Acta Biomaterialia 6 (2010) 1099–1106

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Immobilization of heparin on polysulfone surface for selective adsorptionof low-density lipoprotein (LDL)

Xiao-Jun Huang a,b, Deepak Guduru a, Zhi-Kang Xu b, Jörg Vienken c, Thomas Groth a,*

a Biomedical Materials Group, Department of Pharmaceutics and Biopharmaceutics, Institute of Pharmacy, Martin Luther University Halle-Wittenberg, Halle (Saale) 06120, Germanyb Key Laboratory of Macromolecular Synthesis and Functionalization (Ministry of Education), Department of Polymer Science and Engineering, Zhejiang University,Hangzhou 310027, Chinac Fresenius Medical Care Deutschland GmbH, BioSciences Department, Bad Homburg 61352, Germany

a r t i c l e i n f o

Article history:Received 30 June 2009Received in revised form 11 August 2009Accepted 31 August 2009Available online 3 September 2009

Keywords:PolysulfoneSurface modificationHeparinLDL adsorptionApheresis

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.08.039

* Corresponding author. Tel.: +49 3455528461; faxE-mail address: thomas.groth@pharmazie.uni-hall

a b s t r a c t

A versatile method was developed to immobilize heparin covalently on polysulfone sheets (PSu) toachieve selective adsorption of low-density lipoprotein (LDL). This was achieved by activation of PSu withsuccessive treatments of chlorodimethyl ether and ethylenediamine, and subsequent chemical binding ofheparin with bifunctional linker molecules. A heparin density up to 0.86 lg cm�2 on a dense PSu film wasachieved. The modified PSu films were characterized by attenuated total reflectance Fourier transforminfrared spectroscopy and X-ray photoelectron spectroscopy. The hydrophilicity of the PSu film wasimproved greatly by covalent immobilization of heparin. The water contact angle of PSu film wasdecreased from 86.6 ± 3.7� to 50.5 ± 3.2� after binding of 0.36 lg cm�2 heparin. An enzyme-linked immu-nosorbent assay was used to measure the binding of LDL on plain and modified PSu films. It was foundthat the heparin-modified PSu film could selectively recognize LDL from binary protein solutions. Fur-thermore, it was possible to desorb LDL from heparinized PSu, but not from plain PSu, with heparin,sodium chloride or urea solution, which indicates a selective but reversible binding of LDL to heparin.The results suggest that heparin-modified PSu membranes are promising for application in simultaneoushemodialysis and LDL apheresis therapy.

� 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

The abnormal elevated level of cholesterol in human blood hasbeen confirmed to be a main risk factor for the process of coronaryheart diseases, atherosclerosis and cerebral thrombosis. Choles-terol is carried in plasma by a series of lipoproteins, such as verylow-density lipoproteins (VLDL), low-density lipoproteins (LDL),intermediate density lipoproteins (IDL) and high-density lipopro-teins (HDL). It is now believed that LDL is one of the major causa-tive elements in the development of atherosclerosis [1,2]. Highlevels of LDL, especially coupled with low HDL concentrations,can promote cholesterol accumulation in the intra- and extracellu-lar arterial wall that in turn leads to plaque formation causing var-ious heart and vascular diseases [2–4]. Therefore, the reduction ofLDL level in blood is used therapeutically to lower the risk of car-diovascular/cerebrovascular diseases [5–8].

In recent years, LDL has been efficiently eliminated by extracor-poreal LDL apheresis systems [6–8]. Various general LDL apheresisprocedures exist for routine clinical LDL elimination, such as unse-

ia Inc. Published by Elsevier Ltd. A

: +49 345 5527 379.e.de (T. Groth).

