JOURNAL OF CELLULAR PHYSIOLOGY 146:131-140 (1991,

Thrombin Increases Proliferation and Decreases Fibrinolytic Activity of Kidney Glomerular Epithelial Cells CI-JIANCHE,* ERIC RONDEAU, ROBERT 1. MEDCALF, ROGER LACAVE, WOLF-DIETER SCHLEUNINC, AND JEAN-DANIELSRAER lnserm Unit6 64, H6pital Tenon, 75020 Paris, France (C.-I.H., E.R., R.I., l.-D.S.1, Centre Hospitalier Universitaire Vaudois, 10 1 7 lausanne, Switzerland (R.L.M.), and Schering AG, Biochemistry Institute, D- 1000 Berlin 65, Germany (W.-D.S.j Human glomerular epithelial cells (GECs) in culture synthesize single-chain, urokinase-type plasminogen activator (SC-uPA), tissue-type plasminogen activator (t-PA), and plasminogen activator inhibitor 1 (PAI-1) and possess specific membrane-binding sites for U-PA. Using purified '251-alpha thrombin, we demonstrate here the presence of two populations of specific binding sites for thrombin on GECs (1.Kd = 4.3 5 1.0 x lo-'' M, 5.4 2 1.4 x l o 4 M sites per cell, 2. Kd = 1.6 5 0.5 X l o p 8 M, 7.9 2 1.8 x 10' sites per cell). Purified human alpha thrombin promoted the proliferation of GECs and induced a timeand dose-dependent increase of SC-uPA, t-PA, and PAL1 antigens released by GECs. Thrombin-mediated increase in antigen was paralleled by an increase in the levels of corresponding U-PAand PAL1 messenger RNA. In contrast, thrombin decreased u-PA activity in conditioned medium. This discrepancy between u-PA antigen and U-PA activity was explained by a limited proteolysis of SC-uPA by thrombin, leading to a two-chain form detected by immunoblotting and that could not be activated by plasmin. Thrombin also decreased the number of U-PA binding sites on GECs (p < 0.05) without changing receptor affinity. Hirudin inhibited the binding and the cellular effects of thrombin, whereas thrombin inactivated by diisopropylfluorophosphate had no effect, indicating that both membrane binding and catalytic activity of thrombin were required. We conclude that thrombin, through specific membrane receptors, stimulates proliferation of GECs and decreases the fibrinolytic activity of GECs both at the cell surface and in the conditioned medium. These results suggest that thrombin could be involved in the pathogenesis of extracapillary proliferation and persistency of fibrin deposits in crescentic glomerulonephritis.

Fibrinolysis is a process mediated by plasmin, a serine protease, which is generated from plasminogen by two distinct plasminogen activators (PA): urokinase-type plasminogen activator (u-PA) and tissuetype plasminogen activator (t-PA). Specific inhibitors, plasminogen activator inhibitor 1 (PAI-1) and 2 (PAI2), further regulate PA activity (for review see Collen and Lijnen, 1986). Recently we demonstrated that glomerular epithelial cells (GECs)of the human kidney release t-PA, PAI-1, and u-PA, which is released as inactive single-chain proenzyme (SC-uPA). We also demonstrated that GECs possess membrane-binding sites for SC-uPA and u-PA that are partially autosaturated by SC-uPA released by GECs (Rondeau et al., 1989). In some pathological conditions GECs proliferate and fibrin deposition is observed in kidney glomeruli, with subsequent infiltration by inflammatory cells (Sraer et al., 1988). The pathophysiology of these phenomena remains unelucidated. Fibrin formation results from local thrombin generation and thrombin is also known to interfere with fibrinolysis. It can induce a limited 8 1991 WILEY-LISS, INC.

proteolysis of single-chain urokinase (SC-uPA), which then cannot be activated to u-PA (Ichinose et al., 1986). Furthermore, thrombin can also inactivate PAI-1 (Knudsen, 1987) and has been shown to enhance the synthesis of both t-PA and PAI-1 in human endothelial cells (Levin et al., 1984; Gelehrter and Szycer-Laszuk, 1986; Dichek and Quetermous, 1989). However, the resulting effect on the balance between plasminogen activator and plasminogen activator inhibitor activities was not investigated. Thrombin has also been shown to increase DNA synthesis and proliferation of various cell types (Gospodarowicz et al., 1978; Zetter and Antoniades, 1979; Gordon and Carney, 1986; Carney et al., 1985).These cellular effects of thrombin were shown to be mediated through a complex interaction with high affinity binding sites on the cell membrane, which re uired both binding capacity and catalytic activity ( arney et al., 1986). Here, we present evi-

%

Received April 19, 1990; accepted October 4, 1990. 'To whom reprint requestdcorrespondence should be addressed.

HE ET AL.

