Aberrant CpG island methylation of the p16INK4a ... - Semantic Scholar

3 downloads 0 Views 240KB Size Report
activated carcinogens form a DNA adduct at CpG sites in codons 157, 248 and 273 of the p53 gene, the major sites of mutation in this gene for lung cancer (2).
Carcinogenesis vol.23 no.2 pp.335–339, 2002

Aberrant CpG island methylation of the p16INK4a and estrogen receptor genes in rat lung tumors induced by particulate carcinogens

Steven A.Belinsky1,4, Sheila S.Snow1, Kristen J.Nikula2, Gregory L.Finch3, Carmen S.Tellez1 and William A.Palmisano1 1Lovelace

Respiratory Research Institute, Lung Cancer Program, 2425 Ridgecrest Drive SE, Albuquerque, NM 87108, USA

Present addresses: 2Pharmacia, St. Louis, MO 63137 and 3Pfizer, Groton, CT 06340 4To

whom correspondence should be addressed Email: [email protected]

Recent studies by our laboratory indicate that the p16INK4a gene is frequently methylated in lung tumors induced by genotoxic carcinogens and that the frequency for methylation of the estrogen receptor α (ER) gene varies as a function of carcinogenic exposure. The purpose of the current investigation was to define the role of these two genes in lung tumors induced by the particulate carcinogens carbon black (CB), diesel exhaust (DE) or beryllium metal. Methylation of p16 was observed in 59 and 46% of DE and CB tumors, respectively. In contrast, the ER gene was inactivated in only 15% of DE or CB tumors. Methylation of the p16 and ER genes was very common (80 and 50%, respectively) in beryllium-induced lung tumors; both genes were methylated in 40% of the tumors. Bisulfite sequencing revealed dense methylation throughout exon 1 of the ER gene. The inhibitory effect of methylation on gene transcription was confirmed through RT–PCR expression studies in which p16 gene expression was 30–60-fold lower in methylated than unmethylated tumors. Residual expression in methylated tumors was consistent with contamination by stromal and inflammatory cells. Results indicate that tumors induced by these particulate carcinogens arise, in part, through inactivation of the p16 and ER genes. Furthermore, the inactivation of the p16 gene by these carcinogenic exposures supports a possible role for oxidative stress and inflammation in the etiology of human lung cancer.

Introduction Significant advances have been made over the past two decades in identifying genetic and epigenetic changes in the development of non-small lung cancer. However, the complexity of tobacco smoke, which contains over 3500 compounds (1), makes it difficult to directly link any of these carcinogens to a specific gene dysfunction. Two of the most likely candidates are the genotoxic carcinogens benzo[a] pyrene and 4-methylnitrosamino-1-(3-pyridlyl)-1-butanone Abbreviations: ER, estrogen receptor α; p16, p16INK4a; CB, carbon black; DE, diesel exhaust; H&E, hematoxylin and eosin; MSP, methylation-specific PCR; NNK, 4-methylnitrosamino-1-(3-pyridlyl)-1-butanone; NSCLC, nonsmall cell lung cancer; RT-PCR, reverse transcriptase-polymerase chain reaction. © Oxford University Press

