Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

MYCN/c-MYC-induced microRNAs repress coding gene networks associated with poor outcome in MYCN/c-MYC-activated tumors


Increased activity of MYC protein-family members is a common feature in many cancers. Using neuroblastoma as a tumor model, we established a microRNA (miRNA) signature for activated MYCN/c-MYC signaling in two independent primary neuroblastoma tumor cohorts and provide evidence that c-MYC and MYCN have overlapping functions. On the basis of an integrated approach including miRNA and messenger RNA (mRNA) gene expression data we show that miRNA activation contributes to widespread mRNA repression, both in c-MYC- and MYCN-activated tumors. c-MYC/MYCN-induced miRNA activation was shown to be dependent on c-MYC/MYCN promoter binding as evidenced by chromatin immunoprecipitation. Finally, we show that pathways, repressed through c-MYC/MYCN miRNA activation, are highly correlated to tumor aggressiveness and are conserved across different tumor entities suggesting that c-MYC/MYCN activate a core set of miRNAs for cooperative repression of common transcriptional programs related to disease aggressiveness. Our results uncover a widespread correlation between miRNA activation and c-MYC/MYCN-mediated coding gene expression modulation and further substantiate the overlapping functions of c-MYC and MYCN in the process of tumorigenesis.


Activated signaling of MYC gene-family members (c-MYC, MYCN, MYCL) is a hallmark of many cancer types and contributes to tumorigenesis by promoting cell growth, metastasis, angiogenesis and genomic instability (Adhikary and Eilers, 2005). MYC genes can activate and repress transcription of target genes through distinct mechanisms. Transcriptional activation is well understood and depends on the binding of MYC to its consensus DNA recognition sequence 5′-IndexTermCACGTG-3′ or E-box. In contrast, the mechanism of transcriptional repression is more obscure and seems independent of E-box binding. One mechanism relies on the binding of MYC with the cofactor Miz-1, which tethers MYC to promoters (Kleine-Kohlbrecher et al., 2006) whereas another involves the induction of gene silencing through MYC-mediated promoter hypermethylation (Gartel, 2006).

Recently, the MYCN/c-MYC transcriptional network has been shown to also include microRNAs (miRNAs). These small non-coding RNAs have an effect on virtually every aspect of tumorigenesis and function as negative regulators of messenger RNA (mRNA) levels and translation (Bartel, 2009). Some miRNAs, such as those belonging to the oncogenic miR-17-92 cluster, have been shown to be important players in MYCN/c-MYC signaling (O'Donnell et al., 2005; Dews et al., 2006; Chang et al., 2008; Fontana et al., 2008; Schulte et al., 2008; Northcott et al., 2009). However, a more general insight in the relationship between miRNAs and mRNAs within the MYCN/c-MYC transcriptional network remains to be examined.

Neuroblastoma qualifies as an excellent model system to study the MYCN/c-MYC transcriptional network. Activated MYCN/c-MYC signaling is a hallmark of poor prognosis for neuroblastoma tumors (Fredlund et al., 2008) and can be caused either by MYCN amplification (Seeger et al., 1985) or increased c-MYC expression in stage 4 MYCN non-amplified neuroblastoma tumors (Westermann et al., 2008). This increase in MYCN/c-MYC expression results in the activation of coding target genes related to poor patient prognosis independently of MYCN amplification (Fredlund et al., 2008; Westermann et al., 2008).

In this study, we identified a comprehensive miRNA signature induced by increased MYCN/c-MYC signaling in neuroblastoma and show that MYCN and c-MYC have overlapping functions in the induction of this miRNA signature. Most importantly, we provide evidence that MYCN/c-MYC-activated miRNAs are correlated to widespread mRNA downregulation pointing at MYCN/c-MYC-regulated gene expression modulation beyond transcriptional control. Transcriptional programs that are affected in this way underlie distinct prognostic subgroups of both neuroblastoma as well as other tumor entities with activated MYCN/c-MYC signaling.


MYCN/c-MYC microRNA signature delineation

For the identification of miRNAs differentially expressed between MYCN-amplified (MNA) on the one hand, and MYCN single-copy low-risk (SL) or high-risk (SH) tumors on the other (see Supplementary Material for definition), two independent unpublished patient cohorts (n=56 and n=39) were analyzed for a total of 430 miRNAs. When comparing MNA with SL tumors, we found 49 miRNAs significantly differentially expressed in both cohorts. Comparing MNA with SH resulted in 12 differentially expressed miRNAs. In total, we identified 50 unique miRNAs (16 upregulated and 34 downregulated) differentially expressed between MNA and MYCN single-copy tumors in the two independent data sets (Supplementary Table S1). Supporting our result, differential expression has earlier been shown for several of the listed miRNAs in other cancer entities with increased MYCN/c-MYC signaling (Supplementary Table S2). Cluster analysis based on this 50 miRNA signature on both cohorts distinguishes three major clusters with a clear separation between MNA and MYCN single-copy tumors (Figure 1a, Supplementary Figure S1). Interestingly, the 50 miRNA signature also separates SH from SL tumors despite the fact that differential miRNA expression between these groups was not selected for. This additional separation suggests that MYCN/c-MYC signaling rather than MYCN levels alone underlies the differential expression of this miRNA signature. Indeed, c-MYC expression is significantly increased in SH tumors (Supplementary Figure S2), confirming the inverse relationship between MYCN and c-MYC (Vandesompele et al., 2003), and has been shown to drive MYCN/c-MYC mRNA target gene expression in these tumors (Westermann et al., 2008). In addition, other genetic factors than increased MYCN/c-MYC activity could also contribute to the observed miRNA expression differences. MNA tumors are frequently associated with 1p deletions (80.6% in MNA subgroup and 17.9% in SH subgroup) while SH tumors are associated with 11q deletions (89.3% in SH subgroup and 32.3% in MNA subgroup). Factors in either of these regions affecting miRNA expression can contribute to the observed signature. The three main clusters, defined by the miRNA signature, significantly correlate to overall (P<0.001) and event-free (P<0.001) patient survival confirming its ability to define subgroups with differential prognosis (Figures 1b and c). We identified six typical patterns of expression within the 50 miRNA signature when comparing MNA, SH and SL tumors (Figure 2). Four of these are linked to increased MYCN/c-MYC signaling with miRNA expression being intermediate in SH tumors (Figures 2a and b) or either high (or low) in both MNA and SH tumors (Figures 2c and d). The remaining two patterns represent miRNAs with a differential expression that is restricted to MNA tumors alone (Figures 2e and f). These miRNA expression patterns are similar to those that have been described for protein coding MYCN/c-MYC target genes (Westermann et al., 2008).

