Quartz Crystal Microbalance-with Dissipation Monitoring (QCM-D) For

Biosensors and Bioelectronics 21 (2005) 79–86

Quartz crystal microbalance-with dissipation monitoring (QCM-D) forreal time measurements of blood coagulation density and immune

complement activation on artificial surfaces

Marcus Anderssona,∗, Jonas Anderssonb, Anders Sellborna, Mattias Berglina,Bo Nilssonb, Hans Elwinga

a Department of Cell and Molecular Biology/Interface Biophysics Lundberg Laboratory, G¨oteborg University, Box 462, SE-405 30 G¨oteborg, Swedenb Department of Oncology, Radiology and Clinical Immunology, Section of Clinical Immunology, Rudbeck Laboratory C5,

Uppsala University Hospital, SE-751 85 Uppsala, Sweden

Received 24 June 2004; received in revised form 24 September 2004; accepted 29 September 2004Available online 11 November 2004

Abstract

taneousa on the sensors consequenceo our QCM-D tz sensors lmetacrylate)a t we couldd n rate, (3)t ity can bea owever,t e QCM-Dw

the surfacea ed well fort©

K

1

s(r

oped(8

ad-iththens,

r

0d

A recently developed variant of quartz crystal microbalance (QCM) called QCM-with dissipation monitoring (QCM-D) allows simulnd simple measurements of changes in adsorbed mass as well as the viscoelastic property (D-factor) of deposited protein layersurface. We have taken the QCM-D technology a step further and demonstrated its advantages in the study of protein assembly as af surface induced immune complement activation, or contact activated blood coagulation. In the present study we have continuedinvestigations of surface assembly of fibrin clot formation and complement activation and incubated differently modified quar

urfaces in blood plasma and sera. Polymer surfaces used were spin-coated polyethylene, poly(ethylene terephtalate), poly(methynd poly(dimethylsiloxane). Also used were sputtered titanium and heparin grafted surfaces. In this investigation we found thaescribe the surface induced coagulation with four independent parameters: (1) Time of onset of coagulation, (2) fibrin depositio

otal frequency shift at stable plateau, and (4) fibrin clot density. The most important finding was that the blood plasma clot densssessed with the use ofD determinations and that the clot density varied significantly with the chemical composition of the surface. H

he D-factor did not give any new analytical information about the possible complement activation mechanisms. Nevertheless, thas found to be a reliable tool for the analysis of surface induced complement activation.We also compared the QCM-D technique with traditional enzyme immuno assay (EIA) measurements of soluble products from

ctivation of the complement and coagulation systems. We found that the results from EIA and QCM-D measurements correspondhe complement activation but not for the coagulation, probably due to the biological complexity of the coagulation system.

2004 Elsevier B.V. All rights reserved.

eywords:Quartz crystal microbalance; Blood coagulation; Complement activation; Enzyme immuno assay

. Introduction

Quartz crystal microbalance (QCM) is a high frequencyurface sensitive method for various biosensor applicationsMarx, 2003). Most applications monitor only changes ofesonant frequency (�f) of the adsorbed layers on the sensor

∗ Corresponding author. Tel.: +46 31 7732566; fax: +46 31 7732599.E-mail address:[email protected] (M. Andersson).

surfaces. Recently, a variation of QCM has been develwhich simultaneously allows registration of dissipationD)of the sensor signal (Rodahl et al., 1995; Hook et al., 199).The D-factor is related to the viscoelastic properties of thesorbed layer. The new QCM modification, called QCM wdissipation monitoring or QCM-D, has greatly increasedanalytical range of the instrument in different applicatiofor example in studies of single protein adsorption (Hook etal., 1998) and adsorption of phospholipid layers (Glasmasta

956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved.oi:10.1016/j.bios.2004.09.026

80 M. Andersson et al. / Biosensors and Bioelectronics 21 (2005) 79–86

et al., 2002). The benefit of simultaneous registration ofDandf has been demonstrated in several investigations, whereQCM-D measurements were combined with optical surfacesensitive methods such as ellipsometry and surface plasmonresonance (SPR) (Hook et al., 2001).

Our present interest in QCM-D is mainly focused on mea-surements of protein adsorption to biomaterials. Previously,our group has used QCM-D to study complex protein–surfaceinteractions such as immune complement activation andcontact activated blood coagulation (Sellborn et al., 2003;Andersson et al., 2002; Berglin et al., 2004). Surface inducedactivation of the complement system and coagulation systemresult in a massive deposition of proteins on the surface,which is of great importance for biomaterial compatibility(Tang et al., 1998; Kalltorp et al., 1999). Traditional methodsinvestigating complement and blood coagulation systemsare mostly based on measurements of soluble products.The introduction of the QCM-D method presents a way ofstudying new mechanisms of bio-interface research.

