Aberrant p16 promoter methylation in smokers and former smokers with nonsmall cell lung cancer

Hypermethylation of cytosines in CpG‐rich islands of the promoter regions of regulatory genes has been discovered as a common mechanism of gene silencing during carcinogenesis. We analysed 64 primary lung carcinomas for promoter methylation of the tumour suppressor genes (TSGs) p16 (p16INK4a/CDKN2A) and p14 (p14ARF) by methylation‐specific PCR, in order to evaluate aberrant methylation as a potential biomarker for epigenetic alterations in tobacco‐related lung cancer. Methylation of p16 was observed in 34% (22/64) of the lung tumours examined. In particular, p16 methylation occurred in nonsmall cell lung cancer (NSCLC) only, with 41 % (22/54) of the tumours being positive. The highest frequency was found in large cell carcinoma (5/7, 71%), followed by adenocarcinoma (9/25, 36%) and squamous cell carcinoma (7/21, 33%). Methylation of the p14 gene was less frequent in lung cancer (4/52, 8%). When association with tobacco smoking was analysed, 42% (21/50) of NSCLC from ever smokers exhibited p16 methylation. Interestingly, the analysis revealed a significantly higher risk of p16 methylation in former smokers as compared to current smokers [odds ratio (OR) 5.1; 95% confidence interval (CI) 1.3–22]. The difference was retained after adjustment for age (OR 3.7; 95% CI 0.9–17). The promoter methylation results were then combined with data on genetic alterations determined previously in the same set of tumours. This data similarly showed that p16 methylation in parallel with p53 gene mutation or p14 methylation occurred more frequently in former smokers than in current smokers (44% vs. 14%; P = 0.035). Taken together, our data suggest that analysis of promoter methylation in TSGs may provide a valuable biomarker for identification of groups with an elevated risk of cancer, such as smokers and ex‐smokers. © 2003 Wiley‐Liss, Inc.

Hypermethylation of cytosines in CpG-rich islands of the promoter regions of genes is one of the mechanisms of gene silencing during the development in mammals. 1 In cancer, hypermethylation of the promoter regions is associated with transcriptional inactivation and loss of expression of tumour suppressor and other regulatory genes, constituting an alternative, epigenetic way to the loss of gene function. 2,3 Furthermore, recent data suggest that abnormal methylation of cancer-related genes occurs in human cancer in a tumour-type and gene-specific manner. 4 The tumour suppressor gene p16 (p16 INK4a /CDKN2A) is a cell cycle regulator that is frequently inactivated in many different types of malignancies, including lung cancer. [5][6][7] It has been demonstrated that p16 gene may be transcriptionally inactivated by aberrant promoter methylation during tumorigenesis. 8 -10 The gene product, p16 protein, is a molecular component of the retinoblastoma protein (pRB) regulatory pathway, functioning as a specific inhibitor of cyclin-dependant kinases 4 and 6. 11 Inactivation of p16 enables initial phosphorylation of pRB that releases it from transcription factor and allows subsequent progression of cells into S phase. 12,13 The p14 gene (p14 ARF /ARF), an alternative reading frame of the p16 gene locus using a separate promoter region, encodes an important regulatory protein functioning in the p53 pathway. 14 p14 and p16 proteins are encoded by 2 different first exons and the encoding is continued in alternative reading frames through a common exon 2. 15 Although encoded by the same locus, both proteins act in different cell cycle inhibition pathways. p14 neutralises MDM2-dependant p53 degradation, thereby increasing p53 stability. 16 Methylation of p14 promoter has been shown to occur in human tumours. 4 Exposure to carcinogens from tobacco smoke is the primary cause of lung cancer, with occupational exposure to pulmonary carcinogens, such as asbestos, being the second major etiological factor. 17 In addition to data that indicates that various types of genetic damage are associated with exposure to mutagenic and carcinogenic components from tobacco smoke, 18 -20 recent studies have suggested an association between p16 promoter methylation and tobacco smoking in lung cancer. [21][22][23][24] In our work, we determined the frequency of p16 and p14 promoter region methylation in primary lung tumours by methylation specific PCR (MSP), a method allowing sensitive detection of the methylation status of cytosine residues of CpG islands within a gene promoter. 25 In addition, samples of nonmalignant peripheral lung tissue from the majority of the patients were also examined. We then analysed p16 tumour methylation data in regard to demographic, clinico-pathological and exposure characteristics of the cases. Finally, we analysed the promoter methylation data in combination with previously determined gene mutation data in current and former smokers.