132

dence for specific binding sites for thrombin on GECs and the effects of thrombin on proliferation and fibrinolytic activity of GECs. MATERIALS AND METHODS Materials The following materials were obtained collagenase type IV, alpha amanitin, cycloheximide, hirudin, and Triton X 100 (Sigma, St. Louis); bovine thrombin (Hoffman-LaRoche, Basel, Switzerland); low melting point agarose (BRL); polyvinyl chloride U microtiter plates (Dynatech Laboratories, Alexandria, VA); purified human fibrinogen, plasminogen, plasmin, and a synthetic substrate D-Val-Leu-Lys-pNA 6 2 2 5 1 ) (AB Kabi, Stockholm, Sweden); human urokinase reference standard grade from Choay (Paris, France); human t-PA, rabbit polyclonal antihuman t-PA, and antihuman u-PA antibodies (Biopool, Sweden); Iodogen, 1,3,4,6tetrachloro 3 6 diphen$glycoluril (Pierce Chemical Co; Rockford, IL); alpha 2P dCTP > 300 Ciimmol (Amersham, UK); diisopropyl fluorophosphate, Fluka, AG (Chem Fabrik Buchs); and polyclonal and monoclonal (MA1 12) anti-PAI-1 antibodies (Biopool). Purified human alpha thrombin was a generous gift from Dr. Freyssinet (3,000 U/mg, lU/ml= lo-’ M, Centre Departemental de Transfusion Sanguine, Strasbourg, France). DFP-inactivated thrombin had less than 1%of the procoagulant activity of native thrombin as measured by the coagulation time of a 2 mg/ml fibrinogen solution. Cell Cultures The glomerular epithelial cells were cultured from normal human kidneys (n = 6)judged to be unsuitable for transplantation. Glomerular epithelial cells were obtained by colla enase digestion of the isolated glomeruli, as previous y described with slight modifications (Rondeau et al., 1989). Briefly, glomeruli were isolated from renal cortex by differential sieving through a 180-pm sieve to exclude vessels and tubules and were then washed over a 105-pm sieve that allowed cellular debris to pass through but retained glomeruli. Isolated glomeruli were then suspended in Waymouth’s medium (Flow Laboratories, Irvine, UK) containing 500 U/ml collagenase for 30 min at 37°C. The suspension was passed over a 25-pm sieve that retained partially digested glomeruli. Epithelial cells in the filtrate were cultured at 37°C in Waymouths medium containing 20% fetal calf serum in a humidified atmosphere of 5% C02 in air. They were purified using cloning rings in order to isolate groups of homogenous cells and identified by their morphologic features as previously reported (Stricker et al., 1980; Rondeau et al., 1989). Epithelial cells are polyhedral at confluency. By immunofluorescence, the epithelial origin of these cells was confirmed by positive staining with anticytokeratin antibodies. All the cells were positively stained by antiurokinase antibodies but not by antimyosin or antifactor VIII-related antigen antibodies, excluding mesangial and endothelial contamination. These cells were also positively stained with an anti-CALLA monoclonal antibody (I0 T5 from Immunotechnology),

P

which binds to podocytes and to brush border of normal human kidney. Finally, these cells were stained by a monoclonal antihuman GECs antibody (PHM5 from Artarmon, Australia). By electron microscopy, GECs grown on plastic appeared flat with rare villi. WeibelPalade bodies were not observed. Incubation procedures Incubations were performed on the third subculture of glomerular epithelial cells. For the third passage the cells were grown in multidish wells (Nunclon, Nunc, Roskilde, Denmark) in hormonally defined medium (DMEM 50%, HAM F12 50%, HEPES 10 mM, glutamine 2 mM, insulin 5 pg/ml, dexamethasone 5.10-’ M, transferrine 5 p,g/ml, selenium 3.1OP8M, penicillin 10 pg/ml, streptomycine 100 pg/ml) with 1%fetal calf serum. Before cells reached confluence, each well was washed three times with serum-free defined medium to remove serum-derived plasminogen activators and plasminogen activator inhibitors and incubated 24 h in minimum defined medium (DMEM 50%, HAM F12 50%,HEPES 10 mM, glutamine 2 mM). Adherent cells were then incubated with the compound to be tested or with vehicle alone at 37°C in 1ml of minimum defined medium in 5% C02 atmosphere. At the end of incubation, the culture media were collected separately for determination of plasminogen activator activity. Cells were counted after detachment with trypsin. Care was taken that the viability of the cells did not differ between experimental and control conditions: in all cases more than 95% of the cells excluded trypan blue dye. Levels of cell-associated SCu-PA, t-PA, and PAI-1 were determined in cellular lysates after disruption in 1%Triton X 100. Thrombin receptor binding and urokinase receptor binding assays Purified human urokinase and human alpha thrombin were iodinated with 1251 by the Iodogen method as previously reported (Rondeau et al., 1989). Iodinated proteins were separated by chromatography on a Sephadex G50 column. Elution fractions (100 p.1) were then analyzed by SDS-PAGE and autoradio aphy. The fraction with the most active 53 Kd l2PI u-PA signal was used from the binding assay, corresponding approximately to 5.107 cpm per picomoles of u-PA. Similarly 1251thrombin was prepared for binding assay, corresponding to lo5 cpm per picomoles of alpha thrombin. Binding experiments for u-PA receptors were performed as described (Stoppelli, 1986). Adherent glomerular epithelial cells (5.104/well) were washed with Waymouth‘s medium containing 1mg/ml bovine serum albumin (BSA) and then incubated at 4°C or 22°C with lZ5I-u-PAin Waymouth’s medium containing 1 mg/ml BSA (approximately 1to 2.105 cpm/well corresponding to 4-8 10-l’ M labeled ligand). At the end of a 30-min incubation, the supernatant was removed, cells were washed three times with 1 mg/ml BSA in PBS buffer and cells were dissolved in 1 M NaOH and radioactivity counted in a gamma ray counter. Each experiment was carried out in duplicate. Specific binding was calculated by subtracting the

133

THROMBIN AND FIBRINOLYSIS BY GLOMERULAR EPITHELIAL CELLS

counts bound in the presence of 3.10'6 M unlabeled u-PA. In order to unmask u-PA receptors, the cells were treated for 3 rnin with 50 mM glycine-HC1buffer, pH 3, containing 0.1 M NaC1, then quickly neutralized with 50 mM Hepes buffer (pH 7.5), 0.1 M NaCl (21). Acidtreated cells were then subjected to the binding assay in the absence or presence of competitor. Binding experiments to thrombin receptors were carried out at 4°C or at 22°C. Iodinated thrombin (2 to 3 x lo5 cpdwell) corresponding to 2.lO-' M labeled ligand was added in Waymouth's medium containing 1mg/ml BSA. Specific binding was calculated by subtracting the counts bound in the presence of M unlabeled thrombin. In the absence of cells, the binding of radiolabelled ligand was less than 100 cpdwell both for u-PA and thrombin. Measurement of 3Hthymidine incorporation and proliferation of GECs After a 24-h incubation in minimum defined medium, resting cells were challenged with the compounds to be tested or their vehicle for 24 to 48 h. Six h before cell count, 1 $X3H thymidine was added to each well. At the end of incubation, cells were washed, incubated 30 min in minimum defined medium containing 1 mglml nonlabeled thymidine, and then harvested by trypsinization. Cells were filtered on 0.45 p,m Millipore filter, and the radioactivity of the filter was counted in a p-scintillation counter. Aliquots of de-

