当前位置：文档库 › Binding of a model regulator of complement activation (RCA) to a biomaterial surface surface-bound f

Binding of a model regulator of complement activation (RCA) to a biomaterial surface surface-bound f

*Corresponding author.Tel.:#46-18-6114907;fax:#46-18-553149.E-mail address:bo.nilsson @klinimm.uu.se (B.Nilsson).

Abbre v iations:SPDP *N -succinimidyl 3-(2-pyridyldithio)propionate;RCA *regulator of complement activation;QCM-D *quartz crystal microbalance-dissipation;EIA *enzyme immunoassay;PAV *poly-

allylamine

Biomaterials 22(2001)2435}2443

Binding of a model regulator of complement activation (RCA)to a biomaterial surface:surface-bound factor H inhibits

complement activation

Jonas Andersson ,Rolf Larsson ,Ralf Richter ,Kristina Nilsson Ekdahl ,Bo Nilsson *

Section of Clinical Immunology,Uni v ersity Hospital,S-75185U ppsala,Sweden

Q-Sense AB,Holtermansgatan 1,S-41292Gothenburg,Sweden

Department of Chemistry and Biomedical Sciences,Uni v ersity of Kalmar,S-39182Kalmar,Sweden

Received 8March 2000;accepted 7December 2000

Abstract

The complement system is an important in #ammatory mediator during procedures such as cardiopulmonary bypass and hemodialysis when blood is exposed to large areas of biomaterial surface.This contact between blood and the biomaterials of implants and extracorporeal circuits leads to an in #ammatory response mediated by the complement system.The aim of this study was to assess the ability of a complement regulator (factor H)immobilised on a biomaterial surface to inhibit complement cascade mediated in #ammatory responses.The cross-linker N -succinimidyl 3-(2-pyridyldithio)propionate was used to immobilise factor H on a model biomaterial surface without a !ecting the biological activity of the inhibitor.Binding of factor H was then characterised using quartz crystal microbalance-dissipation (QCM-D)and enzyme immunoassays for products of complement activation:bound C3fragments and soluble C3a,sC5b-9,and C1s-C1INA.Immobilised factor H reduced the amount C3fragments deposited on the biomaterial surface after incubation with serum,plasma,or whole blood.In addition,lower levels of soluble C3a and sC5b-9were generated after incubation with whole blood.In summary,we have demonstrated that complement activation on a highly activating model surface can be inhibited by immobilised factor H and have de "ned prerequisites for the preparation of future biomaterial surfaces with immobilised regulators of complement activation. 2001Elsevier Science Ltd.All rights reserved.

Keywords:Biocompatibility;Blood;Complement activation;Factor H

1.Introduction

Contact between blood and a biomaterial surface trig-gers an in #ammatory response against the material.Whole blood }biomaterial contact takes place in many di !erent types of devices,including oxygenators,plas-mapheresis equipment,hemodialysers,catheters,stents,vascular grafts,miniature pumps,sensors,and heart aids [1,2].Contact between blood and biomaterial also oc-curs,at least initially,during the implantation of bio-materials into soft and hard tissues.

Although progress has been made in reducing side e !ects,many procedures are still associated with undesir-able in #ammatory responses.In particular,during car-diopulmonary bypass (CPB)and hemodialysis,exposure of large areas of biomaterials to blood gives rise to a systemic in #ammatory response.A major challenge is CPB,during which blood comes in contact not only with a biomaterial but also with the gas surface [3].A second complicating factor is the ischemia/reperfusion damage that occurs during and after the treatment.These factors generate an in #ammatory response which may cause hemostatic problems leading to neurological de "ciencies,cognitive disturbances and,in the worst cases,organ failure [4,5].Hemodialysis is also associated with whole body in #ammation,which is thought to contribute to the accelerated arteriosclerosis that occurs in these patients [6,7].

Many of these adverse reactions are related to activa-tion of the cascade systems of which the complement

system is an important element[8].Complement activa-tion causes activation of leukocytes and platelets,and conversely,inhibition of its activity leads to reduced platelet and leukocyte activation[9}12].Heparin coat-ings have been used to reduce the activation of the coagulation and the complement systems on bio-materials surfaces in contact with whole blood.However, the extent to which they decrease complement activation is not always satisfactory,particularly with regard to bound complement fragments[13,14].Conjugation of a surface with biologically active inhibitors of comp-lement activation is a potential approach to lowering the complement activation on the material surfaces. Factor H is a serum protein of150kDa which belongs to a family of soluble and membrane-bound regulators of complement activation(RCA).It has a"brillic structure consisting of20short consensus repeats(SCR)[15].The size of factor H is495A;34A,but the protein has a high degree of#exibility and by electron microscopy has been shown to be folded back on itself[16].Factor H inhibits complement activation by two mechanisms:First,by displacing of factor Bbfrom the alternative pathway convertase(decay accelerating activity)and,second,by serving as a cofactor of factor I,which mediates cleavage of C3b[17}19].There have also been reports that factor H inhibits the classical pathway by interfering with the C1complex;however,the importance of this interaction has not been fully investigated[20,21].The aim of the present study was to covalently bind factor H to a model biomaterial surface in order to evaluate the ability of surface-bound factor H to inhibit complement activation on a biomaterial surface in contact with whole blood.

