Purpose: This study was designed to determine whether DNA methylation biomarkers are associated with recurrence and survival in colon cancer patients.

Experimental Design: A retrospective analysis of 82 patients who received curative surgical resection for American Joint Committee on Cancer (AJCC) high-risk stage II or III colon cancer (1999–2007) was conducted. DNA methylation status was quantitatively evaluated by the pyrosequencing method. We preselected three tumor suppressor genes and one locus of interest; CHFR, ID4, RECK, and MINT1. Mean methylation levels of multiple CpG sites in the promoter regions were used for analysis; 15% or more was defined as methylation positive. The association of recurrence-free survival (RFS) and overall survival (OS) with methylation status was analyzed by the log-rank test, Kaplan–Meier method, and Cox proportional hazards model.

Results: Methylation levels of ID4, MINT1, and RECK did not correlate with RFS or OS. CHFR was methylation positive in 63% patients. When methylation status was dichotomized (negative or low: <30%, high: ≥30%), patients with CHFR methylation-high (44%) had worse RFS (P = 0.006) and reduced OS (P = 0.069). When stratified by stage, CHFR methylation-high was associated with reduced RFS (P = 0.004) and OS (P = 0.010) in stage III patients. CHFR methylation-high was commonly associated with N2 disease (P = 0.04) and proximal tumors (P = 0.002). Multivariate analysis indicated AJCC T4 disease and CHFR methylation-high (P = 0.001 and P = 0.015, respectively) were independent predictors for recurrence.

Conclusions: The extent of CHFR promoter methylation correlates with RFS, indicating it is a promising epigenetic marker for recurrence. Clin Cancer Res; 17(13); 4531–40. ©2011 AACR.

Translational Relevance

This study demonstrates that CHFR hypermethylation is significantly associated with disease recurrence in American Joint Committee on Cancer (AJCC version 6) stage III colon cancer after curative surgical resection. Identification of patients with high risk of disease recurrence may have implications for clinical management of colon cancer following curative surgical resection. Our analysis indicates that CHFR methylation status is a potential prognostic biomarker for colon cancer recurrence.

Colorectal cancer remains the third most common malignancy and the second leading cause of cancer mortality for both men and women combined in the United States (1). Despite curative surgical resection, it is reported that 20% to 50% of all patients with American Joint Committee on Cancer (AJCC) stage II or III colon cancer will develop disease recurrence; 80% of all recurrences occurring within the first 3 years following surgical resection (2). Historical controls indicate that the absolute benefit for adjuvant 5-flurouracil (5-FU)–based chemotherapy in all stage II patients is minimal (3).

Accepted clinicopathologic risk factors for stage II colon cancer patients at high-risk of recurrence include T4 disease (tumor perforation or invasion of an adjacent organ), identification of less than 12 lymph nodes in the surgically resected specimen, microvascular invasion, and poorly differentiated histology (4–6). In this select group of stage II patients, discussion of the possible benefits of adjuvant chemotherapy is encouraged but largely remains up to the discretion of the treating physician and patient (7).

In stage III patients, the 5-year overall survival (OS) rate varies widely from 28% to 60% (2), indicating that a significant amount of tumor heterogeneity exists and that the current clinical staging system does not necessarily reflect the tumor's biologic behavior. Although the AJCC staging system (version 7) has been revised, predictive and prognostic markers were not included for locally advanced disease and remain an area under continued development.

Currently, all stage III disease patients and select stage II disease patients are considered for either 6 months of adjuvant 5-FU and leucovorin (LV) or a combination of oxaliplatin, infusional 5-FU, and LV (FOLFOX) yielding an absolute OS rate benefit of less than 3% and 10%, respectively (3, 8). Unfortunately, few chemotherapy advances have developed for the treatment of locally advanced colon cancer. Furthermore, despite the use of novel molecular-targeted therapeutics (i.e., bevacizumab, cetuximab, and panitumumab) in patients with metastatic disease, their use in stage II and III patients is not supported at this time. Recent results from 3 phase III trials were negative. The NSABP C-08 trial of FOLFOX ± bevacizumab (stage II and III) and the N0147 trial of FOLFOX ± cetuximab (stage III only) failed to determine any additional benefit beyond standard FOLFOX for an improvement of disease-free survival (DFS; refs 9, 10). Preliminary results from the international phase III AVANT trial (11) of adjuvant FOLFOX or FOLFOX + bevacizumab or XELOX + bevacizumab also failed to meet the primary endpoint of DFS with improved outcomes for the control arm. The results from these phase III trials clearly indicate that the underlying biology in locally advanced colon cancer is dissimilar from that of metastatic colon cancer.

Thus, for the time being a therapeutic plateau in locally advanced colon cancer has been reached. Identification of a contemporary biomarker for recurrence risk, OS, and potential therapeutic benefit would result in optimal clinical management of locally advanced colon cancer patients.

Colon cancer is clearly a molecularly diverse cancer given the variability in OS and recurrence (12). Many studies have explored the prognostic value of various molecular markers in both early and advanced colorectal cancer (13–15). Although a series of molecular markers for colorectal cancer is available for clinical use, many are still being validated in ongoing clinical trials and currently none are conventionally accepted for locally advanced colon cancer. Recent data has indicated that the presence of the K-ras mutation is likely not prognostic, but is predictive of potential benefit of molecular-targeted therapy. Currently, K-ras is the only validated predictive marker for colon cancer, but its role is limited to the setting of metastatic disease (16).

Epigenetic aberrations have been reported to be important mechanisms in human carcinogenesis (17–19). Inactivation of tumor suppressor genes by promoter hypermethylation has been implicated in colorectal carcinogenesis (20). Prior studies have indicated the predictive and prognostic utility of gene methylation in primary colorectal tumors (21–23). However, it is still unclear whether the methylation profile of tumor suppressor genes can serve as prognostic biomarkers for tumor recurrence and survival in patients with colon cancer following curative surgical resection.

Promoter hypermethylation of tumor suppressor genes including CHFR (checkpoint with forkhead-associated and RING finger domains), ID4 (inhibitor of DNA binding), and RECK (reversion-inducing cysteine-rich protein with Kazal motifs) has been reported in colorectal cancer and is associated with reduced mRNA and/or protein expression (24–26). In this study, we evaluated select tumor suppressor genes and the MINT1 (methylated-in-tumor) locus (27) to analyze the association of tumor recurrence and patient OS with methylation status.

