Human Lung Microsomal Cytochrome P4501A1 (CYP1A1) Activities: Impact of Smoking Status and CYP1A1, Aryl Hydrocarbon Receptor, and Glutathione S-Transferase M1 Genetic Polymorphisms¹

Most chemicals that initiate lung cancer, including those found in tobacco smoke, require bioactivation in the lung to their “ultimate” genotoxic metabolites that interact with DNA. Individual differences in the ability to bioactivate carcinogens may contribute to host susceptibility and, thus, may play a role in lung cancer risk (1, 2, 3).

In 1973, Kellerman et al.(4) first described an association between high AHH³ inducibility in cultured lymphocytes and bronchogenic carcinoma. Subsequently, numerous studies established a link between AHH inducibility and lung cancer risk (reviewed in Ref. 1). Much of human lung microsomal AHH activity has been attributed to a single member of the P450 multigene superfamily, CYP1A1, and cigarette smoking is considered to be the most important factor related to the pulmonary expression of this enzyme (5, 6, 7, 8, 9, 10). CYP1A1 bioactivates PAHs, a major class of tobacco procarcinogens (3). The CYP1A1 gene exhibits inducibility in human lung through ligand binding to the AHR, and CYP1A1 can be induced via an AHR-dependent mechanism by PAHs and other similar planar compounds (Refs. 5, 11, 12, 13 and references therein). A positive correlation exists between CYP1A1 activity and pulmonary-PAH-associated DNA adduction, implicating this enzyme as an important factor in the etiology of lung cancer (reviewed in Refs. 14, 15).

In recent years, the genetic determinants of individual differences in CYP1A1 expression and their association with lung carcinogenesis have been examined. CYP1A1 contains two prominent polymorphic sites associated with lung cancer: a 3′ flank T to C transition known as the MspI mutation (CYP1A1*2A; Ref. 16); and an exon 7 heme binding region A to G transition resulting in an Ile to Val substitution (CYP1A1*2C; Ref. 17). Individuals possessing both mutations are denoted as CYP1A1*2B (Ref. 18 and CYP1A1 allele nomenclature website).⁴

Numerous epidemiological studies have addressed correlations between CYP1A1 genotype and lung cancer susceptibility, yielding conflicting results (Refs. 16, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51; reviewed in Table 1). Early studies first pointed toward an association between variant genotype and lung cancer risk in Japanese populations (16, 19, 21, 25), with no associations found in Caucasian populations (20, 22), conceivably because of relatively low allele frequencies in the latter. A number of studies have found associations with lung cancers in individuals carrying at least one copy of either variant allele (27, 28, 31, 35, 39, 40, 42, 49), a more likely scenario in Caucasian populations, in which variant allele frequencies are low (52). The assumption is that genetically determined alterations (polymorphisms) in the expression of CYP1A1 affect related enzyme activities and, thus, the manner in which carcinogens are metabolized.

Studies examining the effects of the CYP1A1*2A and CYP1A1*2C mutations on enzyme activity in various models have also produced conflicting results (Refs. 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66; Table 1). Multiple studies also have examined the association of CYP1A1 genetic polymorphisms with biomarkers or biological outcomes relevant to lung cancer and lung cancer risk. Association of CYP1A1 variants with DNA-adduct levels in lung tissues (67, 68, 69, 70, 71, 72) and circulating lymphocytes (73) from smokers, pulmonary CYP1A1 expression (74), lung cancer prognosis (75), and p53-tumor suppressor gene mutations (76, 77, 78, 79, 80) have all been investigated, yielding conflicting results (Table 1). A second exon 7 heme binding region polymorphism, a C to A transversion resulting in a Thr to Asn substitution, in close proximity to CYP1A1*2C has been identified (81). Although not associated with lung cancer susceptibility (81), recent reports from expression systems (65, 66) suggest differences in substrate affinities and catalytic activities associated with the CYP1A1*4 variant protein CYP1A1.4 relative to CYP1A1.1 (Table 1).

