Podocyte protease activated receptor 1 stimulation in mice produces focal segmental glomerulosclerosis mirroring human disease signaling events

Abstract

About 30% of patients who have a kidney transplant with underlying nephrotic syndrome (NS) experience rapid relapse of disease in their new graft. This is speculated to be due to a host-derived circulating factor acting on podocytes, the target cells in the kidney, leading to focal segmental glomerulosclerosis (FSGS). Our previous work suggests that podocyte membrane protease receptor 1 (PAR-1) is activated by a circulating factor in relapsing FSGS. Here, the role of PAR-1 was studied in human podocytes in vitro, and using a mouse model with developmental or inducible expression of podocyte-specific constitutively active PAR-1, and using biopsies from patients with nephrotic syndrome. In vitro podocyte PAR-1 activation caused a pro-migratory phenotype with phosphorylation of the kinase JNK, VASP protein and docking protein Paxillin. This signaling was mirrored in podocytes exposed to patient relapse-derived NS plasma and in patient disease biopsies. Both developmental and inducible activation of transgenic PAR-1 (NPHS2 Cre PAR-1^Active+/-) caused early severe nephrotic syndrome, FSGS, kidney failure and, in the developmental model, premature death. We found that the non-selective cation channel protein TRPC6 could be a key modulator of PAR-1 signaling and TRPC6 knockout in our mouse model significantly improved proteinuria and extended lifespan. Thus, our work implicates podocyte PAR-1 activation as a key initiator of human NS circulating factor and that the PAR-1 signaling effects were partly modulated through TRPC6.

Keywords: PAR-1; circulating factor; nephrotic syndrome; podocyte; proteases.

Figure 1. Protease-activatedreceptor 1 (PAR-1) is highly expressed in the podocyte, and its activation is detrimental.

(a) Data from Sigma’s Protein Atlas show an enrichment of PAR-1 expression within the glomeruli of the kidney. All densitometry graphs show the optical density of the band, normalized to the control and corrected by β-actin load control. An example blot is shown beneath each graph. Wild-type conditionally immortalized human podocytes were treated with a PAR-1 agonist at a dose of 15 μM for the indicated time points. There was significant phosphorylation of VASP (b), p38 mitogen-activated protein kinase (MAPK) (c), JNK (d), and Paxillin (e) Bonferroni’s multiple comparison test, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. A range of pathways were interrogated that were not significantly stimulated (f). A wound-healing assay was performed to assess the ability of the PAR-1 agonist to induce a motile phenotype (g). Data shown n = 3 in duplicate, normalized to untreated control. Both the PAR-1 agonist and thrombin treatments significantly increase podocyte motility (1-tailed Mann-Whitney test P = 0.0011 and 0.0022, respectively). pJNK, phospho–c-Jun N-terminal kinase; pVASP, phospho–vasodilator-stimulated phosphoprotein.

Figure 2. Protease-activated receptor 1 (PAR-1)–associated signaling pathways are activated in human nephrotic syndrome.

PAR-1 agonist treatment at a concentration of 15 μM showed significant stimulation of (a) p38 mitogen-activated protein kinase (MAPK), (b) JNK, and (c) vasodilator-stimulated phosphoprotein (VASP) signaling pathways (Bonferroni’s multiple comparison test). The transient receptor potential cation channel subfamily c member 6 (TRPC6) inhibitor SAR7334 (10 nM with 30-minute preincubation) was capable of significantly dampening the response of the podocyte to PAR-1 agonist treatment (a–c) (Bonferroni’s multiple comparison test). Podocytes treated with relapse plasma demonstrated significant stimulation of (d) pp38 MAPK, (e) phospho–c-Jun N-terminal kinase (pJNK), and (f) phospho–vasodilator-stimulated phosphoprotein (pVASP) (Bonferroni’s multiple comparison test) compared with podocytes treated with remission plasma. A further treatment with SAR7334 significantly decreased the stimulation (d–f) (paired 1-way t test). The densitometry shown in (a)–(c) is based on 4 western bots, whereas that shown in (d)–(f) is based on 6, normalized to β-actin load control and relative to control lane. A representative blot is shown beneath each graph. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

Figure 3. Protease-activated receptor 1 (PAR-1)–associated signaling pathways in human kidney biopsies.