Patients
Primary lung tumour samples (n ϭ 64) were obtained from lung cancer patients who underwent a surgical resection, as described previously. 26 For 58 of the cases, a specimen of nonmalignant peripheral lung tissue was also available for investigation. In addition, leukocyte DNAs from cancer patients (n ϭ 15) were examined. Ten of the lung cancers were diagnosed as SCLC, and 54 as NSCLC, including 23 adenocarcinoma, 21 squamous cell carcinoma, 2 adeno-squamous cell carcinoma and 7 large cell carcinoma cases. One case was a tumour carcinoid. The mean age of the patients was 63.6 Ϯ 8.7 years, 15 of them were females and 49 males. Five patients were lifelong nonsmokers, 21 former smokers (who ceased smoking at least 1 year before the diagnosis) and 38 current smokers. Smokers had consumed in average 21.3 Ϯ 8.9 cigarettes per day and had a mean cumulative exposure of 40.4 Ϯ 19.9 pack-years. Data on occupational asbestos exposure were available for 61 patient, with 22 cases exposed to asbestos at work. 26

Methylation specific PCR
Genomic DNA was extracted from frozen samples of lung tumour tissue, peripheral lung tissue and peripheral blood leukocytes using proteinase-K digestion and phenol/chloroform purification followed by ethanol precipitation. Methylation specific PCR. 25 was used for analysis of the methylation pattern in 5Ј region of the p16 and p14 genes. Briefly, 2 g of DNA in a final volume of 50 l were denatured with 2 M NaOH (Merck) for 10 min at 37°C, and after that exposed to bisulfite modification (converting unmethylated cytosines to uracils) with 3 M sodium bisulfite (Sigma Chemical Co., St. Louis, MO) plus 10 mM hydroquinone (Fluka, Milwaukee, WI) for 16 hr at 50°C. Modified DNA was purified with the Wizard DNA Clean-up System (Promega, Madison, WI) in vacuum manifold. DNA modification was completed by 3 M NaOH treatment for 5 min at room temperature. DNA was precipitated with 70% ethanol and dissolved in 40 l of sterile water. Bisulfite modified DNA was analysed for methylation immediately or after the storage at Ϫ80°C. PCR primers specific for methylated (M) and unmethylated (U) sequences within the 5Ј region of the p16 and p14 genes were synthesised according to published sequences 25,27 and used to determine methylation in promoter region. A set of PCR primers for nonmodified DNA (W) was included in study to control the bisulfite conversion of DNA. Primer sequences for p16 gene analysis were as follows: M primers, 5Ј-TTATTAGAGGGTGGGGCGGATCGC (sense) and 5Ј-CCACCTAAATCGACCTCCGACCG (antisense); U primers, 5Ј-TTATTAGAGGGTGGGGTGGATTGT (sense) and 5Ј-CCAC-CTAAATCAACCTCCAACCA (antisense); W primers, 5Ј-CAGAGGGTGGGGCGGACCGC (sense) and 5Ј-CGGGCCGCGGCCGTGG (antisense). PCR product amplified with M, as well as with U primers, was 234 bp in length. Primer sequences for p14 gene analysis were as follows: M primers, 5Ј-GTGTTAAAGGGCGGCGTAGC (sense) and 5Ј-AAAAC-CCTCACTCGCGACGA (antisense); U primers, 5Ј-TTTTTG-GTGTTAAAGGGTGGTGTAGT (sense) and 5Ј-CACA-AAAACCCTCACTCACA-ACAA (antisense); W primers, 5Ј-CTGGTGCCAAAGGGCGGCGCAGC (sense) and 5Ј-CGAAAACCCTCACTCGCGGCGG (antisense). For p14 gene, PCR product amplified with M primers was 122 bp and that with U primers was 132 bp in length.
