Regenerative and Proinflammatory Effects of Thrombin onHuman Proximal Tubular Cells

GIUSEPPE GRANDALIANO, RAFFAELLA MONNO, ELENA RANIERI,LORETO GESUALDO, and FRANCESCO P. SCHENA,WITH THE TECHNICAL ASSISTANCE OF

CARMELA MARTINO AND MICHELE URSIDivision of Nephrology/Department of Emergency and Transplantation, University of Bari, Italy.

Abstract.Interstitial fibrin deposition is a common histologicfeature of tubulointerstitial diseases, which suggests that thecoagulation system is activated. Thrombin, generated duringthe activation of the coagulation cascade, is a powerful acti-vating factor for different cell types. Although proximal tubu-lar cells are potential targets for this coagulation factor, noinformation is available on the effect of thrombin on thesecells. Thus, the expression of protease-activated receptor-1(PAR-1), the main thrombin receptor, was investigated inhuman proximal tubular cells (hPTC)in vivo and in vitro. Adiffuse expression of PAR-1 was observed by immunohisto-chemistry along the basolateral membrane of PTC in normalhuman kidney. This observation was confirmedin vitro incultured hPTC. Because tubular damage and monocyte infil-tration are two hallmarks of tubulointerstitial injury, the effectof thrombin on DNA synthesis and monocyte chemotacticpeptide-1 (MCP-1) gene and protein expression was evaluatedin cultured hPTC. Thrombin induced a significant and dose-

dependent increase in thymidine uptake and a striking upregu-lation of MCP-1 mRNA expression and protein release into thesupernatant. Although PAR-1 is a G protein-coupled receptor,its activation in hPTC, as in other cell systems, resulted in atransient increase in cellular levels of tyrosine-phosphorylatedproteins. An increased level of tyrosine-phosphorylated c-srcsuggested the activation of this cytoplasmic tyrosine kinase inresponse to thrombin and its potential role in thrombin-inducedprotein-tyrosine phosphorylation. Interestingly, thrombin-in-duced DNA synthesis and MCP-1 gene expression were com-pletely blocked by genistein, a specific tyrosine kinase inhib-itor, but not by its inactive analogue daidzein, demonstrating acentral role for tyrosine kinase activation in the thrombineffects on hPTC. Moreover, the specific src inhibitor PP1abolished the thrombin effect on DNA synthesis. In conclu-sion, thrombin might represent a powerful regenerative andproinflammatory stimulus for hPTC in acute and chronic tu-bulointerstitial diseases.

Fibrin deposition in the peritubular capillaries and along thetubular basement membrane is commonly observed in severalrenal diseases characterized by tubular and/or interstitial dam-age, such as ischemic tubular necrosis, obstructive nephropa-thy, experimental lupus nephritis, and acute as well as chronicrenal allograft rejection (1–6). This observation indirectly sug-gests the activation of the coagulation cascade at the interstitiallevel in the development of acute and chronic renal damage.There is an increasing body of evidence supporting an involve-ment of the coagulation system in the pathogenesis of glomer-ular lesions both in human and in experimental glomerulone-phritides, whereas the consequences of the coagulation cascadeactivation in interstitial diseases have not received much atten-tion (7–9).

The activation of the coagulation cascade leads to the partialproteolysis of prothrombin with subsequent local thrombin

generation (10). This serine protease may then be accumulatedin its active form within the fibrin clots and released locallyover a prolonged period of time (11). Thrombin, besides itsphysiologic action in the clotting cascade, is known to modu-late a variety of cell functions through the interaction withspecific cell surface receptors (12,13). All of the known throm-bin receptors belong to the protease-activated receptor (PAR)family and are characterized by a peculiar proteolytic mecha-nism of activation (14–16). Indeed, receptor activation occurswhen thrombin cleaves the extracellular domain of the receptorexposing a tethered ligand (14). Among the receptors of thePAR family, thrombin can interact specifically with PAR-1, -3,and -4 (14–16). Of these three signaling receptor proteins,however, only PAR-1 has been shown to be expressed in thekidney (14,17). Although proximal tubular cells (PTC) mayrepresent a potential target for thrombin, no information isavailable on the presence of PAR-1 on their surface as well ason their activation by thrombin.

Tubular damage and monocyte infiltration are two of thehistopathologic hallmarks of acute and chronic tubulointersti-tial injury (18,19). Tubular necrosis and atrophy, especially inthe setting of acute tubulointerstitial damage, are potentiallyreversible, although the mechanisms responsible for the regen-erative response are still poorly understood (18). Monocyteinterstitial infiltration is a key step in the pathogenesis of

Received November 16, 1998. Accepted October 18, 1999.Correspondence to Dr. Giuseppe Grandaliano, Division of Nephrology/Depart-ment of Emergency and Transplantation, University of Bari, Polyclinic, PiazzaGiulio Cesare, 11, 70124 Bari, Italy. Phone:139 080 5592787; Fax:139 0805575710; E-mail: g.grandaliano@nephro.uniba.it

1046-6673/1106-1016Journal of the American Society of NephrologyCopyright © 2000 by the American Society of Nephrology

J Am Soc Nephrol 11: 1016–1025, 2000

tubulointerstitial damage and monocyte chemotactic peptide(MCP-1), expressed by tubular cells, may be involved in themonocyte influx into the interstitial space (19,20).

In the present study, PAR-1 expression by human PTC wasinvestigatedin vivoandin vitro. Moreover, the thrombin effecton human PTC mitogenesis and MCP-1 gene and proteinexpression was evaluated.

