Summary
Detection of lung cancer at early stages could potentially increase survival rates. One promising approach is the application of suitable lung cancer-specific biomarkers to specimens obtained by non-invasive methods. Thus far, clinically useful biomarkers that have high sensitivity have proven elusive. Certain genes, which are involved in cellular pathways such as signal transduction, apoptosis, cell to cell communication, cell cycles and cytokine signaling are down-regulated in cancers and may be considered as potential tumor suppressor genes. Aberrant promoter hypermethylation is a major mechanism for silencing tumor suppressor genes in many kinds of human cancers. Using quantitative real time PCR, we tested 11 genes (3-OST-2, RASSF1A, DcR1, DcR2, P16, DAPK, APC, ECAD, HCAD, SOCS1, SOCS3) for levels of methylation within their promoter sequences in non-small cell lung cancers (NSCLC), adjacent non-malignant lung tissues, in peripheral blood mononuclear cells (PBMC) from cancer free patients, in sputum of cancer patients and controls. Of all the 11 genes tested 3-OST-2 showed the highest levels of promoter methylation in tumors combined with lowest levels of promoter methylation in control tissues. 3-OST-2 followed by, RASSF1A showed increased levels of methylation with advanced tumor stage (P<0.05). Thus, quantitative analysis of 3-OST-2 and RASSF1A methylation appears to be a promising biomarker assay for NSCLC and should be further explored in a clinical study. Our preliminary data on the analysis of sputum DNA specimens from cancer patients further support these observations.
Keywords: Real time PCR, Tumor suppressor gene, Non-small cell lung cancer
1. Introduction
Lung cancer is the leading cause of cancer-related deaths in the world [1]. The survival rates for lung cancers have changed little over the past two decades [2]. A major factor in the high mortality of lung cancer patients is the presence of metastases in approximately two-thirds of patients at the time of diagnosis [3]. It has been estimated that detection of lung cancer at earlier stages could potentially increase survival rates by 10–50-fold [4]. Lung cancer screening by chest X-ray and sputum cytology have proven ineffective in increasing patient survival [3,5], leading to the search for more sensitive and specific tests [6]. One promising approach is the identification of lung cancer-specific biomarkers and non-invasive methods for the detection of these biomarkers at an early stage. However, there is lack of a sufficient number of biomarkers especially for NSCLC (~80% of the total lung cancer cases) that are both cancer specific and highly sensitive. Thus, investigations into identifying biomarkers that can effectively separate cancer cells from normal cells with high sensitivity are needed.
Aberrant promoter hypermethylation is a major mechanism for silencing tumor suppressor genes in many kinds of human cancers [7]. Gene promoter hypermethylation potentially provides a non-invasive screen for early cancer detection and offers opportunities to develop novel therapies for reversing aberrant methylation [8]. The development of the real time methylation specific PCR has simplified the study of the genes inactivated by promoter hypermethylation in human cancer [9,10]. It has the advantage of increasing specificity and sensitivity (attributable to the use of an internally binding fluorogeneic hybridization probe for each gene) over conventional PCR and of great importance, provides quantification (qPCR). Recent publications have demonstrated the presence of promoter hypermethylation of various genes in clinical specimens containing exfoliated tumor cells (such as malignant effusions, sputum, serum, etc.) [11–14]. We believe that markers that are tumor specific, with high levels of methylation in tumor but low or absent methylation in adjacent non-malignant tissue and in blood mononuclear cells will prove to be clinically more useful. There is severe shortage of markers with high frequency and high levels of methylation for NSCLC that can also separate cancerous from adjacent non-malignant tissue.
Certain genes, which are involved in cellular pathways such as signal transduction (3-OST-2), apoptosis (DcR1, DcR2, DAPK), adhesion (APC, E-CAD, H-CAD), cell cycle (P16, RASSF1A), cytokine signaling (SOCS1, SOCS3), etc. are epigenetically inactivated in cancers and may be considered as potential tumor suppressor genes (Table 1) [15–32].
Table 1.
Summary data of genes analysed for methylation with quantitative PCR
| Gene | Gene name | Chromosomal location | Known function | Refs. |
|---|---|---|---|---|
| 3OST2 | Heparan sulfate (glucosamine) 3-O-sulfotransferase | 6p12 | Heparan sulfate biosynthesis signal transduction involved in multiple physiological pathways | [15,16] |
| DcR1 | Decoy receptor 1 | p22 | Apoptosis | [18,19] |
| DcR2 | Decoy receptor 2 | 8p22 | Apoptosis | [18,19] |
| RASSF1A | Ras association domain family protein1a | 3p21 | Growth signals | [17] |
| P16 | Cyclin-dependent kinase inhibitor 2A | 9p21 | Cell cycle | [20] |
| APC | Adenomatous polyposis coli gene | 5q21 | Adhesion | [21] |
| DAPK | Death-associated protein kinase | 9q34 | Apoptosis | [22] |
| SOCS1 | Suppressor of cytokine signaling-1 | 16q13 | Suppression of cytokine signaling | [23] |
| SOCS3 | Suppressor of cytokine signaling-3 | 17q25 | Suppression of cytokine signaling | [23] |
| H-CAD | H-Cadherin | 16q24 | Adhesion | [25] |
| E-CAD | E-Cadherin | 16q22 | Adhesion | [25] |
In an effort to identify new clinically useful markers for lung cancer detection, we used qPCR assays to determine the methylation frequencies of 11 genes implicated in critical cellular pathways, for their ability to separate lung cancer from normal cells. In a preliminary study, we applied our panel of markers to sputum specimens from patients suspected of having lung cancer.