Figure 1

A MYCN/c-MYC microRNA (miRNA) signature delineates prognostic neuroblastoma subgroups. (a) Hierarchical clustering of 95 primary neuroblastoma tumors based on the expression of the MYCN/c-MYC miRNA signature. MYCN-amplified (MNA), MYCN single-copy low-risk (SL) and MYCN single-copy high-risk (SH) tumors are indicated in black, light gray and dark gray, respectively. (b, c) Kaplan–Meier analysis for overall (OS) and event-free (EFS) survival of patients according to the three main clusters in (a).

Figure 2

MicroRNA (miRNA) expression patterns. Overview of the six main miRNA expression patterns (af) observed in MYCN-amplified (MNA), MYCN single-copy low-risk (SL) and MYCN single-copy high-risk (SH) tumors. Expression patterns are represented schematically using vertical bars. The height of each bar reflects the relative miRNA expression in each group. Black, gray and white bars represent significant expression differences. For each pattern, one representative miRNA is shown. (a) Upregulated in MNA, intermediate in SH, (b) downregulated in MNA, intermediate in SH, (c) upregulated in MNA and SH, (d) downregulated in MNA and SH, (e) upregulated in MNA, (f) downregulated in MNA.

Further inspection of the classification of MNA, SH and SL tumors revealed some misclassified cases. Two MYCN single-copy tumors cluster within the MNA cluster. For one of these tumors, fluorescence in situ hybridization analysis revealed low level gain of c-MYC in combination with c-MYC translocation (Supplementary Figure S3A). Fluorescence in situ hybridization results for the t(8;14) translocation described in Burkitt lymphoma were negative suggesting a translocation partner other than chromosome 14 (data not shown). These aberrations at the c-MYC locus ultimately result in an increased expression of c-MYC (12.8-fold) and of key protein coding c-MYC target genes (Supplementary Figure S3B and C). To our knowledge, this is the first report of a primary neuroblastoma tumor with c-MYC translocation. In addition, this tumor presented with genetic abnormalities (such as 1p deletion and 17q gain, data not shown) typically found in MNA tumors thus supporting the notion that MYCN and c-MYC have overlapping functions in vivo and in the process of tumorigenesis (Malynn et al., 2000). c-MYC translocation provides an additional mechanism for neuroblastoma cells to obtain high level c-MYC expression, together with the earlier reported c-MYC amplification in SJNB-12 and MP-N-TS neuroblastoma cells (Saito-Ohara et al., 2003; Van Roy et al., 2006). The second MYCN single-copy tumor did not show any aberrations at the c-MYC locus suggesting that other mechanisms are responsible for increased activity of MYCN/c-MYC in this tumor. Both of the patients died of disease. Finally, eight MNA tumors clustered outside the MNA cluster. All but one grouped within the SH cluster, presumably because of the overlapping miRNA expression patterns between MNA and SH tumors (Figure 2).

To further examine the miRNA signature, we also examined a panel of MNA (n=5), MYCN single-copy (n=5) and c-MYC-amplified (n=1) neuroblastoma cell lines. For several miRNAs, expression levels in the c-MYC-amplified cell line SJNB-12 were similar to those in the MNA cell lines (Supplementary Figure S4). These data are in line with the miRNA expression pattern that was observed in the c-MYC- translocated primary neuroblastoma tumor and further support the apparent overlapping functions for MYCN and c-MYC with regard to miRNA regulation in these cancer cells.

