Urokinase Promotes Redundantly Intratubular C3a-Formation But Not ENaC-Driven Hypertension in DOCA/Salt Kidney Injury

Abstract

BACKGROUND:

uPA (urokinase-type plasminogen activator) inhibitors mitigate salt retention, plasmin, and complement activation in acute proteinuric kidney diseases. We hypothesized that in chronic kidney injury with albuminuria, uPA contributes to hypertension and complement-dependent tissue inflammation and injury.

METHODS:

Wild-type and uPA KO (knockout) mice underwent either sham surgery or unilateral nephrectomy, followed by insertion of deoxycorticosterone acetate (DOCA)- or sham pellets and a high (4%) or control (0.5%) sodium chloride diet for 21 days. Glomerular filtration rate was estimated by transcutaneous fluorescein-isothiocyanate–sinistrin, and arterial blood pressure was recorded continuously by indwelling femoral catheters. Urine was analyzed for albumin, plasmin(ogen), electrolytes, kidney injury markers (neutrophil gelatinase-associated lipocalin), and complement proteins (C3, C3a). Kidney tissue was examined for neutrophil gelatinase-associated lipocalin, epithelial sodium channel, C3, C3a, C5a, cytokines, inflammation, and macrophage polarization markers (CD16⁺CD32⁺, CD163⁺).

RESULTS:

DOCA-salt increased diuresis, Na⁺ excretion, albuminuria, tubular injury markers, and single-kidney glomerular filtration rate, with no genotype-dependent differences. Blood pressure increased by ≈30 mm Hg within 2 days and then stabilized, with no genotype difference for up to 15 days. DOCA-salt wild-type mice showed elevated urinary protease activity, plasmin, C3, and C3a, while these were mitigated in KO mice. In the kidney, DOCA-salt increased IL-6 (interleukin 6), MCP-1 (monocyte chemoattractant protein-1), MIP-1α (macrophage inflammatory protein 1 alpha), and TNF (tumor necrosis factor), while TNF and IP-10 (interferon gamma-induced protein-10) were reduced in KO mice. CD16⁺CD32⁺ macrophages predominated over CD163⁺ macrophages in DOCA-salt kidney tissue across genotypes.

CONCLUSIONS:

While not essential for filtration barrier injury, glomerular filtration rate decline, and hypertension in DOCA-salt–induced kidney injury, uPA, through plasmin, generates anaphylatoxins with effects on specific cytokines.

NOVELTY AND RELEVANCE.

What Is New?

• While urokinase is a dominant activator of urinary plasminogen in albuminuria, it does not drive glomerular injury, glomerular filtration rate decline, or hypertension in mineralocorticoid-high salt kidney injury.

• Deoxycorticosterone acetate-salt increases complement C3 and plasma cytokines, activating proinflammatory macrophages in kidney tissue.

• Urokinase-plasmin contributes to intratubular conversion of C3 to C3a and accumulation of cytokines IP-10 (interferon gamma-induced protein-10) and TNF (tumor necrosis factor) in kidney tissue.

What Is Relevant?

• We reveal urokinase’s proinflammatory role, linking the intratubular generation of anaphylatoxins to proinflammatory macrophage-associated cytokine response in kidney tissue.

Clinical/Pathophysiological Implications?

• Inhibition of protease-mediated complement activation in chronic proteinuric kidney injury—using broad-spectrum protease inhibitors—may reduce tissue inflammation, slow disease progression, and ultimately help preserve nephron function.

The association between hypertension and chronic kidney disease is well established and has important clinical implications. The 2 conditions often exist in a bidirectional relationship, leading to a progressive vicious cycle. The underlying mechanisms that couple hypertension and chronic kidney disease remain unclear. A shared important biomarker for adverse renal and cardiovascular outcomes is urinary albumin excretion, and reducing albuminuria lowers adverse cardiovascular and renal outcomes.^1–4 Albumin and cofiltered plasma components may have a pathogenic role in the association between hypertension and renal disease through adverse tubular effects.⁵

