Collagen-induced Binding to Human Platelets of Platelet-derived ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 8, Issue of March 15, pp. 4336-4341,1989 Printed in U.S.A.

Collagen-induced Bindingto Human Platelets of Platelet-derived Growth Factor Leading to Inhibition of P43 and P20 Phosphorylation* (Received for publication, February 29,

1988)

Marie ClaudeBryckaertz, Francine Renduz, Gerard TobelemSg,and Ake Wastesonll From the SH6pital Lariboisihre,Znstitut National de la Santi et de la RechercheMedicak U 150, Centre National de la Recherche Scientifique UA 334, 6 rue Guy Patin, 75010 Paris, France and the 7Department of Cell Biology, Faculty of Health Science, Uniuersity of Linkoping, 858 185, Linkoping, Sweden

Platelet-derived growth factor (PDGF) is known to inhibit collagen-induced platelet aggregation. Collagen-induced binding of 12’I-PDGF to human washed platelets wastherefore investigated. It was found 1) to be time-dependent, reaching a plateau at 20 “C after 30 min, 2) collagen concentration-dependent, 3) specifically inhibited by unlabeled PDGF, and 4) saturable. Scatchard plot analysis showed a single class of sites with 3000 f 450 molecules bound/cell and an apparent KO of 1.2 f 0.2 lo-* M. The effects of PDGF on collagen-induced phosphoinositide breakdown and protein phosphorylation were alsoinvestigated. At 50 ng/ml PDGF, a concentration which completely inhibited collagen-induced aggregation, the breakdown of [32P]phosphatidylinositol4,5-biphosphate (PIP2) and [32P]phosphatidylinositol4-phosphate (PIP)was observed, but the subsequent replenishment of [32P]PIP2 was inhibited. The same PDGF concentration totally inhibited collagen-induced phosphatidic acid formation. PDGF also completely prevented phosphorylation of P43 and P20, as a result of protein kinase C activation consecutive to phosphoinositide metabolism. These results suggest that (i) a specific PDGF receptor can be induced by collagen, and (ii) PDGF can affect the early eventsof collagen-induced platelet activation by inhibiting PIP2 resynthesis and P43 and P20 phosphorylation. It is concluded that PDGF might be involved in a negative feed-back control of platelet activation.

Platelet-derived growth factor (PDGF)’ from human platelets, is a cationic glycoprotein of about M , 30,000 (1). PDGF is an cy granule component (2) released during the platelet release reaction induced by thrombin, collagen, or ADP (3), or when platelets adhere to sites of blood vessel injury. In addition to PDGF, other growth factors are located in platelets, including epidermal growth factor (EGF) (4),transforming growth factor (5), hepatocyte growth factor (6), and endothelial growth factor (7). The existence of target cells for PDGF hasbeen demonstrated in connective tissue, fibroblasts

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. To whom reprint requests should be addressed. The abbreviations used are: PDGF, platelet-derived growth factor; EGF, epidermal growth factor; SDS, sodium dodecyl sulfate; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; Temed, N,N,N’,N’-tetramethylethylenediamine;DG, diacylglycerol; IP3, inositol trisphosphate; PIP, phosphatidylinositol 4-phosphate; PIP,, phosphatidylinositol 4,5-biphosphate; PI, phosphatidylinositol; PA, phosphatidic acid.

(8), smooth muscle cells (9), glial cells (lo), and trophoblasts (11).

It has been shown that different (Y granule components, like platelet factor 4 (PF4) (12), factor VIII/von Willebrand (13), thrombospondin (14,15),fibrinogen (16), and fibronectin (17) bind to the platelet membrane during activation. We thus postulated that like these cy granule components, PDGF might also bind to the platelet membrane and interfere with the platelet functions. We previously showed that PDGF alone did not induce any platelet aggregation or secretion, but did modify the phosphoinositide metabolism of platelets prelabeled with [32P]orthophosphate (18). In addition, PDGFinhibited thrombin and collagen-induced platelet aggregation and [‘*C]serotonin release in a dose-dependent manner (18). These findings suggested that PDGF binds to platelets and alters platelet responses. In view of these earlier data, we examined here the possibility of PDGF binding to theplatelet membrane. The results demonstrate that PDGF binds specifically to this membrane at a site on the platelet surface. As a result of this binding, PDGF inhibitssome of the initial eventsof platelet activation, including the collagen-induced turnover of inositol phospholipids, and theconsequent phosphorylation of P43 and myosin light chain P20. MATERIALSANDMETHODS