MYCN/c-MYC binds to microRNA promoters

To confirm that the 16 upregulated miRNAs are indeed responsive to MYCN/c-MYC, we profiled their expression in two independent MYCN model systems. On MYCN activation, 10 out of 16 upregulated miRNAs showed a >1.5-fold induction in expression in at least one model system whereas 6 out of 16 were induced in two model systems (Supplementary Figure S5). Varying results between the between model systems and primary tumor cells might be due to cell-specific differences in target genes and cell-dependent variations in the composition of MYCN transcriptional complexes (Cappellen et al., 2007). Differences between the model systems could be because of the level of induction of the MYCN gene, culture effects or treatment-related events. To analyze miRNA regulation by MYCN/c-MYC, we assessed binding of MYCN or c-MYC using chromatin immunoprecipitation (ChIP)-chip to genomic regions of the upregulated miRNAs. Besides binding of MYCN or c-MYC, we also defined the epigenetic marker status of these genomic regions in six neuroblastoma cell lines. We used H3K4me3 for active and H3K27me3 for repressed regions. In addition, H3K36me3 was used for transcript elongation. For 10 out of 11 miRNAs that are broadly covered by the array probe set, we observed MYCN or c-MYC binding as well as an epigenetic marker state that is in line with transcription of the respective miRNAs (Table 1). For two miRNAs (mir-601 and mir-610) that are not covered by the array probe set, we observed binding of MYCN and c-MYC to the promoter regions of the hosting genes (DENND1A and KIF18A). Transcription of the genes in neuroblastoma cell lines indirectly suggests that the miRNAs within these genes are also transcribed (data not shown). Our results confirm c-MYC/MYCN binding to the promoters of miR-17-92 (O'Donnell et al., 2005; Fontana et al., 2008; Mestdagh et al., 2009), miR-181a (Mestdagh et al., 2009), and miR-9 (L Ma et al., submitted, personal communication) and suggest that the promoters of miR-15b, miR-130a and miR-214 are also bound and activated by MYCN/c-MYC (Supplementary Figure S6). These results are in line with the hypothesis that, for the majority of the miRNAs, increased expression in tumors with activated MYCN/c-MYC is a direct effect of MYCN/c-MYC binding and transactivation.

Table 1 MYCN/c-MYC binding to miRNA promoters

A microRNA target identification strategy

For high-throughput miRNA target identification, we devised a strategy based on the integration of a unique mRNA (obtained through Affymetrix exon gene expression analysis) and miRNA expression data set together with miRNA target recognition characteristics. In a subset of 40 tumors, we calculated correlations between the expression of each of the 50 miRNAs and 15 000 mRNAs. Candidate target mRNAs were defined as being significantly negatively correlated to the expression of the miRNA (Spearman's rank, Benjamini Hochberg multiple testing correction) with a rho value below −0.5 and with the occurrence of at least one 3′UTR seed. Three different seed matches were considered: at least one 7mer seed (7mer-A1 or 7mer-m8), at least one 8mer seed or at least one conserved 7 or 8mer seed. To confirm that these criteria preferentially select miRNA target genes, we established a new model system with tetracycline inducible miR-17-92 expression in the SHEP neuroblastoma cell line (SHEP-TR-miR-17-92) (Supplementary Figure S7). Expression of the miR-17-92 cluster is activated by MYCN/c-MYC and five miRNAs belonging to this cluster (miR-18a, miR-18a*, miR-19a, miR-20a and miR-92a) are among the 16 upregulated miRNAs in our signature. For three of these (miR-19a, miR-20a and miR-92a) we can define candidate miRNA targets using our selection strategy on the mRNA and miRNA data from the 40 tumors. We then compared differentially expressed mRNAs in treated and untreated SHEP-TR-miR-17-92 cells with our predicted targets for miR-19a, miR-20a and miR-92a using gene set enrichment analysis (Subramanian et al., 2005). Three custom gene set enrichment analysis gene lists were established based on the different seed matches (7mer, 8mer and conserved 7 or 8mer) of miR-19a, miR-20a and miR-92a. Interestingly, two gene lists (8mer, P=0.01 and conserved 7 or 8mer, P=0.005) showed significant enrichment among the mRNAs that were downregulated on miR-17-92 induction in the SHEP-TR-mir-17-92 cells hereby validating our selection strategy (Supplementary Figure S8). The fact that only predicted targets with 8mer seeds and conserved seeds are enriched confirms the higher efficacy of these seeds (Baek et al., 2008). In addition, several predicted targets defined by this strategy, such as TGFBR2, ATXN1 and HIPK3, have already been reported as direct targets of miR-20a, miR-19a and miR-92a, respectively (Volinia et al., 2006; Landais et al., 2007; Lee et al., 2008). We also compared our target predictions for miR-20a in neuroblastoma to a set of validated miR-20a targets obtained by ribonucleoprotein immunoprecipitation-gene Chip in Hodgkin lymphoma cells (Tan et al., 2009). From the 37 miR-20a 8mer targets that we identified, 14 were reported by Tan et al. (2009) (Fisher's exact test, P<0.001). These results indicate that, although some of the predicted targets could be indirect, our miRNA target identification strategy definitely selects a substantial number of direct miRNA target genes.

miRNA activation as a mechanism for MYCN/c-MYC-induced mRNA downregulation

We next set out the search for target mRNAs downstream of the MYCN/c-MYC miRNA signature. If a miRNA affects target gene expression, the negatively correlated mRNAs should be enriched for 3′UTR miRNA seeds. Strikingly, 3′UTR seed enrichment was found for almost all upregulated miRNAs (14 out of 16) but not for the downregulated miRNAs (1 out of 34) (Fisher's exact test, P<0.0001). For the upregulated miRNAs, cumulative distribution plots show a significant difference between the Spearman's rank rho value distribution (representing mRNA:miRNA correlation) of mRNAs with a 3′UTR seed compared with mRNAs without a 3′UTR seed (Kolmogorov–Smirnov, P<0.001) (Figure 3a). For the downregulated miRNAs this is not the case (Figure 3b). These findings show that MYCN/c-MYC-activated miRNAs rather than MYCN/c-MYC repressed miRNAs have a widespread effect on differential mRNA target gene expression. This observation therefore suggests that activated miRNA expression could serve as a mechanism for MYCN/c-MYC-induced mRNA repression. To further support this hypothesis, we established a core set of predicted miRNA targets using our target identification strategy and compared the number of predicted targets between the up- and downregulated miRNAs. On average, upregulated miRNAs have four times more predicted targets when considering 7mer seeds and 6 to 18 times more predicted targets when considering 8mer or conserved seeds respectively (Figure 3c). This huge discrepancy in predicted target genes between up- and downregulated miRNAs thus supports the notion that MYCN/c-MYC-activated miRNAs predominantly drive differential gene expression in high-risk neuroblastomas. Moreover, it confirms that activated miRNA expression is correlated to widespread mRNA repression. In line with this, MYCN and c-MYC downregulated genes should contain 3′UTR seeds for MYCN/c-MYC-activated miRNAs. To evaluate this, known MYCN and c-MYC down- and upregulated genes were extracted from the MYCNot database ( and the MYC target gene database, (Zeller et al., 2003) respectively. The total occurrence of seeds from all 16 upregulated miRNAs were calculated and compared between down- and upregulated genes. As expected, MYCN downregulated genes were significantly enriched for seeds from the upregulated miRNAs (Fisher's exact test, P<0.001) (Figure 3d) confirming the observed correlation between miRNA activation and MYCN target gene repression. Strikingly, this enrichment was also apparent for c-MYC downregulated genes (Fisher's exact test, P<0.01) (Figure 3d). Although entries in MYCNot mainly represent MYCN targets in neuroblastoma, the MYC target gene database contains c-MYC targets across multiple other tumor entities. Despite the fact that the upregulated miRNAs were selected from primary neuroblastoma tumors, their seed signature is also manifested in other tumors with activated MYCN/c-MYC signaling, underlining the general relevance of this MYCN/c-MYC signature for tumor biology.