Aim of this investigation was to refine the QCM-Dtechnique for quantification of both the complement and thecoagulation system on solid surfaces. Preliminary studiesindicated that the clot density is dependent on the chemistryof the surface. Thus, our second aim was to perform a moredetailed analysis of the D-parameter in relation to fibrindeposition. The QCM-D was used to investigate materialsc (Ti),p A),p ne)( -Dr onalm lublef

1c

te-r thef oodv do andc spec-t .,2e -t ntm sur-f usu-a greeo din(e asb ambem r-f ment

of the soluble thrombin–antithrombin complex (TAT) withEIA as the standard method. However, the QCM-D instru-ment has turned out to be a very interesting alternative fordetermination of surface induced blood coagulation, sincethe fibrin formation can be detected in situ and in real timeon the inductive surface. The frequency shift gives informa-tion about the general clotting kinetics, such as time of onsetand fibrin deposition rate, while the dissipation factor pro-vides valuable data concerning viscoelastic properties, suchas clot density. This way of measuring surface induced bloodcoagulation has been further refined in this study.

The immune complement system is present in blood andserum and consists of about 20 different proteins. On contactwith a foreign surface, e.g. a bacterial surface, the comple-ment system is activated in a cascade fashion, resulting indestruction of the bacterial surface or release of bioactivedegradation products, causing inflammatory reactions in thesurrounding tissue, or both. However, the activated comple-ment cascade can be a major problem concerning medicalimplants and extra corporal blood treatment devices such asoxygenators, dialysators, etc. (Tang et al., 1998; Kalltorp etal., 1999). It also seems to have long term implications, e.g.for in wear debris from total joint arthroplasty (DeHeer etal., 2001) and xenotransplantation (Bengtsson et al., 1998).Determination of complement activity usually involves EIAmeasurements of fragments from degradation of complementf n ac-t mentf ple-m sur-f oundc facea s, us-is ,2

2

2

redo rich,W ich,2 14-7 nce,C istedi W-4 oldsp tely2 singQ ur-p mersw faces

ommonly used in medical implants such as titaniumolyethylene (PE), poly(methyl methacrylate) (PMMoly(ethylene terephtalate) (PET), poly(dimethylsiloxaPDMS) and heparinized surfaces (Hep). The QCMesults were compared to results obtained from traditiethods such as enzyme immuno assays (EIA) of so

actors and platelet counting.

.1. Short description of the coagulation andomplement systems in relation to biomaterials

Activation of the blood coagulation system on biomaial surfaces may result in fibrin clots that may impairunction of various medical devices such as artificial blessels, stents (Christensen et al., 2001), blood catheters anther extra corporal devices. Blood coagulation kineticslot characterisation are studied in various ways such asrophotometry (Sanchez et al., 2002), QCM (Andersson et al002), ellipsometry (Walivaara et al., 1996), SPR (Hanssont al., 2002; Vikinge et al., 2000a, 2000b) and free oscilla

ion rheometry (Hansson et al., 2002). Platelets are importaediators in the coagulation cascade and are involved in

ace induced blood coagulation. The platelet number islly determined by cell-counting methods, and their def activity is measured by the amount of thrombosponTSP) released, using EIA or equivalent methods (Bergsetht al., 2000). In our setup, EIA is performed on blood that heen incubated on coated glass slides using the slide chodel (Hong et al., 1999b, 2001). Measuring potential su

ace induced blood coagulation often involves measure

r

actors in blood or serum that has been in contact with aivating surface. Examples of such products are compleactor 3a and 5a (C3a and C5a) and the terminal com

ent complex (TCC). Complement activation on a solidace can also be detected by the amount of surface bomplement factor 3 (C3) on the activating surface. Surssociated C3 can be measured via anti-C3 antibodie

ng a method such as ellipsometry (Elwing et al., 1986),urface plasmon resonance (SPR) or QCM (Sellborn et al.003).

