung cancer (12) in individuals with the highest PAH–DNA adduct levels. Weak associations between exposure to PAHs and breast cancer have been reported in adults and post-menopausal women subjected to significant PAH exposures during early life (13–17).

PAHs, including BP, are activated to DNA binding species by cytochrome P450s (CYPs) and epoxide hydrolases (1,4,18). CYP1A1 and CYP1B1 are involved in the initial metabolic activation of PAHs, and CYP1B1 was 10-fold more active, compared to CYP1A1, in converting BP to the BP-7,8-diol (19). In addition to the CYPs, BP metabolism is impacted by the aldo-keto reductases (20), the glutathione transferases and NAD(P)H:quinone oxidoreductase 1 (NQO1). NQO1 was reported to exert a protective effect, reducing the level of PAH–DNA binding (21) and mutagenesis (22). In addition to the metabolic processes involving CYP1A1, CYP1B1 and NQO1, steady state levels of PAH–DNA adducts, often referred to as the biologically effective dose, are impacted by other metabolic enzymes (epoxide hydrolase, glutathione transferases), apoptosis, cell proliferation, DNA repair and peroxidase-mediated radical cation formation.

Previously, our studies employing microarrays and relative real-time PCR to measure fold-changes in gene expression, revealed up-regulation of CYP1A1, CYP1B1 and NQO1 expression in primary normal human mammary epithelial cells (NHMECs) exposed to BP (23). In addition, when a series of NHMEC strains were exposed to 4.0 µM BP for 12h, the fold-inductions for both CYP1A1 and CYP1B1 expression weakly correlated with BPdG adducts formed in these cell strains (24). Whereas monitoring fold-changes in gene expression, in response to exposure, is excellent for screening purposes, pre-existing variations in basal levels of gene expression are not measured and are therefore overlooked. Here we have evaluated the actual numbers of gene transcripts, both before and after 12h of exposure to BP, for CYP1A1, CYP1B1 and NQO1, in NHMEC strains from 16 women, and the results represent human interindividual variability on a small scale. We compared the gene transcript levels with BPdG levels, which were measured in the same cell strains and reported previously (24). Finally, in a subset of four of these NHMEC strains, chosen for their different levels of BPdG formation, we evaluated protein levels of CYP1A1, CYP1B1 and NQO1 by western blot, and enzyme activities. Ethoxyresorufin-O-deethylase (EROD) assay, was used to measure CYP1A1 and CYP1B1 activities together, and NQO1 activity was also measured by enzyme assay. The BPdG values were then compared to the RNA cpn, protein levels and enzyme activities.

Materials and methods

Cell culture and chemicals

NHMECs were isolated from normal breast tissue, and collected at reduction mammoplasty by a process involving mechanical and enzymatic disruption (25). The tissue was obtained through the Cooperative Human Tissue Network, which is sponsored by the National Cancer Institute and the National Disease Research Interchange. Human Studies Review Board approval was sought at the National Institute for Occupational Safety and Health, where the tissue was received and the cells were derived, and a waiver was granted because no unique identifiers accompanied the tissues. Uniform cultures of epithelial cells were obtained by growing enzyme disrupted cells for six passages in serum free mammary epithelial growth medium (MEGM) (Clonetics™, Walkersville, MD). MCF-7 cells, purchased from the American Type Culture Collection (ATCC, Manassas, VA) were grown in minimal essential medium containing 2mM l-glutamate (Clonetics™) and 10% fetal bovine serum.BP (99% purity) was purchased from the National Cancer Institute Chemical Carcinogen Reference Standard Repository, Midwest Research Institute (Kansas City, MO).

Exposure of NHMECs and MCF-7 cells to BP, cell survival and DNA preparation

For these experiments semi-confluent NHMEC strains from 16 different individuals, and MCF-7 cells, were exposed to 4.0 µM BP for 12h (23,24). Cell survival, examined by Cell Titer-Glo Luminescent Assay (Promega, Madison, WI) in the MCF-7 cells and in four of the NHMEC strains, was ≥87%. For the complete set of 16 NHMEC strains, and the MCF-7 cells, BP exposures were performed twice, and RNA transcript analysis and BPdG analyses were performed three times on each sample.

