Effect of Anticoagulation with Citrate versus Heparin on theAdsorption of Coagulation Factors to Blood Purification Resins withDifferent ChargeCarla Tripisciano,† Andre Leistner,‡ Ingrid Linsberger,† Aniela Leistner,‡ Dieter Falkenhagen,†

and Viktoria Weber*,†

†Center for Biomedical Technology, Danube University Krems, Dr.-Karl-Dorrek-Straße 30, 3500 Krems, Austria‡Polymerics GmbH, Landsberger Allee 378, D-12681 Berlin, Germany

*S Supporting Information

ABSTRACT: In liver failure, hydrophobic toxins accumulatein the blood circulation. To support hepatic function,extracorporeal blood purification systems have been devel-oped, in which both cationic and neutral adsorbents are usedto remove albumin-bound metabolites from blood. An issue ofthese systems is the additional removal of coagulation factorscontaining negatively charged γ-carboxyglutamate (Gla)domains, which, in physiological conditions, are shielded bycalcium ions. We hypothesized that complexation of calciumions by citrate leads to exposure of negative Gla domains,resulting in their binding to the positively charged adsorbents.The data presented here confirm that the binding of coagulation factors containing Gla domains to positively charged polymers isenhanced in the presence of citrate as compared to heparin. This effect increased with increasing charge density of the polymerand has important implications for the clinical application of positively charged polymers.

■ INTRODUCTIONLiver failure is a pathological condition characterized byimpaired synthetic and metabolic hepatic functions, with aresulting accumulation of uncleared hydrophobic metabolites inthe bloodstream.1 Extracorporeal liver support systems havebeen developed to purify blood from such harmful substances.Whereas small water-soluble toxins, such as ammonium, can beremoved via hemodialysis or hemofiltration, the depletion ofpoorly water-soluble, albumin-bound compounds, such asbilirubin, bile acids, phenolic derivates, and aromatic aminoacids, requires combined membrane-adsorption devices,2,3 inwhich toxins are removed using activated carbons or syntheticorganic polymers.Polystyrene divinylbenzene (PS-DVB) copolymers are

employed as adsorbents in liver support devices2,4,5 andefficiently bind molecules containing aromatic rings due tohydrophobic interactions. It has been shown that the pore sizeof neutral polystyrene divinylbenzene adsorbents is crucial toincrease their efficiency in removing strongly albumin-boundsubstances and that the minimal pore size of neutral polymerfor efficient bilirubin removal is about 5.5 nm.6 In addition toneutral polymers, anion exchange resins based on PS-DVB havebeen developed and are clinically used in liver support devicesto optimize the adsorption of bilirubin, which contains carboxylgroups.Coagulation is frequently severely compromised in patients

with disturbed liver function, because coagulation factors are

synthesized in the liver.7,8 Therefore, the blood compatibility ofadsorbents is crucial and their binding of coagulation factors oractivation of the coagulation cascade should be minimized. Inthis context, it is important to mention that a number of factorsof the coagulation cascade, as well as the anticoagulant proteinC, are of similar molecular mass as albumin. Thus, adsorbentswith a pore size optimized for removal of albumin-boundsubstances will also bind these coagulation factors.There have been reports in the literature showing that

fractionated plasma was deprived of factors II, VII, and X,protein C, and protein S upon direct contact with polystyrene-divinylbenzene-based anion exchange resins.9 A commonfeature of all these factors is that they contain γ-carboxyglutamic acid-rich (Gla) domains, which are generatedby vitamin K-dependent post-translational carboxylation of theγ-carbon of glutamic acid residues.10 Under physiologicalconditions, Gla domains are shielded by free calcium ions inblood, with the resulting formation of a structure that mediatesthe interaction of Gla domains with membrane phospholi-pids.11−13 In this structure, Ca2+ plays a dual role, maintainingthe Gla domain conformation and serving as a bridge betweenthe negatively charged Gla domains and the negatively chargedphosphatidylserine residues14 exposed at membrane surfaces

Received: October 29, 2011Revised: December 19, 2011Published: January 9, 2012

Article

pubs.acs.org/Biomac

© 2012 American Chemical Society 484 dx.doi.org/10.1021/bm201529z | Biomacromolecules 2012, 13, 484−488

after cell damage or activation.15 The binding of factors on thephospholipid surface in close proximity renders them moreactive15 and facilitates their interaction, which is necessary forthe efficient proceeding of the coagulation reaction, that is,thrombin generation.16,17