. Materials and methods

.1. Surfaces

The following surfaces were used: titanium (sputten gold sensor surfaces in vacuum), polyethylene (AldI, USA, 9002-88-4), poly(ethylene terephtalate) (Aldr

9154-49-2) poly(methyl methacrylate) (Aldrich, 9011-), and poly(dimethylsiloxane) (Rhodia silicones, FraAF 2534). Solvents and polymer concentrations are l

n Table 1. The polymers were spin-coated (spin-coater KA Chemat Technology Inc., Northridge, CA, USA) on gensor surfaces (Q-sense, Goteborg, Sweden) using 50�lolymer solution (polyethylene was heated to approxima00◦C before casting). The thickness was estimated uCM-D in air (Table 1). Titanium plates and microscope pose glass slides coated with the above-mentioned polyere used in the slide chamber model (see below). Sur

M. Andersson et al. / Biosensors and Bioelectronics 21 (2005) 79–86 81

Table 1General polymer characteristics; mean values of advancing and receding contact angle measurement as measured by the Wilhelmy plate method (n= 2)

Polyethylene(PE)

Poly(ethyleneterephtalate) (PET)

Poly(methylmethacrylate) (PMMA)

Poly(dimethylsiloxane)(PDMS)

Advancing contact angle (◦) 103 70 75 103Receding contact angle (◦) 50 48 59 91Glass transition temperature,Tg (C◦) −125 81 95 −123Melting transition temperature,Tm (C◦) 138 252 Amorphous −60QCM sensor surface spin-coating premises 1% in decaline 1% in phenol 4% in toluene 1% in toluene

3000 rpm 5000 rpm 2000 rpm 2000 rpm2 min 2 min 2 min 20 s

Estimated polymer thickness 20–30 nm 20 nm 15 nm 84 nm

Glass transition temperature (Tg) and melting transition temperature (Tm) values were provided by the manufacturer. Polymer thickness was estimated byQCM-D measurements in air.

were used only once, except the titanium surfaces which werereused after they were thoroughly washed with sodiumdode-cylsulfate (2%, 30 min), then excessively rinsed with MilliQwater and placed in an UV-ozone oven (10 min) and finallyrinsed again in MilliQ water.

Heparinized surfaces (Corline Systems AB, Uppsala,Sweden) were used as a non-complement activating sur-face and served as negative control for all experiments(Christensen et al., 2001). The heparinized surfaces wereprepared by irreversible adsorption of a macromolecularconjugate that according to the manufacturer contains 70heparin molecules covalently linked to a 50 kDa polymercarrier (polymeric amide) (data on file, Corline SystemsAB, Uppsala, Sweden). The conjugate was then adsorbed byelectrostatic coupling to a polymeric amide coated surface.The surface concentration of heparin was approximately 0,5�g/cm2, which is equivalent to a film thickness of about10 nm. The antithrombin (AT) binding capacity reachesabout 2–4 pmol AT/cm2.

2.2. Dynamic and static contact angle measurements

Advancing and receding contact angle was determinedwith the Wilhelmy plate method (Garbassi et al., 1994) usingde-ionized water (γ = 72.8 mJ/m2) at 20◦C. The instrumentwas a DCA-322 from Cahn operating at 100�m/s. Data wase lls thes 1t an ti-t

2

nor( ital,G andf xedw tedf dert rod-u

30 min in 37◦C. The pellet was discarded and the serumwas collected in Eppendorf vials and immediately frozen(−70◦C) until further use. Thawed plasma and serum wasused immediately.

Blood used in the slide chamber model was drawn in thelab (50 ml) from the same donor as mentioned above, intoa heparinized falcon vial (Corline Systems AB) and usedimmediately.

2.4. The QCM technique and measurements

In brief, the QCM-D technique (Rodahl et al., 1995;Hook et al., 1998; Voinova et al., 1999; Marx, 2003) isbased on a quartz crystal sandwiched between two gold elec-trodes. When voltage is placed over the crystal, a lateraloscillation starts with a fundamental frequency of 5 MHz.The QCM-D also registers the frequencies at the third, fifthand seventh overtone (15, 25 and 35 MHz, respectively),where the third overtone often was the most stable. Whenmolecules are adsorbed, the frequency (f) drops in a linearproportion to the adsorbed mass (normally in the ng/cm2

range) according to the Saurbrey equation (Saurbrey, 1959).By switching the voltage on and off, a passive registra-tion of the viscoelastic damping (D) can be performed. AllQCM-D measurements were obtained during no-flow condi-tions.

2c

Ms odelQ atedf n ofp )a f thec y af agu-l e oft en-s andt (seeF

valuated with WinDCA Version 1.03 (Cahn, WI, USA). Aurfaces were tested with regard to hydrophobicity byessile water drop technique (Dahlgren and Sundqvist, 198)o ensure complete coating coverage and to ensure cleanium surfaces.

