Evacuation of microscopic air bubbles from Dacron reduces complement activation and platelet aggregation Peter G. Kalman, M D , F R C S ( C ) , Denise A. McCullough, and Charles A. Ward, PhD, Toronto, Ontario, Canada Complement activation by biomaterials may play an important role in vascular graft failure since the physiologically active polypeptides, C3a and C5a, have several relevant properties. C3a promotes platelet aggregation and release, and C5a activates neutrophils, which may stimulate platelet aggregation by liberation of platelet activating factor or by a direct neutrophil platelet interaction. Microscopic air bubbles (nuclei) are found in the surface roughness or pores of most biomaterials, and their number and size are related to the surface tension of the material. Therefore two interfaces can be postulated to exist when Dacron is exposed to blood: (1) a blood/biomaterial, and (2) a blood/air interface. These air nuclei in the surface and the biomaterial itself are capable of activating complement. The purpose of these experiments was to eliminate these surface nuclei from Dacron by a process termed denucleation and subsequently to determine the effect of this intervention on complement activation and platelet aggregation in vitro. Dacron was denucleated by pretreatment that involved serial rinsing with ethanol and degassed buffer that results in replacement of the air nuclei by buffer. Both control and denucleated pieces of Dacron (2, 4, and 6 cm 2) were then incubated in human plasma. Each plasma sample was assayed for complement activation products (C3a, C5a, and C4a) by means ofradioimmunoassays, and the degree of autologous platelet aggregation that resulted from the addition of a portion of each incubated plasma sample to an autologous platelet suspension was measured. There was a significant reduction in C3a and C5a in the plasma samples incubated with denucleated Dacron as compared to control Dacron (p ~ 0.001, analysis of variance [ANOVA]). Also, the plasma incubated with denucleated Dacron caused reduced platelet aggregation as compared to the plasma incubated with control Dacron (p ~ 0.001), ANOVA) when added to a platelet suspension. (J VAse SURG 1990;11:591-8.)
When a biomaterial is exposed to blood, complex events occur that may result in cellular adhesion ( ~htelets, neutrophils) and lead to thrombus formarion. Complement activation by polymer surfaces is known to occur with hemodialysis,~ cardiopulmonary bypass surgery, 2 hemofiltration, 3 after cardiac valve replacement, 4 and with prosthetic vascular grafts, s,6 Whether complement activation proceeds via the classic or alternative pathway, the physiolog-
From the Department of Surgery and Division of Vascular Surgery, Toronto General Hospital, and the Department of MechanicalEngineering,Universityof Toronto. Supported by the Physicians'Services Incorporated Foundation (PSI) of Ontario. Presented at the BreakfastProgram, Societyfor Vascular Surgery and the InternationalSocietyfor CardiovascularSurgeryAnnual Meeting, New York, N.Y. June 20, 1989. Reprint requests: P. G. Kalman,MD, Toronto GeneralHospital, Divisionof VascularSurgery,9 Eaton North - Room 211, 200 Elizabeth St., Toronto, Ontario, M5G 2CA Canada.
z4/6/18048
ically active inflammatory mediators C3a and C5a are released. C3a generation is important since it promotes platelet aggregation and release. 7 C5a is an extremely potent mediator of neutrophil chemotaxis, s aggregation, 9 and adhesion. ~° Activated and adherent neutrophils may contribute to vascular graft thrombosis by augmenting coagulation, n This may occur by stimulation of platelet aggregation through release o f platelet activating factor, ~2 or by a direct neutrophil/platelet interaction. ~3 If prosthetic vascular grafts significantly activate complement locally, then reduction of the activation would be expected to cause a decrease in platelet and neutrophil adhesion and perhaps thrombus formation. Since microscopic air bubbles can become trapped in the surface roughness of a biomaterial when the material is in contact with blood, ~4 two interfaces can be postulated to exist: (1) a blood/biomaterial and (2) a blood/air bubble interface. Elimination o f these surface air bubbles has successfully decreased cellular adhesion to blood contact 591
Journal of VASCULAR SURGERY
592 Kalman, McCullough, and Ward
surfaces, as Furthermore, complement activation by air bubbles has been observed in both rabbit 16 and human a7 plasma. Since microscopic bubbles of air can be expected to be trapped in the surface roughness of Dacron 18 as they are in other materials, ~4'xs,~9 they would be expected to give rise to complement activation when the material is brought in contact with plasma, and plasma incubated with Dacron would be expected to give rise to platelet aggregation. Furthermore, if Dacron is subjected to a procedure to remove the air nuclei before exposure to plasma, complement activation would be expected to be decreased, and if it is complement activation that induces platelet aggregation it would be expected that if the air nuclei were removed before incubation with plasma, then this would result in reduced platelet aggregation. Our results confirm that air bubbles activate complement and augment platelet activation in vitro. Moreover, denucleation of Dacron strongly reduces both complement activation and platelet aggregation. A recent review of the factors controling cell adhesion to a synthetic material suggests that complement activation plays a mediating role, 2° therefore denucleation of Dacron may be a means by which the biocompatibility of this commonly used biomaterial can be improved.
