Comparative Oncology Program Laboratory

Division Of Basic Sciences - Nci

Investigators

Linked publications & trials

Abstract

Laboratory activities and projects carried out by the COP laboratory have the specific goal of improving the understanding of the impact of anti-cancer agents on comparative aspects of metastasis biology by virtue of parallel study of murine, canine and human cell lines in a variety of in vitro, ex vivo/in vivo (Pulmonary Metastasis Assay or PuMA) model systems. Data generated in this manner improves understanding of naturally-occurring canine osteosarcoma (OS) models and could be employed to answer unique in vivo questions regarding the anti-metastatic potential of agents, via the COTC clinical trial mechanism. To extend our investigations into the comparative aspects of naturally-occurring OS in dogs to humans, we have recently initiated the DOG2 project: Decoding the Osteosarcoma Genome of the Dog. The DOG2 project fulfills the main mission of the NCI-COP, which is to strategically position the canine cancer patient in studies of cancer biology and drug development, in order to improve outcome for both dogs and humans. The COP has a longstanding focus in osteosarcoma (OS) biology and clinical trials. OS is an aggressive pediatric/AYA malignancy and the most common malignant bone tumor in children and adolescents. OS is also a common naturally-occurring canine cancer with strikingly similar clinical presentation and natural history; preliminary studies suggest a shared molecular landscape. While dogs largely develop OS in adulthood, the similar genomic features and clinical disease characteristics underscore the notion that age does not distinguish canine OS from the disease in children. We recognize that osteosarcoma is a complex disease and success is unlikely with a single approach. Therefore, to expand this work it is necessary to identify additional targets and drugs. Collectively this work is designed to address the following strategic priorities: 1. Improved knowledge of comparative OS cancer biology to enhance dog to human translation 2. Discovery of new targets and companion biomarkers in support of drug development 3. Assessment and prioritization of new therapeutic strategies in preclinical models 4. Harmonization of comparative canine and human oncology clinical trials to advance new therapeutic concepts Biospecimens. We will utilize an existing high-quality clinically-annotated biospecimen repository of 400 canine OS patient sample sets, unique to the NCI, gathered from ongoing and completed canine OS clinical trials. This repository contains multiple aliquots of both tumor and matched normal tissues, preserved as frozen, RNAlater, and formalin-fixed specimens. Samples have been derived from treatment-naive dogs that all underwent the same treatment protocol (amputation + 4 cycles of carboplatin chemotherapy). We have full clinical data on every patient including demographics, treatment details/adverse events, and event-free and overall survival. We propose beginning with a subset of n = 100 patient samples selected from the ends of a Kaplan-Meier curve that represent early treatment failures (progression-free survival less than 90 days) vs. elite responders (progression-free survival greater than 360 days). We will also query n = 25 matched pairs of tumors with their metastases that developed despite therapy. We have not attempted to perform viability testing for PDX engraftment in NSG mice due to the technical challenges and low rate of metastasis associated with PDXs; rather we use 4 unique canine OS cell lines developed in our lab for dog-in-mouse preclinical testing and are in the process of creating and curating a collection of over 25 additional canine OS cell lines from extramural collaborators. All specimens will be subjected to pathology review by an expert board of veterinary and physician sarcoma pathologists to verify adequate sample quality for further genomic study. For genomic profiling, we will employ the following platforms to execute a combination of experiments to understand relationships between genomic complexity and integrity (CGH and WES to identify structural rearrangements, single nucleotide variants, copy number gain/loss), and gene expression (through RNAseq). The specific breed of dog will be verified through use of a germline SNP array recently made available through MARS/Wisdom Health. Samples for genomic profiling will be sent to our collaborator at NCSU (Matthew Breen) for DNA and RNA isolation, library prep, CGH for genome-wide DNA copy number profiling (Agilent CGH array), and for DNA whole exome sequencing (150b paired end reads at 100x for tumors/30x for normal, NovaSeq6000 platform with Roche dog exome kit), methyl-DNA resequencing, as well as RNA sequencing (125b paired end reads, Illumina HiSeq2500 platform). Computational approach. This project is an extension of an existing collaboration with NCSU and the NCATS Division or Preclinical Innovation (DPI) and Therapy for Rare and Neglected Diseases (TRND) program. Through experiments we have already conducted in n = 12 canine OS sample sets similar to those described above as a discovery cohort, the TRND bioinformaticians have developed a strategy that identifies distinct gene co-expression models between and specific to a set of 5 selected canine cancer types, of which OS is one, from which they derive cancer-specific gene panels. A similar analysis is performed on existing RNAseq data publicly available from human tumor samples from the same tumor types to produce cancer-specific human gene panels. These cancer-specific shared gene panels are compared, and genes common to both are retained to facilitate dog-to-human translation of genomic alterations. The RNAseq data, after alignment to reference, is then compared back to CGH and whole-exome DNA sequencing carried out in the same tumor sample to identify genomic changes (CNV, SNV) that are linked to expression changes and that may represent new druggable targets. These oncogenes or other drug targets in given pathway(s) were used to identify inhibitors, suppressors or agonists utilizing Pharos, which interfaces to a central database containing information about the targets collected by the Illuminating the Druggable Genome (IDG) program and the Comparative Toxicogenomic Database (CTD). This combined experimental/computational approach has already been tested using both proteomic and RNAseq data generated from the first n = 12 OS tumor/normal pairs. The data has been analyzed at NCATS to match changes in gene expression to drugs known to modulate the pathways associated with the associated genes. Canine OS cell line screening is already underway through NCATS to determine the in vitro impact of 40 drugs, including 30 synergistic combinations, based on the respective drugs' pathway modulation profile. These data serve as a proving ground for additional work carried out in a larger, outcome-linked, collection of tumor samples proposed herein. Single agents or combinations of drugs identified in Phase I will be tested in our existing panel of human and canine OS cell lines and screened rapidly through available xenograft human-in-mouse and canine-in-mouse models within the COP laboratory. This preclinical animal testing is critical to ranking and advancing appropriate candidates for further study and to understand the comparative features of how drugs behave in our in vivo models. We have expertise in both primary tumor and metastasis modelling to evaluate drug impact at several stages of tumor development and progression. These results will be the basis (go/no go decisi *TRUNCATED*