Podocyte injury: the role of proteinuria, urinary plasminogen, and oxidative stress

Abstract

Podocytes are the key target for injury in proteinuric glomerular diseases that result in podocyte loss, progressive focal segmental glomerular sclerosis (FSGS), and renal failure. Current evidence suggests that the initiation of podocyte injury and associated proteinuria can be separated from factors that drive and maintain these pathogenic processes leading to FSGS. In nephrotic urine aberrant glomerular filtration of plasminogen (Plg) is activated to the biologically active serine protease plasmin by urokinase-type plasminogen activator (uPA). In vivo inhibition of uPA mitigates Plg activation and development of FSGS in several proteinuric models of renal disease including 5/6 nephrectomy. Here, we show that Plg is markedly increased in the urine in two murine models of proteinuric kidney disease associated with podocyte injury: Tg26 HIV-associated nephropathy and the Cd2ap^−/− model of FSGS. We show that human podocytes express uPA and three Plg receptors: uPAR, tPA, and Plg-RKT. We demonstrate that Plg treatment of podocytes specifically upregulates NADPH oxidase isoforms NOX2/NOX4 and increases production of mitochondrial-dependent superoxide anion (O₂⁻) that promotes endothelin-1 synthesis. Plg via O₂⁻ also promotes expression of the B scavenger receptor CD36 and subsequent increased intracellular cholesterol uptake resulting in podocyte apoptosis. Taken together, our findings suggest that following disruption of the glomerular filtration barrier at the onset of proteinuric disease, podocytes are exposed to Plg resulting in further injury mediated by oxidative stress. We suggest that chronic exposure to Plg could serve as a “second hit” in glomerular disease and that Plg is potentially an attractive target for therapeutic intervention.

normal urine is essentially protein free. Proteinuria of increased severity is associated with progression of chronic kidney disease (CKD) independent of the etiology and baseline glomerular filtration rate (45, 58). Specialized endothelial cells, the glomerular basement membrane (GBM), and podocytes work in partnership to preserve the permselectivity of the kidney filter (24, 67, 74, 79). Podocytes, the component of the glomerulus most susceptible to injury (8, 20, 22, 24, 36, 74), are highly differentiated cells that adhere to the GBM. Podocytes are critical for maintaining the integrity of the kidney filter (13, 22, 24, 36, 54, 60, 67). Independent of the cause of injury, podocyte response follows similar pathways including disorganization of the actin cytoskeleton, foot process effacement, loss of the slit diaphragm (SD), detachment and death (24, 36, 67, 74, 75, 93). Podocytes' ability to undergo cell division and compensate for injury and loss is limited (20, 24, 36, 67, 74, 92, 93). When podocyte loss exceeds 30%, persistent nonselective proteinuria, progressive glomerulosclerosis, and progressive CKD ensue (20, 36, 91–93).

Clinically, proteinuria accompanied by continued podocyte loss and CKD progression is a hallmark of many glomerular diseases of different etiology including diabetic nephropathy and focal segmental glomerular sclerosis (FSGS) (8, 13, 24, 36, 54, 60, 67). To date, it remains unclear whether persistent injury is driven by persistence of the initial pathogenic process or by a superimposed pathogenic process, a so-called “second hit model of podocyte injury.” The latter raises the question of whether continuous trans-glomerular passage of large circulating macromolecules might contribute to podocyte damage as a “second hit” (20, 24, 36, 46, 47, 54, 68, 93). In this context, studies have shown that albuminuria itself may contribute to tubule-interstitial disease and to glomerulosclerosis (11, 53).

Clinically, most of the mechanisms linking proteinuria to CKD progression remain to be fully defined (9, 13, 22, 54, 68). Data from clinical and animal models have documented significant quantities of biologically active urinary plasminogen (Plg) and the protease plasmin in nephrotic range proteinuria of diverse etiologies (6, 81). Furthermore, in patients in remission from nephrotic syndrome the urinary Plg/plasmin content is minimal (1), thereby supporting the concept that under physiologic conditions podocytes are not exposed to large concentrations of Plg; a fact further supported by its molecular size (81 kDa), which is larger than albumin (3).