Materials and MethodsReagents

Dulbecco’s modified Eagle’s medium (DMEM)/F12, trypsin, pen-icillin, and streptomycin were obtained from Mascia Brunelli (Milan,Italy). Fetal bovine serum (FBS),L-glutamine, sodium pyruvate, non-essential amino acids, insulin, transferrin, and selenium were fromSigma Cell Culture (Milan, Italy). Bovine thrombin, genistein, daid-zeim, herbimycin A, prostaglandin E1, hydrocortisone, and T3 werefrom Sigma Chemical Co. (Milan, Italy). Epidermal growth factor(EGF), PP1, and PP3 were obtained from Calbiochem (La Jolla, CA).The monoclonal antibody directed against the extracellular domain ofhuman PAR-1 was kindly provided by Dr. L. F. Brass (ThomasJefferson University, Philadelphia, PA). Polyclonal rabbit anti-humansrc antibody was purchased from Santa Cruz Biotechnology (Heidel-berg, Germany). The monoclonal anti-phosphotyrosine antibody Py20was obtained from Upstate Biotechnology (Lake Placid, NY). Thehorseradish-peroxidase-conjugated sheep anti-mouse and sheep anti-rabbit antibodies were obtained from Amersham (Buckinghamshire,United Kingdom). [32P]dCTP and [methyl-3H]-thymidine were pur-chased from Amersham. All other chemicals were reagent grade.

Cell Isolation and CultureHK2, an immortalized PTC line from normal adult human kidney

(21), was obtained from American Type Culture Collection (Manas-sas, VA). Cells were grown to confluence in DMEM/F12 mediumsupplemented with 5% FBS, 100 U/ml penicillin, 100mg/ml strepto-mycin, 2 mM L-glutamine, 5mg/ml insulin, 5 mg/ml transferrin, 5ng/ml sodium selenite, 5 pg/ml T3, 5 ng/ml hydrocortisone, 5 pg/mlprostaglandin E1, and 10 ng/ml EGF. For passage, confluent cellswere washed with phosphate-buffered saline (PBS), removed with0.05% trypsin/0.02% ethylenediaminetetra-acetic acid in PBS, andplated in DMEM/F12 medium.

Immunohistochemistry and ImmunocytochemistryApparently normal kidney portions from 10 patients undergoing

nephrectomy for renal cell carcinoma and cultured HK2 cells platedon 8-well multitest slides (ICN, Aurora, OH) were used to investigatePAR-1 protein expression. The detection of this thrombin receptorwas performed on frozen 4-mm-thick kidney sections and on subcon-fluent HK2, fixed in 4% paraformaldehyde, using specific mousemonoclonal anti-human PAR-1 antibody directed against an epitope(corresponding to PAR-1 residues 51 to 64) in the N terminus of thereceptor that is retained after PAR-1 cleavage at 1:200 dilution (17).HK2 cells were serum-starved overnight and incubated in serum-freemedium with or without thrombin (5 U/ml) for 15 min, beforefixation. The anti-PAR-1 antibody recognizes the extracellular do-main of the inactive as well as of active PAR-1. Immobilized mouseantibodies were detected by the immunoalkaline phosphatase(APAAP) method with affinity-purified rabbit anti-mouse IgG (Dako,Glostrup, Denmark) and APAAP complex (1:50 dilution; Dako),following a two-step technique as described previously (19). Alkalinephosphatase was developed with New Fuchsin (Sigma). Negative

controls were performed by omitting the primary or secondary anti-bodies, and using nonimmune mouse or rabbit antiserum as first layer.

3H-Thymidine IncorporationDNA synthesis was measured as the amount of [methyl-3H]-thy-

midine incorporated into TCA-precipitable material, as describedpreviously (22). Briefly, HK2 were plated in 24-well dishes at adensity of 43 104 cells/well, grown to confluence, and made quies-cent by being placed in serum-free medium for 48 h. The cellmonolayer was then incubated with thrombin at the indicated concen-trations for 24 h at 37°C. In separate sets of experiments, cells werepreincubated with genistein (25mM), daidzein (25mM), PP1 (25mM), and PP3 (25mM) for 18 h before adding thrombin. At the endof the incubation period, cells were pulsed for 4 h with 1.0mCi/ml3H-thymidine. The medium was then removed, the cells were washedtwice in ice-cold 5% TCA, and then incubated in 5% TCA for 5 min.The monolayer was solubilized by adding 0.75 ml of 0.25N NaOH in0.1% sodium dodecyl sulfate (SDS). Half-milliliter aliquots were thenneutralized and counted in scintillation fluid using a beta counter.

In parallel experiments, cell proliferation was determined by directcell counting after 24 and 48 h of incubation with thrombin (5 U/ml),as described previously (23).

RNA Isolation and Northern Blot AnalysisHK2 cells were plated in 75-mm2 flasks and cultured as described

above. After reaching confluence, cells were serum-starved for 48 hand then incubated for the indicated time periods with thrombin (5U/ml). In separate sets of experiments, cells were preincubated withgenistein (25mM), daidzein (25mM), PP1 (25mM), and PP3 (25mM)for 18 h before adding thrombin. At the end of incubation, cells werelysed with 4 M guanidinium isothiocyanate containing 25 mM sodiumcitrate, pH 7.0, 0.5% sarcosyl, and 0.1 mMb-mercaptoethanol. TotalRNA was isolated by the single-step method, using phenol and chlo-roform/isoamyl alcohol (24).

MCP-1 gene expression was studied by Northern blotting, as de-scribed previously (24). Briefly, electrophoresis of 20mg of totalRNA from each experimental condition was carried out in 1% agarosegel with 2.2 M formaldehyde. The RNA was then transferred over-night onto a nylon membrane (Schleicher & Schuell, Dassel, Germa-ny). The membrane was stained with ethidium bromide to evaluate the28S and 18S ribosomal bands and prehybridized at 42°C for 2 h in50% formamide, 0.5% SDS, 53 SSC, and 0.1 mg/ml salmon spermDNA. A 0.7-kb fragment of the baboon MCP-1 cDNA was used as aprobe (25). The DNA fragment was labeled by random priming usinga commercially available kit (Amersham) and [32P]dCTP (specificactivity, 3000 Ci/mmol). The probe (106 cpm/ml) was added to 10 mlof prehybridization solution, and the blots were hybridized for 16 h at42°C. The membranes were then washed once in 23 SSC, 0.1% SDSat room temperature for 5 min, once in the same buffer at 55°C for 30min, and in 13 SSC, 0.1% SDS at 55°C for an additional 30 min.After drying, membranes were exposed to a Kodak X-OMAT filmwith intensifying screens at270°C.

