A Hallmarks of Cancer-based Oncology Models Fidelity Score.

ABSTRACT:
Understanding the molecular mechanisms of disease pathogenesis and clinical therapeutics are central to treating cancer. For decades, mouse models have been a foundation of research efforts in oncology. New tools for genomic manipulation such as CRISPER and other technological advancements have served to accelerate the development of a multitude of novel and powerful animal models. However, challenges remain that impede further progress in translating results from animal models of cancer into improvements in patient care. Animal models do not always faithfully recapitulate human cancer, as evidenced by numerous therapeutic agents that show promise in preclinical studies but fail in clinical trials. Therefore, it is critically important for researchers to have tools that enable them to understand how accurately an animal model reflects the human cancer they wish to study. We have developed a new scoring system called the Oncology Model Fidelity Scores (http://comphealth.ucsf.edu/hallmarks) to address the issue of the fidelity of animal models of cancer. This analytic tool uses gene expression data from RNA-Seq or microarray studies for comparison. The conceptual framework and visualization for the scoring system is based on Hanahan and Weinberg’s Hallmarks of Cancer papers which describe a set of characteristic traits that define the transformation into malignancy, but which can also be used to classify cancer therapeutics based on the perturbation of a particular Hallmark. Therefore, to give researchers a better understanding of the biologic processes in each animal model, we have designed the Oncology Model Fidelity Scores to create metrics for each of the Hallmarks of Cancer. We have applied this scoring system to animal models available through the National Cancer Institute’s Oncology Model Forum (http://oncologymodels.org/), and analyzed how these compare to the relevant human cancer based on data from The Cancer Genome Atlas. We present case examples of the application of the Oncology Model Fidelity Scores to cancer models from a variety of different cancer types, and show that our scoring system enables rapid identification of biological processes that are driving the malignancy or leading to drug resistance. Analysis of tumor samples from a mouse model for lung adenocarcinoma identify the biologic processes that are essential for driving the malignant evolution from primary tumor into metastatic lesions. We also show that head and neck squamous cell cancer that has developed erlotinib resistance is driven by upregulation of pathways involved in the immune response and inflammatory pathways. Ultimately, the development of the Oncology Model Fidelity Scores will help to advance patient care through efficient identification and validation of mouse models for a variety of applications, from pre-clinical testing of novel therapeutics to the use of patient-specific animal models.