MYCN is a major driver of neuroblastoma tumorigenesis and MYCN amplification is the worst prognostic indicator of aggressive NB. To identify potentially therapeutic tumor suppressor microRNAs for aggressive NB, we utilized a conditional MYCN system to simulate MYCN-amplified and nonamplified tumor types and performed a genome-wide search for MYCN target microRNA promoters differentially repressed under high MYCN conditions. We identified 20 gene promoters hosting 30 microRNAs that were directly bound and differentially regulated by MYCN. Eleven of these genes showed significant clinical correlations for neuroblastoma with 4 genes linked with better survival and 7 genes linked with poor survival. Surprisingly, expression analysis of host genes and microRNAs demonstrated that 8 of 11 pairs were repressed by high levels of MYCN regardless of the clinical correlation of the host gene. We therefore predicted these intronic microRNAs would be tumor suppressors. In fact, detailed gain of function studies for two miRs, miR-591 and miR-558, confirmed potent tumor suppressive effects for miR-591 in orthotopic neuroblastoma xenografts. However, miR-558 markedly increased colony formation, proliferation, and tumor growth in vivo. Our data reveal host-gene independent functions of MYCN-target microRNAs and demonstrate that MYCN represses both tumor suppressive and proproliferative microRNAs. Cancer Res; 71(11); 3841–51. ©2011 AACR.

Neuroblastoma remains a major therapeutic challenge in pediatric oncology, accounting for almost 15% of all childhood cancer deaths (1). To identify potential new therapeutic strategies for this aggressive embryonal malignancy of young children, we investigated the MYCN-dependent regulation and function of microRNAs (miRNAs) in neuroblastoma cell lines and xenograft models. miRNAs are essential regulators of embryogenesis and tissue differentiation, modulate stem cell biology, and act as oncogenes and tumor suppressors (2). These are the same cellular processes that are often perturbed by oncogenes in embryonal malignancies of childhood such as neuroblastoma (3, 4). Furthermore, multiple studies now show that miRNAs collaborate with oncogenic transcription factors to alter cell proliferation and cell cycle control (5–7), thereby contributing to malignant transformation. miRNAs also act as tumor suppressors either by directly suppressing oncogenes such as MYCN (8), activating apoptosis (9), or by indirectly modulating oncogenic transcriptional profiles (10). The tissue-specific functions of miRNAs in multiple types of tumors is thought to be essential for the expression of MYC gene family (c-MYC, MYCN, MYCL) related cellular phenotypes (11, 12) and oncogenic or tumor suppressive roles for individual miRNAs in neuroblastoma continue to be discovered (13).

The MYCN oncogene plays a central role in the biology of neuroblastoma and several other cancers, both as a negative prognostic factor and as a driving oncogene for tumorigenesis (14). In fact, MYCN amplification remains the strongest negative prognostic indicator for neuroblastoma (15) and neural crest targeted expression of MYCN is sufficient to induce neuroblastoma in transgenic animals (16). However, as a bHLH transcription factor, MYCN also directly activates genes which limit proliferation and increase apoptosis and the net effect on cellular phenotypes is due to a complex balance of pro and anti-oncogenic effects. For example, MYCN directly activates p53 transcription and high levels of wild-type p53 are found in MYCN amplified neuroblastoma (17) and a major downstream target of p53, p21 (CDKN1A), is rapidly up-regulated on DNA damage in neuroblastoma. In contrast, MYCN also directly activates transcription of the oncogenic miR-17-92a miRNA cluster, which in turn targets and inhibits p21 translation (18). Thus, the oncogenic consequences of MYCN function is the net integrated result of activation and repression of multiple downstream pathways with contrasting tumor repressive or tumor promoting cellular effects (12).

To identify downstream MYCN-driven genes and miRNAs that are associated with aggressive NB we carried out a genome-wide search for MYCN-binding sites showing differential promoter occupation at high versus low MYCN condition. Using ChIP-Seq in a tetracycline inducible MYCN cell line (Tet-On) we identified genes and miRNAs that are co-repressed under conditions of high MYCN expression with tumor suppressive and, surprisingly, proproliferative effects in neuroblastoma tumor models. Detailed in vitro and in vivo characterization of 2 miRNAs strongly repressed by MYCN reveals miR-591 to be a potent tumor suppressor. In contrast miR-558, which was also down-regulated by MYCN, was strongly oncogenic. Our data highlight 2 distinct opposing effects of miRNAs downstream of MYCN. Dissecting the complex interactions of oncogenes such as MYCN and miRNAs which are also potent modulators of gene functions is critical to the rational incorporation of miRNA-mediated therapeutics to current treatment strategies.

Chromatin immunoprecipitation

ChIP was performed on MYCN3 cells either untreated or after addition of doxycycline (1 μg/mL) for 24 hours using a previously described method (19) with the following modifications. Cells were fixed in 1% formaldehyde for 10 minutes followed by 0.125M glycine for 15 minutes to prevent over fixation. Following lysis in a 1% SDS buffer, the samples were sonicated in 35 cycles of 30 seconds intervals in a Bioruptor sonicator (Diagenode). The samples were then diluted to 0.1% SDS cleared with protein A-agarose beads blocked with 0.1 mg/mL BSA by rotating for 45 minutes at 4°C. Using 2 μg mouse anti-MYCN or anti-actin antibodies, the samples were immunoprecipitated overnight with protein A-agarose beads blocked with 0.1 mg/mL BSA. The negative control was immunoprecipitated without antibody. Following immunoprecipitation, the samples were washed and the DNA was eluted and purified as described earlier. Input was generated by purifying DNA from the sonicated lysate of treated and untreated cells.

Quantitative PCR (qPCR) was performed using the Qiagen QuantiTect SYBR Green PCR Kit according to the manufacturer's instructions (Qiagen) using 5 μL of ChIP DNA per 25 μL reaction and 60°C annealing temperature for all primer sets. To control for the difference in DNA concentration between samples, a standard curve of amplification for each primer pair was generated using serial dilution of treated and untreated input. Both anti-MYCN and anti-actin samples were amplified in triplicate from treated and untreated samples.

The qPCR primers used to validate MYCN binding to the E-boxes within the MDM2 and MIRGH1 gene promoters are listed in Table 1.

Table 1.