Enzyme-Linked Immunosorbent AssayHK2 cells plated in 6-well dishes and grown to 70% confluence

were serum-starved for 24 h and then incubated for 24 h in serum-freemedium with or without thrombin (5 U/ml). At the end of theincubation, the supernatant was harvested, centrifuged for 10 min at10003 g to remove the cell debris, and stored at280°C until used.MCP-1 measurement in the supernatant was performed using a com-mercial human MCP-1 enzyme-linked immunosorbent assay (ELISA)

J Am Soc Nephrol 11: 1016–1025, 2000 Thrombin and Proximal Tubular Cells 1017

kit (Quantikine; R&D Systems, Abingdon, United Kingdom). This isa multiple sandwich solid-phase enzyme immunoassay that uses aspecific monoclonal antibody raised against human MCP-1. The sen-sitivity of the ELISA is 5 pg/ml. The MCP-1 concentration of theunknown samples was determined by interpolation into a standardcurve developed with known amounts of recombinant human MCP-1protein. MCP-1 protein concentration was normalized to cell counts.

Western BlotHK2 were plated in 60-mm2 Petri dishes and grown to confluence

in DMEM/F12 medium supplemented with 10% FBS. The cells wereincubated for 48 h in serum-free medium and then exposed to throm-bin (5 U/ml) for the indicated time periods. At the end of thetreatment, the cell monolayer was rapidly rinsed twice with ice-coldPBS and lysed in 100ml of RIPA buffer (1 mM phenylmethylsulfonylfluoride, 5 mM ethylenediaminetetra-acetic acid, 1 mM sodium or-thovanadate, 150 mM sodium chloride, 8mg/ml leupeptin, 1.5%Nonidet P-40, and 20 mM Tris-HCl, pH 7.4). The lysates were set onice for 30 min and centrifuged at 10,0003 g at 4°C for 5 min. Thesupernatants were collected and stored at280°C until used. Aliquotscontaining 7.5mg of proteins from each lysate were subjected toSDS-polyacrylamide gel electrophoresis on a 7.5% gel under reducingconditions and then electrotransferred onto nitrocellulose membrane(Hybond™ C; Amersham). The filter was blocked overnight at roomtemperature with 2% bovine serum albumin in PBS containing 0.1%Tween 20 (TBS) and incubated with monoclonal anti-phosphoty-rosine antibody at room temperature for 4 h. The membranes werewashed twice in TBS and incubated for 2 h atroom temperature withhorseradish peroxidase-conjugated sheep anti-mouse IgG at 1:1500dilution in TBS. The membranes were washed three times at roomtemperature in TBS and then once with 0.1% SDS in PBS. The ECLenhanced chemiluminescence system (Amersham) was used for de-tection.

ImmunoprecipitationConfluent HK2 cells in 60-mm2 culture dishes were placed in

serum-free medium for 48 h. Thrombin (5 U/ml) was then added forthe indicated time periods. Cells were washed twice with ice-cold PBSand lysedin situ with RIPA buffer for 30 min at 4°C. The cell lysatewas centrifuged at 10,0003 g for 30 min at 4°C. One hundredmicrograms of protein from the supernatant was first incubated withanti-phosphotyrosine antibodies for 2 h on arocking platform at 4°Cand then with agarose-linked protein A for 2 h at4°C. The immuno-beads were washed twice with RIPA buffer and twice with 0.5 mMLiCl, 0.1 mM Tris-HCl, pH 7.5, 1 mM sodium orthovanadate. Thebeads were then resuspended in sample buffer and boiled. The im-munoprecipitated proteins were separated by electrophoresis on a7.5% polyacrylamide gel and transferred onto a nitrocellulose mem-brane. The membrane was blocked as described previously and incu-bated with rabbit anti-src antibody (1:1000) for 4 h atroom temper-ature, washed, and incubated with horseradish peroxidase-conjugatedmouse anti-rabbit IgG (1:1500). The ECL system was used for detec-tion of the horseradish peroxidase-coupled antibodies.

Statistical AnalysesData are presented as mean6 SD and compared by ANOVA.P ,

0.05 was considered significant.

ResultsPAR-1 expression has been demonstratedin vivoandin vitro

in glomerular cells, whereas no information is available on its

presence in tubular cells (17,25–27). To address this issue,PAR-1 protein expression was investigated in normal kidneysections by immunohistochemistry, using a specific monoclo-nal antibody that recognizes the extracellular domain of thisthrombin receptor. Although glomerular cells represented themain site of PAR-1 expression within the kidney, PTC werealso specifically and significantly stained (Figure 1, A throughC). Indeed, all of the glomeruli and 30 to 40% of the corticaltubular sections were positive for PAR-1. The new fuchsindeposits within PTC were mainly localized to the basolateralmembrane, although in a few tubular sections an apical expres-sion was also observed (Figure 1, A through C). Proximaltubular sections were recognized by cell morphology and im-munohistochemistry (Tamm–Horsfall-negative, aquaporin2-positive sections; data not shown). To confirm this observa-tion, PAR-1 protein expression was studiedin vitro, by immu-nocytochemistry, in an immortalized and well characterizedhuman PTC line (21). PAR-1 protein was strongly expressedby untreated cultured human PTC (hPTC) (Figure 2A). More-over, as shown previously in mesangial and endothelial cells(28,29), also in hPTC, incubation with thrombin (5 U/ml)induced an almost complete downregulation of PAR-1-specificstaining within 15 min (Figure 2B).

