Purpose: Epigenetic aberrations have been shown to play an important role in the pathogenesis of most cancers. To investigate the clinical significance of epigenetic changes in neuroblastoma, we evaluated the relationship between clinicopathologic variables and the pattern of gene methylation in neuroblastoma cell lines and tumors.

Experimental Design: Methylation-specific PCR was used to evaluate the gene methylation status of 19 genes in 14 neuroblastoma cell lines and 8 genes in 70 primary neuroblastoma tumors. Associations between gene methylation, established prognostic factors, and outcome were evaluated. Log-rank tests were used to identify the number of methylated genes that was most predictive of overall survival.

Results: Epigenetic changes were detected in the neuroblastoma cell lines and primary tumors, although the pattern of methylation varied. Eight of the 19 genes analyzed were methylated in >70% of the cell lines. Epigenetic changes of four genes were detected in only small numbers of cell lines. None of the cell lines had methylation of the other seven genes analyzed. In primary neuroblastoma tumors, high-risk disease and poor outcome were associated with methylation of DCR2, CASP8, and HIN-1 individually. Although methylation of the other five individual genes was not predictive of poor outcome, a trend toward decreased survival was seen in patients with a methylation phenotype, defined as ≥4 methylated genes (P = 0.055).

Conclusion: Our study indicates that clinically aggressive neuroblastoma tumors have aberrant methylation of multiple genes and provides a rationale for exploring treatment strategies that include demethylating agents.

Neuroblastoma, a childhood neoplasm arising from neural crest cells, is characterized by a diversity of clinical behaviors ranging from spontaneous remission to rapid tumor progression and death (1). Genetic abnormalities in neuroblastoma tumors, such as MYCN oncogene amplification, allelic losses of chromosomes 1p and 11q, and gain of chromosome 17q, are predictive of outcome (2). More recent studies have indicated that epigenetic aberrations also contribute to neuroblastoma pathogenesis (3, 4). Abnormal hypermethylation of several genes has been reported in neuroblastoma including the angiogenesis inhibitor gene thrombospondin-1 (TSP-1; ref. 3), the CD44 adhesion receptor gene (5), the tumor suppressor gene RASSFIA (6, 7), the nuclear receptor 1I2 (8), and CASP8 (9), as well as other genes involved in the tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) pathway to apoptosis (10). Although the prognostic significance of epigenetic changes of single genes remains controversial, global methylation studies have indicated that poor prognosis is associated with a CpG island methylator phenotype (4, 11, 12).

In the present study, we examined the methylation status of 19 genes that are known to be aberrantly hypermethylated in adult cancers in 14 genetically heterogeneous neuroblastoma cell lines with disparate growth characteristics. We also evaluated the methylation status of eight genes that have a high incidence of epigenetic aberrations in neuroblastoma cell lines, in 70 primary neuroblastoma tumors, 5 benign ganglioneuromas, and normal adrenal tissue. Abnormal methylation of three individual genes (DCR2, CASP8, and HIN-1) was found to be statistically associated with high-risk factors and poor outcome. The methylation status of the other five individual genes we examined (HIC-1, RASSF1A, BLU, TMS-1, and TIG-1) had no effect on survival. However, a trend was seen associating decreased survival in patients with tumors with a methylation phenotype, defined as ≥4 methylated genes. Our results suggest that aberrant methylation of multiple genes is likely to contribute to neuroblastoma pathogenesis.

Cells and culture conditions. The biological and genetic characteristics of the neuroblastoma cell lines used in this study have previously been described (7). Neuroblastoma cell lines were grown at 5% CO2 in RPMI 1640 (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen), l-glutamine, and antibiotics.

Patients and tumor specimens. DNA from 70 primary, untreated neuroblastoma tumors was obtained from the Children's Oncology Group (COG) Neuroblastoma Tumor Bank. Patients who met the following criteria were identified: patient had an available tumor specimen; patient had an available serum specimen; and patient either died or had >1 year of follow-up time. From this group of patients, 70 patients were selected at random (Table 1). The COG tumor samples were linked to clinical and biological information in the COG Statistical and Data Center and the laboratory investigators were blinded to these data. DNA was also isolated from five ganglioneuromas from children diagnosed at Children's Memorial Hospital in Chicago. This study was approved by the Northwestern University Institutional Review Board, the Children's Memorial Institutional Review Board, and the COG Neuroblastoma Biology Committee.

Table 1.

