Comprehensive profiling of functional Epstein-Barr virus miRNA expression in human cell lines

Background Epstein-Barr virus (EBV) establishes lifelong infections in its human host. The virus is associated with a broad range of malignancies of lymphoid and epithelial origin, including Burkitt’s lymphoma, post-transplant lymphoproliferative disease, nasopharyngeal carcinoma and gastric carcinoma. During the latent phase of its life cycle, EBV expresses more than 40 mature miRNAs that are highly abundant in tumor cells and may contribute to oncogenesis. Although multiple studies have assessed the relative expression profiles of EBV miRNAs in tumor cells, data linking these expression levels to functional target knockdown are mostly lacking. Therefore we set out to systematically assess the EBV miRNA expression levels in EBV + tumor cell lines, and correlate this to their functional silencing capacity in these cells. Results We provide comprehensive EBV miRNA expression profiles of the EBV + cell lines C666-1 (nasopharyngeal carcinoma), SNU-719 (gastric carcinoma), Jijoye (Burkitt’s lymphoma), and AKBM (Burkitt’s lymphoma) and of EBV − cells ectopically expressing the BART miRNA cluster. By deep sequencing the small RNA population and conducting miRNA-reporter experiments to assay miRNA potency, we were able to compare the expression profiles of the EBV miRNAs with their functional silencing efficacy. We observe a strong correlation between miRNA expression levels and functional miRNA activity. There is large variation in expression levels between EBV miRNAs in a given cell line, whereas the relative expression profiles are well maintained between cell lines. Furthermore, we show that miRNA arm selection bias is less pronounced for gamma-herpesvirus miRNAs than for human miRNAs. Conclusion We provide an in depth assessment of the expression levels and silencing activity of all EBV miRNAs in B- and epithelial cell lines of different latency stages. Our data show a good correlation between relative EBV miRNA expression levels and silencing capacity, and suggest preferential processing of particular EBV miRNAs irrespective of cell-type. In addition to encoding the largest number of precursor miRNAs of all human herpesviruses, EBV expresses many miRNAs precursors that yield two functional miRNA strands, rather than one guide strand and a non-functional passenger strand. This reduced strand bias may increase the size of the EBV miRNA targetome.