Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Chapter 6 Comprehensive CADM1 promoter methylation analysis in NSCLC and normal lung specimens Remco M. van den Berg, Peter J. F. Snijders, Katrien Grünberg, Clarissa Kooi, Marieke D. Spreeuwenberg, Chris J.L.M. Meijer, Pieter E. Postmus, Egbert F. Smit, Renske D.M. Steenbergen In press, published online Lung Cancer 2010. Abstract Methylation-mediated silencing of the tumour suppressor CADM1 has been functionally linked to lung cancer development. We aimed to determine whether CADM1 promoter methylation is a candidate early detection marker for lung cancer. To this end frozen tissue samples of 36 non-small cell lung cancers, 26 corresponding tumour distant normal tissue samples as well as 6 samples of normal lung from non-lung cancer patients were tested for DNA methylation at three different regions within the CADM1 promoter (M1, M5 and M9) using methylation specific PCR followed by methylation specific reverse line blot analysis. Sixty-four % of tumour samples tested positive at the M1 region, 47% at M5 and 74% at the M9 region, compared with 65% (M1), 23% (M5) and 46% (M9) of paired normal tissue samples. Methylation of each of these promoter regions was also detected in the majority of non-lung cancer control samples. Dense methylation, defined as methylation at ≥2 promoter regions, was detected in 66% of tumour samples compared with 38% of paired normal tissues and 67% of non-lung cancer control samples. Within the small subgroup of female patients dense methylation was found in all tumour samples but only 22% of paired normal samples. Neither methylation of individual sites nor dense methylation was correlated with disease free survival. In conclusion, CADM1 promoter methylation is a frequent event in NSCLC as well as normal lung, both of lung cancer and non-lung cancer patients. Hence, CADM1 methylation analysis is unlikely to have diagnostic value for the early detection of lung cancer in an unselected population. However, a diagnostic value for selected subjects, such as females, cannot be excluded. 88 Introduction Lung cancer claims more lives than any other malignancy. Two main categories are distinguished: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). The latter, an umbrella term for a number of histological subtypes that comprise about 80% of lung cancers, can be cured by surgery in its earlier stages. However, this disease often becomes symptomatic only after the tumour has metastasised. Efforts to find screening tests that enable early diagnosis, when curative treatment is still an option, should be pursued. Malignant transformation is caused by a stepwise accumulation of molecular changes. Theoretically, detection of these changes by sensitive nucleic acid amplification assays that are available nowadays could serve as a molecular marker for pre-symptomatic lung cancer detection, e.g. by sputum analysis of risk groups. Hypermethylation of tumour suppressor gene promoter regions has generated attention since this is a frequent event in the onset of many cancers 1, 2. Different methods to determine DNA methylation have been developed 3. Of these, methylation specific PCR (MSP) is one of the most sensitive methods and consequently well suited to detect low target frequencies of tumour DNA in sputum samples. Support for a functional involvement of methylation events in lung cancer development has come from a recent study of Damiani et al., who demonstrated that transformation of immortalized bronchial epithelial cells upon exposure to carcinogens was causally related to an increase of DNA Methyl Transferase 1 (DNMT1) protein and concomitant de novo methylation of tumour suppressor genes 4. Indeed, a large number of genes has been found to be methylated in lung cancer, of which p16, CDH13 and RASSF1A are amongst the most commonly methylated genes being described in multiple studies 5. Analyses on sputum samples revealed that a population with increased risk of lung cancer might be discerned using methylation