Figure 3

MicroRNA (miRNA) seed enrichment and target identification. (a, b) Cumulative distribution of Spearman's rank rho values, representing mRNA:miRNA correlation, for mRNAs with no seed (dark gray) and mRNAs with a seed (light gray) shown for a representative MYCN/c-MYC-activated miRNA and MYCN/c-MYC repressed miRNA, respectively. (c) Average number of identified targets for up- and downregulated miRNAs when considering 7mer, 8mer or conserved 3′UTR seeds. (d) Fraction of genes listed in the MYCNot database and the MYC target gene database (MYCdb) containing a 3′UTR seed. Results for upregulated (UP) and downregulated (DN) genes are listed as separate bars.

MYCN/c-MYC-activated microRNAs act in concert

The entire network of MYCN/c-MYC upregulated miRNAs and their predicted target mRNAs are significantly enriched for genes listed in the neuroblastoma gene server NBGS (Fisher's exact test, P<0.001), a database containing genes reported as differentially expressed in neuroblastoma tumors ( Center for Medical Genetics, Ghent, Belgium). In addition, it reveals that a significant number of mRNAs are putatively regulated by multiple miRNAs (Figure 4). Over 30% of the mRNA targets are predicted to be under the regulation of two or more miRNAs indicating a concerted mode of action of miRNAs toward their predicted target genes. The significance of this cooperative regulation between miRNAs and mRNAs in the network were evaluated with respect to repeated sampling of an equally sized list of randomly selected mRNAs or miRNAs (Supplementary Figure S9 and Supplementary Material). We consistently observed a higher degree of cooperative regulation between MYCN-activated miRNAs and their targets as compared with random selections of mRNAs/miRNAs (Kolmogorov–Smirnov, P<0.0001) suggesting that cooperative regulation is a likely feature within the network of MYCN-activated miRNAs. Cooperative regulation of gene expression by co-expressed miRNAs has been suggested before and could serve as a mechanism to fine tune gene expression (Sampson et al., 2007). Alternatively, it might be that one binding site is insufficient for proper repression. An increasing number of 3′UTR binding sites have indeed been shown to be more effective toward target regulation (Selbach et al., 2008).

Figure 4

Interaction network of MYCN/c-MYC-activated microRNAs (miRNAs) and target mRNAs. MYCN/c-MYC-activated miRNAs (red triangles) are connected to their predicted target mRNAs (dots). mRNAs that are targeted by multiple miRNAs from within the network are indicated in black, mRNAs targeted by only one miRNA in gray. Only miRNAs with at least three mRNA targets are shown. A full colour version of this figure is available at the Oncogene journal online.

MicroRNA target gene expression correlates to patient survival

As MYCN/c-MYC-activated miRNAs appear essential in mediating MYCN/c-MYC-induced transcriptional repression, we asked whether the predicted mRNA targets of these miRNAs would have any prognostic value. The combined activity of the predicted 8mer miRNA targets (n=193) was evaluated using a rank-based pathway score in two independent neuroblastoma microarray data sets (Oberthuer et al., 2006; Wang et al., 2006). To assess the effect on patient outcome, samples were divided into quartiles according to the pathway activity score (Fredlund et al., 2008). Kaplan–Meier analysis of the Oberthuer data based on these quartiles revealed a significant correlation to both event-free (EFS, P<0.001) and overall survival (OS, P<0.001) (Figures 5a and b). Patients in the first quartile, representing tumors with the lowest expression of predicted 8mer miRNA targets, have a particular poor prognosis as compared with those in the fourth quartile. Of interest, 48% of the tumors in the first quartile (n=66) were MYCN single-copy confirming that MYCN/c-MYC signaling rather than MYCN amplification underlies the poor outcome of these patients (Fredlund et al., 2008; Westermann et al., 2008). Moreover, 8mer miRNA-mediated target repression but not MYCN amplification status was an independent factor in a multivariate Cox regression model for EFS and OS (Supplementary Table S3). In the Wang data set, Kaplan–Meier analysis confirms the observed correlation to OS (P<0.001) and EFS (P<0.001) (Figures 5c and d). In this study, 8mer miRNA target signaling was no independent factor for OS but was for EFS (Supplementary Table S3). A possible explanation for this discrepancy is that the Wang data set is highly biased toward MNA high-risk patients. Taken together, these data clearly show that the process of mRNA repression, mediated by MYCN/c-MYC upregulated miRNAs, are highly correlated to tumor aggressiveness and ultimately patient survival.