hours at room temperature, and if it is disturbed after this time no bubbles are seen to separate from the surface. The ethanol is decanted, leaving enough to cover the Dacron. Phosphate buffer (pH 7.37, osmolarity 0.297) is degassed for a minimum of 6 hours with a vacuum and magnetic stir bar at room temperature. The Dacron is then slowly and continuously washed for 6 hours with degassed buffer, gradually replacing all ethanol. The specimens were kept in degassed buffer until incubation with plasma. Incubation o f samples
Samples were placed in 2.70 ml capacity selfcapped tubes and were subsequently incubated on a rotator at 22 rpm in a 37 ° C water bath for 30 minutes. 16 Plasma incubated with air bubbles. Plasma (2.5 ml) was placed in a 2.70 ml tube, and after the tube was capped small bubbles were formed by vigor0*'~ly shaking it. The tube was placed on the rotator, and bubbles were maintained during incubation by sharply thumping one end during each rotation by use of a device added to the rotator. Plasma incubated with zymosan. Zymosan activated plasma served as a strong positive control and was used to determine the upper limit of complement activation for each plasma sample. The zymosan MATERIAL AND METHODS (Sigma Chemical, St. Louis, Mo.) was boiled in saCollection o f blood line solution for 15 minutes, cooled, and washed three times, then resuspended in saline solution Eleven healthy male volunteers (age range 19 to (5 mg weight per milliliter). A 2.70 ml tube was 36 years) donated blood after approval by the human filled with 2.50 ml of the zymosan-saline suspension ethics committee at our institution. Polypropylene then centrifuged (1033g, 4 ° C, 10 minutes) leaving syringes and incubation tubes were used for all exa pellet of zymosan. Plasma was added to the pellet periments, since polypropylene only minimally actito form a suspension (5 mg zymosan per milliliter vates complement. One hundred milliliters of blood plasma) and was then gently shaken to ensure adewas collected from each volunteer by venipuncture, quate mixture. The suspension was then incubated 90 rnl in heparin (10 units per milliliter of blood) on the rotator with the other samples, but not and 10 ml in acid citrate dextrose (ACD) (6 parts thumped. blood to 1 part ACD). Plasma was separated in the Plasma exposed to Dacron. Pieces of weaveknit heparinized sample by centrifugation (1033 g, 15 Dacron (Meadox Medicals Inc. Oakland, N.J.) with minutes, 4 ° C). The blood collected in ACD served surface areas of 2, 4, or 6 cm 2 were cut from a sample as a source of platelets for the aggregation experiof a tube of the material (10 mm outer diameter, ments. 8 mm inner diameter). The Dacron pieces to be inDenucleation o f Dacron cubated were placed in tubes, overfilled with plasma The denucleation process was very similar to that without introducing air bubbles, then capped and described by Ward et a1.19The cut pieces of Dacron sealed. Three of the tubes each contained a sample were placed in a glass beaker containing 99% ethanol. • of control Dacron, and three contained denucleated Dacron (2, 4, 6 cm2). All were incubated on the Visible bubbles leave the Dacron rapidly at first, then they slowly form and enlarge on the material surface, rotator but not thumped. Plasma control sample. One 2.70 rnl tube was If the beaker is shaken, bubbles are seen to separate. overfilled with plasma and capped to ensure that n,~ The material is left in ethanol for a minimum of 24
Volume 11 Number 4 April 1990
Complement and denudeation 593
Zymosan
Air Bubble
Fig. 1. Platelet aggregation assay. The incubated plasma samples have been added to a platelet suspension. The vertical deflection in centimeters on the recording is related to the light transmitted through the platelet suspension; the greater the aggregation, the greater the de-i flection. The height of deflection after 30 seconds is read in centimeters above the baseline and is noted for each sample. air bubbles were introduced, then incubated on the rotator but not thumped. At the completion of incubation all plasma samples were centrifuged (1033g for 4 ° C, 10 minutes), and the clear supernatant was transferred to new polypropylene tubes. Radioimmunoassay for complement activation Complement activation was determined by measuring the stable fluid-phase metabolites C3a desArg, C4a des-Arg, and C5a des-Arg, with commercially available kits (Amersham, Oakville, Ontario, Canada). We assume that measurements of C3a desArg, C4a des-Arg, and C5a des-Arg represent equir, 51ar concentrations of C3a, C4a, and C5a respectively, produced in plasma during the incubation. The method on which the commercial kits are based has been described,2~'22and it consists of precipitating the C3, C4, or C5 from each of the incubated plasma samples with a supplied agent. After the precipitation step the supematant contains the C3a des-Arg, CAa des-Arg, and C5a des-Arg. The concentrations of these components were determined by a standard radioimmunoassay (RIA) by use of iodine 125labeled purified peptides and specific antibodies. Assay for platelet aggregation The 10 ml blood sample collected in ACD served as the source of the platelet suspension for the aggregation experiments. The blood was centrifuged at room temperature (1033 g, 10 minutes), and the ~-l¢ar platelet-poor plasma was saved on ice. The re-
mainder containing plasma, red blood cells, and platelets was further centrifuged (2195g for 2 minutes), and the platelet-rich plasma (PRP) was separated and again centrifuged (2195 g, 15 minutes). The PRP was carefully removed, and an attempt was made not to contaminate it with red blood cells. The last two steps were repeated until we were certain that the PRP was free of red blood cells. A platelet concentration of 7 to 10 × 106 platelets/ml was reached using the platelet-poor plasma for dilution as necessary. The plasma samples for incubation were prepared in the same fhshion as described herein. Platelet aggregation by the different samples was determined using a dual channel aggregometer. Each sample was added to the autologous platelet suspension at the completion of incubation. Platelet aggregation was directly related to light transmission through the suspension as measured by the deflection (in centimeters) on a strip chart recording. All measurements were made 30 seconds after addition of the incubated plasma sample to the platelet suspension. An example from a strip chart recording is illustrated in Fig. 1. Statistical analysis The Statistical Analysis System program (SAS Institute Inc., Cary, N.C.) was used. 23 Comparisons between control and denucleated groups o f Dacron were made with a two-way analysis of variance (ANOVA) (main effects group and surface area). Differences between control and denucleated Dacron
Journal of VASCULAR SURGERY
594 Kalman,IVlcCullough,and Ward
10-
G5
C3 z
o
I.I,.,(,5 ,< Z ,< (i) O N
0 PLASMA CONTROL
AIR BUBBLE
PLASMA CONTROL
AIR BUBBLE
Fig. 2. C3a des-Arg and C5a des-Arg concentration comparing air bubble stimulated plasma with plasma control. There was no difference observed for C4a des-Arg.
60
50
< O
40 ¸
30-
Z
_o I-..O
20
W LI_ I.U D
PLASMA CONTROL
AIR BUBBLE
Fig. 3. Platelet aggregation comparing plasma stimulated with air bubbles with plasma control.
for each size were specified with a paired t test. The results are presented in the figures as the mean + standard error. RESULTS Since zymosan activated plasma represented the strong positive control and determined the upper limit of activation for each plasma sample, the RIA and platelet aggregation results are expressed as the percent of zymosan activation. Fig. 2 shows the RIA results, and Fig. 3 shows the platelet aggregation results comparing air bubble stimulated plasma
to plasma control. Plasma incubated with air bubbles had a significantly higher C3a (p = 0.002) and higher C5a (p = 0.05) concentration than the plasma control (Fig. 2), but no difference was detected for C4a (p = 0.794). There was a significant increase in platelet aggregation when plasma incubated with air bubbles was added to the autologous platelet suspension as compared to addition of control plasma (p = 0.001) (Fig. 3). The RIA and platelet aggregation results for control and denucleated Dacron are shown in Figs. 4 through 6. The amount of C4a-des Arg after incu-
Volume 1 1 Number 4 April 1 9 9 0
Complement and denudeation 595 25 C3
20
[]
Control Dacron
]
Denucleated Dacron
'-F-
15
L
10.