markers 6-11. One of the promising methylation marker panels described for sputum consists of APC, p16, 3OST2 and RASSF1A 11. However, combinations of markers with better sensitivity and specificity than those that have been studied so far are required to allow implementation in screening tests 5. Thus, there is a need for additional markers in lung cancer that can complement currently established panels to increase their efficacy. One candidate marker is methylation of the CADM1 gene. CADM1 (Cell adhesion molecule 1, originally called TSLC1 for tumour suppressor in lung cancer 1, and also referred to as IGSF4) is located in a region of chromosome 11q, which is commonly affected by deletions in all types of lung cancer 12. It encodes a transmembrane protein of the immunoglobulin superfamily involved 89 in cell-cell interactions between epithelial cells. The CADM1 gene has been shown by functional complementation studies to inhibit tumorigenesis of lung- and cervical cancer cell lines 13, 14. Reduced expression and/or promoter methylation of CADM1 has been found in carcinoma of the lung, cervix, nasopharynx, oesophagus and breast 14-18. The gene is silenced in most cases through methylation of its promoter, either or not in combination with allelic loss 15, 18. In a series of 48 NSCLCs examined using methylation-specific single strand conformation polymorphism (MS-SSCP) and bisulfite sequencing, 21 tumours showed CADM1 promoter hypermethylation, whereas no abnormal methylation in the adjacent normal lung tissue could be detected15. CADM1 methylation was more frequently detected in high-grade lung tumours 14, 15 and also found to be associated with tobacco smoking and poor prognosis19. A recent study investigated CADM1 as part of a panel of 27 methylation markers in NSCLC using semi-quantitative real-time ms-PCR (methylight). Although CADM1 was more frequently methylated in tumour tissue than in normal tissue (18% vs. 4% respectively), it was not included in the panel of 8 markers that the authors deemed most relevant because any added sensitivity did not outweigh the decrease in specificity 20. Although CADM1 has proven tumour suppressive activity in a lung cancer cell line13, and expression in lung cancers is often reduced 21, 22, the reported frequencies of promoter methylation of this gene are highly variably amongst different studies. Therefore, no firm conclusions can be drawn yet about the potential value of CADM1 as biomarker for early lung cancer detection. For this a more comprehensive analysis is warranted. Towards this aim we examined methylation of three CADM1 promoter regions in lung cancer tissue specimens and adjacent normal lung tissue samples, as well as normal lung tissue of non-lung cancer patients using methylation specific PCR (MSP) and a methylation specific reverse line blot hybridisation read-out for PCR products23. Previous application of this methodology to cervical carcinomas and precursor lesions, revealed an association between methylation at ≥2 of the CADM1 promoter regions (i.e. dense methylation) and both reduced CADM1 expression and increased severity of the lesion23. 90 Materials and methods Tissue samples Tissue samples from 36 patients who had surgery for NSCLC between July 1987 and September 2004 as well as lung samples from 6 patients who underwent surgery for lung diseases other than primary lung cancer were obtained from the lung tissue bank of the department of Pathology of the VU University Medical Center (table 1). Specimens had been collected within an hour after surgical resection and tissue of the tumour had been snap-frozen and stored in liquid nitrogen until use. For 26 patients normal lung tissue taken from the surgical specimen at maximum distance from the tumour had been collected and snap-frozen as well. Normal lung tissue samples from non-lung cancer controls were obtained from patients undergoing surgery for lesions that turned out to be pulmonary metastases of colorectal carcinoma (n=3) or non-malignant disease (fibrous