(NNK) (1). The benzo[a])pyrene diol expoxide and other activated carcinogens form a DNA adduct at CpG sites in codons 157, 248 and 273 of the p53 gene, the major sites of mutation in this gene for lung cancer (2). In addition, inactivation of the p16 gene by aberrant promoter hypermethylation has been detected in non-small cell lung cancers (NSCLCs) induced in the rat by NNK and associated with tobacco in humans (3–5). In addition to direct damage to DNA by carcinogens, tobacco smoke also elicits a marked inflammatory response that involves the recruitment of neutrophils and alveolar macrophages (6). This response may be mediated by nitric oxide within the particulate phase of cigarette smoke and/or by direct effects of the tar on the bronchial epithelium to release chemokines such as interleukin 8, which leads to recruitment of inflammatory mediators (7,8). This process generates reactive oxygen species that can lead to DNA nicking and single-strand breaks (9,10). A role for oxidative damage in affecting specific gene targets in lung cancer has not been substantiated because of the wide spectrum of DNA damage that may ensue and the lack of studies with a comparative animal model. The rat has proven to be a good model in which to evaluate the effects of chronic inflammation and subsequent oxidative damage on lung cancer development. The particulate carcinogens diesel exhaust (DE), carbon black (CB) and beryllium metal all elicit a chronic inflammatory response and induce lung cancer in the rat through mechanisms that could also be involved in tobacco-associated lung cancer (11,12). A single inhalation exposure of beryllium metal results in the appearance of tumors by 14 months after exposure and a 64% incidence of lung tumors over the lifetime of the rat (11). In contrast, chronic exposure to either DE or CB results in a tumor incidence of ~18% with the majority of tumors appearing after 18 months of exposure (12). Our previous studies have indicated that lung tumors induced by these carcinogens do not arise through mutation of either the K-ras or p53 genes that are commonly altered in human NSCLC (12,13). However, recent studies by our group and others indicate that an epigenetic mechanism of gene inactivation, aberrant promoter hypermethylation of regulatory genes, occurs in human NSCLC and is recapitulated in rat lung tumors induced by genotoxic carcinogens (3,14,15). Aberrant methylation of the p16 gene is common (65–90%) in rat lung tumors induced by NNK and radiation, while the frequency for methylation of the estrogen receptor α (ER) gene that has growth suppressive properties in multiple tumor types varies significantly (20–90%) depending on the carcinogenic exposure (16). The purpose of the current study was to determine whether the induction of lung tumors by the particulate carcinogens DE, CB and beryllium metal in the rat involves inactivation of the p16 and ER genes. Materials and methods Tumor induction Details of the exposure conditions have been described (11–13). Briefly, for DE and CB, male and female F344/N rats (7–9 weeks old) were exposed

