In vivo thrombus formation induced by complement activation on polymer surfaces K. Hayashi," H. Fukumura,t and N. Yamamoto Government Industrial Research Institute, Osaka, 1-8-31, Ikeda, Osaka 563, lapan To clarify involvement of complement activation in thrombus formation on polymer surfaces, in vitro complement activation was evaluated for polyethylene (PE) tubes radiation-graft c o p o l y m e r i z e d w i t h acrylamide (AAm), acrylic acid (AC), 2hydroxyethyl methacrylate (HEMA), Nvinylpyrrolidone (NVP), and vinyl alcohol (VOH), and compared to their in vivo antithrombogenicity and cell adherence in canine peripheral veins. The complementactivating surfaces (NVP and VOH) cause preferential adhesion of leukocytes and were more thrombogenic than the low

complement-activating surfaces (AAm, PE, and HEMA). Infusion of naja haje cobra venom factor depressed leukocyte adhesion, followed by a marked decrease in thrombogenesis, for the strong classicalpathway-activating surface (NVP). Although estimation of in vitro activation for AC was inconclusive because of a large effect of adsorption, AC behaved like VOH in vivo. These results suggest that C5a(des Arg) mediated activation of leukocytes may play a role in thrombus formation by complement activation on polymer surfaces.

INTRODUCTION

The phenomenon of leukocyte adhesion at the interface between blood and materials has often been observed under dynamic condition^.'-^ Although the role of complement activation in cell adhesion was ~uggested,~ the involvement of complement activation in thrombus formation has not been clarified yet. Previously, we studied in vivo antithrombogenicity for polymer surfaces modified by radiation-graft copolymerization in terms of the relative patent time (RPT) in canine peripheral veins,4 and noticed leukocytes preferentially adhering to some of these surfaces. In the present study, we evaluated in vitvo complement activation for several graft-copolymerized polyethylene surfaces, compared it to both RPT and cell adhesion in the peripheral veins of dogs with and without infusion of CVF, and will discuss the relation between complement activation and thrombus formation on polymer surfaces.

*To whom correspondence should be addressed. tPresent address: Kyoto Institute of Technology, Matsugasaki, Kyoto 606, Japan. Journal of Biomedical Materials Research, Vol. 24, 1385-1395 (1990) CCC 0021-9304/90/101385-11$04.00 0 1990 John Wiley & Sons, Inc.

HAYASHI, FUKUMURA, AND YAMAMOTO

1386

EXPERIMENTALS

Tubular specimens Low-density polyethylene (PE) tubing supplied from Sumitomo Electric Industries, Ltd, (Sumikathane C215, Sumitomo chemicals, Ltd.), 3 mm I.D. and 0.25 mm thick, was cut into 17-mm lengths, washed several times with watedethanol mixtures as the ethanol concentration was increased gradually, and dried in air. The specimens of PE tubes graft-copolymerized with acrylamide (AAm), acrylic acid (AC), 2-hydroxyethyl methacrylate (HEMA), and N-vinylpyrrolidone (NVP) were prepared by a radiation oxidation method with y-rays. The specimens grafted with poly(viny1 alcohol) (VOH) were prepared by mutual radiation-graft copolymerization with vinyl acetate in bulk at room temperature and by subsequent alkaline hydrolysis.' The PE tube of 1 mm I.D. was cut into 10-cm lengths and graft-copolymerized in the same way. A 3-mm specimen of original PE and a 1-mm specimen of glass capillary were used as a standard for the evaluations of in vivo antithrombogenicity and in vitro complement activation, respectively.

In vitro complement activation The complement activation was evaluated from the residual hemolytic activity of the serum after contact with the material surfaces. Canine serum was used in consideration of the in vivo experiments using dogs. The tubular specimens of 1 mm I.D. were used to increase the area contact with serum. Ten centimeters of the 1-mm specimen give about 80 pL of capacity and about 40 cmZ/mLof the ratio of contact area to serum volume. The specimen, covered with Tygon S50-HL tube (Norton Company) when the specimen was hydrophilic, was bent in a U-shape and primed with gelatin veronal buffer containing 0.15 mM Ca" and 0.5 mM Mf (GVB, pH 7.4) by a syringe. One hour after, GVB was substituted with serum by injection and overflow of 1 mL of serum from one side of the U-tube. Three kinds of sera were used: (1) normal serum, (2) about 90%-diluted serum containing 5 mM Mg2' and 10 mM ethylene glycol-bis (P-aminoethyl ether)-N,N,N,N,-tetraacetic acid (EGTA), and (3) the diluted serum containing 10 mM ethylenediamine tetraacetic acid (EDTA). After incubation for 2 h at the canine biological temperature of 39"C, the reaction was quenched by cooling to 0°C. The reason for adoption of the longer incubation time of 2 h was in consideration of the low contact area of tubular samples. The decrease in hemolytic activities of sera after 1h incubation were generally lower than those of 1.5 h incubation in hollow fibers.6 Complement titer (CH,,) was assayed by the method of M a ~ e rand , ~ the hemolytic activity of the fourth component of complement (C4) was determined by that of Gaither et al.' The percentage decrease in hemolytic activity was calculated by assuming the hemolytic activity of the control (glass) as 100%.