scar tissue n=3). Diagnosis of the lesion was obtained from the pathology report. A board certified pathologist confirmed the histology of the frozen material. Data on staging and follow up were obtained from patient files. This study followed the ethical guidelines of the Internal Review Board of the VU University Medical Center. Table 1 Patient characteristics according to histology AC n=11 SCC n=19 LC-other n=6 Non-LC n=6 Age 60±11 69.6±7.9 59.2±21.6 67.0±7.7 Follow-up time (yrs) 4.2±3.1 3.9±4.5 1.3±1.9 1.3±1.3 Male Disease free survival (yrs)* 5 (45) 5.0±1.2 14 (74) 9.9±2.2 5 (83) 1.9±1.0 5 (83) Recurrence 6 (55) 6 (32) 3 (50) Data presented as mean ± standard deviation or n (%) unless otherwise indicated. AC: tumours with adenocarcinoma features (adenocarcinoma (n=8), adenosquamous carcinoma (n=2), bronchioloalveolar carcinoma (n=1)), SCC: squamous cell carcinoma, LC-other: lung cancer-other: i.e. LCC (n-4), MEC (n=1) and carcinoid (n=1) * Mean ± standard error 91 Laboratory procedures DNA extracted from frozen tissue was subjected to bisulfite modification using the EZ DNA methylation kit™ (Zymo Research, Orange, Ca, USA). 500 ng of DNA was modified and 25 ng of modified DNA was used for MSP. Laboratory procedures and MSP primers and probes for three regions in the CADM1 promoter (indicated as M1, M5 and M9, respectively) were as described previously 23. In short, PCR mixtures contained 25ng modified DNA, 0.5μM primers, each dNTP at 200μM, FastStart Taq PCR buffer (Roche Diagnostics), 1.5mM MgCl2, and 1.25U FastStart Taq DNA polymerase (Roche Diagnostics). All U-targeted primers were specific for unmethylated DNA, while M-targeted primers only detected methylated DNA. Antisense primers were biotinylated for reverse line blot (RLB) detection, in which denatured PCR-products were hybridized to oligoprobes specific for unmethylatedand methylated-DNA, respectively. In each run H2O, unmodified DNA and unmethylated DNA (primary keratinocytes) were included as negative controls and DNA isolated from the cervical cancer cell line SiHa was included as a positive control. Analytical sensitivity of each MSP-RLB assay, as determined on a dilution series of methylated DNA in unmethylated DNA, was 1% for M1, 5 % for M5 and 0.5% for M9 23. Statistics Proportions of methylated sites in tumour and tumour-distant normal samples of NSCLC patients were compared using the McNemar exact test and the relation of ‘dense methylation’ (i.e. MSP positivity at ≥2 sites) with clinical characteristics was analysed using the Fisher exact test. The presence of a relation between dense methylation in tumour samples and different patient characteristics was tested using stepwise binary logistic regression analysis with forward selection. Relation between clinical characteristics of NSCLC patients, dense methylation or methylation of individual sites and disease free survival was analysed using Kaplan Meier survival analysis with a log rank test. A p-value of ≤0.05 was considered statistically significant. 92 Results Patient characteristics according to histology are summarized in table 1. Eleven patients had tumours with adenocarcinoma features; adenocarcinoma (AC; n=8), adenosquamous carcinoma (ASC; n=2) and bronchioloalveolar carcinoma (BAC; n=1). Other histological subtypes were squamous cell carcinoma (SCC, n=19); large cell carcinoma (LCC, n=4); mucoepidermoid carcinoma (MEC, n=1) and carcinoid tumour (n=1). Methylation results of individual samples are shown in table 2. Sixty-four percent of tumour samples (23/36) showed CADM1 promoter methylation at the M1 site, 47% (17/36) at the M5 site and 74% (26/35) at the M9 site. Dense promoter methylation (defined as methylation of 2 sites or more, and previously associated with reduction of CADM1 expression in cervical lesions23) was present in 66% (23/35) of tumour samples. Dense methylation was significantly more frequent in female patients tumour specimens (6AC, 5 SCC and 1 carcinoid); all (12/12) specimens of female patients revealed dense methylation