Protease‐activated receptor‐1 contributes to renal injury and interstitial fibrosis during chronic obstructive nephropathy

Abstract End‐stage renal disease, the final stage of all chronic kidney disorders, is associated with renal fibrosis and inevitably leads to renal failure and death. Transition of tubular epithelial cells (TECs) into mesenchymal fibroblasts constitutes a proposed mechanism underlying the progression of renal fibrosis and here we assessed whether protease‐activated receptor (PAR)‐1, which recently emerged as an inducer of epithelial‐to‐mesenchymal transition (EMT), aggravates renal fibrosis. We show that PAR‐1 activation on TECs reduces the expression of epithelial markers and simultaneously induces mesenchymal marker expression reminiscent of EMT. We next show that kidney damage was reduced in PAR‐1‐deficient mice during unilateral ureter obstruction (UUO) and that PAR‐1‐deficient mice develop a diminished fibrotic response. Importantly, however, we did hardly observe any signs of mesenchymal transition in both wild‐type and PAR‐1‐deficient mice suggesting that diminished fibrosis in PAR‐1‐deficient mice is not due to reduced EMT. Instead, the accumulation of macrophages and fibroblasts was significantly reduced in PAR‐1‐deficient animals which were accompanied by diminished production of MCP‐1 and TGF‐β. Overall, we thus show that PAR‐1 drives EMT of TECs in vitro and aggravates UUO‐induced renal fibrosis although this is likely due to PAR‐1‐dependent pro‐fibrotic cytokine production rather than EMT.

pathogenesis of renal fibrosis. After injury, TECs undergo EMT in order to avoid imminent cell death and to aid tissue repair. [2][3][4][5] Dysregulated repair due to persistent injury, however, leads to a switch from a regenerative process into a detrimental fibrotic response. 6 Although the concept of TECs undergoing mesenchymal transition upon kidney damage was raised more than a decade ago, the molecular mechanisms that control EMT of TECs remain largely unidentified.
Interestingly, the family of protease-activated receptors (PARs) recently emerged as key players in EMT. 7 PARs are G-protein coupled receptors that are activated by coagulation proteases thereby enabling these proteases to influence a range of pathophysiologic processes. 8 As opposed to classical G-protein coupled receptors, PAR activation requires proteolytic cleavage rather than ligand binding. Indeed, after proteolytic removal of the N-terminal extracellular region, a novel tethered ligand that interacts with the body of the receptor is unmasked.
Subsequent transmembrane signalling leads, amongst others, to EMT of alveolar epithelial cells and retinal pigment epithelial cells. [9][10][11] Moreover, PAR-1-dependent signalling drives fibroblast proliferation and extracellular matrix production in vitro, whereas PAR-1 deficiency limits liver, lung and skin fibrosis in experimental animal models. [12][13][14][15] In the kidney, PAR-1 is expressed by endothelial cells, podocytes, mesangial cells, and tubular epithelial cells 16 and we recently showed that PAR-1 potentiates diabetic nephropathy by inducing mesangial cell proliferation and extracellular matrix production. 17 Based on the key role of PAR-1 in fibroproliferative disease, it is thus tempting to speculate that PAR-1 may be a key factor driving the pathogenesis of renal fibrosis. We challenged this hypothesis by evaluating renal fibrosis in wild-type and PAR-1-deficient mice subjected to the well-established murine unilateral ureter obstruction (UUO) model.

| Mice
PAR-1-deficient mice, generated on a C57Bl/6 background, were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) whereas wild-type C57BL/6 mice were purchased from Charles River
Six wild-type and six PAR-1-deficient mice received a sham operation, in which all procedures were followed apart from ligation of the ureter. Eight mice of each genotype were killed either 1, 3, 7, or 10 days after surgery. Blood and kidneys were harvested and prepared for further analysis. Each kidney was divided in halves; one half was fixed in 4% formalin and embedded in paraffin and the other half was homogenized for protein and RNA analysis. Contralateral non-obstructed kidneys served as control.

| RNA isolation and RT qPCR
For gene expression analysis, mRNA was isolated from kidney homogenates or cultured cells using TriReagent (#11667165001; Roche Diagnostics) according to the manufacturer's recommendations. All Expression levels were normalized using the average expression levels of β-actin, GAPDH and TBP. Primer sequences are listed in Table 1.

| Western blot
Cells were seeded at a density of 20 000 cells/well in 24-well plates.
After stimulation for 24 hours, cells were washed in ice-cold PBS and lysed in Laemmli buffer. Kidney homogenates were lysed in Greenberger Lysis buffer. Lysates were next separated on 8%-12% SDS-PAGE gels and transferred onto Immobulin-P membranes (Millipore) as described before. 21 Membranes were blocked for 1 hour at room temperature in 5% bovine serum albumin (BSA) in TBS+0.

