DESCRIPTION (provided by applicant): The use of animal models for preclinical therapeutic testing is a powerful approach, if the model(s) used are accurate mimics of the human disease in question. Many researchers have used histological assessments of tumors coming from Genetically Engineered Mice (GEM) as a means of linking specific mouse models to known human disease subtypes. We have developed a complementary approach based upon genome-wide gene expression profiling where we objectively compare the genomic profiles of mouse tumors coming from a given organ system (i.e. breast) versus human tumor profiles, and have been able identify specific GEM models which best mimic human disease subtypes. We propose to continue and extend these genomic validation studies to include additional models of breast carcinoma and to extend our studies to models of lung cancer with the goal of using the most appropriate models for pre- clinical therapeutic trials of immune checkpoint inhibitors given the tremendous clinical importance in this area. GEM models present an advantage over other pre-clinical models like PDX or human cell line models in that they are host is immunologically intact. We propose to study the immune microenvironment in these GEM models to identify mechanisms of immune evasion and predictive signatures for response to immune checkpoint inhibitors that can ultimately be applied to human clinical trials.

PUBLIC HEALTH RELEVANCE: Mouse models of cancer have been used to study cancer biology for decades. The expansion of genomic analysis coupled with technological advances in basic science has resulted in many new models of Genetically Engineered Mice. These models have been increasingly used in translational research on disease progression, and treatment response. However there is no accepted systemic methodology to confirm how relevant a new mouse model is to its human counterpart. Previously in the Perou lab the gene expression of over 27 murine models of breast cancer consisting of approximately 350 tumors was compared to hundreds of human breast cancer gene profiles representing 6 distinct subtypes. This across species analysis identified key features for many of the mouse models that were shared with the human subtypes. It also suggested which models had a higher correlation with a particular human subtype of breast cancer. We now propose to extend this method of validation to lung models and other GEM models available through the Oncology Models Forum. Furthermore, as immunotherapy can only be studied in animals with an intact immune system, we propose to use these genomically validated mouse models to develop key immune cell signatures/features that will then be used for murine-based Co-Clinical trials of immune checkpoint therapies.