Dr V. Videm, MD, PhD, Department of Immunology and Transfusion Medicine, St. Olavs Hospital, Trondheim University Hospital, N-7006 Trondheim, Norway. E-mail: firstname.lastname@example.org
Complement plays a vital role in the body’s defence systems. Cardiopulmonary bypass induces a detrimental inflammatory reaction in which the complement system is known to participate through direct effects as well as through activation of neutrophils, platelets and endothelial cells. On the other hand, it has been suggested that in the setting of cardiopulmonary bypass, complement may be activated by neutrophils, perhaps due to fragmentation caused by the heart–lung machine. We therefore investigated whether intact or fragmented neutrophils were able to activate the complement system, and whether neutrophil–platelet interaction could influence such complement activation. Lepirudin-anticoagulated plasma was incubated at 37 °C with resting or activated intact neutrophils or neutrophils combined with platelets, or increasing amounts of fragmented neutrophils. Complement activation was evaluated by measurement of C1rs-C1 inhibitor complexes, C4bc, C3bBbP, C3bc, C5a and sC5b-9. We found significant activation of complement only by unphysiological doses of fragmented neutrophils or supernatant from fragmented neutrophils, consistent with a limited clinical significance related to neutrophil destruction during cardiopulmonary bypass. Unstimulated neutrophils induced C3bPBb formation but little formation of other activation products, indicating an increased C3 hydrolysis which was kept under control by regulatory mechanisms. Neutrophils and platelets combined increased classical activation and decreased alternative activation, similar to the findings with platelets alone. Our data confirm that in the setting of acute neutrophil fragmentation or activation, complement activation is much more important in the inflammatory network as an event upstream to neutrophil activation than vice versa.
Complement is an essential part of the innate immune system and has various important physiological activities, including host defence against infections and interplay between innate and adaptive immunity. It takes part in the disposal of immune complexes from tissue and clearance of apoptotic cells and other products of an inflammatory injury or ‘used’ self. Finally, complement is necessary for tissue repair and regeneration [1, 2]. Many complement proteins are involved in initiating and completing these functions, among others C1q and the anaphylatoxins C3a and C5a as well as the membrane-attack complex (C5b-9).
Complement activation was early associated with the complex inflammatory reaction observed during and after cardiopulmonary bypass, which also encompasses activation of neutrophils, platelets and endothelial cells. The aetiology of complement activation is complex with the foreign surface of the extracorporeal circuit as well as the use of heparin and protamine, release of endotoxins and tissue injury being involved [3–8]. A late phase of complement activation occurs post-operatively and seems to be associated with increasing levels of acute phase proteins like protein secretory phospholipase A2 and C-reactive protein [9, 10]. The activation of the complement system in relation to cardiopulmonary bypass involves all three pathways, the classical, mannose-binding lectin and alternative pathway [11, 12] and results in tissue injury either through activation of inflammatory cells like neutrophils or a direct damaging effect on parenchymal cells [11, 13].
The complexity of the inflammatory reaction in cardiopulmonary bypass has resulted in speculations regarding other mechanisms of complement activation. Thrombin can function as a C5 convertase independently of C3 . Neutrophils have also been implicated as they can release enzymes capable of inducing conformational changes of the complement component C5 to a functionally active C5b-like conformation [15–17]. Another hypothesis is that the destruction of neutrophils in the extracorporeal circuit results in release of intracellular components as mitochondria and content of granules that activate the complement system.
The aim of the present study was to investigate whether fragmented neutrophils or isolated resting or activated neutrophils could contribute to complement activation in vitro. We also studied whether platelets–neutrophil interaction could modify neutrophil-dependent complement activation as suggested in an earlier study .
Materials and methods
Isolation of neutrophils and platelets. Neutrophil granulocytes and platelets were isolated from fresh buffy coats from informed volunteer donors, blood type A, Rh D positive (Blood Bank, St. Olav’s Hospital, Trondheim, Norway). For platelet isolation, equal volumes of buffy coat and pooled lepirudin-anticoagulated plasma (50 μg/l Refludan, Schering AG, Berlin, Germany) from four blood type A, Rh D-positive donors who screened negatively for irregular blood group antibodies, were carefully mixed and then centrifuged for 10 min at 200 × g. Lepirudin, a recombinant hirudin was used for anticoagulation because it has no effect on complement activation . The use of pooled plasma permitted employment of the same batch in all experiments, removing a potential source of variation. Platelet-rich supernatant was stored with gentle agitation until use. Platelet isolation and storage were performed at room temperature (20–23 °C). Neutrophils were isolated from the buffy coats as previously described .