Thrombin is a powerful mitogen for several cell types inculture (26,27,30). It is conceivable that thrombin, frequentlyactivated in the setting of acute and potentially reversibletubular damage, may represent a regenerative stimulus forhPTC. To support this hypothesis, the effect of thrombin onhPTC DNA synthesis was investigated. As shown in Figure 3,the serine protease induced a dose-dependent increase in triti-ated thymidine uptake that reached statistical significance at0.05 U/ml and peaked at 5 U/ml. The proliferative effect ofthrombin was further confirmed by direct cell counting. In-deed, incubation with this protease, at the concentration of 5U/ml, caused a statistically significant increase in PTC numberafter 24 and 48 h (Figure 4).

Interstitial monocyte infiltration is a histopathologic hall-mark of acute and chronic tubulointerstitial disease (19).MCP-1 is a specific and powerful chemotactic factor for mono-cytes, and its expression is strikingly upregulated during thedevelopment of tubulointerstitial damage (20,31). The expres-sion of this chemokine has been demonstrated to be induced bythrombin in vascular smooth muscle and in endothelial andmesangial cells, but not in epithelial cells (25,32,33). Thus, theeffect of thrombin on MCP-1 gene expression was investigatedin cultured hPTC. Thrombin at the dose that maximally stim-ulated DNA synthesis caused a marked upregulation of MCP-1mRNA abundance that was evident already at 3 h and was stillpresent after 24 h (Figure 5). To determine whether the in-creased MCP-1 mRNA levels correlated with an increasedtranslation, the MCP-1 protein concentration was evaluated, byELISA, in the supernatant of serum-starved hPTC after 24 h ofincubation in serum-free medium in the presence and in theabsence of thrombin (5 U/ml). As shown in Figure 6, thrombininduced a statistically significant increase in the MCP-1 proteinreleased into cell supernatant.

Tyrosine phosphorylation of growth factor receptors plays

1018 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 1016–1025, 2000

an important role in their mitogenic and cell-activating effect(34). In the past few years, different G protein-coupled receptoragonists, including thrombin, have been shown to induce ty-rosine phosphorylation of several cellular proteins (13,34). Toinvestigate the effect of thrombin on tyrosine phosphorylationof cellular proteins in hPTC, equal amounts of protein fromunstimulated and thrombin-stimulated cells were separated bySDS-polyacrylamide gel electrophoresis and analyzed by im-munoblotting, using a specific anti-phosphotyrosine monoclo-nal antibody. Thrombin, at the dose that maximally stimulatedDNA synthesis, caused a transient increase in the cellularlevels of tyrosine-phosphorylated proteins, with the mostprominent phosphorylated bands of 60, 70, and 90 kD (Figure7). This early cellular effect of thrombin has been shown indifferent in vitro systems to be dependent on the activation ofcytoplasmic and/or receptor tyrosine kinases. In platelets,thrombin stimulation induced a strong and rapid activation ofc-src, a ubiquitous cytoplasmic tyrosine kinase, whereas infibroblasts, keratinocytes, and COS-7 cells, thrombin has beenshown to cross-activate the EGF receptor, a transmembranetyrosine kinase (35,36). To investigate whether thrombin acti-vates c-src in hPTC, the state of tyrosine phosphorylation ofthis enzyme was investigated as indirect evidence of its acti-vation. For this purpose, cell lysates from unstimulated andstimulated PTC were immunoprecipitated with anti-phospho-tyrosine antibodies and blotted with anti-src antibody. Asshown in Figure 8, thrombin induced a time-dependent in-crease in tyrosine-phosphorylated c-src that peaked at 30 min.PTC expressin vivo and in vitro the EGF receptor (37). Thus,the phosphorylation of this receptor in response to thrombin

was also evaluated, but no EGF receptor tyrosine phosphory-lation was observed upon thrombin stimulation (data notshown).

To determine the role of the early tyrosine kinase activationin thrombin-induced DNA synthesis and MCP-1 gene expres-sion, the effect of a specific tyrosine kinase inhibitor, genistein,on these two cellular responses was evaluated. Genistein at aconcentration of 25mM, a dose that completely blocks tyrosinephosphorylation, abolished completely thrombin-inducedDNA synthesis as well as MCP-1 expression, whereas itsinactive analogue daidzein was unable to influence both throm-bin effects (Figures 9 and 10). The central role of tyrosinekinase activation in thrombin-elicited DNA synthesis was fur-ther confirmed using a second specific tyrosine kinase inhibi-tor, herbimycin A, with a mechanism of action different fromgenistein (Figures 9 and 10). Herbimycin A inhibited thymi-dine incorporation and MCP-1 expression induced by thrombinto the same extent as genistein. To better define the role of srcactivation in thrombin-induced DNA synthesis, the effect of aspecific src inhibitor, PP1, was investigated. Preincubation ofPTC with PP1 at the concentration of 25mM significantlyinhibited the increase in DNA synthesis caused by thrombin,whereas PP3, the inactive analogue of PP1, at the same molarconcentration, was unable to influence the thrombin prolifera-tive effect (Figure 11).

DiscussionThe potential role of the coagulation system in glomerular

diseases has been investigated extensively over the past 30 yr(7–9). The accumulated evidence suggests that the activation

Figure 1.Protease-activated receptor-1 (PAR-1) protein expression in normal renal tissue sections. A specific immunohistochemical stainingfor PAR-1 can be detected in proximal tubular cells (A through C). The new fuchsin deposits are mainly distributed at the basolateral membrane(B), but can be observed in a few tubular sections at the apical level (C). (D) Microphotograph of a tissue section stained with non-immunemouse antiserum. Magnification:3100 in A and D;31000 in B and C.

J Am Soc Nephrol 11: 1016–1025, 2000 Thrombin and Proximal Tubular Cells 1019

of the clotting cascade, revealed by fibrin deposition, may playa significant role in the development of several glomerularlesions (7–9). Although the local priming of the clotting cas-cade within the tubulointerstitium has been suggested by sev-eral studies (1–6), the potential involvement of the coagulationsystem in the pathogenesis of tubulointerstitial damage hasbeen underestimated.