2. Materials and methods
Surgically resected non-small lung cancers (NSCLC, n=40) and their adjacent normal lung tissues were obtained from MD Anderson Cancer Center, Houston TX. Institutional review board permission and informed consent were obtained from all patients. Tumor staging was performed according to the International Union Against cancer (UICC) TNM classification. Peripheral blood mononuclear cells (PBMC) were obtained from 40 healthy individuals with a family history of cancer. Three day pooled preferably morning sputum samples (uninduced) were collected in saccomanno’s fixative (2% polyethylene glycol in 50% ethanol). Sputum samples were obtained from 13 patients with NSCLC and 25 individuals without lung cancer (COPD patients and patients treated for NSCLC, subsequent sputum collection and follow-up of minimal 1 year without signs of lung cancer, i.e. normal controls, without malignancy) in the Canisius Wilhelmina Hospital, Nijmegen, the Netherlands (Table 2).
Table 2.
Clinico-pathological information and quantitative methylation data for sputum samples
| Sputum sample no. | Age | Gender | Diagnosis | Stage | m3-OST-2 | mRASSF1A | mP16 | mAPC |
|---|---|---|---|---|---|---|---|---|
| CWDNA10 | 71 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA11 | 68 | Female | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA14 | 80 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA15 | 54 | Female | No lung cancer | NA | 0 | 0 | 3.99 | 0.28 |
| CWDNA16 | 63 | Female | No lung cancer | NA | 3.21 | 0.42 | 0 | 0 |
| CWDNA17 | 78 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA20 | 64 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA22 | 71 | Male | No lung cancer | NA | 0 | 0 | 7.67 | 0 |
| CWDNA23 | 54 | Female | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA26 | 79 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA27 | 66 | Male | No lung cancer | NA | 0 | 1.12 | 0 | 0 |
| CWDNA29 | 52 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA33 | 54 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA35 | 63 | Female | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA36 | 80 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA37 | 69 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA39 | 54 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA24 | 45 | Female | Breast carcinoma | NA | 0 | 0 | 0 | 0 |
| CWDNA2 | 77 | Male | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA6 | 50 | Female | No lung cancer | NA | 0.85 | 0 | 0 | 0 |
| CWDNA8 | 61 | Female | No lung cancer | NA | 0 | 0 | 0 | 0 |
| CWDNA12 | 73 | Male | Prior lung cancer* | ND | 0 | 0 | 0 | 0 |
| CWDNA18 | 68 | Male | Prior lung cancer* | IA | 0 | 0 | 0 | 0 |
| CWDNA30 | 73 | Male | Prior lung cancer* | IB | 2.98 | 0 | 0 | 0 |
| CWDNA1 | 58 | Male | Prior lung cancer* | IIIA | 0 | 0 | 0 | 0 |
| CWDNA3 | 81 | Male | Squamous cell carcinoma | IIB | 0 | 0 | 512.86 | 0 |
| CWDNA4 | 71 | Male | Adenocarcinoma | IV | 0 | 159.59 | 0 | 5.4 |
| CWDNA5 | 79 | Male | Squamous cell carcinoma | IIIA | 33.11 | 125.6 | 0 | 0.32 |
| CWDNA7 | 55 | Male | Adenocarcinoma | IV | 100 | 0 | 0 | 0 |
| CWDNA9 | 68 | Male | Squamous cell carcinoma | IIIA | 0 | 0 | 0 | 0 |
| CWDNA13 | 83 | Male | Adenocarcinoma | IV | 0 | 0 | 0 | 0 |
| CWDNA19 | 78 | Male | Squamous cell carcinoma | IIIA | 0 | 50.84 | 29.32 | 4.86 |
| CWDNA21 | 52 | female | Adenocarcinoma | IA | 0 | 0 | 0 | 0 |
| CWDNA25 | 77 | Male | Squamous cell carcinoma | IIIB | 0 | 0 | 0 | 0 |
| CWDNA31 | 39 | Male | Adenocarcinoma | IIB | 4.62 | 36.73 | 14.45 | 0.91 |
| CWDNA32 | 68 | Male | Squamous cell carcinoma | IA | 13.49 | 0 | 0 | 0 |
| CWDNA38 | 65 | Male | Squamous cell carcinoma | ND | 0 | 35.48 | 0 | 0 |
| CWDNA40 | 73 | Male | Squamous cell carcinoma | IV | 0.42 | 0 | 0 | 0 |
All the controls were free of lung cancer at the time the sputum was collected. Four contol patients (*) had lung cancer >1 year before the sputum was collected. Control patient CW24 had breast cancer at the time the sputum was collected. NA, not applicable; ND, information not available.
3. DNA extraction and methylation methods
Genomic DNA was extracted from cell lines, primary tumors and non-malignant cells by using a DNA extraction kit (Serological Corporation, Norcross, GA) [33]. Chemical modification of unmethylated (but not methylated) cytosines to uracil within CpG islands using sodium-bisulfite treatment was performed as described previously [34]. The modified DNA was used as a template for qPCR analysis. Genomic DNA was extracted from sputum samples according to the previously proposed protocol [35]. Dithiothreitol was used to dissolve mucus. Genomic DNA was extracted from sputum samples by using the QIAamp DNA blood mini kit (QIAgen, Venlo, the Netherlands). The DNA was bisulfite treated as described above [34]. The modified DNA was used as a template for qPCR analysis.