Reagent~-[~~P]Orthophosphate(370 Mbq/ml) and carrier-free [‘251]sodium iodide werepurchased from CEA (Saclay, France). [“C] Serotonin (1850MBq/mmol) and a thromboglobulin/radioimmunoassay kit were purchased from Amersham (Les Ulis, France). Collagen type I was obtained from Diagnostica Stago (AsniBres, France), and EGF from Collaborative Research (Waltham, MA). Metrizamide came from Nyegaard (Oslo, Norway). Flurbiprofen was purchased from Boots Dacour (Courbevoie,France). Thin layer chromatography plates werefrom Merck (France) and acrylamide, bisacrylamide, Temed, sodium persulfate, glycine, and sodium dodecyl sulfate were purchased from Bio-Rad; x-ray films and x-omatic intensifying screens came from Eastman Kodak Co., and dibutyl phthalate and dioctyl phthalate from Aldrich Chemical Co. (Beerse, Belgium). The scintillation fluid, Dynagel, waspurchased from Baker (Netherlands). Purification and Radioiodination of PDGF-PDGF purified from human platelets was 90-95% pure as estimated from analytical NaDodS04-polyacrylamide gel electrophoresis. PDGF was radiolabeled using a modified chloramine-T method (19),with a specific activity of about 20,000-40,000 cpm/ng (20).Under these conditions, ‘“I-PDGF had about 80% of the biological activity of native PDGF (27). Binding Assay-Binding of Y - P D G F was assessed on fresh human platelets. Human blood was obtained from normal volunteers, and collected on acid/citrate/dextrose. Platelets were isolated and washed by the method of Patscheke and Worner (21).With this method, the number of contaminating cells is less than 1 leukocyte/ 1000 platelets. Platelets were resuspended in 0.02 M sodium phosphate, 0.15 M sodium chloride, and 0.3% bovine serum albumin, pH 7.4.Aliquots of 0.3 ml of platelet suspension (final concentration 2.10

4336

Binding of PDGF to

Platelets: Inhibition of P43 Phosphorylation

X 10' platelets/ml) were preincubated for 5 min at room temperature with 2.5 ng of lz5I-PDGF (40.000 cpm/ng) as tracer, and 100 pl of collagen (final concentration 20 pg/ml) was then added. The final sample volume in each tube was 0.5 mlcorresponding to 10' platelets/ tube. Nonspecific binding was performed with a 100-fold excess of unlabeled PDGF. Incubation with gentle stirring was stopped after 45 min. At selected times thereafter, 0.4 ml of the platelet suspension was layered onto 0.5 ml of a mixture of dibutyl phthalate/dioctyl phthalate (2.5:2, v/v) and centrifuged at room temperature for 3 min at 12,000 X g i n a Beckman microfuge. The supernatantwas aspirated and the'I in the platelet pellets was measured. P-Thromboglobulin and [14C]Serotonin Release-0-Thromboglobulin and ['4C]serotonin release were measured during the above binding assay. Platelets were labeled with 0.6 p~ ["Clserotonin and washed as above. The reaction was stopped by transferring the suspension (0.4 ml) into 0.1 ml of 0.1 M EDTA and was immediately centrifuged for 1 min in an Eppendorf microcentrifuge. 0-Thromboglobulin and ["C]serotonin were measured in the supernatant by a radioimmunoassay using a P-thromboglobulin/radioimmunoassay kit, and by liquid scintillation counting in a spectrometer, respectively. Phospholipid Metabolism and Protein Phosphorylation-Plateletrich plasma was incubated for 90 min at 37 "C with [32P]orthophosphate (37 MBq/lO ml). After incubation, platelets were washed and isolated on metrizamide gradients as previously described (22). Platelets were then resuspended at 4 X 10' platelets/ml in 10 mM Hepes, 140 mM NaCl, 3 mM KCl, 0.5 mMMgC12, 12 mM NaHC03, 10 mM glucose, pH 7.4. Platelet samples (0.4 ml) were transferred to the aggregometer and preincubated for 1 min at 37'C with or without human PDGF (50 ng/ml). Next, platelets were stimulated with 1020 pg/ml collagen. The reaction was stopped at different times by transferring the sample into glass tubes containing 3.75 volumes of chloroform, methanol, 12 N HCl, 0.1 M EDTA (20:40:1:2, v/v/v/v) at 4 "C. The preparation was then centrifuged and the phospholipids extracted and separated. The proteins whichwere located at the phospholipid interphase were analyzed on the same sample as phospholipids (23). The organic phase was evaporated under Nz and resuspended in chloroform/methanol/H*O (75:25:2, v/v/v). Inositol phospholipids were separated by thin layer chromatography on silica plates according to Jolles et al. (24). [32P]PIP2,[32P]PIP, [32P]PI, and [32P]PAwere located by autoradiography and quantified by scraping the identified spots off the plates and counting in a scintillation counter. 32P-labeled proteins were solubilized for 1 h at 56 "C in 0.6 M Tris-HC1 buffer, 2% SDS (w/v), 20% glycerol (v/v) and 0.01% bromophenol blue (w/ v), pH 10, and reduced with 5% (v/v) P-mercaptoethanol for 30 min at 37 "C. They were then applied onto 13% SDS-polyacrylamide gels. Gels were dried and exposed for autoradiography using Kodak XAR film. Autoradiograms were scanned with an LKB ultroscan.