It is well-established that after binding onto cell surface receptors (42, 51, 52, 59) Plg is protected from inhibitors, undergoes proteolytic activation, and converts to the serine protease plasmin. In many cell types, plasmin initiates responses linked to both its proteolytic as well as its nonproteolytic actions including cleavage of matrix proteins (62, 87), inactivation of vascular endothelial growth factor (69), induction of reactive oxygen species (ROS) (85), as well as activation of transforming growth factor (TGF)-β (23, 34, 43).

Here, we demonstrate that Plg is increased in the urine in two murine models of proteinuric kidney disease associated with podocyte injury: the Tg26 HIV-associated nephropathy mouse (37) and the Cd2ap^−/− model of focal segmental glomerulosclerosis (76). We show that immortalized human podocytes constitutively express and synthesize urokinase-type plasminogen activator (uPA) as well as the Plg receptors uPAR, tPA, and the novel trans-membrane receptor Plg R-KT.

Our studies show, to our knowledge for the first time, that podocyte binding of Plg upregulates NADPH oxidase isoforms NOX2 and NOX4 and elicits mitochondrial production of superoxide anion (O₂⁻) accompanied by 1) upregulation of the multifunctional scavenger receptor CD36 (77, 96) resulting in increased podocyte uptake of oxidized LDL (oxLDL) with subsequent apoptosis and 2) induction of podocyte synthesis of endothelin-1, a molecule that acts in a paracrine/autocrine fashion and has been implicated in podocyte-endothelial cell cross talk and glomerular disease progression (14, 35, 41, 61). Epsilon aminocaproic acid (EACA), an inhibitor of cellular Plg binding and activation (51, 52), prevented all the injurious actions elicited by exposure of podocytes to Plg. In addition, we demonstrated that Apocynin, AMPkinase, and Mito Tempo, molecules that mitigate production of ROS (18, 19, 40), provided similar podocyte protection.

These data strongly support the conclusion that Plg-plasmin is a molecule that via previously unrecognized mechanisms that include oxidative stress can play a role as a promoter/contributor of a “second hit” type of podocyte injury in proteinuric nephropathies. Our studies may contribute to development of therapeutic interventions in proteinuric kidney disease.

METHODS

Mouse studies.

Urine was collected from two proteinuric kidney disease models that have been well-characterized: 5-wk-old Cd2ap^−/− mice (25, 72, 76, 90, 94) and 6-wk-old Tg26 mice (2, 4, 21, 32, 37, 38, 65). Urinary Plg was measured using 2 μl urine with Elisa Kit (ICL E-90PMG) read at absorbance at 450 nm. Mouse creatinine was measured with an enzymatic assay kit (Crystal Chem: no. 80350).

Human podocyte culture.

Human podocytes initially generated by MA Saleem Children's Renal Unit and Academic Renal Unit, University of Bristol, UK (71) were generously provided by S. Merscher-Gomez and A. Fornoni, University of Miami, who also provided initial technical supervision of podocyte culture as reported (50, 64, 71). Briefly, human podocytes were cultured and differentiated in RPMI 1640 culture medium containing 10% FBS, 1% penicillin/streptomycin with or without 1% ITS, 50 U/ml human interferon as described (50, 71). The immortalized normal human podocytes were propagated at 33°C and then thermoshifted for differentiation for 14 days at 37°C. Terminally differentiated podocytes were starved in serum-free RPMI 1640 medium for 24 h before the experiments were performed (24, 50, 64, 71).

Determination of O₂⁻ production.