Clinical and biological characteristics of 70 neuroblastoma patients from COG

CharacteristicNo. patients (%)3-y overall survival rate (SE)P*
Sex    
    Boys 37 (53) 77 (8)  
    Girls 33 (47) 100 (0) 0.005 
Age at diagnosis, y    
    <1 24 (34) 100 (0)  
    ≥1 46 (66) 82 (7) 0.027 
Diagnosis    
    Ganglioneuroma 0 (0)   
    Ganglioneuroblastoma 11 (16) 83 (15)  
    Neuroblastoma 59 (84) 89 (4) 0.806 
International Neuroblastoma Staging System stage    
    I 16 (23) 100 (0)  
    IIa 3 (4) 50 (35)  
    IIb 10 (14) 100 (0)  
    III 21 (30) 91 (6)  
    IVs 3 (4) 100 (0)  
    IV 17 (24) 76 (11) 0.009 
MYCN status    
    Amplified 9 (13) 56 (17)  
    Nonamplified 60 (89) 93 (4) <0.001 
Risk group    
    Non–high risk 47 (67) 96 (4)  
    High risk 23 (33) 73 (9) <0.001 
CharacteristicNo. patients (%)3-y overall survival rate (SE)P*
Sex    
    Boys 37 (53) 77 (8)  
    Girls 33 (47) 100 (0) 0.005 
Age at diagnosis, y    
    <1 24 (34) 100 (0)  
    ≥1 46 (66) 82 (7) 0.027 
Diagnosis    
    Ganglioneuroma 0 (0)   
    Ganglioneuroblastoma 11 (16) 83 (15)  
    Neuroblastoma 59 (84) 89 (4) 0.806 
International Neuroblastoma Staging System stage    
    I 16 (23) 100 (0)  
    IIa 3 (4) 50 (35)  
    IIb 10 (14) 100 (0)  
    III 21 (30) 91 (6)  
    IVs 3 (4) 100 (0)  
    IV 17 (24) 76 (11) 0.009 
MYCN status    
    Amplified 9 (13) 56 (17)  
    Nonamplified 60 (89) 93 (4) <0.001 
Risk group    
    Non–high risk 47 (67) 96 (4)  
    High risk 23 (33) 73 (9) <0.001 
*

Log-rank test.

Stage IV versus non–stage IV.

DNA isolation and bisulfite modification. Total genomic DNA was extracted from the neuroblastoma cell lines and ganglioneuroma tumors using the Genomic-tip and DNeasy tissue kit (Qiagen) and modified by sodium bisulfite using the CpGenome DNA Modification Kit (Intergen Co.). Genomic DNA from human normal adrenal tissue was purchased from BioChain Institute, Inc. As previously described (3), 1 μg of genomic DNA was denatured by NaOH and modified by sodium bisulfite, which converts all unmethylated cytosines to uracils. The modified DNA was desulfonated with NaOH and purified.

Methylation analysis. Bisulfite-modified DNA was amplified, as previously described (3, 13), using primers specific for methylated and unmethylated sequences of 19 gene promoter regions. The PCR assays were done using conditions previously described (4, 9, 1430). The PCR products were separated by electrophoresis on a 2.5% agarose gel and visualized under UV illumination using ethidium bromide staining. Universal Methylated DNA (Intergen), which is enzymatically methylated human genomic DNA, was used as a positive control. Primers and PCR conditions are shown in Supplementary Table S1.

Statistical analysis. Data about the methylation status of the eight genes assayed in the 70 primary neuroblastoma samples were forwarded to the COG Statistical and Data Center for statistical analyses. The clinical data for this analysis were frozen in April 2006. Patients were classified as high risk if they had stage IV disease and were over the age of 1 year or had stage III MYCN-amplified tumors. All other patients were considered non–high risk. The methods of Kaplan and Meier were used to estimate 3-year overall survival rates, and survival curves across groups were compared using the log-rank test (31). Survival rates are presented as the rate ± SE, where the SEs are computed using Greenwood's formula (32). Associations of the methylation status (unmethylated versus partially or completely methylated) of the different genes were examined using a two-sided Fisher's exact test (33). Log-rank tests were used to identify the number of methylated (either partially or completely) genes that was most predictive of overall survival. For a given gene, a Cox proportional hazards model was used to estimate the effect of methylation status (0, unmethylated; 1, partially methylated; 2, completely methylated) on the relative hazard of risk for death after adjustment for MYCN amplification (34). As appropriate for small exploratory studies, no adjustment for multiple comparisons was made in this study, and P < 0.05 was considered statistically significant.

A methylation phenotype was defined on the basis of results of log-rank tests of the effect of number of methylated (either completely or partially) genes on overall survival. The lowest P value in Table 2 was used to identify an optimal cutoff of the number of methylated genes.

Table 2.

