The ability of breast cancers to metastasize is the primary determinant of mortality from this disease. Accordingly, the identification of molecules that are required for metastasis is essential for developing more effective approaches to cancer therapy. We previously identified and cloned a novel SNF1/AMPK-related protein kinase, HUNK, from a murine HER2/neu-induced mammary tumor and demonstrated that this kinase is upregulated in the mammary gland during pregnancy. In humans, HUNK is overexpressed in aggressive subsets of carcinomas of the breast, ovary, and colon, and gene expression signatures associated with high HUNK expression predict decreased metastasis-free survival in women with breast cancer. The function of this kinase, however, is unknown. To determine the physiological role of this kinase, we used germline targeting to generate HUNK-deficient mice. Analysis of these mice revealed that HUNK is required for metastasis of c-myc and Polyoma middle T-antigen-induced mammary tumors in vivo, yet is entirely dispensable for normal development. Consistent with this, c-myc-induced tumor cell lines derived from HUNK-deficient mice exhibit decreased migration, invasion, and metastasis. Strikingly, each of these defects can be rescued by HUNK in a manner that depends upon its kinase activity. Moreover, a gene expression signature in mice that distinguishes HUNK-wild type from HUNK-deficient tumors accurately predicts decreased metastasis-free survival in women with node-negative breast cancer. Additionally, we find that HUNK is required for primary HER2/neu-induced mammary tumorigenesis and for spontaneous tumor recurrence. These findings constitute the first demonstration of a function for HUNK and identify this kinase as a potential therapeutic target in human cancer. The goal of this proposal is to understand the mechanisms underlying the requirement for HUNK in metastasis and tumor recurrence. The first aim of this proposal will use in vitro and in vivo reconstitution approaches to elucidate the cellular and molecular basis for HUNK-dependent metastasis, including a structure-function analysis of HUNK domains required for its pro-metastatic effects, identification and evaluation of candidate pathways that contribute to HUNK-dependent metastasis, and determination of whether the impaired invasive behavior of HUNK-deficient tumor cells is related to altered interactions of tumor cells with their stromal microenvironment. The second aim of this proposal will use in vivo genetic approaches, including constitutive and conditional HUNK-deficient mice in combination with inducible transgenic mouse models for breast cancer, to address the contribution of HUNK to metastasis, the establishment of residual neoplastic disease, and the re-emergence of residual neoplastic cells as recurrent cancers. In addition, this aim will build upon prior findings by determining the association of HUNK expression with relevant clinical and histopathological features of human breast cancers, including metastasis-free survival.