2.Materials and methods

2.1.Puri x cation of factor H

Following precipitation of euglobulins,factor H was isolated from2units of fresh frozen plasma essentially according to the protocol of Hammer et al.The plasma was pooled from two individuals[22,23].

2.2.SPDP conjugation of proteins

Factor H or control bovine serum albumin(BSA; Intergen Company,Oxford,UK),at approximately 0.5mg/ml,in10m M phosphate-bu!ered saline pH7.4 (PBS),was mixed with an equal volume of0.1M phos-phate bu!er,pH7.9.After addition of2%(v/v)2mg/ml of N-succinimidyl3-(2-pyridyldithio)propionate(SPDP; Pierce Chemical Company,Rockford,IL,USA)in meth-anol,the mixture was incubated for30min at room temperature(RT).The unconjugated SPDP was re-moved on a PD-10column(Amersham Pharmacia Biotech,Uppsala,Sweden)equilibrated in PBS.2.3.Immobilisation of proteins on polystyrene

SPDP was immobilised on the surface of polystyrene Nunc-Immuno F96MaxiSorp plates(Nunc A/S,Ros-kilde,Denmark).Unless stated otherwise,all incubations were performed at RT.The plates were washed three times with water from a MilliQ UF plus(Millipore AB, Sundbyberg,Sweden)after each incubation step.

1.To provide primary amines the surface was coated

with a high molecular weight polyallylamine,PAV (Corline Systems AB,Uppsala,Sweden)which had been modi"ed to increase its adhesiveness.PAV

(0.025mg/ml in25m M borate bu!er,pH9.0)was

incubated with the microtitre plates for15min.

2.The plates were dried at503C for20min,and2mg/ml

SPDP in methanol was bound to the primary amines by incubation for30min.

3.Excessive amino groups were blocked by incubation

with acetic anhydride(10 l/ml in25m M borate bu!er, pH10.5)for5min.

4.The SPDP bound to the surface was reduced by

incubation with10mg/ml dithiothreitol(DTT)in water for30min.The surfaces were washed six times with water to produce the activated surfaces(there-after referred to as the model surface).

5.Native or SPDP-conjugated factor H or BSA,at the

desired concentrations,was incubated in PBS with the plates for1h.

6.The plates were washed three times with washing

bu!er(PBS containing0.05%(v/v)Tween20and

0.02%(v/v)Antifoam(Pharmacia,Uppsala,Sweden))

and three times with PBS.Thereafter,the plates were stored in PBS at43C until use.

2.4.Preparation of blood,plasma,and serum for complement assays

Serum was obtained from blood drawn from healthy blood donors which was allowed to clot for1h in glass tubes or for1.5h in polystyrene tubes and then centri-fuged at2000g for10min.Serum from15di!erent individuals were pooled.Alternatively,whole blood was collected in polystyrene tubes with a Corline heparin surface(Corline AB)and containing3IU/ml of soluble heparin(Bio Iberica,Barcelona,Spain)[14,24]. This blood was either incubated in surface-modi"ed plates as described below,or centrifuged at2000g for 10min to obtain plasma.Plasma and serum were stored at!703C.

https://www.wendangku.net/doc/f77160933.html,plement acti v ation

Samples of plasma or serum(150 l,diluted in veron-al-bu!ered saline(VBS),containing0.15m M Ca >and

2436J.Andersson et al./Biomaterials22(2001)2435}2443

0.5m M Mg >)or whole blood(250 l),were incubated at

373C for the desired time in polystyrene wells coated with

factor H or BSA.After incubation,the samples were

transferred to tubes containing EDTA at a"nal concen-

tration of10m M and centrifuged at2000g for10min.The

samples were frozen at!703C pending analysis.In

some experiments the complement inhibitors EDTA,

Mg-EGTA(0.2M EGTA,50m M MgCl )and Com-pstatin were added to the serum at a"nal concentration

of10m M,10m M,and55 M,respectively,prior to incu-

https://www.wendangku.net/doc/f77160933.html,pstatin was a kind gift from Dr https://www.wendangku.net/doc/f77160933.html,m-

bris,Dept.of Pathology,University of Pennsylvania,

Philadelphia,USA.It is a cyclic synthetic peptide which

is known to preferably inhibit the alternative pathway of

complement at this concentration[12,25].

