PTEN, RASSF1 and DAPK site-specific hypermethylation and outcome in surgically treated stage I and II nonsmall cell lung cancer patients
Article first published online: 30 SEP 2009
Copyright © 2009 UICC
International Journal of Cancer
Volume 126, Issue 7, pages 1630–1639, 1 April 2010
How to Cite
Buckingham, L., Penfield Faber, L., Kim, A., Liptay, M., Barger, C., Basu, S., Fidler, M., Walters, K., Bonomi, P. and Coon, J. (2010), PTEN, RASSF1 and DAPK site-specific hypermethylation and outcome in surgically treated stage I and II nonsmall cell lung cancer patients. Int. J. Cancer, 126: 1630–1639. doi: 10.1002/ijc.24896
- Issue published online: 28 JAN 2010
- Article first published online: 30 SEP 2009
- Manuscript Accepted: 27 AUG 2009
- Manuscript Received: 1 JUL 2009
- lung cancer;
- DNA methylation;
The primary objective of this study is to identify prognostic site-specific epigenetic changes in surgically treated Stage I and II nonsmall cell lung cancer (NSCLC) patients by quantifying methylation levels at multiple CpG sites within each gene promoter. Paraffin-embedded tumors from stage Ib, IIa and IIb in training and validation groups of 75 and 57 surgically treated NSCLC patients, respectively, were analyzed for p16, MGMT, RASSF1, RASSF5, CDH1, LET7, DAPK and PTEN promoter hypermethylation. Hypermethylation status was quantified individually at multiple CpG sites within each promoter by pyrosequencing. Molecular and clinical characteristics with time to recurrence (TTR) and overall survival (OS) were evaluated. Overall average promoter methylation levels of MGMT and RASSF1 were significantly higher in smokers than in nonsmokers (p = 0.006 and p = 0.029, respectively). Methylation levels of the p16 promoter were significantly higher in squamous cell carcinoma than in adenocarcinoma (p = 0.020). In univariate analysis, hypermethylation of RASSF1 at CpG sites −53 and −48 and PTEN at CpG site −1310 were the significantly associated with shorter TTR (p = 0.002 and p < 0.000, respectively). Hypermethylation of PTEN at −1310 and DAPK at −1482 were most significantly associated with outcome in multivariate analysis. These results show that methylation of specific promoter CpG sites in PTEN, RASSF1 and DAPK is associated with outcome in early stage surgically treated NSCLC.
Genetic and epigenetic events largely determine tumor phenotype and ultimately patient outcome. In carcinogenesis, tumor suppressor genes are frequently inactivated in two steps, starting with functional loss of one allele by a variety of mechanisms including mutations and methylation followed by deletion of the remaining allele. Methylation is the predominant mechanism for promoter inactivation by postreplicative epigenetic modification. Normally, stable or transient methylation at CpG dinucleotide promoter sequences coordinates gene expression during cell cycling.1, 2
Aberrant methylation in tumor suppressor genes has been linked to disease course and outcome.3–5 Gene expression profiling in primary tumors and metastases have revealed molecular signatures associated with a metastatic phenotype. For example, a 17-gene signature, characteristic of primary adenocarcinoma and associated with poor clinical outcome (p < 0.03), showed that cells preconfigured to metastasize are present in some primary tumors even at the time of diagnosis.6–8
For this study, promoter methylation of genes involved in a variety of cellular functions including adhesion (CDH1), DNA repair (MGMT), gene expression (LET7), cell division (p16, RASSF1, RASSF5) and survival (DAPK, PTEN), was compared with time to recurrence and survival in surgically treated early stage lung cancer patients. The genes studied were selected based on reports of associations between loss of function and outcome in lung and other cancers. Hypermethylation of p16, DAPK, MGMT and RASSF1 in bronchial epithelium and sputum have been linked to increased risk of lung cancer and hypermethylation of CDH1 and DAPK promoters were observed in lymphocytes from smokers.9, 10 Studies on cell lines investigating the methylation status of the Ras association domain family proteins, RASSF1 and RASSF5, support the disruption of apoptotic pathways in the metastatic phenotype.11–13 Aberrant methylation of a member of the Let-7 microRNA gene family, Let-7a-3 has been implicated in modification of expression of multiple genes that may affect outcome in ovarian14 and lung cancer.15 In previously reported studies, we observed that PTEN expression is associated with a positive effect on prognosis in late stage lung cancer patients treated with gefitinib.16
Hypermethylation was quantified using pyrosequencing.17 This methodology provides more detailed information of CpG methylation than other methods that qualitatively analyze one or two CpG sites, revealing critical methylations sites that may not have been detected previously.