tached cells were counted in Malassez plates and control of cell viability assessed by Trypan blue dye exclusion. Each experimental condition was made in triplicate. Electrophoretic procedures and zymography SDS-polyacrylamide gel electrophoresis (SDSPAGE) was carried out according to Laemmli (1970). The stacking and separating gels contained 4% and 10% acrylamide, respectively. Marker proteins (Pharmacia, Uppsala, Sweden) were run in parallel for molecular weight determination by staining the gels with Coomassie Blue dye. Commercial u-PA and t-PA were also run in parallel with each gel. Zymographic procedures were carried out as described by Loskutoff et al. (1983). Plasminogen activator activities of conditioned culture medium were detected on a fibrin-agar underlay, which contained 10 m /ml fibrinogen, 300 mU/ml thrombin, 1%agarose, an 20 kg/ml plasminogen. Incubation was performed at 3 7 T in humidified atmosphere. Immunoblotting analysis After electrophoresis of fivefold concentrated culture medium conditioned by unstimulated or thrombin stimulated cells, transfer on a nitrocellulose filter was performed as previously described (Lacave et al., 1989) in a 25 mM Tris-HC1 buffer pH 8.3 containing 20%

%

p I -x-*

0 .

5

. . 10 15

.. I

30

'A 11-1

45

'

60

b

Time, minutes Fig. 1. A. Binding of '251-thrombin to human glomerular epithelial cells. Iodinated thrombin (2 to 3 x lo5 cpdwell) was added to GECs in Waymouth's medium containing 1 mg/ml BSA. Binding of lZ5Ithrombin was measured by counting the cell-associated radioactivity after dissolution of cellular proteins in 1M NaOH. Specific binding ( 0 ) was calculated by subtracting the counts bound in the presence of M unlabeled thrombin. Maximal binding was reached after 15 min at 22°C. Addition of 1Uiml hirudin prevents almost completely thrombin binding even after 60 min of incubation ( X ). The binding of thrombin M unlabeled thrombin (A). was reversible after addition of Addition of u-PA (W or t-PA ( 0 )at similar concentrations had no effect.

tY

r;

L)

10

9

8

7

Unlabeled T H R , - l o g M At P C , maximal binding was slightly lower and reached also a plateau after 15 min (0). B. Competitive inhibition of 1251-thrombin binding to human glomerular epithelial cells by unlabeled thrombin. Iodinated thrombin (2 to 3 x lo5 cpdwell) was added to GECs in Waymouths medium containing 1 mg/ml BSA. Unlabeled thrombin was added simultaneously at various concentrations (0 to W 7 M ) . After 15 min of incubation at 22"C, cells were washed and cellassociated radioactivity was measured after dissolution of cellular proteins in 1M NaOH. Scatchard plot analysis is shown (inset).Each point represents the mean of three separate experiments. The intraassay variation was less than 10%.

HE ET AL.

134 A

6

9

0 Thrombin

( x10F8M)

*

1 2 3 Time, days

Fig. 2. Effect of thrombin on DNA synthesis and proliferation of glomerular epithelial cells. A. Dose-response effect of thrombin on H-thymidine incorporation by GECs. Mean SEM of 3 experiments made in duplicate are presented. Half-maximal effect was obtained at 10' M thrombin. * p < 0.05. B. In some experiments cells were

counted after 1, 2, or 3 days of addition of 1%FCS ( 0 ) or 10.' M thrombin ( A ) . No cell growth was observed in serum and hormonefree medium (). Each point represents the mean of two separate experiments. * p < 0.05.

methanol, 150 mM glycine, and 0.07% SDS. The filter was then colored by Red Ponceau t o identify marker proteins, washed, and incubated twice for 30 min in phosphate buffer saline pH 7.4 (PBS) containing 1.75% gelatin. Then it was incubated with polyclonal goat anti-u-PA antibody (20 pgiml) in PBS Tween 0.5% for 4 h at 37°C and 16 h at 4°C. After three washings in PBS Tween 0.5%, the filter was incubated with biotinylated mouse antigoat Ig (Vector Laboratories, Burlingame, CAI for 4 h at 37°C. After washing, the filter was incubated with streptavidine-peroxydase(dilution 1/200, Amersham, UK) in PBS-Tween 0.1% for 2 h. Finally after washing, peroxydase was revealed by addition of 1.5 ml chloronaphtol(6 mg/ml in methanol), 45ml Tris HC10.01 M pH 7.4, and 45 p.1 H,02 30%. Enzyme linked immunosorbent assay (ELISA) for t-PA, u-PA, and PAI-1 Determination of u-PA, t-PA, and PAI-1 antigens were performed by ELISA as previously described (Lacave, 1989). ELISA for PAI-1 was slightly modified microtiters polyvinylchloride plates were coated with specific goat polyclonal antihuman PAI-1 IgG (Biopool, Sweden) and free and complexed PAI-1 were detected with a monoclonal anti-PAI-1 antibody (250 ng/ml), which binds to both free PAI-1 and complexed PAI-1 (MA1 12, Biopool, Sweden). The assay was then continued as described previously (Lacave, 1989). The detection limits for PAI-1, u-PA, and t-PA were 0.4 ng/ml, 0.9 ng/ml, and 0.14 ng/ml, respectively. Measurement of plasminogen activator activity As previously shown (Rondeau, 1989), the PA activity of the culture medium conditioned by GEC is due to u-PA, whereas few t-PA are present in an inactive form complexed to PAT-1 (see Fig. 6). Thus PA activity was measured by determination of the u-PA activity. Samples were centrifuged (1,500 g x 5 min at 4°C) to remove cellular debris. The supernatants were collected and Triton X 100 was added (final concentration 0.1%). Urokinase activity was determined as follows: the paranitroanilide liberated from the plasmin sensi-