.3. Blood products

Citrated blood plasma from an apparently healthy dotested by the blood central, Sahlgren’s University Hospoteborg, Sweden) was distributed to Eppendorf vials

rozen (−70◦C). Serum was prepared from plasma miith CaCl2 (to a final concentration of 25 mM) and incuba

or 30 min in 37◦C. This procedure was carried out in oro remove the fibrin from the plasma. The un-clotted pct was centrifuged (Universal 16R, 5200 rpm, 3000×g) for

.4.1. QCM-D measurements of contact activatedoagulation

Citrated plasma (100�l) was added to a modified QCensor surface placed in an open measurement cell (mWiC 301, Q-Sense, Gothenburg, Sweden) and incub

or 5 min in order to achieve a stable baseline. A solutiolasma (90�l) and CaCl2 (10�l, 0.25 M) was mixed (30 snd added to the stabilized surface. The propagation ooagulation is observed as an initial lag time followed brequency shift, which describes the kinetics of the coation. After the coagulation is completed, the final valuhef andD can be analyzed to attain a description of the dity of the formed clot. Time of onset, slope of the curveotal coagulation time formed the basis of this analysisig. 1).


Fig. 1. Graphs of representative coagulation experiments performed on testsurfaces. As indicated there are three parameters: (A) Onset time, measuredas time from addition of Ca2+ to apparent frequency decrease. (B) Fibrindeposition rate (slope of the curve at inflection). (C) Total frequency shift atplateau. Polymers used in this experiment are described inTable 2. Heparinsurfaces (Hep) and titanium surfaces (Ti) were also used. It was noted thatHep surfaces measured no QCM-D registration of fibrin deposition or visualcoagulation. Each type of surface was tested at least three times without anymajor differences.

2.4.2. QCM-D measurement of surface deposited C3fragments

Serum diluted (1:5) in veronal buffered saline (VBS)(0.15 M NaCl, 1.8 mM Na-5,5-diethylbarbiturate, 3.1 mM5,5-diethylbarbituric acid, 0.5 mM NaN3, 0.5 mM MgCl2with 0.15 mM CaCl2 pH 7.4) was added to the modifiedsensor surfaces which was placed in a temperature con-trolled measurement chamber (22.0± 0.2◦C) (model QAFC301, Q-Sense, Goteborg, Sweden). The serum was incu-bated for 30 min. After rinsing with VBS (5 min), rabbit anti-human C3c (DakoCytomation A/S, Glostrup, Denmark) wasadded (1:50 in VBS, 20 min), followed by another VBS rinse(5 min). The frequency shift produced by the added antibod-ies was measure and used as an estimate of the degree ocomplement activation (see insetFig. 3). All QCM-D mea-surements were performed in triplicates. In previous studies(Sellborn et al., 2003) it has been shown that there are noC3 fragments to be found on the titanium surfaces, althoughthe solution contains high levels of C3a (Hong et al., 1999a).Hence, in this study, titanium was omitted from C3 fragmentsmeasurements. However, an upcoming study focusing on thissubject will be presented in the near future.

2.5. The slide chamber model and its measurements

thod( d-u d co-a MAr cmfi me

of 1.65 ml. For this experiment, the wells were coated withheparin and filled with 1.3 ml blood and 1 IU/ml soluble hep-arin (Løvens Kemiske Fabrik, Ballerup, Denmark). A secondslide, the test surface, was placed on top. The whole construc-tion was then sealed with two clips. O-rings were used be-tween the wells and the test surface to prevent leaking. Whensealed, the construction was placed on a horizontally rotat-ing disc in a water bath (37◦C, 30 min). After incubation, thechambers were opened and the blood was collected into twoheparinized vials containing EDTA (65�l, 0.2 M, pH 7.4). Ablood sample was taken prior to incubation and directly putinto two heparinized vials and used as control. The remainingplatelet count was measured in a cell counter (Coulter Ac·Tdiff analyzer, Coulter Corporation, Miami, FL, USA). Theblood was then centrifuged (Minifuge T, 4000 rpm, 3450×g,5◦C, 25 min) and the cells discarded. The remaining plasmawas further analysed for C3a, TCC, TAT and TSP. These fac-tors were measured with EIA standardized tests. The EIAmicrotiter plates were analysed in a Multiskan MCC/340 mi-crotiter reader at 492 nm (Labsystems, Vantaa, Finland). Twoslides for each surface modification were analysed, resultingin four measurements per surface type. Heparin coated slideswere used as negative control.

