Project Summary Exposure to pathogenic viruses that the immune system cannot rapidly clear leads to chronic viral infections. Cytotoxic T cell (CTL) responses against such viruses gradually adopt a epigenetically fixed, hypofunctional state commonly referred to as T cell exhaustion, which is thought to prevent immunopathological tissue damage from continuous high-level inflammation driven by antiviral CTL. A similar form of adaptation occurs in CTL responses against malignant tumors, and leads to increasing immunological tumor tolerance. In both chronic viral infection and malignancy, antibody blockade of so-called immune checkpoints, including the inhibitory PD-1 receptor expressed on activated CTL, can often restore exhausted CTL responses. However, the mechanisms by which this restoration occurs are incompletely understood. At the single cell level, CTL populations at effector sites of exhausted T cell responses consist of stem-like memory cells that express the transcription factor TCF-1 as well as TCF-1neg effector-like cells that transiently acquire full-fledged effector and cytokine-secreting function before eventually adopt a hypofunctional state of terminal differentiation. It recently emerged that PD-1 blockade does not restore the function of terminally differentiated cells, but acts primarily on TCF-1pos stem-like cells and causes them to divide, lose TCF-1 expression, and give rise to the highly proliferative transitory effector-like subset that mediates anti-viral or anti- tumor control. In prior work for this project we have characterized the critical role of the chemokine receptor CXCR6 to optimize the interaction of transitory effector-like CTL with a rare activated state of conventional dendritic cells called DC3, which trans-present IL-15 to CTL to promote their survival and accumulation in tumor tissue. DC3 densely cluster around micro-vessels in the tumor stroma to form perivascular niches, which we suggest may act as critical gatekeepers for CTL responses at sites of chronic inflammation. We propose to test the hypothesis that these perivascular niches are also sites for the maintenance of TCF-1pos stem-like CTL in tumor tissue, and for their conversion into effector-like cells. We will use highly multiplexed histological imaging of whole tumors and intravital microscopy to identify the chemotactic guidance cues by which TCF-1pos stem- like CTL are recruited to and retained in these niches and directed to interact with DC3, and we will explore the mechanism sby which the niches are formed in response to PD-1 blockade.