335

S.A.Belinsky et al. chronically (16 h/day, 5 days/week, 24 months) by inhalation to one of two particle concentrations or to filtered air (sham control). Means (⫾ standard error) of weekly mean particle concentrations from daily filter samples were 2.4 ⫾ 0.02 and 6.33 ⫾ 0.04 mg/m3 for the DE exposure chambers, and 2.46 ⫾ 0.03 and 6.55 ⫾ 0.06 mg/m3 for the CB chambers. Twelve-week-old F344/N rats received a single, nose-only exposure to a beryllium metal aerosol. Four different exposure levels were evaluated, and group mean lung burdens were 40, 110, 360 and 430 µg. Lung tumors were induced by DE, CB and beryllium metal at all exposure concentrations and contributed to the sample set being used for the current study. Lung neoplasms were fixed in 4% buffered paraformaldehyde or in 10% neutral buffered formalin. These samples were embedded in paraffin, cut at 5 µm, stained with hematoxylin and eosin (H&E) and examined by light microscopy for histologic diagnosis. For the DE and CB tumors, serial sections were also cut for DNA analysis. For the beryllium-induced tumors, a small portion was also frozen in liquid nitrogen and stored at –80°C. Microdissection of tumors and DNA isolation The frozen beryllium-induced tumors were not microdissected. Sequential sections (5 µm) were prepared from tumors induced by DE or CB, deparaffinized and stained with toluidine blue to facilitate dissection. A 25 gauge needle attached to a tuberculin syringe was used to remove the lesions under a dissecting microscope. Because some tumors are contaminated with normal tissue or are very small, it was essential to include normal appearing cells to ensure that after bisulfite modification and column clean up of the DNA template, enough sample remained to conduct the methylation-specific PCR (MSP) assay as described below. Thus, because the goal of the study was to determine whether p16 methylation was present and not to quantify methylation levels, microdissection was used to enrich the samples examined. The sensitivity (one methylated allele in 1000 unmethylated alleles) of the MSP assay that is used to assess methylation state is amenable to specimens where the isolated DNA is from both tumor and contaminating normal tissue. DNA was isolated from frozen tumors and microdissected tumors by digestion with Pronase in sodium dodecyl sulfate (1%), followed by standard phenol– chloroform extraction and ethanol precipitation. RNA isolation and reverse transcriptase–polymerase chain reaction (RT–PCR) RNA was isolated by the TRIZOL™ method (Life Technologies, Gaithersburg, MD) from beryllium-induced lung tumors. Contaminating DNA was removed by digestion with DNase I (Life Technologies) followed by phenol extraction and ethanol precipitation. RNA was quantified spectrophotometrically at 260 nm. First-strand cDNA was generated at 42°C from 3 µg of total RNA using a SuperScript™ kit (Life Technologies). RT minus controls, where the RT enzyme was omitted, were also generated and analyzed for each sample. PCR primers described previously (14) were designed to cross exon splice junctions. Amplification products comprising a portion of exons 1 and 2 for p16 and exons 1 and 2 for β-actin were generated from the cDNA template in parallel PCRs. The amount of cDNA used for the PCR reaction was serially diluted, and 0.13, 0.25, 0.5 and 1.0 µl (from a 20 µl RT–PCR reaction) was used for amplification of the p16 and β-actin transcripts. PCR conditions have been described (14). PCR products were analyzed on 2% agarose gels by visualization with ethidium bromide. The relative amount of p16 transcript to β-actin transcript was compared over the range of cDNA template by densitometry. All results were reproduced in three separate experiments. MSP and bisulfite sequencing The methylation status of the p16 and ER genes was determined by MSP (3,14). Primer sequences and PCR conditions have been described for the p16 gene (14). ER primers designed to recognize unmethylated alleles were as follows: forward, 5⬘-TGTTGTTGTTGTGTGGTTGTTGG-3⬘; reverse, 5⬘-TC AACAAACTAAACAACACAC-3⬘. The sequence for methylated primers was as follows: forward, 5⬘-TTAATTATTTCGAGGGCGTC-3; reverse, 5⬘AAAACGACGACACGTACGAC-3⬘. The PCR amplification protocol was as follows: denature at 95°C for 10 min, then denature at 95°C for 15 s, anneal at either 56°C (unmethylated primers) or 61°C (methylated primers) for 15 s, extension at 72°C for 15 s for 40 cycles followed by a final extension at 72°C for 5 min. Primers were localized to regions in and around the transcription start site of the p16 and ER genes, a region that correlates with loss of gene expression (14,16). PCR amplification was performed using ~250 ng of bisulfite-modified DNA as the template. Normal lung DNA was used as a control in all PCRs. The products were visualized on 2% agarose gels. Amplification of the ER gene using DNA from DE or CB often did not result in any PCR product. This is probably due to the DNA degradation associated with formalin fixation and the resulting low concentration of amplifiable DNA. This was not a problem for amplification of p16. Therefore, we developed a nested, two-stage PCR approach for assaying the methylation state of the ER gene as described (17). Outer PCR primers were designed to

336

Table I. Prevalence for p16 and ER methylation in lung tumors induced by particulate carcinogens Exposure

P16

ER

DE CB Beryllium metala

13/22 (59) 10/22 (45) 16/20 (80)

3/22 (14) 3/22 (14) 10/20* (50)

Prevalences are expressed as the number of tumors positive for methylation/ total number of tumors analyzed. Percentage positive is indicated in parentheses. aFour tumors induced by beryllium metal appeared to be partially methylated at the p16 locus. *P ⬍ 0.05 when comparing methylation prevalence in beryllium-induced tumors to DE- or CB-induced tumors.

recognize the bisulfite-modified template, but not to discriminate between methylated and unmethylated alleles. Primer sequences were as follows: forward, 5⬘-GATGTTTATGGAGAGGGTTTG-3; reverse, 5⬘-CAAATAATAAAACACCTAATAACC-3⬘. The PCR amplification protocol was as follows: denature at 95°C for 10 min, then denature at 95°C for 30 s, anneal at 54°C for 30 s, extension at 72°C for 30 s for 40 cycles followed by a final extension at 72°C for 5 min. An aliquot (1 µl) from the first PCR was then used in a second-stage PCR that employed either unmethylated or methylated MSP primers as described above. Exon 1 of the ER from normal lung tissue and from three berylliuminduced lung tumors shown to be positive for methylation was amplified using ER MSP primers specific for unmethylated or methylated bisulfite-modified DNA, respectively. The PCR products were ligated into the PCR II vector using the TA cloning kit (Invitrogen, San Diego, CA). Four to five clones from each sample were sequenced.