The qPCR primers used to validate MYCN binding to the E-boxes within the MDM2 and MIRGH1 gene promoters

PrimerForwardReverse
MDM2 E-box 5′-agcctttgtgcggttcgtg-3′ 5′-cccccgtgacctttaccctg-3′ 
MIRGH1 E-box 1 5′-tctcccgagggcgagagttaaagc-3′ 5′-caccctcgcgcgtacaaagtttgg-3′ 
MIRGH1 E-box 2 5′-tggcggctacgcggagaatcgcag-3′ 5′-gaccgggcaggaacacccgagact-3′ 
MIRGH1 E-box 3 5′-tcctgctagtattgctcgactctt-3′ 5′-gtgcagttaggtccacgtgtatga-3′ 
MIRGH1 E-boxes 4 and 5 5′-aagtggaagccagaagaggaggaa-3′ 5′-tggcctctcccaaatggattgaca-3′ 
PrimerForwardReverse
MDM2 E-box 5′-agcctttgtgcggttcgtg-3′ 5′-cccccgtgacctttaccctg-3′ 
MIRGH1 E-box 1 5′-tctcccgagggcgagagttaaagc-3′ 5′-caccctcgcgcgtacaaagtttgg-3′ 
MIRGH1 E-box 2 5′-tggcggctacgcggagaatcgcag-3′ 5′-gaccgggcaggaacacccgagact-3′ 
MIRGH1 E-box 3 5′-tcctgctagtattgctcgactctt-3′ 5′-gtgcagttaggtccacgtgtatga-3′ 
MIRGH1 E-boxes 4 and 5 5′-aagtggaagccagaagaggaggaa-3′ 5′-tggcctctcccaaatggattgaca-3′ 

Parallel sequencing

Chromatin immunoprecipitated DNA was prepared for sequencing following a modified version of Illumina's Genomic DNA protocol. Sheared DNA fragments undergo end repair by T4 DNA polymerase to fill in 5′ overhangs, Klenow polymerase to remove 3′ overhangs, and T4PNK to phosphorylate the 5′-OH. For Illumina adapter ligation, a single adenine nucleotide overhang must be added to the polished DNA. A 1:30 dilution of the Adaptor Oligo Mix (20) was used in the ligation step. The resulting adapted constructs are amplified by 18 cycles of PCR enriching the DNA fragments with adapters flanking each end. The amplified product was quantified using a spectrophotometer at 260 nm (Nanodrop) then a portion of final sample is size separated on a 6% TBE (Invitrogen) to verify correct target range amplification. The validated DNA library is submitted at a 10 nmol/L concentration for sequencing. All protocols for Illumina library sequence preparation, sequencing, and quality control are provided by Illumina (20).

Mapping sequencing data

Between 3.7 and 3.9 × 106 sequence tags for each immunoprecipitation condition were mapped onto the human genome reference sequence (HG18) using PASH (21) in an approach similar to Ahn and colleagues (22). Uniquely mapping reads were selected and the read coverage computed over 100bp windows tiling across the entire genome. The read density was then normalized employing a quantile normalization step, ensuring a similar distribution across the 3 data sets. Next we computed then the difference between each sample (MYCN-induced cell line, a MYCN–non-induced cell line) and the control, and generated P values assuming a normal distribution of the differences. The false discovery rate for each window was computed and an FDR below 0.001 was selected, and used to apply a segmentation algorithm to the remaining windows. Finally, we selected genomic features such as gene promoters, miRNA promoters, CpG islands, etc. and computed enrichment over background to identify promoters that are differentially occupied by MYCN in either the MYCN-induced or MYCN–non-induced sample. The analysis results were stored and visualized using the Genboree Discovery Platform (23).

PCR assays

To induce the MYCN expression, MYCN3 cell line was treated with doxycycline (1 μg/mL) for 12, 24, and 48 hours. Snap frozen tumor tissues were processed using RNALater Ice before the RNA isolation. Total RNA including miR fraction was isolated using Qiagen miRNeasy kit as per the manufacturer's protocol. For exon/intron/stemloop PCR, total RNA was reverse transcribed using TaqMan Reverse Transcription Reagent and amplified using Taq polymerase in ABI Veriti Thermocycler. Primers (Integrated DNA Technologies) for the respective exons, introns, and stemloops were designed as explained in Figure 3. PCR products were analyzed after 1% to 3% agarose gel electrophoresis.

Figure 1.

Chromatin immunoprecipitation with MYCN conditional cells. A, the MYCN3 cell line rapidly upregulates MYCN expression on exposure to 1 μg/mL doxycycline. QPCR for mRNA expression shows 20-fold increase in message level with concomitant increase in protein. B, induced MYCN is functional as shown by alteration in cell cycle distribution 24 hours after exposure to doxycycline. C and D, chip-PCR results for E-boxes of 2 known MYCN target genes MIR17HG and MDM2, confirming the specificity of the immunoprecipitation and relative enrichment for MYCN target gene promoters.

Figure 1.

Chromatin immunoprecipitation with MYCN conditional cells. A, the MYCN3 cell line rapidly upregulates MYCN expression on exposure to 1 μg/mL doxycycline. QPCR for mRNA expression shows 20-fold increase in message level with concomitant increase in protein. B, induced MYCN is functional as shown by alteration in cell cycle distribution 24 hours after exposure to doxycycline. C and D, chip-PCR results for E-boxes of 2 known MYCN target genes MIR17HG and MDM2, confirming the specificity of the immunoprecipitation and relative enrichment for MYCN target gene promoters.

Close modal
Figure 2.

Correlations of microRNA host genes expression with clinical outcome in neuroblastoma. Representative Kaplan–Meier survival curves for genes illustrating both positive (higher expression = worse outcome) and negative (lower expression = worse outcome) for genes bound by MYCN identified by ChIP-seq. Clinical characteristics, MYCN status, and distribution of patients for this data set have been previously published (27). Survival probability and P values calculated using Kaplan–Meier and Mantel–Haenszel methods as described previously (for log-rank testing, n = 28 samples in each group). Overall, 50% of genes containing microRNAs had significant correlations when queried with this data set (http://pob.abcc.ncifcrf.gov/cgi-bin/JK; Table 3).

Figure 2.

Correlations of microRNA host genes expression with clinical outcome in neuroblastoma. Representative Kaplan–Meier survival curves for genes illustrating both positive (higher expression = worse outcome) and negative (lower expression = worse outcome) for genes bound by MYCN identified by ChIP-seq. Clinical characteristics, MYCN status, and distribution of patients for this data set have been previously published (27). Survival probability and P values calculated using Kaplan–Meier and Mantel–Haenszel methods as described previously (for log-rank testing, n = 28 samples in each group). Overall, 50% of genes containing microRNAs had significant correlations when queried with this data set (http://pob.abcc.ncifcrf.gov/cgi-bin/JK; Table 3).

Close modal
Figure 3.

MYCN-dependent expression of genes and intronic microRNAs. A, diagram demonstrating PCR strategy to evaluate the expression pattern of miR stemloops, miR hosting introns, and surrounding exons. B, composite agarose gel figures from MYCN-induced (+) and -uninduced (−) MYCN3 cell lines, demonstrating significant loss of expression of corresponding introns, exons, and stem loops in the induced line compared to the uninduced line. C, composite TaqMan MicroRNA Assay data from MYCN-induced MYCN3 cell lines, demonstrating significantly reduced expression of mature miRs on MYCN induction (P values calculated by Kruskal–Wallis method; mean ± SEM; n = 3 in each group, *, P < 0.05, **, P < 0.01).

Figure 3.

MYCN-dependent expression of genes and intronic microRNAs. A, diagram demonstrating PCR strategy to evaluate the expression pattern of miR stemloops, miR hosting introns, and surrounding exons. B, composite agarose gel figures from MYCN-induced (+) and -uninduced (−) MYCN3 cell lines, demonstrating significant loss of expression of corresponding introns, exons, and stem loops in the induced line compared to the uninduced line. C, composite TaqMan MicroRNA Assay data from MYCN-induced MYCN3 cell lines, demonstrating significantly reduced expression of mature miRs on MYCN induction (P values calculated by Kruskal–Wallis method; mean ± SEM; n = 3 in each group, *, P < 0.05, **, P < 0.01).

