Cancer Letters 273 (2009) 336–346
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
Aberrant methylation of candidate tumor suppressor genes in neuroblastoma Jasmien Hoebeeck a, Evi Michels a, Filip Pattyn a, Valérie Combaret b, Joëlle Vermeulen a, Nurten Yigit a, Claire Hoyoux c, Geneviève Laureys d, Anne De Paepe a, Frank Speleman a, Jo Vandesompele a,* a
Center for Medical Genetics, Ghent University Hospital, Ghent, Belgium Molecular Oncology Unit, Centre Léon Bérard, Lyon, France c Department of Pediatric Oncology, Hôpital de la Citadelle, Liège, Belgium d Department of Pediatric Hematology and Oncology, Ghent University Hospital, Ghent, Belgium b
a r t i c l e
i n f o
Article history: Received 19 June 2008 Received in revised form 19 July 2008 Accepted 18 August 2008
Keywords: Neuroblastoma Methylation MSP
a b s t r a c t CpG island hypermethylation has been recognized as an alternative mechanism for tumor suppressor gene inactivation. In this study, we performed methylation-speciﬁc PCR (MSP) to investigate the methylation status of 10 selected tumor suppressor genes in neuroblastoma. Seven of the investigated genes (CD44, RASSF1A, CASP8, PTEN, ZMYND10, CDH1, PRDM2) showed high frequencies (P30%) of methylation in 33 neuroblastoma cell lines. In 42 primary neuroblastoma tumors, the frequencies of methylation were 69%, CD44; 71%, RASSF1A; 56%, CASP8; 25%, PTEN; 15%, ZMYND10; 8%, CDH1; and 0%, PRDM2. Furthermore, CASP8 and CDH1 hypermethylation was signiﬁcantly associated with poor event-free survival. Meta-analysis of 115 neuroblastoma tumors demonstrated a signiﬁcant correlation between CASP8 methylation and MYCN ampliﬁcation. In addition, there was a correlation between ZMYND10 methylation and MYCN ampliﬁcation. The MSP data, together with optimized mRNA re-expression experiments (in terms of concentration and time of treatment and use of proper reference genes) further strengthen the notion that epigenetic alterations could play a signiﬁcant role in NB oncogenesis. This study thus warrants the need for a global proﬁling of gene promoter hypermethylation to identify genome-wide aberrantly methylated genes in order to further understand neuroblastoma pathogenesis and to identify prognostic methylation markers. Ó 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Neuroblastoma is a childhood tumor originating from sympathetic nervous system cells. The molecular basis of neuroblastoma development and progression is still poorly understood. The best-characterized genetic alterations include ampliﬁcation of the proto-oncogene MYCN, gain of chromosome arm 17q and losses of 1p, 3p, and 11q. Based on these genetic aberrations, neuroblastoma is currently classiﬁed into three major genetic subgroups * Corresponding author. Tel.: +32 9 3325187; fax: +32 9 3326549. E-mail address: [email protected]
(J. Vandesompele). 0304-3835/$ - see front matter Ó 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.canlet.2008.08.019
[1–3]. Many studies focused on the identiﬁcation of culprit tumor suppressor genes located within regions of recurrent loss in order to gain insight into the molecular defects governing neuroblastoma development. Thus far, no bona ﬁde tumor suppressor genes with a classic 2-hit inactivation have been identiﬁed. Other strategies including gene expression proﬁling also remained unsuccessful with respect to the identiﬁcation of genes and pathways implicated in neuroblastoma. As presently, epigenetic modiﬁcations, such as DNA methylation or chromatin modiﬁcations, are considered as an integral part of the process of malignant transformation and progression of cancer cells [4–7] further studies of the neuroblastoma
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
epigenome are needed. In human cancer, many genes implicated in pathways controlling growth, genomic stability, and survival were reported to be silenced by promoter hypermethylation. Of fundamental and possible also therapeutic interest is the fact that genes silenced by DNA methylation can be reactivated by treatment with demethylating agents, such as 5-aza-20 -deoxycytidine (DAC). The methyltransferase inhibitor DAC is a cytosine analogue and is incorporated into DNA during replication. Incorporation triggers covalent binding and inhibition of DNA methyltransferase (DNMT), leading to genome wide demethylation. In addition to CpG hypermethylation based gene silencing, compact and inactive chromatin characterized by the presence of hypoacetylated histones and methylation of speciﬁc lysines residues also mediates gene silencing. The level of acetylation of histones depends on the activity of histone deacetylase (HDAC) and histone acetyltransferase enzymes. The HDAC inhibitor trichostatin A can be used to reactivate genes that were silenced due to condensed chromatin. In order to further asses the role of epigenetic modiﬁcations in neuroblastoma, we selected 10 genes for analysis of promoter hypermethylation in primary neuroblastoma tumors and cell lines using methylation-speciﬁc PCR (MSP). In a second step, the effect of the methylation status on gene expression was investigated through measurement of the mRNA expression before and after treatment with a demethylating agent and a histone deacetylase inhibitor in 22 NB cell lines. To achieve this, we optimized the workﬂow of a re-expression experiment by determining optimal concentration and duration of treatment and by validating the reference genes for proper normalization of the gene expression data. 2. Materials and methods 2.1. Neuroblastoma tumors and cell lines Forty-two primary neuroblastoma tumor samples (at least 60% tumor cells) were collected prior to therapy at the Ghent University Hospital (Ghent, Belgium). The patients were randomly selected. Median age at diagnosis was 17.8 months and median follow-up time for survivors (n = 24) was 76.4 months. Twelve tumors were classiﬁed as stage 1, four as stage 2, six as stage 3, 18 as stage 4, and two as stage 4S according to the International Neuroblastoma Staging System . An overview of the clinical and biological parameters of the patients is given in Table 1. In addition, 33 well-characterized neuroblastoma cell lines were included in this study [2,9– 13]. DNA was isolated using the QIAamp DNA mini kit (Qiagen). 2.2. Methylation-speciﬁc PCR (MSP) MSP was performed according to Herman et al. , with minor modiﬁcations. Brieﬂy, 1 lg of genomic DNA was denaturated by NaOH in a volume of 20 ll (ﬁnal concentration 0.4 M). Five microliters of 30 mM hydroquinone and 156 ll of 5 M sodium bisulﬁte (pH 5.0) and 5 ll water were added and incubated at 50 °C for 16 h. DNA samples