Results Methylation of p16 Sixty-four lung tumors induced by DE (n ⫽ 22), CB (n ⫽ 22) or beryllium metal (n ⫽ 20) were available for analysis of gene-specific promoter methylation. Three tumors from each exposure group were classified as squamous cell carcinomas, while the remaining tumors were adenocarcinomas. Methylation of p16 was detected in 13 of 22 and 10 of 22 tumors induced by DE and CB, respectively (Table I). Unmethylated alleles were also simultaneously detected in these samples because the tumors were contaminated with stromal and inflammatory cells (contamination ranged from 5 to 70%). Only unmethylated alleles were detected in DNA isolated from sham-exposed rat lung (not shown). Methylated p16 was very common in lung tumors induced by beryllium metal, being present in 16 of 20 tumors (Figure 1A; Table I). p16 methylation was detected in seven of nine squamous cell carcinomas available for analysis from the three different carcinogenic exposures. Interestingly, four tumors induced by beryllium metal (nos 7, 10, 15 and 17) exhibited relatively faint amplification by the MSP primers specific for methylation, suggesting that methylation of p16 may be only limited in these tumors. As a result of the lack of homology for exon 1α between the human and rat p16 gene (14), commercial antibodies to human p16 protein are not reactive to the rat p16 protein. To determine whether the intensity of the methylated PCR product correlated inversely with the level of p16 transcript, semiquantitative RT–PCR was conducted. Several beryllium metal-induced tumor samples, showing strong methylationspecific amplification product (nos 2, 3, 13 and 20), weak product (no. 15) or no product (nos 9 and 16), were selected for analysis. All tumors selected exhibited ⬍25% contamination with stromal and inflammatory cells based on examination

Promoter hypermethylation in particle-induced cancer

Fig. 1. (A) Methylation of the p16 gene in beryllium metal-induced rat lung tumors. Bisulfite-modified DNA from lung tumors (lanes 1–20) and normal rat lung (lane 21) was amplified using MSP primers for the p16 gene. Strongly methylated PCR products are present in lanes 2–4, 6, 8, 11–14 and 18–20. Weaker products are seen in lanes 7, 10, 15 and 17. As expected, no methylated p16 alleles were detected in DNA from normal lung (lane 21). (B) Semi-quantification of p16 expression. The expression of p16 in tumors 3, 9, 13 and 15 from (A) were compared by titrating the amount of cDNA template in the PCR reaction. The relative amount of p16 product to β-actin product was compared over this range by densitometry.

Fig. 2. Methylation of the ER gene in lung tumors induced by DE and CB. Bisulfite-modified DNA from lung tumors induced by DE (lanes 1–10), CB (11–20) and normal fixed lung DNA (lane 22) was amplified using MSP primers that recognize either unmethylated or methylated alleles. Methylation of the ER gene was detected in DNA from lanes 1, 4, 8, 13, 17 and 18. Unmethylated alleles were present in all samples except lane 4.

of H&E stained tumor sections. Shown in Figure 1B are representative results from analyses of four tumors (nos 15, 9, 3 and 13). Expression of p16 was compared across samples by titrating the amount of cDNA template (1, 0.5, 0.25 and 0.13 µl) in the PCR reaction. Expression was quantified by densitometric comparison between β-actin and p16 levels as expression of both genes became linear. Robust expression of p16 was detected in tumors 15 (minimal p16 methylation product) and 9 (no p16 methylation product) and was still apparent even with the lowest amount of cDNA template. In contrast, expression of p16 was near limits of detection with the highest amount of cDNA in both tumors (nos 3 and 13) where methylated alleles were readily detected by MSP (Figure 1A and B). Levels of p16 transcript in the two tumors with an unmethylated or weakly methylated p16 gene were 28– 60-fold greater than in tumors where methylation appeared abundant. A similar relationship of p16 transcript levels to p16 methylation status was also seen for tumor samples 2, 16 and 20 (not shown). The minimal expression seen in tumors methylated for the p16 gene is probably from contaminating stromal and inflammatory cells as seen in human tumors stained immunohistochemically for this protein (5). Methylation of ER Methylation of the ER was infrequent in lung tumors induced by either DE or CB. It was seen in three tumors associated with each exposure (Figure 2, Table I). As expected, using the nested, two-stage PCR approach, unmethylated alleles were also detected in the majority of tumors positive for ER