Close modal

TaqMan MicroRNA Assay probes for hsa-miR-558 and hsa-miR-591 were used to reverse transcribe and quantify using ABI StepOne RealTime PCR System. StepOne Software, Version 2.1 was used to analyze data. Kruskal–Wallis test (1-way ANOVA) was used for the statistical analysis.

MTT proliferation assays

Three thousand cells/well were plated into 96-well microtiter plates. Plates were incubated for 5 days. Proliferation was assessed at baseline and daily in triplicate. To assess proliferation, 15 μL of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) were added to each well and plates read at 550 nm.

Colony formation in soft agar

Colony formation assays were performed using standard conditions as in ref. 24. Briefly, soft agar plates were set-up and plated with 5000 cells (CHLA-255) and 10,000 cells (SH-SY5Y) per well (6-well plates) and incubated for 10 days. Colonies were stained with Thiazolyl Blue Tetrazolium Bromide (Sigma-Aldrich M5655) and visualized and counted with (BioRad Versadoc, Image 1 software).

Animal model

The Institutional Animal Care and Use Committee of Baylor College of Medicine approved the experiments described below. Four- to 6-week-old female NCr nude mice (Taconic Farms) were housed in a barrier facility and acclimated to 12-hour light/dark cycles for at least 48 hours before experimental use.

After tumor cells were counted, they were washed and resuspended in sterile PBS at a concentration of 107 per mL. Mice (n = 30) were anesthetized with intraperitoneal ketamine (50 mg/kg) and xylazine (5 mg/kg) and 2% inhaled isofluorane. The left flank of mice was surgically prepared in a sterile fashion, and a transverse incision was performed to expose the left kidney. An inoculum of 106 tumor cells in 0.1 mL of PBS was injected under the renal capsule using a 27-gauge needle. Tumor cells lines used were CHLA-255 (n = 10), CHLA-255 miR558 (n = 10), and CHLA-255 miR591 (n = 10). The flank musculature was closed with 4-0 Vicryl suture (Ethicon), and the skin incision was closed with a skin staple.

Tissue harvest.

At the time of sacrifice, tumors and contralateral kidneys were resected and weighed. Tumors were visualized under fluorescent microscopy (Olympus SZX16) to assess GFP expression. Portions of tumor were also stored in RNAlater (Ambion Inc.) for quantitative miRNA analysis. Portions of tumor were also preserved for histology.

Cell lines.

The neuroblastoma cell lines SH-SY5Y and SHEP were obtained from ATCC. Luciferase positive SH-SY5Y was a kind gift from Dr. Darrell Yamashiro, Columbia University, NY. The MYCN inducible MYCN3 cell line was generated from SHEP as previously described (19). CHLA-255 were obtained from Dr. Leonid Metelista. All cell lines used were validated for GD2 expression and MYCN expression by flow cytometry and Western blot. All cell lines were confirmed negative for mycoplasma by routine testing.

A genome-wide search reveals novel MYCN targets that are repressed under conditions of high MYCN and therefore putative tumor suppressors of aggressive neuroblastoma

To identify gene and miRNA promoters that show increased MYCN binding under high MYCN conditions in the same genetic background, we used a MYCN-conditional cell line (MYCN-3, Tet-on) with rapid and robust inducible expression of functional MYCN for anti-MYCN ChIP (Fig. 1A; ref. 25). Following cross-linking of protein/DNA adducts, we isolated cell lysates under MYCN-induced (24 hours) and non-induced conditions. DNA was isolated, sheared, and MYCN-bound DNA was immunoprecipitated with a MYCN antibody as previously described (19). To confirm the specificity of our immunoprecipitation conditions, the assay was validated using qPCR analysis for the known MYCN target genes MDM2 (19) and MIR17HG (26) and isotype control antibodies (Fig. 1B).

Next, purified ChIP DNA was sequenced on the Illumina GA-1 sequencing platform and sequence reads mapped on the human genome reference sequence (hg18, NCBI build 36) using the PASH algorithm (22). For the MYCN-induced cells, of 2,751,696 uniquely mapped sequences, 15,322 tags mapped to E-box DNA sequences. For the noninduced (MYCN-low) cells, we obtained 2,606,250 sequence tags of which 12,986 mapped to E-boxes. After normalization to ensure comparable genomic distribution between data sets, we calculated P values for each enrichment site and then determined false discovery rate (FDR) values for each genomic element sequenced. We next filtered all windows with a stringent FDR (<0.001) to identify MYCN binding to those E-boxes either (a) within 3kb of a known miRNA or (b) E-boxes within promoters of genes hosting intronic miRNAs (see Methods and Supplementary Figs. S1–S4 and Tables ST1–ST3, for details). With this approach we identified a list of 20 genes encoding 29 intronic miRNAs with clear enrichment of E-box–specific MYCN binding (FDR < 0.001) shown in Table 2.

Table 2.

