Binding and internalization of soluble fibrin by thrombocytes

THROMBOSIS RESEARCH 58; 175-l 83,199O 0049-3848190 $3.00 + .OO Printed in the USA. Copyright (c) 1990 Pergamon Press pk. All rights reserved.

BINDING AND INTERNALIZATION OF SOLUBLE FIBRIN BY THROMBOCYTES

Helmut

Hormann,

Max Planck-Institut

(Received

14.12.1989; (Received

Hartmut

Richter

fur Biochemie, German Federal accepted

and Viktorija

Martinsried Republic

in original form 10.1.1990

by the Executive

Jelinid

bei Mtinchen,

by Editor S. Witte)

Editorial Office 15.2.1990)

ABSTRACT Binding of soluble ‘*‘I-fibrin to platelets was investigated with thrombocytes separated by gelfiltration or by sedimentation as well as in a thrombocyte concentrate. Gelfiltered platelets failed to retain ‘251-fibrin within 16 hours unless they had been pretreated with factor XIIIa. In addition, a 30 kDacomponent derived from the N-terminal fibronectin domain was required as a mediator. Platelets isolated by sedimentation bound some 1251-fibrin even in the absence of those cofactors. The 30 kDa-component improved binding and only this increase was sensitive to putrescine inhibition. Evidently, centrifuged platelets unlike gelfiltered ones express two pathways of fibrin binding. In a thromboc te concentrate with platelets in their plasmatic environment ’Y51-fibrin was partially internalized. Engulfed radioactivity was detectable only for a limited period between 4-6 hours after substrate application suggesting that 1251-fibrin was intracellularly degraded followed by release of the fragments. The 30 kDa-component promoted internalization, while factor XIIIa improved the capacity. Thrombin inhibitors suppressed the uptake.

Key Words: Abbreviations:

Platelet, thrombocyte, fibrin, factor XIII, putrescine. BSA, bovine serum albumin; BSS, Hanks’ balanced salt solution; factor XIIIa, plasma transglutaminase; PPACK, D-phenylalanyl-L-prolyl-L-arginyl-chloromethylketone. 175

176

SOLUBLE FIBRIN AND PLATELETS

Vol. 58, No. 2

Soluble fibrin is rapidly removed from circulatin blood. Miiller-Berghaus et al. (1) determined a half-life of ca. 12 hours for l3 igI-fibrin injected i.v. into rabbits, whereas 1251-fibrinogen was eliminated with a half-life of ca. 46 hours. Evidently, there is a special system for cellular recognition of soluble fibrin possibly followed by internalization and intracellular degradation. Cells capable of taking up fibrin are not yet known. Sessile cells of the reticula-endothelial system as well as mobile particles circulating in blood may be involved. A cellular system recognizing soluble fibrin but not fibrinogen was for the first time described on macrophages (2). It required as a cofactor a free N-terminal 30 kDa-domain of fibronectin as well as cell-attached transamidase (3). Apparently, this macrophage-derived ligase represented the receptor, as polystyrene beads coated with plasma transamidase (coagulation factor XIIIa) bound 12%-fibrin under the same conditions as macrophages (4). Intact fibronectin was ineffective as cofactor. There is evidence that a compound related to free N-terminal 30 kDa-domain is circulating in blood although a quantitative determination of this factor is still difficult (5) mainly due to the weak immunogenicity of that domain. Binding of soluble fibrin was also studied with thrombocytes which had been depleted from adherent plasma proteins b{ gelfiltration. On those cells the free fibronectin domain mediated binding of I2 I-fibrin as well (6). Again, there was evidence for the involvement of a cell-attached transamidase whose origin was not yet clarified. Thrombocytes contain plenty of transamidase in their cytosolic compartment; their surface, however, should be devoid of it (7). The present investigation demonstrates on gelfiltered platelets that plasma transamidase is to be considered as a second cofactor besides the 30 kDa-component for the binding of fibrin. Thrombocytes, however, which only had been separated from a concentrate by centrifugation and which still contained a halo were found to bind fibrin still by another of associated plasma proteins, mechanism in addition to the above-mentioned one. Finally, in a thrombocyte concentrate platelets even internalized soluble fibrin.

