Single-Nuclei Multiome (ATAC + Gene Expression) Sequencing of a Primary Canine Osteosarcoma Elucidates Intra-Tumoral Heterogeneity and Characterizes the Tumor Microenvironment

Abstract

Osteosarcoma (OSA) is a highly aggressive bone tumor primarily affecting pediatric or adolescent humans and large-breed dogs. Canine OSA shares striking similarities with its human counterpart, making it an invaluable translational model for uncovering the disease's complexities and developing novel therapeutic strategies. Tumor heterogeneity, a hallmark of OSA, poses significant challenges to effective treatment due to the evolution of diverse cell populations that influence tumor growth, metastasis, and resistance to therapies. In this study, we apply single-nuclei multiome sequencing, encompassing ATAC (Assay for Transposase-Accessible Chromatin) and GEX (Gene Expression, or RNA) sequencing, to a treatment-naïve primary canine osteosarcoma. This comprehensive approach reveals the complexity of the tumor microenvironment by simultaneously capturing the transcriptomic and epigenomic profiles within the same nucleus. Furthermore, these results are analyzed in conjunction with bulk RNA sequencing and differential analysis of the same tumor and patient-matched normal bone. By delving into the intricacies of OSA at this unprecedented level of detail, we aim to unravel the underlying mechanisms driving intra-tumoral heterogeneity, opening new avenues for therapeutic interventions in both human and canine patients. This study pioneers an approach that is broadly applicable, while demonstrating significant heterogeneity in the context of a single individual's tumor.

Keywords: 10× genomics; canine; dog; heterogeneity; multiome; oncology; osteosarcoma; sequencing; single-nuclei; tumor.

Figure 1

High-power microscopy to evaluate single-nuclei quality. Phase contrast (A,C) and AO/PI-stained nuclei under fluorescent microscopy (B,D) show that nuclear membranes are intact with minimal blebbing.

Figure 2

Cellular heterogeneity in primary canine OSA reflected by nine cell clusters. UMAP plot shows nine total clusters (c0–8) for GEX (A), ATAC (B), and weighted nearest neighbor (WNN) graph which combines both modalities (C). Each dot represents a single nucleus, and the color corresponds to the cluster.

Figure 3

Cluster annotation with known cell markers and markers derived from differential analysis of bulk RNA sequencing of the same OSA tumor and patient-matched normal bone. Cell cluster annotation based on known single-cell marker genes (A) and marker genes from bulk RNA sequencing of the same primary OSA tumor and patient-matched normal bone (B). Each dot represents a single nucleus, and the color corresponds to the annotated cell group name.

Figure 4

Heatmap of CNVs in osteoblastic clusters. Heatmap of the CNVs identified in the osteoblast clusters 0, 1, and 7 using the remaining clusters as the reference (A). Expression values for the normal cell clusters (depicted in the top heatmap) are subtracted from tumor cluster expression data (depicted in the bottom heatmap) to visualize differences. Rows are individual nuclei, columns are genes (ordered from left to right across the chromosomes); amplifications are colored red and deletions are colored blue.

Figure 5

Differentially expressed genes define clusters in primary canine OSA. A plot of the top 3000 variable genes in the dataset with the top 20 most highly variable genes labeled. Red dots indicate the differentially expressed genes (n = 3000), and black dots represent the non-variable genes (n = 20,784) (A). Heatmap of the top 5 differentially expressed genes in each cluster. Clusters are identified by color and number on the top x-axis, gene symbols are listed on the y-axis; yellow indicates up-regulation and pink/purple indicates down-regulation (B).

Figure 6

Gene set enrichment analysis among clusters using Hallmark and Canonical pathways. Heatmap of the Z-scores produced from gene set enrichment analysis using Hallmark Pathways (A) and Canonical Pathways (B). Red indicates up-regulation, and blue indicates down-regulation. Clusters are numbered on the x-axis.

Figure 7

Enriched GO Biological Processes among clusters. The top 5 enriched GO Biological Processes and their associated fold enrichment and false discovery rate (FDR) for each cluster. Dot size correlates to the corresponding fold enrichment.

Figure 8

Annotation of the Immune Subclusters with Canine Leukocyte Markers. Sub-clustering and annotation of the immune cells present in the tumor microenvironment using known canine leukocyte markers reveals five distinct subclusters. Each dot represents a single nucleus, and the color corresponds to the annotated cell group name.

Figure 9

Expression of the top up-regulated genes from the bulk OSA tumor/normal bone RNAseq results. Expression of marker genes derived from the bulk OSA tumor/normal RNAseq dataset for the top up-regulated genes in the group (A). Marker gene expression for the top up-regulated genes in the same patient from the bulk OSA tumor/normal RNAseq dataset (B). Darker shades of purple indicate up-regulated expression, and gray indicates zero change.