The present study demonstrated for the first time that PTCexpress bothin vivoandin vitro PAR-1, and, therefore, may beconsidered potential targets for the modulatory action ofthrombin. Interestingly, the thrombin receptor was localizedmainly at the basolateral level. This localization, facing theinterstitial space, may be pathogenically relevant. Indeed, fi-brin deposits, the potential sites of thrombin accumulation, aremainly described in the peritubular capillaries and along thetubular basement membrane (1–6). In addition, in only a fewtubular sections was luminal expression of PAR-1 observed.Although there are no reports on the presence of thrombinwithin the intratubular protein casts in proteinuric glomerular

diseases, we hypothesize that thrombin generated within theglomerular tuft may reach the urinary space and activate thePTC by interacting with PAR-1 present at the luminal level.

Figure 2. PAR-1 protein expression in unstimulated (A) and throm-bin-stimulated (B) cultured human proximal tubular cells (hPTC).Unstimulated cultured cells present a diffuse and specific staining forPAR-1 (A). Within 15 min, incubation with thrombin (5 U/ml) causedan almost complete downregulation of the PAR-1 protein expression(B). Magnification:3400.

Figure 3.Thrombin-induced DNA synthesis in hPTC. Confluent PTCwere serum-starved for 48 h and then stimulated with the indicateddoses of thrombin for 28 h. DNA synthesis was evaluated as theincorporation of 3H-thymidine into TCA-insoluble material. Datarepresent mean6 SD of four separate experiments each done intriplicate wells. cpm, counts per minute. *P , 0.01versuscontrol.

Figure 4.Thrombin-induced cell proliferation in hPTC. SubconfluentPTC were serum-starved for 48 h and then incubated in the presenceand in the absence of thrombin (5 U/ml) for the indicated time periods.The cells were then trypsinized and counted as described in Materialsand Methods. Results are representative of three experiments. *P ,0.001versuscontrol.

1020 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 1016–1025, 2000

The interaction of thrombin with the PAR-1 present on PTCleads to an increase in DNA synthesis and MCP-1 gene andprotein expression. Both of these responses may be relevant inthe setting of acute as well as chronic tubulointerstitial damage.Most of the acute conditions associated with interstitial fibrindeposition are characterized by potentially reversible tubulardamage (2,4,6). Acute renal ischemia was the first pathologiccondition in which this association was demonstrated (2). Theligation of the renal artery causes tubular necrosis and theintrarenal activation of the coagulation system with the subse-quent extensive deposition of fibrin within the interstitial space(2,38). In this scenario, the ability of thrombin to act as a“growth factor” for PTC and stimulate a regenerative responsemay represent the first step toward the potential recovery, oncethe ischemic injury has been removed.

The ability of thrombin to induce MCP-1 gene and proteinexpression could represent a key event in the development ofacute allograft rejection and in the progression of chronicrejection. Indeed, both acute and chronic renal graft rejection

are characterized by a diffuse monocytic infiltrate (39). Mono-cytes, once recruited within the interstitial space, may representa reservoir of cytokines and growth factors that can prime andmaintain the activation of resident cells (40). Although mono-cytes may play a pivotal role in the pathogenesis of interstitial

Figure 5. Thrombin-stimulated monocyte chemotactic peptide-1(MCP-1) gene expression in hPTC. Quiescent confluent PTC werestimulated with thrombin (5 U/ml) for the indicated time periods. Thecells were then harvested and total RNA was isolated as described inMaterial and Methods. MCP-1 expression was evaluated by Northernblotting (top panel). 28S and 18S ribosomal RNA bands on ethidiumbromide-stained gels were used to control the RNA loading (bottompanel). Results are representative of three experiments.

Figure 6.Thrombin-stimulated MCP-1 protein production and releasein hPTC. Subconfluent and quiescent PTC were stimulated withthrombin (5 U/ml) for 24 h. At the end of the incubation period, thesupernatant was collected and the cells were trypsinized and counted.MCP-1 protein concentration was determined by enzyme-linked im-munosorbent assay as described in Materials and Methods and nor-malized to cell number. Data represent mean6 SD of three differentexperiments. *P , 0.01versuscontrol.

Figure 7. Thrombin-induced protein tyrosine phosphorylation inhPTC. Confluent quiescent PTC were stimulated with thrombin (5U/ml) for 10, 30, and 60 min and then lysed in RIPA buffer. Equalamounts of protein from each cell lysate (7.5mg) were separated bysodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were then transferred onto nitrocellulose filters andprobed with mouse monoclonal anti-phosphotyrosine antibody asdescribed in Materials and Methods. Molecular mass markers are onthe left. The arrows on the right indicate the main tyrosine-phospho-rylated bands.

J Am Soc Nephrol 11: 1016–1025, 2000 Thrombin and Proximal Tubular Cells 1021

damage, the mechanisms underlying their influx into the inter-stitial space are still largely undefined. The local release ofchemokines may represent the initial step in this event (19). Inthe growing chemokine family, MCP-1 represents the mostspecific and powerful chemotactic and activating factor formonocytes (41). We have recently demonstrated an increasedMCP-1 expression at the tubular level in acute transplantrejection that was significantly correlated with monocyte infil-tration (31). The infiltrating monocytes represent a majorsource of tissue factor, and thus powerful inducers of clottingcascade activation (42). The subsequent release of thrombinand its induction of MCP-1 production by PTC may stimulatea further influx and activation of circulating monocytes, clos-ing a positive feedback loop and amplifying the phenomenon.Indeed, in human renal allograft rejection, activated interstitialmacrophages are closely associated with fibrin deposits (43).Moreover, the hypothesis of a relevant role for thrombin in thesetting of acute dysfunction of the renal allograft is furthersupported by the recent observation that proximal tubular cellsexpress and produce anti-thrombin III, and the depletion oftubular anti-thrombin in the donor kidney is correlated with thedegree of allograft function at 3 d after transplantation (44).