lective plasma exchange [9], semiselective double-filtration plas-mapheresis (DFPP) [10], and there are other more selectivetechniques based on immunogenic or electrostatic interactionssuch as immunoadsorption using anti-LDL antibodies (LDL-Thera-sorb, Miltenyi Biotec, Bergisch Gladbach, Germany) [11], heparin-induced extracorporeal LDL precipitation (HELP, B Braun, Melsun-gen, Germany) [12,13], adsorption on dextran sulfate–cellulose(Liposorber, Kaneka Corporation, Osaka, Japan) [14], and directadsorption of lipids from whole blood using polyacrylamide beadscoated with polyacrylic acid (DALI-system, Fresenius, Bad Hom-burg, Germany) [15]. Normally, these apheresis systems, exceptfor DALI, require the separation of blood cells and plasma by plas-ma filtration into a secondary circuit as a first step. For the HELPsystem, the process is even more complex and costly and a filtra-tion is needed to remove the precipitated LDL complexes. Further-more, a simultaneous LDL apheresis and hemodialysis procedure isrequired especially in treatment of patients with chronic renal fail-ure and LDL-induced coronary heart disease [16]. In these cases,the membrane filtration must be applied during LDL apheresis pro-cedures. Overall, procedures to carry out LDL apheresis are compli-cated by the additional burden for the patient. Simultaneoustreatments of ESRD patients to continuously remove LDL duringhemodialysis sessions without resorting to a combination of

ll rights reserved.

1100 X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106

different methods would not only reduce costs, but probably alsoprovide more safety and comfort for the patients.

For fabrication of LDL adsorbents, selection of an appropriateligand and matrix is presently a major area of concern. Heparin,a highly sulfated polysaccharide, is known as one of the most effi-cient LDL ligands, which can interact with apolipoprotein B of LDLvia electrostatic interactions [17,18]. For this reason different hep-arin-based adsorbents had been suggested, such as heparin-cou-pled Sepharose [8] and heparin-modified poly(vinyl alcohol)(PVA) granules [19]. Furthermore, heparin was used as ligand toprepare sensors for the measurement of LDL in blood [20,21].

Based on the complexity of the contemporary LDL apheresisprocedures and the obvious advantage of a simultaneous LDLapheresis–hemodialysis treatment, we suppose that direct immobi-lization of a highly selective LDL ligand such as heparin on hemodi-alysis membranes may be useful for selectively adsorbing LDL inESRD patients with hypercholesterolemia. As hydrophilized poly-sulfone (PSu) is currently the most widely used polymer for hemod-ialysis membranes, and enables efficient removal of small tomedium-sized molecules [22,23], PSu was used here as matrix forheparin immobilization. The objective of this study was to exploredifferent chemical routes to immobilize heparin on PSu films cova-lently with a high yield. Herein, a method to tether heparin onto aPSu surface is described, and the surface properties were analyzedby Fourier transform infrared spectroscopy (FTIR) and X-ray photo-electron spectroscopy (XPS) and water contact angle measurements.The ability of modified PSu to bind LDL selectively was studied by en-zyme-linked immunosorbent assays (ELISAs). It was found thatmodification of PSu with heparin results in a material surface thatcan bind selectively and reversibly high quantities of LDL.

2. Materials and methods

2.1. Materials

Pure PSu granules were supplied by Fresenius Medical Care.Tetrahydrofuran (THF), tin(IV) chloride (SnCl4), chlorodimethylether, ethylenediamine (EDA), 1-ethyl-3-(dimethyl-aminopro-pyl)carbodiimide hydrochloride (EDC), N-hydroxysuccinimide(NHS), toluidine blue (TB), bovine serum albumin (BSA), humanserum albumin (HSA, �99%), primary antibody: anti-b-lipoprotein(LDL, antibody produced in goat whole antiserum), secondary anti-body: anti-goat IgG (peroxidase antibody produced in rabbit),blocking buffer and Tween 20 were purchased from Sigma–Aldrichand used as received. LDL (5.8 mg ml�1, Millipore), 3,30,5,50-tetra-methylbenzidine (TMB, Care Roth GmbH), heparin (sodium salt,166.0 U mg�1, VWR International GmbH) and flat ELISA plates(Greiner Bio-One, Germany) were purchased and used as received.Water used in all experiments was deionized and ultrafiltered to18 MX with a TKA MicroPure Water system.