Results of log-rank tests for effect of number of methylated genes on overall survival in 70 neuroblastoma patients

No. methylated genesNo. patients with profileP
≥1 70  
≥2 69 0.719 
≥3 58 0.214 
≥4 46 0.055 
≥5 29 0.361 
≥6 17 0.106 
≥7 10 0.059 
≥8 0.439 
No. methylated genesNo. patients with profileP
≥1 70  
≥2 69 0.719 
≥3 58 0.214 
≥4 46 0.055 
≥5 29 0.361 
≥6 17 0.106 
≥7 10 0.059 
≥8 0.439 

Profile of promoter hypermethylation in neuroblastoma cell lines. Methylation-specific PCR was used to examine the promoter hypermethylation profile in 14 neuroblastoma cell lines. Gene methylation frequencies varied from 0% to 100%. Aberrant methylation of 8 of the 19 genes analyzed was detected at high frequency in the cell lines [DCR2 (100%); TMS-1 and BLU (93%); HIC-1 (86%); E-cadherin, DCR4, and TIG-1 (79%); and HIN-1 (71%)]. Four genes (ENDRB, MGMT, CDH13, and IRF-7) were methylated in 1% to 30% of the cell lines. The remaining seven genes (RB1, p27, Dkk-3, VHL, MCT-1, PTEN, and BRCA1) were unmethylated in all of the cell lines (Supplementary Fig. S1 and Fig. 1).

Fig. 1.

Methylation profile of 19 genes in 14 neuroblastoma cell lines. Red, genes with complete methylation; yellow, genes with partial methylation; green, genes that are unmethylated.

Fig. 1.

Methylation profile of 19 genes in 14 neuroblastoma cell lines. Red, genes with complete methylation; yellow, genes with partial methylation; green, genes that are unmethylated.

Close modal

Clinical and biological characteristics of patient cohort. To investigate the clinical significance of epigenetic aberrations in neuroblastoma, we examined the methylation status of 8 genes in 70 primary tumors, 5 ganglioneuromas, and normal adrenal tissue. The genes selected for these experiments all have a high frequency of epigenetic aberrations in neuroblastoma cell lines. Six of the genes (DCR2, TMS-1, BLU, HIC-1, TIG-1, HIN-1) were found to be methylated in >70% of the neuroblastoma cell lines we examined in this study. We also evaluated the methylation status of two additional genes (CASP8 and RASSF1A) that are known to have tumor suppressor function and have previously been shown to be methylated at high frequency in neuroblastoma (7, 14, 35). Due to limited quantities of available tumor DNA, we were unable to evaluate two of the genes (E-cadherin and DCR4) that were methylated at high frequencies in the neuroblastoma cell lines in the primary tumor samples. All the genes selected for the primary tumor studies are known to have tumor suppressor activity and/or epigenetic changes of these genes have been shown to be clinically relevant in other types of cancer.

The clinical and biological features of the patients and tumors are summarized in Table 1. The patients were diagnosed between February 2000 and July 2003, and 3-year overall survival for the entire cohort was 88 ± 4%, with median follow-up time for living patients of 4 years. Children ≥1 year of age at diagnosis had a statistically worse outcome than infants (3-year overall survival rate of 82% versus 100%; P = 0.027), and patients with stage IV disease had decreased survival compared with those with locoregional tumors (3-year overall survival rate of 76% versus 92%; P = 0.009). As expected, MYCN amplification was also an unfavorable prognostic factor in the cohort. The estimated 3-year overall survival rate for patients with MYCN-amplified tumors was 56 ± 17%, compared with 93 ± 4% for those with nonamplified neuroblastomas (P < 0.001). Patients were classified as high risk if they had stage IV disease and were over the age of 1 year or had stage III MYCN-amplified tumors. The estimated 3-year overall survival rate for patients with non–high-risk disease was 96 ± 4%, compared with 73 ± 9% for high-risk patients (P < 0.001).

Profile of promoter methylation in neuroblastoma tumors. Methylation frequency of the eight genes in the neuroblastoma tumors ranged from 21% to 99%. Ninety-nine percent of the tumors showed methylation of HIC-1. RASSF1A methylation was detected in 90% of the neuroblastoma tumors. Methylation of CASP8, BLU, TMS-1, and DCR2 was detected in 56%, 54%, 46%, and 44% of the tumors, respectively, whereas methylation of TIG-1 and HIN-1 was observed in a smaller cohort (23% and 21%, respectively). All 70 tumors had methylation of at least one gene (Fig. 2). Seven genes (CASP8, DCR2, HIN-1, RASSF1A, BLU, TMS-1, and TIG-1) were methylated in a subset of neuroblastoma tumors but remained unmethylated in the ganglioneuromas and adrenal tissue (Fig. 2). The other gene analyzed (HIC-1) was methylated in neuroblastoma tumors as well as the ganglioneuromas and adrenal tissue. Further analyses revealed significant associations between methylation of BLU and CASP8 (P = 0.030), TIG1 and CASP8 (P = 0.022), DCR2 and CASP8 (P = 0.030), HIN-1 and CASP8 (P = 0.042), TMS-1 and BLU (P < 0.001), TIG-1 and BLU (P = 0.003), HIN-1 and BLU (P = 0.039), TIG1 and TMS-1 (P = 0.044), and HIN-1 and DCR2 (P = 0.018; Supplementary Table S2).

