Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Whereas the risk for bronchogenic carcinoma is markedly increased in smokers compared with nonsmokers, less than 10% of heavy smokers actually develop bronchogenic carcinoma. This observation implies considerable inter-individual variation in risk. This variation has been attributed to an interaction between exogenous and endogenous factors (1). Exposure to urban air is a potentially important exogenous determinant of risk (2). Polymorphisms in xenobiotic metabolism enzyme (XME)² genes responsible for activation and subsequent inactivation of procarcinogens (5) are important endogenous determinants (10). For example, polymorphisms in cytochrome p450 (CYP)1A1 and glutathione transferase µ are associated with increased risk for bronchogenic carcinoma in some studies (11, 12). As opposed to measuring polymorphisms in or near the coding regions of XME genes and correlating them with other determinants of risk, another approach is to measure inter-individual variation in quantitative levels of gene expression. For example, inter-individual variation in quantitative levels of expression of aromatic hydrocarbon hydroxylase (AHH) activity occurs (13, 14) and is associated with inter-individual variation in the amount of benzo(a)pyrene adducts formed in bronchial epithelial DNA (15). It is to be expected that understanding the risk a particular cell type encounters following exposure to inactive procarcinogens will require specific and independent measurement of multiple activating and inactivating enzymes. For example, cytochromes p450 1A1 and 1B1 both metabolize polycyclic aromatic hydrocarbons (PAH), are inducible (16, 17), and are expressed in many human tissues, including bronchial epithelial cells (BEC) (18). When cells express CYP1A1 activity but no microsomal epoxide hydrolase (mEH) activity, benzo(a)pyrene is metabolized primarily to noncarcinogenic 7,8 benzophenolic products (19). In contrast, when both CYP1A1 and mEH are expressed, the primary product is the carcinogenic 7,8-diol, 9,10-epoxide derivative. Thus, activation of PAH procarcinogens is dependent on the relative expression of key activating XME.

The genes that were evaluated in this study are candidate markers for exposure to environmental xenobiotics and/or susceptibility to bronchial disease (e.g., bronchogenic carcinoma). For a gene to serve as a measure of exposure, it must be expressed in the tissue of interest. In addition, it must be inducible to a degree that is detectable by current methodology and over a range that allows determination of exposure/response relationships. In contrast, an indicator of susceptibility need not be inducible as long as it is expressed at different constitutive levels in different individuals. XME genes that are responsible for the metabolism of procarcinogens may potentially serve as indicators both for exposure to procarcinogens and for susceptibility to cancer.

Because there are marked inter-tissue differences in expression and regulation of genes, it is important to assess expression of XME genes in BEC, the progenitor cells for bronchogenic carcinoma. It may then be possible to correlate levels in more readily accessible tissues, such as lymphocytes (20), with those in BEC for epidemiologic studies. Efforts to evaluate gene expression in BEC have been limited in the past by the small number of cells recoverable for analysis. While RNAse protection methods are more sensitive than Northern analysis, they are still 100- to 1,000-fold less sensitive than reverse transcriptase polymerase chain reaction (RT-PCR). Immunohistochemical methods for measuring gene expression are highly sensitive and have led to important observations regarding distribution of cells expressing different XME genes within the lung and the bronchial epithelium (13, 21), but they are not quantitative. Recently developed RT-PCR methods are both quantitative and highly sensitive (22); thus, detailed quantitative measurement of XME gene expression in BEC is now possible. CYP1A1, -1B1, and -2B7, mEH and NADPH oxidoreductase (NADPH OR) are expressed in BEC (18). Inter-individual variation in expression of these genes was evaluated in BEC using modifications of previously described methods for quantitative RT-PCR (23).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

PCR buffer (10X) was obtained from Idaho Technology, Inc. (Idaho Falls, ID); Taq polymerase (5 u/µl), M-MLV RT, M-MLV RT 5X buffer (250 mM Tris-HCl, pH 8.3; 375 mM KCl; 15 mM MgCl₂; and 50 mM DTT), oligo dT primers, RNAsin, PGEM size marker, and dNTPs were obtained from Promega (Madison, WI); LE and Nusieve agarose were obtained from FMC Bioproducts (Rockland, ME); and Tri Reagent was obtained from Molecular Research Center, Inc. (Cincinnati, OH). RNAse-free water was obtained from Gibco-BRL (Grand Island, NY). All other chemicals and reagents were molecular biology grade.

Cell Lines

The BEP2D cell line was a human papillomavirus-immortalized human bronchial epithelial cell line. Establishment, methods of maintenance, and characterization of this cell line have been described previously (25).