2.2. Fabrication of PSu dense film

Pure PSu was dissolved in THF at about 25 �C for 24 h with vig-orous stirring to form a 10 wt.% homogeneous solution. After airbubbles were removed completely, the solution was cast onto aclean glass plate using a casting knife with a 100 lm gate opening.The glass plate with the nascent film was directly dried for 24 h at60 �C under vacuum. The dense film was peeled off and then driedfor another 24 h at 80 �C under vacuum to ensure dryness beforesurface modification was conducted.

2.3. Immobilization of heparin on PSu surface

PSu was chemically activated according to the method of Higu-chi et al. [24]. The PSu film was immersed briefly in a solution of

chlorodimethyl ether, hexane and SnCl4 at 25–30 �C for variousreaction times (10 min–24 h) and then washed in methanol for2 h. Thereafter, the chloromethylated PSu (PSu-CH2Cl) was im-mersed into EDA at 25 �C to obtain amino groups for the subse-quent chemical binding of heparin. Finally, EDA-modified PSu(PSu-NH2) was submerged into heparin and EDC/NHS solution(5 mg ml�1 of heparin and EDC in citrate buffer solution: 0.2 MNa2HPO4 and 0.1 M citric acid, adjusted to pH 4.7 with 1 M NaOH,molar ratio of EDC to NHS = 1:1) for 24 h at 25–30 �C to bind hep-arin covalently. The modified PSu was taken out and washed threetimes with phosphate-buffered saline (PBS, 0.03 M Na2HPO4,0.02 M KH2PO4, and 0.137 M NaCl, adjusted to pH 7.0 with 1 MNaOH).

2.4. Surface characterization

2.4.1. Heparin density on the modified PSu surfaceThe quantity of heparin bound to PSu surface was assayed by

the toluidine blue (TB) colorimetric method according to the liter-ature [25,26]. The assay is based on the fact that TB will irrevers-ibly bind to a polyanion substrate (e.g. heparin-modified PSu).The amount of immobilized heparin can be calculated by compar-ison with a standard curve using soluble heparin of knownconcentration.

For a calibration curve, TB was dissolved in 0.01 mol l�1 hydro-chloric acid containing 0.2 wt.% NaCl to prepare 0.005% TB solution.A series of heparin solutions with concentrations varying from 0 to25 lg ml�1 were prepared by dissolving heparin in an aqueous0.2 wt.% NaCl solution. Heparin standard solution (0.5 ml) wasadded to TB solution (0.5 ml) and then agitated for 30 s. Next,n-hexane (1 ml) was added, and the mixture was shaken well sothat the heparin–TB complex was extracted into the organic layer.The non-extracted TB remaining in the aqueous phase was deter-mined by measuring the absorption at 631 nm. A linear correlationbetween the concentration of heparin in the aqueous solution andthe absorbance at 631 nm caused by the residual TB was obtainedand used as a calibration curve to determine the quantity of immo-bilized heparin.

Accordingly, the heparin-modified PSu sheet film was cut intoround size with a diameter of 1.35 cm and immersed in TB solutionand incubated for 30 min. Subsequently, n-hexane was added andthe mixture shaken well to ensure uniformity in treatment. Afterremoving the PSu from the solution, the absorbance of aqueouslayer at 631 nm was measured by UV–visible spectrophotometry.The quantity of immobilized heparin was calculated from theabove constructed calibration curve. Each value was an averageof five independent measurements.

2.4.2. FTIR and XPSAttenuated total reflectance (ATR)-FTIR measurements were

carried out on a Vector 22 FTIR (Brucker Optics, Switzerland)equipped with an ATR cell (KRS-5 crystal, 45�). Sixteen scans weretaken for each spectrum at a normal resolution of 2 cm�1. XPSspectra were recorded on a PHI-5000C ESCA system (Perkin-Elmer,USA) with Al Ka excitation radiation. The pressure in the analysischamber was maintained at 10�6 Pa or lower during measure-ments. To compensate for the surface charging effect, all surveyand core-level spectra were referenced to the C1s hydrocarbonpeak at 284.6 eV.