TABLE 1. Effect of growth factors and thrombin on 3H-thymidine incorporation in G E C ~

*

3H-thymidine count/5.104 cells Basal FCS (1%) Thrombin (lo@ M) Thrombin (lo@ M) FCS (1%) Insulin (2.5 rg/ml) Thrombin (lo@ M) insulin (2.5 rg/ml)

+

+

984 t 186 2,118 f 405* 1,856 k 355* 3,652 i 581** 995 5 216 3,184 f 522**

'Glomerular epithelial cells were incubated for 24 h in the presence or absence of growth factors and/or thrombin. 3H-thymidine (1 pCi) was added 6 h before the end ofincubation. At the endofincubation, cell-associatedradioactivity and thenumber of cells were determined. Mean SEM of 3 experiments made in duplicate are presented. *p < 0.01; **p < 0.001 as compared to control conditions.

+

tive synthetic chromogenic substrate D-Val-Leu-LyspNA 2 HCl(S-2251, AB Kabi, Stockholm, Sweden) was spectrophotometrically measured in microtiter polyvinylchloride plates. To detect total u-PA activity, supernatants were incubated for 2 h at 37'C with plasmin (final concentration 100 ng/ml), which transforms the inactive pro-uPA into the active form. In these cases standard curves were constructed using commercial u-PA, which had also been preincubated with plasmin (100 nglml). The samples were tested in a final volume of 100 p.1 containing 1.2 kg plasminogen and 0.5 mM S2251. The absorbance at 405 nM was measured with a micro ELISA counter after incubation at 37°C. For each experiment, the absorbance of fresh Waymouths medium containing 0.1%Triton X 100,0.5 mM S2251, and 1.2 pg plasminogen, plus the compound to be tested was determined and subtracted from the corresponding test value. These compounds had no direct effect on the apparent enzymatic activity of exogenous urokinase measured by the chromogenic assay. Northern blot analysis After incubation of glomerular epithelial cells, total RNA was extracted by the phenol-chloroform method

135

THROMBIN AND FIBRINOLYSIS BY GLOMERULAR EPITHELIAL CELLS TABLE 2. Effect of hirudin and DFP-thrombin on cell proliferation and synthesis of fibrinolytic components by GECs' 3H thymidine incorporation (cpm/5.104 cells)

Culture treatment Control Thrombin (1 U/ml) Hirudin (1 U/ml) Hirudin (1 U/ml) + thrombin (1 U/ml) DFP-thrombin (1U/ml)

Fibnnolytic components of GECs u-PA (ng/ml) t-PA (ng/ml) PAI-1 (ng/ml)

*

1036 k 150 1991 i 270' 1205 230 1139 i 115''

40.5 f 8.0 105.5 f 8.6* 39.3 i 2.8 41.3 f 7.6**

2.00 f 0.30 5.71 f 1.20' 1.92 i 0.34 2.13 k 0.45**

120 12 375 f 21' I25 11 134 12**

* 137"

39.8 f 9.6**

2.08 f 0.55**

122 f Il**

+

1057

*

+

'Mean i SEM of at least 3 different experiments made in triplicate are given

*p < 0.01 as compared to control; **p < 0.01 as compared to thrombin alone.

B / F x 0.001

t

1

5

10

Thrombin ( x l o - 8 M ) Fig. 3. Effect of thrombin on 1251-urokinasebinding on glomerular epithelial cells and competitive inhibition of lZ5I-u-PAbinding. A. Dose-response effect of thrombin on the binding lZ5I-u-PAon GECs. Specific binding is shown ( A ) after subtraction of total count with bound radioactivity in the presence of 3 . W M unlabeled u-PA. After a 24 h incubation with thrombin, a dose-dependent decrease of u-PA binding on GECs was demonstrated, with a half maximal inhibition induced by M thrombin. Addition of hirudin during the incubation prevented the decrease of u-PA binding induced by thrombin (A).B. The cells were incubated in the presence of 1251-~-PA (4.10-l' M) and

'% i r8

7

6

5

-

Unlabeled u-PA,--bg M increasing amounts of unlabeled u-PA (from 0 to 3 . W 6 MI. After 30 min, '''1-u-PA binding was measured by counting the total cellassociated radioactivity after dissolution of cellular proteins in 1 M NaOH. At the higher unlabeled u-PA concentration, nonspecific binding of lZ5I-u-PAwas reached, corresponding to 200-300 cpmil0" cells. Binding experiments were performed on unstimulated cells ( 0 ) and on cells pretreated by lo-@M thrombin for 24 h (0). Scatchard plot analysis is shown (inset), indicating that thrombin decreased the number of u-PA receptors without changing affinity.

as described (Genton et al., 1987) and precipitated by Li (sodium salt citrate, 20 x SSC = 175,3 g NaC1, 88,2 g C1 3 M. Total RNA (10 pgitrack) was separated by sodium citrate for 1liter HzO, pH 7.0) containing 0.1% electrophoresis in 0.9% agarose gel containing 20% formaldehyde and transferred to a nylon membrane (Gene Screen Plus). The membrane was prehybridized for 3 h at 42°C in a hybridization buffer containing 50% formamide, 600 mM NaC1,0.5%SDS, 50 mM Tris-HC1, pH 7.4, 5 mM EDTA, l x Denhart's (100 mg Ficoll, 50 mg polyvinylpyrolidone,50 mg BSA in 10 ml H20) and 100 Fgiml salmon sperm DNA, and then hybridized at 42°C for 16 h in the same buffer containing an alpha 32P labeled human cDNA probe specific for u-PA or PAI-1 as described (Medcalf et al., 1988). The membrane was then washed three times for 30 min at 42°C in 2 x SSC

SDS and used for autoradiography. Filters were exposed to Fuji RX X-ray films at -80°C using intensifying screens for 72 to 96 h. Autoradiographs were quantitated with a LKB 2202 Ultrascan Laser Densitometer .