| Immunocytochemistry
Cells were seeded at a density of 20 000 cells/well on coverslips in 24-well plates. After the indicated stimulation for 24-72 hours, cells were washed in ice-cold PBS, fixed in 4% paraformaldehyde in PBS and stained as described before. 22 In short, cells were washed with PBS and incubated for 30 minutes in 0.1% Triton X-100% and 0.5% BSA in PBS for blocking and permeabilization. Subsequently, cells were incubated with the following primary antibodies: rabbit-anti-ZO-1 (1:200; Thermo Fisher; 617300) or mouse-anti-α-SMA (1:1000; Santa Cruz; sc-32251). After overnight incubation, cells were washed with PBS and incubated with Alexa488-linked secondary anti-mouse or anti-rabbit antibodies for 1 hour. The cytoskeleton was stained using phalloidin (100 nM in PBS, 30 minutes) and nuclei were stained with DAPI (10 minutes). Coverslips were mounted onto object glass slides using prolong gold antifade reagent (Thermo Fisher). Micrographs were made using a Leica DM5000B fluorescent microscope with LAS X software (Leica).

| (Immuno)histochemistry
Formalin-fixed, paraffin-embedded, 4-μm-thick kidney slides were stained with periodic acid-Schiff-diastase (PAS-D) and picrosirius red following routine procedures. The PAS-D slides were subsequently scored by a pathologist in a blinded fashion as previously described. 18 Specific immunostainings were performed as described before 17

| Statistics
All values are expressed as mean ± SEM. All groups were tested for normality using the D'Agostino-Pearson omnibus normality test.
Detected outliers were excluded from analysis. Differences between two groups were analysed using a t test if data were normally distributed, or a Mann-Whitney U test for non-parametric data. Multiple comparisons were analysed using one-way-ANOVA analysis or Kruskal-Wallis test (for nonparametric values), followed by Bonferroni's or Dunns multiple comparison tests, respectively. All analyses were performed using GraphPad Prism version 5.01.

| PAR-1 activation induces EMT in tubular epithelial cells in vitro
To test the hypothesis that PAR-1 signalling induces EMT of TECs,

| PAR-1 deficiency reduces UUO-induced renal fibrosis in mice
Based on our in vitro findings showing that PAR-1 activation triggers EMT of TECs, we hypothesized that PAR-1 would contribute to renal interstitial fibrosis. To prove or refute this hypothesis, PAR-1deficient mice were subjected to UUO and fibrotic responses were compared to wild-type controls. We first assessed PAR-1 expression levels in control (contralateral) and obstructed kidney sections of wild-type mice. As shown in Figure 2A, showing PAR-1 induction after UUO using immunohistochemistry 23 ).
As expected, no PAR-1 mRNA was measured in the PAR-1-deficient mice during the experiment ( Figure 2A). As shown in Figure Figure 3F, confirms that collagen I was significantly reduced in PAR-1-deficient mice. Finally, we show that diminished collagen deposition was not accompanied by reduced fibronectin production.
Although fibronectin mRNA levels were reduced in PAR-1-deficient mice compared to wild-type mice at t = 10, western blot analysis showed that protein levels did not differ between wild type and PAR-1-deficient mice ( Figure 3G). Overall, these data show that PAR-1 deficiency limits UUO-induced fibroblast accumulation with subsequent collagen deposition. WAASDORP ET AL. | 1271

| No evidence that PAR-1 deficiency preserves the epithelial phenotype of tubular epithelial cells after UUO
It is tempting to speculate that the observed differences in vimentin and α-SMA levels between wild-type and PAR-1-deficient mice (Figure 3A, B) suggest that PAR-1 also drives EMT after the induction of UUO. Indeed, vimentin and α-SMA are both well-known markers of EMT, but vimentin and α-SMA expression may also originate from infiltrating/proliferating fibroblasts rather than from transitioned TECs. 24,25 To discriminate between these processes, immunohistochemical analysis of α-SMA and vimentin was performed to localize the α-SMA and vimentin expressing cells. Interestingly, both these markers were mainly expressed in the renal interstitium and hardly any positive TEC was found ( Figure 4A, B) suggesting that the observed reduction of vimentin and α-SMA more likely reflects diminished interstitial fibroblast accumulation, rather than reduced EMT of TECs. Moreover, no difference in protein expression levels of the epithelial markers e-cadherin, SGLT2, and AQP-1 was observed between wild-type and PAR-1-deficient mice ( Figure 4C, D). Finally, we observed that SNAI1 expression (a transcription factor specific for EMT) was significantly induced at day 10 after the induction of UUO but no difference between wild-type and PAR-1-deficient mice was found ( Figure 4E). Overall, we only observe minimal signs of EMT in both wild-type and PAR-1-deficient mice suggesting PAR-1-dependent EMT is of minor importance in UUO-induced pathology.