Possible associations between the AHR gene polymorphisms and CYP1A1 inducibility also have proven to be controversial (Refs. 64, 82, 83, 84, 85; Table 2). Kawajiri et al.(83) found that the AHR codon 554 Arg to Lys polymorphism (AHR⁵⁵⁴) in exon 10 (receptor transactivation domain) was not associated with either CYP1A1 inducibility in cultured lymphocytes or with lung cancer incidence in a Japanese population. Contrary to this observation, Smart and Daly (64) found that cultured lymphocytes from individuals carrying at least one AHR⁵⁵⁴ Lys-variant allele had significantly elevated CYP1A1 activity and expression. However, a recent study by Wong et al.(85) suggested no evidence for altered biological function or CYP1A1 induction associated with AHR⁵⁵⁴ Arg to Lys.

Pulmonary glutathione S-transferases (GSTs) play a major role in detoxification of reactive electrophiles via conjugation with glutathione and, thus, are integral components in the balance between carcinogen activation and detoxification in the lung. Numerous studies have assessed the combined effects of CYP1A1 genetic polymorphisms and the GSTM1 deletion polymorphism in relation to lung cancer (21, 25, 27, 33, 39, 41, 42, 45, 47, 49, 51, 75, 76, 78, 80), the majority of which suggest increased risk when variant genotypes are combined (Table 1). We have previously demonstrated that the GSTM1 gene deletion results in a functional deficiency in GSTM1–1 activity in human lung cytosols (86). Vaury et al.(87) suggested that high inducibility of CYP1A1 is associated with the GSTM1-null genotype in cultured cell lines, possibly because of persistence of CYP1A1 inducers attributable to lowered GST-catalyzed metabolism. However, these results were not corroborated in two other studies using cultured lymphocytes for in vitro phenotyping (64, 88).

What the majority of previous studies have failed to address is whether polymorphisms actually translate into significant alterations in CYP1A1 enzyme activity/induction in the target tissue for pulmonary carcinogens. Such knowledge is necessary to validate the presumed basis for reported associations between genotype and lung cancer susceptibility and to explain conflicting data in the literature. As such, objectives of this study were to determine whether CYP1A1 and/or AHR genotypes correlate with CYP1A1-related activities in human lung microsomes, and to determine the impact of the GSTM1 gene deletion polymorphism and smoking status on these activities.

Chemicals were obtained as follows: Taq polymerase, deoxynucleoside triphosphates, MgCl₂, and PCR buffer from Life Technologies, Inc., Gaithersburg, MD,; MspI and BsaI from New England Biolabs, Mississauga, ON, Canada; MeaIII from Boehringer Mannheim, Dorval, PQ, Canada; 7-ERF and resorufin from Molecular Probes, Eugene, OR; PCR primers from Cortec DNA Services Laboratories, Kingston, ON, Canada; Metaphor agarose from FMC Bioproducts, Rockland, ME. All other chemicals were reagent grade or higher and were obtained from common commercial suppliers.

Human lung tissue, devoid of macroscopically visible tumors, was obtained from Kingston General Hospital, in accordance with procedures approved by the Queen’s University Research Ethics Board. After informed consent, sections of peripheral lung (20–100 g) were removed during clinically indicated lobectomy. Immediately after removal, the tissue was placed in 0.9% NaCl solution and kept on ice. Elapsed time between surgical resection and tissue manipulation was ∼15 min. Initially, 0.5 cm³ was removed from the tissue, snap-frozen in liquid N₂, and stored at −80°C for DNA isolation, and 1.5 cm³ sections were removed from the cut specimen surface and placed in 10% neutral buffered formalin. The fixed tissue was dehydrated and embedded in paraffin, and 5-μm sections were stained with H&E (89). They were then examined by light microscopy to confirm the absence of microscopic tumors. Patients were characterized with respect to age, gender, surgical diagnosis, possible occupational exposure to carcinogens, drug treatment 1 month prior to surgery, and self-reported smoking history (Table 3). Patients were classified as former smokers if smoking cessation was greater than 2 months before surgery. This time interval was chosen to eliminate the inductive effects of cigarette smoke on CYP1A1(5).