In diseases with filtration barrier injury, albumin is aberrantly filtered along with plasma proteases and complement proteins.⁶ Proteases—particularly plasminogen/plasmin, kallikrein,⁷ TMPRSS2 (transmembrane protease, serine 2),⁸ and prostasin⁹—activate the epithelial sodium channel (ENaC) by proteolytic cleavage of the extracellular domain of the γ-subunit.^10–12 The intratubular activation of plasminogen depends on the uPA (urokinase-type plasminogen activator).^11,13 Urine from patients and rodent models with proteinuria contains proteases that activate aprotinin- and amiloride-sensitive Na⁺ currents in collecting duct cells in vitro.^13–17 ENaC contributes to sodium and water retention, which is alleviated by the diuretic drug amiloride either by direct inhibition of ENaC or by off-target inhibition of uPA activity.^18,19 The broad-spectrum serine protease inhibitor aprotinin alleviates Na⁺ and water retention in acute proteinuria mouse models²⁰; however, plasmin or kallikrein deficiency or selective inhibition has only modest or no effect on fluid retention.^7,13,21 Thus, multiple urinary proteases may act in concert to contribute to salt and fluid retention in acute proteinuria.^20,22 In chronic settings, continuous activation of ENaC may be central to the pathogenesis of hypertension, as in, for example, Liddle syndrome. However, the role of uPA-plasmin versus other proteases (kallikrein, prostasin, TMPRSS2) in this process remains unclear, although amiloride has significant blood pressure (BP)-lowering effects.^19,23 In the context of hypertension, inflammation is increasingly regarded as an independent contributor.²⁴ We have previously shown that, in addition to ENaC activation, uPA-plasmin activates the complement C3 in urine from patients with albuminuria.^6,25,26 Plasmin can cleave C3 to C3a and C5 to C5a, generating potent proinflammatory anaphylatoxins in plasma and urine through a noncanonical pathway.^6,26,27 Inhibition of uPA by amiloride lowers urinary anaphylatoxin excretion in patients with albuminuria uPA.²⁶ The C3a and C5a receptors (C3aR, C5aR1, C5L2) are expressed in renal epithelial cells.^28,29 Intratubular anaphylatoxins are associated with tubulointerstitial inflammation, although the activation pathway is unclear. C3a-C3aR stimulates macrophage polarization toward the proinflammatory (M1) phenotype through the ERK-signaling pathway.³⁰ Preclinical studies show that C3 and C3aR deficiency attenuates the accumulation of proinflammatory macrophages,³¹ the development of diabetic nephropathy,³² and unilateral ureteral obstruction-induced kidney fibrosis.³⁰ Interestingly, plasminogen deficiency also reduces kidney fibrosis.³³ These observations support the concept that uPA-plasminogen promotes kidney inflammation driven by proinflammatory macrophages through noncanonical intratubular complement activation and ENaC-driven hypertension. Consistent with this, mice lacking C5 or the ability to assemble C5b-9 are protected from deoxycorticosterone acetate (DOCA)-salt kidney injury compared with control mice, despite similar BP.³⁴ We designed the present study to explore the role of uPA-plasminogen in chronic kidney injury with hypertension, albuminuria, and inflammation. The study elucidated filtration barrier injury, intratubular and tissue complement activation, inflammation, and ENaC cleavage, as well as their impact on BP. Experiments were designed to address the hypotheses that uPA deletion abolishes the intratubular activation of plasminogen to plasmin, which in turn reduces (1) albuminuria, (2) ENaC-driven hypertension, (3) intratubular C3a production, and (4) kidney proinflammatory macrophages and related cytokines. To investigate this, we used the DOCA-high–salt model in FVB-strain female and male mice, as much of the previous data has been gathered using this model.

Methods

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. Detailed in vivo protocols, CONSORT-like diagrams, and specific materials and methods are provided in the Supplemental Methods and Materials. All Western blots are provided in the unedited gel file.

Animal Studies

Ethics Permission

All experimental procedures were approved by the Danish Animal Experiment Inspectorate (2020-15-0201-00470) and complied with Danish national guidelines and NIH standards.

Mice were housed at the Biomedical Laboratory, University of Southern Denmark, on a 12:12-hour light-dark cycle at 22 °C with daily independent veterinarian inspections. Metabolic cage studies were conducted in a dedicated room at 32 °C.

The DOCA-Salt Model

We adopted protocols published by several other groups.³⁵ Mice with functional deletion of the gene PLAU encoding uPA (FVB; Plautm1lg/Mmus) and FVB/NHANHsd wild-type (WT) mice, aged 8 to 14 weeks, were included and genotyped by polymerase chain reaction using ear DNA (Experiment 1—CONSORT-diagram Figure S1).

Measurement of the Glomerular Filtration Rate

Glomerular filtration rate (GFR) was measured transcutaneously using fluorescein- isothiocyanate-sinistrin clearance (MediBeacon; Experiment 2—CONSORT-diagram Figure S1).

Continuous BP Measurements in Conscious, Unrestrained, Freely Moving Mice

uPA, WT, and KO mice underwent sham/DOCA-salt treatment or ANGII (angiotensin II) infusion. Arterial pressure (mean arterial pressure, systolic and diastolic pressure, and heart rate) was recorded continuously for 15 days via femoral artery catheters in conscious, freely moving mice (previously described¹⁹; Experiments 3–4—CONSORT-diagram Figure S1).

Statistical Analysis

Normality was tested by D’Agostino and Pearson and data were presented as mean±SEM. Nonnormally distributed data were log-transformed and presented as mean±interquartile range (IQR). Two-way ANOVA with Bonferroni post hoc tests was used; P<0.05 was considered significant. Analyses were performed in GraphPad Prism 9 (San Diego, CA).

Results

Effect of DOCA-Salt on Body Weight, Diuresis, Water Intake, Kidney Injury, and Function in WT Mice and in Mice With Deletion of uPA

The protocol included 32 female and 34 male mice. All female mice completed the study, while 1 male mouse died; thus, 33 male mice completed experiment 1. For further details, see Figure 1A and Figure S1A. In response to DOCA-salt treatment, 24-hour water intake, diuresis, and urine Na⁺ and NH₄⁺ excretion increased significantly compared with sham, with no difference between genotypes (Table S1). Urine potassium excretion was not significantly different between groups. At termination (day 21), DOCA-salt–treated mice had similar body weight gain (Δ body weight) as sham-control mice, with no genotype difference (Table S1). DOCA-salt–treated mice exhibited significant high–molecular-weight glomerular proteinuria (Figure 1B). Twenty-four-hour albumin excretion increased ≈1250-fold after DOCA-salt treatment, with no difference between genotypes (Figure 1C). The single remaining right kidney of the DOCA-salt–treated mice showed a 2-fold increase in organ/body weight ratio compared with a single kidney from sham mice for both genotypes. DOCA-salt–treated mice had cardiac hypertrophy compared with sham, with no difference between genotypes (Table S1). The female mice subjected to the DOCA-salt model exhibited the same phenotype as the male mice, with no significant differences between sexes (Table S2). The subsequent experiments and analyses were conducted only on male mice to reduce the number of animals used.