Material and Methods
Patients and clinical assessment
Tumor material used in this study was from Stage Ia, Ib, IIa and IIb surgically treated NSCLC patients. Analysis was first performed on tumors from a training group of 75 patients (35 adenocarcinoma, 26 squamous cell carcinoma, 14 other) and then an independent validation group of 57 patients (28 adenocarcinoma, 23 squamous cell carcinoma and 6 other). Reference DNA from 22 normal leukocyte samples and 20 nonmalignant lung specimens was also tested. For all 132 NSCLC patients and 20 nonmalignant lung samples, fixed tumor specimens were obtained from the Department of Pathology, Rush University Medical Center (Chicago, IL). Diagnosis of NSCLC was acquired from pathology reports and histologic evaluation. Clinical data were established from chart review. Followup included radiographic imaging with histological verification of recurrence. Time to recurrence (TTR) and overall survival (OS) were measured in months from the date of diagnosis to the time of disease progression or death. Recurrent cases included any recurrence. Seventy nine percent of recurrences were less than 40 months after surgery. Eighty two percent of deaths from lung cancer occurred in less than 60 months. Equal numbers of nonrecurrent cases were selected for this study, with followup from 1 to 12 years after surgery. All cases were staged according to the TNM classification criteria (6th edition).18 This study was approved by the Rush University Medical Center Institutional Review Board with waiver of individual consent.
Immunohistochemistry (IHC) specimens were 5.0 μm sections of formalin-fixed paraffin-embedded tumor tissue or sections from cytology cell blocks. Immunostaining methods and reagents were described previously.16 Staining frequency and intensity of all tumor cells on each slide were estimated on scales of 0 to 4 without knowledge of clinical patient data. Intensity was judged from background and relative density of staining. For frequency, less than 1% positive tumor cells per field was scored as 0, 1–10% as 1, 11–35% as 2, 36–70% as 3, and over 70% as 4. IHC expression was dichotomized into 2 levels: positive (intensity × frequency > 4) and negative (intensity × frequency ≤ 4).
Fluorescence in situ hybridization
For fluorescent in situ hybridization (FISH), specimen slides were prepared as previously described.16 The prepared specimen slides were hybridized with two-color FISH probe solutions (PTEN/CEP7 LSI) in a HYBrite™ automated codenaturation oven (Abbott Diagnostics, Downers Grove, IL). DAPI stain was applied to the specimen for visualization of the nuclei. PTEN/CEP7 ratios <0.75 were interpreted as PTEN loss.
DNA was extracted by proteinase K digestion of tumor and normal cells manually microdissected from paraffin-embedded tissue samples. DNA from white blood cells was isolated by inorganic extraction (Qiagen, Inc., Valencia CA). Bisulfite treatment of DNA was performed using the Qiagen Epitech system according to manufacturer's protocol. This system is designed to minimize adduct-induced strand breakage during and after conversion. Following bisulfite treatment, the converted DNA was amplified using primers designed for the altered sequences (Biotage, Inc., Uppsala, Sweden; EpigenDx, Worcester, MA). PCR was performed with HotStarTaq® (Qiagen) according to the manufacturer's protocol using modified primers (see accompanying Supporting Information). Amplicons were resolved by agarose electrophoresis to confirm proper amplification and quality of product.
The detection of the C/T polymorphisms that result from PCR amplification of bisulfite-treated DNA was performed by Pyrosequencing™ (Biotage) on a PyroMark MD pyrosequencer. Luminescent signal generated from the incorporation of nucleotides complementary to the test template is normalized and quantified. For each C/T polymorphism, the relative percent C (methylated) vs. T (unmethylated) at that position is reported. Non-CpG cytosines, which should be 100% converted, are included in each sequence to confirm complete conversion. This method was used to detect and quantify the degree of methylation for each CpG within each sample promoter being investigated. The following previously reported regions (CpG islands) of hypermethylation were analyzed (A of ATG of the start of translation = +1): CDH1, −160 to −121; p16, −64 to −40; LET7, −158 to −116; RASSF1, −57 to −36; RASSF5, −224 to −177; DAPK, −1518 to −1406; MGMT, −37 to −19; PTEN, −1333 to −1276. The output data includes site-specific percent methylation for each CpG in the analyzed area as well as overall average percent methylation, which is the average of the CpG levels within the promoter.