were puriﬁed using Microcon-100 columns (Millipore) and eluted into 50 ll water. Modiﬁcation was completed by NaOH (ﬁnal concentration 0.3 M) treatment for 5 min at room temperature, followed by precipitation. Bisulﬁte modiﬁed DNA was resuspended in 30 ll water. MSP was performed using primers shown in Table 2. Newly developed primers were designed using the web-based MSP design software MethPrimer (http://www.urogene.org/ methprimer) , followed by in silico speciﬁcity assessment using our in-house-developed methBLAST software (http://medgen.ugent.be/methblast/, ). All primers are available in the public methPrimerDB database (http:// medgen.ugent.be/methprimerdb/, ) (see Table 2). PCR was carried out in a 50 ll reaction containing 50 ng bisulﬁte modiﬁed DNA, 1 Platinum Taq PCR buffer (Invitrogen), 6 mM MgCl2, 200 lM of each dNTP, 1.25 U Platinum Taq polymerase (Invitrogen), 300 nM of each primer. Three percent of DMSO was added to increase speciﬁcity for the MSP reactions of CD44, PTEN, ZMYND10, PRDM2, ROBO1 and TP73. The PCR cycling conditions consisted of an initial enzyme activation step at 93 °C for 4 min, followed by 35– 40 cycles of denaturation of 30 s at 93 °C, annealing of 30 s at annealing temperature shown in Table 2, extension for 30 s at 72 °C, followed by a ﬁnal extension step of 4 min at 72 °C. For PCR detection of RASSF1A (both methylated (M) and unmethylated (U) alleles) and for the methylated alleles of PTEN, a touchdown protocol was used. The PCR cycling conditions consisted of an initial enzyme activation step at 93 °C for 4 min, followed by six (RASSF1A U) or seven (RASSF1A M and PTEN M) touchdown cycles of 20 s at 93 °C, 40 s at temperature X °C shown in Table 2, extension for 20 s at 72 °C, X-1 °C per cycle and then 35 cycles of 20 s at 93 °C, 30 s at X-6 °C (RASSF1A U) or X-7 °C (RASSF1A M and PTEN M) and 30 s at 72 °C and ﬁnally, an extension step of 4 min at 72 °C. PCR products were loaded onto a 2% TBE agarose gel, stained with ethidium bromide, and visualized under UV illumination. SssI methylase (New England Biolabs, Beverly, MA, USA) treated DNA (M-DNA), following the manufacturer’s instructions and normal human genomic DNA were used as a positive and negative control for methylation after bisulﬁte modiﬁcation, respectively. Samples were scored as methylated when an ampliﬁcation product was clearly visible using the primers speciﬁc for the methylated allele. If neither the methylated nor the unmethylated product was present, results were excluded. 2.3. Statistical analysis of MSP data Univariate survival analysis was performed with the Kaplan–Meier method and log-rank statistic using SPSS 15.0 software to estimate overall (OS) and event-free survival (EFS). Event-free survival was deﬁned as the time between initial diagnosis and relapse or death of disease, or time between diagnosis and last follow-up if no event had occurred. The relationship between the methylation status and clinico-genetic parameters was determined using Fisher’s exact test. Power analysis was performed with the STPLAN software version 4.3 (Houston, TX: Department of Biostatistics, The University of Texas M.D. Anderson Cancer Center).
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
Table 1 Clinical characteristics of neuroblastoma patients and genetic alterations and stage of their tumor Sample
W91–239 T97–96 01T130 T99–84 T98–47 T94–64 T96–82 T98–33 T98–17 T99–9 T98–104 01T149 T97–26 01T198 T94–40 01T146 W90–160 T99–119 T00–121 01T96 T98–24 T96–21 T00–35 T96–81 T95–60 T01–25 01T28 01T143 01T94 T95–24 W91–5 T01–31 SL–32 T95–4 W92–145 T99–13 W93–64 T97–10 SL–28 T94–25 01T208 01T15
1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4S 4S
No No No No No No No No No No No No No No No No Loss Loss No No No No Loss No No Loss Loss No No Loss Loss No Loss No – No Loss Loss – No Loss No
No No No No No No No No No No No Loss No Loss Loss No No No No Loss No No No Loss No No No No No Loss No No No No – No No No – No No No
No No No No No No No No – – – Loss No No – No No No No No No No No Loss No No No No No Loss No No Loss No – No No No – No No No
No No No No No No No No No No No No No No No No Amp Amp No No No No Amp No No Amp Amp No Amp Amp No No No No No No amp No – No No No
6.6 18.4 1.6 1.9 65.0 23.2 10.0 3.0 46.6 5.7 0.9 8.0 14.1 0.1 10.4 15.5 20.1 22.4 7.5 43.8 17.3 7.5 11.7 27.67 49.9 29.5 25.4 1.3 26.4 57.5 113.9 31.5 62.9 56.7 100.7 54.6 34.0 16.9 39.6 3.3 8.5 3.7
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 DOT 0 1 1 1 1 1 0 0 1 1 0 DOT 1 1 0 1 1 1 1 0 0
159.4 104.2 31.4 70.7 103.2 133.8 113.9 97.0 78.8 84.6 74.1 52.2 112.3 53.1 159.4 61.3 21.0 17.8 69.0 64.0 – 107.6 13.6 31.8 18.7 5.5 9.9 65.8 66.1 24.8 27.0 11.7 – 21.5 19.2 87.2 28.0 9.4 23.6 71.2 51.5 72.6
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 1 0 1 1 1 1 1 0 0 1 1 1 – 1 1 0 1 1 1 1 0 0
23.2 104.2 31.4 70.7 103.2 133.8 113.9 97.0 78.8 84.6 74.1 52.2 112.3 53.1 159.4 61.3 14.4 15.1 34.2 64.0 6.4 107.6 9.4 29.3 13.1 – – 65.8 66.1 9.2 22.8 8.9 – 18.2 – 87.2 25.7 – 19.0 51.6 51.5 72.6
a b c d e f g
No, no loss, –, not tested or not informative. Amp, ampliﬁcation; no, no ampliﬁcation; –, not tested. age at diagnosis (in months). 0, alive; 1, dead; DOT, death of toxicity. Overall survival (in months from the date of diagnosis till disease-related death or last follow-up). 1, tumor progression, relapse or death; 0, no event; –, not available. Event-free survival (months) (see Section 2); –, not available.
