Temporal studies on the deposition of complement on human colostrum IgA and serum IgG immobilized on methylated silicon

The blood compatibility challenge. Part 2: Protein adsorption phenomena governing blood reactivity

The adsorption of proteins is the initiating event in the processes occurring when blood contacts a "foreign" surface in a medical device, leading inevitably to thrombus formation. Knowledge of protein adsorption in this context has accumulated over many years but remains fragmentary and incomplete. Moreover, the significance and relevance of the information for blood compatibility are not entirely agreed upon in the biomaterials research community. In this review, protein adsorption from blood is discussed under the headings "agreed upon" and "not agreed upon or not known" with respect to: protein layer composition, effects on coagulation and complement activation, effects on platelet adhesion and activation, protein conformational change and denaturation, prevention of nonspecific protein adsorption, and controlling/tailoring the protein layer composition. STATEMENT OF SIGNIFICANCE: This paper is part 2 of a series of 4 reviews discussing the problem of biomaterial associated thrombogenicity. The objective was to highlight features of broad agreement and provide commentary on those aspects of the problem that were subject to dispute. We hope that future investigators will update these reviews as new scholarship resolves the uncertainties of today.

The inflammatory response to cardiopulmonary bypass: a therapeutic overview

The demographic of cardiac surgery patients continues to evolve to include older, sicker candidates, all the while maintaining an expectation of excellent outcomes. These latter results can only be achieved by the parallel advancement and re-examination of the technology of cardiopulmonary bypass (CPB); the key tool used daily by surgical teams worldwide. In this review, we will provide an overview of integrated therapeutic strategies that can be utilized to minimize the complex and myriad changes related to inflammation after CPB with the understanding that this may abrogate the detrimental end-organ and systemic effects of blood activation. Therapeutic strategies specifically related to the technology can be classified into those targeting biomaterial dependent or independent processes. The former can be addressed by the utilization of currently available biocompatible surfaces such as with heparin-coated circuits, phosphorylcholine-coated circuits (‘biomembrane mimicry’) and circuits composed of copolymers containing surface-modifying additives. The most important strategies related to biomaterial independent activation include the modification of techniques related to cardiotomy blood management and blood filtration. Finally, all of these strategies must be integrated and tailored with complementary pharmacologic agents such as aprotinin and steroids to optimize anti-inflammatory synergism. Only if we are armed with a comprehensive knowledge of the molecular and cellular basis for these strategies will we be able to continue to evolve our treatment in parallel with our patients to achieve these goals.

Complement inhibition in biomaterial- and biosurface-induced thromboinflammation

Therapeutic medicine today includes a vast number of procedures involving the use of biomaterials, transplantation of therapeutic cells or cell clusters, as well as of solid organs. These treatment modalities are obviously of great benefit to the patient, but also present a great challenge to the innate immune system, since they involve direct exposure of non-biological materials, cells of non-hematological origin as well as endothelial cells, damaged by ischemia-perfusion in solid organs to proteins and cells in the blood. The result of such an exposure may be an inappropriate activation of the complement and contact/kallikrein systems, which produce mediators capable of triggering the platelets and PMNs and monocytes, which can ultimately result in thrombotic and inflammatory (i.e., a thrombo-inflammatory) response to the treatment modality. In this concept review, we give an overview of the mechanisms of recognition within the innate immunity system, with the aim to identify suitable points for intervention. Finally, we discuss emerging and promising techniques for surface modification of biomaterials and cells with specific inhibitors in order to diminish thromboinflammation and improve clinical outcome.

Currently available biomaterials for use in cardiopulmonary bypass

Cardiopulmonary bypass (CPB) represents one of the most important technical innovations in healthcare history, yet the systemic responses to CPB remain a fundamentally unresolved problem. Study of the blood-biomaterial interaction and development of biocompatible materials is intimately related to efforts to optimize patient outcome following CPB. This article reviews the design innovations in biomaterial surfaces that have been introduced into clinical practice in an attempt to ameliorate the detrimental consequences of CPB, contrasting the actual clinical improvements and patient benefits achieved against those predicted on the basis of theory and in vitro testing. Some discussion of the underlying mechanisms of action as presently understood is provided and the current limitations of biomaterial-dependent strategies to improve outcome following CPB are addressed.