O₂⁻ production was determined by using lucigenin-enhanced chemiluminescence as described in our previous publications (96). Human podocytes were treated with Plg (1 μmol/l), the concentration of Plg found in human serum (51, 52), for 6 h as determined in experiments of time responses to Plg up to 24 h (not shown).

To determine the source of ROS, the cells were preincubated with the following compounds for 30 min before incubation with Plg: 1) NADPH oxidase inhibitors diphenyleneiodonium (DPI, 1 μmol/l, Sigma) and APOCY (100 μmol/l, Sigma) (12, 39, 40, 96), 2) AICAR, an agonist to AMP-activated protein kinase (100 μmol/l, Sigma) (19, 52, 86), 3) EACA (0.2 mol/l, Sigma) (42, 52), and 4) amiloride (75 μM, Sigma).

Immunoblot analyses.

The cells were harvested with lysis buffer containing a cocktail of protein inhibitor (Complete Mini, Roche Diagnostics GmbH, cat. no. 11836153001). Protein was quantified by Bio-Rad assay, and 30 mg of total protein were first subjected to SDS-PAGE and then transferred to nitrocellulose membranes. The membranes were incubated overnight in a cold room with primary rabbit anti-CD36, anti-tPA, anti-WT1, and anti-uPAR (Santa Cruz Biotechnology), anti-phospho-AMPK, anti-AMPK, rabbit anti-NOX 4 (Santa Cruz Biotechnology), and anti-NOX2 (AbCam) followed by incubation with a peroxidase-conjugated secondary antibody for 1 h. All the Ab used equivalence-of-protein loading and transfer was confirmed by reblotting the samples with anti-β-actin antibody. Immune reactive bands were detected by chemiluminescence and quantified by densitometry. Relative quantities of each protein were normalized by β-actin (Santa Cruz Biotechnology) and expressed as fold increase vs. control (96).

Determination of total cellular cholesterol content.

Human podocytes in six-well dishes were incubated with or without Plg (1 μmol/l) for 24 h followed by incubation with oxLDL (50 μg/ml) for another 24 h (96). Total cellular cholesterol was extracted by adding 1 ml hexane/isopropanol (3:2, vol:vol) to the wells. The samples were evaporated by using a SpeedVac to remove the solvent and then redissolved in 50 ml ethanol. Cholesterol was determined with an enzyChrom AF cholesterol assay kit according to the manufacturer's instructions (Bioassay Systems, Hayward, CA). The assay results were normalized according to the sample protein content.

Apoptosis assay.

A total of 2 × 10⁶ cells was seeded in 75-cm² flasks and treated for 24 h with one of the following conditions: vehicle, Plg (1 μmol/l), oxLDL (50 μg/ml), or incubation with plasminogen followed by oxLDL (50 μg/ml). The cells were trypsinized and washed with PBS twice. Apoptosis was measured using the annexin V-fluorescein isothiocyanate apoptosis detection kit (Roche Diagnostics, Indianapolis, IN) followed by flow cytometry.

Statistics.

The results were expressed as means ± SD. Statistical analyses were performed by ANOVA with Bonferonni's correction for multiple comparisons, followed by Scheffé's test. Significance was assumed at P < 0.05.

RESULTS

Urinary plasminogen is increased in proteinuric mouse models.

We measured urinary Plg normalized to urinary creatinine in two independent models of proteinuric kidney disease. Cd2ap^−/− mice provide a well-characterized model of focal segmental glomerulosclerosis with proteinuria developing between 2 and 3 wk of age. Urinary Plg from 5-wk-old Cd2ap^−/− mice (mean urine protein/cr ratio = 125.4 mg/mg) was markedly elevated compared with wild-type littermate controls (30.96 vs. 0.09 ng/mg; P: 0.007; Fig. 1). We also detected elevated urine Plg in 6-wk-old mice from the Tg26 model (mean urine protein/creatinine ratio 83.1 mg/mg) of HIV nephropathy (2.56 vs. 0.49 ng/mg; P: 0.008).