Methods were described previously for establishment and maintenance of human lymphocyte cell lines genetically engineered to express different defined levels of human XME genes (26). The cell lines used in these studies include the parent AHH-1 human -lymphoblastoid cell line, MCL5, and h1A1v2 which contain CYP1A1 expression vectors, and h2E1/OR which contains a human CYP2E1/NADPH OR expression vector.

Bronchoscopy

Ten nonsmoker and nine smoker volunteers were recruited from the university environment for BEC recovery via video-assisted fiberoptic bronchoscopy through the oropharynx according to previously described protocols (18). Informed consent was obtained from each subject. The protocol was approved by the Research Subjects Review Board at the University of Rochester. The subjects were free of respiratory symptoms for at least 4 wk prior to cell recovery and no subject expressed symptoms consistent with chronic or episodic diseases.

Sample Preparation and RNA Extraction

For both the bronchoalveolar lavage and bronchial brush samples, aliquots were taken for total and differential cell counting and viability analyses. The remaining cells were centrifuged at 300 × g for 10 min at 4°C, the supernatant was poured off, and 1 ml of Tri Reagent added for each 10⁷ cells to lyse the cells, denature the proteins, and release nucleic acids. The solution then was either processed for RNA extraction or frozen at 70°C. RNA extraction and RT were carried out according to previously described methods (18, 27) and methods described in the Tri Reagent manufacturer protocol (28). Frozen samples could be stored for as long as 72 h without loss of RNA. Cell count and viability were determined by trypan blue exclusion on a hemacytometer.

Primers

Information on the primers for amplification of native cDNA sequences have been published previously (18). Competitive templates (CT) were prepared according to previously described methods (29), and primer sequences were selected based on two criteria: that PCR product lengths were sufficiently shorter than the native sequence to allow separation by agarose gel electrophoresis, and that they provided optimal PCR efficiency with an annealing temperature of 58°C based on Oligo software analysis. The upstream primers were the same as those previously reported (18). The downstream primers were designed for synthesis of the CTs and contained the original 20 bp downstream primers (18) at their 5' ends and 20 bp lengths of internal sequence at their 3' ends. The 40 bp CT primer sequences and references for cDNA sequences listed in Genbank are as follows: CYP1A1 5' ACA GCA GGC ATG CTT CAT GGT TCT CAC CGA TAC ACT TCC GCT 3' (30); CYP1B1 5' CGT TCG GGC TGA GGC TGG TGC CGT CAA CAG GAA CCC GCA GGC 3' (17); CYP2B7 5' CCA TGT GGA GCA GGT AGG TGG TGT GCC CCA GGA AAG TAT T 3' (31); mEH 5' GGG TGA AAC GGA ACT TAT CGG CCC CCA CCA CCC ATC TTC AA 3' (32); and NADPH OR 5' AGA AGT CCT GGG CAT TGT CGG CAG GTC GGC CAG GTC ATA CTC 3' (33). Although these CT primers may bind to either the 3' or internal sequence of the template, the only sequences amplified were those resulting from annealing at the internal sequence (29). The resulting mutant PCR products contained the previously published downstream primer sequences (18) at positions closer to the upstream primers than in native PCR products. As a result, amplification of these CTs with the previously reported pairs (18) resulted in products that are shorter than the native sequences and therefore easily separated from the native sequences for analysis.

Competitive Template Standard Mixture

Each of the CTs was amplified in multiple 100-µl reactions in an MJ thermocycler, electrophoresed on preparative low-melt agarose gels, cut out; phenol extracted, taken up into 20-µl of TE (10 mM Tris-Cl, pH 7.4, 0.1 mM EDTA) buffer, and quantified by electrophoresing triplicate samples on an agarose gel adjacent to 1 µg of PGEM marker. Aliquots of each of the XME gene CT stock solutions sufficient to give a concentration of 10⁸ M were combined with TE buffer in a final volume of 20 µl. A separate solution of -actin at 10⁸ M in TE buffer was prepared. Aliquots of the -actin 10⁸ M solution and the xenobiotic gene 10⁸ M solution were mixed in different proportions (see next section). These CT mixtures were used as internal standards for quantitative PCR. CT mixtures derived from the same original stock were used for all of the data reported here.

Since, in this study, each of the primer pairs was specific for a particular gene and had no significant homology to other gene sequences, no interference between competitive templates was expected. In confirmation of this, each of the primer pairs for a particular gene amplified a single band from the CT mix when no native cDNA was present. If all of the CTs in this study were amplified simultaneously, it would be possible to distinguish a separate band for each geneexcept it would not be possible to distinguish between CYP1A1 and CYP2B7 CTs (each is 248 bp in length).