~
5,
0
'
2cm2
4cm2
6cm2
Fig. 4. C3a des-Arg concentration comparing denucleated Dacron with control Dacron pieces (2, 4, 6 cm 2) (different at all sizes, p = 0.042, p = 0.019, 0.027, respectively). 8"
C5
7" Z o
6:
[]
Control Dacron
[]
Denucleated Dacron
I-
5"
z< O >N
4-
/
03
3 2 1
0
2 cm2
4 cm 2
6 crn2
Fig. 5. C5a des-Arg concentration comparing denucleated Dacron with control Dacron pieces (2, 4, 6 cm 2) (different at all sizes, p = 0.012, p = 0.023, p = 0.003, respectively). bation with Dacron, was not significantly different from that found after plasma was incubated with zymosan. There was a significant difference by A N O V A between denudeated Dacron and control Dacron for C3a-des Arg (p = 0.001) and for C5ades Arg (p = 0.001). The differences were also significant by paired t test at each material size (p < 0.05 for C3a-des Arg andp < 0.01 for C5a-des Arg). The degree o f activation was significantly different when analyzed for material size (ANOVA) for C5a-des Arg (p = 0.017), and the trend was the same for C3ades Arg (p = 0.111). The differences were signific~ant when platelet aggregation was analyzed with
respect to group (denucleated Dacron vs control Dacron) (p = 0.001) and material size (2, 4, and 6 cm 2) (p = 0.001) (ANOVA). The values for C3a des-Arg and C5a des-Arg for the experiments are summarized in Table I. DISCUSSION Our results confirm that air bubbles activate complement in human plasma, and that removal o f the microscopic bubbles trapped in the surface roughness o f Dacron can result in reduced complement activation and in vitro platelet aggregation. It has been previously shown that incubating human plasma
596
Journal of VASCULAR SURGERY
Kalman, McCullough, and Ward
120 •
100 < O9 ©
]
Control Dacron
]
Denucleated Dacron
80-
>N
60" Z
o
}--
40 L~ UJ
20
2 cm2
6 cm 2
Fig. 6. Platelet aggregation comparing denucleated Dacron with control Dacron pieces (2, 6 cm2) (different at all sizes,p = 0.001, p = 0.0001, respectively).
Table I. Values (ng/ml) for C3a des-Arg and C5a des-Arg for control Dacron and denucleated Dacron for various sizes Group Control
Denucleated
Size (cma) 2 4 6 2 4 6
C3a des-Arg (ng / ml) 1576.2 3107.9 3982.5 733.1 830.3 971.2
-+ + + + +-+
392.7 1312.1 1673.5 328.4 338.5 325.9
C5a des-Arg (ng / ml) 155.7 199.9 269.6 82.8 90.8 104.5
_+ 47.6 + 57.4 -+ 69.6 + 36.8 _+ 39.1 +_ 41.9
with air bubbles results in complement activation along the alternative pathway, 17 and introduction of nitrogen bubbles directly into PRY' results in platelet aggregation.24 It is known that when an air bubble enters plasma the bubble surface becomes coated with proteins, z°'2s The possibility exists that the plasma proteins at this interface become distorted, and that the complement system simply reacts to these proteins, although native in origin, as it would in the presence of foreign proteins. Incubation of rabbit plasma with air bubbles and then injection of the incubated plasma into an autologous platelet st/spension was found to induce platelet aggregation, and the degree of platelet aggregation was proportional to the degree of complement activation. 26 When the rabbits were decomplemented in vivo with cobra venom factor before the blood collection, it was found that after plasma was incubated with air bubbles, the previously observed aggregation of an autologous platelet suspension was inhibited. 26 This would indicate that complement activation plays an
essential role in platelet aggregation induced by air bubbles. Osada et al.lS demonstrated in a sheep model that denucleation of the silicone rubber blood contact surface in an extracorporeal membrane oxygenator significantly reduced the platelet loss during perfusion as well as platelet adhesion and thrombus formation on the membrane surface. However, the mechanism for these observations was not clear. Herzlinger and Cumming27 performed experiments in dogs to investigate the role of complement activation in mediating neutrophil and platelet adhesion to polymer surfaces. After dogs were decomplemented by means of cobra venom factor, a sharp reduction in the adhesion of platelets and neutrophils to both nylon 6,6 and polymethyl methacrylate was observed. 27 Tb,'~:e results strongly implicate the complement system for cellular adhesion that occurs after a biomaterial is exposed to blood. As mentioned earlier, C3a promotes platelet aggregation and release, 7 and C5a attracts and activates neutrophils, s-l° Meuer et al. 28observed that C5a was approximately 50 times more active in producing serotonin release by platelets than was C3a in guinea pigs, They postulated distinct receptors on platelets for C3a and C5a and therefore two independent pathways o f platelct activation. 28 Kornecki et al. 13 on the other hand, suggested that neutrophil/platelet interactions may play a role in hemostasis and thrombosis, and observed that elastase secreted from human neutrophils exerted a direct stimulatory effect on human platelet fimction. Ward et al. 16 used the leukocyte aggregation test as a qualitative measure of complement activation in
Volume 11 Number 4
April1990
r~obits, and demonstrated that the complement system was activated by air bubbles in plasma and serum, and that the response could be inhibited if the rabbits were decomplementcd. Since complement activation occurred with both serum and plasma samples, it was concluded that fibrinogen did not have an important role in the activation. 16 Although the leukocyte aggregation technique is a sensitive measure of C5a in plasma, 29 RIA is much more quantitative and allows precise measurement of not only C3a and C5a but C4a as well. Since CAa levels can be measured, the pathway to activation can be determined. The mechanism by which vascular grafts and other biomaterials activate complement is not yet clear. It has been suggested that Dacron activates complement by both the classic and alternative pathways. ~,6 C4a is detected only after activation of the classic pathway. Since zymosan activates via the alternative pathway, low levels of C4a are expected. CAa in the plasma samples inc~bated with Dacron were not different from the zymosan stimulated samples. This would suggest that Dacron also activates complement via the att~i~rlative pathway. Classic pathway activation occurs 6nly after preformed antibodies bind to their target antigens, whereas the alternative pathWay is activated in nonspecific ways to all0v~ ~ccognition of foreign particulate matter. Classic pathway activation by biomaterials has been attributed to immunoglobulin G adsorption onto polymer surfaces having amino, carboxyl, cyano, or phenyl groups. 3° Recent evidence indicates a relationship between platclct and neutrophil adhesion to biomatcrials and complement activation. 2° Shoenfcld et al? ~observed that Dacron grafts were responsible for increased platelet deposition on polytetrafluorocthylenc grafts at ~remote site using an ex vivo shunt in the baboon. Although the authors did not postulate a mechanism responsible for the increased platelet deposition, they did suggest that their observations were due to systemic platelet activation by the Dacron31; one can only speculate that complement activation may be contributing. The denuclcation procedure first exposes the Dacron surface to ethanol, and subsequently the ethanol is displaced by degassed buffer. Because of the small surface tension of ethanol, it will penetrate crevices of the Dacron surface and will displace the air; then when Dacron is carefully rinsed with degassed buffer, the ethanol present will be diluted by the buffer since they are infinitely soluble in one another, and as the rinsing process continues the ethanol will be completely replaced by buffer. Therefore when denucleated Dacron is exposed to plasma, plasma/mate-
Complement and denucleation 5 9 7
rial, and plasma/buffer, interfaces will be present. The plasma/air bubble interface of the control sampies of Dacron will have been completely eliminated. The differences observed between the control and denucleated samples of Dacron can thus be attributed to the replacement of the plasma/air bubble interface by the plasma/buffer interface. The fact that the denucleation of the Dacron surface leads to reduced complement activation means that in this regard Dacron behaves in the same way as does polytetrafluoroethylenc, silicone rubber,, and cellophane. 19 Therefore control of local complement activation and cellular adhesion to the graft surface might have implications for graft survival. Pharmacologic manipulations to interfere with complement activation have been attempted? z Our experiments on the other hand, demonstrate that complement activation by Dacron can be significantly reduced by denucleation, and the significant reduction of in vitro l~latelet aggregation paralleled that of C3a and C5a. Although our data do not prove a mechanism whereby the changes in complement activation and platelet aggregation are linked, they nevertheless suggest that a common mechanism may be shared. Our results show that complement activation and platelet aggregation vary with material surface area and may help explain why long prosthetic grafts such as an axillofemoral or femorotibial grafts have less than ideal results. Use of this or a similar approach to alter the capacity of a biomaterial to activate complement may eventually improve prosthetic vascular graft biocompatability and graft patency. We performed the experiments reported in this paper to determine the degree of complement activation in human plasma when incubated with Dacron, to determine the degree of platelet aggregation thatresults from injecting incubated plasma into an autologous platelet suspension, and to determine how the degree of complement activation and platelct aggregation can be reduced when the air nuclei arc removed from the surface of the Dacron before the material is exposed to plasma. As a first step, we report the results of a set of control experiments in which human plasma was incubated with air bubbles. The degree of complement activation, as well as autologous platelet aggregation induced by air bubbles was measured. The authors thank Miss Elaine Caon for her assistance in the preparation of this manuscript. REFERENCES 1. Craddock PR, Fehr J, Dalmasso AP, Brigham KL, Jacob HS. Heodialysis leukopenia: pulmonary vascular leukostasis re-
598
2.
