Circulating Micrornas as Predictive Biomarkers for Lupus Nephritis Relapse

Abstract

Background: Lupus nephritis (LN) remains a major cause of morbidity in systemic lupus erythematosus (SLE). Up to 40–60% of patients with SLE develop LN and a substantial proportion experience renal flares after apparent remission, accelerating chronic damage. Existing serologic and urinary protein markers (e.g., anti dsDNA, complement, proteinuria) incompletely capture incipient renal inflammation. Circulating microRNAs (miRNAs)—small, stable, non coding RNAs detectable in plasma/serum and urine (cell free or exosomal)—have emerged as attractive “liquid biopsy” candidates for predicting LN relapse.

Methods: We narratively synthesize evidence on circulating miRNAs associated with LN activity and relapse risk, emphasizing longitudinal and prognostic studies. We summarize assay platforms, pre analytical issues, and analytical normalization; collate key candidate miRNAs (e.g., miR 146a, miR 21, miR 29c, miR 150, miR 155, let 7 family); and outline how miRNA panels might be integrated with clinical variables and machine learning models to anticipate renal flares.

Results: Across studies, urinary and serum/exosomal miRNAs track with histologic activity and chronicity indices and, in some cohorts, shift months before clinical relapse. Urinary exosomal miR 146a correlates with disease activity and flare occurrence over multi year follow up, supporting its use in longitudinal surveillance. Multimarker urinary panels comprising miR 21/miR 29c/miR 150 reflect fibrotic pathways and chronicity; while primarily prognostic for progression, they may enrich risk stratification for post remission flare when combined with clinical data. Additional candidates (e.g., miR 155, miR 203) show promise but require validation in prospective, adequately powered cohorts.

Conclusion: Circulating miRNAs—especially urinary exosomal miR 146a and composite panels involving miR 21/miR 29c/miR 150—are biologically plausible, analytically feasible biomarkers that could enhance prediction of LN relapse when embedded in multimodal models. Standardized pre analytical workflows, agreed flare definitions, and prospective multi center validation with time to event endpoints are the next steps toward clinical deployment.