We hypothesized that regional anticoagulation with citratepromotes the adsorption of Gla domain-containing factors toanion exchangers due to the complexation of Ca2+ and theexposure of negatively charged Gla domains.Therefore, the aim of this study was to assess the influence of

anticoagulation (heparin vs citrate) on the adsorption ofvitamin K-dependent coagulation factors.Our results confirm that these factors bind preferentially to

cationic adsorbents in the presence of citrate due to a localdepletion of Ca2+ and exposure of anionic Gla domains. Incontrast, neutral polymers showed a lower binding of thesefactors and minimal variations in binding in citrate versusheparin anticoagulated plasma.

■ MATERIALS AND METHODSPlasma and Chemicals. Unfractionated heparin was purchased

from Baxter (Vienna, Austria). Citrated and heparinised plasma wereobtained by drawing fresh blood from healthy volunteer donors intotubes containing trisodium citrate or heparin to a calculatedconcentration of 11 mM or 5 IU/mL, respectively, in whole blood(Vacuette, Greiner Bio-One, Vienna, Austria). Blood was immediatelycentrifuged at 3600 g for 10 min at room temperature (RT) toseparate plasma from cellular components.Bovine serum albumin, 4-(2-hydroxyethyl)-1-piperazineethanesul-

fonic acid (HEPES), o-phenylenediamine, bilirubin, tryptophan,phenol, and cholic acid were purchased from Sigma Aldrich (Vienna,Austria).Adsorbents. Adsorbents were provided by Polymerics GmbH,

Berlin, Germany. Two divinylbenzene copolymers were used, namely,a neutral divinylbenzene-styrene (DVB-ST) resin and a divinylben-zene-vinylimidazole (DVB-VI) copolymer. They were produced aspreviously described.18 In addition, commercial polystyrene divinyl-benzene-based resins were tested in this study, namely, Prometh 01and 02 (Fresenius Medical Care, Bad Homburg, Germany). Prometh01 (further designated as NP, from Neutral Polymer) is a styrenedivinylbenzene resin whereas Prometh 02 (further designated as CP,from Charged Polymer) is a polystyrene divinylbenzene copolymersubstituted with trimethylamine groups.4

Characterization of Adsorbents. Particle morphology wasanalyzed by scanning electron microscopy (SEM) using a ZeissSupra 50 VP scanning electron microscope (Carl Zeiss NTS) with anacceleration voltage of 5 kV. Dry NP and CP particles were cut with arazor blade, whereas DVB particles were ground with the help of amortar and pestle. Both series of particles were mounted on a sampleholder and coated with palladium (4 nm thickness). A CILAS 1064laser granulometer (Cilas, France) was used to measure the averageparticle diameter in isopropanol. To determine the specific surface andthe pore size distribution, nitrogen adsorption and desorptionisotherms were recorded at liquid nitrogen temperature (−196 °C)at relative pressures p/p0 between 0.001 and 1.0 using an ASAP 2010porosimetry system (Micrometrics Instrument Corp., Norcross,U.S.A.). The specific surface area (SBET) was calculated using theBrunauer, Emett, Teller (BET) equation.19 The total microporevolume (pore size < 2 nm) was calculated with the Horvath−Kawazoe(H−K) method,20 whereas the mesopore volume (2−50 nm) wasobtained via the Barrett, Joyner, Halenda method (BJH).21 Averagepore size d was calculated as d = 4V/SBET (with V = maximumadsorbed nitrogen volume).The charge density of the characterized adsorbents was evaluated

via titration: 200 mg of DVB-VI adsorbent polymer were washed withmethanol and then suspended in water. Subsequently, 2 mL of 0.1 NHCl were added to the polymer suspension in order to protonate all