Patient enrollment and tissue samples

In a single-institution retrospective analysis, we evaluated patient records from 82 patients with histologically confirmed (AJCC version 6) high-risk stage II (T3–4N0) or III (TxN+) colon adenocarcinomas. High-risk stage II was defined as T4, poorly differentiated histology, less than 12 lymph nodes dissected, and evidence of lymphovascular invasion. Patients with rectal cancer and/or known familial colorectal cancer syndromes were excluded. All patients had undergone curative surgical resection without prior chemotherapy at The University of Texas MD Anderson Cancer Center (Houston, Texas) from December 1999 to August 2007. All tissue specimens were randomly provided by the MD Anderson tissue bank based on eligibility criteria and available primary colon tumor tissue and adjacent normal colonic mucosa. Charts were reviewed using all available electronic medical records and data on tumor recurrence, OS, and clinicopathologic characteristics. All patients had previously provided preoperative informed consent to allow their surgical tissue specimens to be used for investigational purposes. This study was approved by the institutional review board of MD Anderson Cancer Center.

DNA methylation analysis

DNA methylation levels were measured using the highly quantitative pyrosequencing technology (Pyrosequencing AB, Sweden). Surgical specimens were frozen in liquid nitrogen and stored at −80°C, and subsequently used for the extraction of tissue DNA. The same tumor samples were evaluated by hematoxilin-eosin staining to confirm that the dissected area contained more than 50% of tumor cells in all cases. Lymphocyte DNA was extracted from blood samples donated by healthy donors and served as the negative control. DNA purification kits (Promega) were used for DNA extraction. To create a positive control, lymphocyte DNA was treated with CpG Methyltransferase (SssI; New England Biolabs). Each genomic DNA was treated with bisulfite using an Epitect Bisulfite Kit (QIAGEN). Bisulfite-treated DNA (50 ng) was amplified with gene-specific primers in a 2-step PCR. The second step of PCR was used to label single DNA strands with biotin using a universal primer tag (28). PCR was performed in a total volume of 25 μL of 2 mmol/L magnesium chloride (MgCl2), 0.2 mmol/L of deoxyribonucleotide triphosphates (dNTPs), 200 nmol/L PCR primers, and 1.25 U AmpliTaq gold DNA Polymerase and GeneAmp PCR Gold Buffer (Applied Biosystems). The PCR primer sequences and annealing temperatures are listed in Table 1. All primers were designed to assay the methylation status of 3 to 4 CpGs in promoter regions within 0.6 kb from the transcription start sites. The following conditions were used in both PCRs: initial denaturation at 95°C for 10 minutes, followed by 48 cycles for denaturation at 94°C for 15 seconds, annealing at the appropriate temperature for 30 seconds, and extension at 72°C for 30 seconds. The touchdown method was used only for MINT1 amplification. To overcome PCR bias, the optimal annealing temperature for PCR in each primer set was determined by testing some mixed rates (0%, 25%, 50%, 75%, and 100%) of positive-control DNA in negative-control DNA (29). After PCR, the biotinylated strand was captured on streptavidin-coated beads (Amersham Bioscience), incubated with the sequencing primers (Table 1), and run on the PSQ HS 96 pyrosequencing system and Pyro Gold CDT Reagents (Biotage, Sweden), as described previously (28). Pyrosequencing can measure the methylation levels of several CpG sites in a given promoter. The methylation levels were averaged to represent the degree of methylation in each sample for the tested genes. We defined the mean methylation level of 15% or more as methylation positive because values of 15% or less are difficult to distinguish from background noise.

Table 1.

Summary of bisulfite PCR and pyrosequencing primers

Gene or lociBisulfite PCR primersSequencing primersAnnealing temperature (cycles)
CHFRa Forward  55°C (48) 
 AGGGTATTTTTTGAAAGTGAAAGG TTTTTTGAAAGTGAAAGG  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTATCCTACTTCACCCTCTTAAAACC   
ID4 Forward  58°C (48) 
 TTTTTGGGTATATATTAGTTTGGT TGGAGTGTTTTTTTTATTG  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTAAATCTAAAATAACCAACCAATCA   
RECK Forward  55°C (48) 
 GTTGGGTTATAATAAAGAGTTTTG TTATAATAAAGAGTTTTGGT  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTACCCCCTACCTCACAATATTTC   
MINT1b Forward  56°C (3) 
 GGTTTTTTGTTAGAGTTTGTATTT TTTAGTAAAAATTTTTTGGG 54°C (4) 
 Reverse-Universal  52°C (5) 
 GGGACACCGCTGATCGTTTAATTAATCCCTCTCCCCTCTAAACTT  50°C (36) 
Gene or lociBisulfite PCR primersSequencing primersAnnealing temperature (cycles)
CHFRa Forward  55°C (48) 
 AGGGTATTTTTTGAAAGTGAAAGG TTTTTTGAAAGTGAAAGG  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTATCCTACTTCACCCTCTTAAAACC   
ID4 Forward  58°C (48) 
 TTTTTGGGTATATATTAGTTTGGT TGGAGTGTTTTTTTTATTG  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTAAATCTAAAATAACCAACCAATCA   
RECK Forward  55°C (48) 
 GTTGGGTTATAATAAAGAGTTTTG TTATAATAAAGAGTTTTGGT  
 Reverse-Universal   
 GGGACACCGCTGATCGTTTACCCCCTACCTCACAATATTTC   
MINT1b Forward  56°C (3) 
 GGTTTTTTGTTAGAGTTTGTATTT TTTAGTAAAAATTTTTTGGG 54°C (4) 
 Reverse-Universal  52°C (5) 
 GGGACACCGCTGATCGTTTAATTAATCCCTCTCCCCTCTAAACTT  50°C (36) 

NOTE: For those with the Universal-tailed primer, the following additional primer is added: Universal (5′-Biotin) GGGACACCGCTGATCGTTTA.

aComplement sequence was used for designing PCR and sequencing primers.

bTouchdown PCR was used for MINT1 amplification.

