Combined analysis of RNA-sequence and microarray data reveals effective metabolism-based prognostic signature for neuroblastoma.

The relationship between metabolism reprogramming and neuroblastoma (NB) is largely unknown. In this study, one RNA-sequence data set (n = 153) was used as discovery cohort and two microarray data sets (n = 498 and n = 223) were used as validation cohorts. Differentially expressed metabolic genes were identified by comparing stage 4s and stage 4 NBs. Twelve metabolic genes were selected by LASSO regression analysis and integrated into the prognostic signature. The metabolic gene signature successfully stratifies NB patients into two risk groups and performs well in predicting survival of NB patients. The prognostic value of the metabolic gene signature is also independent with other clinical risk factors. Nine metabolism-related long non-coding RNAs (lncRNAs) were also identified and integrated into the metabolism-related lncRNA signature. The lncRNA signature also performs well in predicting survival of NB patients. These results suggest that the metabolic signatures have the potential to be used for risk stratification of NB. Gene set enrichment analysis (GSEA) reveals that multiple metabolic processes (including oxidative phosphorylation and tricarboxylic acid cycle, both of which are emerging targets for cancer therapy) are enriched in the high-risk NB group, and no metabolic process is enriched in the low-risk NB group. This result indicates that metabolism reprogramming is associated with the progression of NB and targeting certain metabolic pathways might be a promising therapy for NB.