Quantitative PCR Amplification

For most experiments, DNA was PCR-amplified in a Rapidcycler air thermocycler (Idaho Technology, Inc., Idaho Falls, ID). Reaction volumes were 10 µl and contained 0.05 µg of each primer, 0.5 units Taq polymerase, 1 µl of 10X PCR buffer (500 mM Tris, pH 8.3; 2.5 mg/ml bovine serum albumin; 5% Ficoll 400; 10 mM tartrazine; and 30 mM MgCl₂), 0.2 mM dNTPs, and water. The reactions were cycled 35 to 38 times, with each cycle consisting of 5 s at 94°C, 10 s at 58°C, and 15 s at 72°C with a slope of 9.9 for a total amplification time of approximately 20 min. For data presented in Figure 3, DNA was PCR-amplified according to previously described methods (23).

View larger version (38K): [in this window] [in a new window]   Figure 3.   Induction of CYP1A1 in BEP2D cells by -napthoflavone. (a) RNA (10 µg) from cells exposed to -napthoflavone (100 µM) for 0, 8, or 24 h were electrophoresed and Northern blotted, then incubated with 32P-labeled cDNA probe for CYP1A1. Equal loading of RNA into each well was confirmed by quantitation of ribonuclear RNA bands on ethidium bromide-stained gels. (b) Aliquots of cDNA samples (1, 3, or 5 µl) extracted from cells incubated for 24 h in the absence or presence of -napthoflavone [the same cells as in (a)] were multiplex RT-PCR-amplified with primers for CYP1A1 and GAPDH in the presence of CTs for CYP1A1 (3 × 104 molecules/reaction) and GAPDH (6 × 105 molecules/reaction). The sizes for the native and CT bands for GAPDH were 788 bp and 582 bp, respectively, and for CYP1A1 were the same as in Figure 2.

View larger version (67K): [in this window] [in a new window]   Figure 2.   Xenobiotic gene expression in BEC from a nonsmoker (Subject 2) and a smoker (Subject 11). Ethidium bromide-stained agarose gels containing PGEM marker, negative control (lane 1), amplified -actin (lane 2), CYP1A1 (lane 3), CYP1B1 (lane 4), CYP2B7 (lane 5), and mEH (lane 6). In each lane, the upper band is native, the lower band is CT. For -actin, three bands are observed. The middle band is a heterodimer consisting of one strand of native-length cDNA and one strand of CT-length cDNA (confirmed by S1 nuclease digestion, data not shown); as expected, it migrates exactly halfway between the native and CT bands. The sizes of the native and CT, respectively, are CYP1A1, 355 bp and 248 bp; CYP1B1, 355 bp and 231 bp; CYP2B7, 352 bp and 248 bp; and mEH 351 bp and 258 bp. The right-pointing arrows indicate the CYP1A1 native and CT products and the left up-pointing arrow the CYP1B1 products, and for each of these reactions the amount of product is too small to be able to detect heterodimer. The left down-pointing arrow indicates the native CYP2B7, which completely out-competes the CYP2B7 CT. The two bands observed in lane 6 represent the native and CT products of mEH PCR amplification. Native and CT mEH bands are both apparent in both subjects. Each of the 10 µl PCR reactions contained 6 × 105 CT molecules for -actin, and 6 × 102 molecules of CT for each of the target genes. Following electrophoresis, digital images were obtained; in lanes where both native and CT bands were visualized, they were analyzed by Collage software (Fotodyne, Inc., Hartland, WI) with the Laplacian edge-sharpening feature activated. Subject 2 is a nonsmoker; Subject 11 is a smoker (see Figure 1).

View larger version (23K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   View larger version (33K): [in this window] [in a new window]   View larger version (24K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   Figure 1.   Quantification of gene expression in BEC. Quantification of CYP1A1 (a), CYP1B1 (b), CYP2B7 (c), mEH (d), and NADPH OR (e) in each of 19 subjects. Subjects 1-10 were nonsmokers and subjects 11-19 were smokers. The bars and error bars represent the average and range of two values or, where indicated by asterisk, the mean and standard deviation of three or more values. The value of the x-axis (10 molecules/106 -actin molecules) is the estimated lowest quantifiable level of gene expression under conditions used. Data were obtained by analyzing digital images of ethidium bromide-stained gels such as those presented in Figure 2 using methods described in MATERIALS AND METHODS.