For DNA preparation, cells were washed with PBS, lysed with 5ml lysis buffer (Qiagen, Valencia, CA) and incubated with RNase A (250 μg, Qiagen) for 20min at 37°C followed by proteinase K (500 μg, Qiagen) for 3h at 50°C. At the end of the proteinase K digestion, an equal volume of 100% ethanol was added to each sample and the DNA was isolated using Qiagen columns. The DNA was eluted in molecular biology grade water (26), and stored at −70°C. The quality and quantity of the DNA was measured by UV spectrophotometry at 260/280nm, and then DNA was reconstituted to the required concentrations using TE buffer (pH 7.4, 1×).

Preparation of RNA and cDNA synthesis

Total RNA was isolated from cells using the RNeasy kit (Qiagen), as per the manufacturer’s protocol. RNA was resuspended in 50 µl of nuclease-free water, and the concentration was measured by spectrophotometry, while the purity and quality were measured by gel electrophoresis. RNA (12 µg) was converted to cDNA using the qPCR kit by Affymetrix (Affymetrix, Santa Clara, CA) with oligonucleotide (dT)₂₄ primers. Unused nucleotides and primers were removed from cDNA by column filtration (Microcon YM-30, Millipore, Billerica, MA) and quantity of cDNA was measured by fluorescence spectroscopy using SYBR green at excitation and emission wavelengths of 492 and 526nm, respectively.

CYP1A1, CYP1B1 and NQO1 RNA cpn by quantitative real-time PCR

Quantitation of RNA cpn was accomplished by quantitative real-time PCR (qRT-PCR) using gene specific primers and SYBR green (Bio-Rad, Hercules, CA, USA). qRT-PCR was performed in a total volume of 25 μl reaction mixture containing 1× iQ SYBR Green Supermix, 333nM gene specific primers and an aliquot of cDNA. A standard curve was generated using human universal cDNA (Clontech, Mountain View, CA, USA). A three-step PCR reaction was performed using the iQ5 real-time PCR detection system (Bio-Rad). Universal cDNA having 46000, 133000 and 25000 copies of CYP1A1, CYP1B1 and NQO1 per microgram of cDNA, respectively, was serially diluted to have 0–5712 copies of CYP1A1, 0–16608 copies of CYP1B1 and 0–3120 copies of NQO1 per reaction vial, and mixed with iQ SYBR Green Supermix (1×) and 330nM gene specific primers (CYP1A1: forward primer 5ʹ-CACCTCCAAGATCCCTACACTGA-3ʹ and reverse primer 5ʹ-ACCAGACAGAAGATGACAGAGGC-3ʹ; CYP 1B1: forward primer 5ʹ-ATGTCCTGGCCTTCCTTTATGA-3ʹ and reverse primer 5ʹ- AGACAGAGGTGTTGGCAGTG-3ʹ; and NQO1: forward primer 5ʹ-ATGGTCGGCAGAAGAGCACT-3ʹ and reverse primer 5ʹ- ACCACCTCCCA TCCTTTCTT-3ʹ) for real-time PCR amplification to obtain cycle threshold (C _T) values. The copy numbers in the universal cDNA were established using varying concentrations of genomic DNA as standard, assuming that one human somatic cell contains 7.12 pg of genomic DNA and two copies of each gene, and the C _T values determined for the genomic DNA. Each NHMEC/MCF-7 cDNA sample was assayed in duplicate and the copy numbers were determined using the universal cDNA standard curve and expressed as RNA cpn.

Quantification of BPdG adducts

The r7, t8-dihydroxy-t-9,10-epoxy-7, 8, 9, 10-tetrahydrobenzo[a]pyrene (BPDE)–DNA chemiluminescence immunoassay (CIA) was performed as described (27) with minor modifications. The BPdG values presented here for NHMECs were previously generated in this laboratory from duplicate experiments and reported (24). For the current studies, the previously reported BPDE–DNA CIA values were used, but the same NHMEC strains were grown again and assayed for RNA cpn. In addition, the MCF-7 cells were exposed to 4 µM BPDE for 12h, the DNA was extracted and assayed by BPDE–DNA CIA. The lower limit of detection of the BPDE–DNA CIA, using 10 μg DNA, was 0.3 adducts/10⁸ nucleotides. The BPDE–DNA standard curve in the CIA showed 50% inhibition at 0.60±0.08fmol BPdG (mean ± SE, n = 30).