Figure 5

MicroRNA (miRNA) target activity correlates to patient outcome in neuroblastoma. Pathway activity score, represented as quartiles, for predicted 8mer targets of MYCN/c-MYC-activated miRNAs significantly correlates to overall (OS) (a, c) and event-free (EFS) (b, d) patient survival in two independent data sets (Oberthuer, Wang). The lowest and highest pathway activity scores are represented by the first quartile (yellow) and fourth quartile (red), respectively. Cox-regression P-values are listed.

MYCN/c-MYC-activated microRNAs repress pathways affecting patient survival across tumor types

To gain more insight in the genetic programs underlying MYCN/c-MYC-activated miRNA signaling, we dissected the list of predicted 8mer and conserved targets into pathways using the Ingenuity Pathway Analysis software (Supplementary Table S4). The analyses showed that the largest fraction of predicted conserved targets appeared to be implicated in cell death and cancer-related processes and these genes were comparatively lower expressed in high-risk as compared with low-risk neuroblastoma tumors (data not shown). Other pathways known to be involved in neuroblastoma, such as CREB signaling (Jiang et al., 2008), CNTF signaling (Peterson and Bogenmann, 2004) and integrin signaling were also identified. Integrin signaling is of particular interest as decreased expression of integrins selectively enhances neuroblastoma survival and metastasis (Stupack et al., 2006). Both integrin receptors and downstream signaling molecules were among the predicted 7mer and 8mer targets of the MYCN/c-MYC-activated miRNAs. In addition, CAV1, a scaffolding protein linking integrin subunits to the tyrosine kinase FYN, is a direct MYC target (Park et al., 2001). Together, these findings suggest that integrin signaling is repressed by increased MYCN/c-MYC signaling. To assess whether integrin signaling intensity reflects neuroblastoma patient survival, a pathway activity score was calculated using all predicted 7mer targets from the integrin signaling pathway (n=19). Kaplan–Meier analysis of the Oberthuer and Wang data sets revealed a significant correlation to OS (Oberthuer: P<0.001, Wang: P<0.01) and EFS (Oberthuer: P<0.001, Wang: P<0.001) (Supplementary Figure S10). In a multivariate Cox regression model using integrin pathway activity quartiles and MYCN amplification status, low integrin signaling remained predictive of both OS and EFS in the Oberthuer data set and of EFS in the Wang data set (data not shown). In summary, these results suggest that miRNA controlled regulation of specific groups of mRNAs could serve as an additional mechanism of MYCN/c-MYC-induced oncogenicity.

As integrin signaling is (in part) MYCN/c-MYC regulated, we evaluated whether it would reflect patient survival in other tumor entities with increased MYCN/c-MYC signaling. One such entity is diffuse large B-cell lymphoma. Approximately, 15% of diffuse large B-cell lymphomas have rearrangements at the c-MYC locus resulting in c-MYC overexpression and poor patient survival (Kramer et al., 1998; Chang et al., 2000). The integrin signaling pathway activity score was calculated for a microarray data set of 255 diffuse large B-cell lymphomas (Lenz et al., 2008). In keeping with our observation in neuroblastoma, Kaplan–Meier analyses indicate that integrin signaling is proportionally correlated to overall patient survival (P<0.001) (Supplementary Figure S11). These results indicate that the integrin signaling pathway, which is negatively correlated to MYCN/c-MYC-activated miRNAs, is conserved between tumor types. Not only do c-MYC and MYCN activate a common set of miRNAs, these miRNAs appear to signal through the same pathways in two entirely different tumor cell types, lending support to a functional redundancy between the two transcription factors.


We have identified a miRNA signature representative for MYCN/c-MYC signaling in neuroblastoma tumors. Using ChIP-chip, we have shown binding of MYCN to the promoter region of several miRNAs suggesting a direct role for MYCN in the transcriptional regulation of these miRNAs. Additional experiments that assess reduced MYCN binding in cells with low MYCN expression are warranted to fully validate these findings. Our results further suggest that c-MYC and MYCN have overlapping functions for the induction of this miRNA signature. The functional overlap between MYCN and c-MYC is further substantiated by the identification of a c-MYC-translocated neuroblastoma tumor with a mRNA, miRNA and genomic profile that is typical for a MYCN-amplified tumor. It is also perfectly in line with previous reports on miRNAs commonly regulated by c-MYC and MYCN (Chen and Stallings, 2007; Chang et al., 2008; Sander et al., 2008; Schulte et al., 2008; Sun et al., 2008; Northcott et al., 2009) and with the observation that MYCN can functionally replace c-MYC in murine development (Malynn et al., 2000). Whether c-MYC is capable of replacing MYCN remains to be determined but could further corroborate these findings. MYCN appears to be a weak transcription factor as miRNA expression levels in MNA and SH tumors are equally high (or low) despite the fact that MYCN expression in MNA tumors is substantially higher than c-MYC expression in SH tumors.

