|
|
| |||||||||||||||||||||||||||||||||||||||
Background Despite optimal and early surgical treatment of non–small-cell lung cancer (NSCLC), many patients die of recurrent NSCLC. We investigated the association between gene methylation and recurrence of the tumor.
Methods Fifty-one patients with stage I NSCLC who underwent curative resection but who had a recurrence within 40 months after resection (case patients) were matched on the basis of age, NSCLC stage, sex, and date of surgery to 116 patients with stage I NSCLC who underwent curative resection but who did not have a recurrence within 40 months after resection (controls). We investigated whether the methylation of seven genes in tumor and lymph nodes was associated with tumor recurrence.
Results In a multivariate model, promoter methylation of the cyclin-dependent kinase inhibitor 2A gene p16, the H-cadherin gene CDH13, the Ras association domain family 1 gene RASSF1A, and the adenomatous polyposis coli gene APC in tumors and in histologically tumor-negative lymph nodes was associated with tumor recurrence, independently of NSCLC stage, age, sex, race, smoking history, and histologic characteristics of the tumor. Methylation of the promoter regions of p16 and CDH13 in both tumor and mediastinal lymph nodes was associated with an odds ratio of recurrent cancer of 15.50 in the original cohort and an odds ratio of 25.25 when the original cohort was combined with an independent validation cohort of 20 patients with stage I NSCLC.
Conclusions Methylation of the promoter region of the four genes in patients with stage I NSCLC treated with curative intent by means of surgery is associated with early recurrence.
Epigenetic gene silencing is a molecular mechanism of silencing a gene by methylating its promoter region. Epigenetic silencing is involved in the initiation and progression of several types of cancer, including lung cancer.7,8 The detection of epigenetic alterations with the use of a method like the methylation-specific polymerase-chain-reaction (PCR) assay9 may allow for the molecular staging of cancer. With the methylation-specific PCR assay, relatively few genes are required to analyze each type of cancer.9 The method can detect occult micrometastases in lymph nodes from patients with esophageal,10 colorectal,11 gastric,10 prostate,12 or lung cancer,10,13 but its value for predicting the recurrence of early-stage, resected NSCLC has not been examined.
We designed a nested case–control study of early-stage NSCLC (T1–2N0) to test the association between clinical outcome and the DNA methylation status of tumor, regional lymph nodes confined to the pleural space, and mediastinal lymph nodes. We studied seven genes: the cyclin-dependent kinase inhibitor 2A gene p16, the H-cadherin gene CDH13, the adenomatous polyposis coli gene APC, the Ras association domain family 1 gene RASSF1A, the O6-methylguanine–DNA methyltransferase gene MGMT, the PYD and CARD domain–containing gene ASC, and the death-associated protein kinase 1 gene DAPK. These genes are thought to be important in the biologic development of lung cancer and are frequently methylated in lung cancer.14,15,16,17,18 We hypothesized that the methylation-specific PCR assay could be used to define patterns of DNA methylation that can delineate the behavior of the primary tumor and to detect micrometastases in histologically negative lymph nodes. Our results show that aberrant patterns of promoter methylation in the primary tumor, and in regional and mediastinal lymph nodes, can be used to identify patients with stage I NSCLC who have an increased risk of recurrence.
Methods
Patients
Evidence of recurrent disease was evaluated in 715 patients with pathologically verified stage I (T1–2N0) cancer who received a diagnosis of NSCLC (codes 162.3 to 162.9 of the International Classification of Diseases, Ninth Revision, Clinical Modification) and who underwent lobectomy or greater resections at the Johns Hopkins Hospital between January 1, 1986, and July 31, 2002. The case patients were 71 patients at our institution in whom the tumor recurred within 40 months after surgery, by which time approximately 80% of NSCLC recurrences occur.4 Follow-up of all 71 case patients was performed at Johns Hopkins Hospital with the use of radiographic imaging and, usually, histologic verification of recurrence. On the basis of age, NSCLC stage, date of surgery (±5 years), and sex, we matched the case patients to 158 controls with stage I NSCLC in whom there was no recurrence during the 40-month follow-up period. In this phase of the study, tissue samples from 51 of the 71 case patients and 116 of the 158 matched controls were available for methylation analysis. Seven of the 116 controls had a recurrence more than 40 months after surgery. Neither case patients nor controls received adjuvant chemotherapy; between 1986 and 2002, guidelines did not recommend adjuvant therapy for patients with stage IB NSCLC.19,20
All cases of cancer were staged according to the revised TNM classification criteria,5 including the histologic status of mediastinal lymph nodes sampled from levels II, IV, VII, VIII, IX, and X on the right side and levels V, VI, VII, VIII, and IX on the left side. Regional lymph nodes confined to the pleural space were resected en bloc with the tumor. Samples from case patients had no macroscopically or microscopically positive surgical margins, and the patients had had a lobectomy or greater resection.21 We also evaluated 162 paraffin-embedded tissue blocks from the 20 patients in the validation cohort (11 case patients and 9 matched controls). Of these 20 patients, 18 had undergone resection at our institution after August 2002. This study was approved by the Johns Hopkins Institutional Review Board. The requirement of written informed consent was waived.