2.4.3. Water contact angle measurementThe hydrophilicity of the PSu film was characterized on the ba-

sis of water contact angle. Static contact angle was measured atroom temperature on a contact angle goniometer (OCA20, Data-physics, Germany) equipped with video capture. In a typical sessiledrop method, a total of 2 ll of deionized water was dropped onto a

X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106 1101

dry sample with a microsyringe in an atmosphere of air. The con-tact angle image was recorded and determined from the imagewith imaging software. At least five contact angles were averagedto obtain a value.

2.5. LDL adsorption on the PSu films surface measured by ELISA

LDL adsorption on the surface of PSu film was investigated byELISA, as described previously [27,28]. Protein and antibody solu-tions were freshly prepared before the measurement. For the pri-mary antibody, anti-b-lipoprotein was added to 0.1 wt.% BSAsolution at a dilution of 1:10,000. For the peroxidase-conjugatedsecondary antibody, anti-goat IgG was diluted 1:20,000 with0.1 wt.% BSA solution. For the substrate solution (freshly prepared),400 ll of TMB (0.5%, dissolved in DMSO) and 2 ll of hydrogen per-oxide (30%) were added to 10 ml substrate buffer solution (0.2 MNa2HPO4 and 0.1 M citric acid, adjusted to pH 5.0 with 1 M NaOH).

LDL or human serum albumin (HSA) was dissolved in PBS(137 mM NaCl, 1.4 mM KH2PO4, 4.3 mM Na2HPO4, 2.7 mM KCl,pH 7.4). PSu sheet films were cut into discs with a diameter of0.5 cm and placed in 96-well tissue culture plates, then 100 ll ofprotein solution, and PBS (pH 7.4) without protein (blank), wereadded. The plates were incubated for 1 h at 37 �C and then washedthree times with 200 ll PBS (pH 7.4). Blocking solution (100 ll,Sigma–Aldrich) was added and incubated for 0.5 h at 37 �C. Afterrinsing with PBS (pH 7.4), the primary antibody and peroxidase-conjugated secondary antibody were respectively added andincubated for 1 h at 37 �C. Each subsequent step was followed bywashing with Tris-buffered saline (TBS, 137 mM NaCl, 1.4 mMKH2PO4, 4.3 mM Na2HPO4, 2.7 mM KCl, pH 7.4, 0.1% Tween 20).Subsequently, the films were transferred to another 96-well plate,followed by addition of TMB substrate solution at room tempera-

Fig. 1. Scheme of the chemic

ture for 10 min. The reaction was stopped by adding 100 ll of1 M H2SO4. Part of the dye solution (100 ll) was transferred to anew 96-well plate. The optical density (OD) was measured at450 nm vs. 620 nm with a plate reader (Anthos reader, Graz,Austria).

2.6. Desorption of LDL from PSu films

After incubation in protein solution, PSu films were respectivelyimmersed in sodium chloride (0.15–2.0 M, PBS, pH 7.4), heparin(10–1000 lg ml�1, PBS, pH 7.4) or urea (0–8.0 M) solution at roomtemperature for three times, each time using 100 ll solution for10 min. After washing with PBS (pH 7.4), the films were blockedwith blocking solution and the amount of LDL remained on thefilms was determined using the ELISA method described above.

3. Results and discussion

3.1. Chemical activation of and heparin immobilization on PSu

Surface modification is an attractive and efficient approach toendow polymeric membranes with tailored surface properties ina defined, selective way while preserving their original bulk struc-ture. In recent years, several kinds of methods have been devel-oped to activate PSu and immobilize different functional groupsor ligands on PSu membrane surface [24,29,30]. For example, Hig-uchi et al. described a simple Friedel–Crafts reaction to chemicallyactivate PSu for the further covalent attachment of affinity ligandson the surface of the hydrophobic PSu membrane [24]. It wasfound that the degree of substitution on PSu surface could be easilycontrolled over a wide range. In this work, as schematicallydescribed in Fig. 1, heparin was tethered on the PSu surface by

al activation of PSu film.

Fig. 2. Effect of chloromethylation time on the heparin density on PSu dense film. Ahigher heparin density on PSu dense film can be achieved by prolongation of theprimary activation with chlorodimethyl ether.