4337

T i m e (min)

FIG. 1. Time course of '"I-PDGF binding to human platelets. A , time course of lz5I-PDGF binding to collagen-stimulated human platelets. Platelets (2 X 10' cells/ml) were preincubated for 5 min with 2 ng/ml "'1-PDGF and 20 pg/ml of collagen was added. Incubation was performed at 20 "C with gentle stirring. Total '%IPDGF binding was expressed in picograms/lO' platelets (0).Nonspecific binding was tested in the presence of a 100-fold excess of unlabeled PDGF (0)on collagen-stimulated platelets. Specific binding was calculated by subtracting nonspecific binding from total binding (A). Results are means f S.E. of four experiments. B , time course of '"I-PDGF binding to unstimulated human platelets. Platelets (2 X 10' cells/ml) were preincubated for 5 min with 2 ng/ml '"IPDGF. Incubation was performed at 20 "C under gentle stirring. Total '"I-PDGF binding was expressed as picograms/lO' platelets ( 0 ) .Nonspecific binding was tested in the presence of a 100-fold excess of unlabeled PDGF (0).Specific binding was calculated by subtracting nonspecific binding from total binding (A). Results are means f S.E. of three experiments.

RESULTS

Time Course of PDGF Binding-The time course of 1251PDGF binding to activated platelets was studied at 20 "C with 20 pg/ml of collagen (Fig. 1A). This binding was time-dependent and reached saturation after 30 min. Under these conditions, 12 pgof specific '251-PDGFbound to10' platelets. a In these experiments, binding to activated platelets with 100-fold excess of unlabeled PDGF constituted 30-35% of

-A

0

1

1

total binding. O 1 0 10 20 30 40 50 Cdlrgrn l p g h l 1 The timecourse of lz5I-PDGF binding to unstimulated 1251-PDGFbound platelets was alsomonitored(Fig.1B). FIG. 2. Dependence on the collagen concentration of spespecifically t o unstimulated platelets(5 pg/108 platelets). cific PDGF binding to human platelets. Washed platelets were Nonspecific binding was determined with a 100-fold excess of adjusted to 2 X 10' platelets/ml and incubated for 5 min with 2 ng/ unlabeled platelet, also on unstimulated platelets, and consti- ml '251-labeledPDGF. Platelets were then stimulated for 45 min at 20 "C with different concentrations of collagen (0-50 pg/ml). Specific tuted 50% of total binding. When nonspecific binding was binding was calculated for each collagen concentration by subtracting 7.5 pg/108 plateletswere measured on activated platelets, from total binding the nonspecific binding obtained with a 100-fold bound compared to5 pg/108 platelets on unstimulated plate- excess of unlabeled PDGF. Results are means f S.E. of four experilets. ments. "'I-PDGF Binding to Plateletsas a Function of the Collagen Concentration-The binding of '251-PDGF(2 ng/ml) was ex- Above 20 pg/ml, however, binding of '251-PDGF reached a amined as a function of t h e collagen concentration, which plateau at 16 pg/108 platelets. Total binding did not exceed ranged from0 to 50 pg/ml (Fig. 2). For up t o 20 lg/ml, specific 6-2.5% of the total added radioactivity.When a 100-fold 1251-PDGFbinding increased with the collagen concentration.excess of unlabeled PDGF wasadded for each concentration