For experiments presented in Figures 1 and 2, the goal was to evaluate several genes simultaneously in a single cDNA sample and then compare results obtained with those from different cDNA samples. For this purpose, each PCR reaction consisted of separate aliquots of (1) a master mixture containing CTs, buffer, dNTPs, water, taq polymerase, and cDNA from a single sample, and (2) primers for the gene to be amplified. Thus, from a master mixture containing all of the CTs and cDNA from a single sample, multiple different genes could be quantified simply by placing different primer pairs in each of multiple tubes. Because both the cDNA and CTs were in the master mixture, there was no inter-tube variation in the ratio between them, such as may result from pipeting errors. For the experiment presented in Figure 3, in contrast, the goal was to evaluate a single gene (CYP1A1) in multiple different cDNA samples. For this purpose, each PCR reaction consisted of separate aliquots of (1) a master mixture containing CTs, buffer, dNTPs, water, taq polymerase, and primers for both the housekeeping gene (glyceraldehyde-6-phosphate dehydrogenase, GAPDH) and CYP1A1, plus (2) a single cDNA sample.

Selection of the CT mixture with the most appropriate -actin/target gene ratio for an experiment was based on the relative cDNA concentration in the sample being analyzed and on the relative expression of the target gene compared to the housekeeping gene. The housekeeping gene was -actin for data in Table 1 and Figures 1, 2, and 4, and GAPDH for Figure 3b. These housekeeping genes were equally suitable. In current methods, both housekeeping genes are assessed simultaneously and, based on data obtained thus far, there is small inter-individual variation in relative expression between them (data not shown). Typically, 1 µl of cDNA would compete approximately equally with 10⁵ to 10⁶ molecules of -actin or 10³ to 10⁴ molecules of GAPDH and, therefore, allow quantification. To prepare a reaction mixture containing 6 × 10⁵ molecules of -actin CT, 1 µl of stock CT mixture containing -actin CT at a concentration of 10¹² M was added to the reaction mixture to give a final volume of 10 µl and a final concentration of 10¹³ M. For some samples, the concentration of cDNA was lower and therefore a stock CT mixture with a -actin CT concentration of 10¹³ M (6 × 10⁴ molecules/µl) or even 10¹⁴ M (6 × 10³ molecules/µl) was used. For most cDNA samples the appropriate concentration of CT for CYP1B1, mEH, and NADPH OR in the reaction mixture was 10¹⁶ M which, in a 10-µl reaction volume, would amount to 60 molecules; in a few cases 10¹⁶ M was more appropriate. For CYP1A1 and CYP2B7, 10¹⁶ M CTs were usually optimal. Thus, CT working solutions with the following -actin CT/XME ratios were prepared: 10¹² M/10¹⁶ M, 10¹³ M/10¹⁶ M, 10¹⁴ M/10¹⁶ M, and 10¹² M/10¹⁵ M. 

View larger version (57K): [in this window] [in a new window]   Figure 4.   CYP1A1 expression in alveolar macrophages of nonsmokers and smokers. We evaluated CYP1A1 expression in the alveolar macrophages of three nonsmokers (Subjects 1, 2, and 5) and three smokers (Subjects 11, 15, and 17) using methods described in Figure 2. The bands indicated by the left-pointing arrows represent the native and CT -actin bands; the right-pointing arrows indicate the native and CT CYP1A1 bands. For smoker Subject 15, the heterodimer is brighter than the native. This occurs when the amount of CT is far greater than the amount of native DNA. In such cases, the amount of native present in the heterodimer was considered when calculating the ratio between native and CT product.

To conserve cDNA, the best strategy was (1) to determine the lowest amount of target gene CT that is quantifiably visible on an ethidium bromide-stained gel following PCR-amplification, (2) then use only the amount of cDNA in the PCR reaction necessary to compete equally with that amount of target gene CT, and (3) to use the housekeeping gene (in this study, -actin or GAPDH) CT concentration that will compete equally with the amount of native cDNA sample determined in step 2.

Electrophoresis and digital imaging were carried out according to previously described methods (18).

Methods for calculation of gene expression were similar to those previously described (23), the difference being that in the method used here, the heterodimers migrate independently from the competitive template band and may be directly quantified. After digital quantification, half of the heterodimer density value (adjusted for size) is added to the native band density value and half is added to the CT band density value.

Expression of Enzymatic Activity

CYP1A1 activity was measured using 7-ethoxyresorufin O-deethylase activity (34) in whole cells. OR activity was measured in microsomes prepared from the cell lines using cytochrome c reduction (35).

Data Analysis

Inter-individual variation in gene expression was assessed by student's t test analysis of untransformed data.

As a result of large inter-individual variation in expression for each of the genes, the data were not normal; therefore, differences in mean levels of expression could not be assessed statistically without logarithmic transformation. Normality was achieved through logarithmic transformation, the mean and standard deviation of the logarithmically transformed data were determined for each gene, and the expression of each gene was compared by evaluating differences between the means using Duncan's multiple range test (36).