4. Semiquantitative real time PCR analysis
qPCR analysis was performed using the Chromo4 MJ Research Real time PCR system (Waltham, MA). Sodium bisulfite-treated genomic DNA was amplified by fluorescence-based real time MSP using TaqMan technology as described previously [9,10]. In brief, oligonucleotide primers were designed to specifically amplify bisulfite-converted DNA within the region of the test genes that correlates with loss of gene expression [27,36,37] and a probe designed to anneal specifically within the amplicon during extension (Table 3). The identities of the amplicons were confirmed through direct sequencing in forward and reverse directions. The primers used for APC and DAPK and E-CAD and RASSF1A were as previously published [9,13,14]. PCR assays are performed in a reaction volume of 25 μl. The final reaction mixture contained the forward and reverse primers (600 nmol/l of each); the probe (200 nmol/l); each of the four nucleotide phosphates (200 (mol/l for each); 5.5 mmol/l MgCl2; 1XTaqMan Buffer A; 1 U of Hot Star Taq DNA polymerase (Qiagen, Valencia, CA) and 1 μl bisulfite-converted genomic DNA (corresponds to ~50 ng initial DNA used for conversion). PCR was performed under the following conditions: 95 °C for 15 min, followed by 50 cycles at 95 °C for 15 s and 60 °C for 1 min. The non-methylated form of MYOD1 was used as an internal reference standard [9]. Amplification of MYOD1 is independent of its methylation status, whereas the amplification of test genes is proportional to the degree of cytosine methylation within the amplicon. Serial dilution of the Sss1 treatment (which converts all cytosines to their methylated forms) of human genomic DNA from normal human lymphocytes was used to create a standard curve. The fluorescence emission intensities (threshold cycle or Ct values) for the biomarker genes and MYOD1 were calculated using the intercept and the slope of the standard graph. The methylation ratio is defined as the ratio of the fluorescence emission intensity values for the PCR products of the biomarker gene to those of PCR products of MYOD1 multiplied by 100. The ratio is a measure for the relative level of methylation in an individual sample. Fig. 1 illustrates representative fluorescence curves generated from real time PCR analysis of lung tumors and adjacent normal lungs for m3-OST-2, mRASSF1A and MYOD1. The figure also illustrates standard curves for m3-OST-2 and mRASSF1A. The standard curves demonstrate a linear relationship between fluorescence intensity (Log Quantity) and cycle number (Ct cycle) over more than three logs. Similar standard curves were obtained for all genes analysed in this study including MYOD1 and they demonstrated good linear correlation (data not shown). Ct values from three to four independent runs of the same reactions differed by less than 1 cycle. To confirm the reliability of the assay system some of the tumors and adjacent non-malignant lungs were analyzed in a blinded fashion. In addition, a second, independent bisulfite treatment was performed on DNA from some of the tumors and adjacent non-malignant lungs and they were analyzed in a blinded fashion. These data were used to confirm the reproducibility of the assay. The data were found to be reproducible with both high and low methylation level samples analysed in our study. All the sputum DNA samples were coded and shipped from Canisius Wilhelmina Hospital, Nijmegen, the Netherlands and analysed in our laboratory in a blinded fashion.
Table 3.
Primers and TaqMan probes for real time MSP assays
| Gene symbol | Forward primer | TaqMan probe | Reverse primer |
|---|---|---|---|
| m3OST2 | TCGGCGTACGTAAGAGTTTGG | AGCGTTC-GAGTCGTTCGGTTGTTCG | ATCTCCCGATCCTAAACGA-TAAAA |
| mDcR1 | GGAGCGTTTTTTATCGTTAGGGA | CGGTCGTTTGATGGTCGAGG-TAGGGT | CCCGACGCCGTCCTAAAT |
| mDcR2 | GGGATAAAGCGTTTCGATCG | CGAGCGTTCGAGTAGGGCGT-TATTTAGG | TCTAATTCCCGACGCTATCCTAA |
| mRASSF1A | GCGTTGAAGTCGGGGTTC | ACAAACGCGAACCGAAC-GAAACCA | CCCGTACTTCGCTAACTT-TAAACG |
| mP16 | CGCAACCGCCGAACG | CGCGATCCGCCCCACCCT | TTTTTTCGTTAGTATCGCAG-GAAGA |
| mAPC | GAACCAAAACGCTCCCCAT | CCCGTCGAAAACCCGCCC-GATTA | TTATATGTCGGTTACGTGCGTT-TATAT |
| mDAPK | GGATAGTCGGATCGAGT-TAACGTC | TCGGTAATTCGTAGCGG-TAGGGTTTGG | CCCTCCCAAACGCCGA |
| mSOCS1 | GCGAGAGTTTCGATTGTTTTTTC | AGTTGTTGGAGTAT-TACGTGGCGGCG | CCCCAACATACGACGCG |
| mSOCS3 | TCGCGTTTTTTTTTTCGTAGTTT | TCGGGATGCGGTAGCGGTCG | CGCGACCTCCGCACA |
| mH-CAD | GGGATGTTATTTTCGCGGG | CGCGGTTAGTAGGGCGGGT-TAGGG | CCTtCCTAACGCTCCCTCGT |
| mE-CAD | AATTTTAGGTTAGAGGGT-TATCGCGT | CGCCCACCCGACCTCGCAT | TCCCCAAAACGAAACTAACGAC |
| MYOD1 | CCAACTCCAAATCCCCTCTCTAT | TCCCTTCCTATTCCTAAATC-CAACCTAAATACCTCC | TGATTAATTTAGATTGGGTTTA-GAGAAGGA |
Fig. 1.
Fluorescence curves generated from real time PCR reactions for methylated 3-OST-2 (A), methylated RASSF1A (B) and MYOD1(C). The data represent analysis of the 20 normal tumor pairs from our study. A (3-OST-2) and B (RASSF1A), the samples with the lowest Ct value (curves on the far left for both genes, Ct 35.64 for 3-OST-2 and Ct 32.18 for RASSF1A) represent positive controls (SS1DNA) Ct values for of all the adjacent non-malignant lung samples ranged from 42 to 50 (Ct value of 50 means below the threshold) both in 3-OST-2 and RASSF1A groups. The Ct values for methylation positive tumors were always below 42 both in 3-OST-2 and RASSF1A groups. The range of Ct values for MYOD1 of different tumor and normal lungs was much narrower, 31.18–37.88 (Ct value for MYOD of positive control was 36.47). Standard curves are shown for methylated 3-OST-2 (D) and methylated RASSF1A (E). The standard curves demonstrate a linear relationship between fluorescence intensity (log quantity) and cycle number (Ct cycle) over more than three logs.