The PCR mixture for 50 l of total reaction volume contained 2-4 l of modified DNA template, PCR buffer, 0.4 mM of each deoxynucleotide triphosphate, 2.5 mM of MgCl 2 , 1% of dimethyl sulfoxide, forward and reverse primers at the final concentration of 6 ng/l and 1.25 U of AmpliTaq Gold polymerase (Perkin-Elmer, Oak Brook, IL). Modified DNA from human urinary bladder cancer cell line T24 and from colorectal cancer cell line SW48 (both from American Type Culture Collection) were included in all the experiments as a positive control for methylation of p16 and p14, respectively. DNA from leukocytes of healthy controls (n ϭ 2) was used to control unmethylated PCR product, and water controls were also included. DNA from T24 cell line was serially diluted with unmethylated leukocyte DNA from healthy controls to determine sensitivity of the detection. PCR was performed in a thermocycler at following conditions: 10 min at 95°C for the activation of polymerase, 35 cycles at 95°C for 45 sec, 62°C for 45 sec, 72°C for 45 sec and the final extension in 72°C for 10 min. Nine microliters of each of PCR product was loaded onto nondenaturing 7.5% polyacrylamide gel and after ethidium bromide staining visualised under UV illumination.

Statistical analysis
Odds ratios (OR) and the exact or Mantel-Haenszel 95% confidence intervals (CI) for 2 binomial samples were calculated for single variables using the computer software StatXact-4 for Windows (CYTEL Software Corporation, 1998). Owing to small sample sizes, exact logistic regression analysis was carried out to adjust for the effects of the statistically significant single variables. For this, the logistic procedure of the SAS statistical analysis system was used (SAS version 8.2). Two-sided Fisher's exact test and Student's t-test were used for comparison of categorical and continuous variables, respectively. P Յ 0.05 was considered as statistically significant.

Methylation in lung tumours and nonmalignant lung tissue
To set up the methylation specific PCR, we analysed methylated sequences in the promoter regions of the p16 and p14 genes in DNAs from cancer cell lines T24 and SW48, respectively, as positive controls. Clear methylation signals were obtained for both genes by MSP. T24 DNA was diluted up to 20 times with unmethylated DNA from leukocytes, and p16 methylation was detected in the dilution series robustly and reproducibly (Fig. 1).
p16 promoter methylation was detected in 22 of the 64 (34%) primary lung carcinomas studied (Table I). Among those 58 cases for whom both tumour and nonmalignant tissue sample was available, the prevalence of p16 methylation was 31% (18/58) in the tumour and 9% (5/58) in nonmalignant peripheral lung tissue. In 3 cases, p16 methylation was detected in both tissues, while in 2 cases, methylation was observed in peripheral tissue only. All leukocyte DNAs, whether from healthy controls (n ϭ 2) or cancer patients (n ϭ 15), were negative. In all tissue samples positive for methylation, the corresponding unmethylated PCR product from both genes was also detected.
The frequency of p14 promoter methylation was clearly lower in this series of lung tumours as compared to that of p16 gene. Four (8%) out of the 52 cases analysed for p14 exhibited methylation in the tumour (Table I). For 3 of those 4 with tumour methylation, a peripheral lung tissue sample was available for analysis, and 1 of them was methylated in p14 promoter region.
The risk was clearly higher than that in adenocarcinoma (OR 4.2, 95% CI 0.6 -53) but it did not reach statistical significance. Most of the cases that exhibited p16 promoter methylation in the nonmalignant peripheral lung tissue were current smokers (4/5) and presented with squamous cell carcinoma of the lung (4/5).
