A-to-I RNA editing in kidney tissue from patients with nephrotic syndrome

INTRODUCTION RNA editing has been implicated in endogenous double-stranded RNA (dsRNA) sensing and inflammatory disease, but its prevalence, genetic regulation, and consequences in diseased human kidney tissue have not been systematically characterized. Because ADAR enzymes edit multiple neighboring adenosines often in the same transcript, analyzing these sites holistically (as "clusters") may reveal effects missed by single-site analysis. METHODS We profiled both single-site and cluster A-to-I RNA editing in the kidneys of 215 participants from the Nephrotic Syndrome Study Network with focal segmental glomerulosclerosis or minimal change disease who had microdissected glomerular and/or tubulointerstitial RNA-seq and blood genome sequencing. We tested single-site and cluster editing association with estimated glomerular filtration rate, proteinuria, and an interferon stimulated gene expression score. To discover the genetic determinants of editing, we conducted mapping of both single-site, cis-editing QTL and cluster-level editing QTLs (cledQTLs). We then tested cledQTLs for colocalization with kidney eQTLs and kidney-relevant GWAS. RESULTS Greater cluster mean editing in tubulointerstitium was associated with lower interferon-stimulated gene activity (P = 4.81 x 10-9), lower UPCR (P = 0.01) and higher eGFR (P = 8.52 x 10-5). Genetic mapping identified 290 glomerular and 473 tubulointerstitial single-site edQTLs, as well as 21 glomerular and 51 tubulointerstitial cledQTLs. We identified 10 colocalized signals between cledQTL and GWAS and 14 between cledQTL and eQTL. Nine of 51 tubulointerstitial cledQTL clusters were individually associated with eGFR in NEPTUNE. CONCLUSION These results identify A-to-I RNA editing as a measurable and partly genetically regulated molecular phenotype in proteinuric kidney disease and nominate clustered editing of tubulointerstitial transcripts as a putative contributor to attenuated immune activity and higher kidney function. Cluster-level analysis identified additional genetically regulated editing patterns and colocalized signals not detected at individual sites, highlighting the added value of analyzing nearby editing sites as clusters.