CYP1A1 and 1B1 protein levels determined by western blot

Western blotting was employed to evaluate CYP1A1, CYP1B1 and NQO1 protein levels in four NHMEC strains, M98030, M99016, M000012 and M98026, which had BPdG levels of 2.6, 4.1, 5.9 and 10.6 adducts/10⁸ nucleotides, respectively. Proteins from cell lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane (Applied Biosystems, Foster City, CA, USA), and probed with specific antibodies for CYP1A1 (GENTEST BD Biosciences, Woburn, MA, USA), CYP1B1 (GENTEST BD Biosciences), NQO1 (Novus Biologicals, Littleton, CO, USA) and β-actin (Chemicon International, Atlanta, GA, USA). The protein bands were visualised after incubation with secondary antibody conjugated to streptavidin-alkaline phosphatase followed by Nitro-Block Enhancer (Applied Biosystems, Bedford, MA, USA) and CDP-Star (Applied Biosystems). Blots were repeated 4–6 times.

CYP1A1 and 1B1 activity determined by EROD assay

EROD assays were conducted on cells cultured to semi-confluence in 12-well plates with 1ml medium/well and exposed to 4.0 µM BP for 12h. After incubation, the medium was removed and the wells were washed three times with Dulbecco’s phosphate buffered saline (DPBS). EROD activity was measured as previously described (28,29) with slight modifications. Intact cells were incubated in 250 µl of DPBS containing 5 µM ethoxyresorufin and 1.5mM salicylamide for 20min at 37°C. At the end of incubation, duplicate aliquots of 100 µl were transferred to wells of an opaque 96-well plate. The resorufin formed was measured using an Infinite 200 fluorescence reader (Teacan, Männedorf, Switzerland) set at excitation and emission wavelengths of 560 and 592nm, respectively. Values (fmol/min/mg protein) for exposure groups were calculated using a resorufin standard curve generated using the same 96-well plates. After completion of the EROD assay, the wells were washed with PBS, lysed in 250 µl of PBS using sonicator set 20% amplitude in a five 4 s pulses with 5 s of interval to prevent increase of temperature in the wells. An aliquot of the lysate was transferred to polypropylene 96-well plate (Grainer Bio) and used for measuring protein by Pierce Bicinchoninic Acid protein assay as per the manufacturers’ protocol (Thermo Scientific, Rockford, IL) using BSA as standard.

NQO1 activity by enzyme assay

Spectrophotometric analysis (30) of NQO1 was followed at 600nm, using 40 µM 2,6-dicholoroindophenol as substrate, 200 µM NADH with or without 20 µM dicumerol. Activity was expressed as the dicumerol-sensitive decrease in absorbance or µmoles 2,6-dicholo-indophenol reduced/min/mg protein. Protein was measured by Pierce Bicinchoninic Acid protein assay (Thermo Scientific).

Statistical analysis

The statistical significance of the resulting changes was analysed using SigmaStat 3.11 (Systat Software, Inc., San Jose, CA). The BP-exposed cells were compared to the unexposed controls, and when the equality of the variance and the normality of the data were confirmed, they were further analysed by t-test. Correlations between BPdG adduct level and CYP1A1, CYP1B1 and NQO1 RNA cpn, or CYP1A1, CYP1B1 and NQO1 protein expression, or EROD and NQO1 enzyme assays, were determined using the Spearman test and/or the Pearson product moment correlation statistic with a correlation coefficient window of (−1) to (+1).

Results

BPdG levels in the NHMEC strains and MCF-7 cells

The 16 strains of NHMECs, and the MCF-7 cells, were exposed to 4.0 µM BP for 12h and evaluated for BPdG by BPDE–DNA CIA. For the NHMECs, Table 1 shows variability for BPdG levels ranging from 0.2 to 15.8 BPdG adducts/10⁸ nucleotides, with the median at 1.9 BPdG adducts/10⁸ nucleotides. The BPdG values (Table 1) were reported previously (24). Table 1 also shows an adduct value of 790 BPdG adducts/10⁸ nucleotides for MCF-7 cells, which was approximately 400-fold higher than the NHMEC values.

Table I.