1102 X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106

covalent chemical bonds, which include the chemical activation ofPSu with chlorodimethyl ether and EDA, and subsequently chemi-cal binding of heparin in the presence of EDC/NHS. For this pur-pose, PSu was first activated under the optimum conditionsaccording to the method of Higuchi et al. [24]. It was found thatthe primary activation of PSu by chlorodimethyl ether was themain factor affecting the degree of immobilization of heparin.Fig. 2 shows that the heparin density on dense PSu films initiallyincreased markedly with the chloromethylation time and then al-most reached a plateau when the chloromethylation time ex-ceeded 1 h. This is reasonable given that the reactions on PSufilm surface in this work are heterogeneous and the steric hin-drance of reactants is significant during solid-surface reaction. Fur-thermore, it is well known that, given the steric constraints thatexist, carrying out chemical reactions directly between polymersurface and macromolecules is relatively difficult to achieve withhigh efficiency [31]. The steric hindrance of heparin with a molec-

Fig. 3. ATR-FTIR spectra of the pure PSu, PSu

ular weight ranging from 3 to 50 kDa limits its tethering on thefilm surface [32,33]. Similar heparin densities on various surfaceswere reported in the literature. For example, Chen et al. [25] cova-lently immobilized heparin on silicone surface with a density of0.68 lg cm�2 and Kang et al. [26] reported a heparin density ofapproximately 1.1–1.3 lg cm�2 on polyurethane surfaces.

3.2. Characterization and hydrophilicity of the heparin-modified PSusurface

The ATR-FTIR spectra of plain and modified PSu films are shownin Fig. 3. The spectra of all films are similar at wavelengths below1600 cm�1. The 1585 and 1486 cm�1 bands are due to the aromaticgroups, and the 1325/1290 cm�1 doublet and 1150 cm�1 band areassigned to the aromatic sulfone chromophore. Fig. 3 shows that,compared to the plain surfaces, the PSu-heparin surface has anew peak appearing at 1725 cm�1, almost certainly arising fromthe carbonyl group (C@O) of immobilized heparin. Furthermore,PSu-heparin showed a shoulder at around 1050 cm�1, which wasdue to the symmetric stretching of the —SO3

� group of immobi-lized heparin [25,26].

To further verify the chemical structures on the modified PSu,XPS survey scans were performed to assess the elemental compo-sition of the film surface. Fig. 4 demonstrates that, compared toplain PSu, the chloromethylated PSu (PSu-CH2Cl) showed an addi-tional peak corresponding to Cl-2p3 (binding energy 202 eV). Incontrast to this, the peak corresponding to Cl-2p3 disappeared onthe surface of PSu modified with EDA (PSu-NH2) and heparin-mod-ified PSu (PSu-heparin). Furthermore, a new nitrogen componentwas observed on the surface of PSu modified with EDA (PSu-NH2) and heparin-modified PSu (PSu-heparin) based on the termi-nal amine group of tethered ethylenediamine and the amide groupof immobilized heparin [25,26].

The chemical compositions of the different surface-modifiedPSu, calculated from the XPS survey scans, are shown in Table 1.Compared to PSu-CH2Cl, the chlorine content decreased from3.98% to 0% on PSu-NH2 and PSu-heparin, which is consistent withthe XPS spectra shown in Fig. 4. After immobilization of heparin on

-CH2Cl, PSu-NH2 and PSu-heparin films.

Fig. 4. ESCA survey scan spectra of the pure PSu, PSu-CH2Cl, PSu-NH2 and PSu-heparin films.

Table 1Chemical composition of surface-modified PSu film calculated from ESCA scanspectra.

Substrate at.%

C O S Cl N

PSu 78.56 19.19 2.25 – –PSu-CH2Cl 77.29 16.79 1.94 3.98 –PSu-NH2 77.69 12.23 2.25 – 7.83PSu-heparin 67.74 24.39 2.20 – 5.67

Fig. 5. Water contact angle of the modified PSu films with different heparindensities. The higher heparin density on the modified PSu film leads to ahydrophilized surface with a lower water contact angle.

