Analysis of RNA polyadenylation in healthy and osteoarthritic human articular cartilage

Abstract

ABSTRACTAn important transcript structural element is its 3’ polyadenylated (polyA) tail, which defines the 3’ boundary of the transcript’s genetic information and is necessary for transcript stability. The position of the polyA tail can vary, with multiple alternatively polyadenylated (APA) transcripts existing for a single gene. This can lead to different length transcripts which can vary in their 3’ regulatory domains and even by inclusion or exclusion of protein-coding introns. The distribution of polyA tail location on articular chondrocyte transcripts has not been examined before and this study aimed to be the first to define polyadenylation events in human chondrocytes using age-matched healthy and osteoarthritic knee articular cartilage samples. Total RNA was isolated from frozen tissue samples and analysed using the QuantSeqReverse 3’ RNA Sequencing approach, where each read runs 3’ to 5’ from within the polyA tail into the transcript and will contains a distinct polyA site. Initial analysis of differential expression of overall transcript abundance identified by the reads showed significant disruption to transcript levels when healthy samples were compared to osteoarthritic ones. As we expected, differentially regulated genes were enriched with functionalities that were strongly associated with joint pathology. As part of this analysis, we also identified a substantial number of differentially expressed long non-coding RNAs that had not been linked to osteoarthritis before. Subsequent examination of polyA site data allowed us to deifne the extent of site usage across all the samples. This included identification of chondrocyte genes that exhibited the greatest amount polyA site variation. When comparing healthy and osteoarthritic samples, we found that differential use of polyadenylation sites was modest. However, of the small number of genes affected, there was clear potential for the change in polyadenylation site usage elicited by pathology to have functional relevance. We examined two genes, OSMR and KMT2A, in more detail, defining how APA affects transcript turnover and then, in the case of OSMR, identifying that APA is sensitive to inflammatory cytokine stimulation. Overall, we have characterised the polyadenylation landscape of human knee articular chondrocytes but can conclude that osteoarthritis does not elicit a widespread change in their polyadenylation site usage. This finding differentiates knee osteoarthritis from pathologies such as cancer where APA is more commonly observed.