Thrombin stimulates proinflammatory and proliferative responses in primary cultures of human proximal tubule cells

Kidney International, Vol. 67 (2005), pp. 1315–1329

Thrombin stimulates proinflammatory and proliferative responses in primary cultures of human proximal tubule cells DAVID A. VESEY, CATHERINE W. CHEUNG, WADE A. KRUGER, PHILIP PORONNIK, GLENDA GOBE, and DAVID W. JOHNSON Department of Renal Medicine, University of Queensland, Princess Alexandra Hospital, Brisbane, Queenlsand, Australia; School of Biomedical Sciences, University of Queensland, St. Lucia, Brisbane, Queenlsand, Australia; and Department of Molecular and Cellular Pathology, University of Queensland, Royal Brisbane Hospital, Brisbane, Queensland, Australia

Thrombin stimulates proinflammatory and proliferative responses in primary cultures of human proximal tubule cells. Background. Fibrin deposition is frequently observed within the tubulointerstitium in various forms of chronic renal disease. This suggests the presence of active components of the coagulation pathway, which may contribute to the progressive deterioration in renal function. The aim of this study was to investigate the proinflammatory and fibroproliferative effects of the coagulation protease thrombin on human proximal tubular cells (PTC) in culture. Methods. Primary cultures of PTC were established from normal kidney tissue and grown under serum-free conditions with or without thrombin or the protease-activated receptor (PAR) activating peptides TFLLRN-NH 2 , SLIGKV-NH 2 , and SFLLRN-NH 2 (100 to 400 lmol/L). DNA synthesis (thymidine incorporation), intracellular Ca2+ mobilization (fura-2 fluorimetry), fibronectin secretion [enzyme-linked immunosorbent assay (ELISA), immunoblotting], monocyte chemoattractant protein-1 (MCP-1) secretion (ELISA), and transforming growth factor-b1 (TGF-b1) secretion (ELISA) were measured. Reverse transcription-polymerase chain reaction (RT-PCR) was used to assess PAR mRNA expression in these cells. Results. Thrombin enhanced DNA synthesis, fibronectin secretion, MCP-1 secretion, and TGF-b1 secretion in a concentration-dependent manner. Cell injury [lactate dehydrogenase (LDH) release] and cellular protein levels were unaffected. RT-PCR showed that cultures of PTC expressed mRNA transcripts for the thrombin receptors PAR-1 and PAR-3, but not PAR-4. Thrombin and each of the PAR activating peptides enhanced intracellular calcium mobilization. However, the other effects of thrombin were only fully reproduced by the PAR-2–specific peptide, SLIGKV-NH 2 , only partially by SFLLRN-NH 2 , (a PAR-1 peptide that can activate PAR-2), and not at all by the PAR-1–specific peptide, TFLLRN-NH 2 .

Key words: thrombin, proximal tubule cell, transforming growth factor beta, fibronectin, monocyte chemotactic protein-1, protease-activated receptor. Received for publication December 9, 2003 and in revised form July 19, 2004, and October 20, 2004 Accepted for publication November 11, 2004
C

2005 by the International Society of Nephrology

Thrombin-induced DNA synthesis, fibronectin, and MCP-1 secretion were unaffected by a TGF-b neutralizing antibody, the matrix metalloproteinase (MMP) inhibitor, GM6001 and the epidermal growth factor (EGF) receptor kinase inhibitor AG1478. Conclusion. Thrombin initiates both proinflammatory and fibroproliferative responses in human PTC. These responses which are dependent on its protease activity appear not to be mediated by PAR-1 activation, the autocrine action of thrombin-induced TGF-b1 secretion, MMP activation, or EGF receptor transactivation. The proinflammatory and fibroproliferative actions of thrombin on human PTC may help explain the extent of tubulointerstitial fibrosis observed in kidney diseases where fibrin deposition is evident.

Fibrotic disorders of the kidney and other organs (such as the liver and lung) are frequently associated with activation of the coagulation cascade, culminating in the local generation of activated clotting factors, such as thrombin, factor Xa, and fibrin [1–6]. Persistent fibrin deposition has generally been observed within the interstitial space and peritubular capillaries in many forms of renal disease, including ischemic nephropathy, obstructive nephropathy, chronic renal allograft nephropathy, and primary and secondary forms of glomerulonephritis [1–7]. However, the mechanisms responsible for this event and its pathogenetic relevance to tubulointerstitial scarring have not been formally studied. Attempts to prevent renal disease progression with nonspecific anticoagulation have largely been abandoned because of serious side effects, primarily bleeding [2, 8]. A potential link between coagulation, inflammation, and fibrosis has recently been suggested to occur via a novel family of cell surface, G protein–coupled receptors, known as protease-activated receptors (PARs) [9]. Cleavage of the N-terminal exodomain of the receptor reveals a new N-terminus, which acts as a tethered ligand. Four distinct PARs have been described [10, 11]. PAR-1, -3 and -4 are activated by thrombin, while PAR-2

1315

1316

Vesey et al: Thrombin and proximal tubule cells

acts as a receptor for serine proteases other than thrombin (principally mast cell tryptase, trypsin, and clotting factors Xa and VIIa). Activation of PARs by activated clotting factors or other serine proteases has increasingly been recognized in nonrenal studies to result in activation of inflammatory cells such as monocytes, T lymphocytes and mast cells, and proliferation of other cell types, including endothelial cells, fibroblasts, and smooth muscle cells [12, 13]. Thrombin receptors are widely expressed within the kidney. PAR-1 is expressed by glomerular endothelial, mesangial, and tubule cells [7, 14, 15]. The expression of PAR-3 and -4 have been demonstrated in the kidney but they have not been studied in detail [16, 17]. In patients with chronic allograft nephropathy, tubular expression of PAR-1 is markedly up-regulated and correlates with interstitial fibrin deposition, the degree of tubulointerstitial fibrosis, and urinary excretion of transforming growth factor-b (TGF-b) [18]. Moreover, in a murine model of crescentic glomerulonephritis, the use of hirudin (a selective antagonist of thrombin-induced proteolytic activation of PARs), or PAR-1 knockout mice equally effectively reduced crescent formation, glomerular inflammatory cell infiltration, and serum creatinine concentrations. Conversely, administration of a selective PAR-1–activating peptide augments histologic and functional injury [19]. The known cellular actions of PAR activation appear to be mediated either directly via intracellular signaling pathways [including mitogen-activated protein (MAP) kinases, phospholipase C, protein kinase C, cyclooxygenase-2 (COX-2), and peroxisome proliferator activated receptor-gamma (PPARc)] [10, 11] or indirectly via the induction of fibrogenic cytokines, such as TGF-b, platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and prostaglandin E 2 [20–23]. There is also accumulating in vivo evidence that acute or persistent PAR activation contributes to aberrant tissue repair and fibrosis in a number of pathologic conditions, including restenosis and neointima formation following vascular injury [24, 25], atherosclerosis [26], and pulmonary fibrosis [27]. The present study examined the action of the serine protease thrombin on human PTC in culture. We were particularly interested in how thrombin might influence cellular DNA synthesis, the production of extracellular matrix (ECM) protein fibronectin, the chemotactic and proinflammatory molecule monocyte chemoattractant protein-1 (MCP-1) and the potentially fibrogenic growth factor TGF-b1. Using PAR-specific peptides we attempted to link the responses of these cells to thrombin, to PAR-1 activation. In addition, the possibilities that the actions of thrombin are mediated indirectly through the autocrine action of TGF-b, the activation of matrix metalloproteinases (MMP) and/or epidermal growth factor (EGF) receptor transactivation were investigated.

METHODS Chemicals Cell culture tested thrombin purified from human plasma (1000 N IU/mg protein, catalog number T4393) and hirudin (11,500 U/mg, catalog number H0393) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). The PAR activating peptides SFLLRN-NH 2 , TFLLRN-NH 2 , and SLIGKV-NH 2 were synthesized with carboxyl-terminal amidation and purified to >95% via high-performance liquid chromatography (HPLC) by Auspep Ltd. (Melbourne, Australia). The MMP inhibitor, GM6001, and EGF receptor kinase inhibitor, AG1478, were purchased from Merck (San Diego, CA, USA). Cell culture Segments of macroscopically and histologically normal renal cortex (5 to 10 g) were obtained aseptically from adult human kidneys removed surgically for either small (<6 cm) renal adenocarcinomas or benign renal cysts. Patients were otherwise healthy and were on no medications. Informed consent was obtained prior to each operative procedure and the use of human renal tissue for primary culture was reviewed and approved by the Princess Alexandra Hospital Research Ethics Committee. The method for primary culture of human PTC is described in detail elsewhere [28, 29]. Briefly, renal cortical tissue was dissected from the medulla, minced, digested with collagenase (class 2, 200 to 383 U/mg) (Worthington, Freehold, NJ, USA) and passed through a 100 lm mesh. Filtered cells were resuspended in 45% Percoll (Amersham-Pharmacia-Biotech, Uppsala, Sweden) and separated into four distinct bands by isopycnic centrifugation. The lowermost band was removed for PTC culture. PTC were resuspended in serum-free, hormonally defined Dulbecco’s modified Eagle’s medium (DMEM)/F-12 media (Thermo Electron Corporation, Melbourne, Australia), supplemented with penicillin (50 U/mL), streptomycin (50 lg/mL), human transferrin (5 lg/mL), bovine insulin (0.87 lmol/L), hydrocortisone (0.05 lmol/L), EGF (10 ng/mL) (Collaborative Research Inc., Bedford, MA, USA), prostaglandin E 1 (50 nmol/L) (Sigma Chemical Co.), selenium (50 nmol/L) and tri-iodothyronine (5 pmol/L) (Sigma Chemical Co.). The tubular fragments were plated at a density of 1.5 mg pellet/cm2 (approximately 5000 fragments/cm2 ) in 75 cm2 flasks. Cells were incubated in humidified atmosphere of 95% air/5% CO 2 at 37◦ C and media changed every 48 hours. Confluence was reached at approximately 10 days after plating. Cells were passaged using Dispase II (Roche Diagnostics, Indianapolis, IN, USA) and frozen cultures maintained in liquid nitrogen. Cytologic examination of PTC preparations from all donors failed to reveal any evidence of cellular atypia. The morphologic, biochemical, and functional

