||Pancreatic cancer remains one of the poorest prognoses in medicine, with limited treatment options and 5-year patient survival of only ~7%. Activating mutations in the Kras proto-oncogene (KrasG12D) are nearly ubiquitous in human pancreatic cancer and are sufficient to initiate precancerous pancreatic intraepithelial neoplasia (PanINs) when expressed in adult acinar cells of mice. KrasG12D-driven PanINs normally take months to form, which lead to the hypothesis that acinar differentiation determinants may have an inhibitory effect on PanIN formation and subsequent cancer initiation. While targeted drugs that promote cell differentiation are curative in blood cancers like acute promyelocytic leukemia, there is currently minimal evidence suggesting that cell differentiation could play a role in limiting or stopping solid tumor growth. This dissertation aims to test whether loss of acinar differentiation, mediated through the transcription factor Ptf1a, is necessary and/or sufficient to initiate the early stages of pancreatic cancer. Using mouse genetics, we demonstrate that loss of Ptf1a alone is sufficient to promote acinar-to-ductal reprogramming and a cancer-like gene expression profile that is conducive to inflammation and Kras-dependency. As a result, Ptf1a-deficient acinar cells are rapidly transformed into PanINs in the presence of oncogenic KrasG12D. Consistent with PanINs acting as precursor lesions to invasive pancreatic cancer, we demonstrate that loss of Ptf1a allows for rapid progression to carcinoma in situ and mortality from invasive pancreatic cancer in an established mouse model of disease. These data confirm that loss of acinar cell identity hastens tumor development independently of canonical tumor suppressor loss. In contrast, a novel mouse line that sustains Ptf1a expression during pancreatic cancer initiation eliminates formation of KrasG12D-driven precancerous PanINs. Perhaps most strikingly, reintroduction of Ptf1a into established precancerous PanINs reverts them to primitive acinar cells. Our findings therefore suggest that loss of acinar cell differentiation is required for pancreatic cancer origination and progression. This dissertation is among the first studies demonstrating that a cell differentiation program can prevent and reverse premalignant solid tumor cells in vivo and thus introduces a novel paradigm for solid tumor prevention and treatment.