Figure 1.

Deoxycorticosterone acetate (DOCA)-salt increases urinary protein/albumin excretion, kidney injury markers, and single-kidney glomerular filtration rate. A, Illustration of the experimental timeline. B, Coomassie-stained SDS-PAGE gel showing protein separation from urine of wild-type (WT) and uPA (urokinase-type plasminogen activator) KO control and DOCA-salt male mice (n=3 per group) at day 21. BSA (BSA, 10 mg/ml) was used as a positive control. C, 24-hour urine albumin excretion increased ≈1000-fold following DOCA-salt compared with control male mice, with no difference between genotypes (n=7–9 per group). D, Glomerular filtration rate (GFR) was measured transcutaneously using fluorescein-isothiocyanate (FITC)-sinistrin at day 21. Single-kidney GFR (SKGFR) was elevated in DOCA-salt male mice compared with control, with no genotype-dependent differences (n=6–8 per group). E, Immunohistochemical labeling for albumin in kidney sections on day 21. Arrows indicate albumin labeling was localized in the capillaries in sham-control male mice, while there was a marked tubular signal for albumin after DOCA-salt male mice: Scalebar, 50 µm. Significance was tested by 2-way ANOVA. *P<0.05, **P<0.001, ****P<0.0001. KO indicates knockout; and ns, nonsignificant.

Substudy 2 included 32 male mice; 3 mice were excluded because of the absence of valid recordings (Figure S1B, experiment 2). While total GFR was lower in the DOCA-salt protocol (1.46 versus 1.05 mL/min per 100 g body weight) with no genotype-dependent differences (Figure S2A), single-kidney GFR was higher in DOCA-salt–treated mice than calculated single-kidney GFR in sham mice, assuming a 50:50 functional distribution, with no difference between genotypes (Figure 1D). Plasma urea concentration (BUN) did not differ between groups (Table S1). Periodic acid-Schiff staining of kidney sections showed that DOCA-salt treatment led to dilated tubules, thickened basement membranes, interstitial inflammation, glomerular sclerosis, dilated Bowman space, and tubular casts (Figure S2B). Immunostaining of kidney sections for albumin showed distinct punctate labeling of sham kidneys associated with the glomerular lumen and peritubular capillaries, visible at junctions with basolateral epithelial membranes. By contrast, in DOCA-salt–treated mice, albumin labeling was more widespread and associated with the cytoplasm of dilated proximal tubules, glomerular cells, nonproximal tubules, and tubules in the medulla (Figure 1E). The kidney injury marker NGAL (neutrophil gelatinase-associated lipocalin) was elevated in kidney tissue from DOCA-salt–treated mice compared with control, with no genotype difference (Figure S1C-D). Urinary NGAL and KIM-1 protein excretion was significantly higher in urine from DOCA-salt–treated mice compared with control, with no difference between genotypes (Figure S1E and Table S1).

Effect of DOCA-Salt on Urine Protease Activity, Plasminogen Activation, ENaC Cleavage, and BP in WT Mice and Mice With Deletion of Urokinase

DOCA-salt treatment led to the appearance of plasminogen in urine based on comigration with plasminogen in plasma after SDS-PAGE separation and blotting. In WT male mice, plasminogen was activated to plasmin, which migrated at ≈60 kDa, and this was reflected by a significant appearance of protease activity by gel zymography. In mice with deletion of uPA, a similar amount of plasminogen appeared in urine as observed in WT DOCA-salt mice, but active plasmin was absent, and total urine protease activity was markedly reduced. Semiquantification of the relative urine plasmin/plasminogen ratio in samples from DOCA-salt male mice showed a significantly elevated ratio in WT compared with uPA KO (Figure 2A, 2B, and 2C). Western blot analysis of kidney tissue homogenates for γ-ENaC using a C-terminal antibody revealed a dominant band that migrated at ≈75 kDa, likely reflecting furin-cleaved γ-ENaC, with highest abundance in DOCA-salt–treated mice. Immunoreactive bands of lower molecular weight (≈70, 50, and 40 kDa) were present, reflecting cleaved fragments of γ-ENaC, where the ≈50 kDa band represents cleavage by extracellular proteases (Figure 2D and 2E). Semiquantification of the relative ratio between cleaved and furin-cleaved immunoreactive bands showed elevation in the DOCA-salt male mice samples compared with the control, and with no difference between genotypes (Figure 2F).

Figure 2.