Autoregulation of thromboinflammation on biomaterial surfaces by a multicomponent therapeutic coating

Activation of the thrombotic and complement systems is the main recognition and effector mechanisms in the multiple adverse biological responses triggered when biomaterials or therapeutic cells come into blood contact. We have created a surface which is auto-protective to human innate immunity by combining three fundamentally different strategies, all developed by us previously, which have been shown to induce substantial, but incomplete hemocompatibility when used separately. In summary, we have conjugated a factor H–binding peptide; and an ADP-degrading enzyme; using a PEG linker on both material and cellular surfaces. When exposed to human whole blood, factor H was specifically recruited to the modified surfaces and inhibited complement attack. In addition, activation of platelets and coagulation was efficiently attenuated, by degrading ADP. Thus, by inhibiting thromboinflammation using a multicomponent approach, we have created a hybrid surface with the potential to greatly reduce incompatibility reactions involving biomaterials and transplantation.

Surface plasmon resonance in monitoring of complement activation on biomaterials

When artificial materials come into contact with blood, various biological responses are induced. For successful development of biomaterials used in biomedical devices that will be exposed to blood, understanding and control of these interactions are essential. Surface plasmon resonance (SPR) spectroscopy is one of the surface-sensitive optical methods to monitor biological interactions. SPR enables real-time and in situ analysis of interfacial events associated with biomaterials research. In this review, we describe an SPR biosensor and its application to monitor complement activation onto biomaterials surface. We also discuss the effect of surface properties of the material on complement activation.

Solid-phase classical complement activation by C-reactive protein (CRP) is inhibited by fluid-phase CRP–C1q interaction

C-reactive protein (CRP) interacts with phosphorylcholine (PC), Fcgamma receptors, complement factor C1q and cell nuclear constituents, yet its biological roles are insufficiently understood. The aim was to characterize CRP-induced complement activation by ellipsometry. PC conjugated with keyhole limpet hemocyanin (PC-KLH) was immobilized to cross-linked fibrinogen. A low-CRP serum with different amounts of added CRP was exposed to the PC-surfaces. The total serum protein deposition was quantified and deposition of IgG, C1q, C3c, C4, factor H, and CRP detected with polyclonal antibodies. The binding of serum CRP to PC-KLH dose-dependently triggered activation of the classical pathway. Unexpectedly, the activation was efficiently down-regulated at CRP levels > 150 mg/L. Using radial immunodiffusion, CRP-C1q interaction was observed in serum samples with high CRP concentrations. We propose that the underlying mechanism depends on fluid-phase interaction between C1q and CRP. This might constitute another level of complement regulation, which has implications for systemic lupus erythematosus where CRP is often low despite flare-ups.

Fluoroapatite glass-ceramic coating on alumina: surface behavior with biological fluids

The results of a surface analysis performed on a fluoroapatite-based glass ceramic (SAF) also coating a full-density alpha-alumina substrate (SAF-alumina coating) are presented. These two materials have also been evaluated after soaking in simulated body fluid to understand their ability to induce hydroxyapatite growth on them. Aiming to understand the fluoroapatite glass-ceramic interaction with some plasma proteins, in the second part of this study, fibronectin, albumin, immunoglobulin G, IgA, and complement factor C3c SAF binding have been evaluated; surface activity on complement activation has also been quantified. SAF-alumina coating provides good sites for the nucleation and growth of an apatite layer, equivalent to the mineral component of bone and binds preferentially plasma fibronectin, which is well known to enhance cell adhesion and spreading. Moreover, SAF-alumina coating reduces alumina complement activation directly or via reduced IgA binding. Alumina was shown to bind the same C3 fragments as Zymosan, used as complement activating control, and to induce increased levels of serum soluble iC3b and Bb. A mechanical resistant material with enhanced bioactivity, bone integration, and reduced inflammatory potential respect to alumina has been obtained.