accessible imidazole groups. This suspension was then titrated with 0.1N NaOH using a DL25 automated titrator (Mettler-Toledo,Switzerland) resulting in 2 inflection points in the titration curve,the first indicating the amount of excess HCl and the second (pKs 5.9)indicating the quantity of protonated imidazole groups. The number ofprotonated imidazole groups (total ion-exchange capacity) per gram ofthe DVB-VI adsorbent was calculated as c × V/m (with c = molarconcentration of the titrant, V = volume of consumed titrant betweenfirst and second inflection point, m = weight of dry adsorbentpolymer) and determined at 0.1 mval/g. Because the combination ofprotonated imidazole and imidazole forms a weak acid/base pair andthus a buffer system, the pH value of the surrounding mediadetermines the number of actual surface charges on the polymer.Therefore, under the conditions of the adsorption experiments at pH7.4, only 10% of the imidazole groups will be protonated, leading to acharge density of 0.01 mval/g.

For CP adsorbent, the value of charge density was provided by themanufacturer.22

The accessibility of the pores of each adsorbent was characterizedby inverse size exclusion chromatography (iSEC) using the retentiontimes of glucose (180 Da) and dextran standards with molecularmasses between 6 and 2000 kDa23 at a flow rate of 0.5 mL/min in 15%(v/v) isopropanol/water. A Waters HPLC System (Milford, U.S.A.)with a Bischoff 8110 refractive index detector was used to determinethe retention volume (VR) of each dextran standard and glucose.Distribution coefficients (Kd) were calculated as Kd = (VR − V0)/(VT− V0), where VR is the retention volume, VT the total mobile phasevolume, and V0 the interparticle void volume. For each adsorbent, theretention time of glucose was used to calculate VT, and the retentiontime of Dextran 2000 to calculate V0. A Kd of zero means that asubstance is completely excluded from the pores, while a Kd of 1means that the pores are completely accessible to a given substance.

Adsorbent Pretreatment. Prior to use, adsorbents were hydro-philized as described.6 Prior to each batch experiment, adsorbents werewashed three times with 0.9% saline solution and stored at 4 °C in0.9% saline solution until further use.

Assessment of Adsorption Characteristics. The adsorption ofcoagulation factors II, VII, and X, as well as of protein C, was assessedin batch experiments using an adsorbent-to-plasma ratio of 1:9 (v/v).At defined time points (0, 30, 60 min) samples were drawn andcentrifuged at 4600g for 4 min. Supernatants were collected, aliquotedand stored at −80 °C until further characterization. Plasma withoutadsorbents served as a negative control. All experiments wereconducted in triplicates.

Quantification of Factors II, VII, X, and Protein C. Factors II,VII, and X were quantified by enzyme-linked immunosorbent assay(ELISA; Coachrom Diagnostica, Vienna, Austria), according to theinstructions of the manufacturer. A calibrator (Biophen, HyphenBioMed, Neuville-sur-Oise, France) specific for each coagulation factorwas used to prepare the standard curve, whereas a human normalplasma control (Biophen, Hyphen BioMed, Neuville-sur-Oise, France)was used to ensure the validity of the test. Protein C was determinedby ELISA (Technoclone, Vienna, Austria).

Quantification of Heparin and Citrate. Heparin was quantifiedwith the Coamatic Heparin test (Coachrom, Vienna, Austria)according to the instructions of the manufacturer. The test is basedon the addition of a chromogenic substrate and excess of Factor Xa(responsible for hydrolysis of this substrate) to the sample. Theinhibition of Factor Xa by the antithrombin−heparin complexcompetes with the hydrolysis of the substrate by Factor Xa. Theconcentration of heparin is inversely proportional to the absorbancedue to the hydrolysis of the substrate.

Citrate was quantified with an automated analyzer (Hitachi 902),using reagent sets from Roche (Penzburg, Germany).

Adsorption of Bilirubin and Cholic Acid. Citrated plasma wasspiked with bilirubin (300 μM), cholic acid (100 μM), tryptophan(100 μM), and phenol (2 mM). Batch experiments were conductedusing an adsorbent-to-plasma ratio of 1:9 (v/v) for bilirubinadsorption, while a ratio of 1:99 (v/v) was used to measure cholicacid removal. After 0, 15, and 60 min of incubation, samples were