Neutrophil fragmentation. Neutrophils were resuspended in Tyrode’s buffer (137 mm NaCl, 2.8 mm KCl, 1 mm MgCl2, 0.4 mm Na2HPO4, 5.5 mm glucose, 12 mm HCO3, 10 mm Hepes and 0.35% bovine serum albumin) at 30 × 106/ml. Fifteen millilitre suspension was placed in a Cell Disruption Bomb (Parr Instrument Company, Moline, IL, USA) precooled to 4 °C to avoid neutrophil activation and exposed to a nitrogen pressure of 500 psi for 15 min, after which the pressure was rapidly returned to atmospheric. A specimen was examined under a microscope to confirm complete neutrophil fragmentation. A second batch of 15-ml neutrophil suspension was stimulated with 10−7mN-formyl-methionyl-leucyl-phenylanine (FMLP) for 15 min at 37 °C before cavitation as described above. A third batch of 15-ml cell suspension was sonicated on ice until complete cell fragmentation. Samples for measurement of neutrophil degranulation products were removed immediately before and after fragmentation of the neutrophils, centrifuged at 250 × g for 7 min at 4 °C and stored at −70 °C. Three millilitre of each neutrophil cavitate/sonicate was centrifuged at 16,100 × g for 5 min and the supernatants were included in the incubation experiments. These supernatants and remaining uncentrifuged whole cavitates/sonicates were kept on ice until use.
Complement activation by neutrophil cavitate/sonicate. To study complement activation by fragmented neutrophils, increasing amounts (5–250 μl) of complete (uncentrifuged) cavitate/sonicate or supernatants from centrifugation were added to lepirudin-anticoagulated plasma (250 μl) before incubation with slight agitation at 37 °C for 30 min. As controls, one sample of lepirudin-anticoagulated plasma was kept at 4 °C during the experiment and another control incubated at 37 °C for 30 min. Tyrode’s buffer was used to achieve the same total volume of 500 μl in all samples. Ethylenediaminetetraacetic acid (EDTA) was immediately added to a final concentration of 15 mm at termination of the experiment. Samples were stored in aliquots at −70 °C for later analysis. These experiments were performed with neutrophils from four different donors.
Complement activation by neutrophils and platelets. To examine whether intact neutrophils alone or together with platelets could induce complement activation in vitro, lepirudin-anticoagulated plasma with neutrophils or platelet and neutrophils combined was incubated in polypropylene tubes at 37 °C for 15 min. The final concentration of platelets and neutrophils used in the experiments was 200 × 109/l and 2 × 109/l, respectively, i.e. within the physiological range. Lepirudin-anticoagulated plasma alone was used as control. To further investigate the importance of neutrophil activation in complement activation, additional experiments were performed with addition of 10−7m FMLP or 40 ng/ml phorbol myristate-acetate (PMA). Baseline samples of plasma and cell suspensions were removed before incubation and addition of EDTA performed as described above. Samples were immediately centrifuged at 250 × g for 7 min at 4 °C and the supernatants stored in aliquots at −70 °C for later analysis. These experiments were performed with neutrophils from four different donors.
Quantification of complement activation and neutrophil degranulation products. Activation of the classical complement pathway was quantified in an enzyme immunoassay for C1rs-C1-inhibitor complexes , and C4bc, indicating classical as well as lectin pathway activation, was measured as previously described . Activation of the alternative pathway was detected by quantifying the alternative convertase C3bBbP as previously described . Activation of C3 (C3bc) was quantified in an enzyme immunoassay using a monoclonal antibody specific for a neoepitope exposed in C3b, iC3b and C3c . Activation of the terminal pathway was quantified by measuring C5a (CBA kit for human anaphylatoxins; BD Pharmingen, San Diego, CA, USA) and by an enzyme immunoassay for detection of the soluble terminal complement complex (sC5b-9) using a monoclonal antibody specific for a C9 neoepitope exposed in the complex . The degranulation products myeloperoxidase and lactoferrin from the primary and secondary granules of the neutrophils were quantified in enzyme immunoassays as described earlier [24, 25].
Statistics. Non-parametric statistics were used due to non-normal distribution of variables and few observations. Results are reported as median ± 95% confidence interval (six or more observations) or range (less than six observations). By definition, confidence intervals and ranges are always wider than standard errors of the mean. P-values <0.05 were considered statistically significant. For comparisons of parallel interventions within a series of experiments, Friedman’s non-parametric analysis of variance was employed . SPSS-PC software version 13 was used for statistics (SPSS-PC software, Chicago, IL, USA).