versus 46% (11/24) of the male patients tumours (p=0.002) (table 3). Normal lung tissue taken at a maximum distance from the tumour was available for 26 of 36 NSCLC patients. The M1 site was methylated in 65% (17/26), M5 in 23% (6/26) and M9 in 46% (12/26) of these samples. Dense methylation was found in 38% (10/26) of adjacent normal lung tissue samples. When dichotomised into patients older or younger than the median of 68 years, for the purpose of creating table III, significantly more patients with dense methylation of normal tissue were older than 68 (table 3). However, a T-test of mean ages showed no significant difference between patients with and without dense methylation of adjacent normal lung tissue. When comparing methylation at individual sites of paired tumour and normal samples using the McNemar test (table 4), only the proportion of tumour samples methylated at the M9 site was significantly higher than that of corresponding normal samples (p value 0.012). The frequency of dense methylation was also higher in tumour than in corresponding normal samples (p=0.039). This, however, could mainly be attributed to the high frequency of dense methylation in female subjects; all twelve females had densely methylated tumour tissue, whereas in only two out of nine matched normal tissue samples of female patients the corresponding tumour distant normal tissue showed dense methylation as well. In male subjects there was no significant difference between dense methylation in tumour and corresponding normal tissue. The relation between sex and dense 93 94 Sex Age Histology pTNM m m f f m f f f m f m m m m m m m m f f m m m f f m m m f m f m m m m m 72 45 67 47 77 52 51 49 71 63 67 70 59 87 74 75 80 80 62 65 69 70 55 62 67 73 62 75 69 70 45 71 77 74 67 22 AC AC AC AC AC AC AC AC ASC ASC BAC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC atypical carcinoid LCC LCC LCC LCC MEC T1N0 T1N0 T1N0 T1N2 T2N0 T3N0 T3N0 T3N2 T1N0 T2N0 T1N0 T1N0 T1N0 T1N0 T1N0 T1N0 T1N0 T1N0 T1N0 T1N0 T2N0 T2N0 T2N0 T2N0 T2N0 T2N2 T3N0 T3N0 T3N0 T3N1 T1N0 T2N0 T3N0 T3N0 T3N1 T1N0 m m m m m f 77 70 69 70 61 55 control* control* control* control# control# control# M1 Tumor M5 M9 M1 Normal M5 M9 Not available Not available Not available ND Not available Not available Not available Not available Not available Not available Not available methylated unmethylated AC: adenocarcinoma, ASC: adenosquamous carcinoma, BAC: bronchioloalveolar carcinoma, SCC: squamous cell carcinoma, LCC: large cell carcinoma, MEC: mucoepidermoid carcinoma, * pulmonary metastasis, # non-malignant, ND: not done because of insufficiënt material available Table 2 Summary of methylation results on tumour samples, adjacent normal samples and controls 18 I 24 2 4 19 11 67 (16/24) 100 (2/2) 25 (1/4) 58 (11/19) 82 (9/11) 10 (10/18) 72 (13/18) 50 (12/24) 92 (11/12) 64 (23/36) M1 Tumour 68 (13/19) 100 (2/2) 0 (0/4) 85 (11/13) 40 (4/10) 72 (8/11) 60 (9/15) 63 (10/16) 67 (6/9) 65 (17/26) Normal 54 (13/24) 100 (2/2) 25 (1/4) 53 (10/19) 36 (4/11) 33 (6/18) 61 (11/18) 67 (9/24) 67 (8/12) 47 (17/36) M5 Tumour 26 (5/19) 50 (1/2) 0 (0/1) 23 (3/13) 20 (2/10) 36 (4/11) 13 (2/15) 29 (5/17) 11 (1/9) 23 (6/26) Normal 83 (19/24) 100 (2/2) 25 (1/4) 78 (14/19) 91 (10/11) 67 (12/18) 88 (15/18) 56 (15/24) 100 (12/12) 74 (26/35) M9 Tumour 47 (9/19) 0 (0/2) 0 (0/1) 46 (6/13) 60 (6/10) 63 (7/11) 33 (5/15) 47 (8/17) 44 (4/9) 46 (12/26) Normal 67 (16/24) 100 (2/2) 25 (1/4) 63 (12/19) 73 (8/11) 50 (9/18) 77 (14/18) 46 (11/24) 100 (12/12) 66 (23/36) ≥2 sites Tumour 0.72 0.702 0.16 0.002 P1 73 (8/19) 50 (1/2) 0 (0/1) 46 (6/13) 30 (3/10) 64 (7/11) 20 (3/15) 47 (8/17) 22 (2/9) 38 (10/26) Normal 0.66 0.672 0.04 0.39 P1 II and III 12 7 (58) 57 (4/7) 33 (4/12) 14 (1/7) 58 (7/12) 43 (3/7) 58 (7/12) 28 (2/7) Data presented as n or % (n/total) unless otherwise indicated, 1Fisher exact test, 2only using AC and SCC, AC: tumours with adenocarcinoma features, SCC: squamous cell carcinoma, LCC: large cell carcinoma Stage Other LCC SCC AC 18 >68 <68 68 65±12.9 24 12 Median Mean