Noncoding regulatory mutations contribute to the aberrant gene expression in neuroblastoma.

<jats:p>Comprehensive analyses of whole-genome and exome sequencing data from high-risk neuroblastoma tumors have revealed relatively few recurrent, clinically actionable protein-coding driver mutations at initial diagnosis. This observation suggests that noncoding genetic variation, which can alter regulatory sequences like promoters, enhancers and insulators and influence gene expression, may play a critical role in neuroblastoma tumorigenesis. By integrating allele-specific expression (ASE) with somatic mutation profiles from two independent neuroblastoma patient cohorts, we identified a significant and reproducible enrichment of noncoding single-nucleotide variants (SNVs) within regulatory regions of neuroblastoma-specific ASE (NB-ASE) genes. Notably, 63% of these variants disrupted transcription factor binding sites (TFBSs), with FOXJ2 being the most frequently affected transcription factor (TF) across both cohorts. Supporting a functional link between FOXJ2 TFBS SNVs and gene expression dysregulation, NB-ASE genes harboring these variants were significantly enriched among FOXJ2 co-expression partners. These findings nominate FOXJ2 dysregulation via TFBS mutations as a potentially crucial molecular mechanism contributing to aberrant gene expression profiles of neuroblastoma. To prioritize high-impact regulatory mutations associated with NB-ASE genes, we also performed extensive deep learning-based functional predictions and identified 297 TFBS mutations predicted to significantly alter chromatin state. Among these were variants predicted to deactivate enhancers regulating the tumor suppressor genes CASZ1 and PRDM11, both detected in tumors lacking copy-number alteration at the locus, suggesting an alternative, copy number-independent mechanism of downregulation. Collectively, our findings demonstrate that integrating ASE with somatic mutation profiles is a powerful strategy for detecting and interpreting regulatory variations in cancer genomes.</jats:p>