Binding of PDGF Platelets: to Inhibition

4338

of collagen, 1.8-0.7% of the total radioactivity was bound, and represented nonspecific binding. A control experiment was performed without platelets to ensure that the 12'I-labeled PDGF did not bind to collagen (results not shown). Specificity of PDGF Binding-The specificity of the 1251PDGF binding was investigated in two sets of experiments (Fig. 3). In the first, competitive inhibition of this binding was assessed either with an excess of unlabeled PDGF orwith epidermal growth factor. Incubation of I2'I-PDGF with unlabeled PDGF at 20 "C for 45 min gradually inhibited 12'1PDGF binding. With a 200-fold excess of unlabeled PDGF, this inhibition was about 80%. Incubation with an excess of unlabeled EGF, another growth factor also present in platelets, did not interfere with 12'I-PDGF binding. In the second set of experiments, the specificity of 12'1PDGF binding was established by adding to the platelets different amounts of l2'1-PDGF and unlabeled PDGF in order to maintain a constant total PDGFconcentration. The binding was found to be proportional to the total PDGF added (Fig. 4),indicating that thelabeling of PDGF did not alter its affinity for the platelet membrane. Saturation Binding and Scatchard Analysis-The amount of PDGF bound to platelets was quantified with '251-PDGF

-s

c P

of P43 Phosphorylation

PDGF added ( n ~ / l O * p l a l d ~ l s )

FIG. 5. Concentration dependence and Scatchard analysis of 'a51-labeledPDGF binding to collagen-stimulated platelets. Platelets were stimulated with a final concentration of 20 pg/ml of collagen. Various concentrations of PDGF were preincubated for 5 min with platelets (2 X 10' cells/ml) and then incubated a t 20 "C for 45 min with collagen. Results are those of five experiments on five different donors.

TABLE I Effect of flurbiprofen, a platelet release inhibitor, on '26Z-PDGFbinding Platelets (2 X lo8 cells/ml) were pretreated with saline or 10 p~ flurbiprofen and incubated for 5 min under gentle stirring with 6 ng/ m112SI-PDGF.They were stimulated for 45 min at 20 "C with 10-30 pg/ml collagen. Specific binding was calculated as in Fig. 2. ["C] Serotonin release and 8-thromboglobulin release were determined under identical conditions, using unlabeled PDGF in place of "'1PDGF. Results are means & S.E.of three experiments.

~~~~~~

~

~~~

~

PDCF 10

0 n

-b

?o

0

0

0

100

200

9- Fold exce8s of unlabeled growth factors u;

1

36f5 u f 3.2

4.2 3.1

f

45 46

f

9 3 1

5

f

f

4.1

f

51

f

9.3

u

N r

f 4.4

1.1 0.1

5.3 4.2

f

0.1 0.8

6.5 4.4

f

6.2 f 2.1 5.1 f 1.5

6.2 5.1

f

f

f

f

f

3.2 2.3

2.8 3.1

3.2 3.3

aQ

FIG. 3. '"I-PDGF binding to collagen-stimulated human platelets; effect of unlabeled P D G F (0)or EGF (0).Platelets (2 X 10' cells/ml) were incubated for 45 min at 20 "C with 2 ng/ml lZ51PDGF and either unlabeled PDGF or EGF, and a final collagen concentration of 20 pg/ml. Data are expressed (in percent) as a ratio of'"1-PDGF bound to the "'I-PDGF bound in the absence of an added growth factor. Results are means f S.E. of four experiments.

!"..1 k

P

t

t

% of added lZ6 I-PDQF

100

3

c

t:

g

50

c 0

TIME (min)

s \

n

k e

2

N

111

o

o

50 % Of added unlabeled

100 PDQF

TIME (nin)

FIG. 6. Effects of P D G F on variations with time in the amounts of SaP-labeled PIP%,PIP, PI, and P A induced in human platelets stimulated by 10-20 pg/ml collagen. 32P-Labeled platelets (4 X IO8 cells/ml) were stimulated with collagen and incubated a t 37 "C in the absence (0)or presence (0)of 50 ng/ml of PDGF in the aggregometer. Results are expressed as the ratio of the radioactivity of a given phospholipid to the radioactivity of [32P]PA in unstimulated platelets. Results are means f S.E. of four experiments.