Fig. 1.

Urine plasminogen (Plg) levels are increased in mouse models of proteinuric kidney disease. Urine Plg/Cr is significantly increased in 6-wk-old Tg26 and 5-wk-old Cd2ap^−/− mice. WT, wild-type. **P < 0.005.

Plasminogen receptor expression and superoxide upregulation in podocytes.

In cultured human podocytes, we established the expression of uPA as well as three well-characterized Plg receptors that are known to be blocked by the lysine analog EACA: uPAR, tPA, and the novel receptor R-KT (Fig. 2A) (42, 51, 52, 59). Previous studies in vivo and in vitro demonstrated that several isoforms of NADPH oxidase including NOX2 and NOX4 are present in podocytes and that ROS originating from NADPH oxidase play an important role in ROS-mediated podocyte injury. NOX4 is the most abundant isoform in podocytes (10, 19, 31, 80). RTPCR (Fig. 2, B and C) and Western blotting (Fig. 2, D and E) showed that Plg significantly upregulated podocyte NOX4 and NOX2 expression. This was prevented by prior incubation with EACA, thereby demonstrating that Plg binding and activation were required (51, 85).

Fig. 2.

Plasminogen upregulates NADPH oxidase isoforms NOX2 and NOX4: Plg upregulated mRNA and protein expression of NOX2 and NOX4 isoforms of NADPH oxidase that was significantly prevented by inhibition of Plg binding/activation by epsilon aminocaproic acid (EACA). M = marker. Data were expressed as means ± SE. *P < 0.05 vs. control. #P < 0.05 vs. plasminogen.

Exposure of podocytes to Plg increased NADPH oxidase-mediated O₂⁻ production (Fig. 3A) that was prevented by the inhibitors of NADPH oxidase diphenyleneiodonium (DPI) and by apocynin (APOCY) (39, 40).

Fig. 3.

Plasminogen induces production of superoxide anion in human podocytes. A: Plg binds to human podocytes, is activated, and induces production of superoxide anion. Plg (1 μmol/l, the Plg concentration in serum) significantly increased podocyte O₂⁻ production; increased O₂⁻ production was prevented by the NADPH oxidase and flavin-containing oxidases diphenyleneiodonium (DPI) 1 μmol/l, the NADPH oxidase inhibitor APOCY (100 μmol/l), the agonist of AMP-activated protein kinase AICAR (100 μmol/l), as well as the inhibitor of Plg/plasmin activation EACA (0.2 mol/l). Data was expressed as means ± SE. *P < 0.05 vs. control. #P < 0.05 vs. plasminogen; n = 6. B: pAMPK is active AMPK. Exposure of podocytes to Plg reduced pAMPK/AMPK ratio, and increased NADPH oxidase-mediated O₂⁻ production. Activation of AMPK by AICAR, 5-aminoimidazole-4-carboxamide-1-riboside (52–54, 63) significantly prevented these effects by increasing the pAMPK/AMPK ratio and concomitantly reducing O₂⁻ production. AMPK is a ubiquitous serine/threonine kinase that acts as an energy sensor and is expressed in the plasma membrane of podocytes; pAMPK suppresses NADPH oxidase (52–54, 63). Data are expressed as means ± SE. *P < 0.05 vs. control. #P < 0.05 vs. plasminogen; n = 5.

Plg concomitantly reduced the pAMPK/AMPK ratio (Fig. 3B). AICAR, 5-aminoimidazole-4-carboxamide-1-riboside, is known to phosphorylate and activate AMPK (18, 86). Pretreatment of podocytes with AICAR before exposure to Plg significantly prevented the reduction in pAMPK/AMPK ratio and concurrently inhibited O₂⁻ production (Fig. 3A). This suggests an important role of this stress-sensing kinase in the biology of podocyte injury associated with oxidative stress (18, 86).

Expression of the scavenger receptor CD36 is upregulated by Plg.

