NGS Data Repurposing Allows Detection of tRNA Fragments as Gastric Cancer Biomarkers in Patient-Derived Extracellular Vesicles

Abstract

Transfer RNA fragments (tRFs) have gene silencing effects similarly to miRNAs, can be sorted into extracellular vesicles (EVs) and are emerging as potential circulating biomarkers for cancer diagnoses. We aimed at analyzing the expression of tRFs in gastric cancer (GC) and understanding their potential as biomarkers. We explored miRNA datasets from gastric tumors and normal adjacent tissues (NATs) from TCGA repository, as well as proprietary 3D-cultured GC cell lines and corresponding EVs, in order to identify differentially represented tRFs using MINTmap and R/Bioconductor packages. Selected tRFs were validated in patient-derived EVs. We found 613 Differentially Expressed (DE)-tRFs in the TCGA dataset, of which 19 were concomitantly upregulated in TCGA gastric tumors and present in 3D cells and EVs, but barely expressed in NATs. Moreover, 20 tRFs were expressed in 3D cells and EVs and downregulated in TCGA gastric tumors. Of these 39 DE-tRFs, 9 tRFs were also detected in patient-derived EVs. Interestingly, the targets of these 9 tRFs affect neutrophil activation and degranulation, cadherin binding, focal adhesion and the cell–substrate junction, highlighting these pathways as major targets of EV-mediated crosstalk with the tumor microenvironment. Furthermore, as they are present in four distinct GC datasets and can be detected even in low quality patient-derived EV samples, they hold promise as GC biomarkers. By repurposing already available NGS data, we could identify and cross-validate a set of tRFs holding potential as GC diagnosis biomarkers.