Urinary protease activity and increased urinary plasminogen excretion in deoxycorticosterone acetate (DOCA)-salt and changes in γ-epithelial sodium channel (γ-ENaC) relative protein abundance. A, Immunoblotting of 24-hour urine samples from DOCA-salt–treated uPA (urokinase-type plasminogen activator) wild-type (WT) male mice revealed immunoreactive bands corresponding to plasminogen (100 kDa) and active plasmin (70–75 kDa; n=9), which were absent in sham-treated WT (n=8) and KO (n=8) male mice. No active plasmin was present in urine from DOCA-salt uPA KO (n=8). B, Semiquantification showed a significantly higher plasmin/plasminogen ratio in urine samples from DOCA-salt uPA WT compared with DOCA-salt uPA KO male mice. C, Gelatin zymography showed protease activity only in DOCA-salt uPA WT urine samples from male mice, migrating at 75 kDa, consistent with the active plasmin control (n=2 in each group). D, Illustration of γ-ENaC cleavage sites at aa 143 (furin) and 186 (extracellular proteases) and showing the C-terminal antibody (Ab) binding epitope. E, Immunoblotting of whole kidney tissue homogenates from DOCA-salt and sham uPA KO and WT male mice using C-terminal γ-ENaC antibody revealed dominant bands migrating at 75 and 70 kDa with the highest intensity in DOCA-salt samples (n=8 per group). Additional bands migrated at ≈50 and 42 kDa showed similar intensity across all samples. F, Immunoreactive bands were normalized to β-actin. Semiquantification showed an elevation in the ratio between extracellularly cleaved to furin-cleaved γ-ENaC in DOCA-salt–treated male mice compared with controls, with no genotype-dependent differences (n=8 per group). Significance was tested by 2-way ANOVA or unpaired t test. *P<0.05, **P<0.001, ****P<0.0001. L indicates ladder; KO, knockout; and ns, nonsignificant.

Substudy 3 enrolled 33 male mice to record continuous BP in the DOCA-salt model versus sham with respect to genotype (Figure 3A). Mean arterial BP in sham mice with indwelling catheters from day 2 post-surgery was ≈111 ± 6 mm Hg and showed a slight decline over time to reach ≈98 ± 4 mm Hg at day 15. Mean arterial pressure showed a 24-hour diurnal variation between 100 and 110 mm Hg throughout the 15-day recording period (Figure 3B). At termination, sham mice showed a slight decrease in mean arterial and diastolic pressure compared with the start of recordings (Figure S3). By contrast, mean arterial BP increased significantly after 48 hours of DOCA-salt treatment to ≈124±3 mm Hg and reached a plateau around day 6 at ≈135±3 mm Hg. Thereafter, BP stabilized in DOCA-salt–treated mice with no difference in kinetics or absolute levels between genotypes, and recordings were terminated on day 15. There was no difference in diurnal variation in MAP, SBP, DBP, or heart rate over the 15-day recording period (Figure S3). At termination, the difference in BP was ≈40 mm Hg between DOCA-salt and sham (Figure 3B and Figure S3). Heart rate did not change during the 15-day recording time. Systolic and diastolic pressures increased in response to DOCA-salt, with no genotype-dependent differences (Figure S4).

Figure 3.

Continuous measurements of arterial pressure in deoxycorticosterone acetate (DOCA)-salt treated and ANGII (angiotensin II) infused wild-type (WT) and uPA (urokinase-type plasminogen activator)-deficient male mice. A, Illustration of the arterial blood pressure measurement protocol. B, Mean arterial pressure (MAP) increased in both uPA WT (n=4) and uPA KO (knockout; n=4) male mice in response to DOCA-salt over 15 days, peaking at ≈135 mm Hg. MAP was significantly higher in DOCA-salt mice from day 2 onward compared with controls at the end of the experiment (day 15; P<0.05). Four uPA WT and 4 uPA KO male mice were included in the control groups. C, Continuous intravenous infusion of ANGII increased MAP, with no significant differences between uPA WT (n=8) and KO (n=6) mice. The MAP data were analyzed by 2-way ANOVA.

Of the 12 uPA KO male mice included in the study and subjected to DOCA-salt, 4 completed the experiment. One mouse was euthanized, and 7 had incomplete measurements, resulting in a success rate of 33%. Seven uPA WT male mice were allocated to DOCA-salt, of which 2 were euthanized and 1 had incomplete measurements (n=4, success rate 57%). In the sham groups, 6 uPA WT were included, 2 were excluded due to incomplete measurements, and 8 uPA KO male mice were included, of which 2 were euthanized and 2 had incomplete measurements, yielding a success rate of 67% and 50%, respectively (n=4; experiment 3, Figure S1B). To assess whether the pellet implantation, including 2 major surgery sessions, caused the low completion rate, a different approach was used with only 1 surgery session, also to elucidate whether uPA contributes to the direct vascular component of hypertension. ANGII was infused into WT and uPA KO male mice on a standard diet (experiment 4, Figure S1B). Ten uPA WT male mice were included, of which 1 died due to surgical complications and 1 with incomplete measurements (n=8). Twelve uPA KO male mice were included, of which 1 died due to surgical complications and 5 had incomplete measurements (n=6). Here, BPs at baseline were similar and increased with similar kinetics, stabilizing at the same level, 130±6 mm Hg, after ANGII (Figure 3C and Figure S5).

Effect of DOCA-Salt on Complement Factor C3 Abundance and Activation in Plasma, Kidney Tissue, and Urine in WT and Urokinase-Deficient Mice