BPdG adducts/10⁸ nucleotides [from ref. (24)] and RNA copies/ng cDNA^a for CYP1A1, CYP1B1 and NQO1 in 16 strains of NHMECs, and MCF-7 cells, without and with 12h of exposure to 4.0 µM BP

NHMEC	BPdG/108	CYP1A1 a	CYP1A1 a	CYP1B1 a	CYP1B1 a	NQO1 a	NQO1 a
Strain ID	Nucleotides	Unexposed	BP-exposed	Unexposed	BP-exposed	Unexposed	BP-exposed
M99004	15.8	586	9850	1436	33293	16489	27469
M99005	0.2b	446	2352	4744	31980	40112	38060
M99006	1.8	308	1008	3222	12450	20935	30467
M99016	4.1	748	3799	2618	25864	37949	55887
M99021	1.0	291	1022	2388	7487	25017	31395
M99025	0.9	164	793	1763	20608	39600	34696
M00012	5.9	578	7609	1020	13298	15820	19034
M98019	3.1	90	251	2988	8489	20967	22734
M98025	0.2b	175	7166	1379	22366	34023	34326
M98026	10.6	169	13234	661	39433	19110	33729
M98035	1.2	58	285	3806	17864	23704	31356
M98030	2.6	150	2532	4599	57078	29612	36482
M99003	1.3	392	3186	4645	43642	29302	33484
M98021	0.9	836	3408	5587	20255	18651	25583
M98011	2.0	296	6511	1218	10365	20695	35029
M00015	4.3	103	1391	336	4133	5943	4456
Mean (n = 16)	3.5	337	4025	2651	23038	24871	30887
P value
BPdG vs RNA cpn		0.488	0.002	0.059	0.404	0.079	0.572
MCF-7	790.0	8.8	3412.7	1143.7	33792.3	9158.3	14521.4

^aRNA cpn.

^bThese samples were non-detectable and for the purposes of the calculations they were given values of 0.15 (rounded to 0.2), half way between the limit of detection (0.3 BPdG/10⁸ nucleotides) and 0.

RNA cpn for CYP1A1, CYP1B1 and NQO1 in the NHMEC strains and MCF-7 cells

For the genes of interest, RNA cpn values shown in Table 1 include the 16 NHMEC strains and the MCF-7 cells. The Table presents RNA cpn for unexposed cells (basal or endogenous levels), and BP-exposed cells (BP-induced levels). For CYP1A1, in unexposed NHMEC’s, RNA cpn values ranged from 58 to 836, and in BP-exposed cells values ranged from 251 to 13234. The individual fold-increase values for CYP1A1 were between 3-fold and 78-fold. In addition, between the median basal value of 294 RNA cpn, and the median BP-induced value of 2859 RNA cpn, a 10-fold increase was observed for CYP1A1.

In contrast to CYP1A1, for CYP1B1 the NHMEC basal RNA transcripts were more numerous, ranging from 336 to 5587 RNA cpn, with a median of 2503 for unexposed cells. In the BP-exposed cells the induced RNA cpn range was 4133–57077, with a median of 20432. For CYP1B1, fold inductions ranged from 3 to 60, with a median of 9.5. Whereas the fold-increases for BP induction of CYP1A1 and CYP1B1, were in a similar range, Table 1 shows that the numbers of RNA transcripts for CYP1B1 were substantially higher than those for CYP1A1, in the same cell strains.

Table 1 also gives RNA cpn values for NQO1 in unexposed and BP-exposed NHMECs. In contrast to CYP1A1 and CYP1B1, in some cell strains there was no NQO1 induction by BP, and in others the induction was minimal (<2-fold). The median RNA cpn values for unexposed and BP-exposed groups were 23336 and 32440, respectively. Substantial interindividual variability was displayed among NHMEC strains, with regard to change in NQO1 RNA cpn level in response to BP exposure, as 13 strains showed a modest induction with BP, 2 strains showed a decrease with BP, and 1 did not change.

For MCF-7 cells, the CYP1A1 RNA cpn was nine in unexposed cells and 3413 in BP-exposed cells, giving a 379-fold induction in the presence of BP. This was a much greater induction than that found in the NHMECs, and likely contributed to the high BPdG levels observed in these cells. For MCF-7 cells the CYP1B1 RNA cpn was 1144 in unexposed cells and 33792 in BP-exposed cells, giving a 29-fold induction in the presence of BP. This was similar to the induction observed in the NHMECs. For NQO1 RNA cpn the MCF-7 cells showed a 1.6-fold modest induction in the BP-exposed cells compared with the unexposed cells. The substantial CYP1A1 induction in BP-exposed MCF-7 cells suggested a rationale for the high BPdG level (790 adducts/10⁸ nucleotides) observed with the same concentration of BP exposure, compared to NHMECs (mean 3.5 BPdG/10⁸ nucleotides). With the exception of basal levels of CYP1A1, the RNA cpn values shown in Table 1 for the MCF-7 cells were within the range for those seen for the NHMECs.

