Characterization of the Osteosarcoma Genome

Division Of Basic Sciences - Nci

Investigators

Linked publications, trials & patents

Abstract

As osteosarcoma is a rare cancer, obtaining suitable biospecimens is a major obstacle to studying this disease. To overcome this problem it has been necessary to reach outside of the intramural community. A collaboration to obtain specimens has been established with the Childrens Oncology Group (COG), the primary national cooperative clinical trials group for pediatric cancer, to sequence specimens collected from COG trials. As osteosarcoma had been selected by the TARGET project (NCIs Therapeutically Applicable Research to Generate Effective Treatments, http://target.cancer.gov/about/), the TARGET infrastructure has been used to facilitate the COG collaboration. Accrual was limited to previously untreated patients under the age of 40. Clinical annotations were available for most patients. Assays include the sequencing of exomes, transcriptomes, gene panels and, from selected specimens, whole genomes. We have also determined the presence of the ALT telomere maintenance mechanism using the C-circle assay for most samples. Our study also includes array based RNA and DNA methylation data as well as q-PCR miRNA expression data generated in collaborating laboratories. The data have been deposited in public databases. We are now finalizing the data analysis all cases collected through the TARGET program. We have identified clinically relevant genomic biomarkers and continue to identify genes and pathways which may have therapeutic implications. Significant results include the recognition that nearly all osteosarcomas contain mutations in TP53 and/or other cell cycle regulators. Gain of function is mainly the consequence of gene copy number amplification with only infrequent activating mutations of signal transduction genes. We noted a robust pattern of co-occurence and mutual exclusion of molecular properties in the dataset. As a result, we are now able to classify osteosarcoma tumors into groups based on these molecular features. We have also gained insights into the mechanisms and specific events that shape the complex genomes of these tumors. Remarkably, multiple mechanisms leading to genome rearrangements can coexist in one tumor (e.g. kataegis, chromothripsis, extrachromosomal DNA, and complex linear rearrangements), Additionally, RNA, DNA, and methylation signatures have been developed that are associated with high risk tumors. Although osteosarcoma is not responsive to current immunotherapies, we have identified a protective inflammatory signature that strongly suggests that the immune microenvironment is related to clinical outcomes. These data provide a resource for future analyses and point to numerous opportunities for preclinical research. Tumor clonality and divergence is extremely important in the context of precision therapies for osteosarcoma. Diverse clones are likely to have diverse response to targeted therapies with selective pressure from such therapies being likely to promote expansion of pre-existing resistant clones. It is nearly impossible to study clonal evolution in human primary tumors as samples collected at diagnosis are typically too small for these analyses. In the future, single cell analysis of human primary tumors may become possible, but at the present time canine osteosarcoma presents an excellent opportunity to study tumor evolution and clonality. Canine osteosarcoma is similar to the human disease in many respects. Collaborative studies are ongoing in the canine model with the intent of integrating canine and human genomic data to help define the critical oncogenic pathways and mechanisms of genome instability shared between these two species. We are particularly interested in using the canine model to study tumor clonality in primary tumors, a topic that can more easily be addressed in the canine system which does not incorporate neo-adjuvant chemotherapy prior to definitive surgery. Although some differences in mutation frequency distinguish canine tumors from their human counterpart, important aspects of the canine data are likely to translate to the human disease. Comparative genomics is providing insights that are relevant to both canine and human osteosarcoma. At present, our primary goal is to gain understanding tumor clonality in untreated canine primary tumors. In collaboration with the CCR veterinary oncology program, multiple samples are being collected from geographically mapped regions of primary tumors obtained at the time of limb amputation. Using short and long read DNA and RNA sequencing, we are developing a dataset that will provide a novel view of the origins and evolution of canine osteosarcoma. This information will be of interest to veterinary and human oncologists.