Cardiopulmonary bypass technology transfer: musings of a cardiac surgeon

The development of cardiopulmonary bypass (CPB) has been one of the greatest technical advancements in cardiovascular medicine. With heparin anticoagulation, this device can safely replace the circulatory and gas-exchanging functions of the heart and lung, facilitating complex cardiac operations. Limitations still exist however, related to blood reactions at the biomaterial surface, such as cell activation, inflammation and low-grade thrombosis. In this brief review, the thought processes which paralleled the development of CPB biocompatible surfaces such as heparin-coating, will be explored, as well as current theories on the suspected mechanisms by which heparin-coated surfaces act as an anti-inflammatory device during CPB. Results with new surfaces for CPB designed to capitalize on superior protein adsorption properties, such as surface modifying additive (SMA) and poly (2-methoxyethylacrylate) (PMEA), will also be described. Finally, the significance of biomaterial-independent blood activation will be discussed, emphasizing the current need to develop strategies utilizing optimal biomaterials, modified surgical technique and pharmacologic therapy to minimize the systemic complications of CPB.

Interactions between surface-bound actin and complement, platelets, and neutrophils

Actin exists as globular (G) monomers or polymeric filaments (F) in the cytoplasm of eukaryotic cells, mediating cell morphologic changes and motility. Large amounts of this protein may be released out to the extracellular compartment during tissue injury, but little is known about its role in biomaterial-related inflammation. We immobilized actin to methylated glass, methylated and aminated silicon, and gold model surfaces and studied the subsequent blood serum deposition and complement activation, generation of reactive oxygen species (ROS), and adhesion and aggregation of neutrophils and platelets. Null ellipsometry showed that approximately one monolayer of G-actin can be immobilized onto the model surfaces and that actin in buffer polymerized on top of this by the addition of K(+) and Mg(2+) ions to form a thicker layer of firmly bound F-actin. After serum incubation, F-actin bound low amounts of anti-complement factor 1q (anti-C1q). Cell responses upon contact with actin-coated surfaces were analyzed by luminol-amplified chemiluminescence, lumi-aggregometry, and fluorescence microscopy. It was shown that surface-triggered aggregation, spreading, and generation of ROS are down-regulated and comparable to the response by adsorbed albumin. However, F-actin on gold surfaces recruited platelets in a C1q-dependent manner. We conclude that in vitro adsorbed actin is a weak complement, platelet, and neutrophil activator, but that F-actin associates with both C1q and platelets.

Surface Modified Cardiopulmonary Bypass Circuits: Modifying the Inflammatory Response

As a consequence of an aging population demographic, now more than ever, researchers in cardiac surgery must focus on means to improve the methods and technologies related to cardiopulmonary bypass. This review presents a classification of the currently available options for biomaterial modification for cardiopulmonary bypass circuits. Hypotheses are given relating the mechanism of action by which some of these surfaces afford improved biocompatibility. Finally, nonpharmacologic biomaterial-independent strategies for minimizing the effects of cardiopulmonary bypass, such as the use of hemofiltration and leukocyte filtration, and the minimization of the use of cardiotomy suction blood are outlined.

Blood protein adsorption onto chitosan

Chitosan was recently indicated to enhance osteogenesis, improve wound healing but to activate the coagulation and the complement systems. In the present study approximately 10 nm thick chitosan film were prepared on aminopropyltriethoxysilane (APTES) coated silicon. The surfaces were incubated in serum or plasma and subsequently in antibodies towards key complement and contact activation of coagulation proteins. The deposited amounts were compared with those on hydrophilic and hydrophobic silicon, APTES and IgG coated reference samples. Although large amounts of serum deposited to chitosan only a weak transient activation of the complement system and no activation of the intrinsic pathway was observed. Upon acetylation the chitosan layer became a strong activator of the alternative pathway of the complement. After incubation in human plasma anti-fibrinogen deposited onto chitosan but not onto the acetylated chitosan, a finding that may explain previous observations of procoagulant activity by chitosan.

On the binding of complement to solid artificial surfaces in vitro

Since the realization of a complement activation capacity by artificial surfaces upon contact with blood, a common belief has evolved that charged nucleophilic surface groups such as amine (-NH2) and hydroxyl (-OH) react with and eventually bind to the internal thioester in complement factor 3 (C3). A covalent amide or ester linkage is thereby supposed to form between C3b and the surface itself. In this report, we present complement surface binding data by null-ellipsometry for two nucleophilic surfaces (-NH2 and -OH), for surfaces with immunoglobulin G (IgG) covalently bound, and for IgG spontaneously pre-adsorbed to hydrophobic silicon. The results reveal that the plasma proteins that were deposited during complement activation became eluted by sodium dodecyl sulfate. Hence the direct covalent binding between C3 and solid nucleophilic surfaces seems to be only of moderate importance, at least during shorter serum incubations. This strongly suggests that the prevalent covalent linkage model between solid artificial surfaces and C3b is not accurate. Instead we suggest a more pronounced role for C3 associations to other adsorbed proteins and or electrostatic and hydrophobic protein-surface interactions.

