Mapping the Single-Cell Differentiation Landscape of Osteosarcoma

Abstract

Purpose: The genetic intratumoral heterogeneity observed in human osteosarcomas poses challenges for drug development and the study of cell fate, plasticity, and differentiation, which are processes linked to tumor grade, cell metastasis, and survival.

Experimental design: To pinpoint errors in osteosarcoma differentiation, we transcriptionally profiled 31,527 cells from a tissue-engineered model that directs mesenchymal stem cells toward adipogenic and osteoblastic fates. Incorporating preexisting chondrocyte data, we applied trajectory analysis and non-negative matrix factorization to generate the first human mesenchymal differentiation atlas.

Results: This "roadmap" served as a reference to delineate the cellular composition of morphologically complex osteosarcoma tumors and quantify each cell's lineage commitment. Projecting a bulk RNA-sequencing osteosarcoma dataset onto this roadmap unveiled a correlation between a stem-like transcriptomic phenotype and poorer survival outcomes.

Conclusions: Our study quantifies osteosarcoma differentiation and lineage, a prerequisite to better understanding lineage-specific differentiation bottlenecks that might someday be targeted therapeutically.

Figure 1:. Single-Cell Sequencing Reveals Intratumor Heterogeneity in Patient-Derived Xenograft Models of Osteosarcoma.

A: Single-cell gene expression UMAP of three OS PDX samples. B: Louvain clustering identified 15 clusters of cells in the three PDXs. C: Heatmap of gene expression markers (normalized by maximum across all cells) of all individual cells. Genes were grouped according to specific lineage markers or biological processes. D: Average cluster gene expression (normalized by maximum across all clusters) of the set of top 3 differentially expressed genes in each cluster. E: Dot plot of pathway analysis scores for each cluster.

Figure 2:. Constructing a Differentiation Atlas of the Mesenchymal Tissue Landscape.

A: Schematic of datasets utilized to construct an atlas of the Mesenchymal Tissue Landscape (MTL). Osteogenic and adipogenic lineages were experimentally generated over time, culturing on hydrogels of varying stiffness to influence differentiation trajectory. The chondrogenic lineage was sourced from a publicly available dataset (GSE160625), which measured a time course of chondrogenesis in cultured chondroprogenitor cells treated with a combination of TGF-β3 and C59. These two datasets were integrated to construct a mesenchymal differentiation map containing three lineages. B: UMAP of integrated MTL colored by experimental condition specifying the differentiation protocol in which the cells were cultured. C: UMAP colored by experimental time, scaled to the endpoint of each experiment (d21 for osteo, d14 for adipo, d42 for chondro). D: UMAP with designated clusters along three distinct lineages. Clusters were manually annotated as mesenchymal stem cell (MSC), osteogenic (O1-O2), adipogenic (A1-A4), chondroprogenitor (CP), chondrogenic (C1-C3). E: Violin plots of marker genes for early and late stages of each lineage showing distinct temporal patterns.

Figure 3:. Archetype Analysis Defines Signatures of Distinct Mesenchymal Differentiation States.

A: Heatmap of archetype score estimates split by MTL cluster. Archetype scores were estimated by normalized nonnegative matrix factorization. B: Compositions of each MTL cluster based on each cell’s highest expressed archetype signature. C: Average archetype time courses stratified by cell lineage. D: Heatmap of representative gene-archetype correlations. Four representative genes were selected among the top Pearson correlates of each archetype based on their known biological relevance. The complete list of correlates is provided in Supplemental Table S4. On left, each archetype was annotated by lineage(s) and peak time.

Figure 4:. Archetype composition of osteosarcoma tumor samples and PDX models.

A: Single-cell archetype score heatmap of 3 PDX OS models, with hierarchical clustering to accentuate groups of similar cells (dendrogram not shown). Row annotation on the left indicates the lineage of each archetype (same as Fig. 3D). Column annotation indicates the pathologist subtype label based on the predominant cell type. B: Single-cell archetype score heatmap of 11 human OS tumor samples. Similar annotation as panel A. C: Compositions of each PDX based on the maximum archetype score of each cell. D: Compositions of each OS tumor.

Figure 5:. Well-differentiated signature is associated with improved survival in osteosarcoma.

A: Heatmap of archetype expression in TARGET-OS (N=88); each column is a different patient’s OS bulk sample. B: Forest plot of the associated Cox regression model with reported hazard ratio (HR) and confidence interval (CI) using the expressed MTL archetypes as regressors. Significance was determined by a multivariable Cox PH test (N=85). C: Kaplan-Meier plot stratified by the sample-specific Archetype 3 (differentiated) scores grouped as higher or lower than the median. Significance was determined by a univariable Cox PH test (N=85).