Biomacromolecules Article

dx.doi.org/10.1021/bm201529z | Biomacromolecules 2012, 13, 484−488485

drawn and centrifuged (4600g, 4 min), and supernatants werecollected and stored at −20 °C until further analysis. Plasma withoutadsorbents served as a negative control. Experiments were conductedin triplicates. Bilirubin and cholic acid were quantified using a Hitachi902 analyzer (Roche Diagnostic, Penzberg, Germany) and reagent setsfrom Roche for bilirubin and from Trinity Biotech (Wicklow, Ireland)for cholic acid.Statistical Analysis. Data were evaluated with SPSS Statistic

software, version 18.0 (SPSS Inc., Chicago, IL, U.S.A.), using thenonparametric Mann−Whitney U test. Data are expressed as means ±standard deviation. Differences were accepted as significant at p ≤0.05.

■ RESULTS AND DISCUSSION

Physicochemical Characteristics of the Adsorbents.Two series of divinylbenzene-based adsorbents were used inthis study. The first series comprised a neutral and a cationicdivinylbenzene resin with an average particle size of 74 μm(DVB-ST and DVB-VI, Figure 1a,b) and specific surface areasof 641 and 670 m2/g, respectively. These adsorbents containmainly mesopores with an average pore size of 8−9 nm, whichhave been shown previously to be accessible to albumin-boundmolecules.6 The second series of adsorbents were polystyrenedivinylbenzene resins which are already in clinical application,namely Prometh 01, a neutral polystyrene divinylbenzene resin(NP), and Prometh 02, a trimethylamine-substituted styrenedivinylbenzene resin (CP), which functions as a strong anionexchange resin (Figure 1a,c). The physicochemical character-istics for all the tested adsorbents are summarized in Table 1.Scanning electron micrographs (SEM) are presented in

Figure 2 for both series of adsorbents. The outer surface of allthe tested adsorbents appears to be denser and smoother thanthe inner sponge-like structure. Nevertheless, a more compactskin on the surface of NP and CP adsorbents is visible, whichmay impair the pore accessibility of this series of adsorbents.In addition to the assessment of pore size distribution by

nitrogen adsorption (as summarized in Table 1), theaccessibility of pores was evaluated by inverse size exclusionchromatography (iSEC). The distribution coefficients (Kd)values for glucose and dextrans from the iSEC measurementswere plotted against the logarithm of the molecular mass foreach adsorbent (Figure S1). Kd approached zero (0.09) alreadyat molecular mass of 6 kDa for CP and was 0.2 for NP,indicating a poor accessibility of the pores of this series ofadsorbents even to small adsorbates. In contrast, the Kd value ofthe DVB-ST and DVB-VI adsorbents approached zero only atmolecular mass above 100 kDa, suggesting that moleculesunder 100 kDa are able to access the pores of these adsorbents.More than 80% of DVB-ST and 60% of DVB-VI pores wereaccessible to dextran 6 kDa, and nearly 20% of the pores couldbe accessed by albumin molecules.Adsorption of Toxins Related to Liver Failure. The

adsorption efficiency for cholic acid and unconjugated bilirubinfor both series of adsorbents was assessed in vitro under staticconditions. The results are summarized in Table 2. The neutralresin DVB-ST showed a slightly higher adsorption for cholicacid than the weak anion-exchange resin DVB-VI under theexperimental conditions applied, while bilirubin was boundmore efficiently by the DVB-VI resin. The same was true whencomparing the adsorptive efficiency of NP and CP, althoughthe difference in toxin removal between NP and CP was muchbigger, presumably due to the higher charge density of CP.For both metabolites, the amounts removed by DVB-ST and

DVB-VI were significantly higher than for NP and CP (p =

0.05), reflecting the effect of particle size and indicating theimportance of the outer surface area for adsorption. In additionto this difference in particle size, differences in porosity mayalso account for the variation in adsorption characteristics. Thepresence of a dense skin on the surface of the NP and CPpolymers apparently decreases the accessibility to the internal

Figure 1. Structure of the adsorbents used: (a) polystyrenedivinylbenzene adsorbents, DVB-ST and NP; (b) divinylbenzenevinylimidazole adsorbent, DVB-VI; (c) polystyrene divinylbenzenecopolymer substituted with trimethylamine groups, CP. R =N+(CH3)3.