In general, aberrant methylation of p14 gene promoter region was clearly less common than that of the p16 gene (8% vs. 34%, Table I). Similar to p16, all cases exhibiting p14 methylation were histologically NSCLCs (4/46, 9%). The low number of cases positive for p14 methylation did not allow statistical analysis.
Logistic regression analysis was performed to adjust for the 2 variables (age and smoking status) that showed statistically significant associations. The increased risk of p16 methylation in exsmokers as compared to current smokers was retained after adjustment for age (OR 3.7, 95% CI 0.9 -17; pϭ0.08). The association  Data on occupational exposure was available for 51 patients with NSCLC, 21 of whom (41%) had been exposed to asbestos at work. A higher, but statistically nonsignificant (48% vs. 37%; OR 1.6, 95% CI 0.4 -5.6) frequency of p16 promoter methylation was detected in the asbestos exposed group (Table II).

p16 methylation versus other gene alterations
We previously determined p53 and K-ras gene mutations in the same case series. 26 No correlation was observed between p16 methylation and p53 or K-ras mutation (Table II). For p14 gene, promoter methylation occurred concurrently with p16 methylation in 3 tumours, and p14 methylation was presented in parallel with a p53 mutation in 2 tumours. One tumour, LCC by histology, exhibited all 3 alterations, i.e., p16 and p14 methylation as well as p53 mutation.
Finally, we analysed both the genetic and epigenetic alterations in NSCLC in relation to smoking habits. K-ras mutation data were excluded from the analysis due to a lower number of cases analysed. For the analysis, alterations of the p53 and p14 genes, as representing damage in the same pathway, were combined. Totally, in NSCLC, 19% of the tumours had p16 methylation only, 31% p53 mutation and/or p14 methylation only, and 25% had an alteration in both pathways. Again, ex-smokers had significantly more lung tumours with both regulatory pathways affected as compared to tumours from smokers (44% vs. 14%; P ϭ 0.035; Fisher's exact test). The distribution of the alterations between ex-smokers and current smokers is indicated in Figure 3. Of note, 34% of the current smoker tumours but none of the ex-smoker ones were negative for all alterations studied (0% vs. 34%; P ϭ 0.008, Fisher's exact test). DISCUSSION We examined the frequency of promoter region methylation of the p16 and p14 genes in a series of lung tumours and the matching peripheral lung tissue specimens. Overall, 1/3 of the lung tumour samples exhibited p16 methylation. More than 40% of the NSCLCs exhibited p16 methylation in the tumour, whereas all SCLC tumours were negative. Risk of aberrant p16 methylation was elevated not only in current smokers but also in ex-smokers.
Our observation that showed no p16 methylation in SCLC is in agreement with a recent study by Toyooka and co-workers, 28 where a very low rate of p16 methylation was detected in neuroendocrine lung tumours, including SCLC. In NSCLC, p16 methylation is more frequent, with 21% to 58% of the tumours being positive in different studies. 21,22,29 -34 Our data on NSCLC are well in keeping with those findings.