2.4. 5-Aza-20 -deoxycytine and trichostatin A drug treatment and RNA isolation Neuroblastoma cell lines were grown in RPMI 1640 growth medium (Invitrogen) supplemented with 10% FCS at 37 °C and 5% CO2. To select the optimal treatment conditions for demethylation, three neuroblastoma cell lines (NGP, SK-N-AS, and IMR-32) were grown in the presence of different concentrations of 5-aza-20 -deoxycytine (DAC; 0, 1, 3, and 10 lM) during different durations of treatment (3, 5, and 7 days). The medium was changed every 2 days, along with supplementing fresh DAC. For the ﬁnal study to assess reactivation of gene expression, 22 NB cell lines were plated at day 0 and treated 24 h later with either 3 lM DAC (Sigma) for 3 days, or with 3 lM DAC for 3 days
and by 500 nM trichostatin A (TSA) for the last 12 h. Medium was changed on day 2 with addition of new drugs. In parallel, untreated controls were also prepared. Cells (treated and untreated) were harvested and RNA was extracted for real-time quantitative PCR. RNA extraction was performed with the RNeasy Mini kit (Qiagen) according to the manufacturer, accompanied by RNase free DNase treatment on column (Qiagen). 2.5. Real-time quantitative PCR (qPCR) Following an additional RNAse free DNase treatment, cDNA was synthesized using the iScript cDNA synthesis kit from Bio-Rad (Hercules, CA, USA) (http://medgen.ugent.be/CMGG/protocols). mRNA expression was
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346 Table 2 Primer sequences and PCR characteristics for methylation analysis Gene
Forward primer (50 –30 )
Reverse primer (50 –30 )
methPrimerDB ID (http://medgen.ugent.be/methprimerdb/), U, unmethylated; M, methylated; Ta, annealing temperature.
examined by an optimized two-step real-time quantitative PCR assay . The primers were designed with PrimerExpress 2.0 software (Applied Biosystems, Foster City, CA, USA) and are available in the public RTPrimerDB database [18–19] (http://medgen.ugent.be/rtprimerdb/) (gene (RTPrimerDB-ID): HPRT1 (5), YWHAZ (7), GAPDH (3), and HMBS (4), SDHA (7), CASP8 (86), PTEN (363), ZMYND10 (3490), DCC (3491), RASSF1A (3489)). Reactions were run on an iCycler iQ (Bio-Rad). The results were imported into the relative quantiﬁcation software qBase (http://medgen.ugent.be/qbase/)  for further analysis. The transcription levels were normalized using the geometric mean of four stably expressed reference genes (HPRT1, YWHAZ, GAPDH, and HMBS) . A stably expressed reference gene is a control gene that is not variable in expression in the type of cells under investigation and in response to the experimental treatment, as determined by the geNorm algorithm . 3. Results and discussion 3.1. Methylation status in neuroblastoma We selected 10 genes (PRDM2, TP73, CASP8, ZMYND10, RASSF1A, ROBO1, PTEN, CD44, CDH1, and DCC) that previously have been described for undergoing methylationassociated silencing in cancer or that were of interest in neuroblastoma based on their chromosomal location. Five of these genes are located in recurrent regions of loss in neuroblastoma, i.e., PRDM2 and TP73 on 1p, RASSF1A and ZMYND10 on 3p, and DCC on 18q. At the time of study design, silencing through promoter hypermethylation of four genes (i.e., RASSF1A, CASP8, CDH1, and TP73) had been previously investigated in neuroblastoma [21–23]. These
genes were included in order to validate the reported frequencies of methylation in our patient and cell line cohort. Meanwhile, additional studies investigating the methylation status of PRDM2, ZMYND10, PTEN, and CD44 in NB have been published [24–27]. To our knowledge, this report is the ﬁrst study analyzing ROBO1 and DCC methylation in NB. The ROBO1 gene is located on 3p12 and encodes an axon guidance receptor, a member of the NCAM cell adhesion family receptors. Aberrant methylation has been described in breast, cervical and colorectal cancer [28–30]. The DCC gene (18q21) is involved in neuronal development and absent or reduced mRNA and protein expression has been reported in both neuroblastoma cell lines and primary tumors. However, mutation detection frequencies of DCC are low in NB suggesting that other mechanisms for inactivation in addition to deletion may play a role. In other tumor types such as gastric  and breast  cancer, oral squamous cell carcinoma  and acute lymphoblastic leukemia , abnormal DCC methylation has been described. We used MSP to establish the frequency of methylation for these 10 genes in neuroblastoma tumors and cell lines. An overview of the results is given in Fig. 1 and Table 3. Overall, higher methylation frequencies were observed in cell lines compared to primary tumors, in keeping with other reports [35–37]. This could be explained by the fact that neuroblastoma cell lines are mainly derived from more aggressive tumors and this could reﬂect in a higher number of methylated genes. However, we cannot exclude additional in vitro effects. Nevertheless, genes that show high methylation frequencies in neuroblastoma cell lines are also detected to be methylated in primary tumors. All but ﬁve neuroblastoma tumors (88%) presented with methylation in at least one of the investigated regions. Se-
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
Fig. 1. Methylation status of 10 genes in a panel of 33 neuroblastoma cell lines. Abbreviations are as follows: (a) amp, ampliﬁcation; no, no ampliﬁcation; (b) no, no loss; –, not tested.
