Genome-wide mapping of MYCN transcription factor binding sites and identification of clinically relevant targets in neuroblastoma.
Neuroblastoma is a highly genetically heterogeneous form of paediatric cancer with clinical behaviour which can range fiom localised tumours with good prognosis, to widespread metastatic disease with rapid clinical progression and death. High level genornic amplification of the MYCN gene occurs in approximately 20 to 25% of cases and this aberration is the most important genetic indicator of poor clinical outcome. Despite the use of intensive multimodal therapy, patients with MYCN amplified neuroblastoma tumours have less than a 30% chance of 5- year survival. Thus the characterisation of MYCN binding in the NB genome and identification of downstream MYCN targets is critically important for the development of alternative treatment regimens and improving patient survival. The aim of this study was to elucidate MYCN genome-wide binding patterns and to determine its functional effects on downstream target genes, and thereby to identify MYCN function in neuroblastoma disease pathogenesis. To achieve this we performed chIP-chip analysis of MYCN binding sites in neuroblastoma cell lines expressing both high and low levels of the MYCN protein in order to characterise the binding behaviour of MYCN in these states. Our analysis has revealed that MYCN preferentially binds to the non-canonical E-box sequence of CATGTG and to additional motifs when it is overexpressed; potentially indicating that MYCN binding becomes less specific when it is highly abundant. Through gene expression profiling, we also describe a set of novel MYCN target genes which are enriched for genes related to cell cycle regulated upregulated in response to increased levels of MYCN in both tumours and cell lines. Surprisingly, we also observed a striking correlation between MYCN binding and DNA hypermethylation status at interlintragenic sites and within the promoter regions of genes, indicating a novel role for MYCN in the regulation of chromatin structure. We surmised that this association was likely due to the interaction of a methyl-binding protein. ChIP-chip analysis of MECP2 binding sites within the neuroblastoma genome revealed that -70% of hypermethylated MYCN binding sites were also occupied by MECP2. The association was not limited to regions of hypermethylation however, as MeCP2 was also found to co-occupy sites with MYCN which were free of detectable levels of hypermethylation. In conclusion, this study characterises the genome-wide association of MYCN to genomic loci in cell lines expressing both high and low levels of MYCN and characterizes the E-Box motifs to which this oncogenic protein binds. We also describe a set of novel MYCN target genes which are upregulated in response to u increased levels of MYCN which may also play a role in neuroblastoma tumourigenesis. Additionally, we have observed MYCN binding to regions of DNA hypermethylation at a higher than expected frequency and identify the methyl binding protein MECP2 as an interaction partner of MYCN at these loci. Future transcription factor method development may allow a similar such analysis to be performed in primary tumours, thus elucidating finther the mechanism of action of h4YCN in NB pathogenesis.