3.
4.
5.
6.
7. 8.
9.
10.
11. 12.
13.
14. 15.
16.
Kalman, McCullough, and Ward
stilting from complement activation by dialyzer cellophane membranes. J Clin Invest 1977;59:879-88. Hammerschmidt DE, Stronced DF, Bowers TK, et al. Complement activation and neutropenia occurring during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1981; 81:370-7. Bohler J, Kramer P, Gotze O, Schartz P, Scheler F. Leukocyte counts and complement activation during pump-driven and arteriovenous haemofiltration. Cont Hephrol 1983;36:15-7. Ryhanen P, Herva E, Holhnan A, Nuutinen L, Pihlajaniemi R, Saarela E. Changes in peripheral blood leukocyte counts, lymphocyte subpopulations and in vito transformation after heart valve replacement. J Thorac Cardiovasc Surg 1979; 77:259-65. Shepard AD, Gelfand JA, Callow AD, O'Donnell TF. Complement activation by synthetic vascular prostheses. J VASC SUttG 1984;1:829-38. Kotte-Marchant K, Anderson JM, Miller KM, Marchant RE, Lazarus H. Vascular graft associated complement activation and leukocyte adhesion in an artificial circulation. J Biomed Mater Res 1987;21:379-97. Polley MJ, Nachman RL. Human platelet activation by C3a and C3a des-Arg. J Exp Med 1983;158:603-7. Dahlgren C, Hed J, Stendahl O. Chemotaxis of polymorphonuclear leukocytes in response to surface-bound complement derived chemoattractants generalized in situ. Inflammation 1984;8:201-9. Craddock PR, HammerschmJdt D, White JG, Dalmasso AP, Jacob HS. Complement (C5a) induced granulocyte aggregation in vitro: a possible mechanism of complementmediated leukostasis and leukopenia. J Clin Invest 1977; 60:260-4. O'Flaherty JT, Kruetzer DL, Ward PA. Chemotactic factor influences on the aggregation, swelling and foreign surface adhesiveness of human leukocytes. Am J Pathol 1978; 90:537-44. Niemetz J, Muhlfelder T, Chiereg ME, Troy B. Procoagulant activity of leukocytes. Ann NY Acad Sci 1984;95:331-8. Ostermann G, Till J, Thielmann K. Studies on the stimulation of human blood platelets by semi-synthetic platelet-activating factor. Thromb Res 1983;30:127-32. Kornecki E, Ehrllch YH, Egbring R, et al. Granulocyteplatelet interactions and platelet fibrinogen receptor exposure. Am J Physiol 1988;255(Heart Circ Physiol 24):651-8. Ward CA, Ruegsegger B, Stanga D, Zingg W. Reduction in platelet adhesion to biomaterials by removal of gas nuclei. Trans Am Soc Artif Int Organs 1974;20:77-84. Osada H, Ward CA, Du!Ym J, Nelms M, Cooper JD. Microbubble elimination during priming improves biocompatibility of membrane oxygenators. Am J Physio11978;234(Heart Circ Physiol 3):H646-52. Ward CA, Koheil A, McCullough D, Johnson WR, Fraser WD. Activation of complement at plasma-air of serum air interface of rabbits. J Appl Physiol 1986;60:1651-8.
Journal of VASCULAR SURGERY
17. Ward CA, McCullough D, Fraser WD. Relation betwe,~A complement activation and susceptibility to decompression sickness. J Appl Physiol 1987;62:1160-6. 18. Ward CA, Forest T. On the relation between platelet adhesion and roughness of a synthetic biomaterial. Ann Biomed Eng 1979;7:451-69. 19. Ward CA, Koheil A, Johnson WR, Madras PN. Reduction in complement activation from biomaterials by removal of air nuclei from the surface roughness. J Biomed Mater Res 1984;18:225-69. 20. Ward CA, Kalman PG. Complement activation mediates cellular adhesion to synthetic biomaterials. Med Prog Technol 1989;15:63-75. 21. Chenoweth DE, Hugh TE. Techniques and significance of C3a and C5a measurement. In: Nakamura RM, ed. Future perspectives in clinical laboratory immunoassays. New York: Alan R. Liss; Inc., NY, 1980. 22. Gorski JP. Quantitation of human complement fragment CAa in physiological fluids by competitive inhibition radioimmune assay. J Immunol Methods 1981;47:61-73. 23. Ray AA, Sail JP, Salter M. SAS user's guide: Statistics, 1982. Cary, NC: SAS Institute Inc, 1982. 24. Thorsen T, Brubakk A, Orstedal T, Farstad M, Hohnsen H. Method for production of N~ microbubbles in platele~,~ich plasma in an aggregometer-like apparatus, and effect 0:~~the platelet density in vitro. Undersea Biomed Res 1986; 13:27188. 25. Philp RB, Inwood MJ, Warren BA. Interactions between gas bubbles and components of blood: implications in decompression sickness. Aerospace Med 1972;43:946-53. 26. Ward CA, McCullough D, Fraser WD. Role of complement activation in air bubble induced rabbit platelet aggregation (Submitted for publicaton). 27. Herzlinger GA, Cumming RD. Role of complement activation in cell adhesion to polymer blood contact surfaces. Trans Am Soc Artif Intern Organs 1980;26:165-71. 28. Meuer S, Ecker U, Hadding U, Bitter-Suermann D. Plateletserotonin release by C3a and C5a: two independent pathways of activation. J Immunol 1981; 126:1506-9. 29. Hammerschmidt DE, Bowers TK, Lammi-Keefe CJ, Jacob HS, Craddock PR. Granulocyte aggregometry: a sensitive technique for the detection of C5a and complement activation. Blood 1980;55:898-902. 30. Uchida T, Hosaka S, Murao Y. Complement activation by polymer binding IgG. Biomaterials 1984;5:281-283. 31. Shoenfeld NA, Connolly R, Ramberg K, Valeri CR, EldrupJorgensen J, Callow AD. The systemic activation of platelets by Dacron grafts. Surg Gynecol Obstet 1988;166:454-7. 32. Harnmerschmidt DE, White JG, Craddock PR, Jacob HS. Corticosteroids inhibit complement-induced granulocyte aggregation. J Chn Invest 1979;63:798-803.
