84 The Role of the GNAS Oncogene in the Immune Microenvironment of Intraductal Papillary Mucinous Neoplasm

INTRODUCTION: Pancreatic ductal adenocarcinoma (PDA) arises from two main precursor lesions, pancreatic intraepithelial neoplasia (PanIN), and intraductal papillary mucinous neoplasia (IPMN). Extensive studies of The PanIN-to-PDA sequence has identified the immunosuppressive microenvironment and recruitment of pro-tumorigenic inflammatory cells as a key hallmark, whereas the immune microenvironment of IPMN remains unexplored. This proposal aims to test the core hypothesis that GNAS-driven IPMN precursor lesions display a distinct inflammatory program that contributes to tumor initiation and progression. METHODS: We developed novel genetically engineered mouse models (GEMMs) harboring combinations of the most frequent mutations found in human IPMN of benign and malignant disease. We also utilized genetically characterized and carefully demographically annotated human tissues and perform immunological characterization by immunofluorescence, immunohistochemistry and flow cytometry. We also performed cytokine arrays and early functional studies from the preliminary data generated from this characterization. RESULTS: We identified that GNAS-mutant tumors are replete with T-cells and dendritic cells, as compared to the virtual absence of these cell types in “classic” PanIN-PDA (3.0 ± 0.6 vs. 81.4 ± 25.4, P < 0.03) (Figure 1). With the switching off of GNAS, there is an influx of CD8+ T cells 7.2 ± 1.9 vs. 25.4 ± 4.9 ( P < 0.004) and loss of CD206+ macrophages 62.6 ± 10.7 vs. 22.9 ± 6.8, P < 0.02 (Figure 2). We observe that GNAS is a potent regulator of these differences through immunomodulatory cytokines, IL33, Arg1; along with the upregulation of CD105 and Reg3G. The human-relevance of our findings is supported by our initial pilot studies of human pancreatic tumor specimens. These early studies indicated that IPMNs with GNAS mutations (alone or in combination with KRAS) have increased CD3+ and CD4+ T cell infiltrate compared to those lacking GNAS mutations, consistent with our murine data (Figure 3). CONCLUSION: GNAS mutant IPMNs display a distinct immune infiltrate as compared to classic PanIN-PDA lesions, suggesting that these tumors will likely be amenable to immunopreventative and immunotherapeutic strategies. These studies will link immune dynamics to tumor growth, which is profoundly induced in GNAS-mutant IPMN. Deciphering the interplay between the evolving neoplastic cells and immune system underlying these cystic tumors can inform the development of strategies to prevent or treat GNAS-mutant IPMN.