Immunohistochemistry of MLH1 and CHFR

Formalin-fixed, paraffin-embedded tissues were cut into 3 μm sections consecutively. Immunohistochemistry (IHC) of MLH1 was performed on an automated immunostainer (Lab Vision Autostainer 360–2D; LabVision) according to the manufacturer's protocols. Slides were incubated with the primary antibody for MLH1 (G168–15, BD Bioscience) with a dilution of 1:50 for 1 hour, and followed by Peroxidase staining using Lab Vision UltraVision LP Large Volume Detection System HRP Polymer (Ready-to-Use). Each case was evaluated and categorized into MLH negative or positive. CHFR protein expression was measured using the primary anti-CHFR polyclonal antibody (12169-1-AP, Proteintech Group) at a dilution of 1:150. Sections were incubated with the primary antibody at room temperature for 3 hours, followed by incubation with the secondary antibody at room temperature for 1 hour. The antibody complex was detected using an avidin-biotin-peroxidase complex solution and visualized using 3,3′-diaminobenzidine (Zymed Laboratories, Inc.). A negative control was included in each experiment by omitting the primary antibody. The expression level of CHFR was evaluated and categorized into 4 groups; negative, low, medium, and high.

The person who conducted the IHC assay was blinded to the methylation status of the samples.

Statistical methods

Seventy-nine was the minimum sample size required for the Cox proportional hazards model, based on a power of 90%, α = 0.05 (2 sided), standard deviation of 0.50, and event (recurrence) rate of 30% using Stata software, version 10.1 (StataCorp; ref 30).

The association of methylation status with recurrence-free survival (RFS) or OS was analyzed using the log-rank test, Kaplan–Meier plot, and Cox proportional hazards regression models. The multivariate analysis included clinicopathologic factors that had a P < 0.10. Chi-square test or Fisher's exact test was used to analyze the relationship between methylation status, protein expression level, and clinicopathologic variables. To identify cutoff values that had the highest sensitivity and specificity to predict recurrence, receiver operating characteristic (ROC) analysis was used. All statistical testing was conducted with SPSS software, version 17.0 (SPSS Inc.). P < 0.05 was considered statistically significant.

Patient demographics and tumor clinicopathologic features

Table 2 demonstrates the association of clinicopathologic characteristics with tumor recurrence or OS. The median age of the 82 patients was 63.1 years (range, 31–93). Non-Hispanic whites accounted for 78% of the patients. The mean number of lymph nodes dissected ± standard error, was 25.6 ± 2.6 (range: 6–190); 10 cases had less than 12 dissected lymph nodes. Adjuvant 5-FU–based chemotherapy was provided in 61 (74%) patients; FOLFOX was concurrently administered to 18 patients with stage III disease. After a mean follow-up of 60 months, disease recurrence developed in 24 (29%) patients and 12 (15%) had died. Common sites of disease recurrence included liver (67%), lung (25%), peritoneum (4%), and the primary tumor (4%). T4 stage was associated with reduced RFS and OS (P = 0.013 and P = 0.038, respectively, log-rank test). Patients with proximal-sided tumors were noted to have a reduction in OS (P = 0.048).

Table 2.

Association of clinicopathologic characteristics with RFS and OS

RFSOS
VariableNo. of casesNo. of eventsPaNo. of eventsPa
Age, y   0.351  0.126 
 ≤50 15   
 51–60 19   
 61–70 19   
 >70 29 12  12  
Sex   0.824  0.708 
 Male 41 12   
 Female 41 12   
Tumor site   0.109  0.048 
 Proximal 42 15   
 Distal 40   
Tumor size   0.528  0.804 
 <5.0 38 10   
 ≥5.0 44 14   
AJCC tumor (T)   0.013  0.038 
 T3 T3 17   
 T4 11   
AJCC lymph node (N  0.357  0.379 
 N0 28   
 N1 35 10   
 N2 19   
AJCC stage   0.340  0.733 
 II 28   
 III 54 17   
Histologic grade   0.889  0.676 
 Well/moderate 65 19   
 Poor/others 17   
Adjuvant chemotherapy   0.869  0.585 
 Yes 61 18  61  
 No 21   
Oxaliplatin use   0.204  0.995 
 Yes 18   
 No 64 18  11  
RFSOS
VariableNo. of casesNo. of eventsPaNo. of eventsPa
Age, y   0.351  0.126 
 ≤50 15   
 51–60 19   
 61–70 19   
 >70 29 12  12  
Sex   0.824  0.708 
 Male 41 12   
 Female 41 12   
Tumor site   0.109  0.048 
 Proximal 42 15   
 Distal 40   
Tumor size   0.528  0.804 
 <5.0 38 10   
 ≥5.0 44 14   
AJCC tumor (T)   0.013  0.038 
 T3 T3 17   
 T4 11   
AJCC lymph node (N  0.357  0.379 
 N0 28   
 N1 35 10   
 N2 19   
AJCC stage   0.340  0.733 
 II 28   
 III 54 17   
Histologic grade   0.889  0.676 
 Well/moderate 65 19   
 Poor/others 17   
Adjuvant chemotherapy   0.869  0.585 
 Yes 61 18  61  
 No 21   
Oxaliplatin use   0.204  0.995 
 Yes 18   
 No 64 18  11  

Abbreviations: RFS, recurrence-free survival; OS, overall survival.

aLog-rank P value.

Association of DNA methylation with OS and RFS

In the pyrosequencing assay, lymphocyte DNA and adjacent normal tissue DNA were shown to be methylation-negative (<15%). Methylation-positive rates (≥15%) for CHFR, ID4, MINT-1, and RECK in tumor DNA were 63%, 46%, 28%, and 9%, respectively. Methylation status was not significantly associated with RFS or OS in any of the tested genes or in the MINT1 locus (Table 3). When ROC analysis was used to separate all cases into 2 prognostic groups, a cutoff value of 30% for CHFR methylation level recurrence was noted with 62.5% sensitivity and 65.5% specificity. Therefore, we dichotomized the methylation status into methylation-negative or -low (<30%) and methylation-high (≥30%). Patients with CHFR methylation-high status had a significantly reduced RFS and a trend toward reduced OS (P = 0.006 and P = 0.069; Table 3, log-rank test). Kaplan–Meier plots indicated that the probability of RFS at 5 years was 0.55 ± 0.09 in the CHFR methylation-high group and 0.82 ± 0.06 in the CHFR methylation-negative or -low group (P = 0.006, Fig. 1A). Subgroup analysis showed that CHFR methylation-high status was associated with reduced RFS (P = 0.004, Fig. 1B) and reduced OS (P = 0.010) in stage III disease patients, whereas no association with RFS was observed in high-risk stage II disease patients (P = 0.661, Fig. 1C).

Figure 1.

