System-Level Analyses of Multi-Omics Data to Reveal Mechanisms of Head & Neck Cancer - Head and neck squamous cell carcinoma (HNSCC) is a devastating disease with an overall 5-year survival rate of only ~60%. In part, this high death rate is a result of late disease diagnosis, often with regional lymph node metastasis. Recent studies, including The Cancer Genome Atlas (TCGA), and the Cancer Cell Line Encyclopedia (CCLE), mapped genomic alterations in a variety of cancers, and are beginning to provide insights into the dysregulated signals contributing to the onset and progression of HNSCC, but there has been only modest improvement in therapeutic strategies. This lack of progress is partly due to a lack of mechanistic understanding of the signals that drive malignant transformation. Prior and ongoing studies carried out in our laboratories have identified multiple homeostatic pathways whose deregulation contributes to HNSCC HPV(-) initiation, development and progression. In particular, we have shown that pharmacologic inhibition of the interaction between 𝛽𝛽-catenin and the histone acetyltransferase cAMP-responsive element binding (CREB)-binding protein (CBP) inhibits HNSCC cell proliferation and CSC phenotypes, and reduces invasive traits. Conversely, we showed that elevated β-catenin/CBP signaling in primary HNSCC tumors was associated with tumor progression and poor patient survival. We also showed a strong association between WNT/𝛽𝛽-catenin, the oncogenic activity of the paralogous transcriptional regulators YAP and TAZ (YAP/TAZ), and of mTOR, suggesting interactions between these pathways. Thus, our proposal seeks to elucidate, through computational approaches, the molecular links between somatic variation, the activity of selected homeostatic pathways (𝛽𝛽-catenin/CBP, TAZ/YAP, mTOR and others), and EMT and CSC phenotypes in early and advanced stages of HNSCC. To achieve this goal, we will perform integrative analyses of public multi-omics datasets from human and mouse, leveraging our newly developed methods of taxonomy discovery and gene regulatory network (GRN) inference. Application of these methods will allow us to identify important network hubs and regulators, and how their connectivity changes in response to perturbation or between phenotypic groups and disease stages. The hypotheses generated will be validated in in-vitro and in-vivo models we have developed in our laboratories. The results of our analyses will support the identification of novel candidate drivers of the disease amenable to therapeutic targeting. Three aims are thus proposed. First, we will analyze publicly available multi-omics datasets from human HNSCC, including bulk and single-cell RNAseq profiles, proteomics and methylomics profiles, and paired somatic copy number alterations (SCNAs) and mutations. Second, we will analyze bulk- and sc-RNAseq profiles from premalignant lesions in mouse and human, including profiles from mice treated with small molecule 𝛽𝛽-catenin/CBP inhibitors. Finally, we will share all our tools and results through well-documented Bioconductor packages and interactive web-based portals, which will allow other investigators to use the generated resources in their research.