FIG. 4. Effect of radioiodination on PDGF binding to human platelets. The ratio of '"1-PDGF to unlabeled PDGF varied but the total final PDGF concentration was constant. Platelets (2 X lO'cells/ ml) were incubated for 45 min at 20 "C with different amounts of "'1PDGF and unlabeled PDGF (O),and a final collagen concentration of 20 pg/ml. Data are expressed (in percent) as a ratio of lZ51-PDGF as a tracer and different concentrations of unlabeled PDGF. bound to the 1261-PDGFbound in the absence of unlabeled PDGF. When the total PDGF added reached 350 ng, the specific binding curve formed a plateau at 14 ng of PDGF bound per Results were means f S.E. of four experiments.

Binding of PDGF to Platelets: Inhibition of P43 Phosphorylation COLLAGEN P 43

P 20

:i

In the presence of 50 ng/ml PDGF, collagen-induced P43 and P20phosphorylation was completely inhibited (Fig. 7) as was platelet aggregation. When aggregation was partially inhibited, only partial inhibition of P43 andP20 phosphorylation was observed (results not shown). DISCUSSION

4

The present investigation showed that PDGFbinds to both resting platelets andcollagen-stimulated human platelets and that this binding inhibits the stimulation by collagen. PDGF binding to collagen-stimulated platelets was more intensive than toresting platelets. With collagen stimulation, it wasspecific, saturable, and dependent on the collagen concentration. Saturation was obtained with 20 pg/ml of collagen. Its specificity was demonstrated by the capacity of unlabeled PDGF to compete with lz51-labeledPDGF and by the failure of EGF, anothergrowth factor present in platelets, to compete with PDGF. Furthermore, at the level of the platelet membrane, no "transmodulation mechanism" between PDGF and EGF similar to thatdescribed for fibroblasts (26) was observed, since EGF did not compete. PDGF binding to stimulated platelets was saturable with one class of sites comprising 3000 450 molecules/cell, anditsapparent KD was 1.2 X lo-* M. No other class of sites that might correspond to binding of unstimulated platelets was detected. This suggests that either (i) binding sites are identical on resting and activated platelets, but that activation increases the number of sites expressed with the same affinity, or (ii) binding sites are different for resting and stimulated platelets, but are all expressed during activation. If the affinities are about the same, the sites expressed on resting platelets will account for about 40% of PDGF binding to collagen-stimulated platelets. Certain cy granule proteins are adhesive and bind to the platelet surface. Of these, fibrinogen, fibronectin, and factor VIII/von Willebrand factor require platelet activation for binding (13,16,17), whereas PDGF might behave like thrombospondin and bind to resting platelets (14, 15). Here, the nonspecific PDGF binding to unstimulated plateletswas considerable. The use of a lower '251-PDGF concentration with higher specific activity might have reduced this nonspecific binding. However, under these conditions, the biological activity of the PDGF, measured by [3H]thymidineincorporation was found to be greatly reduced. In any case, the present collagen-induced activation enhanced PDGF binding. PDGF inhibited the platelet responses to collagen. The maximal inhibition of collagen-induced platelet aggregation, PA production, and phosphorylations was obtained with 50 ng/ml PDGF. Thus,there is no correlation between this inhibitory effect and the saturation of PDGF receptors which required approximately 350 ng/ml PDGF. Some of the effects of PDGF on activated plateletsmay be due to its binding to the sites expressed on unstimulated platelets. This hypothesis is supported by the fact that porcine PDGF alone modified the metabolism of polyphosphoinositides (18).Moreover, we observed that the inhibitory effect of PDGF on collageninduced aggregation was stronger after plateletpreincubation with PDGF alone, before the addition of collagen. However, we cannot exclude the possibility that some of the binding sites available on activated platelets are also involved. Scatchard analysis showed only one class of sites, suggesting either that the sites on resting and stimulated plateletsare identical but more numerous after stimulation, or that they are different, depending on whether or not theplatelets are activated, even if their affinity is about the same. The fact that binding of PDGF inhibited the metabolic events involved in P43 and P20 phosphorylation in activated but not in resting platelets

~l ~l 0

1

2

3

04

1

2

3

4

TIME fain)

FIG. 7. Time course of the inhibition of PDGF on collageninduced platelet phosphorylation of P43 (le&) and P20 (right).All samples were incubated for 3 min in the aggregometer. the ratio of the radioactivity of P43 and P20 Results are expressed as to that of P43 and P20 in unstimulated platelets. Results are those of four representative experiments on four different donors.

