Abstract Pancreatic cancer -one of the deadliest human malignancies- is preceded by precursor lesions, most commonly Pancreatic Intraepithelial Neoplasia (PanIN). PanINs are characterized by changes in the epithelial architecture, and by the accumulation of a fibroinflammatory stroma. Oncogenic mutation in Kras are almost invariably associated with invasive pancreatic cancer (1-3) and often detected in PanIN lesions (4). Expression of oncogenic Kras in the mouse pancreas recapitulates the step-wise progression of the human disease (5). We have previously published a genetically engineered mouse model that allows inducible, tissue specific and reversible expression of oncogenic Kras in the pancreas, the iKras* mouse (6). By taking advantage of the reversible nature of oncogenic Kras expression in this model, we were able to address the role of this oncogene in PanIN and cancer maintenance. First, we showed that inactivation of oncogenic Kras at the PanIN stage leads to regression of the lesions and recovery of the pancreatic parenchyma. Then, by crossing iKras* mice with a mouse conditionally expressing a mutant form of the tumor suppressor p53 (7), we generated a model of metastatic pancreatic cancer, the iKras*p53* mouse(8). Inactivation of oncogenic Kras* in iKras*p53* mice bearing invasive tumors and metastases led to tumor regression, albeit without complete eradication of the malignant cells. Taking advantage of the reversible nature of oncogenic Kras expression in our model, we have undertaken to investigate the effect of modulating epithelial Kras expression on the tumor stroma. We have previously shown that the extensive myofibroblast population surrounding PanINs and invasive cancer depends on epithelial Kras for its proliferation, active status and maintenance. Recently, we have started characterizing the changes in immune cells infiltrating the pancreas during the PanIN stage. The induction of acute pancreatitis leads to extensive immune infiltration both in wild-type and iKras* mice. However, in wild type animals the immune infiltration is transient, and coincides with the dynamics of tissue repair. In contrast, in iKras* mice the immune infiltration is sustained over time. Moreover, the nature of the infiltrating cells is different in iKras* mice, with increased CD4+ T cells, and immature myeloid cells and fewer CD8+ T cells. This infiltration of immune cells is dependent on oncogenic Kras over time, as inactivation of Kras expression leads to a reduction in the number of inflammatory cells over time. While the nature of infiltrating immune cells over pancreatic carcinogenesis has been characterized before (9), the function of each specific subset is not fully understood. CD4 T cells are prevalent within the Kras-mutant pancreas, and include immune-suppressive cells such as regulator T cells and Th17 cells. Thus, we hypothesized that CD4+ T cells are required in pancreatic cancer. By genetically ablating CD4 T cell in iKras* mice (thus generating iKras*;CD4-/- animals) we determined that CD4 T cells are required for PanIN formation. Furthermore, we showed that CD4+ T cells promote pancreatic carcinogenesis by inhibiting the anti-tumor activity of CD8 + T cells. Thus, our results indicate that a potential anti-cancer immune response is inhibited by CD4+ T cells in response to signals derived from Kras mutant epithelial cells. Given the potential of immune-modulation as part of pancreatic cancer therapy, we plan to follow up on our initial observations by identifying the epithelial signals that modulate the immune response in pancreatic cancer.