COMPLEMENT-INDUCED THROMBUS FORMATION

1387

Zn vivo thrombus formation and cell adhesion Four tubular specimens of 3 mm I.D. and 17 mm long, containing one control (PE) principally, were inserted one by one into four peripheral veins of bilateral external jugular and femoral veins of a dog anesthetized by intravenous injection of Somnopentyl (Pitman-Moore, Inc.). Patent time (PT) till the blood flow was stopped by thrombus formation was obtained by monitoring with a Doppler blood flow meter (Directional Doppler, Parks Electric Co.). RPT was then calculated according to the following equation: RPT(h) = K(PT),/(PT), in which (PT), and (PT), were the PT of sample and control, respectively, and K was a statistical patent time (h) of the control, which is 2 h for PE. For the examination of cell adhesion, some specimens were excised after less than 1 h implantation, dehydrated with acetonejwater, critical-point dried, and coated with gold. Their luminal surfaces were observed by scanning electron microscopy (SEM). Twelve dogs weighing about 10 kg were used in total. Eight of them were intravenously preinjected with 60-70 units/kg of naja haje CVF (Cordis Laboratories, Inc.) about 20 h before operation. Decreases in CH,, of their sera were confirmed before operation. The RPT for AAm, NVP, and VOH was the average two or three measurements with the standard deviation of 0.46.

Contact angle of water Contact angle of water was measured by the sessile droplet technique in which a drop of water was put on the tubular specimen in the chamber of a Kyowa contact angle meter CA-A at 25°C.

RESULTS

In vitro complement activation Figures 1, 2, and 3 show the plots of decreases in residual hemolytic activities vs. grafting degrees for tubular specimens. In these figures, the value at zero grafting is that for the original PE. The decreases in CH,, of normal serum for AAm and HEMA are as small as that for PE over all their grafting degrees (Figs. l(a), (b)’) but those for NVP, VOH, and AC depend on their grafting degrees (Figs. 2 and 3). Moreover, NVP decreases the hemolytic activity of C4 much more than the CHS0of Mg-EGTA serum (Figure 2(a)9), while VOH reduces the CH,, of Mg-EGTA serum predominantly (Figure 2(b)). AC reduces both the C4 activity and CH, of Mg-EGTA (Fig. 3). Decreases in CH, of EDTA-serum are almost zero for all samples except for AC (Figs. 2 and 3).

HAYASHI, FUKUMURA, AND YAMAMOTO

1388 A

6\0

l:l,

._ c :._ > lo\

:W%

0

$< &%

E

0

,~

-10

$

0

,

,~

10

Grafting

o .

-10

30

20

10

0

(X)

20

Grafting

(4

,

30

40

(X)

(b)

Figure 1. Plots of decreases in complement titer (CH,) for the canine normal serum after incubation inside tubular specimens of AAm (a) and HEMA (b)' vs. their grafting degrees.

x

.->

.t;

60

30

40

20

.& 20

10

0

.-0 4-J

x

E

c Q,

vl 0

W

t o Q,

n

0 0

10

20

Grafting

(a)

(X)

30

0

10

20

Grafting

30

(X)

(b)

Figure 2. Plots of decreases in the hemolytic activities vs. grafting degrees: 0,CH3 for normal serum; A, hemolytic activity of C4 for normal serum; 0, CH, for Mg2'-EGTA serum; 0, CHSofor EDTA serum after incubation inside tubular specimens of NVP (a)' and VOH (b).

In vivo antithrombogenicity The results of APT for representative samples are shown in Table I, together with their grafting degrees and contact angles of water. In normal dogs, the RPT for the complement-activating surfaces is shorter than that for the low complement-activating ones. The averages of CH,, for sera of dogs before and 20 h after CVF injection were 40 and 4, respectively. As seen in Figure 4, the decomplementation by CVF causes such a large elongation of RPT for NVP that it becomes almost equal to that for PE (2 h), though that for VOH or AC is not so elongated. AC shows the shortest RPT in both normal and decomplemented dogs.