The FMLP stimulation of neutrophils before cavitation led to significant degranulation as indicated by increased concentrations of myeloperoxidase and lactoferrin in the media (P < 0.05, Fig. 1; Table S1). However, after cavitation or sonication, the concentrations of degranulation products were comparable in the supernatants from unstimulated, FMLP-stimulated and sonicated neutrophils (P = 0.17), indicating similar total release and/or granule damage.
Complement activation by neutrophil cavitate/sonicate and supernatants
There was no formation of C1rs-C1 inhibitor complexes during incubation of lepirudin-anticoagulated plasma with or without addition of neutrophil cavitate, sonicate or supernatants (Table S2A). There was significant formation of the other measured complement activation products in plasma during incubation at 37 °C (Figs. 2 and 3; Table S2B–F). The highest doses of cavitate and sonicate induced the formation of C4bc (P < 0.001) without differences in the maximal C4bc concentration (Fig. 2A). With the supernatants, significant C4bc formation was only found for FMLP-stimulated cavitate (P < 0.001, Fig. 2B). The maximal C4bc concentrations were not significantly different between whole cavitate/sonicate and supernatant for any of the three treatments. The highest doses of cavitates and sonicate also induced formation of C3bPBb (P < 0.05–0.001, Fig. 2C), and the maximal concentrations were lower with the FMLP-stimulated cavitate than with the sonicate (P < 0.001). The supernatants did not induce significant formation of C3bPBb (Fig. 2D), resulting in significantly higher maximal C3bPBb concentrations with the whole cavitates than with the cavitate supernatants (P < 0.05–0.001), but not with whole sonicate compared to sonicate supernatant. C3bc behaved similarly to C3bPBb although less statistical significance was observed (Fig. 2E,F; Table S2D). Taken together, these results may indicate little activation by fragmented neutrophils of the classical pathway but rather activation of the lectin and alternative pathways, as reflected by formation of C4bc and C3bPBb and the lack of increase in C1rs-C1 inhibitor complexes.
The highest dose of cavitates and sonicates induced formation of C5a and the terminal complement complex (P < 0.01–0.001) without differences in the maximal sC5b-9 concentration among the various preparations (Fig. 3A,C; Table S2E,F). Only the supernatants from unstimulated neutrophils induced C5a formation and from sonicated neutrophils induced sC5b-9 formation (P < 0.05–0.01), resulting in significantly higher maximal sC5b-9 concentrations with the whole cavitates than with the cavitate supernatants (P < 0.01), but not with whole sonicate compared to sonicate supernatant. Taken together, these results show that high doses of fragmented neutrophils induced terminal complement activation.
Complement activation by neutrophils and platelets
For plasma alone, plasma with neutrophils and plasma with neutrophils and platelets, there was a significant increase in all the measured complement activation products apart from C1rs-C1 inhibitor complexes in all samples and C5a in the samples with neutrophils and platelets combined during incubation at 37 °C (Fig. 4; Table S3A–F). There were no differences in the concentrations of C1rs-C1 inhibitor complexes whether the samples were stimulated with FMLP or PMA or left unstimulated (Fig. 4A; Table S3A). However, for the stimulated samples, C1rs-C1 inhibitor complex concentrations were significantly higher with neutrophils and platelets combined than with neutrophils alone (P < 0.05–0.01). A similar stimulatory effect of platelets on complement activation was seen with respect to the formation of C4bc, where it was evident also for unstimulated samples (Fig. 4B; Table S3B). Unstimulated neutrophils induced higher concentrations of C3bPBb than plasma alone (P < 0.05), and formation of C3bPBb was lower when neutrophils and platelets were combined whether the samples were stimulated or not (P < 0.01, Fig. 4C; Table S3C). A similar lowering of complement activation by neutrophils and platelets combined was seen for C3bc when the samples were stimulated with PMA (P < 0.001, Fig. 4D; Table S3D).
There were no differences in C5a among the samples incubated at 37 °C (Fig. 4E; Table S3E). The formation of sC5b-9 in the unstimulated samples was slightly lower with platelets and neutrophils combined than for neutrophils alone (P < 0.05, Fig. 4F; Table S3F). This effect was not present in the stimulated samples.
To further evaluate the mentioned differences between neutrophils alone and platelets and neutrophils combined, control experiments comparing complement activation between platelets alone and platelets and neutrophils combined were performed. All results were equivalent between these two interventions (n = 3, data not shown), indicating that the platelets were responsible for the activation.