5. Statistical analyses
Sensitivity and specificity of a given gene are measured as the fraction of tissues whose methylation level exceed a predetermined threshold value (test positive) among all lung cancer tissues and the fraction of tissues test negative among all non-malignant lung tissues. For this study, we had no prior cutoff value for calling a biomarker positive, therefore, we applied the receiver operating characteristic (ROC) curve analysis to determine the cutoff values. The ROC curves, a plot of the sensitivity vs. specificity across all possible cutoff values, were used to identify the accuracy of a marker in discriminating lung cancer from non-malignant lung or adenocarcinoma from squamous carcinoma. A marker with perfect discriminatory capacity (100% sensitivity and 100% specificity) has an ROC plot that pass through the upper left corner, and the higher the ROC curve the higher the accuracy of the test. To measure the accuracy of discrimination, we considered the area under the ROC curve (AUC). An AUC not significantly different from 0.5 indicates that the test is neither sensitive nor specific. The AUC and 95% confidence interval (95% CI) were calculated by a non-parametric method [38]. To combine several markers, we first fitted a logistic regression with binary outcome (cancerous vs. normal tissue, or adenocarcinoma vs. squamous carcinoma) and used a panel of markers as covariates. The resulting linear predictive score from the regression model was then used as a composite marker and the ROC curves and corresponding AUCs were calculated. Cross-validation and permutation were used to assess the accuracy summaries in this case to correct for over fitting.
We also correlated quantitative methylation data for different genes with tumor stage using the Mann–Whitney U-test, which does not require parametric assumptions on the distribution of quantitative methylation. P-values<0.05 were considered significant.
6. Results
We determined the methylation profile of 40 NSCLC tumors (22 adenocarcinoma and 18 squamous carcinomas) and adjacent non-malignant lung tissues using qPCR. Additionally, PBMCs from 40 healthy subjects recruited for genetic epidemiology studies were also quantitatively analyzed. Fig. 2 shows the quantitative methylation data for 6 representative genes of the 11 genes tested. We found aberrant methylation of at least 1 gene in 39 of 40 (98%) of primary tumors. Frequencies of methylation in primary tumors ranged from 36 of 40 (90%) for 3-OST-2 to 16 of 40 (40%) for DcR2. Frequencies of methylation in adjacent non-malignant tissue ranged from 23 of 40 (58%). for APC to 1 of 40 (2.5%) for DAPK. 3-OST-2 had the highest median QR, 14.19 (0–975)) followed by APC, 4.1 (0–331.13). However, 3-OST-2 methylation had the lowest median QR, 0 (0–4.1) in adjacent non-malignant tissue while APC had one of the highest median QR, 0.29 (0–13.95) in adjacent non-malignant tissue. Thus, 3-OST-2 methylation appeared to be both highly tumor specific with a high degree of sensitivity. For all the genes, we set the highest methylation level in non-malignant tissues as the cutoff value to determine methylation positive samples for lung cancer tissues. With this cutoff, the frequency of aberrant promoter methylation in tumors ranged from 70% for 3-OST-2 to 23% for DcR1 (Table 4). For SOC3 and E-CAD the cutoff values were too high for them to be useful tumor markers. The median QRs for all the genes were 0 in PBMCs. However, H-Cad, E-CAD, SOCS1 and SOCS3 did show methylation in some cases (<25%) although at much lower levels than in tumors. Interestingly, APC was unique in the sense that it showed high levels of methylation in non-malignant lung but showed no methylation in PBMCs from cancer-free individuals.
Fig. 2.
Methylation levels of m3-OST-2 (A), mRASSF1A (B), mDcR1 (C), mAPC (D), mSOCS1 (E) and mSOCS3 (F) in NSCLC (tumor), adjacent non-malignant lung and PBMCs from cancer-free individuals. Methylation levels were quantitated by semiquantitative real time PCR. Real time analysis was performed as described in Section 2. Quantitative ratio is defined as the ratio of the fluorescence emission intensity values for the PCR products of the biomarker gene to those of PCR products of MYOD1 multiplied by 100. The ratio is a measure for the relative level of methylation in an individual sample. Because values are expressed on a log scale, completely negative values are expressed as values of 0.01. The solid horizontal bar represents the median value for each group. No solid horizontal bar was shown when the median value for a group is zero.
Table 4.
Median quantitative ratios for methylation of genes in NSCLC and adjacent non-malignant tissue
| Methylated gene | Median quantitative ratio for NSCLC (n=40) | Median quantitative ratio for adjacent non-malignant lung (n=40) | Median quantitative ratio for PBMCs from cancer-free individuals (n=40) | % Samples +ve for methylation in NSCLC | Cutoff value |
|---|---|---|---|---|---|
| 3-OST-2 | 14.19(0–975) | 0(0–4.1) | 0(0–0) | 28/40(70%) | 4.1 |
| RASSF1A | 0.69(0–883.10) | 0(0–4.013) | 0(0–0) | 18/40(45%) | 4.013 |
| P16 | 2.00(0–753.36) | 0(0–2.90) | 0(0–0) | 18/40(45%) | 2.9 |
| DcR1 | 0.19(0–72.44) | 0(0–0.96) | 0(0–0) | 9/40(23%) | 0.96 |
| DcR2 | 0(0–481.94) | 0(0–4.37) | 0(0–0) | 11/40(28%) | 4.37 |
| APC | 4.1(0–331.13) | 0.29(0–13.95) | 0(0–0) | 12/40(30%) | 13.95 |
| DAPK | 0.10(0–4.70) | 0(0–0.22) | 0(0–0) | 17/40(43%) | 0.22 |
| SOCS-1 | 2.72(0–111.39) | 0(0–6.3) | 0(0–16) | 13/40(33%) | 6.3 |
| H-CAD | 2.92(0–148.94) | 0(0–7.59) | 0(0–3.6) | 11/40(28%) | 7.58 |
| SOCS-3 | 1.88(0–912.04) | 0.21(1–186.2) | 0(0–11.73) | 2/40(5%) | 186.2 |
| E-CAD | 3.98(0–467.7) | 2.56(0–39.8) | 0(0–18.84) | 5/40(12%) | 39.8 |
Cutoff value=the highest QR value in adjacent non-malignant lung. For SOCS3 and E-CAD cutoff value is too high and thus no meaningful conclusion can be drawn about % samples+ve for methylation in NSCLC.