The B scavenger receptor CD36, which is present in many cells including podocytes (49), recognizes, binds, and internalizes oxLDL (77, 96). Plg significantly upregulated podocyte expression of CD36 (Fig. 4A). This paralleled the increased production of O₂⁻ induced by Plg (Fig. 3A). The upregulation of CD36 was prevented by EACA as well as by the NADPH inhibitor APOCY and AICAR, the agonist of AMPK (18, 19, 86). Importantly, the significant inhibition of Plg upregulation of CD36 by EACA, APOCY, and AICAR indicates the interdependence between Plg activation, increased oxidative stress, and CD36 upregulation in podocytes.

Fig. 4.

Expression of the scavenger receptor CD36 is upregulated by plasminogen. A: effects of Plg on scavenger receptor CD36 expression in human podocytes. Plg significantly increased CD36 expression, which was prevented by the NADPH oxidase inhibitor APOCY (10 μmol/l), AICAR (10 μmol/l) an agonist of AMP-activated protein kinase, as well as the inhibitor of Plg (plasmin) activation EACA (0.2 mol/l). Data are expressed as means ± SE. *P < 0.05 vs. control. #P < 0.05 vs. plasminogen; n = 5. B: effect of Plg and oxidized LDL (oxLDL) on total intracellular cholesterol content in human podocytes. Incubation of human PODC with oxLDL (50 mg/ml) resulted in a small but significant increase in total intracellular cholesterol content. Addition of Plg in the presence of oxLDL further increased intracellular cholesterol content. Data are expressed as means ± SE. *P < 0.05 vs. control and control + oxLDL. #P < 0.05 vs. control group; n = 6 in each group.

Our group and others have shown that upregulation of CD36 in human macrophages increases uptake of oxLDL accompanied by increased synthesis of ROS (77, 96). It is presently unknown whether upregulation of CD36 plays a similar role in podocytes. Similar to macrophages, podocytes treated with oxLDL for 24 h showed a significant increase in cholesterol accumulation (Fig. 4B) (96), which was further increased in podocytes treated with Plg plus oxLDL thereby suggesting that the increased oxLDL uptake was linked to Plg upregulation of CD36 (77, 96). The off-target effects of the diuretic amiloride as an inhibitor of uPA are well-documented (30, 48, 88). We found that amiloride inhibited Plg-induced upregulation of CD36 (Fig. 5A) and ROS superoxide generation (Fig. 5B).

Fig. 5.

Amiloride regulation of Plg-induced CD36 and reactive oxygen species (ROS) generation in podocytes. Amiloride abrogates Plg induction of CD36 (A) and ROS (B) in podocytes. **P < 0.005, #P < 0.05 vs. plasminogen, ****P < 0.0001.

Plg enhances oxLDL-mediated podocyte apoptosis.

Based on our novel findings that Plg upregulates CD36 (Fig. 4A) and increases oxLDL uptake (Fig. 4B), we investigated by flow cytometry the effects of oxLDL and Plg individually as well as together with oxLDL on podocyte apoptosis. As shown in Fig. 6, both Plg as well as oxLDL alone slightly increased podocyte apoptosis. Importantly, the combination of Plg and oxLDL had a clear additive effect that further significantly augmented podocyte apoptosis.

Fig. 6.

Exposure of human podocytes to Plg plus oxLDL has additive effects in promoting podocyte apoptosis. A: both Plg and oxLDL induced slight but significant podocyte apoptosis after 24 h. The combination of Plg and oxLDL had a clear additive effect on podocyte apoptosis driven by increased oxLDL uptake associated with Plg upregulation of scavenger receptor CD36 (Fig. 4, A and B). B: podocyte apoptosis measured by flow cytometry. Data were expressed as means ± SE. *P < 0.05 vs. control.

Plg promotes endothelin-1 synthesis.