Human biopsy samples were stained for (a) phospho–vasodilator-stimulated phosphoprotein (pVASP) and (d) phospho–c-Jun N-terminal kinase (pJNK). The tissue was sourced from patients diagnosed with either IgA nephropathy (IgAN), membranous nephropathy (MN), minimal change disease (MCD), or focal segmental glomerular sclerosis (FSGS). Subpanels (b) and (e) show all patients tested and their replicates with the error bars showing the SEM for signal within each individual patient. Subpanels (c) and (f) show the means for each patient by disease type; here the error bars show the SEM within each disease type. There is a possible role for a circulating factor in MCD and FSGS. Although there was evident renal damage in IgAN and MN, this damage is probably not caused by the activity of the postulated circulating factor. (c) pVASP was significantly higher in patients with MCD and FSGS relative to patients with non–cystic fibrosis (CF) (patients with IgAN and MN) (P = 0.001 Bonferroni’s multiple comparison test). (f) There was also significantly more glomerular pJNK in FSGS glomeruli relative to non-CF (P = 0.021 Bonferroni’s multiple comparison test). (b,e) Patient identifier shown in an orange box has FSGS with a known genetic cause. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. DAB, 3,3′-diaminobenzidine. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 4. Developmental activation of podocyte protease-activated receptor 1 (PAR-1) in vivo is highly detrimental.

(a) The transgenic form of PAR-1 is tagged with the synthetic peptide FLAG. This allows for the distinction between endogenous PAR-1 and the transgenic PAR-1. Kidneys from 6-day-old mice were harvested and flash frozen. The sections were stained using immunofluorescence. The blue channel shows 4′,6-diamidino-2-phenylindole, the red channel shows nephrin, and the green channel shows FLAG. There is no expression of FLAG in the control mice. The FLAG is clearly visible in the mutant mice. This shows that the transgene is being expressed. (b) These animals are overtly proteinuric at 32 days (P = 0.0005, 1-way unpaired t test). (c,d) They also have significantly higher levels of creatinine (1-tailed Mann-Whitney test, P = 0.0003) and urea in their blood (1-tailed Mann-Whitney test, P = 0.013). (e) The NPHS2 Cre PAR-1^Active+/− mice die around 40 days of age (controls n = 19, median survival 195.4 days, SD 49.46 days; mutants n = 15, median survival 40.8 days, SD 2.23 days; log-rank Mantel-Cox test, P ≤ 0.0001). (f) Tissues from animals culled at 8, 21, 32, and 40 days were periodic acid–Schiff (PAS) and trichrome stained. Bar = 25 μm. In the mutant sections, the PAS staining clearly shows accumulation of extracellular matrix and developing fibrosis over the time course shown. Trichrome staining shows evidence of sclerosis. Representative images shown from electron microscopy studies indicate that the NPHS2 Cre PAR-1^Active+/− mouse is born with a normal phenotype. The ultrastructure of the filtration barrier is well maintained with clear podocyte foot processes. However, this ultrastructure is completely ablated by 26 days of age in the NPHS2 Cre PAR-1^Active+/− mice. (g) With podocyte foot process effacement (*) and thickening of the glomerular filtration barrier (#). Bar = 500 nm. A pathologist (MB) scored 48 images of murine glomeruli, 24 from each of Cre PAR-1^Active+/− mice and control mice. (h) The genotype of the mouse, be it control NPHS2 Cre PAR-1^Active+/−, had a significant impact on the level of glomerulosclerosis (2-way analysis of variance, P ≤ 0.00001). (i) Collating the glomerulosclerosis scores also indicated that there was significantly more total sclerosis in the NPHS2 Cre PAR-1^Active+/− mice compared with wild type (Fisher’s exact test, P ≤ 0.00001). ACR, albumin to creatinine ratio. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 5. Activation of podocyte protease-activated receptor 1 (PAR-1) in maturity is also highly detrimental.

The inducible mice, after induction with doxycycline (dox) treatment, develop a similar phenotype to the developmental model. (a) Albeit over a longer time course, the PAS staining indicates clear fibrosis by 12 weeks, with the trichrome staining showing sclerosis by the same time point. The PAR-1–active inducible animals demonstrate a significant increase in their albumin to creatinine ratio after treatment with doxycycline (2 mg/l 5% sucrose) for 3 weeks (^$Kruskall-Wallis test, P = 0.0160). They are also significantly more proteinuric than their littermate controls who have been subjected to the same doxycycline treatment (**Mann-Whitney U test, P = 0.0043). The same significant interactions are seen at 10 weeks after the commencement of treatment; the PAR-1–active inducible animals maintain a significant increase in proteinuria compared with 0 weeks (^$Kruskall-Wallis test, P = 0.0160), and they also remain significantly more proteinuric than their littermate controls (*Mann-Whitney U test, P = 0.0476). (b) The Pod rtTA Tet O Cre PAR-1 mice have been bred on a single SV129 background. A pathologist (MB) scored 6 mice, 3 control pod rtTA Tet O Cre -ve PAR-1^Active+/− and 3 pod rtTA Tet O Cre -ve PAR-1^Active+/− mice. Both groups received doxycycline treatment. (c) There was significantly more segmental sclerosis in the inducible mice than in the controls (2-way analysis of variance, P = 0.0297). (d) This significance held true when the data for the mice were pooled (Fisher’s exact test P = 0.0154). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. PAS, periodic acid–Schiff; TRI, trichrome. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 6. Protease-activated receptor 1 (PAR-1) signaling response in mouse model biopsies.