Comparison of CYP1A1, CYP1B1 and NQO1 RNA cpn and BPdG adducts in the NHMEC strains

A major goal in designing these studies was to evaluate correlations between BPdG adduct formation and expression levels of metabolic enzymes responsible for BP biotransformation. In Figure 1 we present data for the relationship between BPdG levels and RNA cpn, where the basal or endogenous RNA cpn values were subtracted from the induced cpn values, and plotted as a function of BPdG level. There was a highly-significant relationship between BPdG levels and CYP1A1 RNA cpn for the 16 NHMEC strains (Figure 1A), indicating that BP-induced CYP1A1 levels were strongly positively-associated with BPdG formation (P = 0.002). In contrast, in the same NHMEC strains, induced levels of CYP1B1 RNA cpn (Figure 1B) were not significantly associated with BPdG formation (P = 0.278), suggesting a more complicated relationship between this enzyme and BPdG formation in NHMECs. By comparing the RNA cpn values for CYP1A1 and CYP1B1 in Figure 1A and B, respectively (also Table 1), it is apparent that the BP-induced CYP1B1 RNA cpn levels were about 4-fold higher than those for CYP1A1. Finally, Figure 1C shows a similar graph for NQO1, in which there was no correlation between BP-induced NQO1 RNA cpn and BPdG adduct formation (Pearson P = 0.109; Spearman P = 0.06). Although Figure 1 shows BPdG plotted as a function of RNA cpn where the basal levels of RNA cpn have been subtracted, the graphs of the original data showed the same relationships (data not shown).

Fig. 1.

Correlation between BPdG adduct formation and BP-induced RNA cpn for CYP1A1 (A), CYP1B1 (B), and NQO1 (C), in 16 strains of NHMECs, where the corresponding basal or endogenous levels of RNA cpn have been subtracted. By Pearson and/or Spearman rank correlation statistical analysis: (A) Pearson P = 0.002; (B) Pearson P = 0.278; (C) Pearson P = 0.109 and Spearman P = 0.06.

Western blots for CYP1A1, CYP1B1 and NQO1 in four NHMEC strains

Using four NHMEC strains chosen to span the spectrum of BPdG levels from high to low (n = 4), Western blots were performed to evaluate protein levels for CYP1A1, CYP1B1 and NQO1. Figure 2 shows representative western blots for CYP1A1 in unexposed cells (A) and BP-exposed cells (B) and quantitative (C) CYP1A1 western blot data (n ≥ 4), expressed as a function of β-actin levels. Figure 2D, similar to Figure 1A, shows a strong positive correlation between CYP1A1 protein level and formation of BPdG adducts in the four chosen cell strains (P = 0.0128).

Fig. 2.

Western blot analyses for CYP1A1 (A–D), CYP1B1 (E–H) and NQO1 (I–L) for four NHMEC strains having different levels of BPdG, as well as MCF-7 cells. These are M98030, M99016, M00012 and M98026 having 2.6, 4.1, 5.9 and 10.6 BPdG adducts/10⁸ nucleotides, respectively. A, E and I show representative blots for unexposed cells; B, F and J show representative blots for exposed cells; C, G and K show protein levels (mean ± SD in luminescence units) for ≥4 western blots per protein; D, H and L show the correlation between NHMEC protein level and BPdG adducts, where statistical significance (P = 0.0128) was observed only with D.

Using the same four NHMEC strains we examined protein levels of CYP1B1 by western blot and the results are shown in Figure 2. Again, representative blots for unexposed cells (Figure 2E) and BP-exposed cells (Figure 2F) are shown, as well as CYP1B1 protein quantitation for ≥4 western blots (Figure 2G). Finally, Figure 2H shows the lack of a correlation between BP-induced levels of CYP1B1 protein and BPdG adducts in the four cell strains (P = 0.696).

Similarly, we examined protein levels of NQO1 by western blot and the results are shown in Figure 2. Again, representative blots for NQO1 in unexposed cells (Figure 2I) and BP-exposed cells (Figure 2J) are shown, as well as protein level quantitation for ≥4 western blots (Figure 2K). Figure 2L shows the lack of a correlation between BP-induced levels of NQO1 and BPdG adducts (P = 0.874).