Table 1. Physicochemical Characteristics of the AdsorbentsUsed in This Study

adsorbentcharge density

(mval/g)avg particle

diameter (μm)particle density

(g/mL)SBET

(m2/g)

DVB-ST neutral 75 0.27 640DVB-VI 0.01 73 0.25 670NP neutral 350 0.31 1200CP 3.5 350 0.31 1200

Biomacromolecules Article

dx.doi.org/10.1021/bm201529z | Biomacromolecules 2012, 13, 484−488486

pore structure.22 These findings suggest that the balancebetween porosity and charge density is crucial to improve theperformance with respect to binding of albumin-bound toxins.Adsorption of Coagulation Factors. The removal of the

coagulation factors II, VII, and X and of the anticoagulantprotein C was tested both in citrated and heparinised plasma.Table S1 (see Supporting Information) summarizes the residualpercentages of these molecules after 60 min of incubation in thepresence of citrate and heparin. Table 3 further reduces thesedata to the ratio of residual percentages of the coagulationfactors in citrated versus heparinised plasma.Factors II, VII, X, and protein C, are serine proteases

containing γ-carboxyglutamic (Gla) domains, responsible forhigh-affinity binding of calcium ions, which induce and stabilizethe folding of Gla domains into tightly packed structures withconcomitant exposure of hydrophobic residues that enhancelipid binding.12−14,24 Complexation of calcium by citrateinduces a disordered dynamic state of the Gla domains. Infact, negative Gla residues, not being neutralized by positivecharges of ionized calcium, point away from each other due tocharge repulsions.24 As a result, coagulation factors containingGla domains with exposed negative charges may bind topositively charged groups on ion exchange resins. This effect isless noticed when heparinised plasma is employed, sinceheparin does interact with calcium ions, but not as strongly ascitrate.25,26 A table reporting the molecular masses, number ofGla residues, and isoelectric points for each factor tested in thisstudy is presented in the Supporting Information (TableS2).24,27−32

As shown in Table 3, a higher percentage of Gla domain-containing factors was removed when ion-exchange resins (CPand DVB-VI) were incubated with citrated plasma as comparedto heparinized plasma (p = 0.05). For neutral polymers (NPand DVB-ST), the difference between citrate and heparin wasclearly smaller and not significant (0.2 < p < 1).The efficiency of an adsorbent depends on its pore structure,

that is, the pore size distribution and the specific surface area, aswell as on the hydrophobic/ionic interactions with the targetmolecules. The DVB-ST and DVB-VI polymers used in this

study possess comparable values of specific surface area andmean pore size. As shown by iSEC, their pores are accessible totarget substances in the molecular mass range of albumin,making the particles efficient adsorbents for albumin-boundsubstances and smaller molecules.The presence of vinylimidazole moieties influences the

adsorptive characteristics as illustrated by the enhanced bindingof bilirubin to DVB-VI. However, this goes hand in hand withstronger adsorption of Gla domain-containing factors,especially in citrated plasma.This effect was even more pronounced for CP, which

exhibits a higher charge density than DVB-VI (3.5 vs 0.01mval/g).The binding of Gla domain-containing factors to NP and CP

resins is to be attributed to hydrophobic and ionic interactions,respectively. In both cases, the removal of molecules occursmainly at the outer surface, which is coated by a dense skin,limiting the accessibility of the pores.

Adsorption of Citrate and Heparin. Heparin and citratewere quantified during both sets of experiments (Table S3).Citrate and heparin were not substantially removed by DVB-

VI and there was no adsorption for the neutral beads (DVB-ST). Thus, while neither DVB-ST nor DVB-VI adsorbedsignificant amounts of citrate and heparin, there was significantbinding of both citrate and heparin by the CP (p = 0.05 withrespect to the control values). After 1 h of incubation, citratewas reduced from 23.5 to 14.6 mM and heparin concentrationdropped from 8.3 to 0.7 IU/mL. No such reduction wasobserved for the neutral polymer NP.Overall, an enhanced adsorption of Gla domain-containing

factors to the anion exchanger was observed in the presence ofcitrate. Adsorbents with weak cationic modifications lead to abetter bilirubin removal and a removal of Gla domain-containing factors to a lesser extent than strong ion exchangeresins.