Genomic DNA was isolated by protease digestion followed by standard phenol:chloroform extraction and ethanol precipitation (90). Patients were genotyped for the CYP1A1*2A (MspI) polymorphism by PCR-RFLP (28). Restriction digest products were resolved in an ethidium bromide-stained 1.8% agarose gel. Genotyping for the CYP1A1*2C (Ile to Val) polymorphism was performed using a PCR-based designed-RFLP method (91). Restriction digest 157- and 85-bp fragments for CYP1A1*1 homozygotes and 157-, 136-, and 85-bp fragments for CYP1A1*2C heterozygotes were resolved in a 2% agarose:2% Metaphor (1:1) agarose gel. Given the close linkage of these mutations in Caucasians (22), it is most likely that CYP1A1*2A and CYP1A1*2C variants are occurring on the same allele. Thus, individuals were denoted as CYP1A1*2B when genotyping revealed alleles containing both MspI and Ile to Val mutations. Genotyping for the CYP1A1*4 Thr to Asn polymorphism was modified from Cascorbi et al.(81). Briefly, the PCR product from the CYP1A1*2C genotyping protocol (242 bp) was digested with BsaI, yielding restriction digest 217-bp fragments for CYP1A1*1 homozygotes and 242- and 217-bp fragments for CYP1A1*4 heterozygotes, and were resolved as described for CYP1A1*2C. DNA from an individual genotyped CYP1A1*1/4 and confirmed by direct sequencing (Cortec DNA Services Laboratories, Kingston, ON) was used as a control in all subsequent CYP1A1*4 genotyping assays. Patients were also genotyped for the GSTM1 gene deletion polymorphism after resolving PCR products in a 3% agarose gel (92). Control DNA samples from individuals who were heterozygote variant for both CYP1A1*2A and *2C and GSTM1-null for the gene deletion polymorphisms were generously provided by Dr. D. A. Bell (National Institute of Environmental Health Sciences, Research Triangle Park, NC). Patients were genotyped for the AHR⁵⁵⁴ Arg to Lys polymorphism by PCR-SSCP as described by Wong et al.(85).

EROD was used to assess human pulmonary CYP1A1 activities. Although CYP1A2 and CYP1B1 can also catalyze the dealkylation of 7-ERF, CYP1A2 expression in human lung remains controversial (74, 93), and the CYP1B1 protein is not significantly expressed in normal human lung tissue (94, 95). Furthermore, 7-ERF is a highly selective substrate for CYP1A1, with human recombinant CYP1B1 and CYP1A2-catalyzed oxidation activities being one-tenth and one-forty-fifth those of CYP1A1, respectively (96). Thus, the EROD assay is highly selective for CYP1A1 activity in human lung microsomes.

Peripheral human lung microsomes were prepared from fresh or frozen lung tissues using standard subcellular fractionation techniques (97, 98). In the case of frozen tissues, lung tissue was initially cut into 1.5-cm³ sections, wrapped in aluminum foil and snap-frozen in liquid N₂, and stored at −80°C until microsomal preparation. Protein concentration was determined by the method of Lowry et al.(99). A modified version of the EROD assay (100) was used to assess CYP1A1 activity. Briefly, the 3.0-ml EROD reaction mixture contained 0.1 m Na/KPO₄ buffer (pH 7.6), 5.0 μm 7-ERF (in DMSO), 0.25 mm NADPH, and 0.5 mg microsomal protein/ml. Resorufin formation was monitored spectrofluorometrically over time. Initial reaction velocity was estimated from the linear portion of the product formation curve and was quantitated using a resorufin standard curve.