The effect of DOCA-salt and uPA on the key complement component C3 was assessed by immunoblotting, which showed total C3 (predicted migration at 187 kDa) and C3c (predicted migration at 140 kDa) and, under reduced conditions, an immunoreactive band corresponding to the C3 α-chain appeared (110 kDa). DOCA-salt treatment increased C3 protein in kidney tissue in uPA KO male mice only compared with the control (Figure 4A). Twenty-four-hour urine from DOCA-salt–treated male mice showed full-length C3 protein. Under reduced conditions, C3 migrated as several moieties in the urine from DOCA-salt mice with no apparent genotype differences (Figure 4B). In plasma, total C3 increased in DOCA-salt WT male mice but not in uPA knockout (Figure S6A and S6B). By ELISA, the C3 split product anaphylatoxin C3a protein in kidney tissue homogenates showed increased abundance in response to DOCA-salt with no difference between genotypes (Figure 4C). Anaphylatoxin C5a was similar in all groups (Figure S6C). The 24-hour mass excretion of C3a protein increased in DOCA-salt–treated WT male mice, while in DOCA-salt–treated uPA KO male mice, the increase was significantly mitigated (Figure 4D). As the generation of C3a is lowered but not abolished by deletion of uPA, it was investigated whether redundancy exists with kallikrein.³⁶ Kallikrein generated C3a from C3 when using either physiological saline or urine from healthy subjects as the medium in vitro (Figure 4E and Figure S7A). Kallikrein was absent in 24-hour urine samples from control male mice but appeared in samples from DOCA-salt male mice along with proteinuria (Figure 4F). Mouse plasma (Figure 4F) was included as a positive control for kallikrein and migrates at a similar molecular size to that in urine, indicating aberrant filtration of kallikrein in DOCA-salt. As a positive control, plasmin facilitated cleavage of C3-to-C3 α-, α′-, β-chain, and C3a in a buffer system and urine in vitro (Figure S7B and S7C). Thus, both uPA-plasmin and kallikrein may activate C3 to C3a in plasma and urine.

Figure 4.

Plasmin-urokinase and kallikrein facilitate the cleavage of complement C3 and formation of anaphylatoxin C3a in urine and whole kidney tissue homogenates in vivo and in vitro. A, Immunoblotting of whole kidney tissue homogenate samples showed C3 (≈180 kDa) and C3c (≈140 kDa) under nonreduced conditions. Total C3 protein increased in deoxycorticosterone acetate (DOCA)-salt uPA (urokinase-type plasminogen activator) KO male mice compared with control, with no significant difference between genotypes (n=6 per group). B, Immunoblotting of 24-hour urine samples for C3 protein w/wo reducing conditions. Mouse plasma (MP) was used as a positive control (n=6 per group). C, In-house ELISA analysis of whole kidney tissue homogenates showed significantly increased abundance of C3a protein in DOCA-salt–treated male mice compared with controls (n=8 per group). D, C3a protein was significantly elevated in 24-hour urine samples from DOCA-salt uPA WT (n=8) male mice compared with both control (n=8) and DOCA-salt uPA KO male mice (n=6). E, Kallikrein cleavage of C3-to-C3 α-chain and anaphylatoxin C3a. Lane 1: C3, Lane 2: C3, kallikrein in a buffer system, Lane 3: C3 and kallikrein in urine in nonreduced conditions. Lane 4: C3, Lane 2: C3 and Kallikrein in a buffer system, and lane 6: C3 and kallikrein in urine under reduced conditions. F, Immunoblotting of 24-hr urine samples under nonreduced conditions revealed the presence of kallikrein in urine from DOCA-salt treated WT (n=4) and uPA KO (n=4) male mice with proteinuria, while it was absent in urine samples from control WT (n=2) and uPA KO (n=2) male mice with no proteinuria. Mouse plasma (MP) was used as a positive control for kallikrein. Significance was tested by 2-way ANOVA. *P<0.05, **P<0.001, ****P<0.0001. L indicates ladder; KO, knockout; and ns, nonsignificant.

Effect of DOCA-Salt on Proinflammatory Cytokines, Inflammation Markers, and Fibrotic Changes in the Kidney Tissue of WT and Urokinase-Deficient Mice

Next, it was tested whether the genotype-related difference in urine anaphylatoxin had downstream consequences by analysis of plasma and kidney tissue samples for cytokines associated with macrophages polarized toward a proinflammatory phenotype. In plasma, cytokines IL-15, IL-12p17, and IL-23 (interleukin-15, -12p17, -23) were below the detection limit. IL-1β (interleukin 1 beta), IL-6 (interleukin-6), TNF (tumor necrosis factor), IP-10 (interferon gamma-induced protein-10), and MCP-1 (monocyte chemoattractant protein-1) levels were significantly higher in plasma from DOCA-salt–treated male mice compared with controls, with no genotype-dependent differences, and only MIP-1α (macrophage inflammatory protein 1 alpha) was unchanged (Figure 5A, and 5B and Figure S8A, S8B, S8C, and S8D). In whole kidney tissue homogenates, cytokine levels were normalized to the total protein concentration. IL-6, MCP-1, and MIP-1α were elevated in kidneys from DOCA-salt–treated male mice compared with controls, with no differences between genotypes. IL-23 was unchanged (Figure S8E-H). Notably, TNF and IP-10 were significantly higher in DOCA-salt–treated uPA WT kidneys compared with DOCA-salt–treated uPA KO and controls (Figure 5C and 5D). IL-1β, IL-15 and IL-12p70 were below the detection range in whole kidney tissue homogenate (not shown). CD45 (leukocyte marker), F4/80 and CD68 (general macrophage markers) mRNA abundances were elevated in kidney tissue from DOCA-salt male mice compared with controls, with no difference between genotypes (Figure S9A, S9B, and S9C). CXCL10 (C-X-C motif chemokine Ligand 10) and TNF mRNA levels were significantly elevated in response to DOCA-salt compared with control, with no difference between genotypes. In contrast, iNOS (inducible nitric oxide synthase) mRNA levels were unchanged across groups (Figure S9D, S9E, and S9F). The mRNA abundance of markers of macrophages polarized toward the M2-like phenotype, Chil3 and Fizz1, were unchanged between groups, while CD206 was slightly elevated in DOCA-salt uPA KO compared with KO sham (Figure S9G, S9H, and S9I). Fibrosis markers collagen I and III increased at the mRNA level in DOCA-salt–treated kidney tissue with no difference between genotypes, while the mesenchymal marker, α-smooth muscle actin, was unchanged (Figure S9J, S9K, and S9L). The relative mRNA levels of fibrosis, pro- and antiinflammatory markers in the kidney tissue are summarized in the heatmap (Figure 5E). Immunoblotting revealed increased protein abundance of the NLRP3 inflammasome and collagen I in kidney tissue homogenate from DOCA-salt–treated mice with no difference between genotype (Figure 5F). Masson trichrome staining did not show evident collagen deposition in kidneys from DOCA-salt–treated mice (Figure S10).