2.6.Enzyme immunoassays(EIA)for the detection of factor

H and C3fragments bound to a surface and y uid-phase

C3a,C1s-C1INA,and sC5b-9complexes

In each EIA washing bu!er and working bu!er(i.e.

washing bu!er containing1%(w/v)BSA,were used.The

antibodies used for detection were biotinylated using

biotin-amidocaproate N-hydroxysuccimide ester(Sigma

Chemical Co,St Louis,MO,USA)as described pre-

viously[26].For all assays the plates were stained with

colour solution consisting of10mg phenylendiamine-

dihydrochloride(Sigma)in40ml0.1M citrate}phosphate

bu!er(pH5.0)with10 l30%H O ,for approximately 5min and the absorbance at492nm was measured. Details of each EIA are given below:

2.6.1.Surface-bound factor H

After conjugation with factor H the plates were washed three times with washing bu!er.The wells were saturated with300 l of working bu!er at373C for30min,and then 200 l of anti-factor H diluted1/250(The Binding Site, Birmingham,UK)followed by biotinylated anti-factor H diluted1/250in working bu!er were added to the plates and incubated for30min at373C,respectively. Antibody was detected with100 l HRP-conjugated streptavidin(Amersham,Buckinghamshire,UK)diluted 1/500and incubated for30min at373C,followed by staining.

2.6.2.Surface-bound C3fragments

After incubation with whole blood,serum or plasma as described above,the plates were washed three times with washing bu!er.The wells were saturated with300 l of working bu!er at373C for15min,then incubated with 120 l of anti-C3c diluted1/200(Dako AS,Glostrup, Denmark)and horseradish peroxidase(HRP)-con-jugated anti-C3c diluted1/400(Dako)in working bu!er for30min at373C,followed by staining as described above.2.6.3.C3a

Plasma was diluted1/1000in working bu!er and ana-lysed as described previously[27].The mAb4SD17.3 was used as the capture antibody.Bound C3a was detec-ted with biotinylated rabbit anti-C3a diluted1/150,fol-lowed by HRP-conjugated streptavidin1/500.Zymosan-activated serum,calibrated against a solution of puri"ed C3a,served as standard,and the values are given as ng/ml.

2.6.4.C1s-C1INA

Microtitre plates were coated overnight with anti-C1s antibodies(INCSTAR,Stillwater,MN,USA)diluted 1/200at43C.The plates were then saturated with300 l working bu!er for30min at373C,and100 l of sample, diluted to a"nal serum concentration of1/200in work-ing bu!er containing10m M EDTA,was incubated for1h at RT.Zymosan-activated serum served as a positive control.For the detection of C1s-C1INA complexes, 100 l biotinylated anti-C1INA(Dako)diluted1/200and incubated for1h at RT,and100 l HRP-conjugated streptavidin(Amersham)diluted1/500was then added and incubated with the plates for1h at RT,followed by staining.

2.6.5.sC5b-9

Plasma was analysed for sC5b-9using a modi"cation of the EIA described by Mollnes and coworkers[24,28]. Plasma diluted1/5was added to microtitre plates coated with anti-neoC9mAb MCaE11.sC5b-9was detected by a polyclonal anti-C5antibody diluted1/500(Dako),fol-lowed by HRP-conjugated anti-rabbit immunoglobulin diluted1/500(Dako).Zymosan-activated serum de"ned to contain40000arbitrary units was used as a standard.

2.7.Quartz crystal microbalance-dissipation

(QCM-D)analysis

The QCM-D technique relies on the fact that a mass adsorbed onto the sensor surface of a shear mode oscil-lating quartz crystal causes a proportional change in its resonance frequency.When the adsorbed material is non-rigid,an additional energy dissipation(viscous loss) is also induced.

Analysis of adsorption kinetics,by simultaneous measurement of both the frequency,f,and the energy dissipation,D,was performed on a QCM-D instrument (Q-Sense AB,Gothenburg,Sweden)which is described in detail elsewhere[29].5MHz sensor crystals,spin-coated with hydrophobic polystyrene,were used.Changes in D and f were measured on both the fundamental fre-quency(n"1,i.e.f+5MHz)and the"rst overtone (n"3,i.e.f+15MHz).

