Recent evidence suggests that tumor cells may release DNA into the circulation, which is enriched in the serum and plasma, allowing detection of ras and p53 mutations and microsatellite alterations in the serum DNA of cancer patients. We examined whether aberrant DNA methylation might also be found in the serum of patients with non-small cell lung cancer. We tested 22 patients with non-small cell lung cancer using methylation-specific PCR, searching for promoter hypermethylation of the tumor suppressor gene p16, the putative metastasis suppressor gene death-associated protein kinase, the detoxification gene glutathione S-transferase P1, and the DNA repair gene O6-methylguanine-DNA-methyltransferase. Aberrant methylation of at least one of these genes was detected in 15 of 22 (68%) NSCLC tumors but not in any paired normal lung tissue. In these primary tumors with methylation, 11 of 15 (73%) samples also had abnormal methylated DNA in the matched serum samples. Moreover, none of the sera from patients with tumors not demonstrating methylation was positive. Abnormal promoter methylation in serum DNA was found in all tumor stages. Although these results need to be confirmed in larger studies and in other tumor types, detection of aberrant promoter hypermethylation of cancer-related genes in serum may be useful for cancer diagnosis or the detection of recurrence.
Lung cancer is one of the most common malignancies in the world and is the leading cause of cancer-related deaths in the United States (1) . The most effective treatment for NSCLC 3 is surgical resection, but this modality is limited by the fact that 65% of patients have advanced disease at the time of diagnosis (2) . Many patients undergoing surgical resection will ultimately die of recurrent NSCLC, suggesting the frequent presence of occult metastatic disease (2) . This failure rate is dependent on the stage of the disease, however. For example, the 5-year survival rate following surgical resection is ∼70% for patients with stage I NSCLC but drops to only 30% in patients with stage IIIa disease (3) . Thus, earlier detection of lung cancer would facilitate more effective management of the disease. Standard diagnostic techniques rely on radiological images with a tumor size detection limit of ∼1 cm, which represents 109 cells or a mass of 1 g (4) . Conventional sputum cytology and bronchoalveolar lavage have not been proven effective as screening tools. Therefore, it is important to develop new methods that provide sensitive and reliable diagnostic and prognostic markers of NSCLC.
Nanogram quantities of DNA circulating in blood are present in healthy individuals (5 , 6) . Previous studies have proposed that tumor DNA is also released in plasma and serum (7 , 8) . However, the serum of cancer patients is enriched in DNA, containing, on average, ∼4 times the amount of free DNA, compared to normal controls (9) . On the basis of these observations, several studies have shown that is possible to identify microsatellite alterations in the plasma and serum DNA of patients with head and neck carcinoma (10) and small cell lung carcinoma (11) . In addition, p53 and ras gene mutations have been detected in the plasma and serum of patients with colorectal (12, 13, 14) and pancreatic (15 , 16) carcinomas and hematological malignancies (17) .
In addition to the molecular genetic alterations described above, silencing of tumor suppressor genes by promoter hypermethylation is a common feature in human cancer (18) . Hypermethylation of normally unmethylated CpG islands in the promoter regions of many cancer key genes, including p16, p15, E-cadherin, VHL, and hMLH1, correlates with its loss of transcription in human tumors (19 , 20) . In primary lung carcinomas, the inactivation of the tumor suppressor gene p16 (18 , 21) , the DNA repair gene MGMT (22) , and the detoxifying gene GSTP1 (23) by promoter hypermethylation have been well described. In addition, loss of expression of the DAP kinase gene, reactivated by the demethylating agent 5′-azadeoxycytidine (24) , has also been correlated with metastatic potential in experimental lung cancer models (25) . On the basis of these observations, here, we examined 22 NSCLC patients using the sensitive MSP technique for abnormal promoter hypermethylation in p16, DAP kinase, GSTP1, and MGMT, in primary tumors and paired normal tissue and serum samples.
Materials and Methods
Sample Collection and DNA Preparation.
Normal lung, primary NSCLC, and corresponding serum were obtained from each of the 22 patients included in the study. Serum samples were obtained prior to surgery or at the time of surgery but before the chest incision was made. All patients were diagnosed at Germans Trias i Pujol Hospital (Badalona, Spain). DNA from normal and tumor tissue was isolated using previously published methods (26) . H&E-stained sections from each tumor sample were histologically examined, and only those that were verified to contain >50% neoplastic cells by section evaluation were selected for the study. To obtain serum, blood specimens from each patient were incubated at 36°C overnight and then centrifuged at low speed for 20 min. The serum was stored at −80°C before DNA extraction. Serum DNA was partially purified by phenol-chloroform extraction and ethanol precipitation (10) . After purification, 1 ml of serum yielded an average of 50 ng of DNA (10) .