The mechanisms underlying thrombin-induced cell activa-tion are still poorly understood. All of the known thrombinreceptors, including PAR-1, belong to the G protein receptorsuperfamily (14–16). This class of receptors signal inside thecells through the interaction with one or more heterotrimeric G

protein(s), leading to the activation of the phospholipaseC–protein kinase C pathway on one side and to the modulationof adenylcyclase on the other (12,13,34). Although signalingreceptors have always been rigidly divided in tyrosine kinaseand G protein-coupled receptors, cross-talk between these twosystems frequently occurs in rapidly induced cellular responses(34). In the past 10 yr, an increasing body of evidence sug-gested the activation of different tyrosine kinases in response tothrombin and the relevance of this phenomenon in thrombin-

Figure 8.Thrombin-induced c-src tyrosine phosphorylation in hPTC.Confluent quiescent hPTC were stimulated with thrombin (5 U/ml)for 5, 10, 15, 30, and 60 min and then lysed in RIPA buffer. Equalamounts of protein from each cell lysate (100mg) were immunopre-cipitated with anti-phosphotyrosine antibody, separated by SDS-PAGE, transferred onto nitrocellulose filters, and probed with rabbitpolyclonal anti-src antibody as described in Materials and Methods.Note that the anti-src antibody also recognizes two more tyrosine-phosphorylated bands, namely fyn and yes. Molecular markers are onthe left.

Figure 9. Effect of tyrosine kinase inhibition on thrombin-inducedDNA synthesis in hPTC. Confluent, quiescent PTC were pretreatedwith genistein (25mM), daidzein (25mM) (Panel A), and herbimycinA (Panel B) for 18 h and then stimulated with thrombin (5 U/ml).DNA synthesis was measured as the amount of3H-thymidine uptakeinto TCA-insoluble precipitate as described in Materials and Methods.*P , 0.01versuscontrol; **P , 0.01versusthrombin alone.

1022 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 1016–1025, 2000

induced cell activation (45–47). Interestingly, both cellularresponses described in the present study relied on the activationof the same signaling pathway involving protein-tyrosine phos-phorylation. Indeed, thrombin stimulated the tyrosine phos-phorylation of an array of cellular proteins in cultured hPTC.Although the precise identity of these phosphoproteins remainsto be determined, the 60-kD protein most likely represents oneof the cytoplasmic tyrosine kinases of the c-src family. Inplatelets, thrombin has been shown to activate different ty-rosine kinases of this family, including c-src, fyn, yes, and lyn(45). These observations were, at least partially, reproduced inother cell types (45). However, recently it has been demon-strated that this serine protease can cross-activate the EGFreceptor in several cell lines (36,47). PTC expressin vivo andin vitro both c-src and EGF receptor at high levels and both ofthese kinases play a key role in different physiologic processes(37,48). In the present study, it was demonstrated that c-src, butnot the EGF receptor, is strikingly autophosphorylated in re-sponse to thrombin in human PTC with a time course closelyresembling the one observed for protein-tyrosine phosphoryla-tion. This observation indirectly suggests the activation of c-srcin response to thrombin and its potential role in protein tyrosinephosphorylation induced by this protease.

The mechanisms of c-src activation upon thrombin stimula-tion and its role in thrombin-induced cellular effects are stillcontroversial. Indeed, both Chenet al. and Luttrell et al.demonstrated that the activation of this cytoplasmic tyrosinekinase is, at least partially, G protein-dependent (46,49). More-over, Luttrell et al. reported that src activation may link Gprotein-coupled receptors to the ras-mitogen-activated proteinkinase pathway, the main mitogenic pathway for thrombin(49). On the other hand, Kranenburget al. demonstrated thatsrc activation is G protein-independent and is not necessary forG protein-induced mitogen-activated protein kinase activation(50). In the present study, however, the mitogenic effect ofthrombin seems to be completely dependent on src activation,because the specific src inhibitor PP1 significantly inhibitedthrombin-induced DNA synthesis. On the other hand, we can-not definitively describe the role of src in thrombin-elicitedMCP-1 expression because both PP1 and its inactive analoguePP3 equally inhibited this thrombin effect (data not shown).

In conclusion, PTC expressin vivo and in vitro PAR-1 andrepresent a potential target for thrombin. This serine protease,interacting with PAR-1 and activating c-src, might represent aregenerative and proinflammatory stimulus for PTC in acuteand chronic tubulointerstitial damage.

AcknowledgmentsThis study was supported in part by a Baxter Extramural grant

(eight round, 1996–1998), the Associazione per il Progresso Scienti-fico in Nefrologia e Trapianto (APSNT), the Consiglio Nazionale

Figure 10.Effect of tyrosine kinase inhibition on thrombin-inducedMCP-1 gene expression in hPTC. Confluent, quiescent PTC werepretreated with genistein (25mM), daidzein (25mM), and herbimycinA (4 mM) for 18 h, stimulated with thrombin (5 U/ml) for 6 h, andthen harvested. MCP-1 expression was evaluated by Northern blotting(top panel). 28S and 18S ribosomal RNA bands on ethidium bromide-stained gels were used to control the RNA loading (bottom panel).Results are representative of three experiments.

Figure 11.Effect of src inhibition on thrombin-induced DNA synthe-sis in hPTC. Confluent, quiescent PTC were pretreated with thespecific src inhibitor PP1 (25mM) or its inactive analogue PP3 (25mM) for 18 h and then stimulated with thrombin (5 U/ml). DNAsynthesis was measured as the amount of3H-thymidine incorporationinto TCA-insoluble material as described in Materials and Methods.Data represent mean6 SD of four separate experiments each done intriplicate wells. cpm, counts per minute. *P , 0.01versusthrombinalone.