2.5.1. Soluble C3aPlasma diluted 1:100 in working buffer, PBS contain-

i en2 7.3w ribedC bod-i ta tain1 tiono en inn

2ured

u d byM rk-i AbM i-v ies( sedf ining4 d.

2red

u outhB nti-h an-t . Asa om-b int int

The slide chamber model is a well-documented meHong et al., 1999b) for studying mainly soluble end procts of the complement system and the surface inducegulation system. The model consists of two wells of PMings with a height of 5 mm and an inner diameter of 1.9xed to a PMMA microscope slide. Each well has a volu

f

ng 1% (w/v) bovine serum albumin and 0.1% (v/v) Twe0, was incubated in wells coated with mAb 4SD1hich served as capture antibody. As previously desc3a was detected using biotinylated anti-C3a anti

es followed by HRP-conjugated streptavidin (Nilsson el., 1988). Zymosan-activated serum (shown to con5,000 ng C3a/ml when compared to a calibrated soluf purified C3a) served as standard. The values are givg/ml.

.5.2. Soluble TCC (sC5b-9)The complement activation product sC5b-9 was meas

sing a modification of the previously described methoollnes et al. (1985). Plasma samples diluted 1:5 in wo

ng buffer were added to wells coated with anti-neoC9 mcaE11, a kind gift from Prof. Tom-Eirik Mollnes, Un

ersity of Tromsø, Norway. Polyclonal anti-C5 antibodDako, Denmark) diluted 1:500 in working buffer were uor detection. Zymosan-activated serum, defined conta0,000 arbitrary units per ml (AU/ml) served as standar

.5.3. Soluble thrombin–antithrombin complexThrombin–antithrombin complexes (TAT) were measu

sing antibodies from Enzyme Research Laboratories (Send, IN, USA). TAT was captured in wells coated with auman thrombin diluted 1:200. HRP-coupled anti-human

ithrombin antibody diluted 1:200 was used for detectionstandard purified thrombin mixed with excess of antithrin in the presence of heparin was used. The highest po

he standard curve was 200 ng/ml TAT.


2.5.4. Thrombospondin (TSP)Thrombospondin was measured using monoclonal an-

tibodies directed against TSP-1 (Immunotech S.A., Mar-seille, France). As previously described byBergseth et al.(2000) clone P12 was used as capturing antibody and bi-otinylated clone P10 as detecting antibody. Streptavidinhorseradish peroxidase conjugate (Amersham Biosciences,Buckinghamshire, UK) was used to achieve staining. Nor-mal serum was used as standard.

2.6. Statistical evaluation

Studentst-test two-tailed and unequal variance was used tocompare the results from the different surface modifications.

3. Results

3.1. Coagulation

A summary of representative experiments on test surfacesare shown inFig. 1. We noted that there was a large spanin the results obtained from the different surfaces. As ex-pected, the titanium surface (positive control) displayed thefastest time of onset (A) as well as a very rapid fibrin depo-sition rate (B). On the other hand, the total frequency shifta aring sur-f ep me-t sa in-dp mt ob-s ectedcd ces,s hichi thesb ets’c Thes hardt

TS et time and fibrin deposition rate), EIA (TAT and TSP) and cell-counting (platelets)

S /min) es)

UHTPPPP

M amples

Fig. 2. Representativef/D plots of the current surfaces. The frequency shift(f) is given on theY-axis, registration starts for all surfaces when Ca2+ isadded (f= 0). The corresponding dissipation values are given on theX-axis.The experimental time span is 2 h. We interpret the result as the PET surfaceinducing the clot with the highest density and PMMA with the lowest density.All plots appear linear except the PET plot, which is slightly curved.

As mentioned in the aim, we also set out to study theDparameter in relation to clot formation. An established wayof studying the viscoelastic effect is to plot the frequencyshifts from coagulation kinetic experiments againstD valuesfrom the same experiments in a diagram. Hence, a relativelylow value of dissipation means higher rigidity. Thef/D ratiocan be used as an estimation of the clot density. A highfcombined with a lowD value would indicate a high densityclot. Graphs off/D plots from the current surfaces are shownin Fig. 2. As expected, the Hep plot is very short and hasno defined curvature. The Ti plot appears to illustrate theformation of a low density clot. However, we believe thatthis is a result of the clot’s low adsorption properties to theTi surface, i.e. the clot is not firmly attached to the surfaceand slips over the surface leading to an underestimation ofmass. This may also explain the relatively low frequencyshift (C parameter), i.e. an underestimation of mass of Ti,as seen inFig. 1. The clots were always firmly attachedto the polymer surfaces. PDMS and PMMA displayeda true low density clot while the PET surface showed ahigh density clot. However, the clot formed on the PETsurface seemed to change during time from higher to lowerdensity.