DNA methylation patterns in the CpG island of the MGMT gene were determined by chemical modification of unmethylated but not the methylated cytosines to uracil and subsequent PCR using primers specific for either methylated or the modified unmethylated DNA (27) . Primer sequences of MGMT were for the unmethylated reaction 5′-ATG GTA TGG ATT TTA GTG TTG GTT-3′ (sense) and 5′-ACT ATA CCT AAA CTC AAC AAA CAC CA-3′ (antisense) and for the methylated reaction 5′-GTA TGG ATT TTA GCG TCG GTC-3′ (sense) and 5′-TAA ACT CGA CGA ACG CCG-3′ (antisense). Primer sequences of DAP kinase were for the unmethylated reaction 5′-GGA GGA TAG TTG GAT TGA GTT AAT GTT-3′ (sense) and 5′-CAA ATC CCT CCC AAA CAC CAA-3′ (antisense) and for the methylated reaction 5′-GGA TAG TCG GAT CGA GTT AAC GTC-3′ (sense) and 5′-CCC TCC CAA ACG CCG A-3′ (antisense). Primer sequences of GSTP1 were for the unmethylated reaction 5′-GAT GTT TGG GGT GTA GTG GTT GTT-3′ (sense) and 5′-CCA CCC CAA TAC TAA ATC ACA ACA-3′ (antisense) and for the methylated reaction 5′-TTC GGG GTG TAG CGC TCG TC-3′ (sense) and 5′-GCC CCA ATA CTA AAT CAC GAC G-3′ (antisense). Primer sequences for p16 have been described previously (27) . The annealing temperatures were 60°C for p16 and DAP kinase and 59°C for MGMT and GSTP1. PCR conditions for the four genes were as follows: 95°C for 5 min; then 35 cycles of 95°C for 30 s, the specific annealing temperature for 30 s, and 72°C for 30 s; and a final extension of 4 min at 72°C. All PCR amplification was performed using an Omnigene thermocycler (Hybaid) with tube control for accurate annealing temperatures. Hot start was used. Human placental DNA was treated in vitro with excess SssI methyltransferase (New England Biolabs), to generate completely methylated DNA at all CpGs and was used as positive control for methylated alleles of each gene. DNA from normal lymphocytes was used as the control for unmethylated alleles. Some samples were analyzed for p16 methylation using 40 cycles of amplification, with concordant results.
Briefly, 1 μg of DNA was denatured by NaOH and modified by sodium bisulfite. DNA samples were then purified using Wizard DNA purification resin (Promega), again treated with NaOH, precipitated with ethanol, and resuspended in water. Controls without DNA were performed for each set of PCR. Ten μl of each PCR were directly loaded onto nondenaturing 6% polyacrylamide gels, stained with ethidium bromide, and visualized under UV illumination.
We found that 68% (15 of 22) of NSLCC primary tumors exhibited abnormal promoter hypermethylation in at least one gene (Table 1) ⇓ . p16 was hypermethylated in 41% (9 of 22) of the cases, a similar percentage to that previously described in lung cancer (18) . The presence of abnormal methylation of p16 was found in all of the stages, consistent with the finding that p16 aberrant methylation is an early event in lung carcinogenesis (21) . The incidence of DAP kinase promoter hypermethylation was 23% (5 of 22 cases), suggesting that this is a common epigenetic alteration in lung cancer. Consistent with this data, the treatment with the demethylating agent 5′-azadeoxycytidine restores the expression of DAP kinase in some cancer cell lines (24) , strongly suggesting the methylation-mediated inactivation of DAP kinase in human neoplasia. Abnormal GSTP1 promoter methylation was found in 9% (2 of 22) of cases, consistent with previous findings (23) . Finally, MGMT promoter hypermethylation was present in 27% (6 of 22) of the cases. We recently analyzed a wide spectrum of tumor types and found a similar incidence of MGMT hypermethylation in primary NSCLC (22) . Thus, the incidence of hypermethylation of each gene was consistent with previous work in NSCLC. Moreover, none of the 22 paired lung normal tissues exhibited abnormal promoter hypermethylation of any gene.
Seventy-three % (11 of 15) of the patients showing hypermethylation in tumor DNA also demonstrated abnormal methylation in serum DNA. Representative MSP analyses for p16, DAP Kinase, GSTP1, and MGMT in tumor and serum DNA, as compared with paired normal lung are shown in Fig. 1 ⇓ . Promoter hypermethylation in serum DNA was 33% (3 of 9 cases) for p16, 80% (4 of 5 cases) for DAP kinase, 50% (1 of 2 cases) for GSTP1, and 66% (4 of 6 cases) for MGMT (Table 2) ⇓ . Each of the 11 patients who had abnormal promoter hypermethylation in the serum DNA demonstrated identical alterations in the primary tumor DNA (Table 1) ⇓ . Importantly, only patients whose tumors harbored a hypermethylated marker showed aberrant methylation of the same genes in serum. After abnormal promoter hypermethylation analysis of all specimens was completed, clinical data were correlated with the results. The clinical stage, histology, follow-up (development of recurrence), and clinical outcome in the 22 patients and the methylation status of each gene studied are listed in Table 1 ⇓ . The presence of hypermethylation occurred with all stages, including stage I. Detection of abnormal methylation in serum DNA did not appear to be related to an early recurrence in this small series.