To identify candidate miRNA target genes, we used an integrative approach based on negative correlation analysis between miRNA and mRNA expression in primary tumors, in combination with 3′UTR seed occurrence and a cellular model system. We showed that MYCN/c-MYC-activated miRNAs are correlated to widespread transcriptional repression whereas miRNAs that are downregulated in MYCN/c-MYC-activated tumors were not associated with transcriptional activation. This observation does not imply that mRNAs that do have a seed for downregulated miRNAs are irrelevant. Chang et al. have shown that c-MYC-induced miRNA downregulation can have a profound effect on tumorigenesis (Chang et al., 2008). In line with this, we observed a significant correlation between the predicted targets of the downregulated miRNAs and patient survival (data not shown) suggesting that miRNA downregulation is important in tumorigenesis. The mechanisms by which c-MYC and MYCN induce transcriptional repression are multiple and include binding to Miz-1 and induction of promoter hypermethylation (Gartel, 2006; Kleine-Kohlbrecher et al., 2006). Our results now suggest that MYCN/c-MYC-induced miRNA activation also contributes to coding gene repression. The observed correlations were experimentally verified in neuroblastoma tumors and in silico for published c-MYC and MYCN downregulated genes in different tumor entities suggesting a non-random and conserved feature of the MYCN/c-MYC network affecting specific collections of mRNAs. MYCN/c-MYC-induced miRNA activation has been shown to repress the expression of few coding genes, such as CDKN1A, BCL2L11 and E2F1 (O'Donnell et al., 2005; Fontana et al., 2008) whereas our findings suggest a widespread transcriptional repression. Although we have shown that the observed mRNA downregulation is miRNA dependent in a cellular model system, further experiments are needed to establish direct interactions between the MYCN/c-MYC-activated miRNAs and the repressed mRNAs.

Pathways that are repressed in this way were shown to be of clinical importance in different tumor entities with increased MYCN/c-MYC activity again confirming that MYC family members have overlapping functions that contribute to tumor aggressiveness. It also confirms that assessing entire signaling pathways, rather than individual genes, is highly informative with respect to cancer outcome prediction (Bild et al., 2006; Watters and Roberts, 2006; Liu and Ringner, 2007). For neuroblastoma in particular, combined analysis of multiple signaling networks has been successful and might benefit from the addition of other pathways, such as integrin signaling, to further increase prediction sensitivity (Fredlund et al., 2008). In addition, the observed similarities in miRNA expression between MNA and SH neuroblastoma tumors suggest that both tumor subgroups share a number of pathways deregulated through MYCN/c-MYC-mediated miRNA modulation. The identification of such pathways might prove useful in the search for novel therapeutic targets.

In conclusion, we uncovered a widespread correlation between miRNA activation and c-MYC/MYCN-mediated coding gene expression modulation and further substantiate the overlapping functions of c-MYC and MYCN in the process of tumorigenesis.

Materials and methods

Patient samples and cell lines

A total of 95 primary neuroblastoma tumor samples were collected at the Ghent University Hospital (Ghent, Belgium), the University Children's Hospital, Essen (Essen, Germany) and the Medical School of Valencia (Valencia, Spain) before therapeutic treatment. Patients were staged according to the International Neuroblastoma Staging System. Informed consent was obtained from the patients' relatives. Details on prognostic subgroups and neuroblastoma cell lines are listed in the Supplementary Material.

MicroRNA expression profiling

Total RNA was isolated using the miRNAeasy kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. MiRNA expression profiling and data normalization were performed as described earlier (Mestdagh et al., 2008, 2009).

Messenger RNA expression profiling

Total RNA was isolated from tumor samples and was hybridized to Human Exon 1.0 ST array (Affymetrix, Santa Clara, CA, USA) at the microarray facility of the University Hospital of Essen according to the manufacturer's protocol. Total RNA from tetracycline treated and untreated SHEP-TR-miR-17-92 cells were hybridized to Affy-Hu-Gene1.0ST oligonucleotide chips (Affymetrix) at the microarray facility of the Flanders Institute for Biotechnology (Leuven, Belgium).

Fluorescence in situ hybridization

Fluorescence in situ hybridization was performed according to Van Roy et al. (1994). The following probes were used: LSI MYC dual color, break-apart rearrangement probe and LSI IGH/MYC, CEP8 tri-color, dual fusion translocation probe (Abbott Molecular Products, Des Plaines, IL, USA).

Chromatin immunoprecipitation

Chromatin immunoprecipitation was performed as described earlier using 10 μg of MYCN or c-MYC antibodies (Westermann et al., 2008). For a detailed description see Supplementary Material.


Details on statistical procedures are described in Supplementary Material.

Conflict of interest

The authors declare no conflict of interest.


  1. Adhikary S, Eilers M . (2005). Transcriptional regulation and transformation by Myc proteins. Nat Rev Mol Cell Biol 6: 635–645.

    CAS  Article  Google Scholar 

  2. Baek D, Villen J, Shin C, Camargo FD, Gygi SP, Bartel DP . (2008). The impact of microRNAs on protein output. Nature 455: 64–71.

    CAS  Article  Google Scholar 

  3. Bartel DP . (2009). MicroRNAs: target recognition and regulatory functions. Cell 136: 215–233.

    CAS  Article  Google Scholar 

  4. Bild AH, Yao G, Chang JT, Wang Q, Potti A, Chasse D et al. (2006). Oncogenic pathway signatures in human cancers as a guide to targeted therapies. Nature 439: 353–357.

    CAS  Article  Google Scholar 

  5. Cappellen D, Schlange T, Bauer M, Maurer F, Hynes NE . (2007). Novel c-MYC target genes mediate differential effects on cell proliferation and migration. EMBO Rep 8: 70–76.

    CAS  Article  Google Scholar 

  6. Chang CC, Liu YC, Cleveland RP, Perkins SL . (2000). Expression of c-Myc and p53 correlates with clinical outcome in diffuse large B-cell lymphomas. Am J Clin Pathol 113: 512–518.