The associations between percent methylation and binary covariates were tabulated, and Fisher's exact test was used to measure their significance. For purposes of survival analysis, percent methylation levels at each CpG site, were divided into two classes (yes/no) using site-specific cutpoints. These cutpoints were chosen based on the results from the training data and the same cutpoints used for the validation group. The Kaplan–Meier method was used to estimate the probability of recurrence and the probability of survival as functions of time. Survival differences among comparator groups were analyzed by the log-rank test. Predictors that were statistically significant or marginally significant in univariate analyzes or were deemed to be clinically or biologically important were included as candidate covariates in multivariate Cox proportional hazards (PH) regression models. Statistical analyzes were done using Version 9.1.3 of the SAS software (SAS Institute), SPSS version 15 and the statistical software R. All reported p-values are two sided. P-values between (0.05–0.10), (0.01–0.05) and (<0.01) are respectively reported as marginally significant, significant and strongly significant.
Table 1 shows patient characteristics and clinical status for the training and validation groups. Median ages in these two groups were 62.8 and 65.2 years, respectively. In the combined group, 91 of 132 tumors (68.9%) were more than 3 cm in diameter. Lobectomies accounted for 106/131 (80.9%) surgeries. Other surgical types included 10 (7.6%) segmentectomies, 12 (9.2%) pneumonectomies and 3 (2.3%) sleeve lobectomies. The median Median OS for the combined training and validation groups was significantly shorter for patients who received adjuvant chemotherapy than for those who did not (p = 0.011). OS, but not TTR, was significantly shorter in males than for females (p = 0.039). Patients with tumor size more than 3 cm at diagnosis had significantly shorter OS (p = 0.049). We did not find any other significant association of patient characteristics with either TTR or OS in either patient group. Recurrent and nonrecurrent cases did not differ with regard to age, gender, ethnicity, stage or smoking status. Recurrent cases were significantly more likely to have received chemotherapy (p = 0.000 and p = 0.003 in the training and validation groups, respectively).
Overall average promoter methylation
Methylation was measured quantitatively. Percent methylation varied from gene to gene. The methylation state was dichotomized using gene-specific cutpoints to define hypermethylation. These cutpoints for overall average promoter methylation were selected based on average measurements for each gene in the training data; the same cutpoints (10–50%) were used for the validation and combined data. Table 2 shows the number of cases in the training and validation groups where hypermethylation (percent methylation more than the gene-specific cutpoint) was observed for the genes tested. Frequently hypermethylated (percent methylation above the gene-specific cutpoint) promoters included the PTEN promoter, which was methylated in 49.2% of the cases (training and validation groups combined), the RASSF5 promoter (34.4% of cases) and RASSF1 (37.5% of cases). As has been observed in other studies, the LET7 promoter was heavily methylated, with more than 50% methylation in 99% of cases.15
We did not find significant association of overall average promoter methylation levels with age over 60, gender, ethnicity, adjuvant chemotherapy and stage. Average promoter methylation of p16 was higher in squamous cell carcinoma pathology than in adenocarcinoma (Table 3); the difference was significant (p = 0.04) in the training group, not significant in the smaller validation group, but was significant (p = 0.02) when the two groups are combined. On the other hand, promoter methylation of MGMT was marginally higher in smokers than in nonsmokers in the training group (p = 0.058), but significantly higher in the validation group (p = 0.024) and combined groups (p = 0.006). Overall average promoter methylation levels of RASSF1 was significantly different between smokers and nonsmokers in combined groups (p = 0.029), but not in either of the component groups.