EROD results are based on duplicate values for each patient. Statistically significant differences between non-/former smokers and current smokers were determined by the Mann-Whitney U test, because of heterogeneity of variance. Otherwise, statistically significant differences were determined using Student’s t test. Correlations were examined using Pearson linear correlation analysis. For combined genotype analysis, one-way ANOVA was used, followed by the Newman-Keuls post hoc test, unless Bartlett’s test revealed heterogeneity of variance, when a nonparametric ANOVA (Kruskal-Wallis) was used to analyze the data. In all cases, significance was assigned at P < 0.05.

Patient demographics were obtained and recorded preoperatively (Table 3). EROD activities were assessed in human lung microsomes made from peripheral lung specimens obtained from 45 male and 25 female individuals undergoing lobectomy (average age, 62 ± 9.9 years; range, 40–82 years; Table 3). Although racial demographics were not obtained from individual patients, the population in the catchment area for Kingston General Hospital is predominantly Caucasian.

Mean EROD activities in microsomes from current smokers (12.11 ± 13.46 pmol/min/mg; n = 46) were approximately 15-fold higher than those in microsomes from non-/former smokers (0.77 ± 1.74 pmol/min/mg; n = 24). In fact, in 16 of 24 non-/former smoker tissues, EROD activity was nondetectable (i.e., <0.10 pmol/min/mg; Table 3). Mean EROD activities were not significantly different between tissues from female (8.72 ± 9.75 pmol/min/mg; n = 16) and male (13.93 ± 14.81 pmol/min/mg; n = 30) current smokers. In current smokers for whom self-reported complete smoking histories were available (n = 36), there was no correlation between EROD activity and pack-year consumption (i.e., high lifetime cigarette consumption did not correlate with higher EROD activities; Fig. 1,a). However, there was a significant negative correlation between patient age and EROD activity (Fig. 1 b).

PCR-RFLP-based CYP1A1 genotyping methods revealed both CYP1A1*2A and CYP1A1*2B heterozygotyes and CYP1A1*4 heterozygotes. None of the patients was homozygous for the CYP1A1*2A, *2B,*2C, or *4 mutations. Allele frequencies in this predominantly Caucasian North American population were 0.864, 0.136, 0.036, and 0.014 for CYP1A1*1, CYP1A1*2A, CYP1A1*2B, and CYP1A1*4 respectively. CYP1A1 activities from CYP1A1*1/1 current smokers (n = 33) and heterozygous variant (CYP1A1*1/2A) smokers (n = 10) were not significantly different (Fig. 2,a; Table 4). Only three current smokers were heterozygous CYP1A1*1/2B and exhibited CYP1A1 activities similar to individuals CYP1A1*1/1 (Fig. 2,a; Table 4). None of the patients’ genotypes exhibited CYP1A1*2C alleles (i.e., exon 7 Ile to Val alone). One current smoker was heterozygous variant CYP1A1*1/4 and exhibited relatively high, but comparable, CYP1A1 activities to individuals genotyped homozygous CYP1A1*1 at this locus (patient 6JM, Table 3).

SSCP analysis for AHR⁵⁵⁴ Arg to Lys polymorphism revealed seven AHR⁵⁵⁴Arg/Lys heterozygotes, giving allele frequencies of 0.050 and 0.950 for Lys⁵⁵⁴ and Arg⁵⁵⁴ respectively. None of the patients was homozygous AHR⁵⁵⁴Lys/Lys. There was no significant difference in CYP1A1 activity in lung microsomes from current smokers AHR⁵⁵⁴Arg/Lys (n = 5) compared with microsomes from individuals genotyped AHR⁵⁵⁴Arg/Arg (n = 41; Fig. 2 b).

Mean CYP1A1 activities in microsomes from current smokers with GSTM1 null genotype (n = 28) and those in microsomes from individuals carrying at least one copy of the GSTM1 gene (n = 18) were very similar (Fig. 2 c).

There were no significant differences in EROD activities with various combinations of the CYP1A1, AHR⁵⁵⁴, and GSTM1 genotypes (Table 4).