Figure 5.

Deoxycorticosterone acetate (DOCA)-salt increases kidney tissue inflammation, cytokines, and fibrosis. A and B, Plasma levels of TNF (tumor necrosis factor), and IP-10 (interferon gamma-induced protein-10) proteins increased significantly in DOCA-salt male mice samples compared with controls, with no difference between genotypes (n=6–8 per group). C and D, In whole kidney tissue homogenates, TNF and IP-10 proteins increased in DOCA-salt wild-type (WT; n=8) samples compared with WT control male mice (n=8). TNF (C) and IP-10 (D) were significantly higher in DOCA-salt uPA (urokinase-type plasminogen activator) WT than in DOCA-salt uPA KO male mice (n=8). DOCA-salt did not significantly elevate TNF in uPA KO. E, Heatmap summarizing the relative mRNA abundance of fibrosis, pro- and antiinflammatory markers in the kidney tissue of uPA WT and KO male mice subjected to either control or DOCA-salt treatment (n=6–8 per group). F, Immunoblotting of whole kidney tissue homogenate for NLRP3 inflammasome (n=3 per group) and collagen I (n=6 per group) revealed immunoreactive bands that migrated at ≈110 kDa and >250 kDa, respectively. Semiquantification of the immunoreactive band showed elevated levels of NLRP3 and collagen I in DOCA-salt–treated kidney samples compared with controls from male mice, with no genotype-dependent differences. Significance was tested by 2-way ANOVA. *P<0.05, **P<0.001, ****P<0.0001. ns indicates nonsignificant.

Effect of DOCA-Salt on Polarized Macrophages in the Kidneys of WT and Urokinase-Deficient Mice

Macrophage accumulation and polarization in the kidney tissue were assessed in paraffin-embedded sections (n=3 per group). Kidney tissue was labeled for CD16⁺CD32⁺ and CD163⁺, pro- and antiinflammatory macrophage markers, respectively (Figure 6A and 6B). The liver was used as a positive control, and a negative control with omission of primary antibody was included. Semiquantification of immunoreactive cells per field showed an increased abundance of CD16⁺CD32⁺ cells and CD163⁺ cells in DOCA-salt renal cortex compared with control male mice (Figure 6C and 6D). The CD16⁺CD32⁺ cells localized predominantly in the interstitial compartment of the kidney. CD16⁺CD32⁺ cells were significantly more abundant in the DOCA-salt male mice kidney tissue in the cortex compared with CD163⁺ cells (407 CD16⁺CD32⁺ cells per field versus 6 CD163⁺ cells pr field; Figure 6E). CD163⁺ cells localized primarily in the interstitial space in the wider adventitia of larger vessels in both control and DOCA-salt tissue. In the medulla, there were no immunoreactive signals in the control for CD163⁺ and CD16⁺CD32⁺. In DOCA-salt kidneys, there were abundant CD16⁺CD32⁺ cells in the medulla, while there was limited staining for CD163⁺ cells (250 CD16⁺CD32⁺ cells per field versus 5 CD163⁺ cells per field; Figure S11). Immunoblotting of whole kidney tissue homogenates provided corroborative data for CD16+CD32 and CD163. There was no significant difference in abundance of CD16⁺CD32⁺ across all groups, while CD163 was elevated in DOCA-salt–treated mice compared with control, with no genotype-dependent differences (Figure S12).

Figure 6.

Immunostaining reveals macrophage polarization and abundance in deoxycorticosterone acetate (DOCA)-salt kidney tissue. A and B, Detection of macrophages polarized toward the pro- and antiinflammatory like phenotypes by labeling for CD16+CD32 and CD163, respectively, in the renal cortex (n=3 per group). Negative controls were included by omitting the primary antibody. Liver tissue was used as a positive control. The immunoreactive signals for CD16+CD32 and CD163 were quantified observer-blinded by counting positive cells per field within groups. Eight distinct pictures per cortex were obtained, with each image representing 1 field. C and D, Quantification revealed a greater abundance of CD16+CD32⁺ and CD163⁺ cells in DOCA-salt kidneys compared with control male mice. E, The CD16+32/CDD163 ratio indicated a predominance of CD16+CD32⁺ cells compared with CD163⁺ cells in DOCA-salt–treated kidneys. Scalebar 100 µm.

CD3, a marker for T-cells, showed abundant immunoreactivity in the renal cortex of DOCA-salt male mice compared with controls (116 CD3⁺ cells per field versus 21 CD3⁺ cells per field). The immunoreactive protein was localized in the interstitial compartment. Immunoreactivity for the neutrophil marker Ly6g appeared to be more abundant in the renal cortex of DOCA-salt mice compared with control (15 Ly6g⁺ cells per field versus 3 Ly6g⁺ cells per field), however, to a lesser extent compared with CD3⁺. The immunoreactive Ly6g⁺ signal localized in the basolateral nuclei of tubular cells and was absent from glomeruli. The spleen was used as positive control, and a negative control was included with the omission of the primary antibody. Similar trends in CD3 and Ly6g immunoreactivity were observed in the medulla, though at lower levels (Figures S13 and S14).