When a biomaterial is exposed to blood, complex events occur that may result in cellular adhesion ( ~htelets, neutrophils) and lead to thrombus formarion. Complement activation by polymer surfaces is known to occur with hemodialysis,~ cardiopulmonary bypass surgery, 2 hemofiltration, 3 after cardiac valve replacement, 4 and with prosthetic vascular grafts, s,6 Whether complement activation proceeds via the classic or alternative pathway, the physiolog-
From the Department of Surgery and Division of Vascular Surgery, Toronto General Hospital, and the Department of MechanicalEngineering,Universityof Toronto. Supported by the Physicians'Services Incorporated Foundation (PSI) of Ontario. Presented at the BreakfastProgram, Societyfor Vascular Surgery and the InternationalSocietyfor CardiovascularSurgeryAnnual Meeting, New York, N.Y. June 20, 1989. Reprint requests: P. G. Kalman,MD, Toronto GeneralHospital, Divisionof VascularSurgery,9 Eaton North - Room 211, 200 Elizabeth St., Toronto, Ontario, M5G 2CA Canada.
z4/6/18048
ically active inflammatory mediators C3a and C5a are released. C3a generation is important since it promotes platelet aggregation and release. 7 C5a is an extremely potent mediator of neutrophil chemotaxis, s aggregation, 9 and adhesion. ~° Activated and adherent neutrophils may contribute to vascular graft thrombosis by augmenting coagulation, n This may occur by stimulation of platelet aggregation through release o f platelet activating factor, ~2 or by a direct neutrophil/platelet interaction. ~3 If prosthetic vascular grafts significantly activate complement locally, then reduction of the activation would be expected to cause a decrease in platelet and neutrophil adhesion and perhaps thrombus formation. Since microscopic air bubbles can become trapped in the surface roughness of a biomaterial when the material is in contact with blood, ~4 two interfaces can be postulated to exist: (1) a blood/biomaterial and (2) a blood/air bubble interface. Elimination o f these surface air bubbles has successfully decreased cellular adhesion to blood contact 591
Journal of VASCULAR SURGERY
592 Kalman, McCullough, and Ward
surfaces, as Furthermore, complement activation by air bubbles has been observed in both rabbit 16 and human a7 plasma. Since microscopic bubbles of air can be expected to be trapped in the surface roughness of Dacron 18 as they are in other materials, ~4'xs,~9 they would be expected to give rise to complement activation when the material is brought in contact with plasma, and plasma incubated with Dacron would be expected to give rise to platelet aggregation. Furthermore, if Dacron is subjected to a procedure to remove the air nuclei before exposure to plasma, complement activation would be expected to be decreased, and if it is complement activation that induces platelet aggregation it would be expected that if the air nuclei were removed before incubation with plasma, then this would result in reduced platelet aggregation. Our results confirm that air bubbles activate complement and augment platelet activation in vitro. Moreover, denucleation of Dacron strongly reduces both complement activation and platelet aggregation. A recent review of the factors controling cell adhesion to a synthetic material suggests that complement activation plays a mediating role, 2° therefore denucleation of Dacron may be a means by which the biocompatibility of this commonly used biomaterial can be improved.
hours at room temperature, and if it is disturbed after this time no bubbles are seen to separate from the surface. The ethanol is decanted, leaving enough to cover the Dacron. Phosphate buffer (pH 7.37, osmolarity 0.297) is degassed for a minimum of 6 hours with a vacuum and magnetic stir bar at room temperature. The Dacron is then slowly and continuously washed for 6 hours with degassed buffer, gradually replacing all ethanol. The specimens were kept in degassed buffer until incubation with plasma. Incubation o f samples
Samples were placed in 2.70 ml capacity selfcapped tubes and were subsequently incubated on a rotator at 22 rpm in a 37 ° C water bath for 30 minutes. 16 Plasma incubated with air bubbles. Plasma (2.5 ml) was placed in a 2.70 ml tube, and after the tube was capped small bubbles were formed by vigor0*'~ly shaking it. The tube was placed on the rotator, and bubbles were maintained during incubation by sharply thumping one end during each rotation by use of a device added to the rotator. Plasma incubated with zymosan. Zymosan activated plasma served as a strong positive control and was used to determine the upper limit of complement activation for each plasma sample. The zymosan MATERIAL AND METHODS (Sigma Chemical, St. Louis, Mo.) was boiled in saCollection o f blood line solution for 15 minutes, cooled, and washed three times, then resuspended in saline solution Eleven healthy male volunteers (age range 19 to (5 mg weight per milliliter). A 2.70 ml tube was 36 years) donated blood after approval by the human filled with 2.50 ml of the zymosan-saline suspension ethics committee at our institution. Polypropylene then centrifuged (1033g, 4 ° C, 10 minutes) leaving syringes and incubation tubes were used for all exa pellet of zymosan. Plasma was added to the pellet periments, since polypropylene only minimally actito form a suspension (5 mg zymosan per milliliter vates complement. One hundred milliliters of blood plasma) and was then gently shaken to ensure adewas collected from each volunteer by venipuncture, quate mixture. The suspension was then incubated 90 rnl in heparin (10 units per milliliter of blood) on the rotator with the other samples, but not and 10 ml in acid citrate dextrose (ACD) (6 parts thumped. blood to 1 part ACD). Plasma was separated in the Plasma exposed to Dacron. Pieces of weaveknit heparinized sample by centrifugation (1033 g, 15 Dacron (Meadox Medicals Inc. Oakland, N.J.) with minutes, 4 ° C). The blood collected in ACD served surface areas of 2, 4, or 6 cm 2 were cut from a sample as a source of platelets for the aggregation experiof a tube of the material (10 mm outer diameter, ments. 8 mm inner diameter). The Dacron pieces to be inDenucleation o f Dacron cubated were placed in tubes, overfilled with plasma The denucleation process was very similar to that without introducing air bubbles, then capped and described by Ward et a1.19The cut pieces of Dacron sealed. Three of the tubes each contained a sample were placed in a glass beaker containing 99% ethanol. • of control Dacron, and three contained denucleated Dacron (2, 4, 6 cm2). All were incubated on the Visible bubbles leave the Dacron rapidly at first, then they slowly form and enlarge on the material surface, rotator but not thumped. Plasma control sample. One 2.70 rnl tube was If the beaker is shaken, bubbles are seen to separate. overfilled with plasma and capped to ensure that n,~ The material is left in ethanol for a minimum of 24