1389

COMPLEMENT-INDUCED THROMBUS FORMATION 100

-20

r

L 0

10

20

Grafting

30

40

(X)

Figure 3. Plots of decreases in the hemolytic activities vs. grafting degrees: 0,CHm for normal serum; A, hemolytic activity of C4 for normal serum; 0, CH50for Mf-EGTA serum; 0, CHS0for EDTA serum after incubation inside tubular specimens of AC. TABLE I Grafting Degree, Contact Angle of Water at 25”C, and Relative Patent Time (RPT) in the Peripheral Veins of Normal and Decomplemented Dogs for Original Polyethylene (PE) and Grafted Ones (AAm, HEMA, NVP, VOH, and AC) Grafting

RpT

Contact Angle

(h)

Sample

@.)

(“I

Normal

Decomplemented

PE

0 28.4 14.8 29.2

101 26 63 74 80 70

2 2.9 1.8

3.3 2.0

AAm HEMA NVP VOH AC

21.4

19.8

1.0 0.7 0.4

2

2.0 1.1 0.8

In v i m cell adhesion In the peripheral veins of a normal dog, only platelets adhered to the low activating surfaces like PE and HEMA (Fig. 5(A)) but leukocytes preferentially adhered to the activating surfaces of NVP (Fig. 5(B)), VOH (Fig. 5(C)), and AC (Fig. 5(D)). The surfaces of AAm showed very little adherence of blood cells. The decomplementation by CVF infusion turned distinctly leukocyte adhesion into platelet adhesion on NVP surfaces (Fig. 5(F)),

1390

HAYASHI, FUKUMURA, AND YAMAMOTO

Surfaces of grafted PE Figure 4. RPT obtained by implantation into peripheral veins of normal (solid) and decomplemented (shaded) dogs for various tubular specimens. Refer to the specimens in Table I.

though it hardly affected cell adhesion on the other surfaces (Figs. 5(E), (G), and (H)). DISCUSSION

Complement activation by polymer surfaces The data for the decrease in CH,, in normal serum, which means whole complement activation, show that PE, AAm, and HEMA hardly activate complement but NVP and VOH clearly do (Figs. 1 and 2). The results of both C4 and CH, of Mg2+-EGTAserum assigns the complement activation by NVP to a classical pathway and that by VOH to an alternative pathway (Fig. 2). The decrease in CH, of EDTA serum, namely the decrease by adsorption, is so large for AC that it is impossible to estimate the complement activation so far (Fig. 3). However, the decrease by adsorption is almost negligible for the other specimens (Fig. 2). The fact that complement activation depends on grafting degrees (Figs. 2 and 3) indicates the importance of surface concentration of functional groups for the activation. Activation by HEMA surfaces is almost zero in our experiments (Fig. l(b)), although HEMA-homopolymer and its copolymers with ethyl methacrylate (below 60%) have been reported to activate

COMPLEMENT-INDUCED THROMBUS FORMATION

1391

Figure 5. SEMs of luminal surfaces of tubular specimens of HEMA (A, E), NVP (8,F), VOH (C, G), and AC (D, H) after short implantation (1530 min) in peripheral veins of normal (A-D) and CVF-decomplemented (E-H) dogs. Bar indicates 10 pm.

complement.lo This difference may be due to the low concentration of functional groups on the surface, resulting from the large affinities of monomer and graft chains for the hydrophobic substrate.

