Uncovering Secrets of Small RNAs in cancer: Innovative Methods and their Applications in Expression Profiling

Throughout the years, small noncoding RNAs have emerged as a group of transcripts that do not only contribute to direct protein synthesis, but also fulfill key regulatory functions in genome organization and gene expression. However, the intricate complexity and variety of small RNA structures and their modifications impact the accurate detection and evaluation of their abundance by standard sequencing methods. Within this thesis, several methodological improvements are introduced for small RNA quantification in both cultured cell lines and clinical tissue samples, as well as low-input samples such as blood-derived extracellular vesicles (EVs), with a specific focus on the identification and quantification of transfer RNAs (tRNAs). Previous efforts using standard small RNA sequencing methods have been biased by inefficient reverse transcription of mature tRNAs, mainly due to their rigid secondary structures and extensive RNA modifications. This leads to a disproportionate emphasis on tRNA-derived small RNAs (tsRNAs) and an underrepresentation of full-length species. Such an imbalance makes it challenging to extract meaningful biological insights, especially in understanding disease mechanisms. To address this, we developed adapter-ligated libraries of tRNA-derived sequences (ALL-tRNAseq), which proved its effectiveness in detecting and quantifying full-length tRNA reads by relying on the highly processive MarathonRT. In parallel, we established the IsoSeek method specifically tailored for isomiR quantification in low-input and low-complexity samples. In addition to method development, we also established the NormSeq webserver tool to systematically assess the performance of normalization methods, making the selection of the best performing dataset normalization method key to preserve biological information. Altogether, each of these methods helped to improve detection, identification, and quantification of different types of RNAs in both high- and low abundance samples. Finally, we implemented these tools in two different studies. First, we demonstrated that mature, full-length tRNAs were the most abundant small RNAs within the EV lumen. Strikingly, the vast majority of these EV-tRNA sequences contained a dysfunctional 3’-CCA tail. This 3’ CCA truncation is rendering tRNAs incompetent for amino acid loading, thereby offering a new perspective into the physiological role of secreted EV-RNAs. Lastly, we evaluated the impact of cell proliferation and differentiation on tRNA expression levels and their modifications, as well as in patient-derived diffuse large B-cell lymphoma (DLBCL) samples. Unlike previous studies, while our findings indicated changes in specific tRNA expression levels during differentiation as well as in DLBCL compared to reactive lymph node tissues, we could not confirm the presence of reprogrammed tRNA pools associated with cellular state nor within different subsets of DLBCL.