Discussion

The present study shows that in a mouse model of chronic hypertension with albuminuria caused by reduced renal mass combined with mineralocorticoid (DOCA) and high dietary salt, high–molecular-weight proteinuria was associated with (1) urinary protease activity with plasminogen, active plasmin, and kallikrein, (2) increased abundance and cleavage of γ-ENaC in kidney tissue, (3) increased urinary excretion and renal tissue abundance of complement C3 and anaphylatoxin C3a, and (4) accumulation of tissue injury markers, collagen, proinflammatory cytokines, and CD16⁺CD32⁺ cells. While deletion of the uPA gene did not confer protection from DOCA-salt–induced hypertension, ENaC cleavage, kidney injury, urine kallikrein, and albuminuria, it reduced significantly urine protease activity, presence of active plasmin, urine C3a, and tissue proinflammatory cytokines IP-10 and TNF. Thus, we conclude that uPA-plasmin is not the sole driver of barrier injury, GFR decline, and DOCA-salt hypertension but acts redundantly with other proteases, for example, kallikrein, in intratubular complement activation and kidney tissue inflammation. At the time of termination, mice subjected to DOCA-salt showed high–molecular-weight, glomerular proteinuria, including plasminogen and kallikrein, with hypertension, increased abundance of γ-ENaC, and increased ammonium excretion. While total GFR was lower, single-kidney GFR increased, reflecting the renal functional reserve and the relatively early stage of chronic kidney disease in this model. The renal (and cardiac) hypertrophy was not dependent on uPA. Together, the results indicate that the glomerular hyperfiltration, together with hypertension in DOCA-salt, results in marked kidney glomerular barrier injury and tubular epithelial cell albumin accumulation independent of uPA.

As seen previously in mice with proteinuria, a substantial part of tubular fluid protease activity depends on uPA and plasmin, also in DOCA-salt–treated mice. Despite this finding, the DOCA-driven elevated γ-ENaC protein abundance did not differ in cleavage pattern between genotypes. In vivo and in vitro evidence supports that the 75 kDa immunoreactive band represents furin-cleaved ENaC.^{11–13,20,37} In vivo studies of acute proteinuria identify the extracellular protease-specific cleavage of γ-ENaC at ≈50 kDa,^13,20 which was not changed by uPA deletion, confirming that uPA-plasmin are not solely responsible for cleavage.^20,38 The similar and high levels of ENaC after DOCA were corroborated by the similar BP elevation in the DOCA-salt and ANGII-infusion model.

High molecular glomerular proteinuria is associated with the aberrant filtration of complement protein C3 and its activation split products,^6,25,39 which was confirmed in the present DOCA-salt mice. Active pure plasmin activates C3 in vitro and at sites of coagulation in vivo.²⁷ Of note, endogenous urokinase in urine from healthy individuals is sufficient to activate added plasminogen and subsequently C3 to C3a.²⁶ The present data confirmed nonredundant attenuation of proteinuria-induced urinary C3a excretion in uPA-deleted mice. This confirms that endogenous uPA drives noncanonical complement activation via a plasmin-dependent pathway that bypasses classical C3 and C5 convertases. C5a levels remained unchanged, suggesting a selective activation of the C3 axis in DOCA-salt–induced kidney injury. Apical binding of the anaphylatoxins C3a and C5a to their receptors, C3aR and C5aR, results in the secretion of various proinflammatory cytokines such as TNF and IL-6, promoting immune cell infiltration as well as activating resident immune cells.^40,41 Deletion of uPA mitigated the abundance of cytokines TNF and IP-10 in kidney tissue but not in plasma, indicating a local effect. Given the modest effect of uPA-plasmin on C3a generation and no effect on C5a, along with only mild changes in cytokine levels, no significant differences were observed in tissue macrophage infiltration or fibrosis in DOCA-salt kidneys after 21 days. Since uPA knockout did not affect the severity of hypertension, which by itself is a key driver of inflammation and tissue injury, the limited impact of uPA deletion on renal inflammation and injury is likely attributable to the dominant influence of sustained hypertension.

In vitro, C3a promotes proinflammatory macrophage polarization via ERK signaling, increasing TNF production and contributing to tubular epithelial apoptosis.^42,43 Supporting this, C3-deficient mice subjected to unilateral ureteral obstruction showed reduced TNF, inflammation and renal fibrosis.³⁰ Immunohistochemistry confirmed a predominance of macrophages polarized toward a proinflammatory phenotype independent of uPA and intratubular C3a, suggesting the involvement of parallel or redundant regulatory pathways.

In the kidney, T-cell infiltration is associated with salt-sensitive hypertension, and glomerulonephritis is associated with increased macrophage recruitment but not polarization.^44–46 In accordance, we observe a greater abundance of CD3⁺ cells in the kidney tissue of DOCA-salt mice, with no differences between genotypes. These immune cell dynamics underscore the complexity of the inflammatory response, where variations in individual cytokines may not substantially alter the overall inflammatory state of the kidney.