Vesey et al: Thrombin and proximal tubule cells

1317

characteristics of these cells have been previously studied in this laboratory and found to reproducibly exhibit the features of PTC in vivo [28]. Specifically, these cells exhibit typical cobblestone morphology along with intact tight junctions, parathyroid hormone (PTH) (but not vasopressin), response, cyclic adenosine monophosphate (cAMP) production and polarized transport systems, including pharmacologically distinct apical and basolateral sodium-hydrogen exchange, phlorizin-inhibitible apical Na+ -glucose transport, PTH-inhibitable apical Na+ -phosphate transport, probenecid-inhibitable vectorial organic anion transport, and quinine-inhibitable vectorial organic cation transport. Fibroblast, mesangial, and endothelial cell contamination was negligible, as evidenced by the uniform negative staining of cultures for vimentin, desmin, and factor VIII, respectively [28].

Tween-20 were then added to each well of the plate. After a further 1-hour incubation at room temperature, plates were washed as before and then incubated with 1,2-phenylene diamine dihydrochloride (OPD) substrate (Dako) according to the manufacturer’s instructions. The reaction was stopped with 100 lL of 0.5 mol/L sulphuric acid and the absorbance determined at 492 nm. A standard curve was constructed and unknowns determined from this curve. The intra- and interassay variation at 50 ng/mL was less than 12%.

Experimental protocol

Measurement of TGF-b in conditioned media

All experiments were performed on confluent, passage 2 PTC cultured in 48-, 24-, or 6-well tissue culture plates. Cells were made quiescent by two washes followed by incubation for 24 hours in basic media (DMEM/F12 containing 5lg/mL transferrin). The concentrationdependent effects of thrombin or the PAR-1 peptides, TFLLRN-NH 2 and SFLLRN-NH 2 , or PAR-2 peptide, SLIGKV-NH 2 , on DNA synthesis, fibronectin production, MCP-1 expression and TGF-b1 production were then assessed over a 24-hour period. The influence of the thrombin inhibitor, hirudin, a MMP inhibitor, GM6001, an EGF receptor kinase inhibitor, AG1478, and a TGFb neutralizing antibody on these responses to thrombin was assessed. Media conditioned by PTC were harvested, mixed with a cocktail of protease inhibitors, and stored at −80◦ C until assayed.

TGF-b( levels in PTC-conditioned media were measured using a commercially available ELISA kit (R&D Systems). Total (active plus latent) TGF-b1 was determined following transient acidification of conditioned media. Briefly, each mL of conditioned medium was incubated with 100 lL of 1 mol/L HCl for 15 minutes at room temperature, followed by neutralization with an equivalent volume of 1.2 mol/L NaOH/0.5 mol/L Hepes. The minimum detectable concentration of TGF-b1 by the assay was 15 pmol/L.

Fibronectin determination The fibronectin concentration in the culture supernatant was determined using a sandwich enzyme-linked immunosorbent assay (ELISA) [30, 31]. The wash buffer used was phosphate-buffered saline (PBS) containing 0.05% Tween-20, pH 7.4. The coating buffer was carbonate/bicarbonate (0.05 mol/L carbonate), pH 9.7. Each well of a flat bottomed 96-well microtitre plate (Nalge Nunc International, Naperville, IL, USA) was coated with 100 lL of rabbit antihuman fibronectin (Dako Cytomation, Botany, NSW, Australia) diluted to a concentration of 10 mg/mL in coating buffer for 24 hours at 4◦ C. The plates were then washed four times with wash buffer. Appropriately diluted test samples were then added to the assay plate along with fibronectin standards (1 to 800 ng/mL). The plates were incubated for 2 hours at room temperature and then washed a further four times with wash buffer. One hundred microliters of the secondary peroxidase-conjugated antihuman fibronectin antibody (Dako) diluted 1/4000 in PBS containing 0.05%

MCP-1 determination MCP-1 was measured in the culture medium from PTC using a specific ELISA (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s protocols.

Western blot analysis On some occasions, secreted and cell associated fibronectin were determined by Western blot analysis. Equal amounts of conditioned media, or in the case of cell lysate, cell protein, were diluted in a reducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) sample buffer, heated to 70◦ C for 10 minutes, separated on a 4% to 12% NUPAGE Gel (Invitrogen, Melbourne, Australia) and then electrotransferred to a polyvinlydine difluoride (PVDF) membrane. Membranes were blocked overnight with 5% skim milk powder in PBS containing 0.1% Tween-20. The primary monoclonal fibronectin antibody (Becton & Dickinson, North Ryde, New South Wales, Australia) was added, in blocking buffer for 1 hour. An antimouse peroxidase conjugate (Bio-Rad Laboratories, Regents Park, New South Wales, Australia) was used at recommended dilutions for the secondary antibody. Detection was with enhanced chemiluminescence (ECL)-plus (Amersham-PharmaciaBiotech, Amersham, UK). To control for equal protein loading and transfer the membrane was stripped using Re-Blot Plus Mild (Chemicon, Temecula, CA, USA) according to the supplied instructions and reprobed with a pan-actin primary antibody (NeoMarkers, Fremont, CA, USA) in blocking buffer for 1 hour. The actin bands were visualized as above.

1318

Vesey et al: Thrombin and proximal tubule cells

Cell growth PTC damage was assessed using a cytotoxicity detection kit (Roche, Dee Why, New South Wales, Australia) which measures lactate dehydrogenase (LDH) release into the culture media. Manufacturer’s protocols were followed. Cell counts were performed at the conclusion of selected experiments using a hemocytometer. Tritiated thymidine incorporation into cellular DNA, an index of DNA synthesis, was measured according to a previously described method [30]. Briefly, cells were incubated with control or test media for 24 hours. For the final 4 hours of this culture [methyl-3 H]-thymidine (Amersham-Pharmacia-Biotech, Uppsala, Sweden), 4 lCi (0.15 MBq) per mL of media were added. At the end of the test period cells were washed twice with ice-cold PBS, thrice with ice-cold 10% trichloroacetic acid (TCA) for 20 minutes, and once with methanol. Monolayers were allowed to dry and then dissolved in 300 lL of 0.3 mol/L NaOH containing 1% SDS. Aliquots of cell lysate were taken for liquid scintillation counting in a b counter. Results were corrected for cellular protein content. The protein content of the cell lysate was measured by a commercially available protein assay using bovine serum albumin (BSA) as the standard. Detection of PAR mRNA using reverse transcription-polymerase chain reaction (RT-PCR) Human primary PTCs were grown to confluence as described above. Total RNA was extracted from the cells using the RNeasy Mini Kit (Qiagen, Duncaster, Victoria, Australia) according to the manufacturer’s protocols. During the isolation the RNA was subjected to DNase digestion. Two micrograms of RNA were converted into cDNA using Expand Reverse Transcriptase (Roche) with standard methodologies. PCR was performed in a total reaction volume of 25 lL containing 1 × PCR buffer, 1 lL of cDNA, 12.5 pmol of each primer, 1.5 mmol/L MgCl2, and 0.625 U of Taq DNA polymerase (Red Hot DNA Polymerase) (AB Gene, Epsom, Surrey, UK). The PAR and hypoxanthine phosphoribosyltransferase (HPRT) primer sequences and thermal cycling temperatures have been previously published [30, 32, 33] and are given as follows: PAR-1 primers 1st 5 -TGTGAACTGATCATGGTTATG-3 and 5 -TTCGTAAGATAAGAGAATATGT-3 (708 bp product), 1 × 94◦ C for 5 minutes; 30 × 94◦ C for 1 minute, 55◦ C for 1 minute, and 72◦ C for 1 minute; PAR-1 primers 2nd 5 -CCACAAACGTCCTCCTGATT-3 and 5 TGGGATCGGAACTTTCTTTG-3 (200 bp product), same conditions as for the first set of PAR-1 primers; PAR-2 primers 5 -AGAAGCCTTATTGGTAAGGTT3 and 5 -AACATCATGACAGGTCGTGAT-3 (582 bp product), 1 × 94◦ C for 5 minutes, 35 × 94◦ C for 1 minute, 55◦ C for 1 minute, and 72◦ C for 1 minute; PAR-3 primers 5 -CTGATACCTGCCATCTACCTCC-3 and