Volume 11 Number 4 April 1990
Complement and denudeation 593
Zymosan
Air Bubble
Fig. 1. Platelet aggregation assay. The incubated plasma samples have been added to a platelet suspension. The vertical deflection in centimeters on the recording is related to the light transmitted through the platelet suspension; the greater the aggregation, the greater the de-i flection. The height of deflection after 30 seconds is read in centimeters above the baseline and is noted for each sample. air bubbles were introduced, then incubated on the rotator but not thumped. At the completion of incubation all plasma samples were centrifuged (1033g for 4 ° C, 10 minutes), and the clear supernatant was transferred to new polypropylene tubes. Radioimmunoassay for complement activation Complement activation was determined by measuring the stable fluid-phase metabolites C3a desArg, C4a des-Arg, and C5a des-Arg, with commercially available kits (Amersham, Oakville, Ontario, Canada). We assume that measurements of C3a desArg, C4a des-Arg, and C5a des-Arg represent equir, 51ar concentrations of C3a, C4a, and C5a respectively, produced in plasma during the incubation. The method on which the commercial kits are based has been described,2~'22and it consists of precipitating the C3, C4, or C5 from each of the incubated plasma samples with a supplied agent. After the precipitation step the supematant contains the C3a des-Arg, CAa des-Arg, and C5a des-Arg. The concentrations of these components were determined by a standard radioimmunoassay (RIA) by use of iodine 125labeled purified peptides and specific antibodies. Assay for platelet aggregation The 10 ml blood sample collected in ACD served as the source of the platelet suspension for the aggregation experiments. The blood was centrifuged at room temperature (1033 g, 10 minutes), and the ~-l¢ar platelet-poor plasma was saved on ice. The re-
mainder containing plasma, red blood cells, and platelets was further centrifuged (2195g for 2 minutes), and the platelet-rich plasma (PRP) was separated and again centrifuged (2195 g, 15 minutes). The PRP was carefully removed, and an attempt was made not to contaminate it with red blood cells. The last two steps were repeated until we were certain that the PRP was free of red blood cells. A platelet concentration of 7 to 10 × 106 platelets/ml was reached using the platelet-poor plasma for dilution as necessary. The plasma samples for incubation were prepared in the same fhshion as described herein. Platelet aggregation by the different samples was determined using a dual channel aggregometer. Each sample was added to the autologous platelet suspension at the completion of incubation. Platelet aggregation was directly related to light transmission through the suspension as measured by the deflection (in centimeters) on a strip chart recording. All measurements were made 30 seconds after addition of the incubated plasma sample to the platelet suspension. An example from a strip chart recording is illustrated in Fig. 1. Statistical analysis The Statistical Analysis System program (SAS Institute Inc., Cary, N.C.) was used. 23 Comparisons between control and denucleated groups o f Dacron were made with a two-way analysis of variance (ANOVA) (main effects group and surface area). Differences between control and denucleated Dacron
Journal of VASCULAR SURGERY
594 Kalman,IVlcCullough,and Ward
10-
G5
C3 z
o
I.I,.,(,5 ,< Z ,< (i) O N
0 PLASMA CONTROL
AIR BUBBLE
PLASMA CONTROL
AIR BUBBLE
Fig. 2. C3a des-Arg and C5a des-Arg concentration comparing air bubble stimulated plasma with plasma control. There was no difference observed for C4a des-Arg.
60
50
< O
40 ¸
30-
Z
_o I-..O
20
W LI_ I.U D
PLASMA CONTROL
AIR BUBBLE
Fig. 3. Platelet aggregation comparing plasma stimulated with air bubbles with plasma control.
for each size were specified with a paired t test. The results are presented in the figures as the mean + standard error. RESULTS Since zymosan activated plasma represented the strong positive control and determined the upper limit of activation for each plasma sample, the RIA and platelet aggregation results are expressed as the percent of zymosan activation. Fig. 2 shows the RIA results, and Fig. 3 shows the platelet aggregation results comparing air bubble stimulated plasma
to plasma control. Plasma incubated with air bubbles had a significantly higher C3a (p = 0.002) and higher C5a (p = 0.05) concentration than the plasma control (Fig. 2), but no difference was detected for C4a (p = 0.794). There was a significant increase in platelet aggregation when plasma incubated with air bubbles was added to the autologous platelet suspension as compared to addition of control plasma (p = 0.001) (Fig. 3). The RIA and platelet aggregation results for control and denucleated Dacron are shown in Figs. 4 through 6. The amount of C4a-des Arg after incu-
Volume 1 1 Number 4 April 1 9 9 0
Complement and denudeation 595 25 C3
20
[]
Control Dacron
]
Denucleated Dacron
'-F-
15
L
10.
~
5,
0
'
2cm2
4cm2
6cm2
Fig. 4. C3a des-Arg concentration comparing denucleated Dacron with control Dacron pieces (2, 4, 6 cm 2) (different at all sizes, p = 0.042, p = 0.019, 0.027, respectively). 8"
C5
7" Z o
6:
[]
Control Dacron
[]
Denucleated Dacron
I-
5"
z< O >N
4-
/
03
3 2 1
0
2 cm2
4 cm 2
6 crn2
Fig. 5. C5a des-Arg concentration comparing denucleated Dacron with control Dacron pieces (2, 4, 6 cm 2) (different at all sizes, p = 0.012, p = 0.023, p = 0.003, respectively). bation with Dacron, was not significantly different from that found after plasma was incubated with zymosan. There was a significant difference by A N O V A between denudeated Dacron and control Dacron for C3a-des Arg (p = 0.001) and for C5ades Arg (p = 0.001). The differences were also significant by paired t test at each material size (p < 0.05 for C3a-des Arg andp < 0.01 for C5a-des Arg). The degree o f activation was significantly different when analyzed for material size (ANOVA) for C5a-des Arg (p = 0.017), and the trend was the same for C3ades Arg (p = 0.111). The differences were signific~ant when platelet aggregation was analyzed with
respect to group (denucleated Dacron vs control Dacron) (p = 0.001) and material size (2, 4, and 6 cm 2) (p = 0.001) (ANOVA). The values for C3a des-Arg and C5a des-Arg for the experiments are summarized in Table I. DISCUSSION Our results confirm that air bubbles activate complement in human plasma, and that removal o f the microscopic bubbles trapped in the surface roughness o f Dacron can result in reduced complement activation and in vitro platelet aggregation. It has been previously shown that incubating human plasma
596
Journal of VASCULAR SURGERY
Kalman, McCullough, and Ward
120 •
100 < O9 ©
]
Control Dacron
]
Denucleated Dacron
80-
>N
60" Z
o
}--
40 L~ UJ
20
2 cm2
6 cm 2
Fig. 6. Platelet aggregation comparing denucleated Dacron with control Dacron pieces (2, 6 cm2) (different at all sizes,p = 0.001, p = 0.0001, respectively).