Differences between smokers and nonsmokers were assessed with student's t test and Wilcoxon 2-sample test analysis of logarithmically transformed data.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Subjects

Epidemiologic data is provided in Table 2. The mean age for nonsmokers and smokers, respectively, was 26.9 and 30.8 years. Among the nonsmokers and smokers, respectively, 3 of 10 and 5 of 9 were females. Among the nonsmokers were one teacher and 9 students, whereas a variety of occupations were represented among the smokers. The only subject likely to be exposed to non-tobacco xenobiotics on a consistent basis was subject 15, who was a head cook. Of the nonsmokers 8 of 10 were white (European descent) and of the smokers, 7 of 9 were white. The only medication among the subjects was for birth control, which was used by one nonsmoker and one smoker. All of the smokers were actively smoking at least one-half pack per day and had smoked at their usual rate up to the day of the bronchoscopy. None of the nonsmokers had smoked a cigarette within a week of bronchoscopy. Exposure to secondhand smoke was variable and usually consisted of periodic exposure to public areas where smoking was prevalent. At least one nonsmoker subject (Subject 6) had spent several hours the night prior to bronchoscopy in a public restaurant where there was noticeable secondhand smoke.

Cell and RNA Yield from Bronchoscopy Specimens

Recovery of RNA from cell samples, percent of total cells in the samples that were viable, and recovery of cDNA from the RNA following RT were evaluated in 11 representative BEC samples. There were no significant differences between smokers and nonsmokers with respect to any of these results. Among the 11 cases, 30 ± 10% (mean ± standard deviation) of the total recovered cells were viable, amounting to 2.9 ± 2 × 10⁶ viable cells per sample. RNA was extracted from total cell samples, but yield was quantified on the basis of viable cells. The amount of RNA extracted per 10⁶ viable cells was 1.1 ± 0.9 µg and the total amount of -actin cDNA recoverable from the RNA was 3.8 ± 5.7 × 10⁷ molecules/µg. Approximately 10² to 10⁵ -actin molecules were used per assay, depending on the level of expression of the target gene. At a usage level of 10⁵ -actin molecules/assay, an RNA sample containing 3.7 × 10⁷ -actin molecules was sufficient for 3.7 × 10² assays, or enough to evaluate 10² genes in triplicate.

The cDNA samples were assessed for contamination with genomic DNA that may have been carried over from RNA extracted according to the Tri Reagent protocol. If such genomic DNA were present at large enough levels, it would compete with both the cDNA and the competitive template for primers, and thereby lead to error in measurement of gene expression. Whenever the genomic sequence of a gene was known, the primers were constructed such that they spanned at least one intron. As a result, the amplified genomic template would be longer and migrate slower on agarose than amplified cDNA. While the genomic sequence is not known for all of the genes studied, at least 5 genes for which the genomic template spanned by the primers is longer than the cDNA have been evaluated in all 19 cDNA samples used in this study. No band corresponding to the genomic DNA template was observed for any of the genes in any of the samples. The absence of bands corresponding to genomic DNA template amplification confirmed that there was no significant DNA contamination of any of the RNA samples used in this study. For most genes that expressed less than 10 molecules/10⁶ -actin molecules, PCR was repeated in the absence of competitive templates. In this situation, no bands were observed. The genomic sequence is not known for some of the genes that are expressed at detectable levels. Therefore, theoretically, the genomic and cDNA templates could be the same and not separate on an agarose gel. Because the absence of genomic DNA (to less than at least 10 molecules/10⁶ actin molecules) was confirmed for all genes where the genomic sequence was known, and since the genomic copy number for each gene is two, we can be confident that the level of PCR product for the genomic template of the remaining genes can be no more than 10 molecules/10⁶ molecules of -actin. Such a small amount of genomic template would not significantly alter amplification of the cDNA or competitive template. By all of the above measures, no genomic DNA is detected in any of the cDNA samples thus far studied. Therefore, it is concluded that the Tri Reagent RNA extraction protocol is highly efficient at excluding genomic DNA contamination.

Gene Expression

The mean and range of expression for each gene among smokers, nonsmokers, and both groups are provided in Table 1.

Inter-individual Variation in CYP1A1, CYP1B1, and mEH Expression

CYP1A1 and CYP1B1 were each expressed at greater than 10 molecules/10⁶ -actin molecules in BEC of only one of 10 nonsmokers (Subjects 7 and 5, respectively, in Figures 1a and 1b) but expressed at greater than 10 molecules/10⁶ -actin molecules in BEC of all 9 smokers (Figures 1a, 1b, and 2). Expression among smokers ranged over 2.3 logs for CYP1A1 and 2.0 logs for CYP1B1. In all but one smoker (Subject 11) CYP1A1 was expressed at a higher level (ranging from 5- to 15-fold) than CYP1B1 and the mean level of expression for CYP1A1 was more than 5-fold greater than that for CYP1B1 (P < 0.05) (Table 1).