Preparation of Tumor and Lymph-Node Specimens
All specimen blocks were procured from pathology archives en masse in a blinded fashion with regard to whether they were obtained from case patients or controls. There were no differences in the distribution of tumor and lymph nodes or number of samples per patient between the case patients and controls. Specimens were labeled with study-specific coded identifiers only; laboratory investigators had no knowledge of the patient's group or the source tissue of the DNA. DNA was extracted from a pool of three sequential sections, each 10 µm in thickness, from unstained, paraffin-embedded slides of resected tumors, regional lymph nodes, or mediastinal lymph nodes. For each sample of tumor or lymph-node tissue, adjacent sections were stained with hematoxylin and eosin for histologic confirmation of the presence or absence of cancer. Unstained tissue sections were deparaffinized, and DNA was extracted as described previously.22 The concentration of DNA was measured spectrophotometrically, and 1 µg of DNA was denatured with the use of sodium hydroxide and modified with the use of sodium bisulfite. We then purified the DNA samples by using Wizard DNA purification resin (Promega), treated them again with sodium hydroxide, precipitated them with ethanol, and resuspended them in water.
Methylation-Specific PCR Assay
DNA methylation was evaluated with the use of the methylation-specific PCR assay, performed by three persons working independently. Each of them extracted DNA and performed all steps of the assay separately.23 A combined total of 889 samples of tumor and lymph-node tissues were examined. A multiplex-nested methylation-specific PCR assay was used for all samples, as described previously.24 The nested method initially amplifies bisulfite-modified DNA with the use of flanking PCR primers, without preferentially amplifying methylated or unmethylated DNA. The resulting fragment is then used as the template for the methylation-specific PCR assay. Primer sequences and PCR conditions for p16, MGMT, DAPK, RASSF1A, CDH13, ASC, and APC have been described previously,15,17,24,25,26 including conditions optimized to achieve specific detection of methylation in tumor tissue but not in normal lymphocytes (Table 1 in the Supplementary Appendix, available with the full text of this article at www.nejm.org).15,17,24,25,26 Placental DNA treated with SssI methyltransferase (New England Biolabs) was used as a positive standard. DNA from normal lymphocytes and water (bisulfite-modified water and unmodified water) were used as negative standards. PCR products were separated on 2% agarose gel or 6% nondenaturing polyacrylamide gel and were visually scored as methylated or unmethylated according to the presence or absence of a PCR product, respectively (Fig. 1 in the Supplementary Appendix).23,25,27 If any tumor block or lymph-node specimen was positive for methylation, all of the primary tumor or all associated lymph nodes in that nodal basin, respectively, were scored as positive.
Statistical Analysis
We verified histologic results and deaths or recurrent disease during the follow-up period by reexamining the original hospital records. The primary end point was time to recurrent local or distant disease, measured from the date of surgery to the time of cancer-related death or censoring. Data for controls who were alive and had no evidence of disease at the end of the study were censored for recurrence or death. All deaths of case patients were cancer-related, and no controls were lost to follow-up. Associations among prognostic factors, presence or absence of recurrence, and patient group were assessed by means of univariate and multivariate logistic-regression analysis. The association of risk factors with time-to-event or time-to-censoring end points was analyzed with the use of the log-rank test. Results of all models are reported as odds ratios with 95% confidence intervals. All statistical calculations were performed with the use of Stata statistical software. Two-sided P values of less than 0.05 were considered to indicate statistical significance.