Methylation levels compared to normal tissue
Qualitatively, methylation levels in normal blood cells and nonmalignant lung were low in 6 of the genes tested showing that elevated methylation levels (>10–15%, depending on the gene) observed in these genes was cancer-specific. Two genes, RASSF1 and LET7 had increased overall promoter methylation levels in normal tissues. Increased methylation was observed in 3 of 17 RASSF1 nonmalignant lung specimens compared to 0 of 21 normal blood samples. RASSF1 overall promoter methylation levels averaged 3.86% in leukocytes, 12.2% in nonmalignant lung and 17.1% in NSCLC cases. All three of the nonmalignant lung specimens with elevated RASSF1 promoter methylation levels were from smokers. This is consistent with observations of lower levels of the RASSF1 transcripts found in bronchial epithelia from smokers.19LET7 promoter methylation levels over 50% were present in both leukocytes and nonmalignant lung. LET7 overall average percent promoter methylation levels were 62.0% in leukocytes, 67.5% in nonmalignant lung and 70.4% in NSCLC cases. The overall average and individual CpG promoter methylation levels are provided in the accompanying Supplement.
Methylation levels at individual CpG Sites along each promoter
The site-specific percent methylation of each of multiple CpG cytosines within each promoter was measured. In general, consistent patterns of methylation levels were observed across the promoters for each gene, that is, methylation levels were relatively higher or lower at particular sites within the same promoter for the majority of patients. Graphs of percent methylation vs. promoter site for NSCLC, nonmalignant lung and leukocytes are shown in Supporting information.
Site-specific methylation levels vs. demographics
Among patient characteristics, smoking history was found to be strongly associated with altered tumor methylation status involving cytosine at position −35 of the MGMT gene (p = 0.002, 0.024 and 0.001 in the training, validation and combined groups respectively) and cytosine at position −1507 of the DAPK gene (p = 0.008, 0.033 and < 0.001 in the training, validation and combined groups respectively) (Table 4). Methylation levels of DAPK at position −1486 were lower in stage Ib than in stage II (p = 0.005 and 0.063 in the training and validation groups, respectively). Methylation levels in a separate group of stage Ia patients was 4.71% (data not shown). Methylation of p16 at cytosine site −46 was significantly lower in adenocarcinoma than in squamous cell carcinoma in the training and combined groups (p = 0.034 and 0.009).
Methylation at individual CpG Sites, time to recurrence and survival
The association of site-specific hypermethylation (at individual CpG sites) with TTR is evaluated using the Kaplan-Meier method and log-rank test. Results are shown in Table 5 and Figure 1. Site-specific cutpoints, selected based on percent methylations levels showing significant effects in the training data, were used to define the site-specific hypermethylated state for each cytosine within each gene promoter. Site-specific hypermethylations at five specific CpG sites in four genes, RASSF1, PTEN, RASSF5 and DAPK, were found to be significantly associated with TTR in the training group (Table 5). The directions of association were the same in all five CpG sites; site-specific hypermethylations were associated with shorter survival in all five sites.
Validation of methylation at individual CpG sites, time to recurrence and survival
In the validation group, hypermethylation in the two sites in RASSF1 were marginally associated with TTR. Although the associations with TTR were not significant in the other three genes in the validation group, promoter hypermethylation at the indicated sites was always found with shorter TTR. When the training and validation groups are combined, all 5 CpG sites are significantly associated (three out of the five are strongly) with TTR. The directions of association were again identical at all five sites with hypermethylation being associated with shorter TTR.
Table 5 shows p values for the separate and combined training and validation groups. The marginal significance in the validation group for RASSF1 and PTEN led to a more highly significant association when the groups were combined. Methylation of site −1310 in the PTEN promoter ≥23% was associated with shorter TTR than methylation less than 23% (13.5 mos, 95% CI 9.04 to 18.0 vs. 82.3 months, 95% CI 39.7 to 125; p < 0.001; Fig. 1). Hypermethylation (methylation ≥30%) of site −53 in the RASSF1 promoter was significantly associated with shorter TTR (16.3 mos, 95% CI 0.00 to 33.0) than methylation <30%, (78.8 mos, 95% CI 34.6 to 123; p = 0.002). A similar correlation was seen at an adjacent CpG site −48 (p = 0.003).
With regard to survival, hypermethylation (methylation ≥30%) of RASSF1 at position −53 was also significantly associated with shorter survival than with methylation <30% (median survival 32.1 mos vs. median not reached; p = 0.017). Hypermethylation (methylation ≥30%) of RASSF1 at position −53 was also significantly associated with shorter survival than with methylation <30% (p = 0.002). Hypermethylation of the MGMT (methylation ≥18%) promoter site −35 was correlated with shorter survival (p = 0.003), as was hypermethylation (methylation ≥11%) of p16 at site −49 (p = 0.010; Fig. 1).