Changes in the frequency,f,re#ect the amount of mass coupled to the surface of the crystal.For thin,evenly

J.Andersson et al./Biomaterials22(2001)2435}24432437

Fig.1.E !ect of immobilised factor H and BSA on the binding of C3fragments to a model surface.Serial two-fold dilutions of pooled human serum were incubated for 30min at 373C in wells bearing immobilised native or SPDP-conjugated factor H or BSA as described in Materials and Methods.Binding of C3fragments was measured by EIA with absorbance being read at 492nm.In panel a each point represents duplicates and in panel bsingle samples.Panel a,b inding to im-mobilised factor H or BSA (220pmol/well):native non-conjugated factor H (}?});SPDP-conjugated factor H (}?});no protein (}?});native BSA (}ⅷ});SPDP-conjugated BSA (}*}).Panel b,e !ect of adding varying amounts of immobilised native factor H per well:100pmol (}?});110pmol (}?});55pmol (}*});28pmol (}ⅷ});14pmol (}?});7pmol (} });3pmol (}?});no factor H (}T }).

distributed and rigid "lms,an adsorption-induced fre-quency shift ( f )is related to mass uptake ( m )via the Sauerbrey equation [30], f "!(n /C ) m ,where C (equivalent to 17.7ng cm \ Hz \ )is the mass sensitiv-ity constant and n ("1,3,2)is the overtone number.However,for proteins adsorbed from the aqueous phase,one must also be aware that water hydrodynamically coupled to the adlayer is included in the measured mass uptake [31].The dissipation factor (D )re #ects frictional (viscous)losses induced by deposited materials such as proteins adsorbed on the surface of the crystal.Hence,changes in the viscoelastic properties of adlayers (e.g.induced by conformational changes)as well as di !erences between various protein }surface interactions can be monitored [32,33].

In cases in which the adsorption induces a signi "cant energy dissipation,a quantitative analysis of the data is desirable [34].In the present work we have applied a viscoelastic model developed by Voinova et al.[35].In brief,the model describes the response of the QCM-D sensor crystal to a viscoelastic layer covered by a Newto-nian liquid under no-slip conditions.In this model,the adlayer is represented by a "lm de "ned by a uniform thickness, ,density, ,shear viscosity, ,and shear elastic modulus, .Using the approach described by Ho o k et al.[36],the properties of the added layer can be obtained from the measured changes in frequency and dissipation if both the fundamental frequency (5MHz)and the overtone (15MHz)are measured.One must be aware,however,that the representation of the protein "lm with uniform thickness,density,shear elastic mod-ules and shear viscosity to some extent is a simpli "cation of the true "lm.Accordingly,the estimated values must be regarded as e !ective values describing the average properties of the adlayer.

3.Results

3.1.E w ect of v arious forms of immobilised factor H on the binding of C3fragments

To assess the binding of C3-fragments to immobilised factor H,model surfaces (polystyrene microtitre plates)bearing immobilised SPDP and treated with DTT were reacted with factor H,SPDP-factor H,BSA and SPDP-BSA;a non-protein treated surface served as a further control.Incubation of the plates with serial dilutions of pooled human serum,followed by EIA measurements of the amount of bound C3fragments was assessed revealed that both native and SPDP-conjugated factor H e $-ciently inhibited binding of C3fragments to the surface at all serum concentrations tested (Fig.1a).In contrast,neither native nor SPDP-conjugated BSA had any e !ect on the binding;similarly no inhibitory e !ect was seen for the control surface bearing immobilised SPDP alone.

We then identi "ed the optimal concentration of factor H for inhibiting the binding of C3fragments by varying the amount of factor H (0}220pmol per well):The con-centration of factor H could be lowered to 14pmol per well before the level of C3fragment binding began to increase (Fig.1b).Similar experiments using SPDP-treated factor H showed that the dose requirements were similar or higher (data not shown).Because of its e $cient inhibition of the binding of C3fragments,non-SPDP-treated (native)factor H (at 100pmol per well)was there-fore used for the rest of this study.

3.2.Characterisation of surface-bound factor H

Having established that surface-bound factor H could inhibit the binding of C3fragments to a model surface,we next characterised the immobilisation of factor H in greater detail.Surfaces were prepared with or without SPDP (Section 2.3,step 2,in Materials and methods)and

2438J.Andersson et al./Biomaterials 22(2001)2435}2443

Fig.2.Immobilisation of factor H to various surfaces as detected by EIA.Factor H was immobilised at varying concentrations (from 0to 100pmol factor H per well)to surfaces prepared with or without SPDP,in the presence or absence of DTT.Factor H was reacted in the presence of DTT with the amine surface prepared without SPDP (} })or conjugated with SPDP (}?}).Factor H was also allowed to interact with these same surfaces without DTT treatment (}?},with SPDP;}?},without SPDP).In panel a,factor H was in PBS containing 0.05%Tween;in panel b,factor H was in

PBS.

Fig.3.Binding of factor H (1100pmol/ml)to a SPDP-conjugated polystyrene surface detected by QCM-D.Upper panel:monitoring of factor H binding (changes in frequency, f )to a surface treated with DTT and to one which was not treated.Lower panel:the dissipation changes, D ,for the binding of factor H to the surfaces shown in the upper panel.Only the data for the overtone (15MHz)are shown.

were then incubated with or without DTT (step 4),result-ing in two variants each of the acetylated amine surface and the SPDP-conjugated acetylated amine surface,re-spectively.We then used an EIA for factor H and the QCM-D technique to analyse the binding of factor H to these surfaces.