Table 3 Summary of methylation-speciﬁc PCR results in neuroblastoma tumors Gene
ROBO1 PRMD2 TP73 DCC CDH1 ZMYND10 PTEN CASP8 RASSF1A CD44
41 41 42 41 40 40 20 36 41 32
41 41 42 38 37 34 15 16 12 10
0 0 0 3 3 6 5 20 29 22
0 0 0 7 8 15 25 56 71 69
U, unmethylated, M, methylated; % Me, percentage of methylated samples for each gene.
ven of 10 genes showed methylation in at least one of the investigated tumor samples (DCC, CDH1, ZMYND10, PTEN, CASP8, RASSF1A, and CD44), while no methylation was found for the remaining three genes. In tumors, a high frequency of methylation was found for CD44 (69%), CASP8 (56%), and RASSF1A (71%). Intermediate to low methylation frequencies were found for PTEN (25%), ZMYND10 (15%), DCC (7%), and CDH1 (8%). While PRDM2 and ROBO1 were methylated in 30% and 3% of cell lines, respectively, no methylation could be detected in primary tumor samples. TP73 showed no methylation in neuroblastoma tumors or cell lines. Representative examples of the MSP analysis are shown in Fig. 2. For many tumor samples, we detect
both unmethylated and methylated alleles in the same investigated locus. These mixed results can be explained either by heterogeneity of methylation and by contamination of normal cells. In general, these methylation frequencies match closely with previously reported ﬁndings. Similar methylation frequencies in both cell lines and tumors were found for CDH1 and RASSF1A [22,38–43]. In contrast, large differences were found for CD44 (tumors: 69% vs. 0%; cell lines: 90% vs. 33%) and for PTEN (tumors: 25% vs. 0%; cell lines: 88% vs. 0%) [26–27,44]. Lower methylation frequencies were observed in tumors for ZMYND10 (15% vs. 34–54%) and PRDM2 (0% vs. 25%) while comparable results were described in cell lines [24–25,40–41,44]. For CASP8, lower frequencies (67–92% vs. 97%) were reported in general in neuroblastoma cell lines [23,25,38,40–41,43–46]. For some genes, signiﬁcant lower or higher methylation frequencies were reported in cell lines or tumors. This can possibly be explained by use of different primer pairs and targeted promoter regions that were investigated (e.g., PTEN, Supplementary Figure 1 created using methGraph (http:// medgen.ugent.be/methgraph/; Lefever et al., in preparation), or by use of a different methylation analysis method. 3.2. Methylation status and clinico-genetic parameters Univariate Kaplan–Meier analysis with log rank testing demonstrated that MYCN ampliﬁcation, 1p deletion, age at diagnosis >12 months, and high-stage tumor (stages 3
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
Fig. 2. Representative results of the MSP analysis of genes CASP8, PRDM2, and PTEN. Abbreviations are as follows: U, unmethylated allele; M, methylated allele; 50 bp, 50bp DNA ladder; HgDNA, human genomic DNA; M-DNA, SssI treated positive control for methylation; NTC, no template control; T, tumor sample; C, cell line.
and 4) correlate signiﬁcantly (P < 0.01) with poor survival. Conﬁrmation of predictive power of those known markers demonstrates that our patient cohort is a representative sample of the neuroblastoma patient population. Next, we examined the relationship between the methylation status of genes and clinico-genetic parameters, including stage, age at diagnosis >12 months, overall and event-free sur-
vival time, survival status, 1p, 3p and 11q loss and MYCN ampliﬁcation (Table 4). Kaplan–Meier survival analysis showed that hypermethylation of CASP8 and CDH1 was correlated with poor event-free survival (log rank P = 0.016 and P = 0.038, respectively). The methylation status of the other genes was not correlated with overall or event-free survival. Furthermore, there appears to be an association
Table 4 Cross table: relationship between the methylation status of genes and clinico-genetic parameters RASSF1A
Tumor stage 1, 2, 4S 3, 4
Age at diagnosis <1 year >1 year
Status Alive Dead
Event Yes No
MYCN gene Single copy Ampliﬁcation
Chromosome 1p Normal Deletion
Chromosome 3p Normal Deletion
Chromosome 11q Normal 10 Deletion 1
a b c
Unmethylated. Methylated. P-value, Fisher’s exact test.
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
between CASP8 methylation and the occurrence of an event (P = 0.037). In order to end the long lasting controversy concerning the correlation between CASP8 promoter methylation and MYCN ampliﬁcation, we performed a metaanalysis of three studies in which the same primer pairs were used to study the methylation status of CASP8 (this study, [23,45]). The individual studies could demonstrate only borderline signiﬁcance or no signiﬁcance for this correlation. Our meta-analysis, including 115 neuroblastoma tumors, demonstrated that CASP8 methylation and MYCN ampliﬁcation are signiﬁcantly correlated (Fisher’s exact test: P = 0.0062), whereby a higher frequency of CASP8 methylation in MYCN ampliﬁed versus non-ampliﬁed samples is observed (66% vs. 36%). Our results also indicate that there is a trend towards an association between ZMYND10 methylation and MYCN ampliﬁcation (Fisher’s exact test: P = 0.063). For the other genes, no signiﬁcant correlation between methylation status and the clinico-genetic parameter could be demonstrated. As the power of our study is limited, we cannot exclude weak associations. 3.3. Re-expression of genes after DAC and TSA treatment In order to ascertain whether promoter hypermethylation of the tested genes resulted in silencing, we treated neuroblastoma cell lines with a demethylating agent (5-aza-20 -deoxycytidine, DAC) and a histon-deacetylase inhibitor (trichostatin A, TSA). A pilot study was designed to determine the optimal treatment conditions with
Fig. 3. A representative example of relative mRNA expression levels of the stably expressed reference gene (HMBS) and a methylated gene (RASSF1A) in neuroblastoma cell line SK-N-AS treated with DAC at different concentrations (0, 1, 3, and 10 lM) and durations (3, 5, and 7 days) (untreated cells day 3 are rescaled to 1). A clear time and dosedependent response is visible for RASSF1A reactivation. Error bars denote standard error of the mean (duplicated PCRs, gene of interest divided by geometric mean of four reference genes).