Statistical analysis Results of antigen or activity measurements are expressed in ng or units per 5.104cells and presented as mean ? SEM. Student t-test and two-way variance analysis were used when appropriate.

HE ET AL. TABLE 3. Effect of cycloheximide and alpha-amanitin on thrombin-induced fibrinolytic components svnthesis of GECs' Culture treatment 20

-

u-PA (nglml)

Control Thrombin (1 U/ml) Cycloheximide (0.1 pg/ml) Cycloheximide + thrombin alpha-amanitin (2.5 pg/ml) alpha-amanitin thrombin

~ . = ~. . .. .- ... . ... . , ,~

+

1. '.a

,.

. . .... .. ... . .

~~.~ . . . . .* ............ 1

2 3 4 T H R . U/ml

*+ 0.41 0.82*

* *

33.5 4.2 61.7 f 8.2* 28.2 5.2

1.28 2.88 0.82

31.4 ?C 4.8**

0.86 k 0.25**

195 rt 24**

35.2 f 5.4

1.12 f 0.38

192

* 31

1.23 f 0.42**

200

+ 34**

*

30.4

+ 3.2**

+ 0.22

184 22 450 k 33* 180 21

'Cycloheximide or alpha-amanitin were simultaneously added to GECs with thrombin, and cells wereincubated for 24 h beforeculture medium werecollected for u-PA. &PA, and PAI-1 antigen measurements. Mean f SEM of 3 separate experi. ments are given. *p < 0.001 as compared to control; **p < 0.001 as compared to thrombin alone.

~

100' 0

+

Fibrinolytic components bf G E C ~ t-PA (ng/ml) PAI-1 (ng/ml)

~

5

Time, hours

Fie. 4 Effect of thrombin on the release of fibrinolvtic comoonents byhuman glomerular epithelial cells. Left panel. Dose-responseeffect of thrombin during a 24-h incubation. GECs were incubated in the M)for 24 h presence of various concentrations of thrombin (0to 5 and the concentrations of u-PA, t-PA, and PAI-1 antigens in the conditioned culture medium were measured (A).Mean 4 SEM of 4 different experiments are presented. The cell-associated amounts of these antigens were low ( A )and not modified by thrombin. Right panel. Time-course of the accumulation of u-PA, t-PA, and PAL1 in conditioned medium of GECs treated with 10.' M thrombin (W. As compared to unstimulated cells 0, thrombin-stimulated cells produced significantly more u-PA, t-PA, and PAI-1 after 12 h of incubation (* p < 0.05).

RESULTS Evidence for thrombin binding sites on cultured GECs As shown in Figure 1, specific and saturable binding of labeled thrombin on GECs could be demonstrated. At 22°C a plateau was reached after 15 min of incubation. Unlabeled thrombin M) was able to displace bound labeled thrombin, whereas albumin, plasminogen, uPA, or t-PA at similar concentrations did not. Competitive inhibition of binding and Scatchard plot analysis (Fig. 1 inset) showed two different types of receptors: a high affinity type with a Kd = 4.3 k 1.0 10-l' M and 5.4 t 1.4 lo4 sites per cell and a low affinity type with a Kd = 1.6 t 0.5 M and 7.9 L 1.8 lo5 sites per cell (mean t SEM, n = 3). Hirudin (1 U/ml) prevented completely the binding of labeled thrombin (Fig. 1). Effect of thrombin on DNA synthesis and proliferation of GECs Thrombin increased DNA synthesis in GECs as estimated by 3H thymidine incorporation (Fig. 2). This effect was observed in the absence of cell division (5.4 2 0.5 lo4 versus 5.0 t 0.8 lo4 cells per well) in the first 24 h of incubation. At 48 h, both 3H-thymidine

incorporation and cell number were significantly increased by M thrombin demonstrating that thrombin alone is able to promote mitogenesis in GECs and that the doubling time of GECs in culture is greater than 24 h. Thrombin stimulated in a dose-dependent manner 3H thymidine incorporation by GECs (Fig. 2). Half maximal effect was observed at lo-' M thrombin. Fetal calf serum (1%) also increased DNA synthesis and had additive effect with M thrombin. It has been controlled that thrombin preincubated with 1% FCS had the same coagulant activity on a 2 mg/ml fibrinogen solution. Recombinant human insulin (2.5 pg/ml) alone induced only a small and nonsignificant increase of DNA synthesis in GECs but potentiated the effect of thrombin (Table 1). As shown in Table 2, binding capacity and catalytic activity of thrombin were required to induce cell growth since addition of 1 U/ml hirudin almost completely inhibited thrombin-stimulated 3H thymidine incorporation and since DFP-inactivated thrombin had no effect. Effect of thrombin on u-PA binding sites of GECs Thrombin induced a dose-dependent decrease of binding of labeled u-PA on GECs. As shown in Figure 3A, half maximal inhibition of u-PA binding was observed when cells were incubated in the presence of M thrombin for 24 h. By competitive inhibition of labeled u-PA bindin we found that treatment of GECs with lo-' M thromtin for 24 h induced a significant decrease of the number of u-PA binding sites (1.2 t 0.3 lo4 after thrombin stimulation versus 3.4 k 0.4 lo4 sites per cell in basal conditions, n = 4,g < 0.05)without significant change of Kd (1.2 2 0.5 10- afler thrombin stimulation M in control conditions). Thromversus 1.3 * 0.4 bin incubated in the presence of hirudin had no effect on u-PA binding (Fig. 3); DEP-inactivated thrombin was also ineffective (not shown). Effects of thrombin on components of the fibrinolytic system in GECs Conditioned medium of GECs treated with thrombin contained increased levels of u-PA, t-PA, and PAI-1.

