Identifying Anatomical Subsite Specific Gene Expression Networks in Oral Cavity Squamous Cell Carcinoma

Introduction: Achieving local regional control and improved survival in oral cavity cancer continues to challenging, in part, due to a lack of understanding of the biology of this disease. While depth of invasion and tumor stage has been the focus of treatment, our studies suggest tumor behavior based on subsite location is of clinical importance. We hypothesized that there exist key oncogenic transcriptional networks that are associated within anatomical subsites and may be correlated with clinical stage and outcome.

Methods: To decipher the transcriptomic networks defining the heterogeneity within the oral cavity, we interrogated The Cancer Genome Atlas Oral Squamous Cell Carcinoma (OSCC) RNA-Seq data. Our analysis of OSCC profiles has revealed commonalities and dynamic differences among the various anatomical subsites. These signatures were validated in patient and cell-line derived OSCC RNA-Seq datasets generated by our lab. These analyses were leveraged to identify promising potential treatments targeting the heterogeneity of OSCC.

Results: We demonstrated that differential gene-expression patterns can segregate tumors based on activated transcriptomic networks that are uniquely enriched within each anatomical subsite. Across all subsites, there are two defined main clusters, one containing tumors of higher clinical stage and tumor grade and the other with low grade and stage. Additionally, using predictive drug response analysis, promising new potential therapeutics were identified.

Conclusions: Our studies highlight genes that are systematically over-expressed as well as uniquely enriched genes for each oral cavity subsite. These core signatures highlight crucial aspects of OSCC biology that correlate with patient outcome. Furthermore, we propose new cohorts of targeted therapies. For example, our data identified a subgroup of tongue SCC patients sensitive to vorinostat, a noted HDAC inhibitor. Our approach to characterize tumor variance as a function of anatomy has revealed critical oncogenic pathways, that can be potentially leveraged for personalized cancer treatments.