Male Female Histology Age Sex Total Total Table 3 Percentage methylation per site in tissue samples from NSCLC patients 95 Normal Tumour U M total P* 6 13 19 0.75 U 13 1 14 Total 20 M 10 Dense U 6 M1 M5 M9 U M Total M U Total M Total 3 9 7 4 14 10 16 4 17 5 6 1 7 26 12 26 5 11 21 2 8 12 8 10 26 18 26 0.07 0.012 0.039 U: unmethylated, M: methylated, * 2-sided McNemar test methylation in tumour samples was confirmed by stepwise binary logistic regression analysis with forward selection; no other variables (tested were age, histology and disease stage) were a significant improvement of the model. In six tumour samples, including 4 SCCs and 2 LCCs no methylation was detected at any of the three promoter sites examined, whereas two of these cases (both SCC) had methylation in tumour distant normal samples at one and two sites, respectively. No clinical or histopathological characteristics were associated with dense methylation. There was also no relation between methylation of individual sites or dense methylation in tumour or normal tissue specimens and disease-free survival. The only clinical characteristic associated with disease free survival was disease stage (p=0.036, data not shown). As non-lung cancer controls we used normal tissue samples from patients who underwent surgery for a solitary pulmonary metastasis of a colon carcinoma (n=3) or pulmonary lesions that turned out to be fibrous scar tissue (n=3). Four of these 6 samples (67%) showed dense methylation as well (table 2). 96 Table 4 Comparison of matched tumour and adjacent normal tissue samples Discussion CADM1 silencing is functionally involved in pulmonary carcinogenesis13 and promoter methylation of this gene has been proposed as a candidate marker for early lung cancer detection5. Here, we applied an MSP-based methodology for assessing CADM1 promoter methylation that was previously validated on cervical material. By determining the methylation status at three different regions within the CADM1 promoter, this method predicted gene silencing in cervical tissue, based on the presence of methylation at ≥2 regions or dense methylation23. We found that dense methylation was not only frequently present in lung tumour samples, but also in 38% of corresponding normal samples taken from the same surgical specimen, as well as four out of six normal lung samples from patients without primary lung cancer. In the small subgroup of female patients a clear difference in dense methylation was observed between tumour and paired normal samples, warranting further studies. Whereas previous studies limited their analysis to a single region in the CADM1 promoter, our study included three regions. Methylation of these regions was heterogeneous; no distinct pattern could be discerned and methylation frequencies in each region were different. The M9 region was the only single region where CADM1 methylation was significantly more frequent in tumour samples than in corresponding normal samples. Nevertheless, 12 out of 26 corresponding normal samples were positive for M9 methylation. This relatively high frequency of CADM1 promoter methylation in tumour adjacent normal specimens could be confirmed by application of a newly developed quantitative MSP assay targeting a region largely overlapping that of M9 to a subset of cases (n=21). The overall concordance between both assays was good (86%) (data not shown). Although for the M5 region the difference between methylated tumour and normal samples was only borderline significant (p=0.07), it was the region displaying the highest specificity (the least methylation in corresponding normal samples; 6 out of 26). M5 overlaps the region tested by Fukami et al. and Kikuchi et al.15, 19, resulting in three CpGs shared between the different analyses. Within this region similar methylation frequencies in tumour tissue were found; 47% of primary NSCLCs showed methylation of the M5 region in our study compared with 44% (21/48 and 45/103, respectively) in both other studies. No CADM1 methylation in matched