Kidneys from NPHS2 Cre PAR-1^Active+/− mice and control mice were harvested from 8-day-old mice. These sections were stained for (a) Phospho–vasodilator-stimulated phosphoprotein (pVASP). There is significantly more pVASP in the glomeruli of Cre PAR-1^Active+/− mice relative to control mice (1-tailed unpaired t test, P ≤ 0.001). (b) Phospho–c-Jun N-terminal kinase (pJNK) was also measured and showed a significant increase in the glomeruli of Cre PAR-1^Active+/− mice (1-tailed unpaired t test, P ≤ 0.0291). (c) pPaxillin staining was also performed. Six glomeruli from 6 mice in each group were analyzed; the distribution of the 3,3′-diaminobenzidine (DAB) measurements for pPaxillin is shown in the middle panel. There is significantly more pPaxillin in the glomeruli of Cre PAR-1^Active+/− mice relative to control mice (1-tailed unpaired t test, P ≤ 0.0011). This is an in vivo validation of the in vitro data presented in Figure 1. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. The Pod Cre PAR-1 mice are on a single SV129 background. IHC, immunohistochemistry. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 7. Detrimental effects of protease-activated receptor 1 (PAR-1) activation are mediated through TRPC6.

Calcium influx was measured as a surrogate for transient receptor potential cation channel subfamily c member 6 (TRPC6) activity. R_norm shows the level of calcium influx after 15 μM flufenamic acid treatment. (a) Podocytes treated with relapse plasma saw a significant increase in calcium influx compared with podocytes treated with remission plasma. This suggests the presence of a factor in the relapse plasma that can potentiate the activity of the TRPC6 calcium channel. (b–d) Treatment of wild-type (WT) mouse podocytes with PAR-1 agonist replicated the responses seen in human WT podocytes (Bonferroni’s multiple comparison test). TRPC6 knockout (KO) podocytes were treated with PAR-1 agonist for the indicated time points. These podocytes did not show the signature response of (b) phospho–c-Jun N-terminal kinase (pJNK), (c) phospho–vasodilator-stimulated phosphoprotein (pVASP), and (d) pPaxillin. (e) TRPC6 was knocked out of the Pod Cre PAR-1–active mice (see the Methodssection for model generation); these mice were significantly less proteinuric at 42 days (1-way Mann-Whitney test, P = 0.0402). (f) Pod Cre PAR-1^Active+/− TRPC6 KO mice lived significantly longer (TRPC6 WT n = 13, median survival 33 days vs. 7.2 days; TRPC6^−/− n = 11, median survival 61 days vs. 14.7 days Mantel-Cox test, P ≤ 0.0001). The Pod Cre PAR-1 mice were crossed with whole-body TRPC6 KO mice on a C57/Bl6. A pathologist (MB) scored 6 Pod Cre PAR-1^Active+/− mice on the mixed SV129/C57Bl/6 background versus their Pod Cre PAR-1^Active+/− TRPC6 KO counterparts and found no significant difference in glomerulosclerosis. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. These data can be seen by (g) individual mouse or (h) collated.

Figure 8. Transient receptor potential cation channel subfamily c member 6 (TRPC6) knockout (KO) significantly reduces glomerular phospho–vasodilator-stimulated phosphoprotein (pVASP) and protects filtration barrier ultrastructure.

Glomerular signaling in the Pod Cre protease-activated receptor 1 (PAR-1)^Active+/− TRPC6 wild-type (WT) and Pod Cre PAR-1^Active+/− TRPC6 KO was interrogated using immunohistochemistry (IHC) targeted against (a) phospho–vasodilator-stimulated phosphoprotein (pVASP), (b) phospho–c-Jun N-terminal kinase (pJNK), and (c) pPaxillin. Only pVASP demonstrated a significant difference between the 2 models. Glomerular pVASP was significantly reduced in the TRPC6 KO variant compared with the TRPC6 WT (P = 0.0093, 1-tailed unpaired t test). (d) Representative images shown from electron microscopy (EM) studies that revealed that the structure of the filtration barrier was deranged by 26 days in the Pod Cre PAR-1^Active+/− TRPC6 WT mice. Additional representative images can be seen in Supplementary Figure S4. (e) TRPC6 KO in the Pod Cre PAR-1^Active+/− mice leads to retention of the neat, highly regulated structure of the glomerular filtration barrier. Indeed, analysis of the EM demonstrated a significant decrease in foot process width between TRPC WT and TRPC6 KO variants of the Pod Cre PAR-1^Active+/− model (1-tailed Mann-Whitney test, P = 0.0500). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001. DAB, 3,3′-diaminobenzidine. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.