■ CONCLUSIONS

Our data support previous findings regarding decreased levelsof Gla domain-containing coagulation factors when ion-exchange resins are used during extracorporeal bloodpurification. In addition, we demonstrated that the adsorption

Figure 2. Scanning electron micrographs of representatives of the twoseries of adsorbents used: (a) DVB-ST and (b) CP.

Table 2. Amount of Cholic Acid and Bilirubin Bound to the Adsorbents after 15 and 60 min Incubation

cholic acid adsorbed bilirubin adsorbed

(μmol/g adsorbent) (μmol/mL adsorbent) (μmol/g adsorbent) (μmol/mL adsorbent)

adsorbent 15 min 60 min 15 min 60 min 15 min 60 min 15 min 60 min

DVB-ST 19.8 24.2 5.3 6.5 4.4 7.0 1.2 1.9DVB-VI 19.9 22.4 5.0 5.6 6.7 7.7 1.7 2.1NP 3.4 7.9 1.0 2.4 1.2 1.9 0.4 0.6CP 0.8 1.0 0.3 0.3 2.0 4.1 0.6 1.3

Table 3. Residual Percentages of Protein C and Factors II,VII, and X in Citrated Plasma Referred to HeparinizedPlasma (100%) After 60 min of Incubation with DVB-ST,DVB-VI, NP, and CP, n = 3

NP CP DVB-ST DVB-VI

factor II 93 ± 8 58 ± 7 79 ± 5 60 ± 4factor VII 105 ± 9 74 ± 3 75 ± 2 60 ± 4factor X 95 ± 6 59 ± 8 79 ± 4 63 ± 3protein C 94 ± 15 43 ± 13 56 ± 9 30 ± 7

Biomacromolecules Article

dx.doi.org/10.1021/bm201529z | Biomacromolecules 2012, 13, 484−488487

of Gla domain-containing coagulation factors depended on theanticoagulant used. Positively charged resins exhibited strongeradsorption of Gla domain-containing factors in citrated than inheparinised plasma. This effect was more pronounced forstrongly charged polymers, such as CP. Our results on theadsorption of coagulation factors and of albumin-bound toxinsrelated to liver failure demonstrate that the optimization of theadsorbent characteristics (porosity and charge density) iscrucial to achieve minimal adsorption of coagulation factorsand optimal toxin removal at the same time.

■ ASSOCIATED CONTENT*S Supporting InformationIn Tables S1 and S3, the adsorption of coagulation factors andthe reduction in citrate and heparin levels, respectively, aresummarized. Table S2 reports the structure, number of Gladomains, and isoelectric points of the tested coagulation factors.Figure S1 shows data from iSEC analysis. This material isavailable free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*Phone: ++43 2732 893 2632. Fax: ++43 2732 893 4600. E-mail: viktoria.weber@donau-uni.ac.at.

■ ACKNOWLEDGMENTSThe authors thank Ute Fichtinger for precious technicalassistance, Tanja Buchacher for support with ELISA measure-ments, Stephan Harm for support with iSEC, and Tanja Stoiflfor the electron micrographs, which were taken at the School ofPharmacy and Biomolecular Sciences, University of Brighton,U.K. This research was financially supported by the EuropeanCommission within the “MONACO-EXTRA” project on“MONolithic Adsorbent COlumns for EXTRAcorporealmedical devices and bioseparations” − Marie Curie (FP7Grant Number PIAP-GA-2008-218242).