Fig. 3 shows the ROC curve and AUC (95% CI) which examine the accuracy of the markers to separate cancer from the adjacent non-malignant tissue. The data show that except for SOCS3 and E-CAD all other genes have the discriminatory capacity at separating cancer from adjacent normal lung tissue. Additionally, the combination of 3-OST-2, DAPK and DcR1 best separated cancer from corresponding normal tissue. The accuracy of each gene at separating cancer from PBMCs of cancer free subjects is assessed with a ROC curve and the corresponding AUC (95% CI) (Fig. 4). The data show that all the genes separate cancer from PBMCs, indicative of no or negligible methylation of all the genes in PBMCs. Fig. 5 shows ROC curve and AUC (95% CI) which demonstrates the ability of the genes at separating adenocarcinoma from squamous carcinoma. The combination of DcR1 and 3-OST-2 best separated adenocarcinoma from squamous cell carcinoma.
Fig. 3.
ROC curves for each of the 11 genes (methylated form) at separating cancer from the adjacent non-malignant lung. ROC curves are plots of the true positive rate (vertical axis) against the false positive rate (horizontal axis) for the different possible cutoff points of a diagnostic test. The closer the curve follows the left-hand border and then the top border of the ROC space, the more accurate the test.
Fig. 4.
ROC curves for each of the 11 genes (methylated form) at separating cancer from PBMCs derived from healthy individuals with family history of cancer. The description of ROC curves as in Fig. 3 and in Section 2.
Fig. 5.
ROC curves for each of the 11 genes (methylated form) at separating adenocarcinoma from squamous cell carcinoma. The description of ROC curves as in Fig. 3 and in Section 2.
We also correlated quantitative methylation data for different genes with tumor stage using the Mann–Whitney U-test. For 3-OST-2 increased methylation levels (QR) were observed with increase in stage (stage 1–2, P=0.03), stage 1 and 3, (P=0.0002), stage 1 and 2 combined to stage 4 (P=0.0003) (Table 5). Besides 3-OST-2, increased methylation levels (QR) were also observed to increase with stage for RASSF1A (stage 1 and 2 (P=0.05), stage 1 and 3 (P=0.03), stage 1 and stages 2–4 combined (P=0.03) (Table 6). No statistically significant correlation between methylation levels tumor stage observed for other genes.
Table 5.
Relation between QR of methylated 3-OST-2 and tumor stage
| Stage I | Stage II | Stage III | Stage IV | Stages II–IV |
|---|---|---|---|---|
| 1.36 | 5.74 | 7.9 | 68.39 | 5.74 |
| 18.26 | 97.72 | 316.23 | 114.81 | 97.72 |
| 4 | 758 | 15.49 | 758 | |
| 3.1 | 18.62 | 975 | 18.62 | |
| 4.37 | 12.88 | 219.7 | 12.88 | |
| 1.89 | 45.1 | 501.2 | 45.1 | |
| 0 | 37.16 | 35.48 | 37.16 | |
| 1.7 | 776.28 | 38.9 | 776.28 | |
| 0 | 93.33 | 169.8 | 93.33 | |
| 7.32 | 0 | 5.37 | 0 | |
| 1.2 | 1 | 1 | ||
| 0.6 | 0 | 0 | ||
| 3.21 | 7.9 | |||
| 36.98 | 316.23 | |||
| 3.5 | 15.49 | |||
| 975 | ||||
| 219.7 | ||||
| 501.2 | ||||
| 35.48 | ||||
| 38.9 | ||||
| 169.8 | ||||
| 5.37 | ||||
| 68.39 | ||||
| 114.81 | ||||
| Mean | ||||
| 5.8 | 153.8 | 228.5 | 179.7 | |
Mann–Whitney U-test P<0.05. Stage I vs. stage II P=0.0315, stage I vs. stage III P=0.0002, stage I vs. stages II–IV P=0.0003. For one NSCLC stage was not known.
Table 6.
Relation between QR of methylated RASSF1A and tumor stage
| Stage I | Stage II | Stage III | Stage IV | Stages II–IV |
|---|---|---|---|---|
| 37.15 | 0 | 69.66 | 0 | 0 |
| 0 | 22.8 | 2 | 0 | 22.8 |
| 16.6 | 0 | 92.89 | 0 | |
| 1.37 | 7.94 | 0 | 7.94 | |
| 0 | 147.57 | 0 | 147.57 | |
| 0 | 0 | 416.86 | 0 | |
| 67.44 | 72.11 | 418.78 | 72.11 | |
| 0 | 222.3 | 51.62 | 222.3 | |
| 65.05 | 883.1 | 0 | 883.1 | |
| 0 | 0 | 208.93 | 0 | |
| 0 | 11.07 | 11.07 | ||
| 0 | 222.33 | 222.33 | ||
| 0 | 69.66 | |||
| 0 | 2 | |||
| 0 | 92.89 | |||
| 0 | ||||
| 0 | ||||
| 416.86 | ||||
| 418.78 | ||||
| 51.62 | ||||
| 0 | ||||
| 208.93 | ||||
| 0 | ||||
| 0 | ||||
| Mean | ||||
| 12.5 | 132 | 126 | 119 | |
Mann–Whitney U-test P<0.05. Stage I vs. stage II P=0.0588, stage I vs. stage III P=0.0319, stage I vs. stage II–IV P=0.0349. For one NSCLC stage was not known.