Fig. 4. Relative mRNA expression levels of candidate reference genes GAPDH, HMBS, HPRT1, SDHA, and YWHAZ in four representative neuroblastoma cell lines (before (0) and after treatment with DAC alone (3 lM, 3 days) or combined DAC (3 lM, 3 days) and TSA (500 nM, last 12 h)) demonstrates that expression of SDHA is consistently inﬂuenced by treatment, while for GAPDH, HMBS, HPRT1, and YWHAZ normal variation in mRNA expression is observed (untreated cells are rescaled to 1).
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
respect to duration and dose of the treatment. Based on the mRNA re-expression analysis of RASSF1A and CASP8, two known methylated genes in neuroblastoma, we decided to treat the neuroblastoma cells with 3 lM DAC for 3 days. At these conditions, re-expression of methylated genes is sufﬁciently high. Higher inductions were observed for the 10 lM treatments for one week, but these conditions are likely to induce a more prominent stress response (Fig. 3). In order to accurately normalize the mRNA expression levels, we ﬁrst evaluated ﬁve candidate reference genes (i.e., HPRT1, YWHAZ, HMBS, SDHA, and GAPDH) that were previously shown to be stably expressed in untreated neu-
roblastoma cells . To this purpose, we determined their relative expression in 22 different neuroblastoma cells with or without treatment and analyzed the data using the geNorm applet for Microsoft Excel . Based on a robust gene stability measurement algorithm, the geNorm program determines the most stably expressed reference genes from a set of tested candidate reference genes in a given sample panel. Following this approach, we determined that four reference genes are required for normalization in DAC and TSA treated neuroblastoma cells and that the reference gene SDHA was highly induced in the treated cells (Fig. 4). In this study, we again demonstrate
Fig. 5. Representative examples of mRNA expression analysis for CASP8, PTEN, and ZMYND10 in cell lines untreated (0) or treated with DAC alone (DAC) or in combination with TSA (DAC+TSA) (lowest expression of the represented cell lines is rescaled to 1). U, unmethylated; M, methylated.
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
the necessity of carefully determining which and how many reference genes are required for accurate gene expression proﬁling. After having established suitable reference genes, we performed qPCR-based expression analysis for ﬁve randomly selected genes of the 10 genes that were tested with MSP. Reactivation or upregulation of CASP8 expression was found in all investigated cell lines that were methylated at the intragenic regulatory region between exons 2 and 3, except for NGP and SK-N-AS (Fig. 5). Interestingly, Banelli et al. reported that silencing of CASP8 does not depend directly from hypermethylation of this region . Rather, it likely reﬂects the epigenetic inactivation of genes that transactivate CASP8. Re-expression of RASSF1A after treatment with DAC was detected in most but not all neuroblastoma cell lines. Probably, a more intense treatment (higher dose, longer period) is needed to reactivate expression in those cell lines for which the current treatment did not result in re-expression. Presently, we could not conﬁrm an association between CpG island methylation and loss of gene expression for DCC and ZMYND10. In contrast to DAC treatment alone, gene expression of ZMYND10 was upregulated when cells were treated with both DAC and TSA (Fig. 5) indicating that histone modiﬁcations are also involved. Finally, expression of the PTEN gene was not inﬂuenced by treatment suggesting that the observed hypermethylation might not be linked to the mRNA expression status. These data show that complementary experiments are needed when investigating the silencing of a putative candidate tumor suppressor gene through methylation. In addition to the screening of the promoter region for abnormal methylation, investigating its inﬂuence on gene expression is also required. However, interpreting these mRNA expression data is not always straightforward. Some genes are completely silenced and show no expression in the cancer cell, while other genes present with only reduced expression compared to unmethylated cells. In addition, small experimental and technical variation in gene expression measurement may occur. For these reasons and because of the considerable sensitivity of an optimal real-time quantitative PCR, we considered a gene to be upregulated after treatment when a more than 2.5-fold increase in mRNA expression was observed. In the present study, a NB cell line panel was treated in which we can ﬁrst evaluate mRNA expression before and after treatment with DAC alone or in combination with TSA. In a subsequent step, bisulﬁte sequencing of the promoter region of re-activated genes should be performed to correlate silencing of gene-expression with hypermethylation of speciﬁc CpG dinucleotides in the CpG island. Based on these results, we will be able to determine the region of interest for further analysis using MSP in a cohort of NB tumor samples. In future research, this methylome analysis will aid in the identiﬁcation of genes which contribute to neuroblastoma pathogenesis through methylation and gene silencing. 3.4. Conclusion and perspectives In this study, we report on the gene speciﬁc methylation analysis in neuroblastoma of positional and functional