There was large inter-individual variation in expression of mEH (over 1.8 logs) but there was no significant difference (P < 0.05) in expression between smokers and nonsmokers by student's t test or by Wilcoxon 2-sample test. As expected, due to the large variation in expression of CYP1A1, CYP1B1, and mEH, there was large inter-individual variation in CYP1A1/mEH and CYP1B1/mEH ratios. Since CYP1A1 and CYP1B1 are reported to have similar activities toward PAH substrates, the expression of combined CYP1A1 and CYP1B1 to mEH was compared. The CYP1A1-1B1/mEH ratio varies more than 120-fold, from 2.3 to 281, among the smokers evaluated. In nonsmokers, due to very low CYP1A1 and CYP1B1 values, the CYP1A1-CYP1B1/mEH ratio was 0.3 and 0.7 in Subjects 5 and 7, respectively, and less than 0.1 in the other subjects.

Comparison of Quantitative PCR with Northern Analysis for Measurement of CYP1A1 Induction by -Napthoflavone

It was confirmed that PAH induce CYP1A1 in BEC by incubating populations of the human papillomavirus-immortalized human BEC line BEP2D with -napthoflavone. As measured by Northern analysis (Figure 3a), CYP1A1 induction was observed after 8 h of incubation and was higher after 24 h.

cDNA derived from the same RNA used in Figure 3a was also evaluated by multiplex RT-PCR analysis (Figure 3b; see MATERIALS AND METHODS). Multiplex amplification of both the housekeeping gene GAPDH and the target gene CYP1A1 along with their CTs was used in these experiments because multiple different cDNA samples were being evaluated for one gene (CYP1A1) (see MATERIALS AND METHODS). Loading differences were controlled for by comparing the native GAPDH PCR product to the corresponding CT product in each lane. CYP1A1 expression was quantifiable (all 6 bands measurable) in unexposed BEP2D cells in the reaction containing 3 µl of cDNA and in -napthoflavone-exposed cells in the reactions containing 1 and 3 µl of cDNA. Ideally, the value for CYP1A1 expression relative to GAPDH expression should be the same whether 1, 3, or 5 µl is included in the reaction, as long as all 6 bands are measurable. CYP1A1 expression in control cells was 6.4 × 10³ molecules/10⁶ GAPDH molecules and in -napthoflavone-exposed cells was 4.7 and 4.2 × 10⁴ molecules/10⁶ GAPDH molecules for the reactions containing 1 and 3 µl of cDNA, respectively.

CYP1A1 Expression in Alveolar Macrophages

CYP1A1 was measured in the recovered alveolar macrophages (AM) of 3 nonsmoker and 3 smoker subjects (Figure 4). A faint native CYP1A1 band was observed in 1 nonsmoker (Subject 1) corresponding to 70 molecules/10⁶ -actin molecules, and 2 out of 3 smokers (Subjects 11 and 15), corresponding to 30 and 65 molecules/10⁶ -actin molecules, respectively. While the CYP1A1 band was brighter for Subject 11 compared with Subject 15, it is clear from the -actin lane that considerably more Subject 11 cDNA (7.1 × 10⁵ -actin molecules) than Subject 15 cDNA (2.5 × 10⁵ -actin molecules) was used in the PCR reaction. Thus, when the CYP1A1 values are divided by the -actin values, there is a slightly higher expression in Subject 15.

CYP2B7 and NADPH OR Expression in BEC

The mean level of expression of CYP2B7 was approximately 3-fold greater in smokers than in nonsmokers (Table 1), but there was wide inter-individual variation (2.5 logs) and this difference was not statistically significant (at P < 0.05).

NADPH OR expression (Table 1) also did not differ significantly between smokers and nonsmokers, and the inter-individual variation, while significant, was less than that observed in the other genes evaluated (1.3 logs).

Comparison of Gene RNA Level with Function of Protein Product

The BEC samples obtained were too small to quantify protein and/or enzyme activity for any of the genes studied. Therefore, in order to demonstrate that RNA level quantified by RT-PCR correlates with gene function, human lymphocyte cell lines that had been genetically engineered to express CYP1A1 or NADPH OR at different levels (16) were evaluated. These cells could be expanded to the numbers necessary to provide protein and/or enzyme activity sufficient for analysis. CYP1A1 and NADPH OR enzymatic activity were measured by 7-ethoxyresorufin O-deethylase (34) and cytochrome c reductase activity (35), respectively. The CYP1A1 RNA and 7-ethoxyresorufin O-deethylase enzyme activity levels for AHH-1, MCL5, and h1A1v2 cell lines, respectively, were 3.9 × 10² molecules/10⁶ -actin molecules and 0.07 pmole/10⁶ cells/ min; 5.9 × 10³ molecules/10⁶ -actin molecules and 0.2 pmole/10⁶ cells/min; and 2.3 × 10⁴ molecules/10⁶ -actin molecules and 6 pmole/10⁶ cells/min. The NADPH OR RNA and cytochrome c reductase activity levels for AHH-1 and h2E1/OR, respectively, were 1.7 × 10² molecules/10⁶ -actin molecules and 25 nmoles/mg/min, and 1.2 × 10³ molecules/10⁶ -actin molecules and 80 nmoles/mg/min.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

