COMBINED COPY NUMBER AND GENE EXPRESSION PROFILING IN HIGH RISK NEUROBLASTOMA: A CCLG PILOT BIOLOGICAL STUDY


Year:

Session type:

Alem Gabriel1,Lindi Chen1,Laura Evans2,Sirintra Nakjang1,Nick Bown3,Daniel Williamson1,Deborah Tweddle1
1Newcastle University,2Dana-Farber Cancer Institute,3Northern Genetics Service

Abstract

Background

High-risk neuroblastoma (HRNBL) remains one of the major challenges in paediatric oncology with current 5 year survivorship of <40%. MYCN amplification occurs in half of high risk patients, but is not prognostic within this group. Identification of new biomarkers in HRNBL may lead to better patient stratification and improved outcome. Aims: 1) To compare gene expression profiling by RNA-seq with the Human Transcriptome Array (HTA) 2) To compare single nucleotide polymorphism (SNP) arrays with previously reported multiplex ligation-dependent amplification (MLPA) in HRNBL.

Method

RNA and genomic DNA were extracted from 6 frozen HRNBL, RNA-seq (Illumina TruSeq stranded mRNA library kit) and Affymetrix CytoscanHD SNP-arrays performed for copy number (CN) analysis. From 5 cases RNA was extracted from paraffin tissue for HTA Array (Affymetrix).

Results

5/6 cases were MYCN amplified and 2/5 MYCN amplified cases were ALK co-amplified. Relative fold changes in gene expression in ALK-amplified vs. non-ALK amplified (n=4) cases using HTA arrays detected 54 differentially expressed genes (DEGs). In contrast RNA-seq analysis of the same genomes detected 401 DEGs (adjusted p value < 0.05 at least 2 fold change) reflecting increased sensitivity of RNA-seq. These DEGs are mostly involved in differentiation pathways (including MAPK and WNT pathways) which fits with current understanding of neuroblastoma. CN analysis using SNP arrays on six HRNBL genomes detected novel and recurrent CN abnormalities (CNA), 65 CNAs in total with mean CNAs of 10.8 per case. Our pre-existing MLPA CN data on 4 of these patients showed relative concordance. Our integrated genomewide analysis of expression and CN in a MYCN amplified vs. non-MYCN amplified case (n=2) revealed genomewide expression levels consistent with changes observed in CN.

Conclusion

Our pilot dataset illustrates the feasibility of an integrated genomic approach for frozen HRNBL cases which in a larger study could identify biomarkers associated with relapse and drug resistance.