EROD activities were not detectable in microsomes from 12 of 46 current smokers (Table 3). Three of the 12 patients were CYP1A1*1/2A and the remaining were CYP1A1*1/1. One of the twelve patients was AHR⁵⁵⁴Arg/Lys, and 4 of the 12 patients carried at least one copy of GSTM1. All of the 12 patients were current long-term smokers with at least 20 pack-years of exposure. A number of these individuals also were receiving pharmacotherapy (Table 3), although most of the drugs apparently do not interact with CYP1A1. A possible exception is beclomethasone dipropionate (patients 7DM and 9HM, Table 3) which has been shown to inhibit induction of AHH activity in lung tissue from mice treated with the PAH benzanthracene (101).

Furthermore, EROD activities were detected in microsomes from 8 of 24 non-/former smokers (ranging from 0.19 to 6.62 pmol/min/mg), with the highest activities in microsomes from the three patients who reported cessation of smoking within the previous six months. All three patients were CYP1A1*1/1, with only one individual AHR⁵⁵⁴Arg/Lys. Microsomal EROD activities from the remaining five patients (CYP1A1*1/1, n = 2; CYP1A1*1/2A, n = 1; CYP1A1*1/2B, n = 2; and AHR⁵⁵⁴Arg/Arg, n = 5) were very low with the exception of one patient coded 5 IM, who had EROD activities of 1.77 pmol/min/mg. Among other pharmacotherapies, this individual was being treated for gastric reflux disease with omeprazole, a know inducer of CYP1A1 (Table 3; Refs. 102, 103).

Lung disease may have contributed to EROD activities. Pulmonary microsomes from patient 1DM, a longtime current smoker, exhibited the highest EROD activities in our patient population. This individual, genotyped CYP1A1*1/1, AHR⁵⁵⁴Arg/Arg, and GSTM1-null, showed a positive tuberculosis skin test at the time of surgery, and presented with an active pulmonary Aspergillus fumigatus infection. Initially, lobectomy was indicated for a cavitary lesion believed to be caused by tuberculosis. However, postoperative pathology cultures revealed no sign of Mycobacterium, and did reveal an Aspirigillus fumigatus infection, resulting in a diagnosis of aspirigillosis granulomatosis combined with acute pneumonia.

The association between human CYP1A1 genetic polymorphisms and lung cancer remains controversial despite a relatively large number of epidemiological studies in various populations (Table 1). Similarly, the existence of a human polymorphism in AHR affecting induction of CYP1A1 remains unclear. The majority of studies have failed to address whether such polymorphisms actually translate into significant functional alterations in CYP1A1 enzyme activity/induction in the human lung. The principle aim of the present study was to determine whether a correlation exists between genotype and biotransformation phenotype in human lung microsomes. We have previously demonstrated observable functional differences in polymorphic forms of GST enzymes in human lung cytosols using selective substrates (86, 104).

Numerous epidemiological studies have suggested increased risk for lung cancer in individuals carrying even one copy the CYP1A1*2A variant allele in North Americans (predominantly Caucasian; Refs. 37, 39, 42), in Scandinavians (27), and in other ethnic populations (40, 42). Ishibi et al.(40) also demonstrated an association of heterozygosity for the CYP1A1*2B variant with a 2-fold increase in lung cancer risk in Mexican- and African-Americans, and Dresler et al.(49) found an increased risk for lung cancer in a group of females, the vast majority of which were heterozygous for CYP1A1*2C; the increased risk was particularly apparent when combined with GSTM1-null. In contrast, a recent meta-analysis by Houlston (105) found little evidence to support an association of CYP1A1*2A or *2C variants with lung cancer risk. Several studies examining the functionality of the CYP1A1 genetic polymorphisms in cultured lymphocytes have suggested differences in inducibility/activity measured by AHH or EROD to be associated with heterozygotes possessing either a CYP1A1*2A(36, 53, 54, 59) or a CYP1A1*2C allele (56, 58). However, Cosma et al.(55) found no correlation between CYP1A1*2A variants and gene inducibility measured by mRNA quantitation, and Wedlund et al.(60) found no association between CYP1A1*2A or *2B variants and AHH inducibility in cultured lymphocytes in the same Mediterranean family examined by Petersen et al. (Ref. 53; reviewed in Table 1). Contrary to results of Rojas et al.(69) in a Russian population, Butkiewicz et al.(71) found significantly higher PAH DNA-adduct levels in lung tissues from Polish individuals genotyped CYP1A1*1/2C combined with GSTM1-null genotypes, which suggested that heterozygous variant forms of CYP1A1 may contribute to high levels of carcinogen activation and DNA-adduction. Interestingly, Saarikoski et al.(74) noted that lung tissue from a CYP1A1*1/2A individual demonstrated greater expression of CYP1A1 by immunohistochemical and in situ hybridization analysis, compared with tissues from five CYP1A1*1/1 individuals.