By western blots, in addition to the four strains of NHMECs, we also examined MCF-7 cells. For CYP1A1 (Figure 2C) and for NQO1 (Figure 2K) BP exposure caused large inductions of the enzyme protein levels, but there was no induction of CYP1B1 (Figure 2G).

EROD enzyme activity and NQO1 enzyme activity in four NHMEC strains

Confirmation of the above mentioned relationships was obtained by measurement of enzyme activities in the same four NHMEC strains. Combined CYP1A1/1B1 activity was measured by EROD assay (Figure 3), and NQO1 activity was measured by a dicumerol-sensitive assay (Figure 4).

Fig. 3.

EROD enzyme activity, comprising both CYP1A1 and CYP1B1 activities, for (A) unexposed (□) and BP-exposed (■) NHMECs (n = 4 strains; see legend to Figure 2) and MCF-7 cells; (B) comparison between BPdG adduct formation and EROD activity (P = 0.038) in NHMECs.

Fig. 4.

NQO1 enzyme activity for (A) unexposed (□) and BP-exposed (■) NHMECs (n = 4 strains; see legend to Figure 2) and MCF-7 cells; (B) the comparison between BPdG adduct formation and NQO1 activity in NHMECs was not significant.

Figure 3A shows EROD activity for the four NHMEC strains and MCF-7 cells, and Figure 3B shows a strong correlation between BPdG adduct formation and EROD activity, which is presumably driven by CYP1A1 (P = 0.038). Note that the induced EROD activity in the MCF-7 cells is much higher than that in the NHMECs (Figure 3A), suggesting that this may contribute to the high BPdG levels seen in the MCF-7 cells.

Figure 4A shows NQO1 activity in the NHMECs and MCF-7 cells, indicating that BP-induction of this enzyme is lacking in some cell strains. In the MCF-7 cells, levels of NQO1 activity were similar to those observed in the NHMECs. Figure 4B shows that there may be an inverse correlation between BPdG formation and BP-induced NQO1 enzyme activity (r = −0.524) but the association was lacking statistical significance.

Discussion

In this study, we have used NHMEC strains, which are not cell lines but are primary cells cultured from healthy normal breast tissue obtained at reduction mammoplasty, to evaluate interindividual variability in BP bio-transformation and DNA adduct formation. Comparison of RNA cpn with BPdG adduct level was achieved for CYP1A1, CYP1B1 and NQO1 in the 16 NHMEC strains. Further evaluation of CYP1A1, CYP1B1 and NQO1 proteins by western blot, and enzyme activities by EROD and NQO1 assays was carried out in four NHMEC strains chosen because they had varying levels of BPdG adduct, from high to low. Overall the study shows that, of the three enzymes studied, CYP1A1 induction is the driving force behind BPdG formation in NHMECs. Significant correlations were found between BPdG adduct level and CYP1A1 RNA cpn, CYP1A1 protein level by western blot, and EROD activity, presumably through CYP1A1, though the assay also measures CYP1B1. The study also shows that CYP1B1 RNA transcripts, in both unexposed and exposed cells, are much more abundant than CYP1A1 RNA transcripts, despite the fact that they do not correlate with BPdG formation. NQO1 RNA transcripts are not consistently induced by BP in the NHMECs and therefore there was no correlation, either positive or negative, with BPdG formation.

The data presented here show a novel evaluation of RNA expression for CYP1A1, CYP1B1 and NQO1. The power of this approach, a quantitation of RNA transcript copies before and after induction by chemical exposure, is the capacity to reveal endogenous and xenobiotic-induced levels of RNA transcripts as specific numbers, rather than the usual fold-change documented by microarray or RT-PCR. In the NHMECs we observed that endogenous or basal levels of CYP1A1 are several-fold lower than endogenous levels of CYP1B1, despite the greater magnitude of BP-associated induction typically seen with CYP1A1. This complicates the role of CYP1B1 in BPdG formation in NHMECs. Table 1 shows that endogenous levels of CYP1B1 were associated with BPdG levels (P = 0.031 by Spearman rank correlation), suggesting that this enzyme may play a subtle role in modulating the low levels of BP/PAH exposures that are chronically prevalent in our environment. Reports in the literature have shown CYP1B1 to have both activating and protective effects, and the ultimate role for this enzyme appears to vary with cell type, organ and whole animal (31).

