Identification of Benzo(a)pyrene Diol Epoxide-binding DNA Fragments Using DNA Immunoprecipitation Technique¹

Esophageal cancer remains a major health problem worldwide (1, 2). In the United States, an estimated 13,100 new cases will be diagnosed, which accounts for only 1% of all diagnosed cancers; however, an estimated 12,600 deaths in 2002 make it the seventh leading cause of cancer deaths in American men (3). Morphologically, esophageal cancer is mainly of two types, squamous cell carcinoma and adenocarcinoma (4). Cigarette smoking and alcohol consumption are significant risk factors for both histological types of the cancer (4, 5).

Tobacco is recognized as an important cause of human cancer. It contains at least 2,550 known compounds, more than 40 of which have been found to be carcinogenic in animals (6). These carcinogens are believed to induce DNA adducts and then cause gene mutation, deletion, or DNA methylation in target organs (6). For example, the tobacco carcinogen benzo(a)pyrene undergoes metabolic biotransformation to electrophilic intermediates (BPDE³ is an important one) that react with cellular macromolecules, forming DNA adducts, which are carcinogen metabolites bound covalently to DNA, usually at guanine or adenine residues (7, 8). Major advances over the last two decades in the understanding of DNA adduct structure and its consequences indicate that if DNA adducts escape cellular repair mechanisms and persist, they may lead to miscoding and result in a permanent mutation or gene deletion (7, 8). Cells with damaged DNA may be removed by apoptosis (9). Alternatively, if a permanent mutation occurs in a critical region, an oncogene may be activated, or a tumor suppressor gene may be deactivated. Multiple events of this type lead to aberrant cells with loss of normal growth control and, ultimately, to cancer (7, 8). Although several studies have reported direct evidence of tobacco carcinogen-targeted genes, i.e., hot spot mutations in k-ras and p53 genes that are frequently mutated in esophageal cancer (10, 11, 12, 13, 14), the precise molecular sequence of these events has not been well defined. Therefore, cloning and identification of BPDE-binding DNA fragments will help us to better understand the mechanisms by which tobacco carcinogens induce human cancers, and these genes may serve as biomarkers in future clinical studies.

Normal esophageal epithelial cells (HEEC₂), obtained in collaboration with Dr. Yutaka Shimada, were plated in tissue culture dishes and grown in keratinocyte serum-free medium containing 2.5 μg of epidermal growth factor and 25 μg of bovine pituitary extract (Gibco-Invitrogen). Esophageal cancer cell lines TE-7 and TE-12 were grown in DMEM with 10% fetal bovine serum at 37°C in a humidified atmosphere of 95% air and 5% CO₂. The cells were plated for 24 h in regular medium. This medium was then replaced with either control medium (containing 1 μl of tetrahydrofuran) or 1 μm BPDE (Midwest Research Institute, Kansas City, MO), in which the cells were treated for 12 h, based on our previous BPDE studies (15). BPDE was dissolved in tetrahydrofuran (Sigma Chemical Co., St. Louis, MO; stock solution of 10 mm) diluted into the medium before each experiment.

After the cells were treated with and without BPDE, genomic DNA was extracted using a previously described method (16). Briefly, the cells were scraped from cell culture dishes and pelleted by centrifugation at 1000 rpm. The cell pellets were digested with 100 μg/ml proteinase K at 50°C overnight, and genomic DNA was extracted using phenol/chloroform/isoamyl alcohol. The genomic DNA was then digested with HindIII and BamHI and then further cleaned using phenol/chloroform/isoamyl alcohol.

