Understanding differences and the complex interplay between tumors and healthy normal tissue is essential in finding effective treatments and understanding the genetic basis of cancers. We use a variety of techniques and approaches to study somatic structural variation, the non-coding genome, long non-coding RNAs (lncRNAs), whole proteomes of tumors, and tumor evolution.
Somatic structural variation (SV) in neuroblastoma
Neuroblastoma tumors harbor few recurrent mutations in protein coding genes compared to adult cancers. However, many neuroblastomas exhibit gains and losses of large segments of DNA. These copy number alterations have prognostic significance, and some are used clinically to assign treatment. Copy number alterations represent just one form of structural variation (SV). SVs can also include segments of DNA that are inverted or translocated to alter gene function and promote cancer. With support from the NIH/NCI, our laboratory is studying SVs in tumor and matched normal DNA of neuroblastoma patients using large-scale data from whole genome sequencing (WGS) and single nucleotide polymorphism (SNP) arrays coupled with experimental follow-up to assess biological relevance. In our germline SV studies of blood-derived normal DNA (see Genetic Predisposition), we have identified both common and rare CNVs as genetic risk factors for neuroblastoma. In tumor genomes, chromothripsis (“shattering” of DNA involving potentially thousands of clustered SVs in localized or confined genomic regions) occurs in 19% of high-risk cases. Specific genes are also targeted by SVs in neuroblastoma tumors. Somatic SVs upstream of the telomerase gene (TERT) are observed in 25% of high-risk cases; these SVs drive aberrant TERT expression via enhancer hijacking and influence telomere maintenance to promote cancer. Our work has also identified SHANK2 on chromosome 11 as a novel tumor suppressor gene whose normal function is to promote cellular differentiation; however, the gene is disrupted by recurrent SVs in neuroblastoma tumors. We are currently exploring long-read sequencing technologies to enhance SV detection and extending this work to additional patient cohorts and relapsed tumors.
Role of the non-coding genome in neuroblastoma
Like many childhood cancers, neuroblastoma exhibits a paucity of recurrent somatic mutations within protein coding regions of genes in comparison with adult onset cancers. The observed paucity of recurrent somatic mutations in protein coding genes suggests that the non-coding genome, and alterations of these regions, may play important roles in childhood cancer development. However, only a relatively small number of non-coding variants have been demonstrated to promote tumorigenesis, and the contribution of noncoding RNA, such as long noncoding RNAs (lncRNAs), in cancer remains poorly understood. Therefore, we are implementing integrative (epi)genomic approaches to identify non-coding variants/mutations affecting key regulatory regions and altering transcriptional, epigenetic, or 3D architectural programs in neuroblastoma that may be exploitable for risk prediction and/or therapeutic intervention. Rationale for this approach is provided by our collaborative efforts identifying an LMO1 super-enhancer variant promoting tumor development. In addition, we and others have observed somatically acquired non-coding structural variants (SVs) upstream of the telomerase gene (TERT) in approximately 25% of high-risk neuroblastoma cases; these SVs can drive aberrant TERT expression via enhancer hijacking.
Long non-coding RNAs (lncRNAs) in childhood cancers
Over 70% of the transcribed genome is non-coding, including an estimated 58,648 long noncoding RNAs (lncRNAs). LncRNAs are aberrantly expressed in almost all cancer types and their tissue specific expression makes them attractive therapeutic targets. Functional studies have revealed their role in tumorigenesis stems from active regulation of genes important for cell cycle control, cell survival, and pluripotency. Mechanistically, lncRNAs can function by interacting with transcription factors to modulate target gene expression and with DNA to facilitate chromatin interactions. We are characterizing lncRNA expression in a pediatric cancer cohort (n=1030) comprised of six distinct histotypes using RNA-sequencing from the NCI-TARGET project to determine what role they may play in childhood cancers. With a multi-omic approach, we integrate epigenomic and DNA sequencing data to elucidate the regulation of lncRNAs. We will then validate and assess the biological function of lncRNAs, with a primary focus on neuroblastoma. Our goal is to uncover novel lncRNA biology and potentially discover new cancer targets.
Defining the neuroblastoma proteome
A major complication in the realm of pediatric cancer research has been the fact that RNA levels do not always correspond to protein levels. Moreover, while RNA can be easily measured, even from a single cell, protein estimates are much less reliable and require a far greater sample pool. To better quantify average protein expression levels in neuroblastoma, we are quantifying the whole proteomes of 39 commonly used human-derived neuroblastoma cell lines using mass spectrometry. The transcriptome of these 39 cell lines has already beed assessed using RNA sequencing (Transcriptomic profiling of 39 commonly-used neuroblastoma cell lines. Harenza JL, et al., Sci Data. 2017 Mar 28;4:170033). These resources will inform future experiments and enhance our understanding of the complex relationship between transcription and translation in this important childhood cancer.