TF.R.IA1.S AND METHODS Thrombocyte concentrate was obtained from the Bavarian Red Cross where it was prepared separately by differential centrifugation of a single donation of titrated human blood drawn immediately before. After standing over night at room temperature a small sediment of contaminating erythrocytes was removed. Centrifuged platelets were prepared by 2 min. spinning at lO.OOOxg and were resuspended in Hanks’ balanced salt solution (BSS) containing 0.5 9% bovine serum albumin (BSA) to a concentration of ca. log platelets per ml. Gelfiltered platelets were separated from a concentrate according to Thorsen et al. (8) by passing ca. 6 ml (5-8~10~ cells) over a column of Sepharose CL-2B (Pharmacia, 14x180 mm) as described previously (6). In the eluate platelet concentration was determined from the turbidity measured at 450 nm using Elcm = 0.4 for 108 platelets per ml. This extinction coefficient was also found applicable for counting centrifuged platelets or cells in a thrombocyte concentrate. Fibrinogen KABI was depleted of fibronectin by passage through gelatinSepharose and was purified by gelfiltration over Sephacryl S-300 (Pharmacia). It was labelled with 1251 by the Iodogen method (9) yielding ca. 200 cpm/ng and

Vol. 58, No. 2

SOLUBLE FIBRIN AND PLATELETS

177

in was converted with thrombin to 1251-fibrin. The separated clot was dissolved 1 M KBr, 0.05 M Tris/HCl, pH 5.3, containing 0.025 M E-aminocaproic acid and 10 NIH U/ml hirudin (Sigma) to a concentration of ca. 1 mg/ml. Before use the solution was diluted 1:5000 by BSS containing 0.5 % BSA (BSA/BSS). Free fibronectin 30 kDa-domain was isolated as described (3) from a trypsin digest of fibronectin obtained from human plasma by the method of Miekka et al. (10). The digest was passed over gelatin-Sepharose followed by adsorption of the fragment on heparin-Sepharose. After elution it was purified by gelfiltration over Ultrogel AcA44 (Serva). Its concentration was determined spectrophotometrically at 280 nm using E:F’ = 21 (5). Factor XIII (Fibrogrammin, Behring, 20 U/ml) was activated with 10 NIH U/ml thrombin (Calbiochem) for 1 hour in 0.01 M CaC12, 0.1 M NaCl, 0.05 M Tris/HCl, pH 7.4, at 37’C. Residual thrombin was inhibited by 3.3 pg PPACK (5 min., 37’C). For preincubation 1.3 ml activated factor XIII was added to 6 ml thrombocyte concentrate and stored for 1 hour at 37’C. Subsequently, platelets were separated by gelfiltration. Binding studies were performed in polystyrene tubes (63x11 mm, Sarstedt). 0.05 ml platelet suspension or thrombocyte concentrate, respectively, containing ca. 5x10’ cells was mixed with a solution of 30 kDa-fragment and additives in 0.1 ml BSA/BSS. 200 ng 1251-fibrin in 0.1 ml BSA/BSS was added and the volume was completed to 0.5 ml. After incubation for the time indicated platelets were spun down (2 min., lO.OOOxg), washed with 0.6 ml BSA/BSS and bound radioactivity was determined in a Beckman counter Gamma 4000. Controls run in the absence of cells to measure radioactivity adsorbed to the vessel walls were subtracted from the experimental data. Internalized 1251-fibrin was determined after 1 hour incubation of the platelets with trypsin and chymotrypsin (Sigma, 100 p.g/ml each) at room temperature followed by centrifugation and washing. Each value within a single experiment was determined in duplicate and generally the two data were in close proximity. Results obtained from different platelet preparations may scatter to a much higher extent. Nevertheless, the trend of the data between different preparations remains the same. Therefore, in the tables and figures representative experiments are given.