normal tissue was found in the latter two studies. Other studies on CADM1 methylation in lung specimens have described methylation frequencies in matched normal lung samples varying from 4% to 7% 20, 24. It is likely that these differences in methylation frequency in normal lung samples reflect differences in the sensitivity and targeted promoter regions between 97 the different assays used. Other studies have described varying degrees of cancer-related methylation in histologically normal lung tissue of lung cancer patients, such as adjacent normal tissue of specimens from lung cancer surgery 25 and histologically tumour-free bronchial resection margins 26. Analogous findings have been reported in other organs, such as for MLH1 in colon27. These abnormalities are often explained by the concept of ‘field cancerization’, i.e.(epi)genetically affected tumour adjacent areas that represent a premalignant condition from which new tumours may emerge. However, given the fact that (dense) CADM1 methylation was found in a substantial fraction of normal lung tissue of patients of a similar age who did not suffer from lung cancer, field cancerization is unlikely to be the only explanation for the observed methylation background in normal lung tissue. Instead, methylation observed in normal lung tissue may be the result of environmental insult, aging, or reflect some degree of physiological, epigenetic regulation of the CADM1 gene in certain cell types present in lung tissue specimens. Strikingly, all female patients had dense CADM1 methylation of tumour samples compared with only 46% of male patients. This finding is contrary to the findings of Kikuchi et al.19, who found significantly more male patients with tumour CADM1 promoter methylation. However, they found a relation with smoking index as well. Since we have no information about the patients’ smoking history it could not be included in our analysis, and it is therefore uncertain whether different smoking habits could explain the discrepancy between our results and those of Kikuchi et al. Although this seems irrelevant for our main conclusion, the lack of specificity of CADM1 methylation testing, no firm conclusions can be drawn on the potential biomarker value of CADM1 promoter methylation in non-smokers. Moreover, we could not reproduce the relation between prognosis and CADM1 expression in adenocarcinomas found in the same study19. Conclusion Present study shows that methylation of different sites of the CADM1 promoter is not only common in NSCLC, but can also frequently be detected in normal lung tissue of both lung cancer patients and patients without lung cancer. These data indicate that CADM1 methylation analysis has insufficient specificity to be included in a panel of markers for early lung cancer detection in an unselected population. We also did not find any evidence for CADM1 methylation analysis being a candidate prognostic marker for patients with lung cancer. Nevertheless, a diagnostic value for selected subjects, such as females, cannot be excluded. 98 References 1. F einberg AP, Ohlsson R, Henikoff S. The epigenetic progenitor origin of human cancer. Nat Rev Genet 2006 January;7(1):21-33. 2. Feinberg AP. The epigenetics of cancer etiology. Semin Cancer Biol 2004 December;14(6):427-32. 3. Fraga MF, Esteller M. DNA methylation: a profile of methods and applications. Biotechniques 2002 September;33(3):632, 634, 636-2, 634, 649. 4. Damiani LA, Yingling CM, Leng S, Romo PE, Nakamura J, Belinsky SA. CarcinogenInduced Gene Promoter Hypermethylation Is Mediated by DNMT1 and Causal for Transformation of Immortalized Bronchial Epithelial Cells. Cancer Res 2008 November 1;68(21):9005-14. 5. Anglim PP, Alonzo TA, Laird-Offringa IA. DNA methylation-based biomarkers for early detection of non-small cell lung cancer: an update. Mol Cancer 2008;7:81. 6. Belinsky SA, Liechty KC, Gentry FD et al. Promoter Hypermethylation of Multiple Genes in Sputum Precedes Lung Cancer Incidence in a High-Risk Cohort. Cancer Res 2006 March 15;66(6):3338-44. 