Treatment of melanoma patients by adoptive transfer of in vitro manipulated tumor-reactive T cells has yielded dramatic successes, but extending this strategy to the treatment of other malignancies has proven more difficult. This has reflected in part difficulties identifying immunogenic tumor antigens (Ags) that can be safely and effectively targeted, the low avidity of T cells for the identified tumor Ags that are usually also detected in normal tissues, and immuno-suppressive pathways associated with many progressive tumors that interfere with effective immune responses. Advances in the last decade, including molecular profiling of tumors to find candidate target Ags, methods to genetically engineer high avidity T cell responses, and strategies to disrupt immuno-regulatory pathways and modify the tumor microenvironment, have provided opportunities to pursue T cell therapy. WT1 and mesothelin (MSN) are pro-oncogenic proteins over-expressed by many human tumors with limited expression in normal adult tissues that are being evaluated in human cancer vaccine trials. However, only a small fraction of patients have exhibited clinical benefit, with most failig to achieve T cell responses of large magnitude or high avidity. This could be addressed by adoptive T cell therapy with T cells transduced to express a genetically-engineered high affinity TCR. Our lab has isolated TCRs targeting these Ags and developed new strategies to improve the breadth and qualities of high affinity TCRs that can be generated from naturally isolated lower affinity tumor-reactive TCRs, but clinical translation would benefit from analysis in informative preclinical models since T cells with high affinity TCRs may function well as effector cells in vitro but in vivo may mediate injury to normal tissues expressing these Ags, be deleted, and/or lose function. WT1 and MSN expression in normal tissues, as well as in many malignancies, is very similar in mice and humans. Studies in mice that develop spontaneous tumors expressing these Ags, allowing tumor therapy to be evaluated and modulated in the context of a natural tumor microenvironment, can provide needed insights into the function, toxicities and potential efficacy of T cells expressing TCRs with a range of affinities. The overal goals of this project are to assess the potential to enhance therapeutic efficacy without inducing toxicity with T cells expressing TCRs with affinities that can be an order of magnitude greater than clinically to date, and create new TCR immunotherapy reagents that can effectively target tumors not currently treatable. The specific aims are to: 1) Determine the parameters for engineering TCRs specific for candidate tumor Ags by mutation and positive selection to increase affinity, including changes in on-rates and off-rates, that improve function and therapeutic activity without rendering transduced CD8 T cells toxic to normal tissues; and 2) Determine if adoptive transfer of tumor-reactive T cells expressing high affinity TCRs, can achieve and/or improve therapeutic efficacy, either alone or with reagents modifying the tumor microenvironment, in a murine model of 'spontaneous' pancreas cancer that reproduces most aspects of human disease.