5 -AGAAAACTGTTGCCCACACC-3 (382 bp product), 1 × 94◦ C for 5 minutes, 35 × 94◦ C for 1 minute, 65◦ C for 1 minute, and 72◦ C for 1 minute; PAR-4 primers 5 -ATTACTCGGACCCGAGCC-3 and 5 -TGTAAGG CCCACCCTTCTC-3 (392 bp product), 1 × 94◦ C for 5 minutes; 35 × 94◦ C for 1 minute, 65◦ C for 1 minute, and 72◦ C for 1 minute; and HPRT primers 5 -AAT TATGGACAGGACTGAACGTC-3 and 5 -CGTGGG GTCCTTTTCACCAGCAAG-3 (386 bp product), 1 × 94◦ C for 5 minutes, 25 × 94◦ C for 1 minute, 64◦ C for 1 minute, and 72◦ C for 1minute. PCR products were separated on 1.8% agarose gels containing 1 lg of ethidium bromide per mL in 1 × Tris borate-EDTA (TBE) buffer and viewed and photographed under ultraviolet light. A HaeIII digest of φ × 174 was used as a marker. Measurement of cytosolic Ca2+ Intracellular Ca2+ measurements were performed using the fluorescent ratiometic Ca2+ dye Fura-2 (Fura-2 AM) (Molecular Probes, Eugene, OR, USA) in 96-well plates. Cells were grown to confluence in sterile, tissue culture treated black 96-well plates (Perkin-Elmer Life and Analytical Sciences, Victoria, Melbourne, Australia). Before experimentation cells were incubated for 24 hours in basic media, then washed twice in Tyrode buffer (137 mmol/L NaCl, 5 mmol/L KCl, 2 mmol/L CaCl, 1 mmol/L MgCl, 10 mmol/L D-glucose, 10 mmol/L Hepes, and NaOH, pH 7.4), and loaded with 5 lmol/L Fura-2 in the same buffer for 1 hour at 37◦ C with shaking. After loading, cells were washed twice in Tyrode buffer to remove excess dye. All experiments were performed in a BMG Fluostar Optima (BMG Lab Technologies, Offenburg, Germany) using 340 nm, 380 nm excitation, and 520 nm emission filters (BMG Lab Technologies). An automated injection of agonist was made to give overall concentrations of thrombin (5 U/mL), SFLLRN-NH 2 (100 lmol/L), TFLLRN-NH 2 (100 lmol/L), and SLIGKVNH 2 (100 lmol/L) in the corresponding wells. The ratio of the excitation wavelengths at 340/380 nm represents the change in intracellular Ca2+ in response to the agonist. Statistical analysis All studies were performed in triplicate from PTC cultures obtained from at least three separate human donors. Each experiment contained internal controls originating from the same culture preparation. For the purposes of analysis, each experimental result was expressed as a change from the control value, which was regarded as 100%, and analyzed independently. Results are expressed as mean ± SEM. Statistical comparisons between two groups were made using unpaired t tests. Multiple group comparisons were made by analysis of variance (ANOVA). The software package SPSS version 11.5 (SPSS, Inc., Chicago, IL, USA) was used. P values less than 0.05 were considered significant.

1319

Vesey et al: Thrombin and proximal tubule cells

RESULTS Human PTC expressed thrombin receptors 1 and 3 Using conventional RT-PCR, primary cultures of human PTC were found to express mRNA transcripts for thrombin receptors PAR-1 and PAR-3 but not PAR-4 (Fig. 1A). PAR-2 mRNA is also expressed by these cells. As there has been some controversy [7, 14, 34] with regard to PAR-1 expression in PTC these results were confirmed with a second set of PAR-1 specific primers (Fig. 1B). To exclude the possibility that the PCR products represented amplification of contaminating genomic DNA, all isolated RNA was pretreated with DNase according to the manufacturer’s protocol. In addition, control reverse transcriptase reactions were performed in the absence of reverse transcriptase enzyme, in which case no PAR PCR signals were observed (Fig. 1C). Thrombin stimulates mobilization of calcium from internal stores Calcium mobilization assays have been widely used to study PAR activation. To establish that functional thrombin receptors are present on PTCs, cells were cultured with thrombin or PAR peptides and relative changes in intracellular free calcium levels monitored. Treatment of PTC with either thrombin (5 U/mL), TFLLRN-NH 2 (100 lmol/L), SFLLRN-NH 2 (100 lmol/L), or SLIGKVNH 2 (100 lmol/L) caused a rapid rise in free intracellular calcium concentrations as indicated by enhanced Fura-2 AM fluorescence (Fig. 2A). The relative Fura2 fluorescence ratio reached a peak at 17 seconds after treatment. Of the agonists tested, SLIGKV-NH 2 induced the largest increase in calcium mobilization. Thrombin and SFLLRN-NH 2 both produced similar signals at about 52% of that produced by SLIGKV-NH 2 . The calcium response to TFLLRN-NH 2 was only 34% of that produced by SLIGKV-NH 2 (Fig. 2B). Thrombin stimulates DNA synthesis in human PTC As thrombin and PAR-1 peptides have been shown to stimulate proliferation in a range of fibroblastic and epithelial cell types [15, 21, 22, 35], the effect of thrombin/SFLLRN-NH 2 /TFLLRN-NH 2 on DNA synthesis in primary PTC cultures was investigated. This was achieved by culturing confluent, quiescent cells with thrombin (0.01 to 8 U/mL), SFLLRN-NH 2 (100 to 400 lmol/L), or TFLLRN-NH 2 (100 to 400 lmol/L), and measuring 3 [H]-thymidine incorporation into DNA over the final 4 hours of a 24-hour test period. Thrombin stimulated DNA synthesis in primary cultures of PTC (Fig. 3A). Increases in DNA synthesis were first observed at a concentration of 0.5 U/mL and this response increased in a concentration-dependent fashion up to 8 U/mL; significant increases were reached at concentrations of 2 U/mL

A

603 bp 310 bp

B

PAR-1

708 bp

PAR-1 200 bp C

PAR-1 200 bp

Fig. 1. Reverse transcription-polymerase chain reaction (RT-PCR) analysis of protease-activated receptor (PAR) expression by cultured human proximal tubule cells (PTCs). Total RNA was isolated from confluent passage 2, human PTC, reverse transcribed, and amplified with specific primers for PARs and hypoxanthine phosphoribosyltransferase (HPRT). PCR products were separated on 1.8% agarose gels containing ethidium bromide and photographed. (A) Expression of thrombin receptors (PAR-1, PAR-3, and PAR-4), and PAR-2 by PTC. Lane 1, PAR-1 (200 bp); lane 2, PAR-2 (582 bp); lane 3, PAR-3 (382 bp); lane 4, PAR-4, (392 bp); lane 5, HPRT (386 bp); and lane 6,
X174 DNA/Hae III markers. (B) Expression of PAR-1 in three different PTC cultures by two different sets of PAR-1 primers. Lane 1, PTC 1; lane 2, PTC 2; lane 3, PTC 3; lane 4, negative control; and lane 5,
X174 DNA/Hae III markers. (C) RT-PCR analysis of PAR-1 expression in cells from three different PTC culture in which the RT step is carried out with (RT) or without (NRT) the reverse transcriptase enzyme. Lane 1, PTC 1 RT; lane 2, PTC 1 NRT; lane 3, PTC 2 RT; lane 4, PTC 2 NRT; lane 5, PTC 3 RT; lane 6, PTC 3 NRT; lane 7, negative control; and lane 8,
X174 DNA/HaeIII markers.

and above (P < 0.05). Thrombin at 2, 4, and 8 U/mL increased DNA synthesis by 141.2 ± 24%, 187.7 ± 31%, and 237.2 ± 24% of control values, respectively (Fig. 3A) (Table 1). The PAR peptide SFLLRN-NH 2 , in the range 100 to 400 lmol/L, produced significant but lower increases in DNA synthesis; at 100, 200, and 400 lmol/L, increases in DNA synthesis were 119 ± 17% (NS), 128 ± 7% (P < 0.05), and 140 ± 3% (P < 0.05) of control values, respectively. On the other hand, the specific PAR-1

1320

Vesey et al: Thrombin and proximal tubule cells

33% (NS), and 127 ± 40% (NS) of control values, respectively. The average number of cells per well of a 48-well plate under basal conditions was 173,000 ± 34,600.

A 1.8

[Ca+]i, 340/380 nm

1.7 1.6

Thrombin did not cause cellular damage

1.5

To investigate if thrombin caused damage of PTC cultures, LDH release in response to thrombin treatment was measured. Cells treated with thrombin (0.002 to 8 U/mL) failed to show a significant increase in LDH release at any dose tested (Fig. 3B). Levels of LDH released into the media from both treated and untreated cells were below 8% of total cellular LDH.