Host genes and microRNAs with significantly differential MYCN binding

Host geneMicroRNACytogenetic coordinateFold change at E-boxFold change in promoter bindingE-box sequence
ATF2 hsa-mir-933 2q31.1 3.43 3.43 CACCTG 
NDUFAF3 hsa-mir-191 3p21.31 2.13 2.13 CACCTG 
DALRD3.2 hsa-mir-425 3p21.31 2.13 2.13 CACCTG 
RNF130 hsa-mir-340 5q35.3 2.09 2.46 CACGCG 
FBXL18 hsa-mir-589 7p22.1 2.03 1.68 CACCTG 
SLC25A13 hsa-mir-591 7q21.3 1.93 1.93 CACGTG 
BIRC6 hsa-mir-558 2p22.3 1.83 1.83 CACGCG 
ZNF766 hsa-mir-643 19q13.33 1.82 1.82 CAACTG 
PANK3 hsa-mir-103–1 5q35.1 1.78 1.63 CACGTG 
TLE3 hsa-mir-629 15q23 1.77 1.77 CACGCG 
Intergenic hsa-mir-663 20p11.1 1.64 1.61 CACGCG 
PLEC1.5 hsa-mir-661 8q24.3 1.62 1.44 CATCTG 
LARP7 hsa-mir-302a 4q25 1.56 1.56 CAATTG 
– hsa-mir-302b – – – – 
– hsa-mir-302c – – – – 
– hsa-mir-302d – – – – 
LARP7 hsa-mir-367 – – – – 
ATAD2 hsa-mir-548d-1 8q24.13 1.54 1.54 CACCTG 
PANK2 hsa-mir-103–2 20p13 1.47 1.47 CACGCG 
ABCF1.2 hsa-mir-877 6p21.33 1.37 1.37 CACGTG 
HNRPK hsa-mir-7–1 9q21.32 1.31 1.17 CACGCG 
SUPT3H hsa-mir-586 6p12.3 1.24 1.16 CACGCG 
EIF4H.2 hsa-mir-590 7q11.23 1.17 1.36 CACGTG 
MIR17HG hsa-mir-17 13q31.3 1.15 1.12 CATGTG 
– hsa-mir-18a – – – – 
– hsa-mir-19a – – – – 
– hsa-mir-19b-1 – – – – 
– hsa-mir-20a – – – – 
– hsa-mir-92a-1 – – – – 
SFRS2 hsa-mir-636 17q25.2 1.08 1.22 CAACTG 
Host geneMicroRNACytogenetic coordinateFold change at E-boxFold change in promoter bindingE-box sequence
ATF2 hsa-mir-933 2q31.1 3.43 3.43 CACCTG 
NDUFAF3 hsa-mir-191 3p21.31 2.13 2.13 CACCTG 
DALRD3.2 hsa-mir-425 3p21.31 2.13 2.13 CACCTG 
RNF130 hsa-mir-340 5q35.3 2.09 2.46 CACGCG 
FBXL18 hsa-mir-589 7p22.1 2.03 1.68 CACCTG 
SLC25A13 hsa-mir-591 7q21.3 1.93 1.93 CACGTG 
BIRC6 hsa-mir-558 2p22.3 1.83 1.83 CACGCG 
ZNF766 hsa-mir-643 19q13.33 1.82 1.82 CAACTG 
PANK3 hsa-mir-103–1 5q35.1 1.78 1.63 CACGTG 
TLE3 hsa-mir-629 15q23 1.77 1.77 CACGCG 
Intergenic hsa-mir-663 20p11.1 1.64 1.61 CACGCG 
PLEC1.5 hsa-mir-661 8q24.3 1.62 1.44 CATCTG 
LARP7 hsa-mir-302a 4q25 1.56 1.56 CAATTG 
– hsa-mir-302b – – – – 
– hsa-mir-302c – – – – 
– hsa-mir-302d – – – – 
LARP7 hsa-mir-367 – – – – 
ATAD2 hsa-mir-548d-1 8q24.13 1.54 1.54 CACCTG 
PANK2 hsa-mir-103–2 20p13 1.47 1.47 CACGCG 
ABCF1.2 hsa-mir-877 6p21.33 1.37 1.37 CACGTG 
HNRPK hsa-mir-7–1 9q21.32 1.31 1.17 CACGCG 
SUPT3H hsa-mir-586 6p12.3 1.24 1.16 CACGCG 
EIF4H.2 hsa-mir-590 7q11.23 1.17 1.36 CACGTG 
MIR17HG hsa-mir-17 13q31.3 1.15 1.12 CATGTG 
– hsa-mir-18a – – – – 
– hsa-mir-19a – – – – 
– hsa-mir-19b-1 – – – – 
– hsa-mir-20a – – – – 
– hsa-mir-92a-1 – – – – 
SFRS2 hsa-mir-636 17q25.2 1.08 1.22 CAACTG 

NOTE: ChIP-Seq identified 20 genes and encoding 29 microRNAs and an intergenic loci encoding a single microRNA with significantly increased binding of MYCN on conditional upregulation (FDR <0.001). Sequence tags within 500bp of promoter E-box sequences or within 3 kb of intergenic microRNA loci were quantified and fold change in occupancy calculated. Fold change for the total promoter region was very similar to E-box binding, suggesting that the majority of increased binding was specific for E-boxes (, common data for microRNAs expressed as clusters or coexpressed by 1 promoter).

Significant clinical correlations of MYCN-targeted microRNA host genes

To prioritize further functional experiments, we queried the neuroblastoma prognosis data base which presents microarray gene expression data and annotated clinical data for neuroblastoma and other pediatric cancers (http://pob.abcc.ncifcrf.gov/cgi-bin/JK), to obtain gene expression correlations with survival. Here Kaplan–Meier analysis was performed based on target gene expression dichotomized according to overall median gene expression and “good outcome” (survival >3 years from diagnosis) and “poor outcome” (death from disease). Detailed patient cohort descriptions and methods for this statistical analysis can be found in ref. 27. We found 11 of the 20 MYCN target genes identified in our study demonstrated significant associations with clinical outcome including the well-described oncogenic miRNA cluster miR-17-92 (Fig. 2 and Table 3). As illustrated by representative Kaplan–Meier curves, worse clinical outcomes correlated with both lower gene expression (4 genes) and higher gene expression (7 genes; Fig. 2). We further analyzed the expression of 2 intronic miRNAs, whose host genes had higher expression (SLC25A13, P < 0.001) or lower expression (BIRC5, P < 0.005) correlating with poor outcome (Fig. 2).

Table 3.

Eleven of 20 microRNA hosting genes had increase binding of MYCN and significant clinical correlations with overall survival

Host geneMicroRNAWorse survivalP
NDUFAF3 hsa-mir-191 Higher <0.01 
SLC25A13 hsa-mir-591 Higher <0.001 
SFRS2 hsa-mir-636 Higher <0.01 
ATAD2 hsa-mir-548d-1 Higher <0.0001 
ABCF1 hsa-mir-877 Higher <0.0003 
EIF4H hsa-mir-590 Higher <0.004 
LARP7 hsa-mir-302a-d Higher <0.001 
 hsa-mir-367 Higher  
MIR17HG hsa-mir-17–92 Higher <0.001 
ATF2 hsa-mir-933 Lower <0.01 
BIRC6 hsa-mir-558 Lower <0.005 
HNRPK hsa-mir-7–1 Lower <0.001 
DALRD3.2 hsa-mir-425 NC  
RNF130 hsa-mir-340 NC  
FBXL18 hsa-mir-589 NC  
ZNF766 hsa-mir-643 NC  
PANK3 hsa-mir-103–1 NC  
TLE3 hsa-mir-629 NC  
PLEC1 has-miR-661 NC  
PANK2 hsa-mir-103–2 NC  
SUPT3H hsa-mir-586 NC  
Intergenic hsa-mir-663 No data  
Host geneMicroRNAWorse survivalP
NDUFAF3 hsa-mir-191 Higher <0.01 
SLC25A13 hsa-mir-591 Higher <0.001 
SFRS2 hsa-mir-636 Higher <0.01 
ATAD2 hsa-mir-548d-1 Higher <0.0001 
ABCF1 hsa-mir-877 Higher <0.0003 
EIF4H hsa-mir-590 Higher <0.004 
LARP7 hsa-mir-302a-d Higher <0.001 
 hsa-mir-367 Higher  
MIR17HG hsa-mir-17–92 Higher <0.001 
ATF2 hsa-mir-933 Lower <0.01 
BIRC6 hsa-mir-558 Lower <0.005 
HNRPK hsa-mir-7–1 Lower <0.001 
DALRD3.2 hsa-mir-425 NC  
RNF130 hsa-mir-340 NC  
FBXL18 hsa-mir-589 NC  
ZNF766 hsa-mir-643 NC  
PANK3 hsa-mir-103–1 NC  
TLE3 hsa-mir-629 NC  
PLEC1 has-miR-661 NC  
PANK2 hsa-mir-103–2 NC  
SUPT3H hsa-mir-586 NC  
Intergenic hsa-mir-663 No data  

NOTE: Correlation of gene expression (by microarray) versus survival was assessed using the public data sets for neuroblastoma gene expression at the Oncogenomics Section, Pediatric Oncology Branch, National Cancer Institute (50). Ten of 20 genes listed had significant correlations with survival. Higher expression correlated with worse survival for 7 genes where 3 genes had lower expression correlate with improved survival. Two microRNA clusters, mir-302a-d and miR-17-92, account for 10 microRNAs. Of note, MiR-17-92 is a known MYCN target and negative.