In NSCLC, positive associations have been found between p16 methylation in the tumour and squamous cell histology, 22,28,34,35 male sex, 34 later stage 31 and tobacco smoking. 22 We did not detect predominance of p16 methylation in squamous cell carcinoma as compared to adenocarcinoma. In our series of lung tumours, the cell type showing the highest rate of p16 promoter methylation was large cell carcinoma, with more than 70% of the LCC tumours being positive. In addition, we found that p16 methylation frequently occurred in parallel with p53 mutation (3/5) in LCC, and there was 1 LCC case with alterations detected in all 3 genes studied, i.e., in p16, p14 and p53. Our finding fits well with literature data indicating promoter methylation as well as genetic alterations in multiple genes in LCC. 24,34 Also, a meta-analysis on genetic changes in lung cancer reported the highest rate of p53 mutations in LCC. 36 In the present study, 42% of the NSCLC cases that had ever smoked had methylation in p16 in the tumour. From the 4 nonsmokers with NSCLC studied, 1 exhibited p16 methylation in the tumour DNA. Several recent studies [21][22][23][24]37 have reported associations between aberrant p16 methylation and smoking. Moreover, significant correlations between smoking characteristics, such as duration, pack-years or time since quitting smoking, and p16 methylation in NSCLC were found. 22 In our study, ever smokers had increased frequency of p16 methylation, but we did not observe statistically significant associations between methylation and cumulative tobacco smoke exposure or duration of smoking. For occupational exposure to asbestos, the risk of p16 methylation was elevated in the exposed as compared to the nonexposed cases, but the difference was not statistically significant. A similar trend has recently been reported. 22 Interestingly, we discovered that former smokers with NSCLC had a significantly higher risk of p16 methylation than current smokers did (OR 5.1, 95% CI 1. [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. This difference was retained after adjustment for age, the other variable found to be significantly associated with methylation (OR 3.7, 95% CI 0. 9 -17). Furthermore, when we combined our present data on p16 and p14 methylation with those on p53 mutations analysed earlier, we found that ex-smokers carried an alteration profile different from that in current smokers. All ex-smokers examined had at least 1 of the 3 alterations, whereas 34% of current smokers had none. In addition, p16 methylation occurred significantly more often in combination with alterations in the p53 pathway (i.e., p53 mutation and/or p14 methylation) in tumours from ex-smokers as compared to those from current smokers (44% vs. 14%; p ϭ 0.035; Fischer's exact test).
The occurrence of aberrant methylation in p16 gene in exsmokers is in accordance with data from literature. A high prevalence of p16 methylation has been reported in adenocarcinomas from former smokers, with the detected level equalling to or even exceeding that detected in current smokers. 37 In addition, promoter region methylation of several genes, including p16, has been detected at elevated frequencies in bronchial brush samples or bronchial epithelial cells from former smokers with or without  lung cancer. 38,39 The findings that indicate recurrent abnormalities in smokers, who have quit smoking, are again in accordance with reports on high rates of p53 mutations observed in lung cancers from former smokers. 19,20,40 In summary, these observations suggest that smoking-related genetic and epigenetic abnormalities may persist in tumour cells after cessation of smoking.
We found p16 methylation in 9% (5/58) of the nonmalignant peripheral lung tissue samples, mainly from current smokers with squamous cell carcinomas. Similar rates of methylation (6 -14%) in promoter regions of different cancer-related genes, including p16, APC, RASSF1A, RAR␤, DAPK and FHIT, has been reported for noncancerous lung tissue from NSCLC patients. 34,37,41 Methylation in p16 was shown to occur as an early event in preneoplastic epithelial lesions of the lung 21 and was detectable in bronchial epithelial cells and sputum samples from smokers. 21,38,39 Also, smoking-related genetic damage, such as DNA adducts, has been demonstrated to occur in nonneoplastic lung tissue from cancer-free former smokers. 42 These findings that indicate the presence of various abnormalities in nontumour lung tissue in smokers suggest that larger areas of the histologically normal lung tissue suffer from tobacco-related damage that precedes neoplastic changes, i.e., a field effect. 43,44 Aberrant p16 methylation in tumour was significantly associated with the older age (Ն65 years) at the diagnosis. Such correlation has been detected in colon cancer, 45,46 implicating methylation as an age-associated event. In smokers, however, the increase in methylation along with the older age may, at least partially, reflect increase in cumulative tobacco smoke exposure over the years. The finding that adjustment for smoking diminished the risk related to age supports this notion.
In conclusion, we found that methylation of p16 gene was frequent in NSCLC tumours from both smokers and former smokers. Recent studies have suggested that, in lung cancer patients, aberrant methylation of genes regulating cell proliferation and growth is detectable in samples available through noninvasive sampling methods such as sputum or serum. Therefore, analysis of promoter methylation in such genes may provide a biomarker valuable for identification of groups with an elevated risk of cancer, such as smokers and ex-smokers.