methylation reflecting the contamination with stromal and inflammatory cells. In contrast, 50% of tumors induced by beryllium were methylated at the ER locus (Table I). Previous studies by our laboratory had used Southern analysis following NotI restriction digestion to assay for methylation of the ER gene (16). The MSP primers used in the current study amplify the same portion of exon 1 that encompasses the two NotI sites assayed previously and shown to correlate with loss of transcription. To determine the density of methylation within this region, bisulfite sequencing was conducted on normal lung DNA and DNA from three methylated, beryllium-induced tumors. Twenty-one CpG sites (NotI methylation-sensitive sites at CpG nos 8 and 17) were present within the MSP amplification product. No methylation was detected within this region in DNA from normal lung tissue (Figure 3). Two of the tumors (nos 4 and 12) showed extensive methylation with ~80% of all CpG sites methylated within the clones (Figure 3). More variable methylation was seen for tumor 6 across clones, and overall 33% of CpG sites were methylated. Discussion Findings from this study implicate inactivation of the p16 gene by promoter hypermethylation as a likely contributing factor to the development of lung cancer associated with exposure to the particulate carcinogens, DE, CB and beryllium metal. Methylation changes in p16 correlated directly with loss of gene transcription, a finding consistent with previous studies in radiation-induced, rat lung tumor-derived cell lines (14). 337

S.A.Belinsky et al.

Fig. 3. Bisulfite sequencing of exon 1 of the ER gene from berylliuminduced lung tumors. PCR primers specific to unmethylated and methylated bisulfite-modified DNA were used to amplify exon 1 of the ER gene in normal lung or tumor tissue, respectively. Amplified PCR products were cloned, and four to five individual clones were sequenced to determine the methylation state of the 21 CpGs within the amplified region. Open and filled circles indicate that the CpG site is unmethylated or methylated, respectively. Each row represents one clone.

Inactivation of the ER varied as a function of carcinogenic exposure, being prevalent in tumors induced by beryllium metal, but infrequent in tumors induced by either DE or CB. Methylation was generally dense throughout exon 1 of the ER gene, a region that correlates with loss of gene transcription in lung and breast tumors (16,18). The inactivation of the p16 gene by these particulate carcinogens supports a possible role for oxidative stress and inflammation in the initiation of lung cancer. Non-toxic doses of suspended DE particles have stimulated IL-8 and granulocyte macrophage colony stimulating factor production by human bronchial epithelial cells through activation of NF-kappaB (8). Inflammatory cytokines IL-1β, IL-6 and TNF-α were also induced when bronchial epithelial cells were co-cultured with blood monocytes and exposed to particulate matter in the form of soot particle FR 101 (19). In addition, protein tyrosine kinase activity and NF-kappaB protein activity were increased in these bronchial epithelial cells. The activation of these pathways in persons smoking cigarettes has also been demonstrated. The inflammatory mediators IL-8 and IL-6 were present in higher concentrations in bronchoalveolar lavage from smokers than never-smokers (20). The expression of IL-8 was also increased in epithelial cells collected from the small airways of smokers compared with never-smokers (21). Together these studies suggest that activation of inflammatory mediators may contribute to cancer initiation through oxidative stress and the formation of 8-hydroxyguanosine DNA adducts in rodent and human lung tissue associated with exposure to both DE and cigarette smoke (22–25). 338

