The use of massively parallel sequencing to study the virome, epigenome and genome of canine and feline cancers

Abstract

The remarkable advancements in sequencing technologies have allowed the entire mutational landscape of hundreds of different types of human cancers to be defined. This knowledge gives patients two new categories of treatment options, small molecule inhibitors and targeted immunotherapy. Veterinary oncology has been slow to enter the field of genome-wide studies, largely due to costs. Five years ago, the first whole genome of a canine cancer patient was published. Because of NIH support and the steady decline in sequencing costs, the number of published canine tumor sequences has exploded this past year as these projects are being concluded. This dissertation contributes to this growing knowledge. The scope was diverse and included exploration in metagenomics, epigenomics and transcriptomics in canine and feline cancers. This research affirmed previous notions that papillomavirus is not a common cause of feline oral squamous cell carcinoma (FOSCC) and added to the understanding that it is not caused by any other DNA or integrated dsRNA virus. In canin DLCBL (cDLBCL) we provided more evidence that the subgroups seen in human DLBCL (hDLBCL) cannot be faithfully reproduced in the methylome of cDLCBL, likely at a greater scale than DNA mutations, and neurogenic/neuroendocrine, HOX and Wnt pathways were epigenetically targeted. In canin osteosarcoma (OSA) we built upon the genome model by adding loss of heterozygosity data and illustration of metastatic lesions. We also provided support to target the CDK4, Wnt and mTOR pathways. Finally we highlight unexpected discoveries such as the discovery of Epstein Barr virus in FOSCC, stability of the methylome in Golden Retriever DLBCL, epimutations in the TBX pathway in cDLBCL, dysregulation of human melanoma pathways in canine OSA, absence of dysregulation of HER2 in canine OSA, absence of genomic evidence of a telomere maintenance mechanism, recurrent shattering of chr26 in canine OSA and HSP90AB1, MITF and NOTCH2, as potentially actionable genes. These data will provide a significant resource for the community to generate new hypotheses and answer key questions about the molecular drivers of canine and feline malignancies.