Novel Approaches to the Molecular Characterization of Primary Canine Osteosarcoma

Abstract

Osteosarcoma (OSA) is an aggressive malignant bone tumor that often affects pediatric humans as well as dogs. Canine OSA shares many similarities with the human condition, including clinical presentation and molecular profiles, and therefore serves as an excellent model to study the disease. OSA is a difficult tumor to dissociate and sequence due to its bony and brittle composition. Furthermore, it is characterized by extreme genetic complexity and significant intra- and inter-tumoral heterogeneity. While great strides have been made in identifying key mutations driving carcinogenesis, the treatment and prognosis for OSA have remained largely unchanged for 40 years. New approaches to studying OSA are critical for improving patient outcomes for both humans and dogs. Differential gene expression analysis represents one approach to analyzing important differences in the molecular profiles between normal and tumor tissue. A prerequisite to successful sequencing and differential analysis is the isolation of high-quality RNA from both normal and neoplastic tissue. In the case of OSA, normal bone corresponds to the non-neoplastic and pre-cancerous tissue representative of osteosarcoma. Bone is a dynamic tissue consisting of many different cell types embedded within a rigid matrix. Removal of the bone marrow is an important consideration for isolating RNA that is unique to OSA progenitors for an appropriate differential gene expression analysis. Therefore, a method to isolate RNA from normal canine bone was first established for subsequent transcriptomic sequencing of both unaffected tissue and tumorous lesions. Transcriptomic bulk sequencing of primary OSA provides a global view of the changes involved in OSA tumorigenesis. Differential gene expression analysis identifies dysregulated genes in the tumors compared to the normal bone. Patient-matched OSA and normal bone were collected and sequenced to identify genes that were commonly dysregulated among the group. However, this approach requires a large sample size to accommodate the statistical analysis and fails to account for inter-individual differences. To circumvent this, we supplemented the group differential profile with a novel individual-level analysis by deriving individual fold-change differences from the group’s significant genes. The results confirm the hyper-variability in OSA and the need to study OSA at the individual level. Single-cell sequencing (SCseq) has emerged in the past five years as a valuable tool for interrogating the molecular changes occurring in diseased tissue at the level of individual cells. However, an essential requirement for this approach includes highly viable cells (>60% live cells) dissociated from the primary tissue. OSA has proved difficult to manipulate and dissociate due to the matrix composition, resulting in poor cell viability unsuitable for SCseq. Single-nuclei isolation and sequencing have recently evolved as an alternative using 10x Genomics technology. This approach requires high-quality nuclei but overcomes the need for viable whole cells as input. With protocol adjustments, we have successfully isolated high-quality nuclei from a primary canine OSA used for single-nuclei multiome sequencing. Multiome sequencing includes transcriptomic profiling using RNA sequencing as well as epigenomic profiling using ATAC (Assay for Transposase-Accessible Chromatin) sequencing. This dual sequencing approach was used to characterize the tumor microenvironment and elucidate changes occurring at single-nuclei resolution in an individual patient. This dissertation provides a thorough background of canine OSA, describes a novel method to isolate RNA from canine bone for proper comparative sequencing analysis, presents bulk RNA sequencing of 7 canine OSA tumors along with an individualized approach to differential gene analysis, and concludes with single-nuclei multiome sequencing of a primary OSA tumor to characterize the tumor microenvironment. Importantly, these approaches can be translated to study human OSA for comparison to canine OSA molecular pathways and for the development of more targeted and effective therapies.