137

THROMBIN AND FIBRINOLYSIS BY GLOMERULAR EPITHELIAL CELLS

Fig. 5. Northern blot analysis of u-PA and PAL1 mRNA of glomerular epithelial cells after thrombin stimulation. GECs were incubated for 1, 2, 4, and 24 h in the control conditions (lanes 1 to 4) or after thrombin 10.' M exposure (lanes 5 to 8 ) and then total RNA were extracted. Ten kg of total RNA were run in agarose gel and trans-

The thrombin-mediated increases were time -and dosedependent (Fig. 4). Significant effects were observed after 12 h of thrombin treatment and a two to threefold increase of these antigens were observed after 24 h of incubation. A plateau was reached at lo-*M thrombin. Low levels of these antigens were cell-associated and not significantly modified by thrombin addition. As shown in Table 2, increases of u-PA, t-PA, and PAI-1 stimulated by thrombin were inhibited by addition of hirudin. DFP-inactivated thrombin had no effect. The effect of thrombin on u-PA, t-PA, and PAI-1 productions was significantly blocked by 0.1 pgiml cycloheximide and by 2.5 pg/ml alpha amanitin demonstrating that both de novo protein and RNA synthesis were required (Table 3). By Northern blot analysis, we demonstrated a rapid and transient increase of PAI-1 mRNA levels, which was markedly increased during the two first h of exposure, and then declined until 24 h (Figs. 5,6). Two bands of PAI-1 mFtNA (2.2 and 3.4 Kb) are usually observed in endothelial cells or in mesangial cells (not shown).In glomerular epithelial cells, only one band (3.4Kb) was found in unstimulated conditions, whereas the 2.2 Kb band appeared after thrombin stimulation. The u-PA mRNA level was increased between 2 and 24 h of incubation with lo-' M thrombin (Figs. 5, 6). The resulting: effect of thrombin on the fibrinolstic activity of conaitioned medium is shown in Figurk 7. The u-PA activity in the conditioned medium of unstimulated cells measured after preactivation of SCuPA by plasmin progressively increased with the time of incubation (Fig. 7). Conversely, when cells were incubated with 10.' M thrombin, the u-PA activity of the conditioned medium did not increase and even tended to decrease. This decrease of uPA activity induced by thrombin was dose dependent and half

ferred to a nylon filter. The filter was probed with an u-PA 32PcDNA probe and autoradiographed. After washing, the same filter was rehybridized with a PAI-1 32P cDNA probe. The filter was rehybridized with a mouse p actin 32PcDNA probe to control RNA transfert.

L

2

i

'er+

Hours

k

2

4

'e? Hours

Fig. 6. Quantification of the time course of u-PA and PAI-1 mRNA accumulation after thrombin stimulation. The changes of u-PA (left panel)and PAL1 (rightpanel)mRNA levels have been quantified by densitometric analysis by scoring for each sample the relative intensity of u-PA or PAI-1 mRNA signal reported to those of correspondent mouse p actin mRNA signal. The value of the ratio is expressed using arbitrary units ( A Thrombin stimulation 10.' M; A control).

maximal effect was observed at lo-' M thrombin. Zymographic analysis demonstrated that the 53 Kdu-PA activity detected in conditioned medium of GECs treated by 10.' M thrombin was decreased and that free t-PA (70 Kd) was never found in the conditioned medium (Fig. 8). By immunoblot we demonstrated that GECs released a 53 Kd form of u-PA and also other forms with higher molecular weight (Fig. 9). The 53 Kd form is not modified in reducing conditions indicating that it is a single-chain form of u-PA. Conversely, after

138

HE ET AL.

A

A

I

10

20

30

40

Time hours %

Fig. 7. Decreased u-PA activity of thrombin stimulated GECs culture conditioned medium. A. Time-course study of the urokinase activity of the culture medium of glomerular epithelial cells incubated in the absence ( A ) and presence of 10-8Mthrombin (A ). Mean -t SEM of 3 separate experiments are shown. ** p < 0.001 as compared to

Fig. 8. Zymography of culture medium conditioned by glomerular epithelial cells for 24 h. Lane 1 = 1 U/ml standard u-PA, lanes 2 to 5 = conditioned culture medium of GECs incubated in unstimulated conditions (lane 2), or in the presence of thrombin 10.' M (lane 3), 5.10.' M (lane 4), 10.' M (lane 51, or 2.10.' M (lane 6). As previously described (Rondeau et al., 1989). GECs release mainly 53 Kd u-PA and small amounts of PAIl-t-PA and PAI1-u-PA complexes. Thrombin decreased the activity of 53 Kd u-PA.

incubation with lo-' M thrombin, the conditioned medium still contained a 53 Kd form of u-PA, but in reducing conditions this form is converted into two low molecular weight chains of 30 Kd and 23 Kd approximately. These results and the decreased u-PA activity after addition of thrombin suggested that thrombin partially cleaved SC-u-PA, which could not then be converted to the active form.

DISCUSSION Thrombin receptors of GECs We report here for the first time the presence of two different classes of thrombin receptors on human GECs. Thrombin effects on GECs require both binding

levels observed in the presence of thrombin. B. Dose-response effect of thrombin on u-PA activity of culture medium conditioned by GECs for 24 h. Half maximal inhibition of u-PA activity was observed at M thrombin.