In these studies MCF-7 breast cancer cells were given the same BP exposure conditions as the NHMECs, and the BPdG adduct levels were more than 400-fold higher than those found in the NHMECs. Interestingly, the magnitude of this difference is not completely explicable by differences in RNA cpn for the enzymes evaluated here, as the BP-induced levels for CYP1A1, CYP1B1 and NQO1 in MCF-7 cells were within values generated for the BP-exposed NHMECs. However, the EROD activity levels in BP-exposed MCF-7 cells, in the range of 1200fmol/min/mg protein, were much higher than those observed in the NHMECs, which averaged 125fmol/min/mg protein, suggesting that EROD activity may be an important contributing factor. Interestingly, BP-exposed MCF-7 cells had high levels of NQO1 protein by western blot, but the NQO1 enzyme activity was not induced by BP in these cells. The use of MCF-7 cells alone provides limited information, since they are abnormal and represent only one phenotype, however, the comparison with the NHMECs does provide an interesting contrast. Overall the data suggest that the expressed enzyme levels contribute more to BPdG adduct formation than the numbers of RNA cpn, which were similar in both NHMECs and MCF-7 cells.

These studies were performed using 16 different strains of normal human breast epithelial cells, all obtained from different individuals. The variability observed here provides an indication of human interindividual variability for formation of BPdG adducts. Considering the ubiquitous nature of PAH exposures, and indications that PAH–DNA adducts form in many organs of the human body (32), it is interesting to find only a 15-fold difference in BPdG level across these 16 NHMEC strains. There are many studies in the literature documenting PAH exposures and PAH–DNA adduct formation in the human population, and in some organs PAH–DNA adduct formation indicates both exposure and cancer risk (10,32). However, an association between PAH–DNA adduct formation and an increase in breast cancer risk is not always found, and is at best a weak one (13,14,16,33). Neither the early studies summarised in (34), nor the more recent larger trials, supported an association between increased risk of breast cancer and high levels of bulky DNA adducts measured by ³²P-postlabelling (17,35), or high levels of PAH–DNA adducts measured by BPDE–DNA ELISA. Therefore, the results of this study may tell us more about human interindividual variability with respect to PAH exposure, than human breast cancer risk.

A wealth of published information suggests conflicting roles for BP metabolism by CYPs in different tissues or cells. In the mouse, BP or PAH induction of CYP1A1 was associated with dose-related DNA damage in target organs and tumour incidences (8,36). That paradigm was largely supported in the studies presented here. However, studies utilizing genetically modified mouse models have revealed a complex picture. For example, in CYP1 knockout mice, inducible CYP1A1 was important for detoxification of reactive BP metabolites in intestine and liver, while CYP1B1 was responsible for metabolic activation of BP in spleen and bone marrow, causing immune damage in the absence of CYP1A1 (37). In the cytochrome P450 oxidoreductase null (HRN) mouse model, higher BPdG levels were observed in livers of HRN mice compared to livers of wild type mice (38). Therefore, at least some of the CYP enzymes may have inhibited activation of BP to DNA binding species in the HRN mice. In a study using genetically modified mice with varying levels of the aryl hydrocarbon receptor (AhR), increased levels of BPdG adducts and BP metabolites were found in several organs of AhR null mice, compared to AhR (+/+) and AhR (+/−) mice, also suggesting that Ahr-independent pathways may contribute to BPdG formation in AhR null mice (39).

In conclusion, the data presented here show that, despite the much higher RNA levels of endogenous and BP-induced CYP1B1, there was no association between BPdG formation and BP-induced CYP1B1. CYP1A1 RNA expression, CYP1A1 protein levels, and EROD (CYP1A1 plus CYP1B1) enzyme activity were all associated with BPdG adduct formation, suggesting that CYP1A1 is more important than CYP1B1 or NQO1 in the metabolic events governing the formation of BPdG in NHMECs. The data also suggest that individuals with the lowest levels of CYP1A1 RNA cpn in breast may have the lowest PAH–DNA adduct levels, though it is not clear if this is a breast cancer-related risk factor.

Funding

This research was supported by the intramural research program of the Center for Cancer Research, National Cancer Institute, National Institutes of Health, and the intramural Health Effects Laboratory Division of the National Institute for Occupational Safety and Health, Centers for Disease Control.

Conflict of interest statement: None declared.