Studies in vivo and in vitro showed that endothelin-1 (ET1), acting in paracrine/autocrine fashion, induces oxidative stress and injury of glomerular endothelial cells and podocytes (14, 35, 41). It is currently unknown whether Plg promotes ET1 synthesis by podocytes and whether it is linked to oxidative stress by acting as a second messenger. As shown in Fig. 7B, we demonstrated that Plg promotes synthesis-release of ET1 that was prevented by EACA, APOCY, and by the selective inhibitor of mitochondria ROS Mito Tempo (Fig. 7, A and B) (14). This demonstrated for the first time the important role of Plg-induced ROS, particularly mitochondrial ROS in promoting podocyte ET1 synthesis.

Fig. 7.

Plasminogen-promoted synthesis of endothelin 1 was specifically prevented by Mito-Tempo an inhibitor of mitochondrial ROS. A: Plg induced O₂⁻ production in podocytes that was partially prevented (40%) by Mito Tempo, a selective inhibitor of mitochondrial oxidative stress. B: endothelin 1 synthesis/release was measured by ELISA in the supernatants of the various conditions. Plg increased podocyte synthesis of endothelin 1 that was inhibited by APOCY and by Mito-Tempo. Important to note is that while Mito-Tempo only reduced ROS by 40% (A), it completely prevented endothelin 1 synthesis in response to Plg, suggesting that O₂⁻ from mitochondria plays a critical role in modulating podocyte endothelin 1 synthesis. Data were expressed as means ± SE. *P < 0.05 vs. control. #P < 0.05 vs. plasminogen.

DISCUSSION

Independent of the etiology, chronic severe nonselective proteinuria is a harbinger of progression of CKD (24, 27, 67, 74, 79). It has long been suspected that the relentless trans-glomerular passage of selected plasma macromolecules may contribute to podocyte injury as a “second hit.” Studies have shown that albuminuria itself may contribute to glomerular injury directly and indirectly via podocytes uptake of fatty acids (FFA) bound to albumin (11) as well as by sequestration of retinoic acid, a molecule shown to be critical in protecting podocytes from injury (44, 57). Data from both patients and experimental animals with nephrotic syndrome have consistently shown their urine to contain large quantities of biologically active Plg and the protease plasmin (1, 5, 6, 81–84, 95).

It has been demonstrated that plasmin in nephrotic urine proteolytically activates ENaC, an effect that disappears during remission of the nephrotic syndrome (1, 5, 6, 81–84, 95). In vitro and in vivo studies demonstrated the ability of plasmin to contribute to cellular injury and to inflammation via its proteolytic as well as its nonproteolytic actions (85). Interestingly, previous studies in animals and in patients with proteinuric nephropathies treated with EACA, the inhibitor of Plg activation to plasmin (51, 52, 85), reported a marked reduction of proteinuria and glomerular injury (28).

We utilized two independent well-characterized models of progressive proteinuric kidney disease, the Cd2ap^−/− mouse model of FSGS and the Tg26 model of HIVAN. In both, urinary Plg was significantly elevated (Fig. 1). We also used cultured human podocytes (71) to further assist us in determining whether the Plg/plasmin system could potentially participate as a “second hit” type of podocyte injury linked to proteinuria.

In human podocytes, we confirmed the expression of urokinase plasminogen activator uPA that is synthesized by podocytes and of Plg receptors uPAR, tPA, and Plg R-KT (51, 52) (Fig. 2A). uPAR has high-affinity for uPA, but has no intracellular domain, which limits its signaling capacity. Plg R-KT colocalizes with uPAR and is a unique trans-membrane protein that activates Plg because it exposes a COOH-terminal lysine on the cell's surface (51, 52). In addition, Plg R-KT specifically interacts with tPA, which secures Plg on the cell surface and promotes its activation to plasmin that once bound to the cell surface is protected from its physiological inhibitors (51, 52, 85).