7. Belinsky SA, Palmisano WA, Gilliland FD et al. Aberrant Promoter Methylation in Bronchial Epithelium and Sputum from Current and Former Smokers. Cancer Res 2002 April 1;62(8):2370-7. 8. Gilliland FD, Harms HJ, Crowell RE, Li YF, Willink R, Belinsky SA. Glutathione S-Transferase P1 and NADPH Quinone Oxidoreductase Polymorphisms Are Associated with Aberrant Promoter Methylation of P16INK4a and O6-Methylguanine-DNA Methyltransferase in Sputum. Cancer Res 2002 April 1;62(8):2248-52. 9. Chen JT, Chen YC, Wang YC, Tseng RC, Chen CY, Wang YC. Alterations of the p16(ink4a) gene in resected nonsmall cell lung tumors and exfoliated cells within sputum. Int J Cancer 2002 April 10;98(5):724-31. 10. Palmisano WA, Divine KK, Saccomanno G et al. Predicting Lung Cancer by Detecting Aberrant Promoter Methylation in Sputum. Cancer Res 2000 November 1;60(21):5954-8. 11. Shivapurkar N, Stastny V, Suzuki M et al. Application of a methylation gene panel by quantitative PCR for lung cancers. Cancer Lett 2007 March 8;247(1):56-71. 12. Murakami Y, Nobukuni T, Tamura K et al. Localization of tumor suppressor activity important in nonsmall cell lung carcinoma on chromosome 11q. Proc Natl Acad Sci U S A 1998 July 7;95(14):8153-8. 13. Kuramochi M, Fukuhara H, Nobukuni T et al. TSLC1 is a tumor-suppressor gene in human non-small-cell lung cancer. Nat Genet 2001 April;27(4):427-30. 14. Steenbergen RDM, Kramer D, Braakhuis BJM et al. TSLC1 Gene Silencing in Cervical Cancer Cell Lines and Cervical Neoplasia. J Natl Cancer Inst 2004 February 18;96(4):294305. 15. Fukami T, Fukuhara H, Kuramochi M et al. Promoter methylation of the TSLC1 gene in advanced lung tumors and various cancer cell lines. Int J Cancer 2003 October 20;107(1):53-9. 16. Hui AB, Lo KW, Kwong J et al. Epigenetic inactivation of TSLC1 gene in nasopharyngeal carcinoma. Mol Carcinog 2003 December;38(4):170-8. 17. Ito T, Shimada Y, Hashimoto Y et al. Involvement of TSLC1 in Progression of Esophageal Squamous Cell Carcinoma. Cancer Res 2003 October 1;63(19):6320-6. 18. Allinen M, Peri L, Kujala S et al. Analysis of 11q21-24 loss of heterozygosity candidate target genes in breast cancer: indications of TSLC1 promoter hypermethylation. Genes Chromosomes Cancer 2002 August;34(4):384-9. 19. Kikuchi S, Yamada D, Fukami T et al. Hypermethylation of the TSLC1/IGSF4 promoter is associated with tobacco smoking and a poor prognosis in primary nonsmall cell lung carcinoma. Cancer 2006 March 13. 20. Feng Q, Hawes SE, Stern JE et al. DNA methylation in tumor and matched normal tissues from non-small cell lung cancer patients. Cancer Epidemiol Biomarkers Prev 2008 March;17(3):645-54. 99 21. Uchino K, Ito A, Wakayama T et al. Clinical implication and prognostic significance of the tumor suppressor TSLC1 gene detected in adenocarcinoma of the lung. Cancer 2003 September 1;98(5):1002-7. 22. Goto A, Niki T, Chi-Pin L et al. Loss of TSLC1 expression in lung adenocarcinoma: relationships with histological subtypes, sex and prognostic significance. Cancer Sci 2005 August;96(8):480-6. 23. Overmeer R, Henken F, Snijders P et al. Association between dense CADM1 promoter methylation and reduced protein expression in high-grade CIN and cervical SCC. J Pathol 2008 August;215(4):388-97. 24. Heller G, Fong KM, Girard L et al. Expression and methylation pattern of TSLC1 cascade genes in lung carcinomas. Oncogene 2006 February 9;25(6):959-68. 25. Tsou JA, Shen LYC, Siegmund KD et al. Distinct DNA methylation profiles in malignant mesothelioma, lung adenocarcinoma, and non-tumor lung. Lung Cancer 2005 February;47(2):193-204. 26. Guo M, House MG, Hooker C et al. Promoter hypermethylation of resected bronchial margins: a field defect of changes? Clin Cancer Res 2004 August 1;10(15):5131-6. 27. Nuovo GJ, Nakagawa H, Sotamaa K, Chapelle AL. Hypermethylation of the MLH1 promoter with concomitant absence of transcript and protein occurs in small patches of crypt cells in unaffected mucosa from sporadic colorectal carcinoma. Diagn Mol Pathol 2006 March;15(1):17-23. 100