Additionally, we analysed 38 sputum DNA samples (13 from patients with cancer and 25 from patients without lung cancer) for methylation of the 11 genes using quantitative PCR. Our data showed that, for 4 of the 11 genes, analysis of sputum DNA for methylation may have significant potential as biomarkers in cancer patients (Table 2) (Fig. 6). Analysis of sputum DNA for methylation did not show any significant potential as biomarkers in cancer patients for other genes (data not shown). This may be attributed to very small tumor cell population that might be present in sputum samples. Further, studies may be required with sputum as well as other clinical specimens containing exfoliated tumor cells (effusions and bronchoalveolar lavage) to evaluate the potential of these markers. Fig. 6A shows the quantitative methylation data of 3-OST-2, RASSF1A, P16 and APC for sputum samples from cancer patients (n=13) and normal controls (n=25). Cutoff ratios were established based on the highest QR values in control groups. Based on the cutoff values, methylation was detected in 4, 5, 3 and 3 cancer patients for 3-OST-2, RASSF1A, P16 and APC, respectively (n=13). Thus, based on the same cutoff values none of control samples showed methylation in any of the four genes. When the above quantitative data were analysed for methylation of either 3-OST-2 or RASSF1A 7 of the 13 cancer patient sputa (54%) and 0 of 23 normal controls (0%) (P<0.0002) showed methylation of either of the two genes. When the same quantitative data were analysed for methylation of any of the four genes, 8 of the 13 cancer sputum (62%) and 0 of 23 normal controls (0%) (P<0.0001) showed methylation of atleast one of the four genes. Fig. 6B shows the ROC curve and AUC (95% CI) which examine the accuracy of the markers to separate cancer sputum from the normal sputum. The data show that combination of m3-OST-2, mRASSF1A, mP16, mAPC have excellent discriminatory capacity at separating sputum of cancer patient from that of normal control.
Fig. 6.
(A) Methylation levels of m3-OST-2, mRASSF1A, mp16, mAPC in sputum from NSCLC patients and in sputa from patients without malignancy. Methylation levels were quantitated by semiquantitative real time PCR. Real time analysis was performed as described in Section 2. Quantitative ratio is defined as the ratio of the fluorescence emission intensity values for the PCR products of the biomarker gene to those of PCR products of MYOD1 multiplied by 100. The ratio is a measure for the relative level of methylation in an individual sample. Because values are expressed on a log scale, completely negative values are expressed as values of 0.01. The solid horizontal bar represents the cutoff ratios established based on the highest QR values in control groups. The large open circles with numbers indicates the number of samples below level of detection. The statistical comparison between the methylation frequencies (combined for the four genes) of sputum specimens from cancer patients and those from normal controls was done using Fisher’s exact test. (B). ROC curves to evaluate for the ability of the combination of m3-OST-2. mRASSF1A, mP16 and mAPC in separating sputa of patients with lung cancer from sputa of patients without lung cancer. ROC curves are plots of the true positive rate (vertical axis) against the false positive rate (horizontal axis) for the different possible cutoff points of a diagnostic test. The closer the curve follows the left-hand border and then the top border of the ROC space, the more accurate the test.
7. Discussion
Methylation specific PCR allows one to specifically assay for methylation at a common region within a gene promoter that correlates with loss of transcription [7]. Thus, aberrant methylation of tumor suppressor genes has formed basis in development of biomarkers for early detection of lung cancer [8]. One might conceptualize that markers that show high levels of methylation in tumor but very low methylation in adjacent non-malignant tissue and in blood mononuclear cells will prove to be clinically more useful. Since such assessment needs quantitative assay, in this study we used qPCR we quantitated promoter methylation for 11 genes in NSCLC. The presence of hypermethylation of one or more genes was found in 98% of the NSCLC. Initially, we tested the ability of the panel of markers to separate lung tumors from adjacent non-malignant lung, tumors from PBMCs and adenocarcinomas from squamous carcinomas. We found nine genes (3-OST-2, RASSF1A, P16, DcR1, DcR2, SOCS1, APC, DAPK, H-CAD) whose methylation levels are potentially useful as biomarkers in NSCLC. To our knowledge this constitutes the first report, which determined if levels of methylation of genes such as 3-OST-2, DcR1, DcR2, SOCS1 and SOCS3 could be used as potential biomarkers.
It appears from our data that levels of 3-OST-2, DcR1, DcR2 and SOCS1 methylation may have significant potential as biomarkers in NSCLC. Additionally, we compared the quantitative methylation data of previously untested genes with previously tested genes such as RASSF1A, APC and p16 [11]. RASSF1A was shown to be methylated at high levels in NSCLC (although at a frequency less than 50%) while present at low levels in adjacent non-malignant lung [14] and was considered a promising biomarker. P16 was shown to be methylated at intermediate levels but was also methylated at low levels in adjacent non-malignant lung [14]. APC was shown to be methylated at a high levels and at a high frequency in NSCLC while methylation was also found in high frequency (but at lower levels than tumor) in adjacent non-malignant tissue [11,12]. Our study indicated that methylation of 3-OST-2, DcR1, DcR2, SOCS1 and H-CAD in tumor tissues were one to two orders of magnitude higher than adjacent non-malignant tissue. Thus, the previously untested markers appear to be highly tumor specific. The assay system used here may also serve as a more stringent model for realistically comparing the potential of biomarkers to detect lung cancer in a clinical setting.
Most interestingly, 3-OST-2 showed the highest levels of promoter methylation in tumors (frequency of 70% based on the highest QR in adjacent normal lung tissue as a cutoff) combined with lowest levels of promoter methylation in adjacent normal tissue and also with lowest levels of promoter methylation in PBMCs from non-cancerous individuals. Furthermore, 3-OST-2 promoter methylation levels, and RASSF1A methylation correlated with tumor grade. Generation of polysaccharide chains and enzymatic modification of them to yield the specific saccharide sequences are critical for cell surface heparin sulfate proteoglycans in ligand binding and signal transduction. 3-OST-2 is the enzyme involved in the last of the three sequential steps each transferring a sulfate group to sugar residues [15,16] and of critical importance in functional integrity of the heparin sulfate proteoglycans. Quantitative analysis of 3-OST-2 methylation along with RASSF1A methylation appears to be a promising sensitive and specific biomarker assay for NSCLC and should be further validated in a clinical setting. Based on the preliminary data reported in this manuscript quantitative analysis of 3-OST-2, RASSF1A, P16 and APC methylation in sputum specimens may have excellent potential as a biomarker assay for patients with lung cancer. Methylated DNA has also been found in other diagnostic samples containing exfoliated tumor cells and may vary in the DNA fragment size, levels of PCR inhibitors, relative percentage and types of non-malignant cell types. Our preliminary findings of the application of a methylation panel for the diagnosis of malignancy in sputa need to be expanded. In addition, studies are in progress to evaluate the potential of the methylation gene panel in detection of lung cancers, in malignant effusions, serum and bronchoalveolar lavage samples of lung cancer patients.