1.4 1.3 1.2 1.1 0

5

10

15 20 25 Time, seconds

30

35

40

Thrombin SFLLRN TFLLRN SLIGKV VKGILS B % of maximal activation

120 100 80 60 40 20 0 SLIGKV VKGILS TFLLRN SFLLRN Thrombin Agonist Fig. 2. The effects of thrombin and the protease-activated receptor (PAR) peptides on calcium mobilization in human proximal tubular cells (PTCs). Cells were grown to confluence and loaded with Fura2 AM. They were then exposed to thrombin (5 U/mL), TFLLRNNH 2 (100 lmol/L), SFLLRN-NH 2 (100 lmol/L), SLIGKV-NH 2 (100 lmol/L), or VKGILS-NH 2 (100 lmol/L) and intracellular Ca2+ fluorescence measured. (A) The results are expressed as the change from basal level of the fluorescence ratio (340/380 nm). Each trace is an average response from cells in four different wells and is representative of two different experiments. (B) Results are expressed as a percentage of the SLIGKV-NH 2 signal at 24 seconds (17 seconds after agonist addition).

peptide TFLLRN-NH 2 was unable to induce significant increases in DNA synthesis at any concentration tested (Table 1). Cellular protein levels were not affected by thrombin (Fig. 3F), SFLLRN-NH 2 , or TFLLRN-NH 2 treatment (data not shown for PAR peptides). The PAR2 peptide SLIGKV-NH 2 also increased DNA synthesis. At 100, 200, and 400 lmol/L increases in DNA synthesis were 123 ± 15% (P < 0.05), 210.9 ± 19.6% (P < 0.01), and 298.1 ± 27.5% (P < 0.01) of control values, respectively (Table 1). Thrombin failed to significantly increase PTC numbers after 24-hour culture. Thrombin at 1, 2, and 4 U/mL increase PTC numbers by 125 ± 24% (NS), 128 ±

Thrombin stimulated fibronectin secretion by human PTC Fibronectin is one of the first matrix proteins produced in response to cell injury and inflammation. It has been demonstrated to be important in modulating cell migration, chemotaxis, cell anchorage, and collagen deposition. The present investigation analyzed whether thrombin was able to influence fibronectin secretion by primary PTC. Thrombin enhanced fibronectin secretion by PTC in a concentration-dependent fashion (Fig. 3C). Significant increases in fibronectin secretion were observed at 2 U/mL and reached 263 ± 56% of control values at a concentration of 16 U/mL. The absolute fibronectin secretion rate in the control cultures was 0.459 ± 0.41 lg/mg protein/day (N = 7). Immunoblotting demonstrated enhanced production of a single fibronectin band at an apparent molecular weight of 220 kD in response to thrombin treatment. This is shown in Figure 4A for cells established from the renal cortex of four different kidney samples. Fibronectin fragmentation was not observed. To establish whether this response to thrombin was mediated by PAR-1 activation, fibronectin in the conditioned culture media of PTC exposed to either SFLLRN-NH 2 or TFLLRN-NH 2 was measured. While SFLLRN-NH 2 at all concentrations tested significantly increased fibronectin secretion, the more specific PAR-1 peptide TFLLRN-NH 2 was without an effect (Table 1). In addition, the PAR-2 activating peptide SLIGKV-NH 2 at concentrations of 100, 200, and 400 lmol/L significantly enhanced fibronectin secretion by 160 ± 20% (P < 0.05), 252.7 ± 13% (P < 0.01), and 367.7 ± 48% (P < 0.01) of control values, respectively (Table 1). It has been reported that the appearance of increased amounts of fibronectin in cell culture media in response to thrombin treatment is accompanied by decreases in cell associated-fibronectin [36]. Figure 4B demonstrates that this is the case. However, the decrease in cell-associated fibronectin can not account for the much greater increase in fibronectin released in to the media of thrombin treated cells. By comparison, TGF-b1 at 5 ng/mL significantly increased both

1321

Vesey et al: Thrombin and proximal tubule cells

A

D 350

*

250

*

*

MCP-1 secretion, % control

DNA synthesis, % control

300

200 150 100 50 0 0.001 0.01 0.1 1 10 100 Thrombin, U/mL [log scale]

200

TGFβ1 secretion, % control

LDH release, % control

50

0.1 1 10 Thrombin, U/mL [log scale]

100 50 0 0.001 0.01 0.1 1 10 100 Thrombin, U/mL [log scale]

*

600 *

500 400

*

300 200 100 0 0.1

100

1 10 100 Thrombin, U/mL [log scale]

F

C 350

140

* *

120

300

Cell protein/well, % control

Fibronectin secretion, % control

*

150

E

100

0 0.01

*

*

700

150

200

*

250

B 200

250

* *

300

* *

150 100 50 0 0.001 0.01 0.1 1 10 Thrombin, U/mL [log scale]

100 80 60 40 20

100

0 0.001 0.01 0.1 1 10 Thrombin, U/mL [log scale]

100

Fig. 3. The effect of thrombin on human proximal tubule cell (PTCs) DNA synthesis (A), lactate dehydrogenase (LDH) release (B), fibronectin secretion (C), monocyte chemoattractant protein-1 (MCP-1) secretion (D), transforming growth factor-b1 (TGF-b1) secretion (E), and cell protein levels (F). Confluent passage 2 human PTC were treated with or without thrombin for 24 hours and assayed for the above responses as detailed under the Methods section. Data points are the mean ± SEM from five independent experiments, each performed in triplicate. ∗ P < 0.05 vs. control cells.

cell associated (537 ± 46% of control) and secreted fibronectin (235 ± 12.5% of control). MCP-1 secretion was stimulated in PTCs treated with thrombin There is strong evidence that MCP-1 is an important promoter of tubulointerstitial disease. This is not only due to its potent chemotactic actions for monocytes but also through its direct action on tubulointerstitial cells [37–40]. The present investigation examined

whether thrombin could modulate MCP-1 production in PTCs. We found that thrombin enhanced MCP-1 production by human PTC in concentration-dependent manner (Fig. 3D). A significant increase in MCP-1 secretion was seen at 0.1 U/mL of thrombin. At the highest concentration of thrombin tested (8 U/mL) MCP-1 concentrations were 262 ± 25% of control values. Levels of MCP-1 in the media of cells treated with thrombin reached 2 ng/mL or 10.5 ± 4 ng/mg cell protein/day. To examine if the PAR-1 peptides, SFLLRN-NH 2 and TFLLRN-NH 2 , can induce MCP-1 secretion, cells were cultured with these peptides

1322

Vesey et al: Thrombin and proximal tubule cells

Table 1. Comparative responses of human proximal tubule cells (PTCs) to treatment with thrombin and protease-activated receptor (PAR) peptides Response (% of control) Agonist

Concentration lmol/L

DNA synthesis

Fibronectin secretion

MCP-1 secretion

TGF-b1 secretion

2 U/ml 4 U/ml 8 U/ml 100 200 400 100 200 400 100 200 400 200

141 ± 188 ± 237 ± 24a 110 ± 61 112 ± 41 113 ± 16 119 ± 17 128 ± 7a 140 ± 3a 123 ± 15a 211 ± 20a 298 ± 28a 117 ± 9

130 ± 181 ± 249 ± 55a 127 ± 13 92 ± 28 105 ± 12 156 ± 18a 177 ± 20a 189 ± 21a 160 ± 20a 253 ± 13a 368 ± 48a 103 ± 6.7

217 ± 242 ± 262 ± 25a 110 ± 19 95 ± 25 104 ± 28 137 ± 32 144 ± 30 173 ± 34a 208 ± 17a 245 ± 12a 263 ± 25a 125 ± 13

183 ± 32a 276 ± 24a 524 ± 49a 115 ± 30 125 ± 21 105 ± 10 110 ± 10 109 ± 5 134 ± 9a 170 ± 23a 217 ± 18a 301 ± 45a 98 ± 8

TFLLRN-NH 2 SFLLRN-NH 2 SLIGKV-NH 2 VKGILS-NH 2

24a 31a

18a 27a

24a 27a

DNA synthesis, fibronectin secretion, monocyte chemoattractant protein-1 (MCP-1) secretion and transforming growth factor-b1 (TGF-b1) secretion were measured in confluent passage 2 human PTC treated with thrombin or PAR activating peptides. The results were expressed as a% of the control values (mean ± SEM) for at least three separate experiments that were performed in triplicate. a < P 0.05.

A Media fibronectin C T C

T

C

T

C

T

191 kD 1 Media fibronectin 191 kD

2 B C

3

4

T

Cell fibronectin 191 kD Cell actin 39 kD

Fig. 4. Analysis of fibronectin secreted by or associated with human proximal tubular cells (PTCs) treated with thrombin. Confluent, quiescent, passage 2 human PTC cultures were treated with (T) or without (C) thrombin (5 U/mL). After 24 hours, conditioned media and cells lysate were collected and analyzed by electrophoresis and Western blot analysis (A and B) using a specific fibronectin antibody. (A) Conditioned media from four different PTC cultures were analyzed. A single band at 220 kD was observed. No fibronectin degradation products were apparent.

and MCP-1 secretion into the culture media measured. TFLLRN-NH 2 failed to enhance MCP-1 secretion at all concentrations tested whereas SFLLRN-NH 2 significantly increased MCP-1 secretion at 400 lmol/L (173.4 ± 33.5% vs. control) (P < 0.05) (Table 1). The PAR-2 peptide SLIGKV-NH 2 significantly stimulated MCP-1 secretion at all concentrations tested (Table 1). At 100 lmol/L, the MCP-1 concentration in the conditioned media was increased to 208 ± 16.5% of the control value.

