Identification of urinary extracellular vesicles differentially expressed RNAs in diabetic nephropathy via whole-transcriptome integrated analysis

Background: Diabetic nephropathy (DN) is a common systemic microvascular complication of diabetes and a leading cause of chronic kidney disease worldwide. Urinary extracellular vesicles (uEVs), which are natural nanoscale vesicles that protect RNA from degradation, have the potential to serve as an invasive diagnostic biomarker for DN. Methods: We enrolled 24 participants, including twelve with renal biopsy-proven T2DN and twelve with T2DM, and isolated uEVs using ultracentrifugation. We performed microarrays for mRNAs, lncRNAs, and circRNAs in parallel, and Next-Generation Sequencing for miRNAs. Differentially expressed RNAs (DE-RNAs) were subjected to CIBERSORTx, ssGSEA analysis, GO enrichment, PPI network analysis, and construction of the lncRNA/circRNA-miRNA-mRNA regulatory network. Candidate genes and potential biomarker RNAs were validated using databases and machine learning models. Results: A total of 1684 mRNAs, 126 lncRNAs, 123 circRNAs and 66 miRNAs were found in uEVs in T2DN samples compared with T2DM. CIBERSORTx revealed the involvement of uEVs in immune activity and ssGSEA explored possible cell or tissue sources of uEVs. A ceRNA co-expression and regulation relationship network was constructed. Candidate genes MYO1C and SP100 mRNA were confirmed to be expressed in the kidney using Nephroseq database, scRNA-seq dataset, and Human Protein Atlas database. We further selected 2 circRNAs, 2 miRNAs, and 2 lncRNAs from WGCNAs and ceRNAs and demonstrated their efficacy as potential diagnostic biomarkers for T2DN using machine learning algorithms. Conclusions: This study reported, for the first time, the whole-transcriptome genetic resources found in urine extracellular vesicles of T2DN patients. The results provide additional support for the possible interactions, and regulators between RNAs from uEVs themselves and as potential biomarkers in DN.