Fig. 2.

Methylation profile of 8 genes in 70 neuroblastoma primary tumors and 5 ganglioneuromas. Red, genes with strong methylation; yellow, genes with weak methylation; green, genes that are unmethylated.

Fig. 2.

Methylation profile of 8 genes in 70 neuroblastoma primary tumors and 5 ganglioneuromas. Red, genes with strong methylation; yellow, genes with weak methylation; green, genes that are unmethylated.

Close modal

Methylation profile and clinical outcome. Patients with methylation of CASP8, DCR2, or HIN-1 had significantly decreased overall survival compared with those with tumors that lacked methylation of these individual genes (P = 0.008, P = 0.019, and P < 0.001, respectively; Fig. 3). The methylation status of the other five individual genes was not associated with outcome. However, a trend toward decreased survival was seen in patients with tumors with a methylation phenotype of ≥4 methylated genes (P = 0.055).

Fig. 3.

Kaplan-Meier survival curves for neuroblastoma patients. A, overall survival according to the methylation status of CASP8: methylated (n = 39) and unmethylated (n = 31); P = 0.008. B, overall survival according to the methylation status of DCR2: methylated (n = 31) and unmethylated (n = 39); P = 0.019. C, overall survival according to the methylation status of HIN-1: methylated (n = 15) and unmethylated (n = 55); P < 0.001. D, overall survival curves for all the patients enrolled onto the study according to the methylation profile. Methylated genes ≥4 (n = 46); methylated genes ≤3 (n = 24); P = 0.055.

Fig. 3.

Kaplan-Meier survival curves for neuroblastoma patients. A, overall survival according to the methylation status of CASP8: methylated (n = 39) and unmethylated (n = 31); P = 0.008. B, overall survival according to the methylation status of DCR2: methylated (n = 31) and unmethylated (n = 39); P = 0.019. C, overall survival according to the methylation status of HIN-1: methylated (n = 15) and unmethylated (n = 55); P < 0.001. D, overall survival curves for all the patients enrolled onto the study according to the methylation profile. Methylated genes ≥4 (n = 46); methylated genes ≤3 (n = 24); P = 0.055.

Close modal

Both high-risk disease (P < 0.001) and MYCN amplification (P = 0.001) were associated with the number of methylated genes. When each gene was tested with MYCN in the multivariable Cox analysis, only MYCN was predictive of overall survival with the exception of TIG-1 (Table 3). Lack of TIG-1 methylation was independently predictive of overall survival after adjustment for MYCN status. However, as stated above, TIG-1 was not identified as one of the genes whose methylation status was statistically significantly predictive of survival in univariate analysis (Supplementary Fig. S2; P = 0.840). The hazard ratio of RASSF1A was not estimable in this small sample, either with or without MYCN adjustment, due to the unbalanced distribution of deaths.

Table 3.

Results of Cox regression models for methylation status (0, unmethylated; 1, partially methylated; and 2, completely methylated) of each gene considered separately as a predictor of overall survival, with adjustment for MYCN, in 70 neuroblastoma patients

GeneOverall survival hazard ratioP
HIC-1 0.41 0.090 
CASP8 1.53 0.534 
BLU 0.33 0.109 
TMS-1 0.59 0.251 
TIG-1 0.32 0.034 
DCR2 2.84 0.081 
RASSF1A Not estimable*  
HIN-1 0.372 0.065 
GeneOverall survival hazard ratioP
HIC-1 0.41 0.090 
CASP8 1.53 0.534 
BLU 0.33 0.109 
TMS-1 0.59 0.251 
TIG-1 0.32 0.034 
DCR2 2.84 0.081 
RASSF1A Not estimable*  
HIN-1 0.372 0.065 
*

Hazard ratio is infinitely large.