While it is widely accepted that inhaled chemical procarcinogens in cigarette smoke increase risk for bronchogenic carcinoma, this hypothesis is dependent on the assumption that, through some mechanism, procarcinogens are activated in proliferating BEC. This is because activated carcinogens are highly reactive with proteins as well as nucleic acids, and as a consequence, their effects are generally thought to be restricted to the cells in which they are activated. As reported here, CYP1A1 and CYP1B1 are expressed at higher levels in BEC of smokers compared with nonsmokers, and there is significant inter-individual variation in expression of CYP1A1, CYP1B1, and mEH. Consequently, there also is large inter-individual variation in the ratio of CYP1A1/mEH, CYP1B1/mEH, and (CYP1A1 + CYP1B1)/mEH. Based on in vitro mutagenicity studies, these ratios are assumed to be closely associated with risk for DNA-PAH adduct formation and subsequent DNA mutation in BEC. Such relationships currently are being evaluated directly.

Inter-individual variation in expression may result from (1) hereditary variation in constitutive level of expression and/or inducibility, plus (2) differing exposures to occupational or environmental pollutants or tobacco smoke. Expression of CYP1A1 and CYP1B1 genes in only some nonsmokers may result from variation in exposure to environmental pollutants other than cigarette smoke (such as exhaust from internal combustion engines or heat-generation plants) or from inter-individual variation in constitutive level of expression. The population described here is small and therefore it was not possible to statistically assess effects of different levels of smoking or other epidemiologic criteria (Table 2) on gene expression. Important questions that remain to be addressed include: how soon after smoking are effects on XME induction in BEC observed; how long do they last; what are the cumulative effects of many years of smoking on XME expession; and are CYP1A1 and/or CYP1B1 more inducible in some individuals than in others. In whole lung tissue it was determined that the effects of cigarette smoking on inducible XME genes is observable over several weeks (37). Thus, sorting out the effects of inter-individual differences in smoking history from genetic factors on variation in expression of XME genes in BEC would require a larger, more detailed epidemiologic study than that reported here.

The ratio of CYP1A1/CYP1B1 varied considerably among smokers from 0.9 to 41, clearly indicating that these genes are independently regulated. It is possible that the regulatory sequences vary independently in the population and/or that each individual was exposed to a different mixture of xenobiotics.

The levels of CYP1B1 and CYP1A1 expression in most nonsmokers are very low (Figure 1). It is possible that Subjects 5 and 7 represent individuals at increased risk for bronchogenic carcinoma due to inhalation of non-tobacco sources of procarcinogens, such as urban air. Although it should be possible to make quantitative measurements of CYP1A1 and CYP1B1 in additional nonsmokers among the group studied here as long as sufficient cDNA is used in the assay, it is not clear what the significance of such measurements would be. Other XME genes that are expressed at higher levels in nonsmokers may play a more important role in risk determination. It is possible that a subset of cells, perhaps those in the bronchioles or other sites not sampled in this study, express CYP1A1 and/or CYP1B1 at higher levels. However, since most cigarette smoke-associated bronchogenic carcinomas occur in, or proximal to, the regions sampled (secondary and tertiary bronchi), it is not likely that a different level of expression in terminal bronchioles would significantly affect risk for bronchogenic carcinoma.

The bronchial epithelium comprises multiple different cell types, including basal, ciliated, and mucous differentiated. In the studies described here, mixed populations of BEC were assessed. In murine epidermal cells, there is significant intra-tissue variation in expression of cytochrome p450IA1 following induction with the prototype PAH, -napthoflavone (38). It is important to determine which cells in the BEC are expressing CYP1A1 and/or CYPIB1 at higher levels following exposure to PAH in cigarette smoke. If PAH induce CYP1A1 and/or CYP1B1 expression only in terminally differentiated (e.g., ciliated) cells, it will not likely have an effect on risk for bronchogenic carcinoma. While inter-cellular differences in gene expression were not determined in this study, currently in situ RT-PCR is being used to answer these questions.