Fig. 9. Immunoblotting with goat polyclonal antiurokinase antibodies of GEC culture conditioned medium without (lanes 1 and 4) or with 10~'M thrombin (lanes 2 and 3). Electrophoresis was performed with 5 fold concentrated samples in non reducing (lanes 1 and 2) and in reducing conditions (lanes3 and 4). Goat antibodies were detected by incubation with biotinylated mouse antigoat Ig, and then with avidin peroxydase revealed by chloronaphtol and H,02.

capacity to receptors and catalytic activity of thrombin to induce cell proliferation as reported for fibroblasts (Carney and Cunningham, 1978) or other cell types (BarShavit et al., 1983). We demonstrate here that they are also required for u-PA, t-PA, and PAI-1 antigen productions. DFP-inactivated thrombin, which has the capacity to bind but is enzymatically inactive (Awbrey et al., 1979; Fenton, 19861, had no effect. Hirudin, which blocks thrombin activity and prevents membrane binding (Low and Cunningham, 1982; Fen-

THROMBIN AND FIBRINOLYSIS BY GLOMERULAR EPITHELIAL CELLS

ton and Bing, 1986) also inhibited these effects of thrombin. The exact mechanism of action of thrombin at the cell surface is not known. It is possible that membrane binding of thrombin to specific receptors and proteolytic cleavage of a membrane protein induce two separate intracellular signals that are both required for the cellular effect of thrombin to appear. In fibroblasts, binding of thrombin to high affinity receptors reportedly causes an increase of intracellular CAMP,whereas catalytic activity could induce protein kinase C activation (Gordon et al., 1984; Frost et al., 1987). This is the “two signal” hypothesis that is provided to explain thrombin regulation of mitogenesis in fibroblasts (Carney et al., 1986). Interestingly, the same concentration of thrombin, i.e., loa M, induced half maximal stimulation of thymidine incorporation inhibition of u-PA binding and maximal increased of u-PA, t-PA and PAI-1 antigen release. Further studies are required to determine if cell growth and synthesis of fibrinolytic components are under similar regulatory control in GECs. Thrombin: a possible corner-stone of fibrin deposition and GECs proliferation Thrombin generation in the kidney glomerulus can result from activation of glomerular cells themselves, which produce a procoagulant activity identified as tissue-factor (De Prost and Kanfer, 1985; Kanfer et al., 1987) or from infiltrating inflammatory cells (Holdsworth and Tipping, 1985; Tipping and Holdsworth, 1986). Our results show that thrombin increases u-PA, t-PA, and PAI-1 productions by GECs, but that overall it decreases the fibrinolytic activity. We demonstrate here that thrombin can prevent activation of pro-u-PA by plasmin, probably by inducing a limited proteolysis of pro-u-PA. Immunoblotting analysis demonstrated that in the presence of thrombin the conditioned medium of GECs still contains a 53 Kd form of u-PA that is inactive and is converted into two separate chains in reducing conditions. Half maximal inhibition of u-PA activity was observed at lo-’ M thrombin, an order of magnitude lower than that required to observe the cellular effects of thrombin. Our results are in accord with the previously reported cleavage of pro u-PA in vitro (Ichinose et al., 1986).The increased t-PA release is not associated with a corresponding increase in fibrinolytic activity since t-PA is complexed to PAI-1. For the first time, it is also shown that thrombin decreases u-PA binding sites on GECs. Since this effect requires a minimum of 8 h (not shown) to appear and since affinity was not modified, it is unlikely that thrombin induces conformational changes of u-PA binding sites by a direct proteolytic effect. We previously ruled out a competition between thrombin and u-PA on the same receptor (Rondeau et al., 1989).Our results rather suggest an inhibition of u-PA receptor synthesis or expression at the cell surface. Recently EGF and PMA have been shown to increase the number of u-PA sites on Hela cells and on U-937 cells and to decrease the affinitv of these binding sites. These effects were related to thLdifferentiation the tumoral cells in the presence of PMA and EGF (Estreicher et al., 1989). Further studies are required to

a

139

understand the mechanism of regulation of u-PA binding sites on GECs by thrombin. The growth-promoting effect of thrombin is also described for the first time on GECs. We demonstrate both an increased DNA synthesis and cell division in the presence of thrombin. This effect was potentiated by low doses of serum or insulin, suggesting cooperation between intracellular pathways used by these different growth factors, e.g., tyrosine kinase and protein kinase C. It has been shown recently that insulin can increase PAI-1 synthesis by hepatocytes (Alessi et al., 19881,and these results could indicate a link between cell growth and expression of components of fibrinolytic system. Taken together our results suggest that thrombin could be a mediator of glomerular injury by inducing local fibrin formation, GECs proliferation, and also inhibition of fibrin degradation. ACKNOWLEDGMENTS This work was supported by grants from the Association Claude Bernard, Paris, the Institut National de la Sante et de la Recherche Medicale, and by the Fond National Suisse, grant no 3.334-0.86. We are indebted t o Miss Mina Mallet for excellent secretarial assistance. LITERATURE CITED Alessi, M.C., Juhan-Vague, J., Kooistra, T., De Clerck, P.J., and Collen, D. (1988) Insulin stimulated the synthesis of plasminogen activator inhibitor 1by the human hepatocellular cell line Hep G2. Thromb. Haemost., 60(4):491494. Awbrey, B.J., Hoak, J.C., and Owen, W.G. (1979)Binding ofthrombin of cultured human endothelial cells. J. Biol. Chem., 254:40924095. Bar-Shavit, R., Kahn, A,, Fenton, J.W., and Wilner, G.D. 11983) Receptor-mediate chemotactic response of a macrophage cell line iJ 774) to thrombin. Lab. Invest., 49:702-707. Carney, D.H., and Cunningham, D.D. (1978) Role of specific cell surface receptors in thrombin-stimulated cell division. Cell, 15:1341-1349. Carney, D.H., Scott, D.L., Gordon, E.A., and Labelle, E.F. (1985) Phosphoinositides in mitogenesis: Neomycin inhibits thrombinstimulated phosphoinositide turnover and initiation of cell proliferation. Cell, 42:479-488. Carney, D.H., Herbosa, G.J., Stienberg, J., Bergmann, J.S., Gordon, E.A., Scott, D.L., and Fenton, J.W. (1986) Double signal hypothesis for thrombin initiation of cell uroliferation. Semin. Thromb. Haemostasis, 12931-240. Collen, D., and Lijnen, C. (1986) The fibrinolytic system in man. Critical Review in OncoloaviHematolom. 4:249-301. De Prost, D., and Kanfer,-k (1985) Quantitative assessment of procoagulant activity of isolated rat glomeruli. Kidney Int., 28:566568. Dichek, D., and Quetermous, Th. (1989) Thrombin regulation of mRNA levels of tissue plasminogen activator and plasminogen activator inhibitor-1 in cultured human umbilical vein endothelial cells. Blood, 74:222-228. Estreicher, A,, Wohlwend, A,, Belin, D., Schleuning, W.D., and Vassalli, J.D. (1989) Characterization of the cellular binding site for the urokinase-type plasminogen activator. J. Biol. Chem., 264:1180-1189. Fenton, J.W. (1986) Thrombin. Ann. N.Y. Acad. Sci., 485:5-15. Fent,nn. .~....LW.. . , and Bina. D.H. 11986) Thrombin active site regions. .. Semin. Thromb. Haemost., 12200-208. Frost, G.H., Thommon. W.C.. and Carney, D.H. (1987) Monoclonal antibodv ‘to the thrombin receutor stimulates DNA svnthesis in combination with gamma thrombin or phorbol myristate acetate. J. Cell. Biol., 1052551-2558. Gelehrter, Th.D., and Szycer-Laszuk, R. (1986)Thrombin induction of Dlasminoaen activator inhibitor in cultured human endothelial cells. J. d i n . Invest.. 77165-169 Genton, C , Kruithof, E.K.O., and Schleuning, W.D. (1987) Phorbol ester induces the biosynthesis of glycosylated plasminogen activator ~