X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106 1103

PSu-NH2 surface, the oxygen content on PSu-heparin surface in-creased from 12.23% to 24.39%, while the carbon, sulfate and nitro-gen content was decreased from 77.69%, 2.25% and 7.83% to67.74%, 2.20% and 5.67%, respectively. This is ascribed to the factthat immobilized heparin possesses relatively high oxygen content[25,26].

The wettability of the plain and heparin-modified PSu was esti-mated by water contact angle measurements. Static contact angleson the plain and heparin-modified PSu films are shown in Fig. 5.The water contact angles were obviously decreased with increasedheparin density on the modified PSu films and decreased from86.6 ± 3.7� to 50.5 ± 3.2� after binding of 0.36 lg cm�2 heparin.Overall, the successful modification of PSu with heparin was dem-onstrated by using TB blue assay, the appearance of carbonyl andsulfate groups in ATR-FTIR spectra, new sulfur and nitrogen peaksin XPS scans and the observed changes in water contact angle aftermodification.

Fig. 6. Adsorption of LDL from single protein solution with different concentrations(0.5, 1.0, 5.0, 10 and 20 lg ml�1) on the pure and heparin-modified PSu films. Theadsorption of LDL on both the pure and heparin-modified PSu films surfacefollowed the Langmuir isotherm model. A higher quantity of LDL is adsorbed on theheparin-modified PSu film surface compared to plain PSu.

3.3. LDL adsorption on PSu film measured by ELISA

ELISAs have been demonstrated to be reliable, sensitive andreproducible methods and have been applied to measure theamounts of specific proteins (e.g. albumin, fibrinogen and fibronec-tin) adsorbed on biomaterial surfaces from complex protein mix-tures or plasma [27,28]. Relative values of protein adsorbing tobiomaterials have been reported as optical densities from colori-metric ELISAs. The standard curves of LDL adsorption on hydropho-bic plates measured by ELISA are available in SupportingInformation. There was a linear correlation between the coatingconcentration of LDL on the plates and the absorbance measured

by ELISA. This indicates that this ELISA can also be applied to detectLDL adsorbed on PSu surface.

The adsorption of LDL from single protein solutions on the plainand modified PSu is shown in Fig. 6. It was observed that theamount of absorbed LDL increased with concentration as was ex-pected. The increase shows the characteristics of Langmuir typeof adsorption with a plateau value as was observed for other pro-teins similarly adsorbing on biomaterials [34–36]. Fig. 6 shows alsothat LDL adsorbs on the plain and heparin-modified PSu films. Inthis context, many studies have shown that adsorption of proteinsis increased on surfaces with lower wettability due to hydrophobicinteractions between surface and protein [29,34–37]. It should benoticed that after heparin immobilization, when PSu film becamequite hydrophilic, significantly higher quantities of LDL were ad-sorbed compared to plain PSu. This indicates dominance of theelectrostatic interaction between heparin and LDL, which has beendescribed by Cardin et al. [17] and Gigli et al. [18], is also exploitedduring HELP apheresis [12,13] and allows the adsorption of rela-tively high quantities of LDL.

1104 X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106

Adsorption of proteins to biomaterials from blood or plasmausually involves competition of all the proteins in the mixturefor the available surface. However, because it is very difficult tocontrol all the important variables using complex mixtures of pro-teins, competitive adsorption studies from mixtures of two pro-teins are usually used as an alternative to human blood plasmafor evaluation of protein adsorption to biomaterials [35–38]. Basedon the high concentration of human serum albumin (HSA) in plas-ma, mixtures of LDL and HSA were prepared and plain and hepa-rin-modified PSu films exposed to these mixtures. It is assumedthat this approach simulates the competitive effect of proteinadsorption in (complex) mixtures of proteins. According to theVroman effect [38], for the competitive adsorption from mixturesof proteins, the composition of the adsorbed protein layer isstrongly dependent on the interaction time. The high-concentra-tion proteins initially dominate the surface due to the higher colli-sion rates, but can be replaced by other proteins with highersurface affinity as time passes [34–37]. Hence, in a second experi-ment, the adsorption of LDL was measured from binary proteinsolutions composed of HSA and LDL.