In order to demonstrate the role of transamidases in the platelet-fibrin interaction, a thrombocyte concentrate was preincubated for 1 hour with thrombin-activated factor XIII containing PPACK to neutralize abundant thrombin. Subsequently, the thrombocytes were isolated by gelfiltration and binding of 1251-fibrin was monitored over 16 hours in presence of free 30 kDa fibronectin domain. For comparison, a platelet concentrate preincubated with buffer was treated in the same way. The factor XIIIa-pretreated platelets effectively retained 1251-fibrin without any lag phase (Fig. 1). In contrast, the buffer preincubated platelets bound little 1251-fibrin at least up to 16 hours. Evidently, plasma transamidase has to be considered as an essential cofactor for platelet-fibrin interaction in addition to the 30 kDa-component. In the absence of that free domain gelfiltered platelets failed to bind fibrin even after factor XIIIa pretreatment (Table 1).

SOLUBLE FIBRIN AND PLATELETS

178

Vol. 58, No. 2

TABLE 1

Pretreatment ‘*?-Fibrin (400 Binding of thrombocytes ng/ml) to pretreated with factor XIIIa or buffer followed by gelfiltration. lo8 platelets per ml, 16 h, 2O’C.

12SI-Fibrin bound (%)

30 kDa CLg/ml

FXIIIa

67.4 1.2

30 _-

30 4::

FIG. 1

60-

624

Ti&i~

Influence of factor XIIIa on ‘*9fibrin binding to gelfiltered thrombocytes mediated by 30 kDa-component. o platelets incubated with factor XIIIa followed by gelfiltration, . control platelets incubated with BSA/ BSS and gelfiltered. Binding assay: 400 ng ‘*‘I-fibrin, 20 pg 30 kDa-compound, lo8 cells per ml, 2O’C.

1’6

[hl

FIG. 2 Putrescine inhibition of 1251-fibrin binding to platelets in presence of 30 kDa-component (20 pg/ml). Platelets were isolated by centrifugation (0) or by gelfiltration (0 ) , respectively. In case of gelfiltration the thrombocyte concentrate had been preincubated with factor XIIIa. x Centrifuged platelets in the absence of 30 kDa-component (400 ng ‘251-fibrin, lo8 cells per ml, 18 hours, 2O’C). 0

1.25

5

20

Putrescino

75 [J&I]

300

Vol. 58, No. 2

SOLUBLE FIBRIN AND PLATELETS

179

In another experiment platelets were separated from a thrombocyte concentrate by centrifugation. These cells which still contained a halo of plasma the 30 kDa-domain and variable proteins, were incubated with 1251-fibrin, amounts of putrescine functioning as a competitor of transamidase reactions (11,12). For comparison gelfiltered platelets which had been pretreated by factor XIIIa as outlined above, were subjected to the same procedure. The gelfiltered platelets retained lz51-fibrin and the reaction was completely inhibited by 0.3 mM putrescine indicating that binding exclusively took place by cell-attached factor XIIIa (Fig. 2). Compared to the gelfiltered platelets, the centrifuged ones inhibited bound an elevated amount of 12sI-fibrin but binding was only partially by putrescine. The putrescine sensitive fraction most likely accounts for binding by transamidase which might have been picked up by the cells from their environment and activated by cell-associated proteases. Possibly, it had been released from platelets destructed during centrifugation. Importantly, however, as only partial inhibition was observed with putrescine, one has to conclude that platelets with associated plasma proteins were able to interact with fibrin still by another mechanism independent on surface-attached transamidase. Also, if the 30 kDacomponent was omitted, centrifuged platelets bound 1251-fibrin but only to the extent corresponding to the putrescine-insensitive fraction (Fig. 2). Separate experiments proofed that putrescine had little effect on the binding of 1251-fibrin by centrifuged platelets in the absence of the 30 kDa-domain (not shown). on of 1251-Fibrin bv a Thrombocvte Conceatrate, If a thrombocyte concentrate containing the platelets still in their entire plasmatic environment was incubated with 1251-fibrin, internalization of the labeled substrate was observed. In this system exhibiting coagulation in an uncontrolled way it was not possible to record cell-binding. On the other hand, internalized 1251-fibrin was determined as radioactivity which remained cell-associated after digestion of the centrifuged pellet with trypsin/chymotrypsin. Protease-resistant activity was only detected within a limited period and showed a lag of 2 hours . Depending on the cell preparation it reached a maximum after about 4-6 hours and declined afterwards. Representative experiments are given in Fig. 3 and in the control curve of Fig. 4a. Additional internalization data determined after 4 hours may be cell-associated taken from Table 2. The observation of transient tr psin-resistant l2 r I-fibrin was degraded in inradioactivity might indicate that internalized tracellular compartments followed by release of the radioactive fragments. TABLE Internalization of ‘251-fibrin cells/ml, 4 hours) Additive