Kaplan–Meier plots for RFS in relation to CHFR promoter methylation in (A) all patients, (B) AJCC stage III, and (C) AJCC stage II patients. High methylation level (≥30%) of CHFR promoter region was associated with RFS in all patients and stage III patients (log-rank P = 0.006 and P = 0.004, respectively), but not in stage II patients (log-rank P = 0.661).

Figure 1.

Kaplan–Meier plots for RFS in relation to CHFR promoter methylation in (A) all patients, (B) AJCC stage III, and (C) AJCC stage II patients. High methylation level (≥30%) of CHFR promoter region was associated with RFS in all patients and stage III patients (log-rank P = 0.006 and P = 0.004, respectively), but not in stage II patients (log-rank P = 0.661).

Close modal
Table 3.

Association of DNA methylation with RFS and OS

RFSOS
GeneNo. of casesNo. of eventsPaNo. of eventsPa
CHFR methylation status   0.240  0.558 
 Negative (<15%) 30   
 Positive (≥15%) 52 17   
   0.006  0.069 
 Negative or low (<30%) 47   
 High (≥30%) 35 15   
ID4 methylation status   0.850  0.118 
 Negative (<15%) 44 13   
 Positive (≥15%) 38 11   
MINT1 methylation status   0.190  0.535 
 Negative (<15%) 59 20  10  
 Positive (≥15%) 23   
RECK methylation status   0.087  0.242 
 Negative (<15%) 75 24  12  
 Positive (≥15%)   
RFSOS
GeneNo. of casesNo. of eventsPaNo. of eventsPa
CHFR methylation status   0.240  0.558 
 Negative (<15%) 30   
 Positive (≥15%) 52 17   
   0.006  0.069 
 Negative or low (<30%) 47   
 High (≥30%) 35 15   
ID4 methylation status   0.850  0.118 
 Negative (<15%) 44 13   
 Positive (≥15%) 38 11   
MINT1 methylation status   0.190  0.535 
 Negative (<15%) 59 20  10  
 Positive (≥15%) 23   
RECK methylation status   0.087  0.242 
 Negative (<15%) 75 24  12  
 Positive (≥15%)   

aLog-rank P value.

Association of methylation status with clinicopathologic characteristics

We analyzed the relationship between methylation status and clinicopathologic factors. CHFR methylation-high status was significantly associated with age 70 years or more, proximal tumor, and N2 disease (P = 0.008, P = 0.002, and P = 0.04, respectively; Table 4). Methylation-positive status in both ID4 and MINT-1 was associated with poorly differentiated tumors (P = 0.001 and P < 0.001).

Table 4.

Association between clinicopathologic characteristics and CHFR methylation status

VariableNo. of patients (%)
Methylation-negative or -low (n = 47)Methylation-high (n = 35)P2]
Age, y   0.008 
 <50 7 (47) 8 (53)  
 51–60 13 (68) 6 (32)  
 61–70 16 (84) 3 (16)  
 >70 11 (38) 18 (62)  
Sex   0.503 
 Male 22 (54) 19 (46)  
 Female 25 (61) 16 (39)  
Tumor site   0.002 
 Proximal 17 (40) 25 (60)  
 Distal 30 (75) 10 (25)  
Tumor size   0.585 
 <5.0 23 (61) 15 (39)  
 >5.0 24 (55) 20 (45)  
AJCC primary tumor (T)   0.649a 
 T3 40 (56) 31 (44)  
 T4 7 (64) 4 (36)  
AJCC lymph node (N  0.110 
 N0 17 (61) 11 (39)  
 N1 23 (66) 12 (34)  
 N2 7 (37) 12 (63)  
   0.040 
 N0/N1 40 (63) 23 (37)  
 N2 7 (37) 12 (63)  
AJCC stage   0.654 
 II 17 (61) 11 (39)  
 III 30 (56) 24 (44)  
Histologic grade   0.214 
 Well/moderate 35 (54) 30 (46)  
 Poor/others 12 (71) 5 (29)  
MLH1 expressionb   0.928 
 Positive 26 (52) 24 (48)  
 Negative 8 (53) 7 (38)  
VariableNo. of patients (%)
Methylation-negative or -low (n = 47)Methylation-high (n = 35)P2]
Age, y   0.008 
 <50 7 (47) 8 (53)  
 51–60 13 (68) 6 (32)  
 61–70 16 (84) 3 (16)  
 >70 11 (38) 18 (62)  
Sex   0.503 
 Male 22 (54) 19 (46)  
 Female 25 (61) 16 (39)  
Tumor site   0.002 
 Proximal 17 (40) 25 (60)  
 Distal 30 (75) 10 (25)  
Tumor size   0.585 
 <5.0 23 (61) 15 (39)  
 >5.0 24 (55) 20 (45)  
AJCC primary tumor (T)   0.649a 
 T3 40 (56) 31 (44)  
 T4 7 (64) 4 (36)  
AJCC lymph node (N  0.110 
 N0 17 (61) 11 (39)  
 N1 23 (66) 12 (34)  
 N2 7 (37) 12 (63)  
   0.040 
 N0/N1 40 (63) 23 (37)  
 N2 7 (37) 12 (63)  
AJCC stage   0.654 
 II 17 (61) 11 (39)  
 III 30 (56) 24 (44)  
Histologic grade   0.214 
 Well/moderate 35 (54) 30 (46)  
 Poor/others 12 (71) 5 (29)  
MLH1 expressionb   0.928 
 Positive 26 (52) 24 (48)  
 Negative 8 (53) 7 (38)  

aFisher's exact test.

bSixty-five cases of tissues were available for IHC of MLH1.

MLH1 protein expression

IHC analysis demonstrated that 15/65 (23%) of tissue samples examined were negative for MLH1 staining (Fig. 2). The MLH1 staining status was not associated with CHFR hypermethylation (P = 0.928).

Figure 2.

MLH1 protein expression by IHC (×200). (A) Negative expression of MLH1. (B) Positive expression of MLH1.

Figure 2.

MLH1 protein expression by IHC (×200). (A) Negative expression of MLH1. (B) Positive expression of MLH1.

Close modal

Correlation of CHFR methylation status with protein expression

CHFR protein expression was detected in 30/35 (86%) of the tissue samples analyzed (Fig. 3). The expression level was low in 10 (29%), intermediate in 11 (31%), and high in 9 (26%) of the 35 samples. CHFR expression level was inversely associated with CHFR methylation score (P = 0.008); 12 of 19 (63%) cases with methylation-high (≥30%) had negative or low protein expression level, whereas 13 of 16 (81.3%) cases with CHFR methylation-negative or -low had medium or high protein expression levels.