fter coagulation was comparatively low for Ti. The heprafted surface we used as control is an anti-coagulation

ace and displayed no changes inA, BorC, as expected. Tholymer surfaces displayed differences in all three para

ers (A, B andC). For the polymers theA andC parameterppeared to be related, but theB parameter seemed to beependent ofA andC. Numerical data fromFig. 1 are alsoresented inTable 2together with the results obtained fro

he slide chamber model. In the QCM-D experiments weerved that the two controls Hep and Ti, behaved as exponcerning time to onset. Furthermore, parametersA andBiscriminated fully between the different polymer surfaince the standard deviations were comparatively low, wndicates good reproducibility. However, the results fromlide chamber model (right hand side ofTable 2) differedetween the Hep and Ti for TAT and ‘remaining platelount but not for TSP, where the results were unclear.lide chamber model data for the polymer surfaces wereo interpret due to low reproducibility.

able 2ummary of the coagulation cascade as measured with QCM-D (ons

urface Onset time (A) (min) Fibrin deposition rate (B) (−Hz

ntreated – –eparin No coagulation No coagulationitanium 15± 0.58 144± 33ET 19± 3.5 107± 8.6E 19± 1.7 43± 2.7MMA 24± 2.8 25± 9.2DMS 20± 8.8 38± 20

ean values and standard deviation (±S.D.) are shown. The untreated s

TAT (ng/ml) TSP (ng/ml) Platelets (% of control valu

9± 0 n.d. 100± 1.024± 4.3 39± 70 93± 1.9

51000± 9000 82± 12 0.36± 0.142700± 3000 59± 43 86± 178800± 9300 80± 22 58± 18

17± 4.6 37± 6.4 93± 3.634± 5.9 40± 4.3 88± 2.9

had not been in contact with any surface.


Fig. 3. Results from complement activation on various test surfaces. Empty bars are TCC (AU/ml), grey bars represent amount C3a (ng/ml) and black barsrepresent amount adsorbed anti-C3c (ng/cm2). Note that C3c measurements were not done on titanium. Inset: a graph of a representative C3c measurement(PET surface). Human serum is added after a stable baseline is achieved. After incubation, the serum is rinsed off with VBS buffer and anti C3c is added.Theamount of adsorbed antibodies (ng/cm2) is calculated from the frequency shift (�f) and taken as an arbitrary measure of surface bound complement fragments.

3.2. Complement activation

The principle of determination of complement activa-tion with QCM-D is illustrated in the inset inFig. 3. Whenserum is added to a test surface, a deposited protein layerwill be formed that partly consists of complement factors.After rinsing in buffer, the surface is incubated with anti-complement factor 3 (C3c). The amount of deposited anti-C3c antibodies is a good arbitrary measure of depositedC3.

Adsorption of anti-C3c after serum incubation is shownin Fig. 3 and is compared with results from the fluid phasereactants, C3a and TCC. All surfaces except Hep showed sig-nificantly higher activation than the untreated control for bothC3a and TCC. Titanium, PE, PDMS and PMMA showed sig-nificantly higher activation than PET regarding TCC results.Furthermore, PE showed significantly higher activation thanPMMA (P< 0.05, Studentst-test, two tailed distribution,equal variance). The QCM-D measurement of C3 fragmentis graded as follows: PET≥ PE≥ PDMS≥ PMMA ≥ Hep.PE and PDMS showed significantly higher amounts of C3fragments than Hep (P< 0.05, Studentst-test, two taileddistribution, equal variance). Note that sera incubated onthe PET coated surface resulted in the highest complementactivation amongst the polymers, while the EIA measure-ments ranked PET as the least complement activatings this,t facem Heps 3c)w sultss enta

4. Discussion

The results from the present study clearly indicate that theQCM-D successfully can be used for detailed investigationof both surface induce plasma coagulation and complementactivation. The methodology was found to be easy to use andthe reproducibility was high.