Epigenetic silencing of tumor suppressor genes plays an important role in the pathogenesis of many types of cancer. Although the frequency of aberrant gene methylation in neuroblastoma has not been extensively studied, recent studies indicate that epigenetic changes are likely to influence neuroblastoma phenotype (3, 5, 7, 10). In this study, we used methylation-specific PCR to examine the status of promoter CpG island methylation of 19 genes in 14 neuroblastoma cell lines. We found that 8 (TMS-1, DCR2, HIC-1, BLU, E-cadherin, HIN-1, DCR4 and TIG-1) of the 19 genes were methylated in >70% of the cell lines. Methylation frequency of other 11 genes varies from 0% to 30%. We next evaluated the methylation status of eight genes (HIC-1, BLU, TMS-1, TIG-1, DCR2, HIN-1, CASP8, and RASSF1A), which are frequently methylated in neuroblastoma cell lines, in 70 neuroblastoma tumor samples, 5 ganglioneuromas, and normal adrenal tissue. HIC-1 methylation was detected in more than 95% of the tumors, as well as the adrenal tissue and the 5 ganglioneuroma samples tested. The frequency of methylation of the other seven genes (BLU, TMS-1, TIG-1, DCR2, HIN-1, CASP8, and RASSF1A) was significantly lower in the primary tumors compared with the cell lines (Table 1), suggesting that epigenetic aberrations may provide a growth advantage to neuroblastoma cells cultured in vitro. None of these genes was methylated in the ganglioneuroma or adrenal tissue samples.

To investigate the clinical significance of the epigenetic aberrations, associations between gene methylation, established prognostic factors, and outcome were evaluated. Tumor samples for this study were obtained from the COG Neuroblastoma Tumor Bank, and tumors representative of low, intermediate, and high risk were provided. However, only 24% of the patients had stage IV disease and 33% were classified as high risk. Furthermore, the overall survival for the study cohort was 88 ± 4%, which is significantly better than other large unselected series (36). Although this cohort contained a large number of children with favorable biology tumors, as expected, decreased survival was associated with established unfavorable prognostic clinical and biological factors including advanced stage disease and MYCN amplification.

Epigenetic aberrations of HIN-1, DCR2, or CASP8 were also significantly associated with decreased 3-year survival in this patient cohort. Although the methylation status of the other individual genes was not associated with outcome, after adjustment for MYCN, lack of TIG-1 methylation was found to be predictive of decreased survival. It is likely that this was a spurious occurrence as a result of the small sample size and lack of adjustment for multiple comparisons. Log-rank tests were used to identify the number of methylated genes that was most predictive of overall survival. This analysis showed a trend associating poor survival with a methylation phenotype, defined as ≥4 methylated genes (P = 0.055). It is likely that the prognostic effect of the methylation phenotype would be increased in a more representative cohort of patients.

The HIN-1 tumor suppressor gene has previously been shown to be methylated in neuroblastoma (37). In the study by Shigematzu et al., HIN-1 methylation was detected in 11% of the neuroblastomas analyzed. However, in contrast to our study, these investigators did not examine the prognostic significance of HIN-1 methylation. Aberrant methylation of HIN-1 has also been observed in a variety of other cancers including breast cancer (26, 38), nasopharyngeal carcinoma (39), lung cancer, retinoblastoma, Wilms' tumor, rhabdomyosarcoma, prostate cancer, and lymphomas (37, 40). Although the function of HIN-1 in neuroblastoma remains unclear, in breast cancer cell lines, HIN-1 is capable of inhibiting cell cycle reentry, suppressing migration and invasion, and inducing apoptosis. These effects seem to be mediated by a high-affinity cell-surface receptor and to involve the modulation of the AKT signaling pathway (41).

DCR2 gene methylation has also been shown to be aberrantly hypermethylated in a variety of tumor types including neuroblastoma (14, 23). In a study by Banelli et al. (14), DCR2 methylation was detected in 13 of 31 (42%) neuroblastoma tumors. Similar to our results, none of the 13 benign ganglioneuromas evaluated by Banelli and colleagues had DCR2 methylation. Furthermore, patients with tumors with partial or complete DCR2 methylation had significantly lower overall survival compared with those with tumors that remained unmethylated (14). These investigators also found that the prognostic value of the combination of RASSF1A and DCR2 methylation was increased compared with DCR2 methylation alone. Inactivation of DCR2 should have antitumor effects by rendering the cells more sensitive to TRAIL-induced apoptosis. However, the down-regulation of DCR2 could be considered part of an inefficient defense mechanism activated to inhibit tumor cell growth (14).