10' platelets. Scatchard plot analysis based on five determinations gave 3000 450 molecules/platelet with an apparent Krt of 1.2 & 0.2 lo-' M, suggesting the presence of one class of binding site only. Effect of ~ ~ u r b i p r o ~ae nPlatelet , Release I n ~ ~ i ~on o rlZ5I, PDGF Binding-To investigate the effects of the compounds released from stimulatedplateletsonPDGF binding, the binding of 'T-PDGF (6 ng/ml) induced by different concentrations of collagen was measured in the presence of 10 pM flurbiprofen, a nonsteroid anti-inflammatory drugthat inhibits platelet release andin itsabsence (25). Under these conditions, the release re-inedextremely small (less than 6.5%), ie. about the same as that measured on unstimulated platelets (Table I). Consequently, no significant inhibition by flurbiprofen was measured, and lZ5I-PDGFbinding was not modified by its presence. Phosphoinositide Turnover-The modifications of platelet f32P]phosphoinositidesinduced by 10-20 pg/ml collagen (Fig. 6) showed that thebreakdown of [32P]PIP~ started 1min after collagen addition, at thesame time as the shape change. The amount of [32P]PIP2diminished to 81%of the control value initially measured on unstimulated platelets. The decrease was followed byresynthesis which reached up to 125% of the control value a t 3 min. During the same period, there was also a slight decrease in PIP to 87% of the control value, followed by a progressive increase to 110% above this value at 3 min. No significant variation in PI was observed. As a result of collagen stimulation, [32P]PAformation reached 4 times the control value. A concentration of 50 ng/ml PDGF, which completely inhibited platelet aggregation did not interfere with the collagen-induced reduction in [32P]PIP2 and [32P]PIP (Fig. 6). PDGF inhibited resynthesis of [32P]PIP2significantly and of f3'P]PA completely. No significant variation in the level of PIP or PI was observed. At PDGF concentrations below 50 ngfml, which only partiallyinhibited aggregation, partial resynthesis of PIP, and P A were noted (results not shown). Protein P ~ s p ~ o ~ l a ~ ~ ~ - I n c u b of a t i unstimulated on platelets for 3 min with PDGF had no effect on P43 and P20 phospho~lation,since P43 and P20, respectively, constituted 3% and 1-2% at most of the total radioactivity (Fig, 7). When platelets were stimulated by collagen, P43 and P20 were phosphorylated. Both phosphorylations were time dependent and were maximal 2 min after the addition of collagen.

*

4339

4340

Binding of PDGF toPlatelets: Inhibition of P43 Phosphorylation

suggests that its binding sites are different in the two cases, since saturation of all the receptors was not required to obtain the maximal biological response. In human medullary fibroblasts, PDGF induced maximal DNA synthesis ([3H]thymidine incorporation) at a concentration of 15 ng/ml, but saturation of PDGF receptors was only obtained at 100 ng/ml (27). On platelets, wheat germ agglutinin induced activation when only 17% of the lectin binding sites were occupied (28). When in the present experiments, the contribution of released platelet granule constituents to PDGF binding was estimated on the assumption that the ADP concentration in the dense granules was about 7 p~ (29), a small amount of ADP (0.02 p M ) was found to be secreted after stimulation with collagen. This suggests that PDGF binding is more dependent on platelet activation than granule release. The amount of PDGF secreted was found to be negligible. During a full aggregation response, involving a maximal release of PDGF, 4.10' platelets released 50 ng into the medium. This amount completely inhibited collagen-induced platelet aggregation. Furthermore, cells other than platelets also secrete PDGF: they include activated macrophages (30), endothelial cells (31), and arterial smooth muscle cells (32). In uiuo, these cells might participate in the mechanism of PDGF regulation in platelets. The effects on PDGF binding of other constituents from a granules such as fibrinogen and fibronectin have not been tested. On the platelet membrane, the receptors of these compounds are the GP IIb/IIIa complex, whereas the PDGF receptor on connective tissue cells, including fibroblasts(33), is a 185-kDa transmembrane glycoprotein. In fibroblasts, PDGF binding occurs on the external domain of its receptor, which results in the autophosphorylation of the receptor's cytoplasmic domain, thus showing the presence of tyrosine kinase activity (34, 35). On platelets, no such 185kDa transmembrane glycoprotein possessing tyrosine kinase activity has been identified. Nevertheless, platelets contain considerable tyrosine kinase activity of a different type (36, 37) whichis not correlated with cell proliferation, but may be involved in other cellular functions such as phosphatidylinositol kinase activity which has been shown to be associated with the phosphotyrosine protein pp60""'" (38). Collagen-induced platelet activation involves the activation of polyphosphoinositide-specific phospholipase Candthe subsequent phosphorylation of P43 andto a small extent of myosin light chain'P20 (23). Here, in the presence of PDGF, decreases in collagen-induced PIPzand PIP were observed,suggesting that PDGF did not interfere with phospholipase C activity. In contrast, it inhibited P A synthesis and P43 phosphorylation suggesting that it inhibited diacylglycerol and inositol trisphosphate (IP3) formation, since PIPz hydrolysis by phospholipase C yields the two messenger molecules IP3 and DG. IPS induces mobilization of Ca2+from the dense tubular system to the cytosol. The increase in cytosolic Caz+ primes the protein kinase C which can be activated by DG and phosphorylates P43. Since PI is the main source of DG (39) and since no variation in [32P]PIwas detected during collagen-induced platelet activation, it can be assumed that DG only originated from PIPz and PIP. If PDGF inhibits phospholipase C, the decrease in PIPz still observed might be the consequence of a direct effect of PDGF on the membrane (22). If PDGF hasno effect on phospholipase C, the amount of DG formed will be small, which will hinder (i) phosphorylation of DG to give PA, (ii) the activating effect of DG on protein kinase C, and (iii) the closing of the PI cycle back to PIP and PIP*. This would explain the fact that we did not observe any resynthesis of PIPz or PIP. However, we cannot exclude either stimulation of phosphomonoesterases or inhibition of kinases. The quan-