The results of four EIA analysis indicated that when diluted in PBS,factor H bound to all four types of surfaces (Fig.2b);however,if 0.05%Tween (washing bu !er)was added to the factor H preparation and all the washes,there was no binding to the acetylated amine surfaces without SPDP (Fig.2a).Furthermore,the bind-ing in Tween was greater to the non-DTT-treated SPDP surface than to the DTT-treated one.Thus the most e !ective binding of factor H was to the non-DTT-treated SPDP-conjugated surface,in the presence and absence of Tween.

The binding characteristics of factor H to the two SPDP conjugated surfaces were then investigated in de-tail with the QCM-D technique.The binding kinetics as re #ected by both the frequency change, f ,and the dissipation change, D ,versus time are shown in Fig.3.

The analyses con "rmed that the non-DTT-treated sur-face-bound more factor H, f (15MHz)"!234Hz (1381ng/cm ),than did the DTT-treated one, f (15MHz)"!93Hz (549ng/cm ).Despite binding more factor H,the change in D was smaller ( D (15MHz)"1.4;10\ for the non-DTT treated surface than for the DTT-treated surface (1.9;10\ ),indicating that the layer formed by factor H was less elastic and viscous,and thus was more rigid,on the non-DTT-treated surface (Fig.3).The di !erences in the properties of the complete factor H layers (after the bu !er wash),quanti "ed according to the viscoelastic model,are shown in Table 1.

3.3.E w ect of immobilised factor H on classical and alternati v e pathway acti v ation

In order to investigate which activation pathway was a !ected by immobilised factor H,experiments were per-formed with pooled human serum and speci "c comp-lement inhibitors.The generated activation fragments were monitored using EIAs.

Incubation of human serum in SPDP-treated wells bearing immobilised factor H or BSA showed the depos-ition of C3on immobilised factor H was delayed for about 15min when compared to SPDP-conjugated BSA.

J.Andersson et al./Biomaterials 22(2001)2435}24432439

Table1

Parameters describing the factor H layer on the DTT-treated and the non-DTT-treated SPDP-conjugated surfaces(after bu!er wash),cal-culated according to the viscoelastic model.The uncertainty in the"t between the model and the measured data is estimated to be less than 10%

SPDP-conjugated surface

DTT-treated Non-DTT-treated Thickness, & 5.610.4nm Density, & 1.20 1.38kg m\ Viscosity, & 3.4;10\ 10.2;10\ N s m\ Elasticity, & 1.8;10 6.8;10 N m\

1.9;10\ 1.5;10\ s Characteristic relaxation

time, &" &/ &

Fig.5.Generation of soluble C1s-C1INA complexes in serially diluted serum incubated at 373C for 30min in wells with immobilised factor H (}T }),SPDP-conjugated BSA (} })or no ligand (}ⅷ}

Fig.6.Time-dependent e !ect of immobilised factor H on the binding of C3fragments from heparin plasma and whole blood to the model surface.Immobilised factor H was compared with BSA.Each timepoint represents qaudruplicates in blood and duplicates in plasma.Binding of C3fragments (panel a),generation of C3a (panel b)and generation of sC5b-9complexes (panel c)after incubation of heparin plasma with immobilised factor H (}?})or control BSA (}?})or of whole blood with immobilised factor H (} })or BSA (}?}).

(RCA), e.g.DAF,MCP,C4BP,CR1and factor H [37}41].These proteins are encoded by a cluster of genes on chromosome 1.Although the RCA family mem-ber factor H is a plasma protein,most of its functions are exerted at biological and arti "cial surfaces.Conse-quently,it has been suggested that the binding of factor H to a biomaterial surface de "nes the blood biocompati-bility of a biomaterial with respect to the complement system [42].Furthermore,although many RCA proteins and variants thereof are available in recombinant form,naturally occurring human factor H can be obtained in large quantities.This protein was therefore a particularly suitable candidate of the RCA proteins for immobilisa-tion on a biomaterial surface.

As a model surface we used polystyrene which was coated with a polyamine,followed by the binding of the cross-linker SPDP to the free amines.Remaining free amines were acetylated by acetic anhydride.Our initial experiments showed that SPDP-conjugated factor H could be covalently bound to this surface if the sur-face-bound SPDP was reduced by DTT.The biological e !ect of the coupling was monitored by its ability to reduce the binding of C3fragments to the surface after incubation in human serum.The corresponding surface without SPDP was much less biologically active (not shown)despite the fact it also bound appreciable quanti-ty of factor H (Fig.2b).However,in the presence of the detergent Tween 20,the binding of factor H was negli-gible (Fig.2a),indicating that this binding to surfaces without SPDP was non-covalent.