Figure 3.

CHFR protein expression by IHC (×200). CHFR expression was (A) negative, (B) low, (C) medium, and (D) high. CHFR methylation rate was (A) 71.0%, (B) 52.9%, (C) 18.2%, and (D) 9.1%, respectively.

Figure 3.

CHFR protein expression by IHC (×200). CHFR expression was (A) negative, (B) low, (C) medium, and (D) high. CHFR methylation rate was (A) 71.0%, (B) 52.9%, (C) 18.2%, and (D) 9.1%, respectively.

Close modal

Multivariate analysis

We used the Cox proportional hazards model for multivariate analysis on RFS (Table 5). In all patients (n = 82), CHFR methylation-high status was an independent prognostic factor for recurrence, with HR of 2.99 (95% CI = 1.23–7.23, P = 0.015) after adjusting for T stage, nodal status, use of adjuvant chemotherapy, and adjuvant oxaliplatin use. T4 was significantly associated with RFS (HR = 6.53, 95% CI = 2.22–19.3, P = 0.001). Subgroup analysis indicated that CHFR methylation-high status was a significant predictor for recurrence in stage III patients (HR = 4.13, 95% CI = 1.34–12.8, P = 0.014) but not in stage II patients (HR = 1.45, 95% CI = 0.30–6.88, P = 0.643).

Table 5.

Multivariate Cox regression analysis of RFS

CovariateAll patients (N = 82)Stage III (n = 54)Stage II (n = 28)
HR (95% CI)PHR (95% CI)PHR (95% CI)P
AJCC primary tumor 
 T3 1.0  1.0  1.0  
 T4 6.53 (2.22–19.3) 0.001 6.14 (1.89–19.9) 0.003 2.04 (0.18–22.8) 0.561 
AJCC lymph node 
 N0 1.0      
 N1 3.07 (0.87–10.8) 0.082 1.0  –  
 N2 4.55 (0.89–23.3) 0.069 0.76 (0.22–2.62) 0.664   
Adjuvant chemotherapy 
 Yes 1.0  1.0  1.0  
 No 3.62 (0.87–15.1) 0.078 5.59 (1.09–28.6) 0.039 1.17 (0.12–11.3) 0.893 
Oxaliplatin use 
 Yes 1.0  1.0  a  
 No 0.68 (0.22–2.15) 0.513 0.66 (0.20–2.22) 0.504   
CHFR methylation status 
 <30% 1.0  1.0  1.0  
 >30% 2.99 (1.23–7.23) 0.015 4.13 (1.34–12.8) 0.014 1.45 (0.30–6.88) 0.643 
CovariateAll patients (N = 82)Stage III (n = 54)Stage II (n = 28)
HR (95% CI)PHR (95% CI)PHR (95% CI)P
AJCC primary tumor 
 T3 1.0  1.0  1.0  
 T4 6.53 (2.22–19.3) 0.001 6.14 (1.89–19.9) 0.003 2.04 (0.18–22.8) 0.561 
AJCC lymph node 
 N0 1.0      
 N1 3.07 (0.87–10.8) 0.082 1.0  –  
 N2 4.55 (0.89–23.3) 0.069 0.76 (0.22–2.62) 0.664   
Adjuvant chemotherapy 
 Yes 1.0  1.0  1.0  
 No 3.62 (0.87–15.1) 0.078 5.59 (1.09–28.6) 0.039 1.17 (0.12–11.3) 0.893 
Oxaliplatin use 
 Yes 1.0  1.0  a  
 No 0.68 (0.22–2.15) 0.513 0.66 (0.20–2.22) 0.504   
CHFR methylation status 
 <30% 1.0  1.0  1.0  
 >30% 2.99 (1.23–7.23) 0.015 4.13 (1.34–12.8) 0.014 1.45 (0.30–6.88) 0.643 

Abbreviations: HR, hazard ratio; CI, confidence interval.

aOxaliplatin was not used for stage II patients.

Our current data indicates that the methylation level of CHFR measured by pyrosequencing assay was an independent predictor for recurrence for locally advanced colon cancer patients. By subgroup analysis, CHFR methylation-high was associated with reduced RFS and OS, notably in stage III patients. Though it is well established that stage III colon cancer patients have a relatively high disease recurrence rate (30%–40%) regardless of the completion of adjuvant chemotherapy (31), our results indicate that CHFR methylation status may be a useful prognostic biomarker in stratifying patients for more intensive treatment or surveillance.

CHFR was initially identified as an early mitotic checkpoint protein that delays entry into metaphase of the cell cycle in response to mitotic stress caused by microtubule inhibitors (32). Thereafter, some studies have established that CHFR mRNA expression was lost or decreased in primary tumors and cell lines of colorectal cancer (24), gastric cancer (33), lung cancer (34), and oral squamous cell carcinoma (35); this was predominantly mediated through DNA methylation of the CHFR promoter (24, 33–35). In addition, Yu and colleagues reported that CHFR knockout mice are prone to carcinogenesis (36). These findings suggest an important role of CHFR as an inhibitor of tumorigenesis through regulation of the cell cycle and maintenance of genomic stability.

In our analysis, CHFR promoter methylation was associated with advanced lymph node metastasis (N2). Privette and colleagues demonstrated that low expression of CHFR in human mammary epithelial cells resulted in malignant progression, increased motility, and enhanced invasiveness (37). In prior colorectal cancer studies, low or lost CHFR expression has been associated with a high mitotic index (24), and the checkpoint function of the forkhead-associated domain is a core component of anti-proliferative properties against gastrointestinal carcinogenesis (38). These data suggest that an epigenetic inactivation of CHFR correlates with increased malignant potential of colon cancer.

To date, we are unaware of any published reports demonstrating an association of CHFR methylation level with colon cancer recurrence as observed in our analysis. Two important studies on CHFR function have been recently published. Kashima and colleagues demonstrated that CHFR significantly downregulated IL-8 through the inhibition of nuclear factor κB (NFκB) and that the migration of human endothelial cells was suppressed in the culture medium conditioned from CHFR-expressing cancer cells (39). In a separate study, Oh and colleagues demonstrated that CHFR downregulated histone deacetylase 1 (HDAC1), leading to upregulation of the Cdk inhibitor, p21, and tumor suppressor genes, KAI1 and E-cadherin (40). Taken together, these data support a potential role of CHFR for regulating cancer metastasis.