Protein adsorption to oligo(ethylene glycol) self-assembled monolayers: experiments with fibrinogen, heparinized plasma, and serum

Low protein adsorption is believed advantageous for blood-contacting materials and ethylene glycols (EG)-based polymeric compounds are often attached to surfaces for this purpose. In the present study, the adsorption of fibrinogen, serum, and plasma were studied by ellipsometry on a series of well-defined oligo(EG) terminated alkane-thiols self-assembled on gold. The layers were prepared with compounds of the general structure HS-(CH2)15-CONH-EGn, where n = 2, 4, and 6. Methoxy-terminated tri(EG) undecanethiol and hydroxyl-terminated hexadecanethiol self-assembled monolayers (SAMs) were used as references. The results clearly demonstrate that the adsorption depends on the experimental conditions with small amounts of fibrinogen adsorbing from a single protein solution, but larger amounts of proteins from serum and plasma. The adsorption of fibrinogen and blood plasma decreased with an increasing number of EG repeats and was temperature-dependent. Significantly less serum adsorbed to methoxy tri(EG) than to hexa(EG) and more proteins remained on the latter surface after incubation in a sodium dodecyl sulfate (SDS) solution, indicating a looser protein binding to the methoxy-terminated surface. All surfaces adsorbed complement factor 3 (C3) from serum and plasma, although no surface-mediated complement activation was observed. The present study points to the importance of a careful choice of the protein model system before general statements regarding the protein repellant properties of potential surfaces can be made.

What really is blood compatibility?

The criteria for nonthrombogenicity are classically defined as long clotting times and minimal platelet deposition. The inability to point to unequivocal progress in the development of truly nonthrombogenic materials, highlights the inadequacy if not actually invalidity of these criteria. Our approach is to define nonthrombogenicity in terms of: (1) a thrombin production rate constant, kp < 10(-4) cm s(-1); (2) low platelet consumption and low degree of platelet activation (e.g., microparticle formation); (3) perhaps some platelet spreading; and (4) low complement and leukocyte activation. Only when the target becomes clear, will it be possible to identify clear strategies for producing the materials we need.

In vitro preparation and ellipsometric characterization of thin blood plasma clot films on silicon

The wound-healing process around implants differs from that of a normal healing without the inserted material. In this work, the composition of a natural wound surface was mimicked through clotting of a thin human blood plasma film with approximate ellipsometric thickness of 100 nm onto differently pretreated silicon surfaces. Their stability was investigated by incubations in sodium dodecyl sulphate (SDS) solutions. The enzymatic clot degradation was induced through addition of human tissue plasminogen activator (t-PA) to the plasma and the surface protein remnants after the degradation were analyzed with polyclonal antibodies. The results show that the plasma films were not SDS resistant on hydrophilic silicon. However, stability was obtained after preparation on hydrophobic silicon or when albumin or fibrinogen was immobilized to silicon before the plasma incubations. Different surfaces bound different polyclonal antibodies after the clot film degradation. The methods indicate a simple means to improve or reestablish a normal tissue inflammatory response around biomaterials.

Ellipsometric in vitro studies on the activation of complement by human immunoglobulins M and G after adsorption to methylated silicon

Human serum immunoglobulin M (IgM) or human immunoglobulin G (IgG) were adsorbed to dichlorodimethyl silane (DDS) treated silicon. Subsequently, the model surfaces were incubated in normal-, complement factor 1q (C1q)-complement factor B or complement factor 2 (C2)-depleted human sera at 37 degrees C for up to 1.5 h. The serum deposition and binding of selected polyclonal complement antibodies into this layer were then quantified by null ellipsometry. Both types of precoated surfaces bound large amounts of anti-complement factor 3c (anti-C3c), anti-properdin and anti-C3d, after incubation in normal serum. In contrast to IgG coated surfaces, IgM coated surfaces bound no anti-C1q after the serum incubations and no anti-C3c deposition lag time was observed after incubations in EGTA serum. Upon immersions of IgM coated surfaces in the different sera, a rapid complement activation via a C1q factor B, and Ca(2+)-independent, but C2 dependent pathway, was indicated. When IgM was instead immobilized to APTES/glutaraldehyde surfaces, anti-C3c deposition was lower after incubations in EGTA than normal serum. The results suggest that, under the present experimental conditions, human IgM and IgG activate the complement system differently.