CASP8 methylation was first reported in neuroblastoma tumors by Teitz et al. (9) nearly 6 years ago. These investigators detected methylation of CASP8 in 26 of 42 (62%) neuroblastoma tumors, and an association between CASP8 methylation and MYCN amplification was reported (9). In our series, CASP8 methylation also occurred more frequently in MYCN-amplified tumors, although this association did not reach statistical significance. Others, however, have reported no correlation between CASP8 silencing and MYCN amplification (42). The reasons for the discordant results are not clear but are likely due to differences in patient cohorts and disparities in the regions of CASP8 analyzed. Caspase-8 is a key determinant of sensitivity for apoptosis, and loss of CASP8 expression has recently been shown to render neuroblastoma cells refractory to integrin-mediated death, thereby promoting cell survival in the stromal microenvironment and metastases (43).

Several investigators have shown that restoration of CASP8 expression sensitizes tumor cell lines for death receptor–triggered apoptosis and drug-induced apoptosis (44, 45), suggesting that high levels of CASP8 expression will enhance response to chemotherapy and improve outcome. However, Fulda et al. (46) recently reported that loss of CASP8 protein expression occurs in the majority of tumors and that the level of expression has no effect on survival. In that study, no correlation between CASP8 expression and MYCN amplification or other high-risk features was seen. Although these investigators did not examine CASP8 methylation, the results suggest that epigenetic changes of CASP8 may be independent from gene silencing. It is well known that gene silencing can be caused by multiple mechanisms, and Banelli et al. (42) reported that CASP8 expression is not directly dependent on promoter methylation. Similarly, Abe et al. (11) have reported that methylation of the protocadherin β (PCDHB) gene was not associated with gene silencing. However, a close correlation between PCDHB gene methylation and poor outcome in patients with neuroblastoma was seen.

Epigenetic changes of the other individual genes were not found to be predictive of poor outcome in our cohort. However, a trend associating decreased survival in patients with tumors with a methylation phenotype was seen. A larger series should be analyzed to determine if methylation will be useful in risk stratification of high-risk patients. Recently, global methylation studies have indicated that a CpG island methylator phenotype is predictive of poor outcome in a variety of different cancers including neuroblastoma (11, 47). In a study by Abe and coworkers, a CpG island methylator phenotype was detected in 37 of the 38 MYCN-amplified neuroblastoma tumors studied. Furthermore, this phenotype was strongly predictive of poor survival in the cases without MYCN amplification. Moreover, in that study, a CpG island methylator phenotype was significantly associated with methylation of the RASSF1A and BLU tumor suppressor genes.

An epigenetic mechanism for carcinogenesis was predicted more than 10 years ago (48), and there is now strong molecular evidence supporting this concept. In many types of cancer, abnormal methylation and subsequent silencing of genes known to play important roles in tumor suppression, cell cycle regulation, apoptosis, DNA repair, and metastatic potential are observed at high frequency (49). Our study and others indicate that neuroblastoma phenotype is influenced by epigenetic changes of multiple genes. These results provide a rationale for exploring treatment strategies that include demethylating agents.

Grant support: NIH/National Institute of Neurological Disorders and Stroke grant NS049814, the Neuroblastoma Children's Cancer Society, Friends for Steven Pediatric Cancer Research Fund, the Elise Anderson Neuroblastoma Research Fund, Neuroblastoma Kids, the Robert H. Lurie Comprehensive Cancer Center, NIH National Cancer Institute Core Grant 5P30CA60553, and Children's Cancer Fund Grant (Q. Yang).

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.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

We thank the Children's Oncology Group (COG) Neuroblastoma Biology Committee for approving this study and providing the neuroblastoma DNA samples.