One important question is how the p16 gene is targeted for inactivation in lung cancer. The current study and our previous studies indicate that lung tumors induced by six different carcinogenic exposures, many of which damage DNA through different mechanisms, induce lung cancer that often (45–90% of tumors from a specific exposure) involves aberrant promoter hypermethylation of the p16 gene (3,14,15). The metabolism of NNK leads to the generation of methyl and pyridyloxobutyl DNA adducts, while CB, DE and beryllium metal elicit inflammatory mediators that result in oxidative stress that can damage DNA directly via free radicals or through the formation of 8-hydroxyguanosine DNA adducts (1,24,25). The result of these carcinogenic exposures can be DNA damage manifested primarily as single-strand breaks (23,26). In contrast, plutonium and, to a lesser extent, X-rays appear to cause double-stranded DNA strand breaks in part through formation of oxygen radicals (27). Models of global loss of cytosine methylation and region-specific gain in methylation during malignant transformation involve changes in the methylation status of both the gene-promoter region and the surrounding repetitive DNA. During malignant transformation, essential factors that guide the cytosine-DNA methyltransferases to regions that are to be methylated may be disrupted, and other regions are selected for de novo methylation. The result of this process is the loss of methylation within the repetitive DNA, methylation of CpG sites within the promoter region of the gene, and the formation of chromatin complexes that impede transcription (28). DNA damage by tobacco carcinogens could be one factor leading to regional disruption of normal chromatin structure. This change in genome integrity could lead to inappropriate methylation by the cytosine DNA methyltransferases of CpG sites in the normally protected promoter region of genes such as p16. While such a scenario is clearly speculative, the p16 gene is a paradigm for which such mechanisms could ultimately be elucidated. The prevalence of ER gene hypermethylation in lung tumors induced by particulate carcinogens varied significantly with carcinogenic exposure, being common in beryllium-induced tumors and rare in tumors induced by DE or CB. These findings parallel our previous results where ER methylation was frequent in plutonium-induced tumors and occurred in ⬍25% of NNK-induced rodent tumors and lung tumors from smokers (16). In addition to its involvement in the pathogenesis of breast cancer, methylation of the ER gene has been linked to colon, prostate and hematopoietic malignancies (18,29,30). While methylation of the ER gene is associated with aging in colon cancer, the lung epithelium appears to behave similarly to the breast epithelium where normal cells are unmethylated, and only a subset of tumors shows ER methylation (18). The selectivity for inactivation of the ER gene by specific carcinogenic exposures clearly differs from the common targeting of the p16 gene for methylation by the same carcinogens. A common link as to how plutonium and beryllium target the ER gene for hypermethylation is not readily apparent. Plutonium irradiation of hematopoietic cells has produced a high frequency of non-clonal aberrations in clonal descendents that resulted in the transmission of chromosomal instability to their progeny (31). While similar studies have not been conducted with beryllium, it does interact with non-catalytic sites on DNA polymerase α to alter synthetic fidelity (32,33) and to block cell proliferation prior to DNA replication during G1 phase by modifying normal transcription controls (34). The disruption of replication timing that could ensue because of

Promoter hypermethylation in particle-induced cancer

chromosomal instability or loss of normal transcriptional controls may be a mechanism by which normally unmethylated regions of DNA become aberrantly methylated (35). Thus, the contribution of direct damage to the DNA template in the form of single- and double-strand breaks to affect chromatin structure and the modulation of cellular processes that disrupt normal replication timing may ultimately both prove important in the targeting of gene promoters for aberrant CpG island hypermethylation. Acknowledgement Supported by NIEHS ES08801 under US DOE Cooperative Agreement No. DE-FC04-96AL-76406.