Acknowledgments
The work was supported by grants U01CA084971 (Early Detection Research Network), P50CA70907 (The university of Texas SPORE in Lung Cancer) U01CA096109 and P01CA096964 from the National Cancer Institute, Bethesda, MD.
Abbreviations
- MSP
methylation specific PCR
- NSCLC
non-small cell lung cancer
References
- 1.Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancer statistics. CA Cancer J Clin. 2001;51:15–36. doi: 10.3322/canjclin.51.1.15. [DOI] [PubMed] [Google Scholar]
- 2.Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al. Cancer statistics. CA Cancer J Clin. 2005:10–30. doi: 10.3322/canjclin.55.1.10. [DOI] [PubMed] [Google Scholar]
- 3.Ellis JR, Gleeson FV. Lung cancer screening. Br J Radiol. 2001;74:478–485. doi: 10.1259/bjr.74.882.740478. [DOI] [PubMed] [Google Scholar]
- 4.Wingo PA, Ries LA, Giovino GA, Miller DS, Rosenberg HM, Shopland DR, et al. Annual report to the nation on the status of cancer, 1973–1996, with a special section on lung cancer and tobacco smoking. J Natl Cancer Inst. 1999;91:675–690. doi: 10.1093/jnci/91.8.675. [DOI] [PubMed] [Google Scholar]
- 5.Marcus PM. Lung cancer screening: an update. J Clin Oncol. 2001;19:83S–86S. [PubMed] [Google Scholar]
- 6.Palcic B, Garner DM, Beveridge J, Sun XR, Doudkine A, MacAulay C, et al. Increase of sensitivity of sputum cytology using high-resolution image cytometry: field study results. Cytometry. 2002;50:168–176. doi: 10.1002/cyto.10065. [DOI] [PubMed] [Google Scholar]
- 7.Herman JG, Baylin SB. Promoter-region hypermethylation and gene silencing in human cancer. Curr Top Microbiol Immunol. 2000;249:35–54. doi: 10.1007/978-3-642-59696-4_3. [DOI] [PubMed] [Google Scholar]
- 8.Sidransky D. Emerging molecular markers of cancer. Nat Rev Cancer. 2002;2:210–219. doi: 10.1038/nrc755. [DOI] [PubMed] [Google Scholar]
- 9.Toyooka KO, Toyooka S, Maitra A, Feng Q, Kiviat NC, Smith A, et al. Establishment and validation of real-time polymerase chain reaction method for CDH1 promoter methylation. Am J Pathol. 2002;161:629–634. doi: 10.1016/S0002-9440(10)64218-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shivapurkar N, Stastny V, Takahashi T, Suzuki M, Echebiri C, Reddy J, Gazdar AF. Novel real-time PCR assay using a universal molecular marker for diagnosis of hematologic cancers. Int J Cancer. 2005;116:656–660. doi: 10.1002/ijc.21070. [DOI] [PubMed] [Google Scholar]
- 11.Brabender J, Usadel H, Danenberg KD, Metzger R, Schneider PM, Lord RV, et al. Adenomatous polyposis coli gene promoter hypermethylation in non-small cell lung cancer is associated with survival. Oncogene. 2001;20:3528–3532. doi: 10.1038/sj.onc.1204455. [DOI] [PubMed] [Google Scholar]
- 12.Usadel H, Brabender J, Danenberg KD, Jeronimo C, Harden S, Engles J, et al. Quantitative adenomatous polyposis coli promoter methylation analysis in tumor tissue, serum, and plasma DNA of patients with lung cancer. Cancer Res. 2002;62:371–375. [PubMed] [Google Scholar]
- 13.Harden SV, Tokumaru Y, Westra WH, Goodman S, Ahrendt SA, Yang SC. Gene promoter hypermethylation in tumors and lymph nodes of stage I lung cancer patients. Clin Cancer Res. 2003;9:1370–1375. [PubMed] [Google Scholar]
- 14.Topaloglu O, Hoque MO, Tokumaru Y, Lee J, Ratovitski E, Sidransky D, Moon CS. Detection of promoter hypermethylation of multiple genes in the tumor and bronchoalveolar lavage of patients with lung cancer. Clin Cancer Res. 2004;10:2284–2288. doi: 10.1158/1078-0432.ccr-1111-3. [DOI] [PubMed] [Google Scholar]
- 15.Bernfield M, Gotte M, Park PW, Reizes O, Fitzgerald ML, Lincecum J, Zako M. Functions of cell surface heparan sulfate proteoglycans. Annu Rev Biochem. 1999;68:729–777. doi: 10.1146/annurev.biochem.68.1.729. [DOI] [PubMed] [Google Scholar]
- 16.Esko JD, Selleck SB. Order out of chaos: assembly of ligand binding sites in heparan sulfate. Annu Rev Biochem. 2002;71:435–471. doi: 10.1146/annurev.biochem.71.110601.135458. [DOI] [PubMed] [Google Scholar]
- 17.Vos MD, Martinez A, Elam C, Dallol A, Taylor BJ, Latif F, Clark GJ. A role for the RASSF1A tumor suppressor in the regulation of tubulin polymerization and genomic stability. Cancer Res. 2004;64:4244–4250. doi: 10.1158/0008-5472.CAN-04-0339. [DOI] [PubMed] [Google Scholar]
- 18.Ashkenazi A. Targeting death and decoy receptors of the tumour-necrosis factor superfamily. Nat Rev Cancer. 2002;2:420–430. doi: 10.1038/nrc821. [DOI] [PubMed] [Google Scholar]
- 19.Ashkenazi A, Dixit VM. Apoptosis control by death and decoy receptors. Curr Opin Cell Biol. 1999;11:255–260. doi: 10.1016/s0955-0674(99)80034-9. [DOI] [PubMed] [Google Scholar]