In a previous study, CYP1A1 expression in the normal whole lung tissue of people with lung cancer was positively associated with smoking (39). The inter-individual variation in CYP1A1 expression was > 100-fold, which is the same order of magnitude as the 280-fold variation reported here in BEC of smokers. The level of expression of CYP1A1 relative to -actin was not reported for whole lung in that study and therefore comparisons cannot be made with this current study on BEC.

CYP1A1 expression was far lower in AM cells compared to BEC of smokers (Figures 1 and 4). CYP1A1 was barely detectable in some of the AM samples even in the absence of CT (data not shown). When measured in the presence of CT, 1 of 3 nonsmokers and 2 of 3 smokers expressed quantifiable levels of CYP1A1. The levels observed in the AM of smokers were not higher than that in the nonsmoker, and were lower than those observed in the BEC of all 9 smokers. This may be due in part to the deposition of PAH-containing tobacco smoke particles predominantly at airway surfaces more proximal than the alveoli. Thus, although AM metabolically activate PAHs (40) and tobacco smoking induces AHH activity in AM (41), the level of CYP1A1 activity in AM, including those of smokers, is so low relative to that observed in BEC that it is not likely to significantly alter the risk for DNA-adduct formation in BEC. CYP1B1 was expressed at too low a level to be quantifiable in any of the AM samples. This suggests that, as with CYP1A1, CYP1B1 is expressed at lower levels in AM compared with matched BEC samples.

The amount of CYP2B7 and NADPH OR in whole lung tissue from smokers and nonsmokers was recently evaluated by RNase protection assays (42). The levels of CYP2B7 relative to -actin for BEC (Table 1) were exactly the same as reported previously for whole lung (8.9 × 10³ molecules/10⁶ -actin molecules) but the amount of inter-individual variation was much greater in BEC (> 200-fold versus 6-fold). Further, the NADPH OR expression in BEC (Table 1) was approximately 4-fold less than previously reported for whole lung (42), and again the inter-individual variation was greater in BEC compared to whole lung (28-fold versus 5-fold). It is possible that the greater inter-individual variation in expression of CYP2B7 and NADPH OR in BEC compared with whole lung tissue results from variable degrees of epithelial metaplasia or more extreme variation in exposure to xenobiotics.

RT-PCR with controls used in the manner described here is a highly sensitive and reproducible way to measure messenger RNA levels and should have broad application, for several reasons. First, although this method was developed out of necessity due to the small number of BEC available for analysis, it has proven to be far easier and more reproducible than Northern assays for cell lines as well. Second, since only 1 ng of RNA is necessary for each assay, it is far more sensitive than RNase assays, in which 1- to 10-µg samples of total RNA are commonly used. Third, it is practical to incorporate a virtually unlimited number of CTs into the CT mixture for simultaneous measurement of gene expression. There are CTs for over 60 genes in our mixture at present and it is planned add to these as new ones are prepared.

In summary, CYP1A1 and CYP1B1 expression are each: (1) present in the BEC of all smokers, (2) highly variable relative to mEH expression, (3) independently regulated, (4) expressed in BEC with sufficient inter-individual variation to serve as indicators of inhaled pro-carcinogen exposure and/or risk for bronchogenic carcinoma, (5) expressed at lower levels in AM compared with BEC, and (6) not induced in AM of smokers. The large inter-individual variation in the observed ratio of CYP1A1/ CYP1B1 to mEH expression may contribute significantly to inter-individual variation in risk. Because CYP1A1 and CYP1B1 are highly inducible, they may serve as markers for exposure to inhaled carcinogenic PAHs. Correlating CYP1A1/CYP1B1 expression in lymphocytes with that in BEC following PAH exposure may provide a simpler measure.

The validity of the quantitative RT-PCR methods described here as relevant measures of XME gene expression were supported by two sets of data. First, in the human lymphoblastoid cell lines genetically engineered to express human XME genes at different defined levels, CYP1A1 or NADPH OR RNA level measured by RT-PCR correlates well with expression at the protein level as measured by enzyme activity. While it is possible that there are differences in this relationship in BEC compared with the lymphoblastoid cell lines, similar findings have been reported by other investigators (43). Second, the levels of CYP2B7 and NADPH OR expression reported here for BEC obtained by RT-PCR were close to those obtained in whole lung tissue by RNase protection assay in a previous study (42), although the range of inter-individual variation in expression was greater.

Efforts are underway to correlate inter-individual variation in CYP1A1, CYP1B1, mEH, and CYP1A1/CYP1B1/ mEH levels in BEC with adduct formation and mutagenicity among the same samples reported on here. This will provide more direct information on the role inter-individual variation in XME expression may play in the risk for bronchogenic carcinoma.