References 1. Hecht,S.S. (1999) Tobacco smoke carcinogens and lung cancer. J. Natl Cancer Inst., 91, 1194–1210. 2. Smith,I.E., Denissenko,M.F., Bennett,W.P., Li,H., Amin,S., Tang,M. and Pfeifer,G.P. (2000) Targeting of lung cancer mutational hotspots by polycyclic aromatic hydrocarbons. J. Natl Cancer Inst., 92, 803–811. 3. Belinsky,S.A., Nikula,K.J., Palmisano,W.A., Michels,R., Saccomanno,G., Gabrielson,E., Baylin,S.B. and Herman,J.G. (1998) Aberrant methylation of p16INK4a is an early event in lung cancer and a potential biomarker for early diagnosis. Proc. Natl Acad. Sci. USA, 95, 11891–11896. 4. Merlo,A., Herman,J.G., Mao,L., Lee,D.J., Gabrielson,E., Burger,P.C., Baylin,S.B. and Sidransky,D. (1995) 5⬘ CpG island methylation is associated with transcriptional silencing of the tumor suppressor p16/ CDKN2/MTS1 in human cancers. Nature Med., 1, 686–692. 5. Otterson,G.A., Khleif,S.N., Chen,W., Coxon,A.B. and Kaye,F.J. (1995) CDKN2 silencing in lung cancer by DNA hypermethylation and kinetics of p16INK4a protein induction by 5-aza 2⬘ deoxycytidine. Oncogene, 11, 1211–1216. 6. Cosio,M.G. and Guerassimov,A. (1999) Chronic obstructive pulmonary disease. Inflammation of small airways and lung parenchyma. Am. J. Respir. Crit. Care Med., 160, S21–S25. 7. Pryor,W.A. (1997) Cigarette smoke radicals and the role of free radicals in chemical carcinogenicity. Environ. Health Perspect., 105 (suppl. 4), 875–882. 8. Takizawa,H., Tanaka,M., Takami,K., Ohtoshi,T., Ito,K., Satoh,M., Okada,Y., Yamasawa,F. and Umeda,A. (2000) Increased expression of inflammatory mediators in small-airway epithelium from tobacco smokers. Am. J. Physiol. Lung Cell Mol. Physiol., 278, L906–L913. 9. Pryor,W.A., Stone,K., Zang,L.Y. and Bermudez,E. (1998) Fractionation of aqueous cigarette tar extracts: fractions that contain the tar radical cause DNA damage. Chem. Res. Toxicol., 11, 441–448. 10. Yoshie,Y. and Ohshima,H. (1997) Synergistic induction of DNA strand breakage by cigarette tar and nitric oxide. Carcinogenesis, 18, 1359–1363. 11. Nikula,K.J., Snipes,M.B., Barr,E.B., Griffith,W.C., Henderson,R.F. and Mauderly,J.L. (1995) Comparative pulmonary toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam. Appl. Toxicol., 25, 80–94. 12. Nickell-Brady,C., Hahn,F.F., Finch,G.L. and Belinsky,S.A. (1994) Analysis of K-ras, p53 and c-raf-1 mutations in beryllium-induced rat lung tumors. Carcinogenesis, 15, 257–262. 13. Swafford,D.S., Nikula,K.J., Mitchell,C.E. and Belinsky,S.A. (1995) Low frequency of alterations in p53, K-ras and mdm2 in rat lung neoplasms induced by diesel exhaust or carbon black. Carcinogenesis, 16, 1215–1221. 14. Swafford,D.S., Middleton,S.K., Palmisano,W.A., Nikula,K.J., Tesfaigzi,J., Baylin,S.B., Herman,J.G. and Belinsky,S.A. (1997) Frequent aberrant methylation of p16INK4a in primary rat lung tumors. Mol. Cell. Biol., 17, 1366–1374. 15. Belinsky,S.A., Palmisano,W.A., Issa,J.-P.J., Herman,J.G. and Baylin,S.B.