PTEN methylation, deletion and protein expression
In previous studies,16 PTEN protein expression was associated with longer survival in late stage lung cancer patients treated with Iressa. In the current study, PTEN promoter hypermethylation showed a correlation with shorter TTR in early stage patients. PTEN protein expression and genomic status were analyzed by immunohistochemistry and fluorescence in situ hybridization (FISH), respectively. Although no marker was significantly associated with survival, the results are consistent with deletion, hypermethylation or low PTEN protein expression having an adverse effect on survival (Fig. 2a).
Protein expression was slightly lower with hypermethylation of PTEN at −1310. Similar results were observed for CDH1 and p16 protein expression. It has been reported that LET7 microRNA down-regulates CDH1.21 Methylation of the LET7 promoter at −116 >67% resulted in higher levels of CDH1 protein (frequency × intensity = 6.38) than methylation ≤67% (frequency × intensity = 5.33; Fig. 2b).
Multivariate Cox PH regression was used to estimate the effect of demographic, immunohistochemical and epigenetic study variables on TTR in the combined training and validation groups. Baseline variables such as age at diagnosis, smoking status, ethnicity, histology, stage and methylation status of CpG sites that were significantly associated with TTR in univariate analyzes (Table 5), were included in the multivariate model. Variable selection methods were used to reduce the number of predictors and to arrive at the final model DAPK hypermethylated at −1486 had the smallest adjusted p-value for predicting recurrence in this multivariate model (p < 0.001, HR 1.93, CI 1.06 to 3.53) followed by PTEN at −1310 (p = 0.001, hazard ratio 3.63, 95% CI 1.61 to 8.17, n = 132).
A similar strategy was followed for multivariate Cox PH regression analysis of five year survival. Male gender (p = 0.01, HR = 3.36, 95% CI 1.34 to 8.43), CDH1 hypermethylation at −143 (p = 0.018, HR = 5.07, 95% CI 1.32 to 19.40) and RASSF1 hypermethylation at −48 (p = 0.040, HR = 2.71, 95% CI 1.05 to 7.0) were also significantly associated with shorter survival.
This study reveals a possible relationship between promoter hypermethylation of phosphatidyl inositol 3 kinase inhibitor PTEN, and ras associated protein RASSF1 at specific cytosine residues and shorter TTR and five year survival in two small stage I and II NSCLC patient groups. Furthermore, hypermethylation of a specific promoter cytosine in the death-associated kinase gene, DAPK was associated with shorter TTR in these patient groups. Several studies have been designed to investigate the clinical significance of epigenetic state of selected genes and gene sets in lung cancer. Many methylation studies on lung and other cancers make use of methods that qualitatively analyze particular sites within the gene promoter. This study was performed using a method that quantifies methylation levels at multiple CpG sites within each gene promoter, affording a more detailed survey of the methylation state.
Quantification of methylation by this approach revealed consistent patterns of CpG methylation levels across gene promoters. The patterns observed for the p16 and CDH1 promoter CpG methylation levels were similar to patterns previously reported for these promoters.21, 22 Different CpG sites within the same promoter may show no methylation or high or variable levels of methylation. Studies based on single methylation sites, therefore, may be inconsistent, depending on which sites are being analyzed. Furthermore, the degree of methylation of each CpG site can vary, with respect to actual biological effect. Methylation status predictive of outcome may, therefore, be revealed by testing different levels of methylation at multiple CpG sites.
Two of the 8 tumor suppressor genes tested had significantly higher methylation of at least one promoter CpG site in smokers, compared to nonsmokers. The strongest correlations with smoking were with DAPK at cytosine position −1507 (combined training and validation groups p < 0.001) and MGMT at position −35 (combined training and validation groups p = 0.001). In addition, overall average promoter methylation of RASSF1 was higher in smokers than in nonsmokers (combined training and validation groups p = 0.029). Weakened DNA repair capability (for example by loss of MGMT function) has been connected to better response to nitrosourea compounds in some cancers.23, 24 Loss of apoptotic functions, however, would compromise cell death induced by unrepaired DNA damage. Cells with increased DNA repair (lower promoter methylation of MGMT) and decreased apoptosis (higher promoter methylation of DAPK or RASSF1) might be more resistant to therapeutic intervention.