The majority of clinically significant microRNA–gene pairs are repressed under conditions of high MYCN

To systematically assess the impact of differential MYCN binding at promoter E-boxes, we designed RT-PCR primer pairs to measure independently the levels of an exon of the host gene, the intron in which the miRNA is located and the stem loop precursor from which the intronic miRNA is processed. We found reduction of expression for the majority of MYCN targeted host gene/miRNA pairs in all of the genomic elements measured (Fig. 3). Since MYCN is a key oncogene driving aggressive NB we would expect that these host gene/miRNA pairs repressed by MYCN should be tumor suppressors whose increased expression is linked with better survival. However, increased expression correlated with better survival in only 3 of the 11 clinically significant host gene/miRNA pairs repressed by MYCN. By contrast, increased expression was correlated with worse survival for the majority of these genes (8 of 11 gene–miRNAs pairs; Table 3). These surprising observations led us to further investigate the function of the associated intronic miRNAs, which in some cases can strongly repress or oppose the function of host genes as part of negative regulator feed-back loops (28).

We focused on 2 representative miRNAs from each group, miR-558 (whose host gene correlates with better survival) and miR-591 (whose host gene correlates with worse survival). First we performed real-time qPCR to validate the MYCN-dependent repression for each miRNA. As shown in Figure 3B, MYCN induction in the MYCN-3 line led to a rapid decline in mature miRNA levels of both miR-558 and miR-591 within 12 hours. This change in the host gene, the stem loop precursor, and the mature miRNA persisted over the time course of the experiments.

These data led us to evaluate whether these miRNAs are specifically repressed in the MYCN-amplified subset of high-risk tumors. Therefore, we analyzed miR-558 and miR-591 expression in an independent cohort of 54 patient samples with qPCR according to a previously established protocol (29). Intriguingly, we found very low to undetectable levels of both these miRNAs in all high-risk samples regardless of MYCN status (P < 0.01). Details on clinical characteristics, stage, and MYCN status can be found in Supplementary Table T4 and in ref. 19. Our results suggest that both these MYCN repressed miRNAs may play a role in blocking MYCN-driven tumor initiation, which occurs prior to MYCN amplification as demonstrated by transgenic animal models (3) and suggested to us that both these miRNAs could be potent tumor suppressors.

It is important to note that a proposed tumor suppressor function for miR-591 contrasts with the antiapoptotic function of the host gene SLC25a13. This gene encodes the mitochondrial solute carrier protein Citrin, which is an aspartate–glutamate exchanger and implicated as the causative gene for both juvenile and adult onset citrullinemia (30). Deficiency or mutations of SLC25a13 interfere with NADH mitochondrial transport and lead to progressive liver disease. ShRNA-mediated knockdown of SLC25a13 activates mitochondrial-mediated apoptosis in hepatocytes (31). Similar proproliferative cell protective effects are seen BIRC6, the host gene of miR-558 which encodes BRUCE/Apollon a member of the IAP (inhibitor of apoptosis protein) family. Recently, BRUCE has been shown to promote the proliferation of neural progenitor cells (NPCs) and knockdown reduces growth and increases sensitivity of NPCs to glucocorticoids (32). BRUCE has also been shown to protect cardiac myocytes and breast cancer cells from stress-induced apoptosis (33). As shown later, these observations are consistent with the function of its intronic miR-558 which also strongly promotes proliferation of neuroblastoma cell lines and tumors.

MYCN-repressed miR-558 and miR-591 show contrasting tumor suppressor and tumorigenic functions in neuroblastoma

Because our ChIP-seq, qPCR, and profiling of clinical samples suggested that miR-558 and miR-591 may be potent tumor suppressors despite the opposite effects of their host genes on survival, we sought to determine the functional consequences of overexpression of miR-558 and miR-591 in neuroblastoma. Subclones of the SH-SY5Y and ChLA-255 neuroblastoma cell lines were retrovirally transduced with miR-558 or miR-591 stem loops with 140bp of flanking sequence and selected by FACS sorting for GFP expression to avoid artifacts introduced by extended antibiotic selection (pSuper-GFP). Both of these cell lines form highly undifferentiated tumors and metastasize in vivo, reflective of high-risk neuroblastoma with low to moderate MYCN expression.

The impact of these miRNAs on proliferation and apoptosis was evaluated using MTT assays to measure in vitro proliferation, and colony formation using soft agar colony growth as shown in Figure 4. The miR-591 had little effect on cell proliferation relative to the control miRNA-GFP construct with a mutated 5′ seed region or the host cell line without a miRNA construct. This miRNA did not strongly alter colony formation, although a trend towards decreased colony formation relative to the parental or control cell lines was observed. By sharp contrast, miR-558 dramatically and consistently increased both proliferation and colony formation in both cell lines. These in vitro studies clearly demonstrated that miR-558 expression promoted tumor growth consistent with the proproliferative function of its host gene BIRC6.

Figure 4.

In vitro effects miR-558 and mir-591 on proliferation and colony formation. Stable cell line expression of microRNAs with GFP were compared to control cells. A, proliferation as measured by increase in absorbance from MTT assay. Mir-558 consistently showed increased proliferation rates in both CHLA-255 and SH-SY5Y cells. B, colony formation assays in soft agar—representative wells showing colony growth at 10 days. Colonies were stained and counted by eye or by digital image analysis software. C, colony counts from triplicate experiments showing a pronounced effect of mir-558 relative to miR-591 or control. Mir-591 showed a trend toward decreased colonies than SH-SY5Y. *, P < 0.05 by Kruskal–Wallis, error bars show mean ± SD.

Figure 4.

In vitro effects miR-558 and mir-591 on proliferation and colony formation. Stable cell line expression of microRNAs with GFP were compared to control cells. A, proliferation as measured by increase in absorbance from MTT assay. Mir-558 consistently showed increased proliferation rates in both CHLA-255 and SH-SY5Y cells. B, colony formation assays in soft agar—representative wells showing colony growth at 10 days. Colonies were stained and counted by eye or by digital image analysis software. C, colony counts from triplicate experiments showing a pronounced effect of mir-558 relative to miR-591 or control. Mir-591 showed a trend toward decreased colonies than SH-SY5Y. *, P < 0.05 by Kruskal–Wallis, error bars show mean ± SD.

Close modal

In vivo orthotopic implantation and tumor formation confirm that miR-591 is a tumor suppressor and miR-558 is an oncomiR

Because the effects of miRNAs on cell lines growing in culture may be significantly different from their effects on tumors growing in vivo, we generated orthotopic xenografts of our cell lines in nude mice. We implanted tumor cells directly into the left renal capsule of mice and allowed them to grow for 5 weeks. This model recapitulates the growth of human neuroblastomas, which often arise from adrenal or perirenal locations and are highly vascular and invasive (34, 35). At necropsy, tumors were resected, weighed, and analyzed histologically. As shown in Figure 5, miR-591 significantly repressed tumor growth with average weights less than 50% of control tumors (P < 0.05). These tumors were small and sclerotic and relatively avascular. In contrast, constitutive expression of miR-558 led to highly vascular tumors that were over twice the average weight of tumors from the nontransduced parental line or the mir-558-mutant line containing a mutated 5′ seed sequence (P < 0.05). To confirm expression of the mature miRNAs in our tumors, we performed qRT-PCR for the mature miRNA in multiple tumors. In addition, fluorescent microscopy of the tumor specimens showed strong expression of the GFP marker in all tumors (Fig. 5).