~

~~~~

~

~~~

0,

~

~~

~~~

140

HE ET AL.

inhibitor 2 in high excess over urokinase type plasminogen activator in human U 937 lymphoma cells. J. Cell. Biol., 104:705-712. Gordon, E.A., and Carney, D.H. (1986) Thrombin receptor occupency initiates cell proliferation in the presence of phorbol myristate acetate. Biochem. Biophys. Res. Commun., 142(2):650-656. Gordon, E.A., Fenton, J.W., and Carney, D.H. (1984) Thrombinreceptor occupancy initiates a transient increase in CAMPlevels in mitogenically responsive hamster (NIL) fibroblasts. Ann. N.Y. Acad Sci., 485249-263. GosDodarowicz. D.. Brown. K.B.. Bridwell. C.R.. and Zetter. B.R. (f978) Control of proliferation of human vascular endothelial cells to fibroblast growth factor and thrombin. J. Cell. Biol., 77774-788. Holdsworth, S.R., and Tipping, P.G. (1985) Macrophage induced glomerular fibrin deposition in experimental glomerulonephritis in the rabbit. J. Clin. Invest., 76:1367-1374. Ichinose, A., Fujikawa, K., and Suyama, T. (1986) The activation of pro-urokinase by plasma kallikrein and its inactivation by thrombin. J. Biol. Chem., 261:3486-3489. Kanfer, A., De Prost, D., Guettier, C., Nochy, D., Le Floch, V., Hinglais, N., and Druet, Ph. (1987) Enhanced glomerular procoagulant activity and fibrin deposition in rats with mercuric chloride induced autoimmune nephritis. Lab. Invest., 57138-143. Knudsen, B.S., Harpel, P.C., and Nachman, R.L. (1987) Plasminogen activator inhibitor is associated with the extracellular matrix of cultured bovine smooth muscle cells. J. Clin. Invest., 80:1082-1089. Lacave, R., Rondeau, E., Ochi, S., Delarue, F., Schleuning, W.D., and Sraer, J.D. (1989) Characterization of a plasminogen activator and its inhibitor in human mesangial cells. Kidney Int., 35:80&811. Laemmli. U.K. (1970) Cleavaee of structural Droteins during the assembly of the head of gacteriophage T i Nature (LonYdon), 22 7~680-685. Levin, E.G., Marzed, U., Anderson, J., and Harker, L.A. (1984)

Thrombin stimulated tissue plasminogen activator release from cultured human endothelial cells. J. Clin. Invest., 74:1988-1995. Loskutoff, D.S., Van Mourik, S.A., Erickson, L.A., and Lawrence, D. (1983) Detection of a n unusually stable fibronolytic inhibitor produced by bovine endothelial cells. Proc. Natl. Acad. Sci. USA, 80:29562960. Low, D.A., and Cunningham, D.D. (1982) A novel method for measuring cell surface-bound thrombin. J . Biol. Chem., 257:8504358. Medcalf, R.L., Van Den Berg, E., and Schleuning, W.D. (1988) Glucocorticoid-modulated gene expression of tissue and urinarytype plasminogen activator and plasminogen activator inhibitor 1 and 2. J. Cell. Biol., 106:971-978. Rondeau, E., Ochi, S., Lacave, R., He, C.J., Medcalf, R., Delarue, F., and Sraer, J.D. (1989) Urokinase synthesis and binding of glomerular epithelial cells. Kidney Int., 36593-600. Sraer, J.D., Kanfer, A., Rondeau, E., and Lacave, R. (1988) Glomerular hemostasis in normal and pathologic conditions. Adv. Nephrol., 1727-55. Stoppelli, P., Tacchetti, C., Cubellis, V., Corti, A., Hearing, V.J., Cassani, G., Appella, E., and Blasi, F. (1986) Autocrine saturation of pro-urokinase receptors on human A 431 cells. Cell, 45:675-684. Stricker, G.E., Killen, P.D., and Farin, F.M. (1980) Human glomerular cells in vitro: Isolation and characterization. Transplant. Proc., 12:88-99. Tipping, P.G., and Holdsworth, S.R. (1986) The participation of macrophages, glomerular procoagulant and Factor VIII in glomerular fibrin deposition. Studies in anti-GBM antibody induced glomerulonephritis in rabbits. Am. J. Pathol., 124:lO-17. Zetter, B.R., and Antoniades, H.M. (1979) Stimulation of human vascular endothelial cell growth by a platelet-derived growth factor and thrombin. J. Supranol. Struct., 22:361-370.