Previous studies have identified several tumor-specific genetic alterations in the plasma and serum DNA of cancer patients. Microsatellite shifts (low-level instability) and loss of heterozygosity has been detected in plasma or serum DNA from patients with small cell lung carcinoma (11) and head and neck carcinoma (10) but not in serum of colorectal cancer patients (14) . Mutant K-ras and p53 DNA have also been identified in plasma or serum of patients with colorectal, pancreatic, and hematological neoplasms (12, 13, 14, 15, 16, 17) . Our study demonstrates, for the first time, that it is possible to detect abnormal promoter hypermethylation in serum DNA from cancer patients.
Methylation is the main epigenetic modification in humans (18) , and changes in methylation patterns play an important role in tumorigenesis. In particular, hypermethylation of normally unmethylated CpG islands in many tumor suppressor genes correlates with loss of expression (18) . The recent development of the MSP technique (27) has simplified the study of genes inactivated by promoter hypermethylation in human cancer and its high sensitivity, approaching 1 methylated gene copy in 1000 unmethylated copies in dilution experiments (27) , support the potential power of this approach for molecular detection. Moreover, the MSP technique has demonstrated great specificity in our study, because we did not find any abnormal methylation in serum DNA if this alteration was not present in the primary tumor (Table 1) ⇓ . In other studies, genetic alterations were sometimes detected in the serum, whereas they were absent in the primary specimens (11 , 13) . Our approach with MSP avoids the use of expensive sequencing reagents and radioactivity, allowing the study of multiple markers by rapid analysis. Unlike some tumor suppressor gene mutations (i.e., p53), MSP allows the detection of aberrant DNA methylation as a molecular detection method without prior analysis of the primary tumor. Furthermore, the analysis of abnormal promoter methylation status of several genes in a timely and economic fashion, combined with the study of previously described genetic alterations, may allow the detection of almost all patients with circulating tumor DNA.
Some markers in serum DNA, such as p16 methylation, may be helpful in the monitoring of multiple tumor types because inactivation of p16 by methylation is a common feature in human neoplasia (18) . However, depending upon the particular tumor type, other markers may also be detected in serum DNA. For example, promoter hypermethylation of hMLH1, present in the majority of sporadic colon and endometrial carcinomas with microsatellite instability (19 , 20) , and GSTP1, aberrantly methylated in ∼90% of prostate carcinomas (18) , may be ideal markers for these particular tumor types. In addition, the particular spectrum of genes with promoter hypermethylation in the serum DNA may provide us clues about the biological behavior of the primary tumor. For example, the finding of the inactivation of a putative metastasis tumor suppressor gene like DAP kinase may suggest the metastatic potential of a tumor (25) . The detection of the methylation-mediated silencing of the MGMT gene (22) , which provides resistance to alkylating agents, may modify the chemotherapeutic treatment.
Our results indicate that abnormal promoter hypermethylation of tumor suppressor genes is readily detectable in the serum DNA of cancer patients using MSP analysis. By using the right combination of target genes, MSP may detect aberrant methylated DNA across a broad spectrum of neoplasms. This approach allows sensitive and accurate detection of circulating tumor DNA and may have multiple applications in the follow-up and management of cancer patients.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by NIH Grant 5P50CA58184. M. E. and M. S-C. are recipients of Spanish Ministerio de Educacion y Cultura Awards. J. G. H. is a Valvano Foundation Scholar. S. B. B. and J. G. H. receive research funding and are entitled to sales royalties from Oncor, which is developing products related to research described in this paper. The terms of this arrangement have been reviewed and approved by The Johns Hopkins University in accordance with its conflict of interest policies.
↵2 To whom requests for reprints should be addressed, at Tumor Biology, The Johns Hopkins Oncology Center, 424 North Bond Street, Baltimore, MD 21231. Phone: (410) 955-8506; Fax: (410) 614-9884; E-mail:
↵3 The abbreviations used are: NSCLC, non-small cell lung carcinoma; MGMT, O6-methylguanine-DNA-methyltransferase; GSTP1, glutathione S-transferase P1; DAP kinase, death-associated protein kinase; MSP, methylation-specific PCR.
- Received September 9, 1998.
- Accepted November 9, 1998.
- ©1999 American Association for Cancer Research.