Assessing the diagnostic utility of tRNA-derived fragments as biomarkers of head and neck cancer

Roughly 54,000 individuals are diagnosed with head and neck cancers in the United States yearly. Transfer RNA-derived fragments (tRF) are the products of enzymatic cleavage of precursor tRNAs, and have been proposed for use as biomarkers of head and neck cancer. In this study, we aim to further analyze the utility that tRFs might provide as biomarkers of head and neck cancer. tRF read counts were obtained for 453 tumor and 44 adjacent normal tissue samples and used to construct a gradient boosting diagnostic model. Although we identified 129 tRFs that were significantly dysregulated between these samples, the model achieved a sensitivity of only 69 % and a specificity of 59 %. tRFs are thought to induce the degradation of mRNA transcripts containing a complementary "seed" region. Despite the above performances, we chose to explore this concept of translational regulation by analyzing these tRFs for inverse correlation to the expression of select oncogenes and tumor suppressor genes implicated in head and neck cancer. Among others, CysGCA 5'-half and LysCTT 3'-tRF were upregulated in the tumor samples, and corresponded to decreased expression of PIK3R1, AKT1, and CPEB3. These transcripts were further found to contain numerous significantly complementary sites at which tRF-mediated mRNA degradation might occur. Although these tRFs did appear to correlate to many of the oncogenic metrics analyzed, we believe that additional research is needed before they might be used to improve the diagnosis, treatment, and survival of patients with this disease.