Here, we demonstrated that Plg upregulated mRNA and protein expression of the podocyte NOX2 and NOX4 (Fig. 2) (10, 31), accompanied by significant production of the ROS superoxide anion O₂⁻ as determined by lucigenin chemiluminescence. EACA prevented NOX2 and NOX4 upregulation as well as O₂⁻ production, thereby demonstrating that these effects are associated with Plg binding and activation (Fig. 2, B–E). Next, we determined that Apocynin, an inhibitor of NADPH oxidases, which are considered to be the major source of ROS-mediated podocyte injury (10, 31), prevented the increase in O₂⁻ production induced by Plg (Fig. 3A).

AMPK is a serine/threonine kinase that is present in podocytes and acts as an energy sensor (7, 47). Once AMPK is phosphorylated (pAMPK), it becomes activated and enables cells to reduce NADPH oxidase-dependent ROS formation (18, 19, 70, 86). We determined that exposure of podocytes to Plg reduced pAMPK (Fig. 3B) and that AICAR, an activator of AMPK (18, 19, 70), normalized pAMPK and concomitantly inhibited Plg-driven O₂⁻ production (Fig. 3, A and B). Of note, AICAR has been shown to have beneficial effects in diabetic nephropathy via similar mechanisms to those reported here (18, 19, 86). Hence, AMPK may have a universal role in mitigating podocyte pathobiology linked to oxidative stress (10, 18, 19, 70, 86).

Podocyte ROS originate in the cytosol and particularly in mitochondria, a major source of intracellular ROS production under physiologic and pathologic conditions (10, 14, 18, 31). Many studies have shown that sustained mitochondria ROS accumulation results in mitochondrial dysfunction (10, 14, 18). Here, we showed that Mito Tempo, a selective inhibitor of mitochondrial ROS, reduced by 40% podocyte O₂⁻ induced by Plg (Fig. 6A). Based on our novel findings with AICAR and Mito Tempo, we propose that both agents may have synergistic therapeutic benefits in proteinuric podocytopathies.

CD36 is an innate multifunctional membrane scavenger receptor that plays an important mechanistic role in atherosclerosis and in inflammation (77, 96). The B scavenger receptor CD36 is constitutive in many cells including podocytes, and recognizes, binds, and internalizes oxLDL and FFA (49, 56, 77, 96). We showed in human macrophages that ROS upregulate CD36 (96) and foster increased uptake of OxLDL and foam cell formation. Mayrhofer et al. (49) showed in preparations of glomeruli isolated from normal rats that CD36 is constitutively expressed in podocytes; in the same studies, they demonstrated that in rats with PAN nephrosis podocyte CD36 was upregulated and promoted uptake of FFA and oxLDL resulting in podocyte injury and development of FSGS. Furthermore, a recent study by Souza et al. (78) demonstrated that the apolipoprotein A-I-mimetic peptide 5A, a CD36 antagonist, reduced glomerular and tubular injury as well as CKD progression independent of blood pressure in 5/6 nephrectomized mice subject to ANG II infusion (78). Similar renoprotection was observed in 5/6 nephrectomized CD36KO mice (78). In the aggregate, these in vivo studies support the notion that CD36 plays an important role in the pathogenesis of progressive CKD.

Here, we determined that Plg significantly upregulated CD36 expression (Fig. 4A). This was prevented by inhibition with EACA of Plg binding as well as by inhibition of O₂⁻ synthesis by either APOCY or AICAR (Fig. 4A). Importantly, in subsequent experiments we determined that podocytes take up OxLDL and that incubation with Plg had an additive effect that resulted in increased podocyte oxLDL uptake (Fig. 4B) accompanied by podocyte apoptosis (Fig. 6).