We hypothesized that 40% or more of the case patients would have microscopic disease in the resected lymph nodes, as compared with 20% or less of controls, yielding an odds ratio of 2 for case patients. Under these assumptions, the study would have a statistical power of 80% to detect a significant effect among 168 case patients and controls, matched in a two-to-one ratio.
The authors designed the study; gathered, interpreted, and held the data; wrote the paper; made the decision to publish; and vouch for the completeness and accuracy of the data. There were no agreements concerning confidentiality of data between OncoMethylome Sciences and the authors.
Results
Characteristics of the Patients
Clinical and demographic variables were similar in case patients and controls (Table 1). On the basis of the American Society of Anesthesiologists Physical Status Classification System, case patients and controls were found to be equally fit for surgery. The most frequent site of recurrence was the ipsilateral lung (in 45.1% of patients), followed by metastasis to bone (13.7%), brain (11.7%), and mediastinum (11.7%). Although 14.7% of controls underwent sublobar resections, all pulmonary resections in these patients were done with curative intent.
|
The covariates of pathological stage, age, sex, histologic characteristics of the tumor, smoking status, and race were not associated with the risk of recurrence in patients with histologically negative lymph nodes (Table 2 in the Supplementary Appendix). Although pathological tumor stage showed the strongest association with recurrence, independently of other covariates, the association was not significant. For example, patients with stage IB disease (according to the 1986 classification of the American Joint Committee on Cancer28) had an adjusted odds ratio for recurrence of 1.71 (95% confidence interval [CI], 0.86 to 3.41; P=0.13), as compared with patients with smaller tumors and no pleural invasion (stage IA disease) (Table 2 in the Supplementary Appendix).
Gene Methylation and Recurrence
Methylation profiles of the seven genes were obtained from 727 of the 731 paraffin blocks, from 167 cases and controls. As compared with controls, the largest differences in the univariate distribution among case patients of the frequency of methylation in any type of tissue were found in four genes — p16, CDH13, RASSF1A, and APC — especially in tumors or mediastinal lymph nodes (Table 2). When p16 or CDH13 was methylated in the primary tumor, the adjusted odds ratio for recurrence was 3.50 (95% CI, 1.65 to 7.41; P=0.001) and 2.12 (95% CI, 0.98 to 4.59; P=0.06), respectively (Figure 1 and Table 3). When these same genes were methylated in regional lymph nodes, the odds ratio was 3.62 (95% CI, 1.41 to 9.32; P=0.008) and 1.99 (95% CI, 0.81 to 4.88; P=0.13), respectively. If methylation of p16 or CDH13 was found in mediastinal lymph nodes, the odds of recurrence was 4.67 (95% CI, 1.53 to 14.42; P=0.007) and 3.98 (95% CI, 1.22 to 13.01; P=0.02), respectively (Figure 1 and Table 3). Methylation of RASSF1A or APC in the tumor or mediastinal nodes was not significantly associated with recurrence (Figure 1).
|
|
|
We examined the methylation status of p16, CDH13, APC, and RASSF1A in an independent validation cohort of 20 patients (11 case patients and 9 controls) with stage I NSCLC (Table 1). In this cohort of limited size, we validated the gene pair in the original cohort with the highest odds ratio of recurrence among case patients, as compared with controls, when methylated in both tumor and lymph nodes — p16 and CDH13 — using univariate analysis (Table 3 in the Supplementary Appendix). Multivariate analyses of data in the combined original and validation cohorts showed that the estimated odds of recurrence associated with methylation of p16 and CDH13 in tumor and mediastinal lymph nodes was 25.25 (95% CI, 2.53 to 252.35; P=0.006) for case patients as compared with controls (Table 3).
Kaplan–Meier Estimates
Kaplan–Meier plots indicated that methylation of one or more of four genes — p16, CDH13, RASSF1A, and APC — in any sample from the patient was related to the duration of recurrence-free survival (Figure 2A through 2D). For example, the 5-year recurrence-free survival rates for no methylated genes, one to two methylated genes, and three to four methylated genes in the mediastinal lymph nodes were 77.3% (95% CI, 61.9 to 87.1), 51.4% (34.0 to 66.4), and 30.0% (7.1 to 57.8), respectively (P<0.001) (Figure 2C). Among patients with stage I NSCLC, the 5-year recurrence-free survival rate in the group of patients with two or more of the four methylated genes in the primary tumor and mediastinal lymph nodes was 27.3% (95% CI, 6.5 to 53.9), as compared with 65.3% (53.1 to 75.0) in the group with fewer than two methylated genes at those sites (P<0.001) (Figure 2D).