The administration of amiloride to podocin KO mice with massive proteinuria lowered plasmin activity and urinary C3a-anaphylatoxin excretion. Similarly, monoclonal antibody-mediated uPA inhibition reduced C3a and C5a without affecting albuminuria, kidney tissue inflammation, or fibrosis.²⁶ Redundant mechanisms and factors likely play a role in macrophage polarization beyond C3a-C3aR signaling. Other complement components (eg, C1q, MBL, C3b, C5b-9) may interact with macrophage receptors to regulate cytokine production.⁴² Blocking the uPA-plasmin pathway may provide more substrate for other complement-activating proteases, such as thrombin,^27,47 kallikrein,^7,36 and renin⁴⁸ that could mediate intratubular activation. Indeed, we show that kallikrein cleaves C3 in urine in vitro and is present in urine from mice with proteinuria in vivo. This indicates that significant redundancy exists among the proteases present in tubular fluid in proteinuria, with the ability to generate anaphylatoxins intratubularly. Thus, to target this therapeutically, a broader serine protease class inhibitor should be administered. Emerging data confirm the utility of such approaches, as demonstrated with aprotinin,²⁰ which is effective in Dahl salt-sensitive rats.⁴⁹

The present data indicate that the reduction in BP by administration of amiloride to patients with treatment-resistant hypertension is primarily due to ENaC inhibition, rather than off-target inhibition of uPA.²³ In the 2 different applied hypertension models with variable mineralocorticoid concentrations (DOCA-salt versus ANGII infusion), the deletion of uPA did not alter the kinetics of hypertension development. While the present study does not question that ENaC is significantly involved in the development of hypertension after DOCA-salt⁵⁰ and after ANGII,¹⁹ the data suggest that circulating and aberrantly filtered uPA-plasmin is not essential, likely due to other proteases such as kallikrein with the ability to cleave ENaC.^7,51

Complement anaphylatoxins and proinflammatory cytokines are considered direct prohypertensive mediators.^52,53 However, despite lower C3a and proinflammatory cytokines in renal tissue, the mean arterial BP remained similar between DOCA-salt–treated genotypes, as were plasma cytokines. In line, deficiency of C3aR/C5aR in mice does not protect against ANGII- or DOCA-salt–induced hypertension.⁵⁴ Administration of soluble complement receptor 1 inhibitor also did not affect the developing kinetics of salt-sensitive hypertension and proteinuria in vivo.⁵⁵ Conversely, C3 deficiency or treatment with the TNF blocker etanercept reduces systolic BP in salt-sensitive hypertensive rodents.^56,57 Indeed, DOCA-saltelevated proinflammatory macrophage-associated cytokines IL-6, TNF, IP-10, and MCP-1 in plasma, and accumulation of inflammatory cells in the kidney, could account for the similar BPs. This underscores the complexity of the immune and complement systems in hypertensive end-organ damage.

Conclusions

Urokinase is the primary activator of plasminogen in the tubular lumen during proteinuria. Although uPA is not required for filtration barrier injury or hypertension in mineralocorticoid-salt–induced kidney injury, uPA contributes to intratubular generation of the C3-derived anaphylatoxin C3a in a plasmin-dependent manner and to accumulation of cytokines in kidney tissue. Redundant cleavage of ENaC and activation of C3 by other aberrantly filtered proteases, such as kallikrein, likely limits the specific role for the aberrant presence of uPA-plasmin in inflammation and kidney injury.

Perspectives

The presence of multiple precursors and enzymes in tubular fluid and subsequently in urine in conditions with albuminuria results in urokinase-dependent continuous activation of plasmin, which may contribute redundantly with other proteases, for example, kallikrein to innate immune system complement activation, formation of anaphylatoxins, proinflammatory macrophage accumulation, and activation of ENaC. Urokinase-plasmin proteolytic activity is not necessary for decline in GFR, albuminuria, fibrosis, and hypertension in the DOCA-salt model in the initial phase. Urokinase-plasmin with other proteases may sustain interstitial inflammation and ultimately promote nephron loss and fibrosis. Albuminuria is not only an indicator of filtration barrier injury but also of serine protease-mediated continuous anaphylatoxin and cytokine formation.

ARTICLE INFORMATION

Acknowledgments

The authors thank PhD fellows and MD related to the research group for guidance and advice regarding laboratory practices and discussion of data interpretation. Laboratory technicians are thanked for assistance with assays.

Author Contributions

M.L. Bach and B.L. Jensen drafted the article. M.L. Bach, A. Heinl, and S.S. Thangaraj performed laboratory experiments. M.L. Bach, S.S. Thangaraj, and C. Enggaard performed animal experiments. M.L. Bach performed statistical analyses. P. Svenningsen, Y. Palarasah, and B.L. Jensen designed the study.

Sources of Funding

The study was funded by the Independent Research Fund Denmark (1030-00330B), Augustinus Fonden, A.P. Møller Fonden, The Novo Nordisk Foundation (NNF19OC0058780 and from 2024 NNF23OC0085257), and Helen og Ejnar Bjørnows Fond.

Disclosures

None.

Supplemental Material

Detailed Methods and Materials

Figures S1–S14

Tables S1–S2

Major Resources

Noncropped Versions of Western Blots

Supplementary Material

hyp-83-e25050-s001.pdf ^{(2.2MB, pdf)}

hyp-83-e25050-s002.docx ^{(33KB, docx)}

hyp-83-e25050-s003.pdf ^{(817.2KB, pdf)}

hyp-83-e25050-s004.pdf ^{(22.7KB, pdf)}

hyp-83-e25050-s005.pdf ^{(22KB, pdf)}