candidate genes for which promoter hypermethylation is known in other tumor types. Our MSP data together with the re-expression experiments demonstrated that important tumor suppressor genes are aberrantly methylated suggesting a role for DNA methylation in neuroblastoma pathogenesis. Nowadays, evidence is emerging that subclass discovery through proﬁling of gene methylation offers opportunities for diagnostic and prognostic stratiﬁcation of cancer patients (The Cancer Epigenome Project, [48– 49]). Integration of methylation markers in risk stratiﬁcation of cancer patients seems a powerful approach to improve therapeutic decision making. Most likely, this strategy will result in better survival rates and avoid unnecessary overtreatment of certain patients. For this reason and encouraged by the ﬁndings of the present study, we have embarked on genome-wide proﬁling of methylated genes in NB using the here described platform for validation. 4. Conﬂicts of interest statement The authors declare no competing ﬁnancial or other interests with regard to the submitted manuscript. Acknowledgements We thank Peter Degrave and Geert De Vos for cell culturing. Jasmien Hoebeeck is supported by the Vlaamse Liga tegen Kanker by a grant of the Stichting Emmanuel van der Schueren and by a grant of the Ghent University (BOF 01P07406). Filip Pattyn is supported by the Vlaamse Liga tegen Kanker by a grant of the Stichting Emmanuel van der Schueren. Joëlle Vermeulen is supported by the Belgian Kid’s Fund. Jo Vandesompele is a postdoctoral researcher with the Foundation for Scientiﬁc Research, Flanders (FWO). This study was supported by the Kinderkankerfonds, the Fund for Scientiﬁc Research Flanders (Krediet aan Navorsers 1.5.243.05), the ‘‘Stichting tegen Kanker” Project No. 365B0107, GOA-Grant 12051203, FWO-Grant G.0185.04. This text presents research results of the Belgian program of Interuniversity Poles of attraction initiated by the Belgian State, Prime Minister’s Ofﬁce, Science Policy Programming (IUAP) and the European 6th framework programme EET-pipeline. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.canlet.2008. 08.019. References  G.M. Brodeur, Neuroblastoma: biological insights into a clinical enigma, Nat. Rev. Cancer 3 (2003) 203–216.  E. Michels, J. Vandesompele, K. De Preter, J. Hoebeeck, J. Vermeulen, A. Schramm, J.J. Molenaar, B. Menten, B. Marques, R.L. Stallings, V. Combaret, C. Devalck, A. De Paepe, R. Versteeg, A. Eggert, G. Laureys, N. Van Roy, F. Speleman, ArrayCGH-based classiﬁcation of neuroblastoma into genomic subgroups, Genes Chromosomes Cancer 46 (2007) 1098–1108.
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346  J. Vandesompele, M. Baudis, K. De Preter, N. Van Roy, P. Ambros, N. Bown, C. Brinkschmidt, H. Christiansen, V. Combaret, M. Lastowska, J. Nicholson, A. O’Meara, D. Plantaz, R. Stallings, B. Brichard, C. Van den Broecke, S. De Bie, A. De Paepe, G. Laureys, F. Speleman, Unequivocal delineation of clinicogenetic subgroups and development of a new model for improved outcome prediction in neuroblastoma, J. Clin. Oncol. 23 (2005) 2280–2299.  S.B. Baylin, J.G. Herman, J.R. Graff, P.M. Vertino, J.P. Issa, Alterations in DNA methylation: a fundamental aspect of neoplasia, Adv. Cancer Res. 72 (1998) 141–196.  M. Esteller, Relevance of DNA methylation in the management of cancer, Lancet Oncol. 4 (2003) 351–358.  M. Esteller, Cancer epigenomics: DNA methylomes and histonemodiﬁcation maps, Nat. Rev. 8 (2007) 286–298.  P.A. Jones, P.W. Laird, Cancer epigenetics comes of age, Nat. Genet. 21 (1999) 163–167.  G.M. Brodeur, J. Pritchard, F. Berthold, N.L. Carlsen, V. Castel, R.P. Castelberry, B. De Bernardi, A.E. Evans, M. Favrot, F. Hedborg, et al, Revisions of the international criteria for neuroblastoma diagnosis staging and response to treatment, J. Clin. Oncol. 11 (1993) 1466–1477.  J. Hoebeeck, E. Michels, B. Menten, N. Van Roy, A. Eggert, A. Schramm, K. De Preter, N. Yigit, E. De Smet, A. De Paepe, G. Laureys, J. Vandesompele, F. Speleman, High resolution tiling-path BAC array deletion mapping suggests commonly involved 3p21–p22 tumor suppressor genes in neuroblastoma and more frequent tumors, Int. J. Cancer 120 (2007) 533–538.  