1
Cheung NK, Cohn SL. editors. Neuroblastoma. Heidelberg: Springer-Verlag; 2005.
2
Brodeur GM. Neuroblastoma: biological insights into a clinical enigma.
Nat Rev Cancer
2003
;
3
:
203
–16.
3
Yang Q, Liu S, Tian Y, et al. Methylation-associated silencing of the thrombospondin-1 gene in human neuroblastoma.
Cancer Res
2003
;
63
:
6299
–310.
4
Kuroki T, Trapasso F, Yendamuri S, et al. Allelic loss on chromosome 3p21.3 and promoter hypermethylation of semaphorin 3B in non-small cell lung cancer.
Cancer Res
2003
;
63
:
3352
–5.
5
Yan P, Muhlethaler A, Bourloud KB, Beck MN, Gross N. Hypermethylation-mediated regulation of CD44 gene expression in human neuroblastoma.
Genes Chromosomes Cancer
2003
;
36
:
129
–38.
6
Astuti D, Agathanggelou A, Honorio S, et al. RASSF1A promoter region CpG island hypermethylation in phaeochromocytomas and neuroblastoma tumours.
Oncogene
2001
;
20
:
7573
–7.
7
Yang Q, Zage P, Kagan D, et al. Association of epigenetic inactivation of RASSF1A with poor outcome in human neuroblastoma.
Clin Cancer Res
2004
;
10
:
8493
–500.
8
Misawa A, Inoue J, Sugino Y, et al. Methylation-associated silencing of the nuclear receptor 1I2 gene in advanced-type neuroblastomas, identified by bacterial artificial chromosome array-based methylated CpG island amplification.
Cancer Res
2005
;
65
:
10233
–42.
9
Teitz T, Wei T, Valentine MB, et al. Caspase 8 is deleted or silenced preferentially in childhood neuroblastomas with amplification of MYCN.
Nat Med
2000
;
6
:
529
–35.
10
van Noesel MM, van Bezouw S, Voute PA, et al. Clustering of hypermethylated genes in neuroblastoma.
Genes Chromosomes Cancer
2003
;
38
:
226
–33.
11
Abe M, Ohira M, Kaneda A, et al. CpG island methylator phenotype is a strong determinant of poor prognosis in neuroblastomas.
Cancer Res
2005
;
65
:
828
–34.
12
Issa JP. Opinion—CpG island methylator phenotype in cancer.
Nat Rev Cancer
2004
;
4
:
988
–93.
13
Yang Q, Liu S, Tian Y, et al. Methylation-associated silencing of the heat shock protein 47 gene in human neuroblastoma.
Cancer Res
2004
;
64
:
4531
–8.
14
Banelli B, Gelvi I, Di Vinci A, et al. Distinct CpG methylation profiles characterize different clinical groups of neuroblastic tumors.
Oncogene
2005
;
24
:
5619
–28.
15
Asada K, Miyamoto K, Fukutomi T, et al. Reduced expression of GNA11 and silencing of MCT1 in human breast cancers.
Oncology
2003
;
64
:
380
–8.
16
Lo KW, Tsang YS, Kwong J, et al. Promoter hypermethylation of the EDNRB gene in nasopharyngeal carcinoma.
Int J Cancer
2002
;
98
:
651
–5.
17
Athale UH, Stewart C, Kuttesch JF, et al. Phase I study of combination topotecan and carboplatin in pediatric solid tumors.
J Clin Oncol
2002
;
20
:
88
–95.
18
Zysman MA, Chapman WB, Bapat B. Considerations when analyzing the methylation status of PTEN tumor suppressor gene.
Am J Pathol
2002
;
160
:
795
–800.
19
Akao Y, Seto M, Yamamoto K, et al. The RCK gene associated with t(11:14) translocation is distinct from the MLL/ALL-1 gene with t(4:11) and t(11:19) translocations.
Cancer Res
1992
;
52
:
6083
–7.
20
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 U S A
1996
;
93
:
9821
–6.
21
Roman-Gomez J, Jimenez-Velasco A, Agirre X, et al. Transcriptional silencing of the Dickkopfs-3 (Dkk-3) gene by CpG hypermethylation in acute lymphoblastic leukaemia.
Br J Cancer
2004
;
91
:
707
–13.
22
Sakai M, Hibi K, Koshikawa K, et al. Frequent promoter methylation and gene silencing of CDH13 in pancreatic cancer.
Cancer Sci
2004
;
95
:
588
–91.
23
van Noesel MM, van Bezouw S, Salomons GS, et al. Tumor-specific down-regulation of the tumor necrosis factor-related apoptosis-inducing ligand decoy receptors DcR1 and DcR2 is associated with dense promoter hypermethylation.
Cancer Res
2002
;
62
:
2157
–61.
24
Agathanggelou A, Dallol A, Zochbauer-Muller S, et al. Epigenetic inactivation of the candidate 3p21.3 suppressor gene BLU in human cancers.
Oncogene
2003
;
22
:
1580
–8.
25
Yoshikawa H, Matsubara K, Qian GS, et al. SOCS-1, a negative regulator of the JAK/STAT pathway, is silenced by methylation in human hepatocellular carcinoma and shows growth-suppression activity.