Our examination of the effects of the CYP1A1 genetic polymorphisms on related enzyme activities in human lung suggested that CYP1A1 heterozygous genotypes (CYP1A1*1/2A, CYP1A1*1/2B, and CYP1A1*1/4) do not affect CYP1A1 activities in peripheral human lung. The lack of observed contributions of genotype on EROD activities could have been affected by interindividual variability in microsomal activities or by the absence of homozygous variant microsomal samples and, in the case of the CYP1A1*4 allele, could be limited by the fact that only one current smoker carried one copy of this allele. It is important to note that associations between CYP1A1 variants and lung cancer require high n values because of the multifactorial processes involved in the development of the end point. However, CYP1A1 enzyme activity is much more proximate to CYP1A1 genotype, so relevant differences in phenotype, if present, should be revealed more readily.

Observed allele frequencies for the CYP1A1*1,*2A, and*2B alleles from our patients were similar to those reported in the literature for healthy Caucasians (52) and were lower for the CYP1A1*4 allele (81) and, therefore, do not suggest an association with lung cancer risk. Allele frequencies for AHR⁵⁵⁴Arg and Lys were not higher than expected and were similar to those reported in other mixed North American populations (85).

Mutation of AHR might affect ligand binding and hence CYP1A1 inducibility. However, on the basis of the limited number of Lys⁵⁵⁴ heterozygotes analyzed and consistent with results of Wong et al.(85), we did not see an association between AHR genotype and CYP1A1 activities.

Our GSTM1 results differ from those of an earlier report of enhanced CYP1A1 inducibility with the GSTM1 null genotype in human lymphoblastoid B cells (87), but agree with subsequent studies that did not find a role for GSTM1 genotype in CYP1A1 inducibility (64, 88). Thus, association of the GSTM1 gene deletion polymorphism with lung cancer susceptibility (reviewed in Ref. 2) more likely relates to the demonstrated decreased conjugation of electrophilic substrates (86, 106). Combinations of the CYP1A1 and AHR genetic and GSTM1 polymorphisms also do not appear to affect CYP1A1 related activities in peripheral human lung.

In the present study, human lung microsomes from 70 surgical specimens obtained from a predominantly Caucasian North American population produced EROD activities similar to those reported by Wheeler et al.(6). Consistent with the concept of smoking-mediated CYP1A1 induction in human lung (5, 6, 7, 8, 9, 10), EROD activities were detectable in microsomes from patients who were recorded as current smokers, but were low or undetectable in most of those recorded as former smokers/nonsmokers (Table 3). The detectable levels of CYP1A1 activity in microsomes from a small number of patients who were classified as former smokers could conceivably be attributed to inaccuracies in individual patient smoking status self-reporting and/or to induction by therapeutic agents. In the case of the former, results demonstrating CYP1A1 induction in current smokers compared with non-/former smokers suggested accurate self-reporting on the whole. For the latter, we did observe CYP1A1 activity in microsomes from a longtime former smoker (12 years) treated with omeprazole, a compound known to induce CYP1A1 in cell lines and human tissues (102, 103). Also of potential significance was our observation of very high CYP1A1 activity in lung microsomes from a reported current smoker with asperigillosis. This points to the potential role of inflammatory mediators in altering pulmonary CYP activities. For example, Ohnhaus and Bluhm (107) found high activities for P450 enzyme non-selective 7-ethoxycoumarin-O-deethylase activities in pulmonary microsomes from tuberculosis-positive patients. At present, the contribution that elevated CYP1A1 activity and, hence, altered carcinogen metabolism might make to the observed link between lung disease and lung cancer (108), is not known.