J Am Soc Nephrol 11: 1016–1025, 2000 Thrombin and Proximal Tubular Cells 1023

delle Ricerche (C.N.R., 98.513.04), the CNR target project on Bio-technology, the Ministero dell’Universita´ e della Ricerca Scientifica eTecnologica (60%, 97.7703), and by a grant from the Istituto Supe-riore di Sanita`. We are grateful to Dr. L. F. Brass (Thomas JeffersonUniversity, Philadelphia, PA) for providing the anti-PAR-1 antibody.

References1. Yamamoto K, Loskutoff DJ: The kidneys of mice with autoim-

mune disease acquire a hypofibrinolytic/procoagulant state thatcorrelates with the development of glomerulonephritis and tissuemicrothrombosis.Am J Pathol151: 725–734, 1997

2. Enestrom S, Druid H, Rammer L: Fibrin deposition in the kidneyin post-ischaemic renal damage.Br J Exp Pathol69: 387–394,1988

3. Faulk WP, Gargiulo P, McIntyre JA, Bang NU: Hemostasis andfibrinolysis in renal transplantation.Semin Thromb Hemostasis15: 88–98, 1989

4. Wang Y, Pratt JR, Tam FWK, Hartley RB, Wolff JA, OlavesenMG, Sacks SH: Upregulation of type 1 plasminogen activatorinhibitor mRNA with thrombotic changes in renal grafts.Trans-plantation61: 684–689, 1996

5. Wang Y, Pratt JR, Hartley B, Evans B, Zhang L, Sacks SH:Expression of tissue type plasminogen activator and type 1plasminogen activator inhibitor and persistent fibrin depositionin chronic renal allograft failure.Kidney Int52: 371–377, 1997

6. Wendt T, Zhang YM, Bierhaus A, Kriegsmann J, Deng Y,Waldherr R, Teske T, Luther T, Funfstuk R, Nawroth PP, SteinG: Tissue factor expression in an animal model of hydronephro-sis.Nephrol Dial Transplant10: 1820–1828, 1995

7. Kincaid-Smith P: Coagulation and renal disease.Kidney Int 2:183–190, 1972

8. Sraer JD, Kanfer A, Rondeau E, Lacave R: Glomerular hemo-stasis in normal and pathological conditions.Adv Nephrol17:27–56, 1988

9. Ono T, Kanatsu K, Doi T, Sekita K, Onoe C, Nagai H, Muso E,Yoshida H, Tamura T, Kawai C: Relationship of intraglomerularcoagulation and platelet aggregation to glomerular sclerosis.Nephron58: 429–436, 1991

10. Mann KG, Lundblad RL: Biochemistry of thrombin. In:Ho-meostasis and Thrombosis, edited by Colman R, Hirsh J, MarderV, Salzman E, Philadelphia, Lippincott, 1987, pp 148–161

11. Wilner GD, Danitz MP, Mudd MS, Hsieh K-H, Fenton JW:Selective immobilisation ofa-thrombin by surface-bound fibrin.J Lab Clin Med97: 403–411, 1980

12. Grandaliano G, Gesualdo L, Schena FP: Thrombin: A novelrenal growth factor.Exp Nephrol7: 20–25, 1999

13. Grand RJA, Turnell AS, Grabham PW: Cellular consequences ofthrombin receptor activation.Biochem J313: 353–368, 1996

14. Vu TK, Hung DT, Wheaton VI, Coughlin SR: Molecular cloningof a functional thrombin receptor reveals a novel proteolyticmechanism of receptor activation.Cell 64: 1057–1068, 1991

15. Ishihara H, Connolly AJ, Zeng D, Kahn ML, Zheng YW, Tim-mons C, Tram T, Coughlin SR: Protease-activated receptor 3 isa second thrombin receptor in humans.Nature 386: 502–506,1997

16. Xu WF, Andersen H, Whitmore TE, Presnell SR, Yee DP, ChingA, Gilbert T, Davie EW, Foster DC: Cloning and characteriza-tion of human protease-activated receptor 4.Proc Natl Acad SciUSA95: 6642–6646, 1998

17. Xu Y, Zacharias U, Peraldi MN, He CJ, Lu C, Sraer JD, BrassLF, Rondeau E: Constitutive expression and modulation of the

functional thrombin receptor in the human kidney.Am J Pathol146: 101–110, 1995

18. Wolf G: Cellular mechanisms of tubule hypertrophy and hyper-plasia in renal injury.Miner Electrolyte Metab21: 303–316,1995

19. Van Goor H, Ding G, Kees-Folts D, Grond J, Schreiner GF,Diamond JR: Macrophages and renal diseases.Lab Invest71:456–464, 1994

20. Grandaliano G, Gesualdo L, Ranieri E, Monno R, Montinaro V,Marra F, Schena FP: Monocyte chemotactic peptide-1 expressionin acute and chronic human nephritides: A pathogenetic role ininterstitial monocyte recruitment.J Am Soc Nephrol7: 906–913,1996

21. Ryan MJ, Johnson G, Kirk J, Furstenberg SM, Zager RA,Torok-Storb B: HK-2: An immortalized proximal tubule epi-thelial cell line from normal human kidney.Kidney Int 45:48 –57, 1994

22. Grandaliano G, Choudhury GG, Biswas P, Abboud HE: Mito-genic signaling of thrombin in mesangial cells: Role of tyrosinephosphorylation.Am J Physiol267: F528–F536, 1994

23. Gesualdo L, Di Paolo S, Ranieri E, Schena FP: Trapidil inhibitshuman mesangial cell proliferation: Effect on PDGFb-receptorbinding and expression.Kidney Int46: 1002–1009, 1994

24. Chomczynski P, Sacchi N: Single step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction.Anal Biochem162: 156–159, 1987

25. Grandaliano G, Valente AJ, Abboud HE: A novel biologic ac-tivity of thrombin: Stimulation of monocyte chemotactic pro-tein-1 production.J Exp Med179: 1737–1741, 1994

