Abstract A08: Using directed differentiation of human pluripotent stem cells and gene expression profiling to characterize the cell of origin of neuroblastoma

Purpose: Neuroblastoma (NB) is one of the most common childhood tumors, accounting for ~15% of all pediatric cancer deaths. Improved neuroblastoma disease modeling is needed to improve our understanding and ultimately lead to the development of more effective therapies. NB is an embryonal tumor of the sympathetic nervous system believed to arise from sympathoadrenal (SA) precursor cells belonging to the primitive neural crest. Because the neural crest is a transient embryologic structure, there have been few studies comparing NB tumors to normal human neural crest stem cells (NCSC). However, recent advances have enabled the in vitro differentiation of neural crest cells from pluripotent stem cells. We compared gene expression profiles of pluripotent stem cells, NCSC, and NB tumors to characterize the gene expression profile of the putative cell of origin of NB as the first step in the development of a human NCSC based model of NB. Methods: Human embryonic and induced pluripotent stem (iPS) cells were differentiated to early NCSC and SA precursors in vitro using modifications of previously published protocols. Specifically, undifferentiated human embryonic stem cell GFP reporter and iPS cell lines were induced to differentiate into SA cells in a sequential, stepwise manner over the course of ~4 weeks. After differentiation, SA precursors were identified by expression of the lineage-specific transcription factors ASCL1 and PHOX2B, as well as other markers including tyrosine hydroxylase (TH) and dopamine beta-hydroxylase (DBH). We subsequently compared the gene expression profiles of these cells to early, multi-potent neural crest stem cells, autonomic and enteric neurons, 12 NB cell lines and 39 non-MYCN amplified primary NB tumors (3 stage 1, 5 stage 2, 3 stage 3, 24 stage 4, and 4 stage 4S) using Illumina HT-12 bead arrays. These profiles were also compared to the gene expression profiles of laser-capture microdissected fetal adrenal neuroblasts previously reported by De Preter et. al. (Genome Biology 2006, 7:R84). Results were normalized using quantile normalization and analyzed with principle components analysis (PCA); differential expression analysis was performed using the Linear Models for Microarray Data (Limma) method. Pathway enrichment analysis was performed using DAVID annotation tools and cross-platform normalization was performed using the CONOR package in the R software program. All analyses were corrected for multiple comparisons and results