One of our most important finding was that surface asso-ciated fibrin formation could be described with at least fourparameters using the QCM-D method; (1) time to onset ofclot formation, (2) rate of clot assembly, (3) total frequencyshift at stable plateau and (4) clot density. Few other meth-ods, if any, can describe the surface induced clot formationwith so many parameters. The reason probably lies in thefact that the clot analysis takes place directly on the modifiedsensor surface. In contrast to most other traditional methods,measurements are made separately from the surface. It haspreviously been reported that time of coagulation onset cor-relates with the amount of fibrin deposition (Sanchez et al.,2002). Our findings support this correlation, even though Ti,with the shortest onset time, did not seem to have a partic-ularly high fibrin deposition. The theory was also supportedby the results presented inFig. 1. A new parameter was stud-ied as well; the fibrin clot density, assessed inFig. 2, wherethe frequency shift was plotted against the dissipation. If thishas any implications for the integration of a biomaterial (e.g.p tab-l ana -t d toh hereh fromd fib-

urface amongst the polymeric test surfaces. Excepthe internal correlations between the different surodifications are good. Desorption was noted on the

urface. Probably, the VBS buffer (containing anti-Cashed off the serum proteins from the surface. The rehould be interpreted as a low degree of complemctivation.

orous implants) or for wound healing cannot yet be esished. However, an intriguing finding was presented inrticle (Jorneskog et al., 1996) reporting that clots from pa

ients with insulin dependent diabetes mellitus showeave a less porous fibrin network, i.e. a denser clot. Tave also been reports supporting that subjects sufferingifferent variations of dysfibrinogenemia have a denser


rin clot formation compared to healthy subjects. They alsoshowed decreased endothelial cell ingrowth. The decrease ofcell ingrowth was not related to RGD epitopes (Collen et al.,2001). Both groups of patients have impaired wound healing,although of different origin.

Considering the complement activation system, there wasa good correlation between QCM-D and C3a and TCC deter-minations. This was surprising, since complement activationwas determined in serum with the QCM-D method, whereasC3a and TCC were determined in whole blood in the slidechamber. An inconsistent observation, however, was the dis-crepancy in the PET surface measurements. The EIA methodranked PET as having the lowest complement inducing prop-erties of the polymers, while QCM-D ranked PET as the high-est complement inducing surface. Per definition, the molaramount of C3 fragments can never be more than that of C3a.One explanation of the inconsistency could be that both C3aand TCC adsorbs to the protein/platelet fouled PET surface.An alternative explanation could be that native C3 associatesto the PET surface without being cleaved, hence, there wouldbe low amount of C3a, but high response in C3 as measuredwith the QCM-D.

A general observation when comparing the slide chambermodel and the QCM-D was that the correlation was high inthe complement experiments but low in the coagulation ex-periments. The low correlation in blood coagulation is proba-b ions re isa tion;Q rino-g ntf dif-f e ar-c teletl

, in-c f C3.T TCCw ha-n ed int

5s

ro-v rs att aces.I re-s clotd as-s nos-t fur-t orel

Future studies may include a more detailed investigationof the structure of the different fibrin clots, perhaps by usingconfocal microscopy and/or AFM.

Acknowledgements

We wish to thank Entific medical systems AB, Goteborg,Sweden, for financial support.

References

Andersson, M., Sellborn, A., Fant, C., Gretzer, C., Elwing, H., 2002. J.Biomater. Sci. Polym. Ed. 13, 907–917.

Bengtsson, A., Svalander, C.T., Molne, J., Rydberg, L., Breimer, M.E.,1998. Xenotransplantation 5, 176–183.

Berglin, M., Andersson, M., Sellborn, A., Elwing, H., 2004. Biomaterials25, 4581–4590.

Bergseth, G., Lappegard, K.T., Videm, V., Mollnes, T.E., 2000. Thromb.Res. 99, 41–50.

Christensen, K., Larsson, R., Emanuelsson, H., Elgue, G., Larsson, A.,2001. Biomaterials 22, 349–355.

Collen, A., Maas, A., Kooistra, T., Lupu, F., Grimbergen, J., Haas, F.J.,Biesma, D.H., Koolwijk, P., Koopman, J., van Hinsbergh, V.W., 2001.Blood 97, 973–980.

Dahlgren, C., Sundqvist, T., 1981. J. Immunol. Meth. 40, 171–180.DeHeer, D.H., Engels, J.A., DeVries, A.S., Knapp, R.H., Beebe, J.D.,

2001. J. Biomed. Mater. Res. 54, 12–19.E 156,

G rom

G nterf.

H .L.,

H , R.,

H , B.,

H , B.,

H 001.

H 14,

J sson,23.

K ed.

MM . Im-

N . Im-

R Rev.

S 2002.

SS 003.

