Head and Neck Squamous Cell Carcinoma (HNSCC), in particular in oral cavity cancer, continues to be a challenge in achieving local regional control and improved survival. This is due, in part, to a lack of understanding of the biology of the disease. In this study, we focus on cancers of the oral cavity, which have unpredictable biological behavior. While depth of invasion and T stage has been the focus of clinical treatment, we present data that suggest tumor behavior based on subsite location within the oral cavity is of clinical importance. We hypothesized that there exist key oncogenic transcriptional networks that are associated with unique anatomical subsites and may be correlated with tumor stage and outcome.
To decipher the transcriptomic networks defining the heterogeneity within oral cavity, we examined the gene expression levels of the Cancer Genome Atlas (TCGA) Oral Squamous Cell Carcinoma (OSCC) RNA-Seq dataset. Our comprehensive analysis of the OSCC differential gene-expression profiles has revealed commonalities and dynamic differences among the various anatomical subsites (i.e., Tongue, Floor of Mouth, Alveolar Ridge and the Buccal Mucosa). We demonstrated that differential gene-expression patterns can segregate tumors based on genes uniquely enriched within each anatomical subsite. Interestingly, we found that gene-expression patterns across all subsites defined two main clusters, one containing tumors of higher clinical stage and tumor grade and the other primarily containing tumors of low tumor grade and clinical stage.
Our studies highlight both genes that are systematically over-expressed in each subsite as well as uniquely enriched genes for each oral cavity subsite. These gene-sets have uncovered core gene signatures that highlight crucial aspects of OSCC biology. For example, we observed oncogenic processes such as: HIF1 signaling, angiogenesis and adherens junction signaling for buccal mucosa as well as upregulation of metastasis and cell proliferation genes for oral tongue. Our systematic approach to define and characterize tumor variance as a function of anatomy has thus revealed critical and specific pathways, which can explain the difference in clinical outcomes observed for each subsite and potentially be leveraged for targeted clinical trials.