M. Van Gele, N. Van Roy, A. Jauch, G. Laureys, Y. Benoit, V. Schelfhout, C.R. De Potter, P. Brock, A. Uyttebroeck, R. Sciot, E. Schuuring, R. Versteeg, F. Speleman, Sensitive and reliable detection of genomic imbalances in human neuroblastomas using comparative genomic hybridisation analysis, Eur. J. Cancer 33 (1997) 1979–1982.  N. Van Roy, A. Jauch, M. Van Gele, G. Laureys, R. Versteeg, A. De Paepe, T. Cremer, F. Speleman, Comparative genomic hybridization analysis of human neuroblastomas: detection of distal 1p deletions and further molecular genetic characterization of neuroblastoma cell lines, Cancer Genet. Cytogenet. 97 (1997) 135–142.  N. Van Roy, H. Van Limbergen, J. Vandesompele, M. Van Gele, B. Poppe, H. Salwen, G. Laureys, N. Manoel, A. De Paepe, F. Speleman, Combined M-FISH and CGH analysis allows comprehensive description of genetic alterations in neuroblastoma cell lines, Genes Chromosomes Cancer 32 (2001) 126–135.  J. Vandesompele, E. Michels, K. De Preter, B. Menten, A. Schramm, A. Eggert, P.F. Ambros, V. Combaret, N. Francotte, F. Antonacci, A. De Paepe, G. Laureys, F. Speleman, N. Van Roy, Identiﬁcation of 2 putative critical segments of 17q gain in neuroblastoma through integrative genomics, Int. J. Cancer 122 (2008) 1177–1182.  J.G. Herman, J.R. Graff, S. Myohanen, B.D. Nelkin, S.B. Baylin, Methylation-speciﬁc PCR: a novel PCR assay for methylation status of CpG islands, Proc. Natl. Acad. Sci. USA 93 (1996) 9821–9826.  L.C. Li, R. Dahiya, MethPrimer: designing primers for methylation PCRs, Bioinformatics 18 (2002) 1427–1431.  F. Pattyn, J. Hoebeeck, P. Robbrecht, E. Michels, A. De Paepe, G. Bottu, D. Coornaert, R. Herzog, F. Speleman, J. Vandesompele, methBLAST and methPrimerDB: web-tools for PCR based methylation analysis, BMC Bioinformatics 7 (2006) 496.  J. Vandesompele, K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A. De Paepe, F. Speleman, Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes, Genome Biol. 3 (2002). RESEARCH0034.  F. Pattyn, P. Robbrecht, A. De Paepe, F. Speleman, J. Vandesompele, RTPrimerDB: the real-time PCR primer and probe database major update 2006, Nucleic Acids Res. 34 (2006) D684–D688.  F. Pattyn, F. Speleman, A. De Paepe, J. Vandesompele, RTPrimerDB: the real-time PCR primer and probe database, Nucleic Acids Res. 31 (2003) 122–123.  J. Hellemans, G. Mortier, A. De Paepe, F. Speleman, J. Vandesompele, qBase relative quantiﬁcation framework and software for management and automated analysis of real-time quantitative PCR data 8 (2007) R19.  B. Banelli, I. Casciano, M. Romani, Methylation-independent silencing of the p73 gene in neuroblastoma, Oncogene 19 (2000) 4553–4556.  K. Harada, S. Toyooka, A. Maitra, R. Maruyama, K.O. Toyooka, C.F. Timmons, G.E. Tomlinson, D. Mastrangelo, R.J. Hay, J.D. Minna, A.F. Gazdar, Aberrant promoter methylation and silencing of the RASSF1A gene in pediatric tumors and cell lines, Oncogene 21 (2002) 4345–4349.
 T. Teitz, T. Wei, M.B. Valentine, E.F. Vanin, J. Grenet, V.A. Valentine, F.G. Behm, A.T. Look, J.M. Lahti, V.J. Kidd, Caspase 8 is deleted or silenced preferentially in childhood neuroblastomas with ampliﬁcation of MYCN, Nat. Med. 6 (2000) 529–535.  A. Agathanggelou, A. Dallol, S. Zochbauer-Muller, C. Morrissey, S. Honorio, L. Hesson, T. Martinsson, K.M. Fong, M.J. Kuo, P.W. Yuen, E.R. Maher, J.D. Minna, F. Latif, Epigenetic inactivation of the candidate 3p21 suppressor gene BLU in human cancers, Oncogene 22 (2003) 1580–1588.  M. Alaminos, V. Davalos, N.K. Cheung, W.L. Gerald, M. Esteller, Clustering of gene hypermethylation associated with clinical risk groups in neuroblastoma, J. Natl. Cancer Inst. 96 (2004) 1208– 1219.  P. Lazcoz, J. Munoz, M. Nistal, A. Pestana, I.J. Encio, J.S. Castresana, Loss of heterozygosity and microsatellite instability on chromosome arm 10q in neuroblastoma, Cancer Genet. Cytogenet. 174 (2007) 1–8.  P. Yan, A. Muhlethaler, K.B. Bourloud, M.N. Beck, N. Gross, Hypermethylation-mediated regulation of CD44 gene expression in human neuroblastoma, Genes Chromosomes Cancer 36 (2003) 129–138.  A. Dallol, E. Forgacs, A. Martinez, Y. Sekido, R. Walker, T. Kishida, P. Rabbitts, E.R. Maher, J.D. Minna, F. Latif, Tumour speciﬁc promoter region methylation of the human homologue of the Drosophila Roundabout gene DUTT1 (ROBO1) in human cancers, Oncogene 21 (2002) 3020–3028.  A. Dallol, D. Morton, E.R. Maher, F. Latif, SLIT2 axon guidance molecule is frequently inactivated in colorectal cancer and suppresses growth of colorectal carcinoma cells, Cancer Res 63 (2003) 1054–1058.  G. Narayan, C. Goparaju, H. Arias-Pulido, A.M. Kaufmann, A. Schneider, M. Durst, M. Mansukhani, B. Pothuri, V.V. Murty, Promoter hypermethylation-mediated inactivation of multiple SlitRobo pathway genes in cervical cancer progression, Mol. Cancer 5 (2006) 16.  K. Sato, G. Tamura, T. Tsuchiya, Y. Endoh, O. Usuba, W. Kimura, T. Motoyama, Frequent loss of expression without sequence mutations of the DCC gene in primary gastric cancer, Br. J. Cancer 85 (2001) 199–203.  