In situ complement activation by polyethylene wear debris

A frequent long-term complication of total joint arthroplasty is aseptic loosening, the end result of wear debris accumulation, synovitis, and osteolysis about the implant-bone or cement-bone interface. Complement, an effector system in plasma, synovial fluid, and tissue, has powerful chemotactic, inflammatory, and osteoclast-activating potentials. This study explored the complement-activating ability of polyethylene, a material used in joint implants. In vitro hemolytic assays using sheep red blood cells (E(sh)), human serum, and particulate polyethylene suggested alternative pathway complement activation, as well as polyethylene adsorption of activated complement components. These results were confirmed by enzyme-linked immunosorbent assay (ELISA) quantification of activated complement factors Bb and C3b. In situ double antibody immunoperoxidase staining for factors Bb, C3a, iC3b, and SC5-9 in synovial tissue from revision hip specimens showed localized alternative pathway activation and component adsorption. These results introduce a likely role for complement activation in particle-mediated recruitment, proliferation, and activation of macrophages during early events in osteolysis and implant loosening.

Complement activation on immunoglobulin G-coated hydrophobic surfaces enhances the release of oxygen radicals from neutrophils through an actin-dependent mechanism

Neutrophil granulocytes are among the first cells to encounter a plasma protein-coated implant and may through frustrated phagocytosis release toxic oxidative species. We used two model surfaces, hydrophobic and hydrophilic glass, to investigate the effects of plasma immunoglobulin G (IgG)-complement interactions for neutrophil adhesion and respiratory burst. The respiratory burst was measured with luminol-amplified chemiluminescence and cell adhesion was determined by labeling neutrophils with 2', 7'-bis-(carboxy-ethyl)-5(6)-carboxyfluorescein. We demonstrate that the IgG-triggered neutrophil adhesion and oxygen radical production is augmented in the presence of normal human serum, in particular on hydrophobic surfaces, indicating that complement factors enhance the neutrophil activation. We propose that the complement factors C3, C5a, and C1q are especially important for this amplification, but factor B is probably not. Disturbance of the actin filament dynamics with cytochalasin B or jasplakinolide blocked the neutrophil radical generation on all surfaces. However, these drugs did not affect the number of adherent neutrophils. We suggest that there is a synergistic interaction between adsorbed IgG, and the complement system, which amplifies the neutrophil acute inflammatory responses through a dynamic actin cytoskeleton on synthetic surfaces.

Effects of new polymer-coated extracorporeal circuits on biocompatibility during cardiopulmonary bypass

An inflammatory response due to bioincompatibility of extracorporeal circuits is a major clinical issue during cardiopulmonary bypass (CPB). By using a swine model, we determined whether new polymer-coated circuits, the blood-contacting surfaces of which are coated with poly(2-methoxyethylacrylate) (PMEA), would reduce the inflammatory response during CPB. Plasma bradykinin levels and the percentages of CD35-positive monocytes in PMEA-coated circuits were significantly lower than those in uncoated circuits during CPB. The amount of proteins adsorbed on the PMEA-coated circuits was significantly lower than that on the uncoated circuits (0.30 microg/cm2 versus 3.42 microg/ cm2). Almost no IgG, IgM, or C3c/d was detected in proteins adsorbed on the PMEA-coated circuits although these proteins were clearly detected in proteins adsorbed on the uncoated circuits. We concluded that PMEA coating could reduce complement activation during CPB by suppressing the adsorption of IgG and IgM, which activate C3 via the classical pathway, on the surface of the circuits.

