Role of urokinase plasminogen cascade for the progression of chronic kidney disease

Proteinuria is a common feature in chronic kidney disease (CKD) and an independent risk factor for accelerated disease progression. CKD affects approximately 10% of the global population and leads to hypertension, cardiovascular disease, and increased mortality. Aberrantly filtered plasma proteins, including zymogens and active proteases have potential activity in the tubular lumen. Our group previously demonstrated that plasminogen, a component of the fibrinolytic system, is abundantly present in urine and activated by urokinase-type plasminogen activator (uPA) in proteinuria. uPA is physiologically secreted to the tubular fluid and is significantly increased in urine in proteinuria. Complement factors are found in urine from patients with proteinuria. Plasmin cleaves the key components C3 and C5 in vitro and C5 in vivo during thrombus formation, generating biologically active C3a and C5a, which are potent proinflammatory anaphylatoxins. In the present project, it was hypothesized that i) complement factors are aberrantly filtered from plasma and activated in the tubular lumen via the uPA-plasminogen cascade in proteinuria, ii) complement activation split products are detectable in urine in relation to grade of proteinuria, and deposited in the apical membrane of tubular cells, and iii) uPA-plasmin driven complement activation can be attenuated pharmacologically by amiloride (an off target uPA inhibitor).

Study I Complement proteins were measured in kidney transplant recipients (KTR) with albuminuria using in-house enzyme linked immunosorbent assays (ELISA) with neoepitope-specific monoclonal antibodies. Two cohorts were examined, one cross-sectional and one longitudinal. Urine excretion of mannose-binding lectin (MBL, 10-fold, p<0.001), C3c (2-fold, p<0.05), C3dg (3-fold, p<0.01), and sC5b-9 associated C9 neoantigen (20-fold, p<0.001) were significantly elevated in proteinuria, and C3dg and C9 neoantigen excretion followed changes in albuminuria over time, without changes in plasma concentrations. The fractional excretion of C9 neoantigen was significantly increased compared to albumin in cases, suggesting a net addition of sC5b-9 after filtration or preferential reabsorption of albumin. Nephron segment specific uEV-isolation documented an association of iC3b/C3dg and C5b-9 to proximal tubular apical membranes in albuminuria. The results strongly indicate intra tubular complement activation and deposition in the proximal tubular epithelium. In extension, suggesting that inhibiting complement in the intra tubular space could reduce the negative intra renal effects of proteinuria.

Study II In vitro studies showed that C3 and C5 were activated by uPA-plasminogen but not by uPA or plasminogen alone in physiological and buffered solutions in an amiloride sensitive way. Pooled urine from healthy humans generated C3a from purified human C3 by addition of exogenous plasminogen (activated by endogenous uPA). Pre-incubation of urine with 2 mM amiloride or 500 KIU/ml aprotinin, but not 20 mM ethylenediaminetetraacetic acid (EDTA) inhibited C3a generation. C3a urine excretion was increased significantly KTRs with proteinuria (n = 7) compared to KTR controls (n = 7) but was not reduced by high dose amiloride for one day (p = 0.08). In patients with diabetes type-1 with (n = 16) or without (n = 14) diabetic nephropathy, high dose amiloride treatment for two days reduced C3dg (p < 0.05, n = 5) and C9 neoantigen (p < 0.05, n = 7) excretion rates in cases with detectable urine complement. In proteinuric podocin (-/-) mice, amiloride treatment for four days reduced 24 h urine C3a and C5a excretions (p < 0.05 for both) with no effect on albuminuria compared to vehicle. Inhibition of uPA for 11 days by a monoclonal α-uPA antibody reduced 24 h C5a and C3a excretions by ~50% (p = 0.012 and p = 0.019, respectively). uPAplasminogen play a significant role in intra tubular complement activation and uPA inhibition or direct complement inhibition might slow the progression to end stage kidney disease in proteinuria.

Study III Preeclampsia (PE) is a severe complication of pregnancy with sudden onset of proteinuria and increased blood pressure. Since plasmin activated complement in urine and both complement and plasminogen excretion is elevated in PE, we sought to address the temporal association in an existing prospective cohort of 88 pregnant women with pregestational type-1 diabetes. Participants were included at gestational age (GA) 12 weeks and followed to parturition with urine and plasma sampling at GA 20, 28, 32, 36 and 38 weeks. 14 women developed preeclampsia in the study period and 8 controls with complete sample set were selected. Urine excretion of C3dg (p < 0.001) and C9 neoantigen (p = 0.002) increased gradually and in parallel to albumin in PE. C3dg and C9 neoantigen correlated to plasminogen (r = 0.51, p < 0.001 and r = 0.68, p < 0.001, respectively) and albumin (r = 0.44, p < 0.001 and C9 neoantigen, r = 0.59, p < 0.001, respectively) in urine. Receiver operating characteristic (ROC) analysis showed that C3dg and C9 neoantigen/creatinine ratios (AUC= 0.53 and 0.57 respectively) were inferior to albumin/creatinine ratio (ACR; AUC = 0.86) as biomarkers. uEVs were positive for C3dg in one of three and C9 neoantigen in three of three PE patients at GA 36 weeks but were absent in controls, suggesting tubular cell association. The temporal association is compatible with a direct relation between plasmin and complement activation, but complement was not a superior biomarker for PE.

Conclusion Studies I-III demonstrate that 1) complement split products are aberrantly filtered in relation to albumin and plasminogen, 2) complement precursors are activated in the tubular lumen and deposited on proximal tubular apical cell membranes, 3) uPA-plasminogen activates complement in vitro and 4) uPAplasmin mediated complement activation is attenuated by uPA inhibition and in vitro and in vivo. The findings links proteinuria to intratubular proinflammatory signaling and complement deposition in tubular epithelial cells. This might lead to epithelial cell injury and progression of CKD and suggest a reno-protective effect of amiloride that goes beyond blood pressure reduction.