Early in our study we observed that native factor H also bound to the surface-bound and reduced SPDP with the same or even better biological activity than SPDP-conjugated factor H.We assume that the free thiol

groups of the reduced SPDP molecules attacked acces-sible internal disulphide bonds of the SCR domains of factor H.O !ering a favoured site at which the free thiol groups could react on the non-conjugated factor H mol-ecules as probably also resulted in a more uniform (and favourable)orientation of the molecule compared to the binding to more randomly dispersed SPDP molecules of the SPDP-conjugated factor H.Because of the relative e $ciency of native factor H,it was used in the remaining experiments.

Quantitation by both EIA and QCM-D showed that when the surface-conjugated SPDP was not reduced by DTT,it bound more factor H than did the surface which was subjected to DTT treatment (Figs.2and 3).Despite this higher binding,a much lower biological activity was obtained.The relationship between dissipation and fre-quency for the DTT-treated and the non-DTT-treated

J.Andersson et al./Biomaterials 22(2001)2435}24432441

surfaces showed that the conformation of factor H on the two surfaces di!ered.The dissipation shift was higher on the DTT-treated surface,while the opposite relationship was observed for the frequency shifts.On the surface with reduced SPDP,the viscoelasticity of factor H layer was higher,i.e.,the molecules were organised in a more#ex-ible way.In addition to the viscosity and elasticity,the density and the e!ective(hydrodynamic)thickness of the added layer also varied considerably:The factor H layer on the non-DTT treated surface was almost twice as thick as the layer on the DTT-treated surface(10.4nm compared to5.6nm);the corresponding densities are1.38 and1.2kg m\ .Both values are within the range of values commonly obtained for the density of water (1.0kg m\ )and proteins(1.4kg m\ ).These"ndings indicate that pure binding of factor H to the surface is not enough to reduce complement activation and that the orientation and conformation of the molecule is crucial for the function of factor H on the material surface. This conclusion is consistent with the three-dimensional structure of factor H,which is an extended"brilar molecule.

Incubation of serum,plasma or blood with the factor H-covered surfaces showed that the bound factor H ab-rogated the binding of surface-bound C3fragments and inhibited the generation of soluble C3a and sC5b-9.This inhibition of complement activation by factor H was mediated by both the alternative and the classical path-ways.As expected,when the classical pathway was blocked by EGTA,the deposition of C3fragments was retarded.Furthermore,under conditions in which only the classical pathway was operative(in the presence of Compstatin or at serum concentrations of10%(v/v)or less),immobilised factor H totally inhibited the classical pathway activation,as indicated by an abrogation of the generation of C1s-C1INA complexes and the binding of C3fragments.Although the way in which factor H in-hibits the classical pathway is unclear,these observations support previous studies which have identi"ed factor H as a C1q binding protein with the ability to inhibit the hemolytic activity of#uid-phase C1[20,21].

The degree of complement activation on a biomaterial surface depends upon the properties of the material that is used,e.g.the hydroxyl and amino groups available for C3binding and the balance between hydrophilic and hydrophobic groups.This study and previous studies from our laboratory using Compstatin to block the alter-native pathway activation[12]have shown how swiftly the classical pathway is activated on a material surface. They also show how e$ciently the alternative pathway feedback loop ampli"es the activation.These observa-tions suggest that activation of complement on a biomaterial surface takes place in two steps:First,comp-lement activation is initiated by either the classical or the alternative pathway,depending on the material involved. Second,once C3bis deposited on the surface,the alterna-tive pathway feedback loop has the potential to drasti-cally amplify the activation.This situation presumably explains why biomaterial surfaces are generally thought to mainly activate the alternative pathway independent of which pathway actually triggered the activation. It also suggests that inhibition of the initial activation of the triggering pathway is important if we are to inhibit the whole complement activation on the model surface. This two-step activation suggests that a complement inhibitor immobilised on a biomaterial surface should, like factor H,have the ability to inhibit both activation pathways.

Taken together,our data demonstrate that an RCA protein can be covalently linked to a biomaterial surface while maintaining its functional properties.However, further e!orts need to be made to improve both the inhibitors and the techniques by which they are bonded to the surface.

Acknowledgements

This study was supported by grants from the Go ran Gustafsson Research Foundation,King Gustaf V:s Re-search Foundation,The Swedish Rheumatism Associ-ation,Prof Nanna Svartz'Research Foundation,the Swedish Board for Industrial and Technical Develop-ment,and by Grants Nos.5647,11578and13002from the Swedish Medical Research Council.The authors also wish to thank Foozieh Ghazi for technical assistance in the preparation of factor H,and Ph.D.Graciela Elgue for the analysis of C3a and sC5b-9.