Thrombin induces TGF-b1 secretion by PTC Thrombin can stimulate TGF-b1 production by mesangial cells and the immortalized tubule cell line, human kidney-2 (HK-2) [18, 20]. In the present investigation, we found that thrombin treatment induced a significant increase in TGF-b1 production by primary human PTCs. At 1, 2, 4, and 8 U/mL thrombin increased TGF-b1 secretion by 141 ± 26% (NS), 183 ± 32% (P < 0.05), 276 ± 24% (P < 0.02), and 524 ± 49% (P < 0.01) of control values, respectively (Fig. 3E). The concentration of TGF-b1 in the PTC conditioned media reached about 240 pg/mL, a secretion rate of 1.195 ng/mg protein per day, when treated with 4 U/mL thrombin. The effect of PAR peptides SFLLRN-NH 2 , TFLLRN-NH 2 , and SLIGKV-NH 2 on TGF-b1 secretion was also studied. TFLLRN-NH 2 failed to increase TGF-b1 secretion at any concentration tested, SFLLRN-NH 2 only significantly increased its secretion at 400 lmol/L (134 ± 9% of control) (P < 0.05) and the PAR-2 peptide SLIGKV-NH 2 increased TGF-b1 secretion at each concentration tested (Table 1). The proteolytic activity of thrombin is required for its action on PTC The effect of the serine protease inhibitor hirudin on thrombin-induced PTC fibronectin secretion, DNA synthesis, MCP-1 secretion, and TGF-b secretion was examined. Treatment of thrombin with hirudin (5 U/mL) prior to its addition to the PTC abolished or partially abolished thrombin-induced fibronectin secretion (thrombin 272.9 ± 63.5% vs. thrombin + hirudin 154.1 ± 34.3% of control) (P < 0.01), DNA synthesis (thrombin 127 ± 27% vs. thrombin + hirudin 80 ± 12% of control) (P < 0.01), MCP-1 secretion (thrombin 233.4 ± 27% vs. thrombin + hirudin 113.8 ± 31% of control) (P < 0.01), and TGF-b secretion (thrombin 186 ± 18% vs. thrombin + hirudin 128 ± 11% of control) (P < 0.01) (Fig. 5).

Vesey et al: Thrombin and proximal tubule cells

TGF-b1 was not responsible for thrombin-induced fibronectin secretion

A DNA synthesis, % control

300

#

200 * 100 0

Hirudin − Hirudin +

Control

Thrombin

Fibronectin secretion, % control

B #

300 200

* 100 0

Hirudin − Hirudin +

Control

Thrombin

MCP-1 secretion, % control

C 400

#

300 *

200 100 0

Hirudin − Hirudin +

Control

Thrombin

TGFβ, % control

D 250

1323

As TGF-b1 is produced by PTC in response to thrombin treatment and thrombin is a potent inducer of fibronectin secretion by PTC, we tested the possibility that TGF-b is mediating thrombin-induced fibronectin secretion [30]. As shown in Figure 6A, a pan-specific TGFb neutralizing antibody was able to completely blocked TGF-b1–induced fibronectin secretion and but was unable to influence thrombin-induced fibronectin secretion. TGF-b1 is known to be potent inhibitor of PTC DNA synthesis in PTC and thus increased DNA synthesis in response to thrombin would not be mediated by TGF-b1 [29, 30]. However, it is possible that neutralizing TGF-b1 produced in response to thrombin would further increase DNA synthesis. This is not the case. Figure 6B shows that TGF-b neutralization does not alter thrombin-induced DNA synthesis. Thrombin-induced DNA synthesis, fibronectin secretion, and MCP-1 secretion are not mediated by pro-MMP activation or EGF receptor transactivation There have been several reports which show that thrombin can activate pro-MMPs [41, 42]. In turn MMPs have been shown to mediate sustained release and accumulation of EGF receptor ligands, including TGF-a and heparin-binding EGF, and thus cause EGF receptor transactivation in certain cell types [43, 44]. To explore whether thrombin could induce DNA synthesis, fibronectin secretion, and MCP-1 secretion in PTC cultures by MMP-mediated EGF receptor transactivation, cells were incubated with thrombin in the presence or absence of MMP inhibitor, GM6001 (0.1 lmol/L) or the EGF receptor kinase inhibitor, AG1478 (0.1 lmol/L). Neither GM6001 nor AG1478 significantly inhibited DNA synthesis, fibronectin secretion, or MCP-1 secretion (Fig. 7).

#

200 *

150 100 50

DISCUSSION This investigation has demonstrated that exposure of primary cultures of human PTC to thrombin significantly increases DNA synthesis, fibronectin secretion, MCP-1

0 Hirudin − Hirudin +

Control

Thrombin

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−

E

C

T

T+H

Fig. 5. The effects of hirudin on thrombin-induced DNA synthesis (A), fibronectin secretion (B and E), monocyte chemoattractant protein1 (MCP-1) secretion (C), and transforming growth factor-b (TGF-b) secretion (D). Confluent, quiescent, passage 2 human PTC were treated with or without thrombin (5 U/mL) in the presence (
) or absence () of hirudin (5 U/mL) for 24 hours. Each bar represents the mean ± SEM for three independent experiments each performed in triplicate. ∗ P < 0.01 vs. cells treated with thrombin alone; #P < 0.01 vs, control cells. (E) A representative Western blot of fibronectin secreted by PTC treated with either control media (C), thrombin (5 U/mL) (T), or thrombin (5 U/mL plus hirudin 5 U/mL) (T & H).

1324

Vesey et al: Thrombin and proximal tubule cells

A

A *

600

* #

500 400

*

300 #

200

DNA synthesis, % control

Fibronectin secretion, % control

150

100

50

100 0

0 Con NIL IgG

TGFβ1

Control

Thrombin

NIL

Treatment

Thrombin GM6001

AG1478

B

aTGFβ

* #

B *

180 160 140 120 100 80 60 40 20 0

#

TGFβ1

#

200

100

0

Control NIL

Con NIL IgG

Fibronectin secretion, % control

DNA synthesis, % control

#

Thrombin GM6001

AG1478

Thrombin C

Treatment

* 300

Fig. 6. A pan-specific transforming growth factor-b (TGF-b) neutralizing antibody failed to influence thrombin-induced fibronectin secretion. Fibronectin secretion was measured in conditioned media from confluent, quiescent, passage 2 human proximal tubular cells (PTCs) incubated for 24 hours in serum-free media with TGF-b1 (1 ng/mL), or thrombin (5 U/mL) together with a pan-specific TGF-b neutralizing antibody ( ) (15 lg/mL) or nonimmune control rabbit IgG antibody () (15 lg/mL) or no treatment (
) (NIL). Each bar represents the mean ± SEM for three independent experiments each performed in triplicate. ∗ P < 0.05 vs. control treated cells; #P < 0.05 vs. cells treated with thrombin plus nonimmune control IgG.

MCP-1 secretion, % control

aTGFβ

#

200

100

0 Control NIL

secretion and TGF-b1 secretion. These responses were all concentration-dependent and appeared to require the proteolytic activity of thrombin. Despite the expression of a functional thrombin receptor, PAR-1, in these cells, as demonstrated by RT-PCR and thrombin-induced intracellular calcium mobilization, the inability of the specific PAR-1 activating peptide, TFLLRN-NH 2 , to mimic these responses suggests that they are independent of PAR1 activation. Surprisingly, the PAR-2 activating peptide SLIGKV-NH 2 produced very similar responses to that of thrombin. This implies that the capacity of the less specific PAR-1 activating peptide SFLLRN-NH 2 to partially

#

Thrombin GM6001

AG1478

Fig. 7. The effects of matrix metalloproteinase (MMP) inhibition and endothelial growth factor (EGF) receptor kinase inhibition on thrombin-induced DNA synthesis, fibronectin secretion and monocyte chemoattractant protein-1 (MCP-1) secretion. Confluent, quiescent, passage 2 human proximal tubular cells (PTCs) were treated with or without thrombin (5 U/mL) in the absence (
) or presence of MMP inhibitor GM6001 (0.1 lmol/L) () or EGF receptor kinase inhibitor AG1478 (0.1 lmol/L) ( ) for 24 hours. Each bar represents the mean ± SEM for three independent experiments each performed in triplicate. ∗ P < 0.05 vs. control-treated cells. #P > 0.05 vs. cells treated with thrombin alone.