1392

HAYASHI, FUKUMURA, AND YAMAMOTO

Relation between complement activation and initial cell adhesion The mechanism of complement-induced leukocyte adhesion on material surfaces have been well presumed." Immune adhesion of leukocytes is caused by the interaction of complement fragments bound on polymer surfaces with their receptors CR1 (for C4b, C3b) and CR3 (for C3bi) on leukocyte membrane. However, in canine blood, platelets as well as leukocytes can adhere to complement-activating surfaces because the platelets of nonprimate species have also been reported to have CR1 in contrast to human platelets.12Although both cell adhesions have been observed in the stagnation point flow m e t h ~ dpreferential ,~ adhesion of leukocytes occurred on complement-activating surfaces in the present experiments using the peripheral veins (Figs. 5(B), (C), (D)). Leukocytes could hardly be found to adhere under static condition^.^ The difference in blood flow appears to affect leukocyte adhesion behavior. These phenomena suggest that the diffusible complement fragments C5a and C5adesArg, which can enhance expression of CR1 and CR3 on leukocytes by binding their specific receptors on leukoc y t e ~have , ~ ~ an important role in the preferential adhesion of leukocytes. Since CVF has been confirmed not to cause any substantial reduction in cell numbers in the change in cell adhesion observed for NVP is concluded to be due to the effect of decomplementation. Naja haje CVF depletes the C3 level selectively to 2% of the original amount14by formation of a potent C3 convertase, CVFBb complex.15The residual level of C3 may be too low to produce classical pathway activation but may still be available to produce C5a and C3b by alternative pathway involving an amplification circuit of C3 activation by C3bBbP. The platelet adhesion observed in decomplemented dogs for NVP is considered to result from the fact that the classical pathway activating surfaces also have a strong capacity for platelet adhesion because their surfaces have medium hydrophilicity with a water-contact angle near 70" (Table I).l6 Complement activation as well as cell adhesion is low for the AAm surface. This may be due to the fact that the AAm surface is so hydrophilic17that it depresses adsorption and denaturation of proteins which can trigger initiation of various biological systems. Further investigations of the activation of human complement components by polymer surfaces and investigations of the effects of blood flow rate on cell adhesion are necessary.

Relation between complement activation and thrombus formation Our accumulation of in vivo evaluation of antithrombogenicity by RPT has confirmed that there is good reproducibility in the order of PT among four specimens implanted into the peripheral veins of a dog and that the use of a control permits eliminating the individual differences in dogs and results in reduction of the number of sacrificed dogs.4

COMPLEMENT-INDUCEDTHROMBUS FORMATION

1393

Thrombus formation

6

Complement system

A

Figure 6. Schematic diagram of the process of complement-induced thrombus formation on material surfaces.

Shorter RPTs obtained in normal dogs for complement-activating surfaces mean that they are more thrombogenic than low activating surfaces (Fig. 4). Moreover, the largest elongation of RPT (Fig. 4) resulted when leukocyte adhesion was inhibited by decomplementation (Fig. S(F)). These facts suggest an involvement of the initial adhesion of leukocytes in promotion of thrombus formation. C5a is known to induce the generation of both tissue factor" and plateletactivating factor'' in leukocytes, which can activate coagulation system and platelets, respectively. Besides this, platelet activation by both the C3a and C5b-9 complex and enhancement of platelet-prothrombinase activity by the CSb-9 complexz2may also promote thrombus formation. The process from complement activation to thrombus formation on the material surfaces is diagrammed in Figure 6. AC showed leukocyte adhesion and the shortest RPT in both normal and decomplemented dogs. Since such anionic surfaces as AC are known to activate contact both classical and alternative pathway activation via plasmin in the lysis system initiated by kallikrein and fibrin formationz4is also undeniable for AC. References 1. C. L. Vankampen, D. F. Gibbons, and R. D. Jones, "Effect of implant surface chemistry upon arterial thrombosis," J, Biomed. Mater. Res., 13, 517-541 (1979). 2. K. Kottke-Marchant, J. M. Anderson, K. M. Miller, R. E. Marchant, and H. Lazarus, "Vascular graft-associated complement activation and leukocyte adhesion in an artificial circulation," J. Biomed. Mater. Res., 21, 379-397 (1987). 3. G. A, Herzlinger and R. D. Cummimg, "Role of complement activation in cell adhesion to polymer blood contact surfaces," Trans. A m . SOC.Artif. Intern. Organs, 26, 165-171 (1980). 4. K. Hayashi, N. Yamamoto, and I. Yamashita, "Radiation-graft copolymerization of hydrophilic monomers onto hydrophobic substrates and

HAYASHI, FUKUMURA, AND YAMAMOTO

1394

5.

6. 7. 8.

9.

10. 11. 12. 13.

14. 15.

16. 17.

18.

19.

