Cross-talk of four types of RNA modification proteins with adenosine reveals the landscape of multivariate prognostic patterns in breast cancer

Abstract

Background: Breast cancer (BC) is the most common malignant tumour, and its heterogeneity is one of its major characteristics. N6-methyladenosine (m6A), N1-methyladenosine (m1A), alternative polyadenylation (APA), and adenosine-to-inosine (A-to-I) RNA editing constitute the four most common adenosine-associated RNA modifications and represent the most typical and critical forms of epigenetic regulation contributing to the immunoinflammatory response, tumorigenesis and tumour heterogeneity. However, the cross-talk and potential combined profiles of these RNA-modified proteins (RMPs) in multivariate prognostic patterns of BC remain unknown.Methods: A total of 48 published RMPs were analysed and found to display significant expression alterations and genomic mutation rates between tumour and normal tissues in the TCGA-BRCA cohort. Data from 4188 BC patients with clinical outcomes were downloaded from the Gene Expression Omnibus (GEO), the Cancer Genome Atlas (TCGA), and the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC), normalized and merged into one cohort. The prognostic value and interconnections of these RMPs were also studied. The four prognosis-related genes (PRGs) with the greatest prognostic value were then selected to construct diverse RMP-associated prognostic models through univariate Cox (uniCox) regression analysis, differential expression analysis, Least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox (multiCox) regression. Alterations in biological functional pathways, genomic mutations, immune infiltrations, RNAss scores and drug sensitivities among different models, as well as their prognostic value, were then explored.Results: Utilizing a large number of samples and a comprehensive set of genes contributing to adenosine-associated RNA modification, our study revealed the joint potential bio-functions and underlying features of these diverse RMPs and provided effective models (PRG clusters, gene clusters and the risk model) for predicting the clinical outcomes of BC. The individuals with higher risk scores showed poor prognoses, cell cycle function enrichment, upregulation of stemness scores, higher tumour mutation burdens (TMBs), immune activation and specific drug resistance. This work highlights the significance of comprehensively examining post-transcriptional RNA modification genes.Conclusion: Here, we designed and verified an advanced forecasting model to reveal the underlying links between BC and RMPs and precisely predict the clinical outcomes of multivariate prognostic patterns for individuals.