Inflammatory cell recruitment, distribution, and chemiluminescence response at IgG precoated- and thiol functionalized gold surfaces

The role of complement activation by artificial surfaces relative to inflammatory response is not well understood. This study was performed to evaluate the inflammatory cell recruitment, distribution, and ex vivo metabolic activation of surfaces with different plasma protein adsorption and complement activation properties in vitro. The implants were (1) pure gold (reference), (2) albumin-precoated (3) IgG-precoated gold, and (4) 3-mercapto-1, 2-propanediol [mercaptoglycerol (MG)] and (5) glutathione (GSH) immobilized to gold. The implant disks were inserted subcutaneously in rats for 24 h, and the number of inflammatory cells that were recruited to the implant adjacent to the surrounding fluid phase (exudate) and the surfaces were quantified by DNA measurements. The oxidative burst was analyzed ex vivo using spontaneous and phorbol myristate acetate (PMA)-stimulated, luminol-enhanced chemiluminescence (CL). The in vitro surface-induced anti-rat C3 binding was evaluated by ellipsometry and antibody techniques after plasma incubations for 1 and 30 min. The ellipsometric results showed that immobilized mercaptoglycerol and IgG-coated, but not the immobilized glutathione or the reference Au, bound anti-C3. The in vivo results revealed that the largest amount of cells was associated with the IgG-coated surfaces, followed by immobilized GSH and MG, albumin-coated, and gold surfaces, respectively. No spontaneous ex vivo luminol-enhanced CL was recorded from the cells irrespective of surface functionality or localization. A down-regulation of surface-associated and exudate leukocyte CL was observed ex vivo, irrespective of surface functionality. The results do not indicate a clear relationship between the degree of complement activation in vitro and leukocyte recruitment and adhesion in vivo for differently functionalized surfaces.

Ellipsometric studies in vitro on kinetics of rat complement activation

The role of complement activation may be important during the early interactions between implantable materials and blood and during the acute inflammatory phase, but it is not well understood. This applies especially to rats that are extensively used in in vivo animal models for materials and surface testing. Features of the kinetics of rat complement activation were studied and compared with human complement by the ellipsometry and antibody techniques. The results indicate that the rat classical pathway is rapidly activated, but it is not as fast as the human system. The activation of the alternative pathway was observed within 5 min in the rat system and within 15 min for the human. Thus, the observations indicate substantial differences in the kinetics between the two species. This may influence the choice of the rat experimental model and the tissue response to materials during in vivo conditions.

Immobilized chicken antibodies improve the detection of serum antigens with surface plasmon resonance (SPR)

Surface plasmon resonance (SPR) and other refractive index and mass sensitive methods are, due to complement activation by mouse monoclonal antibodies and with concomitant high background signal, only rarely used for the detection of antibody-antigen interactions in the blood serum milieu. In the present study chicken IgY and mouse IgG were immobilized to a sensor chip CM5 dextran matrix and compared for their background signal and detection of serum antigen. Ellipsometry with antibodies adsorbed to methylated silicon surfaces was used as a complementary detection method. As expected, fundamental differences in binding properties between the two kinds of antibodies were observed. Mouse antibodies bound large quantities of human serum. Human C1q was detected on mouse IgG and the complement system was activated, as seen from the rapid C3 and properdin depositions. Chicken antibodies bound low quantities of human serum and no human C1q. Moreover, C3 and properdin deposited only after prolonged serum incubations. Addition of EDTA to serum reduced the background signal modestly for both IgG and IgY. Serum samples with different concentrations of human C3 were injected over surfaces with immobilized chicken anti-C3, and the response was measured by SPR. Small concentration differences (< 1.25 micrograms/ml) in a physiologically relevant range (1-40 micrograms/ml after 100 times dilution) could then be detected reproducibly. The SPR signal was totally obscured when a mouse monoclonal anti-C3 antibody was used for the detection.

Specific activation of platelets by surface-adsorbed plasma proteins

Platelets were isolated from human blood by Percoll density gradient centrifugation in a low Ca2+/high Mg2+ buffer. The buffer reversibly inactivates the cells during separation. The purity of the isolated cells (> 99%) was determined by flow cytometry, their viability was confirmed by fluorescein diacetate hydrolysis, and their morphology was studied with TEM. Plasma proteins were adsorbed onto hydrophobic glass surface, and pure platelets were added and incubated for up to 30 min at 37 degree C. Platelet activation was determined by cell spreading, formation of microparticles and surface exposure of CD62P indicating the release of alpha-granules. Surface-immobilized IgG was shown to cause the release of microparticles and cell lysis, in accordance with data published by others. Surface-immobilized vWF was shown to induce CD62P exposure on the platelet cell surface. The specificity of this response was demonstrated by adsorbing plasma proteins from normal and factor VIII-deficient plasma.