The off target effect of amiloride as an inhibitor of uPA was documented in early studies reported in the cancer literature (66, 89). In vivo experimental studies in stroke-prone spontaneously hypertensive rats, Dahl salt-sensitive, and in 5/6 nephrectomy rats showed that amiloride significantly reduces glomerular injury and proteinuria independent of blood pressure (26, 33, 73). In patients with type 2 diabetic nephropathy, amiloride lowered urine plasminogen excretion and activation, and albumin/creatinine ratio (55). In the current studies, we determined that amiloride inhibits Plg-induced O₂⁻ production and CD36 upregulation in human podocytes (Fig. 5). This might explain, at least in part, the mechanisms involved in the salutary effects of amiloride in clinical and experimental proteinuric renal disease. These studies also provide strong support to our hypotheses regarding the potential role of Plg-plasmin in proteinuric renal diseases.

Elegant studies of biopsies from patients with FSGS presented evidence of endothelial-mitochondrial oxidative stress associated with podocyte-endothelial cell cross talk mediated by ET1 synthesized by podocytes (14, 15, 35, 41). Here, we demonstrated that Plg-induced synthesis and release of ET1 were inhibited by pretreatment of podocytes with either EACA or the NADPH inhibitor APOCY (Fig. 7B) (39, 40). Importantly, we made the new observation that selective inhibition of mitochondrial ROS with Mito-Tempo was sufficient to completely prevent Plg-induced ET1 synthesis (Fig. 7B) (14).

Cross talk between mitochondria and NADPH oxidases is well-documented and likely represents a feed-forward vicious cycle of ROS production in which activation of NADPH oxidases increases production of mitochondrial ROS and vice versa (16, 17).

In podocytes, Nox4 is predominately localized in mitochondria (19, 31) and has been shown to produce hydrogen peroxide (H₂O₂). Our results with apocyanin that inhibits NOX2 and Mito Tempo would suggest (Fig. 8) that Plg activation of NOX2 generates O₂⁻, which activates mitochondrial Nox4 that reciprocally activates NOX2 via H₂O₂ (16, 17).

Fig. 8.

Working model of the participation of plasminogen-plasmin as a “second hit” type of podocyte injury associated with proteinuria.

Overall, our studies elucidated important new mechanisms that identified apoptosis, linked to the interaction of Plg-CD36-ox-LDL as a potential “second hit” in podocyte injury and death associated with proteinuria (Fig. 8) (20, 36, 47, 54, 68, 91, 93). Based on our studies, we conclude that 1) Plg activation on podocyte membranes can play a significant role in a “second hit” type of podocyte injury associated with chronic proteinuria independent of the initial etiology and 2) that oxidative stress generated by Plg via NADPH oxidases plays a central role in these processes.

At the present time, specific therapies for podocyte injury are not available. Our novel studies suggest that to slow CKD progression of proteinuric glomerular diseases would require multipronged strategies combining pharmacologically new molecules added to other agents such as biologicals that specifically target podocyte antigens, mitochondrial targeted antioxidants, activators of AMPKinase, ET1 receptor blockers/synthesis inhibitors (14, 15, 18, 19, 35, 70), certainly RAAS inhibitors/blockers, (8, 29, 63), and amiloride, cautiously monitoring serum potassium. Further studies will be required to further define the pathogenic role of circulating Plg and potential benefits of its removal or inhibition. In summary, we suggest that our studies provide a new biologic framework for the development of strategies for “comprehensive treatments” to arrest progression of proteinuric podocytopathies.

GRANTS

This work was supported by funds from the South Florida Veterans Affairs Foundation For Research and Education (SFVAFRE) and National Institutes of Health Grant R01DK103022 to K. N. Campbell.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

L.R. and K.N.C. conception and design of research; L.R., R.T., J.S.W., and J.C.H. analyzed data; L.R. interpreted results of experiments; L.R. drafted manuscript; L.R. and K.N.C. edited and revised manuscript; L.R. and K.N.C. approved final version of manuscript; R.T., J.S.W., and J.C.H. performed experiments; R.T., J.S.W., and J.C.H. prepared figures.