At present, we cannot differentiate benign lesions of the pancreas from aggressive tumors without surgical resection and subsequent histological evaluation. The long-term goal of this proposal is to develop a positron emission tomography (PET) imaging agent for clinical use to improve risk stratification among pancreatic neoplasms. The goal of this R21 project is to establish proof of feasibility by demonstrating successful imaging of pre-invasive and invasive tumors in an animal model using our protease activity probes with optical and radionuclide imaging techniques. Recent work has shown cathepsin-like protease activity correlates with pancreatic intraepithelial neoplasm (PanIN) grade. These data raise the possibility that such activity could also be used to differentiate benign lesions from aggressive cancer. We have developed new technology that sensitively and non-invasively detects protease activities. Preliminary studies have shown promise for non-invasively detecting, differentiating, and grading pancreatic diseases including acute and chronic pancreatitis, intraductal papillary mucinous neoplasm (IPMN), PanIN, and pancreatic ductal adenocarcinoma (PDAC). These data support pursuing the following aims to test our hypothesis that we can use proteolysis to activate a membrane binding imaging agent for accurate and quantitative non-invasive imaging of pancreatic cancer to provide functional and dynamic information about tumor status. By using optical imaging to optimize the probe and radionuclide imaging we intend to show proof of principle for ultimately developing a PET imaging agent for clinical use. Aim 1. Optimize and Validate Lead Agents for the Detection of Pancreatic Cancer. The probes will be optimized to increase selectivity and signal to noise and to verify that they work in a variety of cancer cell lines. Probes will be ranked by fluorescence imaging assays in panels of mouse and human pancreatic cancer cell lines. Aim 2. Optimize Probes using Non-Invasive Imaging of Pancreatic Cancer Progression in a Transgenic Model of Pancreatic Cancer. The best probes from Aim 1 will be evaluated in mice. We will use a genetically-engineered mouse model of pancreatic cancer where pancreas-directed expression of constitutively active Kras drives PanIN formation. Probe accumulation, as detected with optical imaging and in sectioned tissue will be correlated to histological grade and oriented by immunofluorescence to cell specific markers. Laser capture microdissection will be used to harvest regions of high proteolytic activity and determine which protease(s) are present with MS. These data will directly link specific protease activities to pancreatic cancer and their precursor lesions in vivo. Based on results from optical imaging, a small number of selected probes will be labeled with radionuclides and PET imaging will be conducted. Pharmacodynamics and biodistribution studies will be carried out to demonstrate translatability to the clinic. These aims will determine whether pericellular proteolysis can be used to identify pancreatic disease and whether different pathologies can be distinguished non-invasively.