Vesey et al: Thrombin and proximal tubule cells

reproduce these responses is due to its ability to activate PAR-2 and is not in this case mediated via PAR-1. Finally, the mechanism of thrombin’s action on these cells was shown not to involve MMP activation, EGF receptor transactivation, or the autocrine action of TGF-b1. The frequent deposition of fibrin in various renal diseases, such as chronic allograft nephropathy, primary and secondary forms of glomerulonephritis, ischemic tubular necrosis, and obstructive nephropathy indicates that the coagulation cascade is activated in these diseases [1–7, 18]. Fibrin is observed within the glomerular and peritubular capillaries, along the tubular basement membrane, within the interstitial spaces, and in the glomerulus itself [7, 8]. While there is a strong positive correlation between fibrin deposits, monocyte/macrophage infiltration and degree of fibrosis, the exact mechanism and relevance of this to the pathogenic processes are unclear [45, 46]. The production of active thrombin from prothrombin with the subsequent conversion of fibrinogen to fibrin represents the final stage of the coagulation cascade. Thrombin associated with fibrin clots can remain active and protected from inhibitors over prolonged periods. Subsequent fibrinolysis can result in thrombin release [18, 47, 48]. The degradation of fibrin is regulated predominantly by the activity of the protease plasmin, which is, in turn, tightly controlled by plasminogen activator inhibitors (PAI-1). Enhanced levels of PAI-1 are reported to contribute to persistent fibrin deposition and also interstitial fibrosis [4, 5, 49]. Tissue factor is the main physiologic initiator of blood coagulation and its abnormal or elevated expression within the kidney in disease conditions appears to be responsible for the activation of the coagulation cascade and fibrin deposition [6, 50]. Activated infiltrating monocytes/macrophages are also known to express elevated tissue factor and there are also a number of reports that demonstrate macrophages can express a prothrombinase activity and have thrombin-like activity [51–53]. In our PTC cultures, mRNA for two of the three known thrombin receptors, PAR-1 and PAR-3, were expressed. In addition, mRNA for PAR-2, a receptor activated by other serine proteases such as trypsin and tryptase, was expressed. Previous studies which examined PAR-1 expression in PTC have differed. In one study where immunohistochemistry and in situ-hybridisation were used, PAR-1 expression could not be detected [14]. In contrast, the presence of PAR-1 in HK-2 cells and normal kidney sections has been reported and elevated PAR-1 expression found in chronic allograft nephropathy [7, 18]. They showed that PAR-1 immunoactivity was predominantly located on the basolateral aspect of PTC but also to some extent on the luminal side. Because of the conflicting evidence of these studies, precautions were taken to eliminate possible false positive results. PAR-1 mRNA expression was assessed using

1325

two different sets of PAR-1–specific primers, the chance of genomic DNA contamination of RNA was reduced by inclusion of a DNase treatment step, and samples for RT-PCR were included in which no reverse transcriptase was present at the RT stage. PAR-3 is reported to lack an intracellular signaling capacity and is thought, at least in mouse platelets, to be a cofactor for PAR-4 [12]. The relevance to PAR-3 expression in the absence of PAR-4 expression is unclear. To confirm that functional PAR receptors are expressed by PTC in culture, cells were challenged with PAR agonists and calcium mobilization monitored. Thrombin and the PAR peptides TFLLRN-NH 2 , SFLLRN-NH 2 , and SLIGKV-NH 2 all induced rapid increases in intracellular Ca2+ , suggesting that functional PAR-1 and PAR-2 are expressed by PTC in culture. The PAR-2–specific peptide, SLIGKV-NH 2 , induced the largest rise in intracellular calcium while the PAR-1– specific peptide, TFLLRN-NH 2 , produced signals at about 34% of SLIGKV-NH 2 . SFLLRN-NH 2 which has been shown to activate both PAR-1 and PAR-2 [54] produced a larger signal at 52% of that due to SLIGKV-NH 2 . A DNA synthetic response to thrombin has been documented in a variety of different cell types, including fibroblasts, airway and vascular smooth muscle cells, keratinocytes, and mesangial cells [14, 55–58]. There has been one report, that we are aware of, that demonstrates that PTCs undergo DNA synthesis in response to thrombin [7]. In this study conducted on the SV40 transformed PTC line, HK-2, it was found that thrombin at 5 U/mL induces a greater than 300% increase in DNA synthesis, which was accompanied by similar increases in cell numbers. Our results support these findings but are more likely to represent PTC in vivo because of our use of low passage, primary nontransformed cells. Ischemia is a common insult in acute renal disease and is often accompanied by activation of the coagulation pathway. It has been suggested that thrombin represents a tubule growth factor responsible for the initial steps toward cellular recovery and regeneration after tubule injury [7]. On the other hand, tubule DNA synthesis and proliferation is often seen in the initial stages of tubule cell activation, inflammation, and fibrosis and can be mediated by fibrogenic growth factors such as insulin-like growth factor (IGF-1) [59]. Clearly, there is a fine-line between the benefits of regeneration after tubule injury and the prolonged uncontrolled inflammatory/proliferative responses derived from repeated cycles of cellular injury. It has been reported that the ability of thrombin to stimulate DNA synthesis in various cell types is mediated indirectly by the induction and autocrine action of a range of growth factors and cytokines. These include TGF-b, PDGF, prostaglandins E 2 , epiregulin, heparinbinding EGF, basic FGF (bFGF), and vascular endothelial growth factor (VEGF). To exclude all these growth

1326

Vesey et al: Thrombin and proximal tubule cells

factors and cytokines as mediators of the DNA synthetic response seen here would be difficult. The enhanced secretion of TGF-b1 noted here in response to thrombin would not account for the enhanced DNA synthesis observed as TGF-b1 inhibits DNA synthesis in PTC [29, 30]. On the other hand the fibronectin secretion could be mediated by the autocrine action of TGF-b1. Fibronectin is a noncollagenous ECM glycoprotein that can be produced by a wide variety of cell types. It is known to play an important part in wound healing, including clot formation, cell adhesion and migration, and matrix deposition. It is a chemoattractant for fibroblasts, keratinocytes, epithelial cells, and monocyte/macrophages, a scaffold protein for endothelial organization and template for collagen deposition [60]. In the present study, thrombin-induced fibronectin could conceivably be involved in any of these processes. We have previously reported that primary PTCs secrete significant amounts of fibronectin and that fibrogenic cytokines such as TGF-b1 and interleukin-1b (IL1b) can greatly enhance its production [30]. This increase was predominantly in the secreted form. The amount of cell-associated fibronectin was significantly increased by IL-1b in renal cortical fibroblasts but not PTCs. In the present study thrombin induced a similar response to these cytokines. Immunoblotting demonstrated that the fibronectin migrated as a single band with an apparent molecular weight of 220 kD, and evidence of fibronectin fragmentation was not observed. Several reports have shown that thrombin is able to cleave fibronectin [36, 61]. The possibility that fibronectin present in the conditioned culture media does not represent increased levels of synthesis and secretion but rather just proteolytic release of cell associated fibronectin was considered. In this case increased fibronectin in the media may be accompanied by reduced amount of cell-associated fibronectin. Immunoblotting demonstrated that increased fibronectin in the media was coupled with decreased cell-associated fibronectin. However, this could not account for all of the fibronectin observed in the media after thrombin treatment. It is probable that increased synthesis and secretion as well as the proteolytic release of cell-associated fibronectin are involved. The finding that hirudin is able to block fibronectin secretion would be compatible with this result. Previous studies in mesangial cells demonstrated that thrombin can enhance TGF-b1 production [20]. We found a similar response to thrombin in PTC. As TGF-b can induce fibronectin secretion by PTC in culture, we tested the possibility that this TGF-b might be responsible for the action of thrombin. A pan-specific TGFb neutralizing antibody was unable to influence this response to thrombin while efficiently blocking the response to TGFb1. TGF-b1 may be secreted in an inactive form but this

does not preclude its ability to act in a paracrine/autocrine fashion in vivo where activating mechanisms would exist. Alternatively, the amount of secreted TGF-b1 may not have been enough to stimulate fibronectin secretion in these cells. MCP-1 is a powerful monocyte chemotactic molecule that can be produced by a range of renal cell types, including tubule epithelium, mesangial, and endothelial cells [40]. Infiltrating monocytes/macrophages in turn secrete cytokines and chemokines, which promote inflammation and fibrosis. However, not only is MCP-1 a chemotactic agent but it has a range of other actions, which are only starting to be described. MCP-1 is able to directly activate tubular cells in vitro, promoting cytokine production (e.g., IL-6), and adhesion molecule expression [e.g., intercellular adhesion molecule-1 (ICAM-1)], via a G protein, protein kinase C, nuclear factor-kappaB (NF-kB)-dependent mechanism [40]. Its importance in renal disease and in particular the development of interstitial lesions is clearly demonstrated in MCP-1 knockout mice which exhibit less tubulointerstitial lesions [62]. Increased monocyte/macrophage infiltration is also a common feature of renal disease and correlates strongly with the presence of fibrin [45]. Our finding which demonstrated that thrombin enhances MCP-1 secretion by PTC supports the view that thrombin enhances MCP-1–induced infiltration of inflammatory cells in interstitial renal disease. As both thrombin and fibrin have also been shown to be chemotactic for monocytes together with MCP-1 and fibronectin the chemotactic signal for monocyte/macrophages infiltration would be amplified. The inability of the PAR-1–specific peptide, TFLLRNNH 2 , to reproduce the actions of thrombin examined here, indicates that these responses are not mediated solely by PAR-1 activation. This is despite the presence of functional PAR-1 on PTC and the apparent requirement for thrombin’s proteolytic activity. That the peptide SFLLRN-NH 2 elicits some of the responses produced by thrombin can most readily be explained by crossactivation of PAR-2 [54]. The other thrombin receptor identified in our PTC cultures was PAR-3. This receptor is not thought to mediate biologic responses in cells independently of another thrombin receptor, PAR-4. We failed to find PAR-4 expression in our PTC cultures and thus PAR-3 is unlikely to contribute to the responses to thrombin observed here. These results imply that a non-PAR thrombin receptor is involved. A number of publications have clearly demonstrated that a noncatalytic domain of thrombin, a site that is responsible for its high affinity binding to cell surfaces, can induce biologic responses that are distinct from those produced by its proteolytic active site [63–65]. Interestingly, Norfleet, Bergmann, and Carney [64] have shown that a 23 amino acid thrombin peptide which lacks any proteolytic