    CAS  Article  Google Scholar 

  7. Chang TC, Yu D, Lee YS, Wentzel EA, Arking DE, West KM et al. (2008). Widespread microRNA repression by Myc contributes to tumorigenesis. Nat Genet 40: 43–50.

    CAS  Article  Google Scholar 

  8. Chen Y, Stallings RL . (2007). Differential patterns of microRNA expression in neuroblastoma are correlated with prognosis, differentiation, and apoptosis. Cancer Res 67: 976–983.

    CAS  Article  Google Scholar 

  9. Dews M, Homayouni A, Yu D, Murphy D, Sevignani C, Wentzel E et al. (2006). Augmentation of tumor angiogenesis by a Myc-activated microRNA cluster. Nat Genet 38: 1060–1065.

    CAS  Article  Google Scholar 

  10. Fontana L, Fiori ME, Albini S, Cifaldi L, Giovinazzi S, Forloni M et al. (2008). Antagomir-17-5p abolishes the growth of therapy-resistant neuroblastoma through p21 and BIM. PLoS One 3: e2236.

    Article  Google Scholar 

  11. Fredlund E, Ringner M, Maris JM, Pahlman S . (2008). High Myc pathway activity and low stage of neuronal differentiation associate with poor outcome in neuroblastoma. Proc Natl Acad Sci USA 105: 14094–14099.

    CAS  Article  Google Scholar 

  12. Gartel AL . (2006). A new mode of transcriptional repression by c-myc: methylation. Oncogene 25: 1989–1990.

    CAS  Article  Google Scholar 

  13. Jiang M, Zhu K, Grenet J, Lahti JM . (2008). Retinoic acid induces caspase-8 transcription via phospho-CREB and increases apoptotic responses to death stimuli in neuroblastoma cells. Biochim Biophys Acta 1783: 1055–1067.

    CAS  Article  Google Scholar 

  14. Kleine-Kohlbrecher D, Adhikary S, Eilers M . (2006). Mechanisms of transcriptional repression by Myc. Curr Top Microbiol Immunol 302: 51–62.

    CAS  Google Scholar 

  15. Kramer MH, Hermans J, Wijburg E, Philippo K, Geelen E, van Krieken JH et al. (1998). Clinical relevance of BCL2, BCL6, and MYC rearrangements in diffuse large B-cell lymphoma. Blood 92: 3152–3162.

    CAS  Google Scholar 

  16. Landais S, Landry S, Legault P, Rassart E . (2007). Oncogenic potential of the miR-106-363 cluster and its implication in human T-cell leukemia. Cancer Res 67: 5699–5707.

    CAS  Article  Google Scholar 

  17. Lee Y, Samaco RC, Gatchel JR, Thaller C, Orr HT, Zoghbi HY . (2008). miR-19, miR-101 and miR-130 co-regulate ATXN1 levels to potentially modulate SCA1 pathogenesis. Nat Neurosci 11: 1137–1139.

    CAS  Article  Google Scholar 

  18. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H et al. (2008). Stromal gene signatures in large-B-cell lymphomas. N Engl J Med 359: 2313–2323.

    CAS  Article  Google Scholar 

  19. Liu Y, Ringner M . (2007). Revealing signaling pathway deregulation by using gene expression signatures and regulatory motif analysis. Genome Biol 8: R77.

    Article  Google Scholar 

  20. Malynn BA, de Alboran IM, O'Hagan RC, Bronson R, Davidson L, DePinho RA et al. (2000). N-myc can functionally replace c-myc in murine development, cellular growth, and differentiation. Genes Dev 14: 1390–1399.

    CAS  PubMed  PubMed Central  Google Scholar 

  21. Mestdagh P, Feys T, Bernard N, Guenther S, Chen C, Speleman F et al. (2008). High-throughput stem-loop RT-qPCR miRNA expression profiling using minute amounts of input RNA. Nucleic Acids Res 36: e143.

    Article  Google Scholar 

  22. Mestdagh P, Van Vlierberghe P, De Weer A, Muth D, Westermann F, Speleman F et al. (2009). A novel and universal method for microRNA RT-qPCR data normalization. Genome Biol 10: R64.

    Article  Google Scholar 

  23. Northcott PA, Fernandez LA, Hagan JP, Ellison DW, Grajkowska W, Gillespie Y et al. (2009). 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 69: 3249–3255.

    CAS  Article  Google Scholar 

  24. O'Donnell KA, Wentzel EA, Zeller KI, Dang CV, Mendell JT . (2005). c-Myc-regulated microRNAs modulate E2F1 expression. Nature 435: 839–843.

    CAS  Article  Google Scholar 

  25. Oberthuer A, Berthold F, Warnat P, Hero B, Kahlert Y, Spitz R et al. (2006). Customized oligonucleotide microarray gene expression-based classification of neuroblastoma patients outperforms current clinical risk stratification. J Clin Oncol 24: 5070–5078.

    CAS  Article  Google Scholar 

  26. Park DS, Razani B, Lasorella A, Schreiber-Agus N, Pestell RG, Iavarone A et al. (2001). Evidence that Myc isoforms transcriptionally repress caveolin-1 gene expression via an INR-dependent mechanism. Biochemistry 40: 3354–3362.

    CAS  Article  Google Scholar 

  27. Peterson S, Bogenmann E . (2004). The RET and TRKA pathways collaborate to regulate neuroblastoma differentiation. Oncogene 23: 213–225.

    CAS  Article  Google Scholar 

  28. Saito-Ohara F, Imoto I, Inoue J, Hosoi H, Nakagawara A, Sugimoto T et al. (2003). PPM1D is a potential target for 17q gain in neuroblastoma. Cancer Res 63: 1876–1883.