|
|
In the original and validation cohorts combined (total, 187 patients), the methylation status of p16 and CDH13 was assayed in primary tumor and mediastinal lymph-node specimens from 91 patients (41 case patients and 50 controls). Of the 11 patients in whom methylation of the two genes was found in both tissue types, 10 had recurrence of the tumor within 30 months after resection, 9 within 17 months, and 8 within 12 months. Patients with two or more methylated genes in tumor and mediastinal nodes also had a worse 5-year rate of recurrence-free survival (23.5% of patients; 95% CI, 7.3 to 44.9) than those with fewer than two methylated genes (63.5%; 95% CI, 52.4 to 72.7; P<0.001) (Figure 3C). When both p16 and CDH13 were methylated in the tumor and mediastinal nodes, there was a significantly lower rate of recurrence-free survival (9.1%; 95% CI, 0.5 to 33.3) than if p16 and CDH13 were unmethylated (61.2%; 95% CI, 49.7 to 70.9; P<0.001) (Figure 3D).
Discussion
Our study indicates that methylation of the promoter regions of certain genes in a resected NSCLC specimen is associated with recurrence of the tumor. A relation between gene methylation and tumor recurrence has been shown.29,30 The four genes of interest in our study are p16, CDH13, APC, and RASSF1A. They are involved in cell-cycle control (p16), invasion and metastasis (CDH13, APC), and Ras signaling (RASSF1A). Other studies of p16 expression or promoter-region methylation in lung cancer have focused mostly on the primary tumors,10,13,18,31,32,33 but we found that molecular examination of lymph nodes improves the assessment of risk of recurrence. The methylation of these genes in histologically normal regional lymph nodes probably indicates the presence of microscopically undetectable micrometastases. Immunohistochemical analyses, for example, may miss a rare cell in the background of normal tissue, whereas the methylation-specific PCR assay is sufficiently sensitive to detect a signal of DNA methylation.
The current method of assessment of the risk of recurrence in patients with stage I (T1–2N0) NSCLC is imprecise — one third of such tumors recur after curative surgery. Our results suggest that the detection of promoter methylation of certain genes may identify cells with a potential for metastatic spread not only within NSCLC but also in lymph nodes. It is possible that the methylated genes in lymph nodes represent tumor DNA that drained to the nodes through the lymphatic system, but this is unlikely because mediastinal nodes are located far from the lung, in the mediastinum. The correlation between short survival and the number of methylated genes in the regional and mediastinal lymph nodes supports the presence of micrometastases in those sites. Recent promising results for predicting the risk of lung cancer34 or its recurrence17 have been obtained by examining changes in gene methylation in sputum.34 Our study was retrospective and conducted at a single institution, and the number of patients studied was small. For these reasons, replication of our findings in a large, prospectively studied cohort is essential before the four-gene panel we investigated can be used in clinical practice.
Supported by grants from the National Cancer Institute (CA058184-10), the Commonwealth Foundation for Cancer Research, OncoMethylome Sciences, and the Hodson Trust.
Drs. Baylin and Herman report receiving consulting fees and research support from OncoMethylome Sciences. Under a licensing agreement between the Johns Hopkins University and this company, the methylation-specific PCR assay was licensed to OncoMethylome Sciences, and the university is entitled to a share of the royalties received by the company from sales of the licensed technology. Dr. Brock reports receiving research support from OncoMethylome Sciences. No other potential conflict of interest relevant to this article was reported.
We thank Dr. Xiaobu Ye, Rick Tracey, and Sharon Blackburn for their assistance with the figures, and Kathy Bender for helping to prepare the manuscript for submission.
Source Information
From Johns Hopkins Hospital, Baltimore (M.V.B., C.M.H., E.O.-M., Y.H., M.G., S.A., S.G., S.P., E.G., G.P., K.P., S.C.Y., S.B.B., J.G.H.); and the Lovelace Respiratory Research Institute, Albuquerque, NM (S.A.B.).
Dr. Brock, Mr. Hooker, and Dr. Ota-Machida contributed equally to this article.
Address reprint requests to Dr. Brock at Johns Hopkins Medical Institutions, 600 N. Wolfe St., Blalock 240, Baltimore, MD 21287, or at mabrock{at}jhmi.edu.
References
| |||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