Vesey et al: Thrombin and proximal tubule cells

activity can induce mitogenesis and chemotaxis in cultured cells and dermal wound healing in animals. A dual signaling model for thrombin activity has been proposed whereby proteolytic cleavage and nonproteolytic binding are required for a full physiologic response [64, 65]. Our findings are compatible with this model. The thrombin induced responses observed here may well require an input from both PAR-1 and thrombin’s nonproteolytic binding site. Activation of PAR-1 by TFLLRN-NH 2 is not sufficient. In our studies, hirudin, which specifically blocks thrombin’s proteolytic activity, was able to block thrombininduced DNA synthesis and MCP-1, fibronectin, and TGF-b1 secretion. As hirudin is a fairly large molecule, it is possible that the noncatalytic domain of thrombin is also blocked by hirudin and that this is the mechanism by which it abrogates the responses of these cells to thrombin. Further studies are warranted to elucidate this possibility. Two recent publications support the possibility of further thrombin receptors. In one study in airway smooth muscle cells, thrombin mediated both growth and proinflammatory actions by a non-PAR-1, G protein–coupled receptor-dependent mechanism [66]. The other showed that thrombin’s antiapoptotic activity in osteoblasts involves thrombin-induced secretion of survival factors but not PAR activation [67]. Another potential mechanism by which thrombin induces the proinflammatory and proproliferative responses seen here, involves pro-MMP activation and EGF receptor transactivation. Lafleur et al [41] have shown that thrombin can activate pro-MMP-2 in human umbilical vascular endothelial cells (HUVEC) [41] and it is well documented that EGF receptor ligands, including heparin-binding EGF and TGF-a, can be produced in various cell types by a MMP-dependent mechanism [43, 44]. In this study, GM6001 (a broad-spectrum MMP inhibitor) and AG1478 (a potent EGF receptor kinase inhibitor) were unable to influence thrombin-induced DNA synthesis, fibronectin secretion, or MCP-1 secretion suggesting some alternative mechanism must exist. It is possible that other growth factors involved in these responses are produced by PTC in response to thrombin [21, 22]. CONCLUSION We have demonstrated that the serine protease thrombin stimulates proinflammatory and fibroproliferative responses in primary cultures of human PTC. These responses, which appear to be mediated by a PAR-1– independent mechanism, may help explain the extent of tubulointerstitial fibrosis observed in renal diseases associated with fibrin deposition. Although the design of therapies that specifically block PAR-1 activation in the

1327

kidney may well prove to be useful strategies to limit the progression of renal fibrosis, combined approaches that also consider the inhibition of serine protease activity, may well have greater potential.

ACKNOWLEDGMENTS The invaluable assistance provided by Ms. Shirley Callaghan (PAH, Tumor Tissue Bank) and the urologists of the Princess Alexandra Hospital in the procurement of human renal tissue are gratefully acknowledged. The study was supported by funds from the National Health and Medical Research Council and Princess Alexandra Hospital Research and Development Foundation. This work was presented in abstract form at the Australian and New Zealand Society for Nephrology Annual Meeting, Perth, Australia, September 2003. Reprint requests to Dr. David A. Vesey, Department of Renal Medicine, Level 2, Ambulatory Renal and Transplant Services Building, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Brisbane Qld 4102, Australia. E-mail: david˙[email protected]

REFERENCES 1. KINCAID-SMITH P: Coagulation and renal disease. Kidney Int 2:183– 190, 1972 2. ENESTROM S, DRUID H, RAMMER L: Fibrin deposition in the kidney in post-ischaemic renal damage. Br J Exp Pathol 69:387–394, 1988 3. FAULK WP, GARGIULO P, MCINTYRE JA, et al: Hemostasis and fibrinolysis in renal transplantation. Semin Thromb Hemost 15:88–98, 1989 4. WANG Y, PRATT JR, HARTLEY B, et al: Expression of tissue type plasminogen activator and type 1 plasminogen activator inhibitor, and persistent fibrin deposition in chronic renal allograft failure. Kidney Int 52:371–377, 1997 5. WANG Y, PRATT JR, TAM FW, et al: Up-regulation of type 1 plasminogen activator inhibitor messenger RNA with thrombotic changes in renal grafts. Transplantation 61:684–689, 1996 6. WENDT T, ZHANG YM, BIERHAUS A, et al: Tissue factor expression in an animal model of hydronephrosis. Nephrol Dial Transplant 10:1820–1828, 1995 7. GRANDALIANO G, MONNO R, RANIERI E, et al: Regenerative and proinflammatory effects of thrombin on human proximal tubular cells. J Am Soc Nephrol 11:1016–1025, 2000 8. GRANDALIANO G, GESUALDO L, SCHENA FP: Thrombin: a novel renal growth factor. Exp Nephrol 7:20–25, 1999 9. CHAMBERS RC, LAURENT GJ: Coagulation cascade proteases and tissue fibrosis. Biochem Soc Trans 30:194–200, 2002 10. HOLLENBERG MD, COMPTON SJ, INTERNATIONAL UNION OF PHARMACOLOGY: XXVIII. Proteinase-activated receptors. Pharmacol Rev 54:203–217, 2002 11. OSSOVSKAYA VS, BUNNETT NW: Protease-activated receptors: Contribution to physiology and disease. Physiol Rev 84:579–621, 2004 12. O’BRIEN PJ, MOLINO M, KAHN M, et al: Protease activated receptors: Theme and variations. Oncogene 20:1570–1581, 2001 13. STRUKOVA SM: Thrombin as a regulator of inflammation and reparative processes in tissues. Biochemistry (Mosc) 66:8–18, 2001 14. RONDEAU E, HE CJ, ZACHARIAS U, et al: Thrombin stimulation of renal cells. Semin Thromb Hemost 22:135–138, 1996 15. HE CJ, RONDEAU E, MEDCALF RL, et al: Thrombin increases proliferation and decreases fibrinolytic activity of kidney glomerular epithelial cells. J Cell Physiol 146:131–140, 1991 16. KAUFMANN R, JUNKER U, NUSKE K, et al: PAR-1- and PAR-3-type thrombin receptor expression in primary cultures of human renal cell carcinoma cells. Int J Oncol 20:177–180, 2002 17. GUI Y, LOUTZENHISER R, HOLLENBERG MD: Bidirectional regulation of renal hemodynamics by activation of PAR1 and PAR2 in isolated perfused rat kidney. Am J Physiol Renal Physiol 285:F95–F104, 2003