(2000) Identifying cancer-susceptible individuals in a radon-exposed population. Radiat. Res., 2, 459–462. 16. Issa,J.-P.J., Baylin,S.B. and Belinsky,S.A. (1996) Methylation of the estrogen receptor CpG island in lung tumors is related to the specific type of carcinogen exposure. Cancer Res., 56, 3655–3658. 17. Palmisano,W.A., Divine,K.K., Saccomanno,G., Gilliland,F.D., Baylin,S.B., Herman,J.G. and Belinsky,S.A. (2000) Predicting lung cancer by detecting aberrant promoter methylation in sputum. Cancer Res., 60, 5954–5958. 18. Ottaviano,Y.L., Issa,J.P., Parl,F.F., Smith,H.S., Baylin,S.B. and Davidson,N.E. (1994) Methylation of the estrogen receptor gene CpG island marks loss of estrogen receptor expression in human breast cancer cells. Cancer Res., 54, 2552–2555. 19. Drumm,K., Messner,C. and Kienast,K. (1999) Reactive oxygen intermediate-release of fibre-exposed monocytes increases inflammatory cytokine-mRNA level, protein tyrosine kinase and NF-kappaB activity in co-cultured bronchial epithelial cells (BEAS-2B). Eur. J. Med. Res., 4, 257–263. 20. Kuschner,W.G., D’Alessandro,A., Wong,H. and Blanc,P.D. (1996) Dosedependent cigarette smoking-related inflammatory responses in healthy adults. Eur. Respir. J., 9, 1989–1994. 21. Takizawa,H., Ohtoshi,T., Kawasaki,S., Abe,S., Sugawara,I., Nakahara,K., Matsushima,K. and Kudoh,S. (2000) Diesel exhaust particles activate human bronchial epithelial cells to express inflammatory mediators in the airways: a review. Respirology, 5, 197–203. 22. Park,E.M., Park,Y.M. and Gwak,Y.S. (1998) Oxidative damage in tissues of rats exposed to cigarette smoke. Free Radic. Biol. Med., 25, 79–86. 23. Asami,S., Manabe,H., Miyake,J., Tsurudome,Y., Hirano,T., Yamaguchi,R., Itoh,H. and Kasai,H. (1997) Cigarette smoking induces an increase in oxidative DNA damage, 8-hydroxydeoxyguanosine, in a central site of the human lung. Carcinogenesis, 18, 1763–1766. 24. Iwai,K., Adachi,S., Takahashi,M., Moller,L., Udagawa,T., Mizuno,S. and Sugawara,L. (2000) Early oxidative DNA damages and late development of lung cancer in diesel exhaust-exposed rats. Environ. Res., 84, 255–264. 25. Tokiwa,H., Sera,N., Nakanishi,Y. and Sagai,M. (1999) 8-Hydroxyguanosine formed in human lung tissues and the association with diesel exhaust particles. Free Radic. Biol. Med., 27, 1251–1258. 26. Hoffmann,D., Rivenson,A., Chung,F.L. and Hecht,S.S. (1991) Nicotinederived N-nitrosamines (TSNA) and their relevance in tobacco carcinogenesis. Crit. Rev. Toxicol., 21, 305–311. 27. Clutton,S.M., Townsend,K.M.S., Walker,C., Ansell,J.D. and Wright,E.G. (1996) Radiation-induced genomic instability and persisting oxidative stress in primary bone marrow cultures. Carcinogenesis, 17, 1633–1639. 28. Robertson,K.D. (2001) DNA methylation, methyltransferases and cancer. Oncogene, 20, 3139–3155. 29. Issa,J.P., Ottaviano,Y.L., Celano,P., Hamilton,S.R., Davidson,N.E. and Baylin,S.B. (1994) Methylation of the oestrogen receptor CpG island links ageing and neoplasia in human colon. Nature Genet., 7, 536–540. 30. Issa,J.P.J., Zehnbauer,B.A., Civin,C.I., Collector,M.I., Sharkis,S.J., Davidson,N.E., Kaufmann,S.H. and Baylin,S.B. (1996) The estrogen receptor CpG island is methylated in most hematopoietic neoplasms. Cancer Res., 56, 973–977. 31. Kadhim,M.A., Macdonald,D.A., Goodhead,D.T., Lorimore,S.A., Marsden,S.J. and Wright,E.G. (1992) Transmission of chromosomal instability after plutonium α-particle irradiation. Nature, 355, 738–740. 32. Sirover,M.A. and Loeb,L.A. (1976) Metal-induced infidelity during DNA synthesis. Proc. Natl Acad. Sci. USA, 7, 2331–2335. 33. Seal,G., Shearman,C.W. and Loeb,L.A. (1979) On fidelity of DNA replication: studies with human placenta DNA polymerases. J. Biol. Chem., 254, 5229–5237. 34. Skilleter,D.N., Barrass,N.C. and Price,R.J. (1991) C-myc expression is maintained during the G1 phase cell cycle block produced by beryllium. Cell Prolif., 24, 229–237. 35. Rountree,M.R., Bachman,K.E., Herman,J.G. and Baylin,S.B. (2001) DNA methylation, chromatin inheritance and cancer. Oncogene, 20, 3156–3165. Received September 27, 2001; revised November 20, 2001; accepted November 27, 2001

339