Table I. Values (ng/ml) for C3a des-Arg and C5a des-Arg for control Dacron and denucleated Dacron for various sizes Group Control
Denucleated
Size (cma) 2 4 6 2 4 6
C3a des-Arg (ng / ml) 1576.2 3107.9 3982.5 733.1 830.3 971.2
-+ + + + +-+
392.7 1312.1 1673.5 328.4 338.5 325.9
C5a des-Arg (ng / ml) 155.7 199.9 269.6 82.8 90.8 104.5
_+ 47.6 + 57.4 -+ 69.6 + 36.8 _+ 39.1 +_ 41.9
with air bubbles results in complement activation along the alternative pathway, 17 and introduction of nitrogen bubbles directly into PRY' results in platelet aggregation.24 It is known that when an air bubble enters plasma the bubble surface becomes coated with proteins, z°'2s The possibility exists that the plasma proteins at this interface become distorted, and that the complement system simply reacts to these proteins, although native in origin, as it would in the presence of foreign proteins. Incubation of rabbit plasma with air bubbles and then injection of the incubated plasma into an autologous platelet st/spension was found to induce platelet aggregation, and the degree of platelet aggregation was proportional to the degree of complement activation. 26 When the rabbits were decomplemented in vivo with cobra venom factor before the blood collection, it was found that after plasma was incubated with air bubbles, the previously observed aggregation of an autologous platelet suspension was inhibited. 26 This would indicate that complement activation plays an
essential role in platelet aggregation induced by air bubbles. Osada et al.lS demonstrated in a sheep model that denucleation of the silicone rubber blood contact surface in an extracorporeal membrane oxygenator significantly reduced the platelet loss during perfusion as well as platelet adhesion and thrombus formation on the membrane surface. However, the mechanism for these observations was not clear. Herzlinger and Cumming27 performed experiments in dogs to investigate the role of complement activation in mediating neutrophil and platelet adhesion to polymer surfaces. After dogs were decomplemented by means of cobra venom factor, a sharp reduction in the adhesion of platelets and neutrophils to both nylon 6,6 and polymethyl methacrylate was observed. 27 Tb,'~:e results strongly implicate the complement system for cellular adhesion that occurs after a biomaterial is exposed to blood. As mentioned earlier, C3a promotes platelet aggregation and release, 7 and C5a attracts and activates neutrophils, s-l° Meuer et al. 28observed that C5a was approximately 50 times more active in producing serotonin release by platelets than was C3a in guinea pigs, They postulated distinct receptors on platelets for C3a and C5a and therefore two independent pathways o f platelct activation. 28 Kornecki et al. 13 on the other hand, suggested that neutrophil/platelet interactions may play a role in hemostasis and thrombosis, and observed that elastase secreted from human neutrophils exerted a direct stimulatory effect on human platelet fimction. Ward et al. 16 used the leukocyte aggregation test as a qualitative measure of complement activation in
Volume 11 Number 4
April1990
r~obits, and demonstrated that the complement system was activated by air bubbles in plasma and serum, and that the response could be inhibited if the rabbits were decomplementcd. Since complement activation occurred with both serum and plasma samples, it was concluded that fibrinogen did not have an important role in the activation. 16 Although the leukocyte aggregation technique is a sensitive measure of C5a in plasma, 29 RIA is much more quantitative and allows precise measurement of not only C3a and C5a but C4a as well. Since CAa levels can be measured, the pathway to activation can be determined. The mechanism by which vascular grafts and other biomaterials activate complement is not yet clear. It has been suggested that Dacron activates complement by both the classic and alternative pathways. ~,6 C4a is detected only after activation of the classic pathway. Since zymosan activates via the alternative pathway, low levels of C4a are expected. CAa in the plasma samples inc~bated with Dacron were not different from the zymosan stimulated samples. This would suggest that Dacron also activates complement via the att~i~rlative pathway. Classic pathway activation occurs 6nly after preformed antibodies bind to their target antigens, whereas the alternative pathWay is activated in nonspecific ways to all0v~ ~ccognition of foreign particulate matter. Classic pathway activation by biomaterials has been attributed to immunoglobulin G adsorption onto polymer surfaces having amino, carboxyl, cyano, or phenyl groups. 3° Recent evidence indicates a relationship between platclct and neutrophil adhesion to biomatcrials and complement activation. 2° Shoenfcld et al? ~observed that Dacron grafts were responsible for increased platelet deposition on polytetrafluorocthylenc grafts at ~remote site using an ex vivo shunt in the baboon. Although the authors did not postulate a mechanism responsible for the increased platelet deposition, they did suggest that their observations were due to systemic platelet activation by the Dacron31; one can only speculate that complement activation may be contributing. The denuclcation procedure first exposes the Dacron surface to ethanol, and subsequently the ethanol is displaced by degassed buffer. Because of the small surface tension of ethanol, it will penetrate crevices of the Dacron surface and will displace the air; then when Dacron is carefully rinsed with degassed buffer, the ethanol present will be diluted by the buffer since they are infinitely soluble in one another, and as the rinsing process continues the ethanol will be completely replaced by buffer. Therefore when denucleated Dacron is exposed to plasma, plasma/mate-
Complement and denucleation 5 9 7
rial, and plasma/buffer, interfaces will be present. The plasma/air bubble interface of the control sampies of Dacron will have been completely eliminated. The differences observed between the control and denucleated samples of Dacron can thus be attributed to the replacement of the plasma/air bubble interface by the plasma/buffer interface. The fact that the denucleation of the Dacron surface leads to reduced complement activation means that in this regard Dacron behaves in the same way as does polytetrafluoroethylenc, silicone rubber,, and cellophane. 19 Therefore control of local complement activation and cellular adhesion to the graft surface might have implications for graft survival. Pharmacologic manipulations to interfere with complement activation have been attempted? z Our experiments on the other hand, demonstrate that complement activation by Dacron can be significantly reduced by denucleation, and the significant reduction of in vitro l~latelet aggregation paralleled that of C3a and C5a. Although our data do not prove a mechanism whereby the changes in complement activation and platelet aggregation are linked, they nevertheless suggest that a common mechanism may be shared. Our results show that complement activation and platelet aggregation vary with material surface area and may help explain why long prosthetic grafts such as an axillofemoral or femorotibial grafts have less than ideal results. Use of this or a similar approach to alter the capacity of a biomaterial to activate complement may eventually improve prosthetic vascular graft biocompatability and graft patency. We performed the experiments reported in this paper to determine the degree of complement activation in human plasma when incubated with Dacron, to determine the degree of platelet aggregation thatresults from injecting incubated plasma into an autologous platelet suspension, and to determine how the degree of complement activation and platelct aggregation can be reduced when the air nuclei arc removed from the surface of the Dacron before the material is exposed to plasma. As a first step, we report the results of a set of control experiments in which human plasma was incubated with air bubbles. The degree of complement activation, as well as autologous platelet aggregation induced by air bubbles was measured. The authors thank Miss Elaine Caon for her assistance in the preparation of this manuscript. REFERENCES 1. Craddock PR, Fehr J, Dalmasso AP, Brigham KL, Jacob HS. Heodialysis leukopenia: pulmonary vascular leukostasis re-
598
2.