- 20.Liggett WH, Jr, Sidransky D. Role of the p16 tumor suppressor gene in cancer. J Clin Oncol. 1998;16:1197–1206. doi: 10.1200/JCO.1998.16.3.1197. [DOI] [PubMed] [Google Scholar]
- 21.Esteller M, Sparks A, Toyota M, Sanchez-Cespedes M, Capella G, Peinado MA, et al. Analysis of adenomatous polyposis coli promoter hypermethylation in human cancer. Cancer Res. 2000;60:4366–4371. [PubMed] [Google Scholar]
- 22.Kogel D, Prehn JH, Scheidtmann KH. The DAP kinase family of pro-apoptotic proteins: novel players in the apoptotic game. Bioessays. 2001;23:352–358. doi: 10.1002/bies.1050. [DOI] [PubMed] [Google Scholar]
- 23.Krebs DL, Hilton DJ. SOCS: physiological suppressors of cytokine signaling. J Cell Sci. 2000;16:2813–2819. doi: 10.1242/jcs.113.16.2813. [DOI] [PubMed] [Google Scholar]
- 24.Nicholson SE, De Souza D, Fabri LJ, Corbin J, Willson TA, Zhang JG, et al. Suppressor of cytokine signaling-3 preferentially binds to the SHP-2-binding site on the shared cytokine receptor subunit gp130. Proc Natl Acad Sci USA. 2000;97:6493–6498. doi: 10.1073/pnas.100135197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Takeichi M. Morphogenetic roles of classic cadherins. Curr Opin Cell Biol. 1995;7:619–627. doi: 10.1016/0955-0674(95)80102-2. [DOI] [PubMed] [Google Scholar]
- 26.Miyamoto K, Asada K, Fukutomi T, Okochi E, Yagi Y, Hasegawa T, et al. Methylation-associated silencing of heparan sulfate D-glucosaminyl 3-O-sulfotransferase-2 (3-OST-2) in human breast, colon, lung and pancreatic cancers. Oncogene. 2003;22:274–280. doi: 10.1038/sj.onc.1206146. [DOI] [PubMed] [Google Scholar]
- 27.Shivapurkar N, Toyooka S, Toyooka KO, Reddy J, Miyajima K, Suzuki M, et al. Aberrant methylation of trail decoy receptor genes is frequent in multiple tumor types. Int J Cancer. 2004;109:786–792. doi: 10.1002/ijc.20041. [DOI] [PubMed] [Google Scholar]
- 28.Toyooka S, Toyooka KO, Miyajima K, Reddy JL, Toyota M, Sathyanarayana UG, et al. Epigenetic down-regulation of death-associated protein kinase in lung cancers. Clin Cancer Res. 2003;9:3034–3041. [PubMed] [Google Scholar]
- 29.Dammann R, Schagdarsurengin U, Strunnikova M, Rastetter M, Seidel C, Liu L, et al. Epigenetic inactivation of the ras-association domain family 1 (RASSF1A) gene and its function in human carcinogenesis. Histol Histopathol. 2003;18:665–677. doi: 10.14670/HH-18.665. [DOI] [PubMed] [Google Scholar]
- 30.Fodde R. The multiple functions of tumour suppressors: it’s all in APC. Nat Cell Biol. 2003;5:190–192. doi: 10.1038/ncb0303-190. [DOI] [PubMed] [Google Scholar]
- 31.He B, You L, Uematsu K, Zang K, Xu Z, Lee AY, et al. SOCS-3 is frequently silenced by hypermethylation and suppresses cell growth in human lung cancer. Proc Natl Acad Sci USA. 2003;100:14133–14138. doi: 10.1073/pnas.2232790100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.He B, You L, Xu Z, Mazieres J, Lee AY, Jablons DM. Activity of the suppressor of cytokine signaling-3 promoter in human non-small-cell lung cancer. Clin Lung Cancer. 2004;5:366–370. doi: 10.3816/CLC.2004.n.015. [DOI] [PubMed] [Google Scholar]
- 33.Shivapurkar N, Takahashi T, Reddy J, Zheng Y, Stastny V, Collins R, et al. Presence of simian virus 40 DNA sequences in human lymphoid and hematopoietic malignancies and their relationship to aberrant promoter methylation of multiple genes. Cancer Res. 2004;64:3757–3760. doi: 10.1158/0008-5472.CAN-03-3307. [DOI] [PubMed] [Google Scholar]
- 34.Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci USA. 1996;93:9821–9826. doi: 10.1073/pnas.93.18.9821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Field JK, Youngson JH. The liverpool lung project: a molecular epidemiological study of early lung cancer detection. Eur Respir J. 2002;20:464–479. doi: 10.1183/09031936.02.00290202. [DOI] [PubMed] [Google Scholar]
- 36.Reddy J, Shivapurkar N, Takahashi T, Parikh G, Stastny V, Echebiri C, et al. Differential methylation of genes that regulate cytokine signaling in lymphoid and hematopoietic tumors. Oncogene. 2005;24:732–736. doi: 10.1038/sj.onc.1208032. [DOI] [PubMed] [Google Scholar]
- 37.Takahashi T, Shivapurkar N, Reddy J, Shigematsu H, Miyajima K, Suzuki M, et al. DNA methylation profiles of lymphoid and hematopoietic malignancies. Clin Cancer Res. 2004;10:2928–2935. doi: 10.1158/1078-0432.ccr-03-0716. [DOI] [PubMed] [Google Scholar]
- 38.Pepe MS. The Statistical Evaluation of Medical Tests for Classification and Prediction. Oxford University Press; New York, USA: 2003. [Google Scholar]