Microsatellite instability (MSI)-high or dMMR (deficient mismatch repair) phenotype of colon cancer occurs infrequently (15%–20%) and is reported to be associated with CHFR hypermethylation (41) or its decreased protein expression (42). It has been also noted that MSI-high colon cancer patients have an overall improved prognosis regardless of stage and do not benefit from single-agent 5-FU (43–45); in fact, if stage II patients are treated with single agent 5-FU they may fare worse for OS (23, 43–45). Although we observed no association between CHFR methylation and MLH1 expression, these previously reported findings appear to be inconsistent with our observation that CHFR hypermethylation correlates with low RFS. Controversy remains when discussing the role adjuvant oxaliplatin-based chemotherapy (FOLFOX) in patients with MSI-high stage III colon cancer (13, 45). Additional assessment of the association of CHFR hypermethylation, adjuvant chemotherapy, and MSI status relative to prognosis (RFS and OS) in a larger patient population is warranted to understand the potential effects and interactions of these factors.

The CpG island methylator phenotype (CIMP) has been named as a distinct molecular colon cancer subgroup that is fundamentally different from other colon cancers (46), and this subtype has been reported to be a prognostic marker of improved OS in colon cancer, according to several studies (23, 47). Although some clinical and molecular characteristics such as high frequency of proximal tumors and MSI have been associated with both CIMP-positive status and CHFR hypermethylation (23, 41, 48), CHFR methylation has not been clustered with CIMP-positive subtype in those studies (23, 41, 47), suggesting the possibility that methylation of CHFR is independent of CIMP.

In our analysis, a cutoff value of 15% was used to define a positive methylation status. An absolute cutoff value of methylation-positive status in a pyrosequencing assay is nonexistent. In addition, methylation status was categorized into 2 groups: methylation-negative or -low (<30%), and methylation-high (≥30%) based on ROC analysis for prediction of recurrence. Considering that CHFR methylation-high (≥30%) was significantly associated with negative or low expression level of CHFR in this study, the cutoff value of 30% is functionally justified for CHFR methylation. Pyrosequencing is a highly reproducible and accurate quantitative assay. Indeed, a method of grouping patients by the extent of methylation has previously been used for improved prognostic stratification (49, 50).

No association of methylation status in ID4, MINT1, or RECK with RFS and OS was observed in our study. ID4 has been thought to act as a dominant-negative inhibitor of gene transcription and to regulate important pathways in cell proliferation and differentiation (22). Hypermethylation of ID4 promoters has been reported to correlate with poorly differentiated tumors and poor OS in a colon cancer study (22). In our study, methylation-positive status of ID4 was associated with poorly differentiated tumors but not with OS and RFS. MINT1 has been defined as one of the CIMP markers (21). CIMP status has been associated with survival of colon cancer patients (21, 23), but it is unclear whether the single effect of MINT1 has significant impact on outcome (20). RECK is an important metastasis suppressor gene, and its epigenetic inactivation has been reported to enhance invasion of human colon cancer cells (26). To date, no association of RECK hypermethylation with recurrence and survival has been demonstrated.

We realize our study's limitations: it is a retrospective, single-institution analysis and is limited by sample size. We also recognize the tissue specimen collection spanned over a 9-year duration which may seem prolonged, but was limited by availability of tissue based on the eligibility criteria. In this analysis, only 1/3 of patients were stage II, resulting in a limited sample size for this subgroup analysis. The greater number of stage III patients reflects the more commonly referred patient population to MD Anderson, of more advanced stage patients. However, these findings provide early insight regarding CHFR hypermethylation and development of disease recurrence in AJCC stage III patients. A larger population analysis will need to be conducted for further validation. If confirmed, methylation status of CHFR may be not only an informative prognostic biomarker to identify patients with increased risk of recurrence, but also a biomarker for therapeutic development. Considering few therapeutic modifications have developed in the management of these patients, identification of additional biomarkers may be of added significance. Hypermethylation may serve as a selected treatment approach, as preclinical and clinical studies have indicated the potential role of hypomethylating agents, such as azacitidine (AZA), in reversing the effects of hypermethylation in solid tumors (51). We are currently conducting a single institution phase I/II study evaluating the role of AZA in combination with oxaliplatin and capecitabine in CpG hypermethylated (MSI-high) metastatic colorectal cancer patients that are refractory to standard therapy.

In summary, we observed a significant association between the extent of CHFR methylation level in stage III disease and the development of disease recurrence. We realize that disease recurrence may be attributable to a multitude of factors, and it is likely there are other undetermined or yet-to-be validated prognostic and predictive markers. Nevertheless, we believe this epigenetic biomarker may have an integral role in disease development and recurrence, and its therapeutic development and significance may be of greater magnitude when combined with other potential biomarkers.

No potential conflicts of interest were disclosed.

We would like to thank Dr. Stanley R. Hamilton for his laboratory support for the pyrosequencing assay.

Philanthropic grant from the Urbieta Family Colon Cancer Research Fund and a MD Anderson Cancer Center Institutional Research Grant (to C. Eng).