1328

Vesey et al: Thrombin and proximal tubule cells

18. GRANDALIANO G, DI PAOLO S, MONNO R, et al: Protease-activated receptor 1 and plasminogen activator inhibitor 1 expression in chronic allograft nephropathy: The role of coagulation and fibrinolysis in renal graft fibrosis. Transplantation 72:1437–1443, 2001 19. CUNNINGHAM MA, RONDEAU E, CHEN X, et al: Protease-activated receptor 1 mediates thrombin-dependent, cell-mediated renal inflammation in crescentic glomerulonephritis. J Exp Med 191:455– 462, 2000 20. YAMABE H, OSAWA H, INUMA H, et al: Thrombin stimulates production of transforming growth factor-beta by cultured human mesangial cells. Nephrol Dial Transplant 12:438–442, 1997 21. WEISS RH, MADURI M: The mitogenic effect of thrombin in vascular smooth muscle cells is largely due to basic fibroblast growth factor. J Biol Chem 268:5724–5727, 1993 22. SHULTZ PJ, KNAUSS TC, MENE P, et al: Mitogenic signals for thrombin in mesangial cells: Regulation of phospholipase C and PDGF genes. Am J Physiol 257:F366–F374, 1989 23. FRUNGIERI MB, WEIDINGER S, MEINEKE V, et al: Proliferative action of mast-cell tryptase is mediated by PAR2, COX2, prostaglandins, and PPARgamma: Possible relevance to human fibrotic disorders. Proc Natl Acad Sci USA 99:15072–15077, 2002 24. RAGOSTA M, BARRY WL, GIMPLE LW, et al: Effect of thrombin inhibition with desulfatohirudin on early kinetics of cellular proliferation after balloon angioplasty in atherosclerotic rabbits. Circulation 93:1194–1200, 1996 25. RADE JJ, SCHULICK AH, VIRMANI R, et al: Local adenoviral-mediated expression of recombinant hirudin reduces neointima formation after arterial injury. Nat Med 2:293–298, 1996 26. NELKEN NA, SOIFER SJ, O’KEEFE J, et al: Thrombin receptor expression in normal and atherosclerotic human arteries. J Clin Invest 90:1614–1621, 1992 27. HERNANDEZ-RODRIGUEZ NA, CAMBREY AD, HARRISON NK, et al: Role of thrombin in pulmonary fibrosis. Lancet 346:1071–1073, 1995 28. JOHNSON DW, BREW BK, PORONNIK P, et al: Transport characteristics of human proximal tubule cells in primary culture. Nephrology 3:183–194, 1997 29. JOHNSON DW, SAUNDERS HJ, BAXTER RC, et al: Paracrine stimulation of human renal fibroblasts by proximal tubule cells. Kidney Int 54:747–757, 1998 30. VESEY DA, CHEUNG CW, CUTTLE L, et al: Interleukin-1beta induces human proximal tubule cell injury, alpha-smooth muscle actin expression and fibronectin production. Kidney Int 62:31–40, 2002 31. CHRISTIANSEN BS, INGERSLEV J, HEICKENDORFF L, et al: Human hepatocytes in culture synthesize and secrete fibronectin. Scand J Clin Lab Invest 48:685–690, 1988 32. ASOKANANTHAN N, GRAHAM PT, FINK J, et al: Activation of proteaseactivated receptor (PAR)-1, PAR-2, and PAR-4 stimulates IL-6, IL8, and prostaglandin E 2 release from human respiratory epithelial cells. J Immunol 168:3577–3585, 2002 33. LIU Y, GILCREASE MZ, HENDERSON Y, et al: Expression of proteaseactivated receptor 1 in oral squamous cell carcinoma. Cancer Lett 169:173–180, 2001 34. RONDEAU E, VIGNEAU C, BERROU J: Role of thrombin receptors in the kidney: Lessons from PAR1 knock-out mice. Nephrol Dial Transplant 16:1529–1531, 2001 35. SABRI A, SHORT J, GUO J, et al: Protease-activated receptor-1mediated DNA synthesis in cardiac fibroblast is via epidermal growth factor receptor transactivation: Distinct PAR-1 signaling pathways in cardiac fibroblasts and cardiomyocytes. Circ Res 91:532–539, 2002 36. GALDAL KS, EVENSEN SA, NILSEN E: The effect of thrombin on fibronectin in cultured human endothelial cells. Thromb Res 37:583– 593, 1985 37. MORII T, FUJITA H, NARITA T, et al: Association of monocyte chemoattractant protein-1 with renal tubular damage in diabetic nephropathy. J Diabetes Comp 17:11–15, 2003 38. LLOYD CM, MINTO AW, DORF ME, et al: RANTES and monocyte chemoattractant protein-1 (MCP-1) play an important role in the inflammatory phase of crescentic nephritis, but only MCP-1 is involved in crescent formation and interstitial fibrosis. J Exp Med 185:1371–1380, 1997 39. DAI C, LIU Z, ZHOU H, et al: Monocyte chemoattractant protein-1

40. 41.

42.

43.

44.

45. 46. 47. 48. 49. 50.

51. 52. 53.

54.

55.

56.

57.

58.

59.

expression in renal tissue is associated with monocyte recruitment and tubulointerstitial lesions in patients with lupus nephritis. Chin Med J (Engl) 114:864–868, 2001 VIEDT C, ORTH SR: Monocyte chemoattractant protein-1 (MCP-1) in the kidney: Does it more than simply attract monocytes? Nephrol Dial Transplant 17:2043–2047, 2002 LAFLEUR MA, HOLLENBERG MD, ATKINSON SJ, et al: Activation of pro-(matrix metalloproteinase-2) (pro-MMP-2) by thrombin is membrane-type-MMP-dependent in human umbilical vein endothelial cells and generates a distinct 63 kDa active species. Biochem J 357:107–115, 2001 DARMOUL D, GRATIO V, DEVAUD H, et al: Activation of proteinaseactivated receptor 1 promotes human colon cancer cell proliferation through epidermal growth factor receptor transactivation. Mol Cancer Res 2:514–522, 2004 UCHIYAMA-TANAKA Y, MATSUBARA H, MORI Y, et al: Involvement of HB-EGF and EGF receptor transactivation in TGF-beta-mediated fibronectin expression in mesangial cells. Kidney Int 62:799–808, 2002 ZHAO D, ZHAN Y, KOON HW, et al: Metalloproteinase-dependent transforming growth factor-{alpha} release mediates neurotensinstimulated MAP kinase activation in human colonic epithelial cells. J Biol Chem 279:43547–43554, 2004 DREW AF, TUCKER HL, LIU H, et al: Crescentic glomerulonephritis is diminished in fibrinogen-deficient mice. Am J Physiol Renal Physiol 281:F1157–F1163, 2001 TIPPING PG, DOWLING JP, HOLDSWORTH SR: Glomerular procoagulant activity in human proliferative glomerulonephritis. J Clin Invest 81:119–125, 1988 GANDOSSI E, LUNVEN C, BERRY CN: Role of clot-associated (-derived) thrombin in cell proliferation induced by fibrin clots in vitro. Br J Pharmacol 129:1021–1027, 2000 WILNER GD, DANITZ MP, MUDD MS, et al: Selective immobilization of alpha-thrombin by surface-bound fibrin. J Lab Clin Med 97:403– 411, 1981 REROLLE JP, HERTIG A, NGUYEN G, et al: Plasminogen activator inhibitor type 1 is a potential target in renal fibrogenesis. Kidney Int 58:1841–1850, 2000 ERLICH JH, HOLDSWORTH SR, TIPPING PG: Tissue factor initiates glomerular fibrin deposition and promotes major histocompatibility complex class II expression in crescentic glomerulonephritis. Am J Pathol 150:873–880, 1997 MANSELL A, REINICKE A, WORRALL DM, et al: The serine protease inhibitor antithrombin III inhibits LPS-mediated NF-kappaB activation by TLR-4. FEBS Lett 508:313–317, 2001 GIUNTOLI DL, RETZINGER GS: Evidence for prothrombin production and thrombin expression by phorbol ester-treated THP-1 cells. Exp Mol Pathol 64:53–62, 1997 PEJLER G, LUNDERIUS C, TOMASINI-JOHANSSON B: Macrophages synthesize factor X and secrete factor X/Xa-containing prothrombinase activity into the surrounding medium. Thromb Haemost 84:429–435, 2000 HOLLENBERG MD, SAIFEDDINE M, AL ANI B, et al: Proteinaseactivated receptors: Structural requirements for activity, receptor cross-reactivity, and receptor selectivity of receptor-activating peptides. Can J Physiol Pharmacol 75:832–841, 1997 CHAMBERS RC, DABBAGH K, MCANULTY RJ, et al: Thrombin stimulates fibroblast procollagen production via proteolytic activation of protease-activated receptor 1. Biochem J 333 (Pt 1):121–127, 1998 FERNANDES D, VLAHOS R, STEWART AG: Thrombin-stimulated DNA synthesis in human cultured airway smooth muscle occurs independently of products of cyclo-oxygenase or 5-lipoxygenase. Pulm Pharmacol Ther 13:241–248, 2000 KANTHOU C, BENZAKOUR O, PATEL G, et al: Thrombin receptor activating peptide (TRAP) stimulates mitogenesis, c-fos and PDGF-A gene expression in human vascular smooth muscle cells. Thromb Haemost 74:1340–1347, 1995 KAUFMANN K, THIEL G: Epidermal growth factor and thrombin induced proliferation of immortalized human keratinocytes is coupled to the synthesis of Egr-1, a zinc finger transcriptional regulator. J Cell Biochem 85:381–391, 2002 NAKAGAWA T, KANG DH, OHASHI R, et al: Tubulointerstitial disease:

Vesey et al: Thrombin and proximal tubule cells

60.

61. 62.

63.

Role of ischemia and microvascular disease. Curr Opin Nephrol Hypertens 12:233–241, 2003 KANG YH, KEDAR VP, MAHESHWARI RK: Thrombin stimulation of synthesis and secretion of fibronectin by human A549 epithelial cells and mouse LB fibroblasts. J Histochem Cytochem 39:413–423, 1991 GOLDFARB RH, LIOTTA LA: Thrombin cleavage of extracellular matrix proteins. Ann NY Acad Sci 485:288–292, 1986 TESCH GH, SCHWARTING A, KINOSHITA K, et al: Monocyte chemoattractant protein-1 promotes macrophage-mediated tubular injury, but not glomerular injury, in nephrotoxic serum nephritis. J Clin Invest 103:73–80, 1999 CARNEY DH, REDIN W, MCCROSKEY L: Role of high-affinity thrombin receptors in postclotting cellular effects of thrombin. Semin Thromb Hemost 18:91–103, 1992

1329

64. NORFLEET AM, BERGMANN JS, CARNEY DH: Thrombin peptide, TP508, stimulates angiogenic responses in animal models of dermal wound healing, in chick chorioallantoic membranes, and in cultured human aortic and microvascular endothelial cells. Gen Pharmacol 35:249–254, 2000 65. HOLLENBERG MD, MOKASHI S, LEBLOND L, et al: Synergistic actions of a thrombin-derived synthetic peptide and a thrombin receptoractivating peptide in stimulating fibroblast mitogenesis. J Cell Physiol 169:491–496, 1996 66. TRAN T, STEWART AG: Protease-activated receptor (PAR)independent growth and pro-inflammatory actions of thrombin on human cultured airway smooth muscle. Br J Pharmacol 138:865– 875, 2003 67. PAGEL CN, DE NIESE MR, ABRAHAM LA, et al: Inhibition of osteoblast apoptosis by thrombin. Bone 33:733–743, 2003