26. He CJ, Peraldi MN, Adida C, Rebibou JM, Meulders Q, SraerJD, Rondeau E: Thrombin signal transduction mechanisms inhuman glomerular epithelial cells.J Cell Physiol150: 475–483,1992

27. Grandaliano G, Poptic E, Barnes JL, Choudhury GG, WoodruffK, Abboud HE: Thrombin stimulates platelet-derived growthfactor production in cultured bovine glomerular endothelial cells.J Am Soc Nephrol9: 583–589, 1998

28. Hoxie JA, Ahuja M, Belmonte E, Pizarro S, Parton R, Brass LF:Internalization and recycling of activated thrombin receptors.J Biol Chem268: 13756–13763, 1993

29. Zacharias U, Xu Y, Haege J, Sraer JD, Brass LF, Rondeau E:Thrombin, phorbol ester, and cAMP regulate thrombin receptorprotein and mRNA expression by different pathways.J BiolChem270: 545–550, 1995

30. Shultz PJ, Knauss TC, Mene´ P, Abboud HE: Mitogenic signalsfor thrombin in mesangial cells: Regulation of phospholipase Cand PDGF genes.Am J Physiol257: F366–F374, 1989

31. Grandaliano G, Gesualdo L, Ranieri E, Monno R, Stallone G,Schena FP: Monocyte chemotactic peptide-1 expression andmonocyte infiltration in acute renal transplant rejection.Trans-plantation63: 414–420, 1997

32. Wenzel UO, Fouqueray B, Grandaliano G, Karamitsos C, KimYS, Valente AJ, Abboud HE: Thrombin regulates expression ofmonocyte chemoattractant protein-1 in vascular smooth musclecells.Circ Res77: 503–509, 1995

33. Colotta F, Sciacca FL, Sironi M, Luini W, Rabiet MJ, MantovaniA: Expression of monocyte chemotactic protein-1 by monocytesand endothelial cells exposed to thrombin.Am J Pathol144:975–985, 1994

34. Pouyssegur J, Seuwen K: Transmembrane receptors and intra-

1024 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 1016–1025, 2000

cellular pathways that control cell proliferation.Annu RevPhysiol54: 195–210, 1992

35. Wong S, Reynolds AB, Papkoff J: Platelet activation leads toincreased c-src kinase activity and association of c-src with an85-kDa tyrosine phosphoprotein.Oncogene7: 2407–2415, 1992

36. Daub H, Wallasch C, Lankenau A, Herrlich A, Ullrich A: Signalcharacteristics of G protein-transactivated EGF receptor.EMBOJ 16: 7032–7044, 1997

37. Gesualdo L, Di Paolo S, Calabro A, Milani S, Maiorano E,Ranieri E, Pannarale G, Schena FP: Expression of epidermalgrowth factor and its receptor in normal and diseased humankidney: An immunohistochemical and in situ hybridizationstudy.Kidney Int49: 656–665, 1996

38. Losonczy G: Intrarenal blood coagulation induced by ischemia inrats: Heparin and thrombocytopenia does not prevent the intra-cortical fibrin formation.Thromb Res34: 87–92, 1984

39. Platt JL, Lebien TW, Michael AF: Interstitial mononuclear cellpopulations in renal graft rejection: Identification by monoclonalantibodies in tissue sections.J Exp Med155: 17–22, 1982

40. Eddy AA: Interstitial macrophages as mediators of renal fibrosis.Exp Nephrol3: 76–79, 1995

41. Leonard EJ, Yoshimura T: Human monocyte chemoattractantprotein-1.Immunol Today11: 97–101, 1993

42. Holdsworth SR, Tipping PG: Macrophage dependent glomerularfibrin deposition in experimental glomerulonephritis in rabbit.J Clin Invest76: 1367–1374, 1985

43. Hancock WW, Gee D, De Moerloose P, Rickles FR, Ewan VA,Atkins RC: Immunohistological analysis of serial biopsies takenduring human renal allograft rejection: Changing profile of in-

filtrating cells and activation of the coagulation system.Trans-plantation39: 430–438, 1985

44. Torry RJ, Labarrere CA, Nelson D, Carter C, Haag B, Faulk WP:Tubular anti-thrombin at transplantation determines subsequentrenal allograft function.Transplantation66: 797–799, 1998

45. Cichowsky K, McCormick F, Brugge JS: p21ras GAP associa-tion with fyn, lyn and yes in thrombin-activated platelets.J BiolCell 267: 5025–5028, 1992

46. Chen YH, Pouyssegur J, Courtneidge SA, Van Obberghen-Schilling E: Activation of src family kinase activity by the Gprotein-coupled thrombin receptor in growth responsive fibro-blasts.J Biol Chem269: 27372–27377, 1994

47. Vaingankar SM, Martins-Green M: Thrombin activation of the9E3/CEF4 chemokine involves tyrosine kinases including c-srcand the epidermal growth factor receptor.J Biol Chem273:5226–5234, 1998

48. Tsuganewa H, Preising PA, Alpers RJ: Dominant negative c-srcinhibits angiotensin II induced activation of NHE3 in OKP cells.Kidney Int54: 394–398, 1998

49. Luttrell LM, Hawes BE, van Biesen T, Luttrel DK, Lansing TJ,Lefkowitz RJ: Role of c-src tyrosine kinase in G protein-coupledreceptor and G beta gamma subunit-mediated activation of mi-togen-activated protein kinases.J Biol Chem271: 19443–19450,1996

50. Kranenburg O, Verlaan I, Hordjik PL, Moolenaar WH: Gi-mediated activation of the ras/MAP kinase pathway involves a100 kDa tyrosine-phosphorylated Grb2 SH3 binding protein, butnot Src nor Shc.EMBO J16: 3097–3105, 1997

J Am Soc Nephrol 11: 1016–1025, 2000 Thrombin and Proximal Tubular Cells 1025