Figure 5.

Contrasting effects on xenograft tumor growth of miR-558 and miR-591. A, representative tumor samples with contralateral kidney (*) for comparison. B, average tumor weight for each cohort ± SEM. Tumors overexpressing miR-558 are significantly larger than control tumors, whereas tumors overexpressing miR-591 are significantly smaller than control (*, Kruskal–Wallis method P < 0.01; mean ± SEM; n = 9 in miR-591 cohort, n = 10 in all other cohorts). C, fluorescent microscopy of tumor specimens at time of necropsy confirming GFP expression in overexpressing miR-558 and miR-591 tumors. D, composite TaqMan microRNA assay data from tumors overexpressing miR-558 and miR-591, demonstrating significantly increased expression of the corresponding mature miRs (Kruskal–Wallis test; mean ± SEM; n = 3 in each group).

Figure 5.

Contrasting effects on xenograft tumor growth of miR-558 and miR-591. A, representative tumor samples with contralateral kidney (*) for comparison. B, average tumor weight for each cohort ± SEM. Tumors overexpressing miR-558 are significantly larger than control tumors, whereas tumors overexpressing miR-591 are significantly smaller than control (*, Kruskal–Wallis method P < 0.01; mean ± SEM; n = 9 in miR-591 cohort, n = 10 in all other cohorts). C, fluorescent microscopy of tumor specimens at time of necropsy confirming GFP expression in overexpressing miR-558 and miR-591 tumors. D, composite TaqMan microRNA assay data from tumors overexpressing miR-558 and miR-591, demonstrating significantly increased expression of the corresponding mature miRs (Kruskal–Wallis test; mean ± SEM; n = 3 in each group).

Close modal

These in vitro and in vivo functional studies for 2 MYCN-repressed miRNAs reveal contrasting tumor suppressive (miR-591) and oncogenic (miR-558) effects on neuroblastoma growth and tumor formation, which either oppose (SLCA25a13/miR-591) or correlate with (BIRC6/miR-558) functions of their host genes. The therapeutic and biologic implications of these data are discussed later in the context of MYCN-driven neuroblastoma.

MYCN amplification is found in about 50% of high-risk neuroblastoma (23% overall) and confers a particularly aggressive phenotype with poor survival (15, 36). By contrast, variations in lower levels of MYCN which are found in nonamplified tumors are clearly not correlated with outcome (37). This suggests that the promoters for MCYN target genes that are differentially regulated at very high levels of MYCN may be particularly important drivers of the highly aggressive MYCN-amplified phenotype.

We simulated MYCN-amplified aggressive neuroblastoma and MYCN-unamplified neuroblastoma by using a conditional cell line that expresses high and low levels of MYCN in the same genetic background and used ChIP-seq to identify gene and miRNA promoters with differential binding of MYCN. Induction of MYCN expression lead to marked changes in promoter occupancy of a subset of genes hosting intronic miRNAs. Interestingly, of the 11 genes identified with significant clinical correlations, 8 were strongly repressed by increased MYCN binding independently of the clinical correlation. By contrast, both MIR17HG which encodes the oncogenic miR-17-92a cluster and EIF4H encoding miR-590 showed increased expression on upregulation of MYCN.

Although MYCN and C-Myc are classically defined as oncogenes which activate transcription of proproliferative cellular networks, it is increasingly clear that they mediate critical oncogenic networks through repression of target genes including miRNAs. Direct inhibitory interactions between MYC, MIZ-1, and SP1 repress p21 (CDNK1A; ref. 38), and MYC oncogenes can directly regulate promoter methylation (39, 40) and increase HDAC expression (41). It has recently been shown that MYCN binding as a SP1/MIZ1/MYCN complex recruits HDAC1 to local histones as well (42). Our data suggests that some of these repressive effects may be critically regulated by the extent of MYCN promoter occupancy, distinguishing MYCN function in amplified and nonamplified neuroblastoma.

Through this work we identified miR-591 as a potent new tumor suppressor for neuroblastoma, a function not predicted by its expression from an intron in the SLC25a13 host gene. This miRNA is expressed at very low levels in high-risk neuroblastoma and strongly suppresses tumor growth of neuroblastoma xenografts. Exogenous in vivo delivery of this miRNA to neuroblastoma or reactivation of miR-591 target pathways could be a novel therapeutic approach for treating aggressive NB. However, exogenous expression of another MYCN-repressed miRNA, miR-558, produced marked pro-oncogenic effects in line with an antiapoptotic function of its host gene BIRC6.

These contrasting tumor suppressive and effects of MYCN-repressed miRNAs highlight the complexity of interactions of oncogenes and downstream miRNAs in aggressive cancers. MYCN expression shortens the cell cycle and increases contact independent growth and metastasis in vitro and in vivo (43, 44). However, it is also clear from in vitro and in vivo models that MYCN can directly upregulate the p53 protein and sensitize cells to apoptosis from genotoxic stress (44, 45). Neuroblastoma is uniformly p53 wild type, and p53 signaling can be robustly stimulated in vitro and in vivo in these cancers (25, 46). These multiple effects of MYCN may manifest themselves at distinct time points of tumorigenesis (3, 47). Hence MYCN function may help to prevent differentiation and apoptosis of neuroblast precursors during tumor initiation, and then later restrain proliferation through p53-dependent and -independent mechanisms to match local nutrient and oxygen supplies within the tumor microenvironment.

The role of miRNAs as effectors of the MYCN phenotype is a major focus of interest due to their widespread effects on transcriptional programs and potential therapeutic utility as both biomarkers and direct therapeutic tools (48, 49). Our findings showing that high-level MYCN expression can repress a new oncogenic miRNA, miR-558, and a new tumor suppressor, miR-591, are consistent with the above hypothesis that MYCN must restrain growth as well as promote it. Thus, repression of miR-558 may be critical to the survival of early neuroblasts expanding under high MYCN levels. In contrast, repression of mir-591 may be a prerequisite for the expansion of these tumor precursors after initiation.

Finally, these studies reveal an additional and new layer of complexity in the role of MYCN as a driving oncogene for neuroblastoma. This transcription factor suppresses the expression of miRNAs with potent proproliferative effects (e.g., miR-558) which could modify the effectiveness of anti-MYCN–targeted therapies. miRNAs engage in “feed-forward” and “feed-back” regulatory pathways in many cellular processes including reprogramming of induced pluripotent cells and maintenance of embryonic stem cell pluripotency. How oncogenes such as MYCN disrupt and co-opt these highly orchestrated functions of miRNAs that define cellular phenotypes and states of differentiation should be carefully considered in the design of novel therapeutic strategies for neuroblastoma and other MYC-driven malignancies.