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.
Jemal
A
,
Siegel
R
,
Xu
J
,
Ward
E
. 
Cancer statistics, 2010. CA
.
Cancer J Clin
2010
;
60
:
277
300
.
2.
Sargent
D
,
Sobrero
A
,
Grothey
A
,
O'Connell
MJ
,
Buyse
M
,
Andre
T
, et al
Evidence for cure by adjuvant therapy in colon cancer: observations based on individual patient data from 20,898 patients on 18 randomized trials
.
J Clin Oncol
2009
;
27
:
872
7
.
3.
Gray
R
,
Barnwell
J
,
McConkey
C
,
Hills
RK
,
Williams
NS
,
Kerr
DJ
. 
Adjuvant chemotherapy versus observation in patients with colorectal cancer: a randomised study
.
Lancet
2007
;
370
:
2020
9
.
4.
Compton
C
,
Fenoglio-Preiser
CM
,
Pettigrew
N
,
Fielding
LP
. 
American Joint Committee on Cancer prognostic factors consensus conference: Colorectal Working Group
.
Cancer
2000
;
88
:
1739
57
.
5.
Le Voyer
TE
,
Sigurdson
ER
,
Hanlon
AL
,
Mayer
RJ
,
Macdonald
JS
,
Catalano
PJ
, et al
Colon cancer survival is associated with increasing number of lymph nodes analyzed: a secondary survey of intergroup trial INT-0089
.
J Clin Oncol
2003
;
21
:
2912
9
.
6.
Krasna
MJ
,
Flancbaum
L
,
Cody
RP
,
Shneibaum
S
,
Ben Ari
G
. 
Vascular and neural invasion in colorectal carcinoma. Incidence and prognostic significance
.
Cancer
1988
;
61
:
1018
23
.
7.
Benson
AB
 3rd
,
Schrag
D
,
Somerfield
MR
,
Cohen
AM
,
Figueredo
AT
,
Flynn
PJ
. 
American Society of Clinical Oncology recommendations on adjuvant chemotherapy for stage II colon cancer
.
J Clin Oncol
2004
;
22
:
3408
19
.
8.
Samantas
E
,
Dervenis
C
,
Rigatos
SK
. 
Adjuvant chemotherapy for colon cancer: evidence on improvement in survival
.
Dig Dis
2007
;
25
:
67
75
.
9.
Allegra
CJ
,
Yothers
G
,
O'Connell
MJ
,
Sharif
S
,
Petrelli
NJ
,
Colangelo
LH
, et al
Phase III trial assessing bevacizumab in stages II and III carcinoma of the colon: results of NSABP protocol C-08
.
J Clin Oncol
2011
;
29
:
11
6
.
10.
Alberts
SR
,
Sinicrope
FA
,
Grothey
A
. 
N0147. A randomized phase III trial of oxaliplatin plus 5-fluorouracil/leucovorin with or without cetuximab after curative resection of stage III colon cancer
.
Clin Colorectal Cancer
2005
;
5
:
211
3
.
11.
De Gramont
A
,
Cutsem
EV
,
Tabernero
J
,
Moore
MJ
,
Cunningham
D
,
Rivera
F
, et al
AVANT: Results from a randomized, three-arm multinational phase III study to investigate bevacizumab with either XELOX or FOLFOX4 versus FOLFOX4 alone as adjuvant treatment for colon cancer
.
J Clin Oncol
2011
;
29
(
suppl 4
):
362[abstract]
.
12.
Cunningham
D
,
Atkin
W
,
Lenz
HJ
,
Lynch
HT
,
Minsky
B
,
Nordlinger
B
, et al
Colorectal cancer
.
Lancet
2010
;
375
:
1030
47
.
13.
Walther
A
,
Johnstone
E
,
Swanton
C
,
Midgley
R
,
Tomlinson
I
,
Kerr
D
. 
Genetic prognostic and predictive markers in colorectal cancer
.
Nat Rev Cancer
2009
;
9
:
489
99
.
14.
Chua
W
,
Moore
MM
,
Charles
KA
,
Clarke
SJ
. 
Predictive biomarkers of clinical response to targeted antibodies in colorectal cancer
.
Curr Opin Mol Ther
2009
;
11
:
611
22
.
15.
Lurje
G
,
Zhang
W
,
Lenz
HJ
. 
Molecular prognostic markers in locally advanced colon cancer
.
Clin Colorectal Cancer
2007
;
6
:
683
90
.
16.
Jiang
Y
,
Kimchi
ET
,
Staveley-O'Carroll
KF
,
Cheng
H
,
Ajani
JA
. 
Assessment of K-ras mutation: a step toward personalized medicine for patients with colorectal cancer
.
Cancer
2009
;
115
:
3609
17
.
17.
Jones
PA
,
Baylin
SB
. 
The epigenomics of cancer
.
Cell
2007
;
128
:
683
92
.
18.
Esteller
M
. 
Epigenetics in cancer
.
N Engl J Med
2008
;
358
:
1148
59
.
19.
Feinberg
AP
,
Tycko
B
. 
The history of cancer epigenetics
.
Nat Rev Cancer
2004
;
4
:
143
53
.
20.
Grady
WM
,
Carethers
JM
. 
Genomic and epigenetic instability in colorectal cancer pathogenesis
.
Gastroenterology
2008
;
135
:
1079
99
.
21.
Shen
L
,
Catalano
PJ
,
Benson
AB
 3rd
,
O'Dwyer
P
,
Hamilton
SR
,
Issa
JP
. 
Association between DNA methylation and shortened survival in patients with advanced colorectal cancer treated with 5-fluorouracil based chemotherapy
.
Clin Cancer Res
2007
;
13
:
6093
8
.
22.
Umetani
N
,
Takeuchi
H
,
Fujimoto
A
,
Shinozaki
M
,
Bilchik
AJ
,
Hoon
DS
. 
Epigenetic inactivation of ID4 in colorectal carcinomas correlates with poor differentiation and unfavorable prognosis
.
Clin Cancer Res
2004
;
10
:
7475
83
.
23.
Ogino
S
,
Nosho
K
,
Kirkner
GJ
,
Kawasaki
T
,
Meyerhardt
JA
,
Loda
M
, et al
CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon cancer
.
Gut
2009
;
58
:
90
6
.
24.
Toyota
M
,
Sasaki
Y
,
Satoh
A
,
Ogi
K
,
Kikuchi
T
,
Suzuki
H
, et al
Epigenetic inactivation of CHFR in human tumors
.
Proc Natl Acad Sci U S A
2003
;
100
:
7818
23
.
25.
Noetzel
E
,
Veeck
J
,
Niederacher
D
,
Galm
O
,
Horn
F
,
Hartmann
A
, et al
Promoter methylation-associated loss of ID4 expression is a marker of tumour recurrence in human breast cancer
.
BMC Cancer
2008
;
8
:
154
.
26.
Cho
CY
,
Wang
JH
,
Chang
HC
,
Chang
CK
,
Hung
WC
. 
Epigenetic inactivation of the metastasis suppressor RECK enhances invasion of human colon cancer cells