References

[1]Courtney JM,Lamba NM,Sundaram S,Forbes CD.Bio-

materials for blood-contacting applications.Biomaterials 1994;15:737}44.

[2]Ratner BD,Ho!man AS,Shoen FJ,Lemons JE.Biomaterials

science.New York:Academic Press,1996.p.484.

[3]Pekna M,Nilsson L,Nilsson-Ekdahl K,Nilsson UR,Nilsson B.

Evidence for iC3generation during cardiopulmonary bypass as the result of blood}gas interaction.Clin Exp Immunol 1993;91:404}9.

[4]Kirklin JK.The postperfusion syndrome:in#ammation and the

damaging e!ects of cardiopulmonary bypass.In:Tinker JH,edi-tor.Cardiopulmonary bypass:current concepts and controver-sies.Philadelphia:W.B.Saunders,1989.p.3026}41.

[5]Greeley WJ,Ungerleider RM.Assessing the e!ect of cardiopul-

monary bypass on the brain.Ann Thorac Surg1991;52:417}9.

[6]Wanner C,Zimmermann J,Quaschning T,Galle J.In#ammation,

dyslipidemia and vascular risk factors in hemodialysis patients.

Kidney Int(Suppl)1997;62:S53}5.

[7]Efendic E,Lindholm B,Bergstrom J,Stenvinkel P.Strong con-

nection between malnutrition,in#ammation and arteriosclerosis.

Improved treatment of renal failure if underlying factors are https://www.wendangku.net/doc/f77160933.html,kartidningen1999;96:4538}42.

[8]Mollnes https://www.wendangku.net/doc/f77160933.html,plement and biocompatibility.Vox Sang1998;

74:303}7.

2442J.Andersson et al./Biomaterials22(2001)2435}2443

[9]Rinder CS,Rinder HM,Smith BR,Fitch JC,Smith MJ,Tracey

JB,Matis LA,Squinto SP,Rollins SA.Blockade of C5a and C5b-9generation inhibits leukocyte and platelet activation dur-ing extracorporeal circulation.J Clin Invest1995;96:1564}72. [10]Himmelfarb J,McMonagle E,Holbrook D,Toth C.Soluble

complement receptor1inhibits both complement and granulocyte activation during ex vivo hemodialysis.J L abClin Med1995;126:392}400.

[11]Larsson R,Elgue G,Larsson A,Nilsson Ekdahl K,Nilsson UR,

Nilsson B.Inhibition of complement activation by soluble recom-binant CR1under conditions resembling those in a cardiopul-monary circuit:reduced up-regulation of CD11band complete abrogation of binding of PMNs to the biomaterial.Immunophar-macology1997;38:119}27.

[12]Nilsson B,Larsson R,Hong J,Elgue G,Nilsson Ekdahl K,Sahu

A,Lambris https://www.wendangku.net/doc/f77160933.html,pstatin inhibits complement and cellular activation in whole blood in two models of extracorporeal circu-lation.Blood1998;92:1661}7.

[13]Nilsson UR,Larm O,Nilsson B,Storm KE,Elwing H,Nilsson

Ekdahl K.Modi"cation of the complement binding properties of polystyrene:e!ects of end-point heparin attachment.Scand J Im-munol1993;37:349}54.

[14]Gong J,Larsson R,Nilsson Ekdahl K,Mollnes TE,Nilsson U,

Nilsson B.Tubing loops as a model for cardiopulmonary bypass circuits:both the biomaterial and the blood}gas phase interfaces induce complement activation in an in vitro model.J Clin Immu-nol1996;16:222}9.

[15]Ripoche J,Day AJ,Harris TJR,Sim RB.The complete amino

acid sequence of human complement factor H.Biochem J 1988;249:593}602.

[16]DiScipio RG.Ultrastructures and interactions of complement

factors H and I.J Immunol1992;149:2592}9.

[17]Weiler JM,Daha MR,Austen KF,Fearon DT.Control of the

ampli"cation convertase of complement by the plasma protein beta1H.Proc Natl Acad Sci USA1976;73:3268}72.

[18]Pangburn MK,Schreiber RD,Muller-Eberhard HJ.Human

complement C3binactivator:isolation,characterization,and demonstration of an absolute requirement for the serum protein

b eta1H for cleavage of C3band C4bin solution.J Exp Med

1977;146:257}70.

[19]Whaley K,Ruddy S.Modulation of the alternative pathway by

the 1H-globulin.J Exp Med1976;144:1147}63.

[20]Easterbrook-Smith SB,Wilson MR,Wines BD.RHP is antigeni-

cally related to factor H and binds to the globular heads of C1q.

Mol Immunol1992;29:1203}7.

[21]Holme ER,Qi M,Ahmed AE,Veitch J,Auda G,Whaley K.