by a thrombocyte

per ml

__

-_

PPACK Hirudin Antithrombin Heparin

0.2 ug 1u 100 ug 20 pg

_-

III

2

__

concentrate

Temperature

(ca.

1251-Fibrin internalized

2O’C

15.0 < 0.1 < 0.1 2.9 < 0.1

4’C 37-c

0.6 14.0

lo8

(%)

SOLUBLE FIBRIN AND PLATELETS

180

Vol. 58, No. 2

Internalization of 1251-fibrin by a thrombocyte concentrate was completely suppressed by thrombin inhibitors including PPACK, hirudin or antithrombin III/heparin (Table 2). Antithrombin III for its own was only partially effective while heparin was also inhibitory in the absence of additives. Most likely, in the thrombocyte concentrate sufficient antithrombin III was available to induce a complete inhibition in presence of heparin. The results suggest that thrombin or related enzymes should be considered as essential cofactors for the internalization of fibrin by thrombocytes. Internalization of 12sI-fibrin was suppressed at

FIG. 3 Internalization of ‘251-fibrin by a thrombocyte concentrate (ca. lo* cells per ml, 2O’C). Each point represents duplicate determinations with results lying closely together. PI

r

2

4 Time

6

8

16

[hl -F XIIIa

30

FIG. 4 Internalization of ‘*‘I-fibrin by a thrombocyte concentrate in presence (0) or absence (. ) of 30 kDa-component. The thrombocyte concentrate (ca. lo* cells per ml) was preincubated (1 h, 2O’C) with BSA/BSS (upper panel) or with 0.4 units/ml factor XIIIa (lower panel) followed by addition of ‘251-fibrin (400 ng/ml) with or without 30 kDacomponent (30 kg/ml).

10

*a

Time

[h]

Vol. 58, No. 2

SOLUBLE FIBRIN AND PLATELETS

Little difference of engulfment from 2O’C to 37’C (Table 2).

4’C.

was observed

if the temperature

181

was raised

The following experiments were done to evaluate the role of factor XIIIa and the free fibronectin 30 kDa-domain in platelet-fibrin internalization. Fig. 4a shows the time course of 1251-fibrin internalization by a thrombocyte concentrate without any addition of factor XIIIa in presence or absence of the 30 kDa-component. In the absence of added cofactors internalization began after a lag phase of two hours and proceeded moderately up to a maximum after 6 hours. In contrast, the samples supplied with the 30 kDa-domain immediately started internalization without any lag phase giving rise to a much higher uptake than in the absence of the cofactor. The biphasic time course of the reaction most likely was due to a heterogeneity of the platelet preparation. Sole addition of factor XIIIa also improved the uptake but failed to influence the lag phase (Fig. 4b). This lag again was overcome by addition of the 30 kDa-domain to the factor XIIIa containing system without further elevation of total uptake. The data indicate that factor XIIIa as well as the 30 kDa-component improve and promote the internalization of fibrin by thrombocytes. However, it cannot be excluded that also other pathways are available for the uptake.

