Background

Transcriptional dysregulation is a driver in the development of malignancy. MYC family members, e.g., MYCN, MYC, and MYCL, are powerful oncogenes that promote tumorigenesis in a wide variety of human tissues [1–6]. Elevated expression of MYC family members can occur through several mechanisms, including amplification of the gene locus, chromosomal translocation with enhancer hijacking, and new enhancer formation through DNA variation [1,6–10]. High expression of MYC family members is associated with rapidly proliferating tumors and poor patient outcomes [11–14]. MYC proteins are helix-loop-helix (bHLH) transcription factors that form heterodimers with MAX and, in normal cells, primarily bind the core promoters of actively transcribed genes [15–17]. Studies in multiple cancer types have demonstrated that high levels of MYC or MYCN increase the binding of MYC proteins to the enhancers and promoters of actively transcribed genes, effectively regulating the entire expressed genome [18–20]. Thus, the predominant effect of high MYC expression levels by cancer cells is increased expression levels of genes actively transcribed by the cells, commonly called “transcriptional amplification” [21].

Transcriptional amplification of the inherent cellular gene expression program through high expression of MYC family members plays a critical role in tumor initiation for many tissue types, including pediatric neuroblastoma. Neuroblastoma is a tumor of the peripheral sympathetic nervous system that originates from neuroblasts of the migratory neural crest. Approximately 20% of high-risk tumors have genomic amplification of MYCN, often as double-minute chromatin bodies, leading to hundreds of copies of MYCN per cell [2]. In an independent subset of high-risk neuroblastoma, MYC is highly expressed via enhancer hijacking [6]. In our previous study [22], we showed that MYCN-amplified neuroblastoma cells undergo growth arrest and differentiation upon treatment with retinoic acid, which leads to a change in the core regulatory transcriptional circuitry with changes in cell state, accompanied by downregulation of MYCN expression. In this study, we assessed differential gene expression between neuroblastoma samples with high and low MYCN expression due to retinoic acid treatment.

To analyze cell states and transcriptional programs, standard RNA sequencing measures the relative abundance of RNAs in the tissue type of interest through massive parallel sequencing of RNA-derived cDNA [23]. Standard analysis of genome-wide RNA expression between two or more samples generally assumes that cells from each sample produce equivalent amounts of RNA per cell. However, previous studies have demonstrated that this analysis algorithm can lead to erroneous interpretations when MYC or MYCN acts as a transcriptional amplifier [18,19,24]. In the case of cancer cells with high levels of expression of MYC or MYCN, each cell produces 2–3× more RNA per expressed gene [18,21,24]. Retinoic acid–treated neuroblastoma cells downregulated MYCN, so we expected 2–3 fold less total RNA per cell [18], potentially confounding genome-wide expression analysis and necessitating a tailored experimental strategy. Here, we describe a step-by-step experimental and computational approach using spiked-in RNA standards from the External RNA Controls Consortium (ERCC) that allowed us to compare each gene's expression in a pool of cells in the context of the effect of downregulated MYCN. Our method can be used for a wide variety of other cancer tissues in which transcriptional amplification through any MYC family member comes into play and profoundly impacts the assessment of cancer cell gene expression signatures.