Comparison of methylation of multiple CpG sites in the p16 promoter vs. histopathology revealed that methylation of p16 at position −46 was higher in squamous cell carcinoma than in adenocarcinoma (combined training and validation groups p = 0.020). Association of promoter hypermethylation of p16 but not RASSF1, RASSF5 nor MGMT with histopathology is consistent with observations made by Blanco et al., supporting an inflammation-induced lung adenocarcinoma model.25
Brock et al.4 observed an association between recurrence in stage Ia patients and promoter hypermethylation of p16 at one of the sites (−64) investigated in this study. We did not observe significantly longer TTR with hypermethylation at −64 nor −59 in the stage Ib, IIa and IIb training and validation groups, however, we did observe an association between hypermethylation of p16 CpG site −64 and shortened survival time in a separate group of stage Ia patients (n = 54; p = 0.046; data not shown). Results reported by Brock et al., unlike ours, did not show significant associations between RASSF1 hypermethylation and recurrence. RASSF1 hypermethylation at −48 was related to shortened survival even in out stage Ia group, further supporting results seen in the later stage training and validation groups (data not shown). Although the statistical power of our results is affected by the small sample numbers, RASSF1 hypermethylation has been linked to worse survival in several independent studies using different methods.26–28
Protein expression as measured by immunohistochemistry was not significantly lower in the hypermethylated specimens (using training group cutpoints to define the hypermethylated state) in the present study. Methodological differences between nucleic acid and protein analysis, in addition to confounding factors may have compromised this direct comparison. The consistently lower direction of protein levels is supportive, however, of a down-regulation due to methylation. Subtle changes are measured in vitro may still have a phenotypic effect.
Quantitative profiling in a study by Vaissiere et al.29 revealed significant associations of RASSF1 hypermethylation with gender, smoking and alcohol intake. Although our results are not entirely consistent with theirs with regard to smoking status, we did observe consistently (but not significantly) lower methylation levels in females than in males at all CpG sites tested in the RASSF1 promoter. Alcohol intake was not addressed in this study. The discrepancies may reflect alternate analysis methods of analysis and interpretation of hypermethylation status as well as different CpG sites investigated. This emphasizes the importance of indicating which locations within the promoter are being tested in defining methylation state. The small number of specimens that can be practically analyzed in such detail is a potential weakness of using promoter sequencing for methylation analysis. Once implicated, however, specific sites can be interrogated by more high-throughput methods such as methylation-specific PCR or array studies.
Global gene expression profiling studies have identified multigene signatures associated with outcome.30, 31 Since different criteria are used for data analysis, no consistent gene panels nor biological pathways have been identified. Comprehensive retrospective studies are being devised to address this issue.32 The correlation between recurrence and survival and hypermethylation of specific RASSF1, DAPK or PTEN CpG sites in this study suggests that hypermethylation of one or a few genes may be prognostic.
- 1Cyclical DNA methylation of a transcriptionally active promoter. Nature 2008; 452: 45–50., , , , , , , , , , , , et al.
- 2Transient cyclical methylation of promoter DNA. Nature 3008; 452: 112–5., , , , , , , , .
- 3Cohypermethylation of p16 and FHIT promoters as a prognostic factor of recurrence in surgically treated resected stage I non-small cell lung cancer. Cancer Res 2006; 66: 4049–54., , , , , .
- 4DNA methylation markers and early recurrence in stage I lung cancer. N Engl J Med 2008; 358: 1118–28., , , , , , , , , , , , .
- 5Epigenetics in cancer. N Engl J Med 2008; 358: 1148–59..
- 6A molecular signature of metastasis in primary solid tumors. Nat Genet 2003; 33: 49–54., , , .
- 7A five-gene signature and clinical outcome in non-small-cell lung cancer. N Engl J Med 2007; 356: 11–20., , , , , , , , , , , .
- 8A genomic strategy to refine prognosis in early-stage non-small-cell lung cancer. N Engl J Med 2006; 355: 570–80., , , , , , , , , , , , .
- 9Aberrant promoter methylation in bronchial epithelium and sputum from current and former smokers. Cancer Res 2002; 62: 2370–7., , , , , , , , , , , , .
