Single-Cell RNA Sequencing of Ewing Sarcoma Tumors Demonstrates Transcriptional Heterogeneity and Clonal Evolution

Abstract

Purpose: Ewing sarcoma is the second most common bone cancer in children, accounting for 2% of pediatric cancer diagnoses. Patients who present with metastatic disease at the time of diagnosis have a dismal prognosis compared with the >70% 5-year survival of those with localized disease. Novel therapeutic approaches that can impact metastatic disease are desperately needed, as well as a deeper understanding of the heterogeneity of Ewing sarcoma tumors.

Experimental design: In this study, we utilized single-cell RNA sequencing to characterize the transcriptional landscape of primary Ewing sarcoma tumors and the surrounding tumor microenvironment in a cohort of seven untreated patients with Ewing sarcoma, as well as in circulating tumor cells (CTC). A potential CTC therapeutic target was evaluated through immunofluorescence of fixed CTCs from a separate cohort.

Results: Primary tumor samples demonstrate a heterogeneous transcriptional landscape with several conserved gene expression programs, including those composed of genes related to proliferation and Ewing sarcoma gene targets, which were found to correlate with overall survival. Copy-number analysis identified subclonal evolution within patients prior to treatment. Analyses of the immune microenvironment reveal an immunosuppressive microenvironment with complex intercellular communication among the tumor and immune cells. Single-cell RNA sequencing and immunofluorescence of CTCs at the time of diagnosis identified TSPAN8 as a potential therapeutic target.

Conclusions: Ewing sarcoma tumors demonstrate significant transcriptional heterogeneity as well as a complex immunosuppressive microenvironment. This work evaluates several proposed targets that warrant further exploration as novel therapeutic strategies.

Figure 1.. scRNA-seq of EwS primary tumors identifies several cell types including cancer cells.

UMAP plot labeled by cell type (A) or sample source (B). Expression of canonical marker genes across cell types are shown as a dotplot (C) or UMAP (D). Inferred copy number alterations in the EwS cells by sample with all other cell types serving as control (E). Expression of select genes within EwS tumor cells from specific patients and aggregate cell types (F).

Figure 2.. Identification of gene expression programs in EwS primary tumors and correlation to patient outcome.

Merged gene expression programs (GEPs) analyzed using over-representation analysis with the FDR of select gene sets displayed by program (A). Heatmap of select genes with merged GEP groups down sampled to 100 cells, with yellow indicating higher expression (B). Heatmaps displaying the expression of the genes in the EWS (C) and Proliferation (D) merged gene expression programs in EwS tumors from GSE17679. Sample scores are plotted at the top with a median split used to classify samples as low (green) or high (red) for each program. Kaplan-Meier curves with a log-rank test comparing the overall survival between the low and high groups identified from EWS (E) or Proliferation (F) gene expression programs in GSE17679 and the validation EwS cohort, GSE63157 (G,H).

Figure 3.. Subclones were identified in EwS primary tumors by analysis of copy number alterations.

EwS primary tumors 005 (A-C) and 061 (D-F) displayed subclones. Inferred copy number alterations (CNAs) in subclones of each sample (A,D). Phylogenic trees inferring evolution of the subclones based on the CNAs with the percentage of cells in each arm indicated (B,E). Subclones displayed as UMAP plots (C,F).

Figure 4.. T and myeloid cell subtypes in EwS primary tumors.

Characterization of T (A-C) and myeloid (D-F) cell subtypes in primary tumors. UMAP plots showing the identified subtypes (A,D). Gene expression defining each subtype (B,E). Proportion of the subtypes across each sample with number of cells indicated (C,F).

Figure 5.. Circulating EwS cells identified in the blood of Patient 005.

Cell types identified in the blood samples displayed as a UMAP (A) and discriminating cell markers as a dotplot (B). Proportion of cell types in each sample (C). Gene expression of EwS-enriched genes displayed as a heatmap in blood (left) and primary tumor cell types from patient 005 (D). Volcano plot highlighting genes significantly enriched in circulating compared to primary tumors cells within 005 (E). ChIP-seq binding profile of FLI1 or input DNA near TSPAN8 in the A673 cell line (F). TSPAN8 staining in EwS cell lines and CTCs (G). The 005_CTC sample is the same source as the 005_CTC scRNA-seq sample. P values from t test: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***).