K. Miyamoto, T. Fukutomi, S. Akashi-Tanaka, T. Hasegawa, T. Asahara, T. Sugimura, T. Ushijima, Identiﬁcation of 20 genes aberrantly methylated in human breast cancers, Int. J. Cancer 116 (2005) 407–414.  K. Ogi, M. Toyota, M. Ohe-Toyota, N. Tanaka, M. Noguchi, T. Sonoda, G. Kohama, T. Tokino, Aberrant methylation of multiple genes and clinicopathological features in oral squamous cell carcinoma, Clin. Cancer Res. 8 (2002) 3164–3171.  K.H. Taylor, K.E. Pena-Hernandez, J.W. Davis, G.L. Arthur, D.J. Duff, H. Shi, F.B. Rahmatpanah, O. Sjahputera, C.W. Caldwell, Large-scale CpG methylation analysis identiﬁes novel candidate genes and reveals methylation hotspots in acute lymphoblastic leukemia, Cancer Res. 67 (2007) 2617–2625.  M.F. Paz, M.F. Fraga, S. Avila, M. Guo, M. Pollan, J.G. Herman, M. Esteller, A systematic proﬁle of DNA methylation in human cancer cell lines, Cancer Res. 63 (2003) 1114–1121.  D.J. Smiraglia, L.J. Rush, M.C. Fruhwald, Z. Dai, W.A. Held, J.F. Costello, J.C. Lang, C. Eng, B. Li, F.A. Wright, M.A. Caligiuri, C. Plass, Excessive CpG island hypermethylation in cancer cell lines versus primary human malignancies, Hum. Mol. Genet. 10 (2001) 1413–1419.  T. Ueki, K.M. Walter, H. Skinner, E. Jaffee, R.H. Hruban, M. Goggins, Aberrant CpG island methylation in cancer cell lines arises in the primary cancers from which they were derived, Oncogene 21 (2002) 2114–2117.  D. Astuti, A. Agathanggelou, S. Honorio, A. Dallol, T. Martinsson, P. Kogner, C. Cummins, H.P. Neumann, R. Voutilainen, P. Dahia, C. Eng, E.R. Maher, F. Latif, RASSF1A promoter region CpG island hypermethylation in phaeochromocytomas and neuroblastoma tumours, Oncogene 20 (2001) 7573–7577.  B. Banelli, I. Gelvi, A. Di Vinci, P. Scarufﬁ, I. Casciano, G. Allemanni, S. Bonassi, G.P. Tonini, M. Romani, Distinct CpG methylation proﬁles characterize different clinical groups of neuroblastic tumors, Oncogene 24 (2005) 5619–5628.  P. Lazcoz, J. Munoz, M. Nistal, A. Pestana, I. Encio, J.S. Castresana, Frequent promoter hypermethylation of RASSF1A and CASP8 in neuroblastoma, BMC Cancer 6 (2006) 254.  M.B. Michalowski, F. de Fraipont, D. Plantaz, S. Michelland, V. Combaret, M.C. Favrot, Methylation of tumor-suppressor genes in neuroblastoma: the RASSF1A gene is almost always methylated in primary tumors, Pediatr. Blood Cancer 50 (2008) 29–32.
J. Hoebeeck et al. / Cancer Letters 273 (2009) 336–346
 I.H. Wong, J. Chan, J. Wong, P.K. Tam, Ubiquitous aberrant RASSF1A promoter methylation in childhood neoplasia, Clin. Cancer Res. 10 (2004) 994–1002.  Q. Yang, P. Zage, D. Kagan, Y. Tian, R. Seshadri, H.R. Salwen, S. Liu, A. Chlenski, S.L. Cohn, Association of epigenetic inactivation of RASSF1A with poor outcome in human neuroblastoma, Clin. Cancer Res. 10 (2004) 8493–8500.  Q. Yang, C.M. Kiernan, Y. Tian, H.R. Salwen, A. Chlenski, B.A. Brumback, W.B. London, S.L. Cohn, Methylation of CASP8, DCR2 and HIN-1 in neuroblastoma is associated with poor outcome, Clin. Cancer Res. 13 (2007) 3191–3197.  P. Gonzalez-Gomez, M.J. Bello, J. Lomas, D. Arjona, M.E. Alonso, C. Aminoso, I. Lopez-Marin, N.P. Anselmo, J.L. Sarasa, M. Gutierrez, C. Casartelli, J.A. Rey, Aberrant methylation of multiple genes in neuroblastic tumours relationship with MYCN ampliﬁcation and allelic status at 1p, Eur. J. Cancer 39 (2003) 1478–1485.  J. Takita, H.W. Yang, Y.Y. Chen, R. Hanada, K. Yamamoto, T. Teitz, V. Kidd, Y. Hayashi, Allelic imbalance on chromosome 2q and alterations of the caspase 8 gene in neuroblastoma, Oncogene 20 (2001) 4424–4432.  B. Banelli, I. Casciano, M. Croce, A. Di Vinci, I. Gelvi, G. Pagnan, C. Brignole, G. Allemanni, S. Ferrini, M. Ponzoni, M. Romani, Expression
and methylation of CASP8 in neuroblastoma: identiﬁcation of a promoter region, Nat. Med. 8 (2002) 1333–1335. author reply 1335. K. Garber, Momentum building for human epigenome project, J. Natl. Cancer Inst. 98 (2006) 84–86. P.A. Jones, R. Martienssen, A blueprint for a Human Epigenome Project: the AACR Human Epigenome Workshop, Cancer Res. 65 (2005) 11241–11246. L.L. Siu, J.K. Chan, K.F. Wong, Y.L. Kwong, Speciﬁc patterns of gene methylation in natural killer cell lymphomas: p73 is consistently involved, Am. J. Pathol. 160 (2002) 59–66. Y. Du, T. Carling, W. Fang, Z. Piao, J.C. Sheu, S. Huang, Hypermethylation in human cancers of the RIZ1 tumor suppressor gene a member of a histone/protein methyltransferase superfamily, Cancer Res. 61 (2001) 8094–8099. J.C. Soria, H.Y. Lee, J.I. Lee, L. Wang, J.P. Issa, B.L. Kemp, D.D. Liu, J.M. Kurie, L. Mao, F.R. Khuri, Lack of PTEN expression in non-small cell lung cancer could be related to promoter methylation, Clin. Cancer Res. 8 (2002) 1178–1184. J.R. Graff, J.G. Herman, S. Myohanen, S.B. Baylin, P.M. Vertino, Mapping patterns of CpG island methylation in normal and neoplastic cells implicates both upstream and downstream regions in de novo methylation, J. Biol. Chem. 272 (1997) 22322–22329.