Consistent with other studies (10, 109, 109, 110, 111, 112), we found marked interindividual variability in pulmonary CYP1A1 activities in microsomes from current smokers. Interestingly, peripheral lung microsomes from some (∼26%) current smokers exhibited nondetectable EROD activities, and this was compatible with reports distinguishing a subset of individuals with no detectable CYP1A1 induction by immunochemistry or by measuring AHH or EROD (7, 74, 113). However, consistent with the results of Anttilla et al.(114), a variant CYP1A1 or AHR⁵⁵⁴ genotype did not account for the lack of detectable CYP1A1 activity as this group comprised individuals CYP1A1*1/1, CYP1A1*1/2A, AHR⁵⁵⁴Arg/Arg, and AHR⁵⁵⁴Arg/Lys.

In contrast to Mollerup et al.(115), who reported that females had significantly higher levels of pulmonary mRNA CYP1A1 expression, we found no significant differences in EROD activities between lung microsomes from males and females. Our results suggest that the association of CYP1A1*2C and increased lung cancer risk in females (49) may not be attributable to higher CYP1A1 activities. The lack of significant correlation between EROD activity and pack-year consumption in current smokers, but a significant decline in EROD activity associated with aging, suggests that, as the body ages, the inducibility/activity of CYP1A1 decreases (Fig. 1, a and b). The lack of correlation of EROD and pack-year consumption is probably affected by the age-related decline in EROD activities, because pack-year values were generally higher in older patients. Similar findings demonstrating a negative correlation between age and CYP activity in human liver have been reported previously (116, 117, 118). On the basis of patient self-reporting, none of our patients recorded as current smokers ceased smoking within the time interval between the day of surgery and the two-month-prior-to-surgery cutoff designated for classification as a “former smoker.” Thus all CYP1A1 activities from current smokers are believed to be representative of tobacco smoke induction up to the day prior to surgery. Patient self-reporting of tobacco consumption was limited to lifetime pack-year values, so it was not possible to assess the relationship of EROD activity and tobacco consumption on a per cigarette basis at the time of surgery.

Recently, kinetic differences have been associated with the CYP1A1*4 variant protein CYP1A1.4 relative to CYP1A1.1 (65), particularly with regards to metabolism of the tobacco carcinogen benzo[a]pyrene (66). However, differences in EROD kinetic behavior were considered as minor (see Table 1). On the basis of results from one current smoker genotyped CYP1A1*1/4, and consistent with the literature, we saw no apparent difference in EROD activity compared with individuals CYP1A1*1/1 at this locus. The impact of this polymorphism and other novel variants (64, 114) on lung cancer susceptibility requires further investigation.

The results of this study suggest that the association of the CYP1A1 genetic polymorphisms with lung cancer susceptibility, especially when an association has been linked to CYP1A1*2A or *2C heterozygotes, is not occurring as a result of altered CYP1A1 activity. The data pertaining to CYP1A1 smoking-induced elevated enzyme activities in smokers are in agreement with other studies, and the variability, as well as lack of detectable CYP1A1 activity among some current long-term smokers, demonstrates the need for further investigation into the mechanism(s) of CYP1A1 regulation. Our results support the suggestion (3) that the association of the CYP1A1 variants with lung cancer in some populations occurs as a result of linkage with some other mutation.

We thank Barbara J. Veley and Carole Fargo for their assistance in conducting this study and Dr. D. A. Bell (National Institute of Environmental Health Sciences, Research Triangle Park, NC) for generously providing the exon 7 Ile to Val designed-RFLP genotyping protocol and control DNA from heterozygotes for both the MspI and exon 7 Ile to Val variant alleles, as well as DNA from a GSTM1-null individual.