■ REFERENCES(1) Hughes, R. D. Int. J. Artif. Organs 2002, 25, 911−917.(2) Falkenhagen, D.; Strobl, W.; Wogt, G.; Schrefl, A.; Linsberger, I.;Gerner, F. J.; Schoenhofen, M. Artif. Organs 1999, 23, 81−86.(3) Mitzner, S. R.; Stange, J.; Klammt, S.; Peszynski, P.; Schmidt, R.;Noeldge-Schomburg, G. J. Am. Soc. Nephrol. 2001, 12, S75−S82.(4) Rifai, K.; Ernst, T.; Kretschmer, U.; Bahr, M. J.; Schneider, A.;Hafer, C.; Haller, H.; Manns, M. P.; Fliser, D. J. Hepatol. 2003, 39,984−990.(5) Krisper, P.; Stauber, R. E. Nat. Clin. Pract. Nephrol. 2007, 3, 267−76.(6) Weber, V.; Linsberger, I.; Hauner, M.; Leistner, A.; Leistner, A.;Falkenhagen, D. Biomacromolecules 2008, 9, 1322−1328.(7) Amitrano, L.; Guardascione, M. A.; Brancaccio, V.; Balzano, A.Semin. Liver Dis. 2002, 22, 83−96.(8) Kujovich, J. L. Crit. Care Clin. 2005, 21, 563−87.(9) Meijers, B. K. I.; Verhamme, P.; Nevens, F.; Hoylaerts, M. F.;Bammens, B.; Wilmer, A.; Arnout, J.; Vanrenterghem, Y.; Evenepoel,P. Am. J. Transplant. 2007, 7, 2195−2199.(10) Mann, K. G.; Nesheim, M. E.; Church, W. R.; Haley, P.;Krishnaswamy, S. Blood 1990, 76, 1−16.(11) Rezaie, A. R. J. Biol. Chem. 1998, 273, 16824−16827.(12) Mizuno, H.; Fujimoto, Z.; Atoda, H.; Morita, T. Proc. Natl.Acad. Sci. U.S.A. 2001, 98, 7230−7234.(13) Sunnerhagen, M.; Forsen, S.; Hoffren, A. M.; Drakenberg, T.;Teleman, O.; Stenflo, J. Nat. Struct. Mol. Biol. 1995, 2, 504−509.(14) Morrissey, J. H.; Pureza, V.; Davis-Harrison, R. L.; Sligar, S. G.;Ohkubo, Y. Z.; Tajkhorshid, E. Thromb. Res. 2008, 122, S23−S26.

(15) Tavoosi, N.; Davis-Harrison, R. L.; Pogorelov, T. V.; Ohkubo, Y.Z.; Arcario, M. J.; Clay, M. C.; Rienstra, C. M.; Tajkhorshid, E.;Morrissey, J. H. J. Biol. Chem. 2011, 286, 23247−23253.(16) Spronk, H. M. H.; Govers-Riemslag, J. W. P.; ten Cate, H.BioAssay 2003, 25, 1220−1228.(17) Huang, M.; Rigby, A. C.; Morelli, X.; Grant, M. A.; Huang, G.;Furie, B.; Seaton, B.; Furie, B. C. Nat. Struct. Biol. 2003, 10, 751−756.(18) Leistner, A.; Leistner, A. U.S. Patent 7,311,845, 2003.(19) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938,60, 309−19.(20) Horvath, G.; Kawazoe, K. J. Chem. Eng. Jpn. 1983, 16, 470−475.(21) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc.1951, 73, 373−378.(22) Vienken, J.; Christmann, H. Ther. Apher. Dial. 2006, 10, 125−131.(23) DePhillips, P.; Lenhoff, A. M. J. Chromatogr., A 2000, 883, 39−54.(24) Whinna, H. C.; Lesesky, E. B.; Monroe, D. M.; High, K. A.;Larson, P. J.; Church, F. C. J. Thromb. Haemost. 2004, 2, 1127−1134.(25) Higgins, C. Med. Lab. Obs. 2007, 39, 16−20.(26) Sachs, C.; Rabouine, P.; Chaneac, M.; Kindermans, C.; Dechaux,M.; Falch-Christiansen, T. Ann. Clin. Biochem. 1991, 28, 167−73.(27) Fujikawa, K.; Legaz, M. E.; Davie, E. W. Biochemistry 1972, 11,4882−4891.(28) Kisiel, W.; Davie, E. W. Methods Enzymol. 1981, 80, 320−332.(29) DiScipio, R. G.; Davie, E. W. Biochemistry 1979, 18, 899−904.(30) Bajaj, S. P.; Rapaport, S. I.; Brown, S. F. J. Biol. Chem. 1891, 256,253−259.(31) Kisiel, W.; Hanahan, D. J. Biochim. Biophys. Acta 1973, 304,103−113.(32) DiScipio, R. G.; Hermodson, M. A.; Yates, S. G.; Davie, E. W.Biochemistry 1977, 16, 698−706.

Biomacromolecules Article

dx.doi.org/10.1021/bm201529z | Biomacromolecules 2012, 13, 484−488488