The DNA was immunoprecipitated using mouse monoclonal anti-BPDE antibodies and MagnaBind goat antimouse IgG beads. Briefly, 10 μg of each mouse monoclonal anti-BPDE antibody (clone 5D11 and 8E11) from PharMingen (San Diego, CA) were added to 1 ml of MagnaBind goat antimouse IgG beads (second antibody) from Pierce Chemical Co. (Rockford, IL). This was incubated at 4°C for 4 h and washed out unbounded first antibody with PBS (pH 7.2) for 5 times using magnetic force to retain first and second antibody conjugate. This first and second antibody conjugate was resuspended in 1 ml of PBS and then added to the above digested genomic DNA and incubated and rotated at 4°C overnight. The mixture was then washed and magnetically separated from unbound DNA by washing with PBS five times at 10 min each. This DNA immunocomplex (BPDE-binding DNA) was collected and heated at 65°C for 10 min, digested with proteinase K at 50°C for 3 h to remove antibodies, and finally cleaned by phenol/chloroform/isoamyl alcohol extraction. The DNA was subjected to subcloning as described in the following section.

The BPDE-binding DNA fragments were cloned into pBluescript vector at the HindIII, BamHI, or HindIII/BamHI site and amplified in Escherichia coli. The positive clones (those with the insert ≥ 300 bp) were sequenced in our institutional sequencing core facility.

Total cellular RNA was isolated from the control and BPDE-treated esophageal cells using Tri-reagent (MRC, Inc., Cincinnati, OH). For Northern blot analysis, 30 μg of total cellular RNA were fractionated on 1.2% formaldehyde agarose gels, stained with ethidium bromide, and transferred in 10× SSC to nylon filters by capillary transfer. cDNA fragments of clones 32 and 70 were used for the probe for Northern blotting. The probe was labeled with ³²P to a specific activity of approximately 2 × 10⁹ dpm/μg using random hexanucleotides as primer. The filters were prehybridized and hybridized at 68°C in Rapid-hyb Buffer (Amersham-Pharmacia, Arlington Heights, IL) with probes used at 10⁷ cpm/filter. The filters were washed with 2× SSC and 0.1% SDS for 10 min at 23°C and with 0.1× SSC and 0.1% SDS for 20 min at 68°C and then exposed to Hyperfilm-MP film (Amersham-Pharmacia) for autoradiography (17).

A previously described method of nonradioactive in situ hybridization was used (17, 18). The quality and specificity of the digoxigenin-labeled antisense and sense riboprobes were determined using Northern blotting, and the specificity of the binding of antisense riboprobes was verified using negative control sections. The stained sections were reviewed and scored with an Olympus microscope. The sections were assigned scores of negative, weak positive, positive, and strong positive staining as 0, 1, 2, and 3, respectively, according to their staining intensity.

To clone BPDE-DNA adduct-related genes (BPDE-binding DNA fragments), we first treated normal esophageal epithelial cells with 1 μm BPDE for 12 h, extracted genomic DNA from these cells, and digested 100 μg of the DNA with HindIII and BamHI endonuclease. BPDE-binding DNA fragments were then immunoprecipitated from the above BPDE-treated genomic DNA by two anti-BPDE antibodies and subcloned into pBluescript cloning vector at the HindIII, BamHI, or HindIII/BamHI site. More than 100 clones were obtained from BPDE-treated cells on LB plates, whereas the control experiment (without BPDE treatment) produced only 4 clones, which had no sequence similarity to clones isolated from BPDE-treated cells (see sequencing and blast search results below). This suggests nonspecific pull-down by the antibodies in the control cells. We next identified more than 70 BPDE-binding DNA fragments of sizes arranging from 200 bp to 3 kb (Fig. 1). We sequenced the DNA fragments with a size ≥300 bp and obtained 48 clones from this experiment. To confirm the reproducibility of this procedure, we then treated the esophageal cancer cell line TE-7 with 1 μm BPDE for 12 h and used 100 μg of the isolated genomic DNA to repeat these experiments. We had 26 clones sequenced, 7 of which were identical to clones in the previous experiment. This represents a high reproducibility, considering the normal versus tumor cell origin of the two samples.