Puri"cation and characterization of RHP(factor H)and study of its interactions with the"rst component of complement.Mol Immunol1992;29:957}64.

[22]Hammer CH,Wirtz GH,Renfer L,Gresham HD,Tack https://www.wendangku.net/doc/f77160933.html,rge

scale isolation of functionally active components of the human complement system.J Biol Chem1980;256:3995}4006.

[23]Nilsson UR,Mu ller-Eberhard HJ.Isolation of 1F-globulin and

from human serum and its characterization as the"fth compon-ent of complement.J Exp Med1965;122:277}98.

[24]Hong J,Nilsson Ekdahl K,Reynolds H,Larsson R,Nilsson B.

A new in vitro model to study interaction between whole blood

and biomaterials.Studies of platelet and coagulation activation and the e!ect of aspirin.Biomaterials1999;20:603}11.

[25]Sahu A,Kay BK,Lambris JD.Inhibition of human complement

by a C3-binding peptide isolated from a phage-displayed random peptide library.J Immunol1996;157:884}91.[26]Storm K-E,Arturson G,Nilsson UR.Puri"cation and character-

ization of porcine C3.Studies of the biologically active protein and its split products.Vet Immunol Immunopathol1992;34: 47}61.

[27]Nilsson Ekdahl K,Nilsson B,Pekna M,Nilsson UR.Generation

of iC3at the interface between blood and gas.Scand J Immunol 1992;35:85}91.

[28]Mollnes TE,Lea T,Froland SS,Harboe M.Quanti"cation of the

terminal complement complex in human plasma by an enzyme-linked immunosorbent assay based on monoclonal antibodies against a neoantigen of the complex.Scand J Immunol 1985;22:197}202.

[29]Rodahl M,Hook F,Krozer A,Brzezinski P,Kasemo B.Quartz-

crystal microbalance setup for frequency and Q-factor measure-ments in gaseous and liquid environments.Rev Sci Instr 1995;66(7):3924}30.

[30]Sauerbrey G.Verwendung von Schwingquarzen zur Wa gung

du nner Schichten und zur Mikrowa gung.Zeitschrift fu r Physik 1959;155:206}22.

[31]Caruso F,Furlong DN,Kingshott P.Characterization of ferritin

adsorption onto gold.J Coll Interf Sci1997;186:129}40. [32]Ho o k F,Rodahl M,Kasemo B,Brzezinski P.Structural changes

in hemoglobin during adsorption to solid surfaces:e!ects of pH, ionic strength,and ligand binding.Proc Natl Acad Sci USA 1998;95:12271}6.

[33]Ho o k F,Rodahl M,Brzezinski P,Kasemo B.Measurements

using the quartz crystal microbalance technique of ferritin mono-layers on methyl-thiolated gold:dependence of energy dissipation and saturation coverage on salt concentration.J Coll Interf Sci 1998;208:63}7.

[34]Rodahl M,Ho o k F,Fredriksson C,Keller CA,Krozer A,

Brzezinski P,Voinova M,Kasemo B.Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion.Faraday Discuss1997;107: 229}46.

[35]Voinova MV,Rodahl M,Jonson M,Kasemo B.Viscoelastic

response of layered polymer"lms at#uid}solid interfaces:con-tinuum mechanics approach.Physica Scripta1999;59:391}6. [36]Ho o k F,Kasemo B,Fant C,Elwing H.A multi-technique study

(QCM-D ellipsometry and SPR)of coupled water and viscoelastic e!ects during adsorption and subsequent cross-linking of an adhesive protein"lm.J Am Chem Soc2001,submitted for publication.

[37]McAleer M,Ripoche J,Day AJ,Sim RB.Alternative splicing in

factor H,a member of the RCA gene family.Genome 1988;30(Suppl1):136.

[38]Rodriguez de Cordoba S,Rubinstein P.New alleles of C4-binding

protein and factor H and further linkage data in the regulator of complement activation(RCA)gene cluster in man.Im-munogenetics1987;25:267}8.

[39]Seya T.Human regulator of complement activation(RCA)gene

family proteins and their relationship to microbial infection.

Microbiol Immunol1995;39:295}305.

[40]Holers VM,Kinoshita T,Molina H.The evolution of mouse and

human complement C3-binding proteins:divergence of form but conservation of function.Immunol Today1992;13:231}6. [41]Nangaku https://www.wendangku.net/doc/f77160933.html,plement regulatory proteins in glomerular dis-

eases.Kidney Int1998;54:1419}28.

[42]Rauterberg EW,Ritz E.Bioincompatibility of dialysis mem-

branes:factor H binding correlates inversely with complement activation indicating a local imbalance of involved pro-teases/anti-proteases.Adv Exp Med Biol1988;240:365}75.

J.Andersson et al./Biomaterials22(2001)2435}24432443