- 10Differential DNA hypermethylation of critical genes mediates the stage-specific tobacco smoke-induced neoplastic progression of lung cancer. Clin Cancer Res 2005; 11: 2466–70., , , , , , , , , .
- 11Association of DNA methylation and epigenetic Inactivation of RASSF1A and beta-catenin with metastasis in small bowel carcinoid tumors. Endocrine 2006; 30: 299–306., , , , , , , , .
- 12Aberrant methylation of E-cadherin and H-cadherin genes in non small cell lung cancer and its relation to clinicopathologic features. Cancer 2007; 110: 2785–92., , , , , .
- 13Nore1 inhibits tumor cell growth independent of Ras or the MST1/2 kinases. Oncogene 2004; 23: 3426–33., .
- 14Hypermethylation of let-7a-3 in epithelial ovarian cancer is associated with low insulin-like growth factor-II expression and favorable prognosis. Cancer Res 2007; 67: 10117–22., , , , , .
- 15The human let-7a-3 locus contains an epigenetically regulated microRNA gene with oncogenic function. 2007; 67: 1419–29., , , , , , , .
- 16The prognostic value of chromosome 7 polysomy in non-small cell lung cancer patients treated with gefitinib. J Thoracic Oncol 2007; 2: 414–22., , , , , , , , , , , , .
- 17Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis. J Mol Diagnostics 2007; 9: 368–81., , , , , , , , , , .
- 18Revisions in the international system for staging lung cancer. Chest 1997; 111: 1710–7..
- 19Nore1 and RASSF1 regulation of cell proliferation and of the MST1/2 kinases. Methods Enzymol 2006; 407: 290–310., , , , .
- 20Let-7 expression defines two differentiation stages of cancer. Proc Natl Acad Sci USA 2007; 104: 11400–5., , , , , , , , .
- 21Detailed DNA methylation profiles of the E-cadherin promoter in the NCL-60 cancer cells. Mol Cancer Ther 2007; 6: 391–403., , , , , , , , , , , , .
- 22Promoter methylation of P16, RARbeta, E-cadherin, cyclin A1 and cytoglobin in oral cancer: quantitative evaluation using pyrosequencing. Br J Cancer 2006; 94: 561–8., , , , , , , .
- 23In vitro drug response and molecular markers associated with drug resistance in malignant gliomas. Clin Cancer Res 2006; 12: 4523–32., , , , , , .
- 24O6-methyl-guanine-DNA methyltransferase methylation in serum and tumor DNA predicts response to 1,3-bis(2-chloroethyl)-1-nitrosourea but not to temozolamide plus cisplatin in glioblastoma multiforme. Clin Cancer Res 2003; 9: 1461–8., , , , , , , , , .
- 25Molecular analysis of a multistep lung cancer model induced by chronic inflammation reveals epigenetic regulation of p16 and activation of the DNA damage response pathway. Neoplasia 2007; 9: 840–52., , , , , , , , , , .
- 26Epigenetic inactivation of RASSF1A in lung and breast cancers and malignant phenotype suppression. J Natl Cancer Inst 2001; 93: 691–9., , , , , , , , , , , , .
- 27Clinicopathological significance of epigenetic inactivation of RASSF1A at 3p21.3 in stage I lung adenocarcinoma. Clin Cancer Res 2002; 8: 2362–8., , , , , , , , , , , .
- 28Hypermethylation of RASSF1A promoter is associated with the age at starting smoking and poor prognosis in primary non-small cell lung cancer. Cancer Res 2003; 63: 3743–6., , , , , , .
- 29Quantitative analysis of DNA methylation profiles in lung cancer identifies aberrant DNA methylation of specific genes and its association with gender and cancer risk factors. Cancer Res 2009; 69: 243–52., , , , , , , , , , , , .
- 30Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci 2001; 98: 13790–5., , , , , , , , , , , , .
- 31A gene expression signature predicts survival of patients with stage I non-small cell lung cancer. PloS Med 2006; 3: e467., , , , , , , , , , , .
- 32Gene expression-based survival prediction in lung adenocarcinoma: a multi-site, blinded validation study. Nat Med 2008; 14: 822–7., , , , , , , , , , , , .
Additional supporting information may be found in the online version of this article.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.