.
J Cell Physiol
2007
;
213
:
65
9
.
27.
Toyota
M
,
Ho
C
,
Ahuja
N
,
Jair
KW
,
Li
Q
,
Ohe-Toyota
M
, et al
Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification
.
Cancer Res
1999
;
59
:
2307
12
.
28.
Colella
S
,
Shen
L
,
Baggerly
KA
,
Issa
JP
,
Krahe
R
. 
Sensitive and quantitative universal Pyrosequencing methylation analysis of CpG sites
.
Biotechniques
2003
;
35
:
146
50
.
29.
Shen
L
,
Guo
Y
,
Chen
X
,
Ahmed
S
,
Issa
JP
. 
Optimizing annealing temperature overcomes bias in bisulfite PCR methylation analysis
.
Biotechniques
2007
;
42
:
48
,
50, 52 passim
.
30.
Hsieh
FY
,
Lavori
PW
. 
Sample-size calculations for the Cox proportional hazards regression model with nonbinary covariates
.
Control Clin Trials
2000
;
21
:
552
60
.
31.
Allegra
C
,
Sargent
DJ
. 
Adjuvant therapy for colon cancer—the pace quickens
.
N Engl J Med
2005
;
352
:
2746
8
.
32.
Scolnick
DM
,
Halazonetis
TD
. 
Chfr defines a mitotic stress checkpoint that delays entry into metaphase
.
Nature
2000
;
406
:
430
5
.
33.
Satoh
A
,
Toyota
M
,
Itoh
F
,
Sasaki
H
,
Ogi
K
,
Kikuchi
T
, et al
Epigenetic inactivation of CHFR and sensitivity to microtubule inhibitors in gastric cancer
.
Cancer Res
2003
;
63
:
8606
13
.
34.
Mizuno
K
,
Osada
H
,
Konishi
H
,
Tatematsu
Y
,
Yatabe
Y
,
Mitsudomi
T
, et al
Aberrant hypermethylation of the CHFR prophase checkpoint gene in human lung cancers
.
Oncogene
2002
;
21
:
2328
33
.
35.
Baba
S
,
Hara
A
,
Kato
K
,
Long
NK
,
Hatano
Y
,
Kimura
M
, et al
Aberrant promoter hypermethylation of the CHFR gene in oral squamous cell carcinomas
.
Oncol Rep
2009
;
22
:
1173
9
.
36.
Yu
X
,
Minter-Dykhouse
K
,
Malureanu
L
,
Zhao
WM
,
Zhang
D
,
Merkle
CJ
, et al
Chfr is required for tumor suppression and Aurora A regulation
.
Nat Genet
2005
;
37
:
401
6
.
37.
Privette
LM
,
Gonzalez
ME
,
Ding
L
,
Kleer
CG
,
Petty
EM
. 
Altered expression of the early mitotic checkpoint protein, CHFR, in breast cancers: implications for tumor suppression
.
Cancer Res
2007
;
67
:
6064
74
.
38.
Fukuda
T
,
Kondo
Y
,
Nakagama
H
. 
The anti-proliferative effects of the CHFR depend on the forkhead associated domain, but not E3 ligase activity mediated by ring finger domain
.
PLoS One
2008
;
3
:
e1776
.
39.
Kashima
L
,
Toyota
M
,
Mita
H
,
Suzuki
H
,
Idogawa
M
,
Ogi
K
, et al
CHFR, a potential tumor suppressor, downregulates interleukin-8 through the inhibition of NF-kappaB
.
Oncogene
2009
;
28
:
2643
53
.
40.
Oh
YM
,
Kwon
YE
,
Kim
JM
,
Bae
SJ
,
Lee
BK
,
Yoo
SJ
, et al
Chfr is linked to tumour metastasis through the downregulation of HDAC1
.
Nat Cell Biol
2009
;
11
:
295
302
.
41.
Brandes
JC
,
van Engeland
M
,
Wouters
KA
,
Weijinberg
MP
,
Herman
JG
. 
CHFR promoter hypermethylation in colon cancer correlates with the microsatellite instability phenotype
.
Carcinogenesis
2005
;
26
:
1152
6
.
42.
Fu
Z
,
Regan
K
,
Zhang
L
,
Muders
MH
,
Thibodeau
SN
,
French
A
, et al
Deficiencies in Chfr and Mlh1 synergistically enhance tumor susceptibility in mice
.
J Clin Invest
2009
;
119
:
2714
24
.
43.
Gryfe
R
,
Kim
H
,
Hsieh
E
,
Aronson
M
,
Holowaty
E
,
Bull
S
, et al
Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer
.
N Engl J Med
2000
;
342
:
69
77
.
44.
Sargent
DJ
,
Marsoni
S
,
Monges
G
,
Thibodeau
SN
,
Labianca
R
,
Hamilton
SR
, et al
Defective mismatch repair as a predictive marker for lack of efficacy of fluorouracil-based adjuvant therapy in colon cancer
.
J Clin Oncol
2010
;
28
:
3219
26
.
45.
Ribic
CM
,
Sargent
DJ
,
Moore
MJ
,
Thibodeau
SN
,
French
AJ
,
Goldberg
RM
, et al
Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer
.
N Engl J Med
2003
;
349
:
247
57
.
46.
Toyota
M
,
Ohe-Toyota
M
,
Ahuja
N
,
Issa
JP
. 
Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype
.
Proc Natl Acad Sci U S A
2000
;
97
:
710
5
.
47.
Nosho
K
,
Irahara
N
,
Shima
K
,
Kure
S
,
Kirkner
GJ
,
Schernhammer
ES
, et al
Comprehensive biostatistical analysis of CpG island methylator phenotype in colorectal cancer using a large population-based sample
.
PLoS One
2008
;
3
:
e3698
.
48.
Deng
G
,
Kakar
S
,
Tanaka
H
,
Matsuzaki
K
,
Miura
S
,
Sleisenger
MH
, et al
Proximal and distal colorectal cancers show distinct gene-specific methylation profiles and clinical and molecular characteristics
.
Eur J Cancer
2008
;
44
:
1290
301
.
49.
Ogino
S
,
Nosho
K
,
Kirkner
GJ
,
Kawasaki
T
,
Chan
AT
,
Schernhammer
ES
, et al
A cohort study of tumoral LINE-1 hypomethylation and prognosis in colon cancer
.
J Natl Cancer Inst
2008
;
100
:
1734
8
.
50.
Dunn
J
,
Baborie
A
,
Alam
F
,
Joyce
K
,
Moxham
M
,
Sibson
R
, et al
Extent of MGMT promoter methylation correlates with outcome in glioblastomas given temozolomide and radiotherapy
.
Br J Cancer
2009
;
101
:
124
31
.
51.
Valentini
AM
,
Armentano
R
,
Pirrelli
M
,
Caruso
ML
. 
Chemotherapeutic agents for colorectal cancer with a defective mismatch repair system: the state of the art
.
Cancer Treat Rev
2006
;
32
:
607
18
.