Abstract B72: Dissecting the heterogeneity of metastatic neuroblastoma cells by single-cell RNA-seq

Abstract

Abstract Introduction: The major cause of death in neuroblastoma (NB) is metastatic relapse, and most therapies do not specifically target metastases. We previously developed a metastatic mouse model using intracardiac injection of tagged human SK-N-AS NB cells, followed by multiple rounds of in vivo selection from bone and brain metastatic sites. Gene expression profiles identified novel genes and pathways dysregulated in metastatic NB cells that, when genetically or pharmacologically manipulated, affected metastasis. A subset of genes up- or downregulated in the enhanced metastatic cells was used to generate a metastatic gene signature that predicted patient survival. We have now asked (1) whether there is a transcriptionally distinct subpopulation of cells in the parental population that selectively metastasizes to bone and brain, and (2) if there are transcriptional differences that can account for the ability of the enhanced metastatic cells to hone to bone/bone marrow vs. brain. Methods: We used 10X Genomics single-cell RNA sequencing (scRNA-seq) to transcriptionally profile parental cells and two in vivo-selected cell lines that showed enhanced metastatic spread to bone/bone marrow (B5) and brain (BR2). Approximately 5,000 genes and 10,000 transcripts per cell were sequenced, with the data analyzed using a unique analysis pipeline incorporating extensive data quality analysis with visualization and clustering methods using evidence-based parameter selection. Genes with high variance were used to compute principal components as inputs to visualize cells in two dimensions (t-SNE plots) in clusters. Results: scRNA-seq data showed that the parental cells clustered separately from B5 and BR2 cells, with clusters in the parental cells that expressed genes enriched in the B5 and BR2 cells that we previously reported drove migration/invasion in vitro and/or metastasis in vivo, including sphingosine kinase 1. These results suggest that subpopulations of the parental cells with enhanced metastatic gene expression are selected in vivo for honing to metastatic sites. We identified genes enriched in B5 and BR2 cells encoding cell surface proteins, chemokines including CCL2, and transcription factors including Msx1 involved in metastasis of other cancers, that we will assess for roles in NB metastasis. While the B2 and BR5 cells were transcriptionally similar to each other, we did identify genes highly expressed in B2 subclusters that are associated with bone cancers including FGF5, and that are highly expressed in BR2 subclusters that are associated with brain cancers including RARRES2. These findings suggest that transcriptional differences between B5 and BR2 cells may influence metastasis to the bone or brain. Conclusion: The identification of genes that are differentially expressed in the more metastatic subpopulations will contribute to our understanding of the molecular mechanisms governing metastases and identify novel targetable pathways that were not detected by profiling bulk tumor populations. Citation Format: Alice R. Shan, Alexander Gont, David R. Kaplan, Meredith S. Irwin. Dissecting the heterogeneity of metastatic neuroblastoma cells by single-cell RNA-seq [abstract]. In: Proceedings of the AACR Special Conference on the Advances in Pediatric Cancer Research; 2019 Sep 17-20; Montreal, QC, Canada. Philadelphia (PA): AACR; Cancer Res 2020;80(14 Suppl):Abstract nr B72.