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Sampson VB, Rong NH, Han J, Yang Q, Aris V, Soteropoulos P et al. (2007). MicroRNA let-7a down-regulates MYC and reverts MYC-induced growth in Burkitt lymphoma cells. Cancer Res 67: 9762–9770.

    CAS  Article  Google Scholar 

  30. Sander S, Bullinger L, Klapproth K, Fiedler K, Kestler HA, Barth TF et al. (2008). MYC stimulates EZH2 expression by repression of its negative regulator miR-26a. Blood 112: 4202–4212.

    CAS  Article  Google Scholar 

  31. Schulte JH, Horn S, Otto T, Samans B, Heukamp LC, Eilers UC et al. (2008). MYCN regulates oncogenic microRNAs in neuroblastoma. Int J Cancer 122: 699–704.

    CAS  Article  Google Scholar 

  32. Seeger RC, Brodeur GM, Sather H, Dalton A, Siegel SE, Wong KY et al. (1985). Association of multiple copies of the N-myc oncogene with rapid progression of neuroblastomas. N Engl J Med 313: 1111–1116.

    CAS  Article  Google Scholar 

  33. Selbach M, Schwanhausser B, Thierfelder N, Fang Z, Khanin R, Rajewsky N . (2008). Widespread changes in protein synthesis induced by microRNAs. Nature 455: 58–63.

    CAS  Article  Google Scholar 

  34. Stupack DG, Teitz T, Potter MD, Mikolon D, Houghton PJ, Kidd VJ et al. (2006). Potentiation of neuroblastoma metastasis by loss of caspase-8. Nature 439: 95–99.

    CAS  Article  Google Scholar 

  35. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102: 15545–15550.

    CAS  Article  Google Scholar 

  36. Sun Y, Wu J, Wu SH, Thakur A, Bollig A, Huang Y et al. (2008). Expression profile of microRNAs in c-Myc induced mouse mammary tumors. Breast Cancer Res Treat 118: 185–196.

    Article  Google Scholar 

  37. Tan LP, Seinen E, Duns G, de Jong D, Sibon OC, Poppema S et al. (2009). A high throughput experimental approach to identify miRNA targets in human cells. Nucleic Acids Res 37: e137.

    Article  Google Scholar 

  38. Van Roy N, Laureys G, Cheng NC, Willem P, Opdenakker G, Versteeg R et al. (1994). 1;17 translocations and other chromosome 17 rearrangements in human primary neuroblastoma tumors and cell lines. Genes Chromosomes Cancer 10: 103–114.

    CAS  Article  Google Scholar 

  39. Van Roy N, Vandesompele J, Menten B, Nilsson H, De Smet E, Rocchi M et al. (2006). Translocation-excision-deletion-amplification mechanism leading to nonsyntenic coamplification of MYC and ATBF1. Genes Chromosomes Cancer 45: 107–117.

    CAS  Article  Google Scholar 

  40. Vandesompele J, Edsjo A, De Preter K, Axelson H, Speleman F, Pahlman S . (2003). ID2 expression in neuroblastoma does not correlate to MYCN levels and lacks prognostic value. Oncogene 22: 456–460.

    CAS  Article  Google Scholar 

  41. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F et al. (2006). A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci USA 103: 2257–2261.

    CAS  Article  Google Scholar 

  42. Wang Q, Diskin S, Rappaport E, Attiyeh E, Mosse Y, Shue D et al. (2006). Integrative genomics identifies distinct molecular classes of neuroblastoma and shows that multiple genes are targeted by regional alterations in DNA copy number. Cancer Res 66: 6050–6062.

    CAS  Article  Google Scholar 

  43. Watters JW, Roberts CJ . (2006). Developing gene expression signatures of pathway deregulation in tumors. Mol Cancer Ther 5: 2444–2449.

    CAS  Article  Google Scholar 

  44. Westermann F, Muth D, Benner A, Bauer T, Henrich KO, Oberthuer A et al. (2008). Distinct transcriptional MYCN/c-MYC activities are associated with spontaneous regression or malignant progression in neuroblastomas. Genome Biol 9: R150.

    Article  Google Scholar 

  45. Zeller KI, Jegga AG, Aronow BJ, O'Donnell KA, Dang CV . (2003). An integrated database of genes responsive to the Myc oncogenic transcription factor: identification of direct genomic targets. Genome Biol 4: R69.

    Article  Google Scholar 

Download references


We acknowledge Q Wang and J Maris for providing the neuroblastoma microarray data set. This research was funded by the Gent University Research Fund (BOF 01D31406 to PM, BOF 01F07207 to FP, BOF 01Z09407 to J Vandesompele), the Fondation pour la recherche Nuovo-Soldati (J Vermeulen), RD06/0020/0102 from RTICC/ISCIII to RN, the Fund for Scientific Research (grant number: G.0198.08 and 31511809), the Belgian Kid's Fund and the Stichting tegen Kanker. KDP is a post-doctoral researcher with the Fund for Scientific Research-Flanders. We acknowledge the support of the European Community under the FP6 (project: STREP: EET-pipeline, number: 037260).

Author information



Corresponding author

Correspondence to J Vandesompele.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Mestdagh, P., Fredlund, E., Pattyn, F. et al. MYCN/c-MYC-induced microRNAs repress coding gene networks associated with poor outcome in MYCN/c-MYC-activated tumors. Oncogene 29, 1394–1404 (2010).

Download citation


  • MYCN
  • c-MYC
  • microRNA
  • neuroblastoma

Further reading


Quick links