| PAR-1 stimulation leads to MCP-1 and TGF-β secretion
Next to its role in EMT, PAR-1 has also been implicated in the production of pro-fibrotic and/or pro-inflammatory mediators, like MCP-1 and TGF-β, during pulmonary fibrosis. 13,26 To determine whether PAR-1 would play a role in the production of these pro-fibrotic and/ or pro-inflammatory mediators in the setting of renal fibrosis as well, we assessed PAR-1-induced cytokine production by TECs in vitro. As shown in Figure 5A, thrombin-dependent PAR-1 activation induced expression levels of MCP-1, TGF-β1, and KC, but not of TNFα, IL6, or TGF-β3 in TECs. In line with these in vitro data, MCP-1 and TGF-β levels were also reduced in obstructed kidneys of PAR-1-deficient mice compared to obstructed kidneys of wild-type mice (Figure 5B, C). Importantly, reduced MCP-1 levels in PAR-1 deficient obstructed kidneys were accompanied by a reduced influx of macrophages (shown as representative pictures in Figure 5E, and quantified by F4/80 expression as shown in Figure 5D). Together, these data show that PAR-1 induces the expression of pro-fibrotic mediators with subsequent macrophage influx thereby providing an alternative explanation for the observed reduced renal fibrosis in PAR-1-deficient mice. Finally, expression levels of the key EMT transcription factor SNAI1

| DISCUSSION
were also similar between wild-type and PAR-1-deficient mice.
Overall, we thus did not obtain any evidence that PAR-1 deficiency preserves the epithelial phenotype of tubular epithelial cells in vivo.
It is important to stress, however, that EMT is difficult to assess in vivo using (epithelial and/or mesenchymal) marker expression and Interestingly, PAR-1 deficiency led to a marked decrease of interstitial macrophage accumulation upon UUO providing an alternative explanation on how PAR-1 potentiates renal fibrosis. As we show that TECs secrete MCP-1 after PAR-1 stimulation, we postulate that PAR-1-dependent macrophage recruitment, which has been observed before in pulmonary fibrosis, 26,29 is likely due to TEC-dependent MCP-1 expression. In addition, PAR-1 may also directly potentiate the intrinsic migratory activity of macrophages. 26,29 Once recruited, macrophages secrete large amounts of pro-fibrotic cytokines like TGF-β, which in turn induce fibroblast proliferation. 7,30 It is thus tempting to speculate that reduced macrophage recruitment in PAR-1-deficient mice results in the observed reduction in fibroblast accumulation with subsequent reduced extracellular matrix production and renal fibrosis. Additionally, PAR-1-dependent TGF-β production by TECs, as observed in our in vitro experiments and as described before for HK2 cells 31 , may further induce fibroblast proliferation and activation thereby further enhancing renal fibrosis. This latter notion is supported by recent findings that PAR-1 stimulation of cardiac fibroblasts also leads to TGF-β production with subsequent myofibroblast accumulation. 32 Although PAR-1 has originally been identified as blood coagulation factor receptor, at least 12 different proteases have already been described to activate PAR-1 in different pathological settings (REF). The identification of the endogenous PAR-1 agonist in the setting of UUO-induced renal fibrosis is therefore a major challenge.
Interestingly, however, during preparation of our manuscript, it was shown that UUO-induced renal damage and tubulointerstitial fibrosis F I G U R E 4 Interstitial expression of mesenchymal markers, in wild-type and PAR-1-deficient mice. Representative images of α-SMA (A)-and vimentin (B)-stained kidney slides of wild-type and PAR-1-deficient mice 7 and 10 d after UUO and in unobstructed (Sham) control kidneys; scale bars represent 50 μm. C, Western blot analysis of E-cadherin, SGLT2 and AQP1 in whole kidney lysates of unobstructed (sham) and obstructed (UUO) kidneys of wild-type (WT) and PAR-1-deficient (PAR-1 −/− ) mice, 7 and 10 d after UUO. GAPDH expression served as loading control. D, Quantification of Western blots depicted in panel C. E, SNAI1 mRNA expression in whole kidney lysates of unobstructed (sham) and obstructed (UUO) kidneys of wild-type (WT) and PAR-1-deficient (PAR-1 −/− ) mice, 7 and 10 d after UUO. *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001 (one-way ANOVA followed by Bonferroni multiple comparisons test) was suppressed in UUO mice treated with the specific FXa inhibitor edoxaban. 23 Although this may pinpoint FXa as the endogenous PAR-1 agonist driving PAR-1-dependent renal injury during UUO, it was not shown that the effect of FX inhibition was PAR-1 dependent.
Overall, we here show that PAR-1 contributes to renal fibrosis, as evident from increased fibroblast accumulation and collagen deposition in the interstitial areas of wild-type kidneys compared to PAR-1-deficient kidneys, and could therefore be a potential target to pursue in the setting of renal fibrosis. Based on both in vivo and in vitro results we propose that PAR-1 potentiates renal fibrosis by regulating the expression of pro-fibrotic mediators MCP-1 and TGF-β subsequently leading to MCP-1-induced macrophage influx and TGF-β-induced extracellular matrix production. Subsequent experiments addressing pharmacological inhibition of PAR-1 should elucidate whether PAR-1 inhibition indeed has clinical potential for renal fibrosis.