SSION A selective elimination of soluble fibrin from circulating blood requires a cellular recognition system discriminating between fibrin and fibrinogen. Originally, a system of this kind was detected on peritoneal macrophages of guinea pigs (2). It was dependent on a free N-terminal 30 kDa fibronectin domain and was inhibited by putrescine or histamine suggesting the involvement of a cell-associated transamidase. The presence of this enzyme on macrophages was confirmed (3). Cell binding of fibrin turned out to be a reaction of moderate rate controlled by the amount of cell-associated transamidase (3). It was, therefore, unlikely that fibrin is retained by sessile cells during the short contact when blood is passing the liver or other organs. Rather, fibrin binding should take place within a longer contact with mobile target cells circulating in blood. Platelets appeared as the most favored ones, especially as they show qualitative and quantitative alterations in latent and manifested blood coagulation. In a previous paper (6) it was reported that gelfiltered platelets bind 1251-fibrin in a reaction dependent on the free 30 kDa-component. The present paper documents that this reaction also requires coating of the platelets by factor XIIIa which, therefore, should be considered as an additional cofactor. Possibly, in the previous experiments the platelets had adsorbed factor XIIIa originating from the cytosolic compartment of dying cells. Now, the experiments demonstrate that gelfiltered platelets fail to bind 12%-fibrin unless they had been treated with factor XIIIa. As this incubation took place prior to gelfiltration, one might assume that factor XIIIa is firmly bound by a platelet receptor. Binding of 1251-fibrin to gelfiltered platelets mediated by the two cofactors 30 kDa-component and factor XIIIa was completely inhibited by putrescine. In contrast, under the same conditions binding to centrifuged platelets still associated with plasma proteins was only partially inhibited suggesting that the halo contains so far unknown factors capable of mediating fibrin binding by still another mechanism. Consequently, at least two pathways should be available for

182

SOLUBLE FIBRIN AND PLATELETS

Vol. 58, No. 2

fibrin-binding to unstimulated platelets. Preliminary data with platelets from different donors give evidence that each of the two pathways may be utilized by a single platelet preparation to a different extent. Further experiments are necessary to clarify which conditions determine the relationship of the two pathways and to characterize all plasmatic factors involved in the binding mechanisms. Finally, a thrombocyte concentrate containing platelets in their plasmatic environment was able to internalize 1251-fibrin by the cells. Gelfiltered or centrifuged platelets, in contrast, failed to engulf this substrate. Evidently, this reaction requires additional plasma components which were insufficiently retained by centrifuged platelets and were diluted after resuspension. Internalized radioactivity was only observable within a limited period between ca. 4-6 hours giving evidence for an intracellular degradation followed by release of the fragments. Most likely, internalization was dependent on traces of bin inhibitors completely abolished the uptake of 1251-fibrin. tin 30 kDa-domain eliminated the lag phase of engulfment, improved the uptake. Further experiments are necessary to lag phase is required by the platelets to release 30 kDa-domain conditions.

thrombin as thromThe free fibronecwhile factor XIIIa clarify whether the under the reaction

Experiments with a platelet concentrate are hampered as many samples undergoe coagulation. Indeed, the best internalization results were obtained with samples showing only little or no coagulation probably due to a limited degradation of their fibrinogen. Completely clotted preparations failed to internalize 1251-fibrin unless an antibody recognizing the membrane glycoprotein IIb/IIIacomplex was added (to be published in more detail). Perhaps, platelets interact with polymerized fibrin or fibrils in another way than with the soluble form possibly by involvement of the GP IIb/IIIa-complex and not resulting in internalization. Interactions of that kind taking place on activated platelets and probably competing with the sequence of binding and internalization are investigated by Hantgan et al. (13) and by Hantgan (14,15).

