Altered methylation of imprinted genes in neuroblastoma: implications for prognostic refinement

Abstract

Abstract Background Neuroblastoma (NB) is a complex disease, and the current understanding of NB biology is limited. Deregulation in genomic imprinting is a common event in malignancy. Since imprinted genes play crucial roles in early fetal growth and development, their role in NB pathogenesis could be suggested. Methods We examined alterations in DNA methylation patterns of 369 NB tumours at 49 imprinted differentially methylated regions (DMRs) and assessed its association with overall survival probabilities and selected clinical and genomic features of the tumours. In addition, an integrated analysis of DNA methylation and allele-specific copy number alterations (CNAs) was performed, to understand the correlation between the two molecular events. Results Several imprinted regions with aberrant methylation patterns in NB were identified. Regions that underwent loss of methylation in > 30% of NB samples were DMRs annotated to the genes NDN, SNRPN, IGF2, MAGEL2 and HTR5A and regions with gain of methylation were NNAT, RB1 and GPR1. Methylation alterations at six of the 49 imprinted DMRs were statistically significantly associated with reduced overall survival: MIR886, RB1, NNAT/BLCAP, MAGEL2, MKRN3 and INPP5F. RB1, NNAT/BLCAP and MKRN3 were further able to stratify low-risk NB tumours i.e. tumours that lacked MYCN amplification and 11q deletion into risk groups. Methylation alterations at NNAT/BLCAP, MAGEL2 and MIR886 predicted risk independently of MYCN amplification or 11q deletion and age at diagnosis. Investigation of the allele-specific CNAs demonstrated that the imprinted regions that displayed most alterations in NB tumours harbor true epigenetic changes and are not result of the underlying CNAs. Conclusions Aberrant methylation in imprinted regions is frequently occurring in NB tumours and several of these regions have independent prognostic value. Thus, these could serve as potentially important clinical epigenetic markers to identify individuals with adverse prognosis. Incorporation of methylation status of these regions together with the established risk predictors may further refine the prognostication of NB patients.