3.
4.
5.
6.
7. 8.
9.
10.
11. 12.
13.
14. 15.
16.
Kalman, McCullough, and Ward
stilting from complement activation by dialyzer cellophane membranes. J Clin Invest 1977;59:879-88. Hammerschmidt DE, Stronced DF, Bowers TK, et al. Complement activation and neutropenia occurring during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1981; 81:370-7. Bohler J, Kramer P, Gotze O, Schartz P, Scheler F. Leukocyte counts and complement activation during pump-driven and arteriovenous haemofiltration. Cont Hephrol 1983;36:15-7. Ryhanen P, Herva E, Holhnan A, Nuutinen L, Pihlajaniemi R, Saarela E. Changes in peripheral blood leukocyte counts, lymphocyte subpopulations and in vito transformation after heart valve replacement. J Thorac Cardiovasc Surg 1979; 77:259-65. Shepard AD, Gelfand JA, Callow AD, O'Donnell TF. Complement activation by synthetic vascular prostheses. J VASC SUttG 1984;1:829-38. Kotte-Marchant K, Anderson JM, Miller KM, Marchant RE, Lazarus H. Vascular graft associated complement activation and leukocyte adhesion in an artificial circulation. J Biomed Mater Res 1987;21:379-97. Polley MJ, Nachman RL. Human platelet activation by C3a and C3a des-Arg. J Exp Med 1983;158:603-7. Dahlgren C, Hed J, Stendahl O. Chemotaxis of polymorphonuclear leukocytes in response to surface-bound complement derived chemoattractants generalized in situ. Inflammation 1984;8:201-9. Craddock PR, HammerschmJdt D, White JG, Dalmasso AP, Jacob HS. Complement (C5a) induced granulocyte aggregation in vitro: a possible mechanism of complementmediated leukostasis and leukopenia. J Clin Invest 1977; 60:260-4. O'Flaherty JT, Kruetzer DL, Ward PA. Chemotactic factor influences on the aggregation, swelling and foreign surface adhesiveness of human leukocytes. Am J Pathol 1978; 90:537-44. Niemetz J, Muhlfelder T, Chiereg ME, Troy B. Procoagulant activity of leukocytes. Ann NY Acad Sci 1984;95:331-8. Ostermann G, Till J, Thielmann K. Studies on the stimulation of human blood platelets by semi-synthetic platelet-activating factor. Thromb Res 1983;30:127-32. Kornecki E, Ehrllch YH, Egbring R, et al. Granulocyteplatelet interactions and platelet fibrinogen receptor exposure. Am J Physiol 1988;255(Heart Circ Physiol 24):651-8. Ward CA, Ruegsegger B, Stanga D, Zingg W. Reduction in platelet adhesion to biomaterials by removal of gas nuclei. Trans Am Soc Artif Int Organs 1974;20:77-84. Osada H, Ward CA, Du!Ym J, Nelms M, Cooper JD. Microbubble elimination during priming improves biocompatibility of membrane oxygenators. Am J Physio11978;234(Heart Circ Physiol 3):H646-52. Ward CA, Koheil A, McCullough D, Johnson WR, Fraser WD. Activation of complement at plasma-air of serum air interface of rabbits. J Appl Physiol 1986;60:1651-8.
Journal of VASCULAR SURGERY
17. Ward CA, McCullough D, Fraser WD. Relation betwe,~A complement activation and susceptibility to decompression sickness. J Appl Physiol 1987;62:1160-6. 18. Ward CA, Forest T. On the relation between platelet adhesion and roughness of a synthetic biomaterial. Ann Biomed Eng 1979;7:451-69. 19. Ward CA, Koheil A, Johnson WR, Madras PN. Reduction in complement activation from biomaterials by removal of air nuclei from the surface roughness. J Biomed Mater Res 1984;18:225-69. 20. Ward CA, Kalman PG. Complement activation mediates cellular adhesion to synthetic biomaterials. Med Prog Technol 1989;15:63-75. 21. Chenoweth DE, Hugh TE. Techniques and significance of C3a and C5a measurement. In: Nakamura RM, ed. Future perspectives in clinical laboratory immunoassays. New York: Alan R. Liss; Inc., NY, 1980. 22. Gorski JP. Quantitation of human complement fragment CAa in physiological fluids by competitive inhibition radioimmune assay. J Immunol Methods 1981;47:61-73. 23. Ray AA, Sail JP, Salter M. SAS user's guide: Statistics, 1982. Cary, NC: SAS Institute Inc, 1982. 24. Thorsen T, Brubakk A, Orstedal T, Farstad M, Hohnsen H. Method for production of N~ microbubbles in platele~,~ich plasma in an aggregometer-like apparatus, and effect 0:~~the platelet density in vitro. Undersea Biomed Res 1986; 13:27188. 25. Philp RB, Inwood MJ, Warren BA. Interactions between gas bubbles and components of blood: implications in decompression sickness. Aerospace Med 1972;43:946-53. 26. Ward CA, McCullough D, Fraser WD. Role of complement activation in air bubble induced rabbit platelet aggregation (Submitted for publicaton). 27. Herzlinger GA, Cumming RD. Role of complement activation in cell adhesion to polymer blood contact surfaces. Trans Am Soc Artif Intern Organs 1980;26:165-71. 28. Meuer S, Ecker U, Hadding U, Bitter-Suermann D. Plateletserotonin release by C3a and C5a: two independent pathways of activation. J Immunol 1981; 126:1506-9. 29. Hammerschmidt DE, Bowers TK, Lammi-Keefe CJ, Jacob HS, Craddock PR. Granulocyte aggregometry: a sensitive technique for the detection of C5a and complement activation. Blood 1980;55:898-902. 30. Uchida T, Hosaka S, Murao Y. Complement activation by polymer binding IgG. Biomaterials 1984;5:281-283. 31. Shoenfeld NA, Connolly R, Ramberg K, Valeri CR, EldrupJorgensen J, Callow AD. The systemic activation of platelets by Dacron grafts. Surg Gynecol Obstet 1988;166:454-7. 32. Harnmerschmidt DE, White JG, Craddock PR, Jacob HS. Corticosteroids inhibit complement-induced granulocyte aggregation. J Chn Invest 1979;63:798-803.