No potential conflicts of interest were disclosed.

We would like to acknowledge the contributions of Dr. Jianhua Yang, Dr. Leonid Metelitsa, and Dr. Malcolm Brenner for their help with review of the article.

This work was supported by a RSG grant from the American Cancer Society, and support from the Hope Street Kids Foundation and Children's Neuroblastoma Cancer Foundation to J.M. Shohet and The Virginia & L.E. Simmons Foundation to J.M. Shohet and P.H. Gunaratne. This work would not have been possible without University of Houston's Illumina/Solexa Genome Analyzer GA-1 acquired through generous funding from the Cullen Foundation to the University of Houston's Institute for Molecular Design (IMD).

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.
Maris
JM
. 
Recent advances in neuroblastoma
.
N Engl J Med
2010
;
362
:
2202
11
.
2.
Croce
CM
. 
Oncogenes and cancer
.
N Engl J Med
2008
;
358
:
502
11
.
3.
Hansford
LM
,
Thomas
WD
,
Keating
JM
,
Burkhart
CA
,
Peaston
AE
,
Norris
MD
, et al
Mechanisms of embryonal tumor initiation: distinct roles for MycN expression and MYCN amplification
.
Proc Natl Acad Sci U S A
2004
;
101
:
12664
9
.
4.
Johnsen
JI
,
Kogner
P
,
Albihn
A
,
Henriksson
MA
. 
Embryonal neural tumours and cell death
.
Apoptosis
2009
;
14
:
424
38
.
5.
Borgdorff
V
,
Lleonart
ME
,
Bishop
CL
,
Fessart
D
,
Bergin
AH
,
Overhoff
MG
, et al
Multiple microRNAs rescue from Ras-induced senescence by inhibiting p21(Waf1/Cip1)
.
Oncogene
2010
;
29
:
2262
71
.
6.
Imam
JS
,
Buddavarapu
K
,
Lee-Chang
JS
,
Ganapathy
S
,
Camosy
C
,
Chen
Y
, et al
MicroRNA-185 suppresses tumor growth and progression by targeting the Six1 oncogene in human cancers
.
Oncogene
2010
;
29
4971
9
.
7.
Li
H
,
Bian
C
,
Liao
L
,
Li
J
,
Zhao
RC
. 
miR-17–5p promotes human breast cancer cell migration and invasion through suppression of HBP1
.
Breast Cancer Res Treat
2011
;
126
:
565
75
.
8.
Wei
JS
,
Song
YK
,
Durinck
S
,
Chen
QR
,
Cheuk
AT
,
Tsang
P
, et al
The MYCN oncogene is a direct target of miR-34a
.
Oncogene
2008
;
27
:
5204
13
.
9.
Hermeking
H
. 
p53 enters the microRNA world
.
Cancer Cell
2007
;
12
:
414
8
.
10.
Kasinski
AL
,
Slack
FJ
. 
Potential microRNA therapies targeting Ras, NFkappaB and p53 signaling
.
Curr Opin Mol Ther
2010
;
12
:
147
57
.
11.
Ma
L
,
Young
J
,
Prabhala
H
,
Pan
E
,
Mestdagh
P
,
Muth
D
, et al
miR-9, a MYC/MYCN-activated microRNA, regulates E-cadherin and cancer metastasis
.
Nat Cell Biol
2010
;
12
:
247
56
.
12.
Mestdagh
P
,
Fredlund
E
,
Pattyn
F
,
Schulte
JH
,
Muth
D
,
Vermeulen
J
, et al
MYCN/c-MYC-induced microRNAs repress coding gene networks associated with poor outcome in MYCN/c-MYC-activated tumors
.
Oncogene
2010
;
29
:
1394
404
.
13.
Northcott
PA
,
Fernandez-LA
,
Hagan
JP
,
Ellison
DW
,
Grajkowska
W
,
Gillespie
Y
, et al
The miR-17/92 polycistron is up-regulated in sonic hedgehog-driven medulloblastomas and induced by N-myc in sonic hedgehog-treated cerebellar neural precursors
.
Cancer Res
2009
;
69
:
3249
55
.
14.
Schwab
M
. 
MYCN in neuronal tumours
.
Cancer Lett
2004
;
204
:
179
87
.
15.
Cohn
SL
,
Pearson
AD
,
London
WB
,
Monclair
T
,
Ambros
PF
,
Brodeur
GM
, et al
The International Neuroblastoma Risk Group (INRG) classification system: an INRG Task Force report
.
J Clin Oncol
2009
;
27
:
289
97
.
16.
Weiss
WA
,
Aldape
K
,
Mohapatra
G
,
Feuerstein
BG
,
Bishop
JM
. 
Targeted expression of MYCN causes neuroblastoma in transgenic mice
.
EMBO J
1997
;
16
:
2985
95
.
17.
Chen
L
,
Iraci
N
,
Gherardi
S
,
Gamble
LD
,
Wood
KM
,
Perini
G
, et al
p53 is a direct transcriptional target of MYCN in neuroblastoma
.
Cancer Res
2010
;
70
:
1377
88
.
18.
Mestdagh
P
,
Boström
AK
,
Impens
F
,
Fredlund
E
,
Van Peer
G
,
De Antonellis
P
, et al
The miR-17-92 microRNA cluster regulates multiple components of the TGF-beta pathway in neuroblastoma
.
Mol Cell
2010
;
40
:
762
73
.
19.
Slack
A
,
Chen
Z
,
Tonelli
R
,
Pule
M
,
Hunt
L
,
Pession
A
, et al 
The p53 regulatory gene MDM2 is a direct transcriptional target of MYCN in neuroblastoma
.
Proc Natl Acad Sci U S A
2005
;
102
:
731
6
.
20.
www.illumina.com.
Illumina Sequencing Applications Platform. 2011 [cited. 2011]
. Available from: http://www.illumina.com/.
21.
Coarfa
C
,
Milosavljevic
A
. 
Pash 2.0: scaleable sequence anchoring for next-generation sequencing technologies
.
Pac Symp Biocomput
2008
;
102
13
.
22.
Ahn
HW
,
Morin
RD
,
Zhao
H
,
Harris
RA
,
Coarfa
C
,
Chen
ZJ
, et al
MicroRNA transcriptome in the newborn mouse ovaries determined by massive parallel sequencing
.
Mol Hum Reprod
2010
;
16
:
463
71
.
23.
Go. Genboree website [Internet]. Bioinformatics Research Laboratory, Baylor College of Medicine; 2011 [updated 2010 Oct; cited 2010 Nov]
. Available from: http://www.genboree.org/java-bin/login.jsp.
24.
Vasudevan
SA
,
Shang
X
,
Chang
S
, et al
Neuroblastoma-derived secretory protein is a novel secreted factor overexpressed in neuroblastoma
.
Mol Cancer Ther
2009
;
8
:
2478
89
.
25.
Barbieri
E
,
Mehta
P
,
Chen
Z
,
Zhang
L
,
Slack
A
,
Berg
S
, et al
MDM2 inhibition sensitizes neuroblastoma to chemotherapy-induced apoptotic cell death