We are indebted to Mrs. V. Legner for technical is also extended to the Deutsche Forschungsgemeinschaft, reich 207, for supporting the investigations.

assistance. Our gratitude Sonderforschungsbe-

1.

MULLER-BERGHAUS, G., MAHN, I., KGVEKER, G. and MAUL, F.D. In vivo behaviour of homologous urea-soluble 1311-fibrin and 1251-fibrinogen in rabbits. The effect of fibrinolysis inhibition. Brit. S, 61-79, 1976.

2.

HGRMANN, H., RICHTER, H. and JELINIC, V. Fibrinmonomer binding to macrophages mediated by fibrin-binding fibronectin fragments. Thromb, Res. 39, 183-194, 1985.

3.

HGRMANN, H., RICHTER, H. and JELINIC, V. The role of fibronectin fragments and cell-attached transamidase on the binding of soluble fibrin to macrophages. Thromb. Res. 46, 39-50, 1987.

Vol. 58, No. 2

SOLUBLE FIBRIN AND PLATELETS

183

4.

HGRMANN, H., RICHTER, H., JELINIC, V. and WENDT, C. N-terminal fibronectin 30-kDa fragment mediates the immobilization of soluble fibrin by factor XIIIa-coated polystyrene beads. Biol.Chem.-Seyb m, 669-674, 1987.

5.

SKRHA, J., RICHTER, H. and HGRMANN, H.: Evidence for the presence of a free N-terminal fibronectin 30-kDa domain in human plasma byquantitaive determination with an indirect immunosorbent-assay. Anal. u, 228-234, 1988.

6.

HGRMANN,H. 30kDa-domain thrombocytes.

7.

LOPACIUK, S., LOVETTE, K.M., McDONAGH, J., CHUANG, H.Y .K. and McDONAGH, R.P. Subcellular distribution of fibrinogen and factor XIII in &, 453-465, 1976. human blood platelets. Thromb.

8.

THORSEN, L.I., BROSSAD, F., GOGSTAD,G., SLETTEN, N. 0. Competitions between fibrinogen with its degradation teractions with the platelet-fibrinogen receptor. Thromb. 1986.

9.

NIEUWENHUIZEN, W., EMELS, J.J., VERMOND, A., KURVER, P. and VAN DER HEIDE, D. Studies on the catabolism and distribution of fibrinogen in rats. Application of the Iodogen labelling technique. Biochem. Bio Dhvs. s, 49-55, 1980.

, RICHTER,H. and JELINIC,V. Free N-terminal mediates binding of soluble fibrin to gelfiltered brnb. w fi, 283-293, 1988.

fibronectin unstimulated

K. and SOLUM, products for in44, 61 l-623 ,

10. MIEKKA, S.I., INGHAM, K.C. and MENACHE, D. Rapid methods m. Res,z, l-14, 1982. lation of human plasma fibronectin.

for iso-

R., JACOBSEN, 11. LORAND, L., RULE, N.G., ONG, H.H., FURLANETTO, A., DOWNEY, J., GNER, N. and BRUNER-LORAND, J. Amine specificity in transpeptidation. Inhibition of fibrin cross-linking. B 2, 12141223, 1968. 12. NILSSON, LUNDEN, inhibitors.

J.L.G., STENBERG, P., LJUNGGREN, C., HOFFMAN, R., ERIKSSON, 0. and LORAND, L. Fibrin-stabilizing &Q, N. Y. AEad,sr;i2Q2, 286-296, 1972.

13. HANTGAN, R.R., TAYLOR, R.G. and LEWIS, fibrin only after activation. -65,1299-1311,

J.C. Platelets 1985.

K.J., factor

interact

with

and fibrinogen binding to ADP-stimula14. HANTGAN, R.R. Fibrin protofibril ted platelets: evidence of a common mechanism. Biochim. BiolZhyS,BEta !&&, 24-35, 1988. 15. HANTGAN, to platelets.

R.R. Localization Biochim. Biophvs.

of the domains of fibrin involved Acta %&, 36-44, 1988.

in binding