tification of IPSwould thus be necessary to establish clearly whether or not phospholipase C activity is inhibited by PDGF, and if so, to clarify the mechanism of this inhibition. PDGF inhibited bothP20 and P43 phosphorylations. This could be the consequence either of the phospholipase C inactivation discussed above, or of a direct effect of PDGF on protein kinase C, since both proteins can be phosphorylated by protein kinase C (39,40). Previous work has shown that in resting platelets, P43 phosphorylation was inhibited in the presence of PDGF, but that phosphorylation took place in its absence (41). This suggests that theplatelets were activated during their preparation. In our unstimulated control platelets, P43 only accounted for 2-3% of total protein phosphorylation. This low level of P43 was normal for the unstimulated state of the control platelets we prepared. The PDGF released from the a granules during platelet activation seems to modulate the platelet response by a negative feed-back control, that inhibits aggregation and secretion as well as the phosphorylation of P43 and P20 which constitute the initial events during platelet activation. In this work, we showed for the first time to our knowledge, that an a granule component, PDGF, has an inhibitory effect on platelet activation. Until now, (Y granule components such as PF 4, factor VIII/von Willebrand, thrombospondin, and fibrinogen have been shown to have amplifying effects. REFERENCES 1. Deuel, T. F., Huang, J. S., Proffitt, R. T.,Baenziger, J. U., Chang, D., and Kennedy, B. B. (1981) J. Bwl. Chem. 256,8896-8899 2. Kaplan, K. L., Broekman, M. J., Chernoff, A., Lesznik, G. R., and Drillings, M. (1979) Blood 53,604-618 3. Witte, L. D., Kaplan, K. L., Nossel, H. L., Lages, B. A., Weiss, H. J., and Goodman, D. S. (1978) Circ. Res. 42,402-409 4. Oka, C. N., and Orth, Y. (1983) J. Clin. Inuest. 7 2 , 249-259 5. Assoian, R. K., Grotendorst, G. R., Miller, D. M., and Sporn, M. B. (1984) Nature 309,804-806 6. Assoian, R. K., Komoriya, A., Meyers, C. A., Miller, D. M., and Sporn, M. B. (1983) J. Biol. Chem. 258,7155-7160 7. Russell, W. E., Mc Govan, J. A., and Bucher, N. L. R. (1984) J. Cell. Physwl. 1 1 9 , 183-192 8. Clemmons, D. R., Isley, W. L., and Brown, M. T. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 164-165 9. Kohler, N., and Lipton, A. (1974) Exp. Cell Res. 87,297-301 10. Ross, R., Glomset, J. A., Kariya, B., and Harker, L. (1974) Proc. Natl. Acad. Sci. U. S. A. 7 1 , 1207-1210 11. Goustin, A. S., Betsholez, C., Pfeife, R., Ohlsson, S., Persson, H., Rydnert, J., Bywater, M., Holmgrew, G., Heldin, C. H., Westermark, B., and Ohlsson, R. (1985) Cell 41, 301-312 12. Capitanio, A. M., Niewiarowski, S., Rucinski, B., Tuszynski, G. P., Cierniewski, C. S., Hershock, D., and Kornecki, E. (1985) Biochim. Biophys. Acta 839, 161-173 13. Green, D., and Muller, H. P. (1978) Thromb. Haemostasis 3 9 , 689-694 14. Wolff, R., Plow, E. F., and Ginsberg, M. H. (1986) J. Biol. Chem. 261,6840-6846 15. Legrand, C., Dubernard, V., Kieffer, N., and Nurden, A. (1988) Eur. J . Biochem. 171,393-399 16. Plow, E. F., and Marguerie, G. A. (1980) Blood 56,553-555 17. Gardner, J. M., and Hynes, R. 0.(1985) Cell 42,439-448 18. Bryckaert, M. C., Rendu, F., Tobelem, G., and Caen, J. (1986) Biochem. Bwphys. Res. Commun. 135,52-57 19. Hunter, W. M., and Greenwood, F. C. (1962) Nature 194,495496 20. Heldin, C. H., Westermark, B., and Wasteson, A. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,3664-3668 21. Patscheke, H., and Worner, P. (1978) Thromb. Res. 12,485-496 22. Rendu, F., Marche, P., Maclouf, J., Girard, A., and Levy-Toledano, S. (1983) Biochem.Biophys.Res.Commun. 1 1 6 , 513519 23. Karniguian, A., Rendu, F., Grelac, F., Lebret, M., and Legrand, Y. J. (1987) Biochem. Biophys. Res. Commun. 146, 277-283