.
Mol Cancer Ther
2006
;
5
:
2358
65
.
26.
Wei
JS
,
Johansson
P
,
Chen
QR
,
Song
YK
,
Durinck
S
,
Wen
X
, et al
microRNA profiling identifies cancer-specific and prognostic signatures in pediatric malignancies
.
Clin Cancer Res
2009
;
15
:
5560
8
.
27.
Wei
JS
,
Greer
BT
,
Westermann
F
,
Steinberg
SM
,
Son
CG
,
Chen
QR
, et al
Prediction of clinical outcome using gene expression profiling and artificial neural networks for patients with neuroblastoma
.
Cancer Res
2004
;
64
:
6883
91
.
28.
Leung
AK
,
Sharp
PA
. 
MicroRNA functions in stress responses
.
Mol Cell
2010
;
40
:
205
15
.
29.
Mestdagh
P
,
Van Vlierberghe
P
,
De Weer
A
,
Muth
D
,
Westermann
F
,
Speleman
F
, et al
A novel and universal method for microRNA RT-qPCR data normalization
.
Genome Biol
2009
;
10
:
R64
.
30.
Mudd
SH
. 
Hypermethioninemias of genetic and non-genetic origin: a review
.
Am J Med Genet C Semin Med Genet
2011
;
157
:
3
32
.
31.
Sawada
S
,
Kinjo
T
,
Makishi
S
,
Tomita
M
,
Arasaki
A
,
Iseki
K
, et al
Downregulation of citrin, a mitochondrial AGC, is associated with apoptosis of hepatocytes
.
Biochem Biophys Res Commun
2007
;
364
:
937
44
.
32.
Sippel
M
,
Rajala
R
,
Korhonen
L
,
Bornhauser
B
,
Sokka
AL
,
Naito
M
, et al
Dexamethasone regulates expression of BRUCE/Apollon and the proliferation of neural progenitor cells
.
FEBS Lett
2009
;
583
:
2213
7
.
33.
Aharinejad
S
,
Andrukhova
O
,
Lucas
T
,
Zuckermann
A
,
Wieselthaler
G
,
Wolner
E
, et al
Programmed cell death in idiopathic dilated cardiomyopathy is mediated by suppression of the apoptosis inhibitor Apollon
.
Ann Thorac Surg
2008
;
86
:
109
14
.
34.
Fuchs
D
,
Christofferson
R
,
Stridsberg
M
,
Lindhagen
E
,
Azarbayjani
F
. 
Regression of orthotopic neuroblastoma in mice by targeting the endothelial and tumor cell compartments
.
J Transl Med
2009
;
7
:
16
.
35.
Henriksson
KC
,
Almgren
MA
,
Thurlow
R
,
Varki
NM
,
Chang
CL
. 
A fluorescent orthotopic mouse model for reliable measurement and genetic modulation of human neuroblastoma metastasis
.
Clin Exp Metast
2004
;
21
:
563
70
.
36.
Brodeur
GM
,
Seeger
RC
,
Schwab
M
,
Varmus
HE
,
Bishop
JM
. 
Amplification of N-myc in untreated human neuroblastomas correlates with advanced disease stage
.
Science
(
New York, NY
) 
1984
;
224
:
1121
4
.
37.
Cohn
SL
,
London
WB
,
Huang
D
,
Katzenstein
HM
,
Salwen
HR
,
Reinhart
T
, et al
MYCN expression is not prognostic of adverse outcome in advanced-stage neuroblastoma with nonamplified MYCN
.
J Clin Oncol
2000
;
18
:
3604
13
.
38.
Gartel
AL
,
Ye
X
,
Goufman
E
,
Shianov
P
,
Hay
N
,
Najmabadi
F
et al 
Myc represses the p21(WAF1/CIP1) promoter and interacts with Sp1/Sp3
.
Proc Natl Acad Sci U S A
2001
;
98
:
4510
5
.
39.
Gartel
AL
. 
A new mode of transcriptional repression by c-myc: methylation
.
Oncogene
2006
;
25
:
1989
90
.
40.
Murphy
DM
,
Buckley
PG
,
Bryan
K
,
Das
S
,
Alcock
L
,
Foley
NH
, et al
Global MYCN transcription factor binding analysis in neuroblastoma reveals association with distinct E-box motifs and regions of DNA hypermethylation
.
PloS One
2009
;
4
:
e8154
.
41.
Marshall
GM
,
Gherardi
S
,
Xu
N
,
Neiron
Z
,
Trahair
T
,
Scarlett
CJ
, et al
Transcriptional upregulation of histone deacetylase 2 promotes Myc-induced oncogenic effects
.
Oncogene
2010
;
29
:
5957
68
.
42.
Iraci
N
,
Diolaiti
D
,
Papa
A
,
Porro
A
,
Valli
E
,
Gherardi
S
, et al
A SP1/MIZ1/MYCN repression complex recruits HDAC1 at the TRKA and p75NTR promoters and affects neuroblastoma malignancy by inhibiting the cell response to NGF
.
Cancer Res
2011
;
71
:
404
12
.
43.
Schweigerer
L
,
Breit
S
,
Wenzel
A
,
Tsunamoto
K
,
Ludwig
R
,
Schwab
M
. 
Augmented MYCN expression advances the malignant phenotype of human neuroblastoma cells: evidence for induction of autocrine growth factor activity
.
Cancer Res
1990
;
50
:
4411
6
.
44.
Fotsis
T
,
Breit
S
,
Lutz
W
,
Rössler
J
,
Hatzi
E
,
Schwab
M
, et al
Down-regulation of endothelial cell growth inhibitors by enhanced MYCN oncogene expression in human neuroblastoma cells
.
Eur J Biochem
1999
;
263
:
757
64
.
45.
Lutz
W
,
Fulda
S
,
Jeremias
I
,
Debatin
KM
,
Schwab
M
. 
MycN and IFNgamma cooperate in apoptosis of human neuroblastoma cells
.
Oncogene
1998
;
17
:
339
46
.
46.
Van Maerken
T
,
Ferdinande
L
,
Taildeman
J
,
Lambertz
I
,
Yigit
N
,
Vercruysse
L
, et al
Antitumor activity of the selective MDM2 antagonist nutlin-3 against chemoresistant neuroblastoma with wild-type p53
.
J Natl Cancer Instit
2009
;
101
:
1562
74
.
47.
Alam
G
,
Cui
H
,
Shi
H
,
Yang
L
,
Ding
J
,
Mao
L
, et al
MYCN promotes the expansion of Phox2B-positive neuronal progenitors to drive neuroblastoma development
.
Am J Pathol
2009
;
175
:
856
66
.
48.
Jackson
A
,
Linsley
PS
. 
The therapeutic potential of microRNA modulation
.
Discov Med
2010
;
9
:
311
8
.
49.
Seto
AG
. 
The road toward microRNA therapeutics
.
Int J Biochem Cell Biol
2010
;
42
:
1298
305
.
50.
Website
POO
[Internet]. Oncogenomics Data center; host site for genomic profiling data of pediatric cancers [cited 2010 Nov]
. Available from: http://pob.abcc.ncifcrf.gov/cgi-bin/JK.

Supplementary data