New activities of the human DEK oncogene

点击翻译按钮获取中文摘要

ABSTRACT. This proposal builds on our past studies of the human DEK gene, which is amplified and overexpressed, and thus implicated in many human cancer types including head and neck squamous cell carcinoma (HNSCC). HNSCC is the sixth most common malignancy worldwide - with dismal outcomes. Identification of biomarkers for early diagnosis and new drug targets remains imperative. DEK is a highly conserved protein that binds nucleic acids and regulates diverse nuclear processes, including transcription. We have shown that DEK is an oncogene, by reporting that DEK overexpression extends the life span of primary human keratinocytes, induces hyperplasia in 3D models of epidermis, cooperates with classical oncogenes to stimulate keratinocyte transformation, and promotes HPV+ and HPV– HNSCC and breast cancer cell proliferation and invasion. Our data also demonstrated that DEK overexpression increases β-catenin activity, and this DEK-β-catenin axis is required and sufficient for some cancer phenotypes. A major hurdle in neoplastic transformation is the ability of cells to meet high bioenergetic and biosynthetic needs for sustained cancer cell growth. Recently, we reported that DEK overexpression increases transcription of key enzymes in glycolysis, lactate fermentation and cholesterol synthesis, and accumulation of metabolites that are glycolytic end products. However, it is not known whether β-catenin is required for DEK-driven metabolic reprogramming, and we do not understand the role of glycolysis and cholesterol synthesis in DEK-driven cancer phenotypes. Finally, in preliminary studies, we discovered that DEK is uniquely targeted to mitochondria, and DEK overexpression stimulated cellular oxidative phosphorylation capacity. In the proposed 3 aims, we test 2 hypotheses. First, that DEK overexpression promotes HNSCC oncogenic phenotypes by 2 discrete pools of DEK which drive metabolic deregulation, one nuclear (Aim1) and one mitochondrial (Aim2). Second, that targeting either β-catenin, or vulnerable nodes of the metabolic signature is an effective strategy to prevent HNSCC phenotypes. In partnership with an expert in stable isotope resolved metabolomics technologies, we will generate an atom-resolved map of metabolic networks controlled by DEK overexpression, and will then use this map to identify and validate novel metabolic flux vulnerabilities for drug targeting and diagnostic biomarkers (Aim3). Taken together, the proposed experiments represent a significant first step towards innovative prevention and treatment strategies to improve the outcomes of HNSCC via metabolic interventions.