Seventy-four fragments of BPDE-binding genes were obtained from the two experiments described above. A blast search was then conducted on the National Center for Biology Information web site⁴ for DNA sequence matches. Seven fragments in the normal cells were identical to 7 in the TE-7 cells, leaving 67 different fragments. The data on the 67 fragments fell into four categories: (a) 15 clones matched known gene sequences (two examples are shown in Fig. 2); (b) 24 matched EST clones (one example is shown in Fig. 3); (c) 22 matched genomic sequences of unknown genes (one example is shown in Fig. 4); and (d) 6 clones did not show any homology with GenBank sequences known to date. As summarized in Table 1, these BPDE-binding DNA fragments belong to DNA repair or apoptosis-related genes, zinc finger protein, cellular enzymes, EST clones, or CpG islands.

Clones HEC-32 and TE7-70 were selected for Northern blotting analyses because they were the only EST clones that were immunoprecipitated in both normal and malignant esophageal cells after BPDE exposure. Normal (HEEC₂) and malignant (TE-7 and TE-12) esophageal cells were treated with or without 1 μm BPDE for 12 h, and total RNA was isolated from these cells. Northern blotting analysis showed a differential expression of these BPDE-binding cDNAs in BPDE-treated (versus untreated control) esophageal cells (normal and malignant), indicating that BPDE reduced the expression of certain genes (Fig. 5).

Using clone HEC-32 and TE7-70 cDNA as templates, digoxigenin probes were labeled for in situ hybridization. Ten pairs of normal and malignant esophageal tissue specimens and non-small cell lung cancer specimens were hybridized with these two probes. Differential expression was found between normal and tumor specimens (Fig. 6). These findings raise the possibility that HEC-32 and TE7-70 may be cDNA fragments of tumor suppressor genes, which will be evaluated in the future by full-length cDNA cloning and characterization.

We were able to identify and clone BPDE-DNA adduct DNA fragments using our newly developed DNA IP technique. This may provide a new tool to study the mechanisms by which tobacco carcinogens cause human carcinogenesis. In 1761, Dr. John Hill first reported that tobacco use can increase human cancer risk (19). In the 1950s and 1960s, several epidemiological and laboratory studies demonstrated that tobacco smoking was associated with lung cancers (20). Subsequently, a number of studies from all parts of the world have demonstrated that tobacco use is a significant risk factor for esophageal cancer (1, 2, 4, 5, 6, 20). The United States Surgeon General and IARC have documented a dose-response relationship between tobacco use and the development of human cancers, including lung and esophageal cancers (20, 21). It has been estimated that tobacco smoking causes 85–90% of lung cancer deaths and about 30% of all cancer deaths in the United States (22).

Induction of human cancer by tobacco carcinogens is thought to be due to their ability to form DNA adducts and cause gene loss, mutation, or DNA methylation. The differential susceptibility of tobacco users to tobacco carcinogens may be due to genetic differences in DNA repair capacity and in the metabolism of tobacco carcinogens (8, 23). Detection of DNA adducts may also help us to understand adduct formation, structure, persistence, and repair, which may contribute to the elucidation of the mechanisms of tobacco carcinogenesis (8, 23). To date, there has been no systematic investigation of BPDE-binding genes. Several studies, however, have examined BPDE-induced mutations in k-ras and p53 genes, which frequently are mutated in esophageal cancer (10, 11, 12, 13, 14), and we and other investigators have started to analyze BPDE-regulated genes (15, 24, 25). Unlike these recent previous studies, which analyzed both direct and indirect effects of BPDE, our current study developed a method that allows the analysis of direct BPDE effects on gene expression.

In summary, we have developed a new technique to identify genes that bind the tobacco carcinogen BPDE. The ability to study molecular alterations in esophageal epithelial cells exposed to BPDE will help us to further elucidate the molecular basis of tobacco-related carcinogenesis and identify important new risk and response biomarkers for future clinical trials in the setting of tobacco-related cancer prevention. A major next step in this area of research will be the full-length cloning and characterization of specific BPDE-binding genes so as to understand their functions and roles in esophageal carcinogenesis.

We thank Hong Wu (University of Texas M.D. Anderson Cancer Center, Houston, TX) for excellent technical assistance and Kendall Morse (University of Texas M.D. Anderson Cancer Center, Houston, TX) for excellent for editorial assistance.