T 41,

ly due to the biological complexity of the blood coagulatystem, involving both soluble proteins and cells. Thelso another important factor that could offer an explanaCM-D measures the phase transition when soluble fiben forms a solid fibrin clot, which is completely differe

rom measuring a single type of protein in solution. Theerences in fibrin clot formation rate, time of onset and thhitecture of the formed clot did not seem to relate to plaoss, neither did TSP or TAT.

The activation of the complement system howeverludes a central activation mechanism, the cleavage ohe appearance of factors like C3a, C3b, C3c, C5a andill quantitatively depend on the central activation mecism. In general, good correlation was therefore expect

he complement activation experiments.

. Conclusion of major findings, limitations, futuretudies

It may be concluded that the QCM-D methodology pides a simple and trustworthy way to study parametehe interface between blood components and solid surft also broadens the possibilities in the biocompatibilityearch area by introducing a direct way of measuring theensity of surface induced coagulation. It is generallyumed that in vitro studies of new materials have low progic values regarding in vivo trials. However, we hope thather studies using the QCM-D methodology will shed might on the bio-interface research field.

lwing, H., Ivarsson, B., Lundstrom, I., 1986. Eur. J. Biochem.359–365.

arbassi, F., Morra, M., Occhiello, E., 1994. Polymer Surfaces: FPhysics to Technology. Wiley, New York, ISBN: 0471938173.

lasmastar, K., Larsson, C., Hook, F., Kasemo, B., 2002. J. Coll. ISci. 246, 40–47.

ansson, K.M., Tengvall, P., Lundstrom, I., Ranby, M., Lindahl, T2002. Biosens. Bioelectron. 17, 747–759.

ong, J., Andersson, J., Ekdahl, K.N., Elgue, G., Axen, N., LarssonNilsson, B., 1999a. Thromb. Haemost. 82, 58–64.

ong, J., Larsson, A., Ekdahl, K.N., Elgue, G., Larsson, R., Nilsson2001. J. Lab. Clin. Med. 138, 139–145.

ong, J., Nilsson Ekdahl, K., Reynolds, H., Larsson, R., Nilsson1999b. Biomaterials 20, 603–611.

ook, F., Kasemo, B., Nylander, T., Fant, C., Sott, K., Elwing, H., 2Anal. Chem. 73, 5796–5804.

ook, F., Rodahl, M., Brzezinski, P., Kasemo, B., 1998. Langmuir729–734.

orneskog, G., Egberg, N., Fagrell, B., Fatah, K., Hessel, B., JohnH., Brismar, K., Blomback, M., 1996. Diabetologia 39, 1519–15

alltorp, M., Askendal, A., Thomsen, P., Tengvall, P., 1999. J. BiomMater. Res. 47, 251–259.

arx, K.A., 2003. Biomacromolecules 4, 1099–1120.ollnes, T.E., Lea, T., Froland, S.S., Harboe, M., 1985. Scand. J

munol. 22, 197–202.ilsson, B., Svensson, K.E., Inganas, M., Nilsson, U.R., 1988. J

munol. Meth. 107, 281–287.odahl, M., Hook, F., Krozer, A., Brzezinski, P., Kasemo, B., 1995.

Sci. Instruments 66, 3924–3930.anchez, J., Lundquist, P.B., Elgue, G., Larsson, R., Olsson, P.,

Thromb. Res. 105, 407–412.aurbrey, G., 1959. Z Phys. 155, 206–222.ellborn, A., Andersson, M., Fant, C., Gretzer, C., Elwing, H., 2

Coll. Surf. B-Biointerf. 27, 295–301.ang, L., Liu, L., Elwing, H.B., 1998. J. Biomed. Mater. Res.

333–340.


Walivaara, B., Askendal, A., Lundstrom, I., Tengvall, P., 1996. J. Bio-mater. Sci. Polym. Ed. 8, 41–48.

Vikinge, T.P., Hansson, K.M., Benesch, J., Johansen, K., Ranby, M., Lin-dahl, T.L., Liedberg, B., Lundstom, I., Tengvall, P., 2000a. J. Biomed.Opt. 5, 51–55.

Vikinge, T.P., Hansson, K.M., Sandstrom, P., Liedberg, B., Lindahl, T.L.,Lundstrom, I., Tengvall, P., Hook, F., 2000b. Biosens. Bioelectron.15, 605–613.

Voinova, M., Rodahl, M., Jonson, M., Kasemo, B., 1999. Phys. Scripta59, 391–396.

Documents

Quartz Crystal Microbalance-with Dissipation Monitoring (QCM-D) For