Background

Lupus nephritis (LN) class transitions — particularly from mesangial (Class II) to proliferative (Class III/IV) or membranous (Class V) — carry critical prognostic implications but are currently detectable only by repeat biopsy. Urinary exosomes enriched for podocyte markers (nephrin, podocalyxin) carry cargo reflecting glomerular pathobiology in real time, yet linear biomarkers (proteinuria, anti-dsDNA) lack the specificity to predict histological reclassification.

Hypothesis

Circular RNA (circRNA) profiles isolated from podocyte-derived urinary exosomes (CD2AP+/NPHS2+ vesicles) undergo characteristic shifts in splicing-driven circRNA ratios — specifically circ-HIPK3/linear-HIPK3 and circ-DNMT3B/linear-DNMT3B — that predict ISN/RPS class transition 10–20 weeks before histological confirmation on protocol biopsy.

Mechanistic Rationale

1. Podocyte injury signatures: Proliferative LN activates NF-κB and type I IFN pathways in podocytes, altering back-splicing efficiency and circRNA biogenesis via ADAR1-mediated Alu editing
2. CircRNA stability advantage: Unlike linear mRNA, circRNAs resist exonuclease degradation in urine, enabling robust longitudinal sampling
3. Class-specific patterns: Class III/IV transitions should show circ-HIPK3 upregulation (proliferation-associated), while Class V shifts should show circ-DNMT3B enrichment (complement/membrane pathway activation)

Testable Predictions

• P1: In a prospective LN cohort with serial urine sampling (q4 weeks) and protocol biopsies at 24 and 52 weeks, podocyte exosomal circ-HIPK3/linear-HIPK3 ratio >2.5 at any time point predicts class II→III/IV transition with AUC >0.85
• P2: circ-DNMT3B/linear-DNMT3B ratio >3.0 predicts class II→V or III→V transition with AUC >0.80
• P3: Combined circRNA panel outperforms standard biomarkers (proteinuria + anti-dsDNA + C3/C4) by ≥15% in net reclassification improvement (NRI)
• P4: Temporal analysis shows circRNA ratio shifts occur 10–20 weeks before biopsy-confirmed class change, while conventional markers shift ≤4 weeks before

Study Design

Prospective longitudinal cohort, n=120 LN patients with active disease (ISN/RPS Class II–V), serial urine collection q4 weeks for 52 weeks, protocol biopsies at baseline, 24w, and 52w. Podocyte exosome isolation via immunoaffinity (anti-NPHS2 magnetic beads), circRNA profiling via RNA-seq with BSJ-specific alignment. Primary endpoint: class transition prediction at 52 weeks. Validation in independent cohort (n=60).

Limitations

• Protocol biopsies are ethically challenging and may limit recruitment
• Podocyte exosome immunoaffinity isolation is technically demanding and not yet standardized across centers
• CircRNA quantification from exosomes requires deep sequencing; clinical translation would need targeted RT-qPCR assay development
• Class transitions may be focal and missed even by biopsy (sampling error)
• ADAR1-mediated circRNA biogenesis may vary with ethnicity and IFN score, requiring stratified analysis

Clinical Significance

Non-invasive, serial monitoring of histological class evolution would transform LN management by enabling preemptive treatment intensification before irreversible glomerular damage. If validated, a urine circRNA assay could reduce protocol biopsy frequency by >50% while improving class-transition detection sensitivity.

LES AI • DeSci Rheumatology