Nat Genet
2001
;
28
:
29
–35.
26
Fackler MJ, McVeigh M, Evron E, et al. DNA methylation of RASSF1A, Hin-1, RAR-β, cyclin D2 and twist in in situ and invasive lobular breast carcinoma.
Int J Cancer
2003
;
107
:
970
–5.
27
Dong SM, Kim HS, Rha SH, Sidransky D. Promoter hypermethylation of multiple genes in carcinoma of the uterine cervix.
Clin Cancer Res
2001
;
7
:
1982
–6.
28
Youssef EM, Chen XQ, Higuchi E, et al. Hypermethylation and silencing of the putative tumor suppressor Tazarotene-induced gene 1 in human cancers.
Cancer Res
2004
;
64
:
2411
–7.
29
Simpson DJ, Hibberts NA, McNicol AM, Clayton RN, Farrell WE. Loss of pRb expression in pituitary adenomas is associated with methylation of the RB1 CpG island.
Cancer Res
2000
;
60
:
1211
–6.
30
Stone AR, Bobo W, Brat DJ, et al. Aberrant methylation and down-regulation of TMS1/ASC in human glioblastoma.
Am J Pathol
2004
;
165
:
1151
–61.
31
Kaplan EL, Meier P. Nonparametric estimation from incomplete observations.
J Am Stat Assoc
1958
;
53
:
457
–81.
32
Greenwood M. The natural duration of cancer. Vol. 33. London: His Majesty's Stationery Office; 1926.
33
Agresti A. Categorical data analysis. 2nd ed. Hoboken (NJ): Wiley; 2002.
34
Kalbfleisch JD and Prentice RL. The statistical analysis of failture time data. 2nd ed. Hoboken (NJ): Wiley; 2002.
35
Alaminos M, Davalos V, Cheung NKV, Gerald WL, Esteller M. Clustering of gene hypermethylation associated with clinical risk groups in neuroblastoma.
J Natl Cancer Inst
2004
;
96
:
1208
–19.
36
Gatta G, Capocaccia R, Coleman MP, Ries LAG, Berrino F. Childhood cancer survival in Europe and the United States.
Cancer
2002
;
95
:
1767
–72.
37
Shigematsu H, Suzuki M, Takahashi T, et al. Aberrant methylation of HIN-1 (high in normal-1) is a frequent event in many human malignancies.
Int J Cancer
2005
;
113
:
600
–4.
38
Krop IE, Sgroi D, Porter DA, et al. HIN-1, a putative cytokine highly expressed in normal but not cancerous mammary epithelial cells.
Proc Natl Acad Sci U S A
2001
;
98
:
9796
–801.
39
Wong TS, Kwong DLW, Sham JST, et al. Promoter hypermethylation of high-in-normal 1 gene in primary nasopharyngeal carcinoma.
Clin Cancer Res
2003
;
9
:
3042
–6.
40
Krop I, Player A, Tablante A, et al. Frequent HIN-1 promoter methylation and lack of expression in multiple human tumor types.
Mol Cancer Res
2004
;
2
:
489
–94.
41
Krop I, Parker MT, Bloushtain-Qimron N, et al. HIN-1, an inhibitor of cell growth, invasion, and AKT activation.
Cancer Res
2005
;
65
:
9659
–69.
42
Banelli B, Casciano I, Croce M, et al. Expression and methylation of CASP8 in neuroblastoma: identification of a promoter region.
Nat Med
2002
;
8
:
1333
–5.
43
Stupack DG, Teitz T, Potter MD, et al. Potentiation of neuroblastoma metastasis by loss of caspase-8.
Nature
2006
;
439
:
95
–9.
44
Fulda S, Dockhorn-Dworniczak B, Debatin K. Sensitization of resistant tumor cells for chemotherapy- or death-receptor-induced, apoptosis by restoration of absent caspase-8 expression through demethylation or caspase-8 gene transfer.
Clin Cancer Res
2001
;
7
:
3707
–8S.
45
Muhlethaler-Mottet A, Bourloud KB, Auderset K, Joseph JM, Gross N. Drug-mediated sensitization to TRAIL-induced apoptosis in caspase-8-complemented neuroblastoma cells proceeds via activation of intrinsic and extrinsic pathways and caspase-dependent cleavage of XIAP, Bcl-x(L) and RIP.
Oncogene
2004
;
23
:
5415
–25.
46
Fulda S, Poremba C, Berwanger B, et al. Loss of caspase-8 expression does not correlate with MYCN amplification, aggressive disease, or prognosis in neuroblastoma.
Cancer Res
2006
;
66
:
10016
–23.
47
Roman-Gomez J, Jimenez-Velasco A, Agirre X, et al. Lack of CpG island methylator phenotype defines a clinical subtype of T-cell acute lymphoblastic leukemia associated with good prognosis.
J Clin Oncol
2005
;
23
:
7043
–9.
48
Holliday R. Epigenetic inheritance based on DNA methylation.
EXS
1993
;
64
:
452
–68.
49
Esteller M, Corn PG, Baylin SB, Herman JG. A gene hypermethylation profile of human cancer.
Cancer Res
2001
;
61
:
3225
–9.