20. 21. 22.

antithrombogenicity of the surfaces resulted,” J. Jpn. SOC.Biomaterials, 1 (2), 59-65 (1983). K. Hayashi, K. Murata, N. Yamamoto, and I. Yamashita, ”Antithrombogenicity of polyethylene radiation-graft copolymerized with hydrophilic monomers,” Kobunshi Ronbunsku, 42, 77-83 (1985). H. Fukumura, S. Yoshikawa, and M. Miya, ”In vitro complement activation by polymeric materials,“ Proceedings of 14th Medical Polymer Symposium, Society of Polymer Science, Japan, 1985, p. 75. M. M. Mayer, Experimental Immunology, E. A. Kobat and M. M. Mayer (eds.), Charles C. Thomas Publishers, Springfield, IL, 1961, p. 133. T. A. Gaither, D. W. Alling, and M. M. Frank, “A new one-step method for the functional assay of the fourth component (C4) of human and guinea pig complement,” J. Immunol., 113, 574-583 (1974). H. Fukumura, K. Hayashi, S. Yoshikawa, M. Miya, N. Yamamoto, and I. Yamashita, ”Complement-induced thrombus formation on the surface of poly(N-vinylpyrro1idone)-graftedpolyethylene,” Biomaterials, 8, 74-76 (1986). M. S. Payne and T. A. Horbett, “Complement activation by hydroxyethyl methacrylate-ethyl methacrylate copolymers,” J. Rimed. .Muter. Res., 21, 843-859 (1987). M. D. Kazattchkine and M. P. Carreno, ”Activation of the complement system at the interface between blood and artificial surfaces,” Biamaterials, 9, 30-35 (1988). C. Bianco and V. Nussenzweig, ”Complement receptor,” Contemp. Top. Mol. Irrrmunol., 6, 145-176 (1977). M. Berger, J. O’Shea, A . S. Cross, T. M. Folks, T.M. Chused, E . J . Brown, and M. M. Frank, “Human neutrophils increase expression of C3bi a5 well as C3b receptors upon activation,” J. CIirz. Invest., 128, 2547-2552 (1982). R. D. Cumming, ”Important factors affecting initial blood-material interactions,” Trans Am. Soc. Artif. Intern. Organs, 26, 304-308 (1980). I. V. Zabern, B. Hinsch, H. Przyklenk, G. Schmidt, and W. Vogt, ”Comparison of naja n. naja and naja h. haje cobra-venom factors: correlation between binding affinity for the fifth component of complement and mediation of its cleavage,” Immunobiologj, 157, 499-514 (1980). Y. Tamada and Y. Ikada, ”T cell adhesion onto polymer surfaces preadsorbed with various proteins,” Polym. Preprints, Japan, 36, 835 (1987). K. Hayashi, H. Fukumura, N. Yamamoto, and I. Yamashita, ”Effect of preirradiation dose and grafting on the antithrombogenicity of polyethylene radiation-graft copolymerized with acrylamide,” Kobunshi Ronbunshu, 44, 681-688 (1987). T. W. Muhlfilder, J. Niemetz, D. Kleutzer, D. Beebe, P. A. Ward, and S. I. Rosenfeld, “C5 chemotactic fragment induces leukocyte production of tissue factor activity. A link between complement and coagulation,” J. Clin. Invest., 63, 147-150 (1979). J. Benveniste, E. Jouvin, E. Pirotzky, B. Arnoux, J. M. Mencia-Huerta, R. Roubin, and B. B. Vargaftig, ”Platelet-activating factor (PAF-acether): molecular aspects of its release and pharmacological actions,” Int. Archs Allergy Appl. Immun., 66 (Suppl. 1) 121-126 (1981). M. J. Polly and R. L. Nachman, ”Human platelet activation by C3a and C3a des-arg,” J. Exp. Med., 158, 603-615 (1983). T. S. Zimmerman and W. P. Kolb, “Human platelet initiated formation and uptake of the C5-9 complex of human complement,” J. Clin. Inzwt., 57, 203-211 (1976). T. Wiedmer, C. T. Esmon, P. J. Sims, “On the mechanism by which complement proteins C5b-9 increase platelet prothrombinase activity,” I. Biol. Chem., 261, 14587-14592 (1986).

COMPLEMENT-INDUCED THROMBUS FORMATION

1395

23. T. Matsuda, ”Biological responses at bloodimaterial interface & molecular design of biocompatible polymer surfaces,” Jpn. I . Artif. Organs, 16, 1252-1256 (1987). 24. G. Murano, ”Hageman connections: Interrelations of blood coagulation, fibrino(geno)lysis, kinin generation, and complement activation,” Am. J. Haematol; 4, 409-417 (1978). Received January 20, 1989 Accepted May 3, 1990

In vivo thrombus formation induced by complement activation on polymer surfaces.

To clarify involvement of complement activation in thrombus formation on polymer surfaces, in vitro complement activation was evaluated for polyethyle...
603KB Sizes 0 Downloads 0 Views