Complement activation by fragmented neutrophils
The present study showed that addition of high doses of whole cavitate or sonicate of neutrophils induced complement activation in plasma. Both activation of the alternative and the lectin pathway as well as terminal complement activation were observed. Activation was substantially less with supernatants from neutrophil cavitate or sonicate, indicating that liberated substances like proteases from the neutrophils were of less importance than for example membrane fragments and organelle components in activation of the complement cascade. There were no differences whether the neutrophils were prestimulated with FMLP or not, indicating that FMLP induced release only and not synthesis of new complement-activating substances. Thus, most likely the amount of degranulation products in the final cavitates was equivalent and cell disruption was total. We therefore suggest that the present model is relevant for clinical situations where neutrophils are inactivated or abruptly activated, like during cardiopulmonary bypass with sudden contact between blood and artificial surfaces. It may not be applicable where neutrophils are continuously activated, change phenotype and induce protein synthesis, like in septicaemia.
The clinical relevance of these findings become clear first when considering the degree of cell disruption needed to achieve this activation. We added 5, 10, 50 or 250 μl of cavitate that contained the fragments of approximately 0.15 × 106, 0.3 × 106, 1.5 × 106 or 7.5 × 106 neutrophils to 250 μl of plasma. With a neutrophil concentration in the circulation of 4 × 106/ml cells, this would amount to disruption of approximately 15, 30, 150 or 750% of neutrophils. Our dose studies therefore demonstrate that neutrophil fragmentation probably has limited impact on complement activation in the clinical setting of cardiopulmonary bypass. The situation may be different locally in tissues, e.g. during formation of an abscess.
Complement activation by intact neutrophils
Our study showed that the addition of neutrophils to plasma could induce alternative complement activation, but very little formation of other activation products. Various mechanisms that could lead to neutrophil activation of complement have been hypothesized. Enzymatic alterations in C5 exposing a binding site for C6 or producing a C5b-like molecule and resulting in formation of a C5b-9 complex has been indicated in various experimental situations, e.g. using isolated human leucocyte elastase  or through the myeloperoxidase-halide system . Another group also found support for neutrophil-dependent activation of C5 in vitro while studying neutrophil activation by the cellulosic dialyser membrane cuprophan . Furthermore, cleavage of C5 by isolated human neutrophils or rat alveolar macrophages, both activated with PMA, has been shown in vitro and indicates that C5a can be generated directly in the lung independent of the plasma complement system . Furthermore, a novel pathway of direct activation of C5 in the absence of C3 was demonstrated, leading to increased coagulation . Considering our results, other mechanisms must operate because cavitated neutrophils and granule constituents liberated to the fluid phase slightly increased the alternative convertase C3bBbP. This could be due to increased C3 hydrolysis, which efficiently is controlled by the regulatory proteins protecting against further activation of the system.
Our study demonstrated that the effects of platelets on complement activation in plasma were more pronounced. The combination of neutrophils and platelets induced more classical complement activation (C1rs-C1 inhibitor complexes and C4bc) than did neutrophils alone, but the effect was opposite for alternative activation (C3bPBb). The control experiments indicated that the differences between neutrophils alone and neutrophils and platelets combined were due to the platelets and not to the neutrophils. As the overall effect was a somewhat blunted response of C3bc formation and no changes in terminal complement activation, the influence of platelets on complement activation was not further investigated. The classical pathway activation induced by platelets is in accordance with a study demonstrating that the classical pathway may be activated by C1q binding to a receptor on the platelet surface . The investigation of platelet effects on the complement system has certain limitations as the thrombin inhibitor hirudin was used for anticoagulation. Thrombin has both pro- and anticoagulant effects and can activate platelets. The choice of anticoagulation is difficult, as alternative anticoagulants like citrate, EDTA and heparin significantly influence complement activation. The results must therefore be evaluated with this in mind. Nonetheless, as the main focus of the study was on neutrophils and the complement system, hirudin was regarded the most suitable anticoagulant.
Altogether our results suggest that complement activation by neutrophil-related mechanisms may be of little significance in clinical situations like cardiopulmonary bypass, as cell destruction is not sufficient to induce activation, and platelets reduce the neutrophil-dependent increase in alternative complement activation. Platelets are, however, able to activate the classical pathway, even though the downstream effect on C3 and the terminal pathway was modest.
Torill Anita Weisethaunet, Grethe Bergseth, Dorte Kristiansen and Judith Krey Ludviksen provided excellent technical assistance. The study was supported by the Central Norway Regional Health Authority, the Family Blix Foundation and the Odd Fellow Foundation.