Binding of PDGFPlatelets: to Inhibition 24. Jolles, J., Schrama, L. H., and Gispen, W. H. (1981) Biochim. Biophys. Acta 666,90-98 25. Levy-Toledano, S., Maclouf, J., Caen, J., Bara, L., Lecrubier, C., and Samama, M. (1983) J. Phurmacol. 14,141-150 26. Sporn, M.,and Roberts, A. (1988) Nature 332,217-219 27. Bryckaert, M.C., Lindroth, M., Lonn, A., Tobelem, G., and Wasteson, A. (1988) Exp. Cell Res. 179, 311-321 28. Rendu, F., and Lebret, M.(1984) Thromb. Res. 36,447-456 29. Holmsen, H., Day, H. J., and Storm, E. (1969) Biochim. Biophys. Acta 186,254-266 30. Shimokado, K., Raines, E. W., Madtes, D. K., Barrett, T. B., Benditt, E. P., and Ross, R. (1985) Cell 43, 277-286 31. Barrett, T. B., Gajdosek, C. M., Shcwartz, S. M., Mc Dougall, J. K., and Benditt, E. P. (1984) Proc. Natl. Acad. Sci. U. S. A . 81,6772-6774 32. Sejersen, T., Betsholtz, C., Sjolund, M., Heldin, C. H., Westermark, B., and Thyberg, J. (1986) Proc. Natl. Acad. Sci.U. A . 83,6844-6848

s.

of P43 Phosphorylation

4341

33. Heldin. C. H.. Wasteson. A.. and Westermark. B. (1982) J. Biol. c h e i . 257; 4216-4221 ' 34. Heldin, C. H., Ek, B., and Rommstrand, L. (1983) J. Biol. Chem. 258,10054-10066 35. Frackelton, A. R., Tremble, P. M., and Williams, L. T. (1984) J. Biol. Chem. 259,7909-7915 36. Phan Dinth Tuv. F., Henry. J., Rosenfeld,. C.,. and Kahn, A. (1983) N a t u r e 3 0 5 ; 435-438 37. Golden, A., Nemeth, S. P., and Brugge, J. S. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,852-856 38. Courtneidge, S. A., and Heber, A. (1987) Cell 5 0 , 1031-1037 39. Majerus, P. W., Wilson, D. B., Connoly, T. M., Bross, T. E., and Neufeld, E. J. (1985) Leuk. Research (Hayaishi, O., and Yamamoto, S., ed) pp. 109-112, Raven Press, New York 40. Inagaki, M., Kawamoto, S., and Hidaka, H. (1984) J. Biol. Chem. 259,14321-14323 41. Gerrard, J. M., Israels, S. J., and Friesen, L. L. (1985) Nouu. Reu. Fr. Hematol. 2 7 , 267-273 42. Ishihara, N., Nagao, K., and Kobayashi, B. (1985) Thromb. Haemostasis 54,579-585 I

.

I