Fig. 7A and B shows the result of this investigation. When theconcentration of LDL was low (10 lg ml�1) it was observed thatthe presence of HSA in solution reduced the LDL adsorption onthe plain PSu to zero provided the concentration of HSA was higherthan 10 mg ml�1. In contrast to this finding, even at higher HSAconcentrations of 25 mg ml�1 adsorption of LDL still occurred onheparin-modified PSu. This represents a ratio of HSA to LDL of1000:1, which is obviously much higher than that of the ratio be-tween HSA and LDL in human plasma for healthy persons (the levelof HSA and LDL was 30–50 g l�1 and 100–120 mg dl�1 [2–5],respectively).

To test further how an increased ratio of LDL to HSA wouldmodulate the binding of LDL, the concentration of HSA was fixedto 1 mg ml�1 and LDL concentration was increased as shown inFig. 7B. Again, the binding of LDL to heparin-modified PSu wasmuch higher than to plain PSu. A greatly increased binding ofLDL was observed with increased concentration, reaching a ratioof 1:100. As mentioned previously, in human plasma the ratio ofLDL to HSA is normally 1:30–50 for healthy persons, but this ratiois much higher in hypercholesterolemic patients (LDL P 160mg dl�2) [2–6]. The ability of heparin-modified PSu to adsorb LDLfrom binary protein solutions means that this material could be a

Fig. 7. Adsorption of LDL on the pure and heparin-modified PSu film from binary protein1.0, 5.0, 10, 25 mg ml�1; (B) HSA = 1.0 mg ml�1, LDL = 0.5, 1.0, 5.0, 10 lg ml�1. The hepsolutions.

potent material for adsorption of LDL under pathological condi-tions. The significant increase in LDL adsorption after heparin mod-ification of PSu can be attributed to the immobilized heparin,which is a polyanion polysaccharide with high density of negativegroups, such as –OSO3– and –COO–. LDL, in which the structuraldomain B100 has a positively charged group of amino acids (lysineand arginine), could specifically bind to heparin by electrostaticinteraction [17–19]. Furthermore, the chain conformation andthree-dimensional structure on the substrates’ surface have a sig-nificant effect on the protein adsorption [36,39]. For heparin-mod-ified PSu film, the immobilized heparin with a molecular weightranging from 3 to 50 kDa [32] can present a three-dimensional sur-face with more binding sites for LDL adsorption than that of theplain pure PSu with a surface that is flat in molecular terms.

3.4. Regeneration of the PSu film containing recognized LDL

The specific interaction between LDL and heparin is mainlydriven by electrostatic forces [17–21]. This type of interaction issignificantly dependent on ionic strength [21,40], while the hydro-phobic interaction between proteins and surfaces remains undis-turbed [35–38]. Apheresis procedures are usually carried outwith regeneration of the adsorber surfaces to allow multiple orrepetitive sessions with the same adsorber while reducing the costof treatment. This requires repeated washing of adsorber anddesorption of bound proteins or other agents. To learn more aboutthe dominating interaction force between LDL and either PSu sur-faces, and to see if heparinized PSu can be regenerated after expo-sure to LDL, desorption experiments were carried out. Washing ofPSu films after exposure to LDL was performed with heparin, NaCland urea solutions. Heparin was used as competitive reagent;strong NaCl solutions disturb the electrostatic interaction force;and urea serves to disrupt the non-covalent bonds in proteinsand is widely applied as a protein adsorber or column regenerationbuffer for chromatography [41].

Fig. 8A shows the result of desorption experiments of LDL fromPSu films exposed to binary protein solutions (LDL = 10 lg ml�1

and HSA = 1.0 mg ml�1) after addition of increasing quantities ofheparin in PBS (pH 7.4). It was found that LDL bound to the plainPSu surface is unaffected by heparin concentration of up to1000 lg ml�1, but LDL could be desorbed from heparin-modifiedPSu surface by heparin solutions with a broad concentration, with

solutions of LDL and human serum albumin (HSA): (A) LDL = 10 lg ml�1, HSA = 0.1,arin-modified PSu film shows a higher ability of binging LDL from binary protein

Fig. 8. Desorption of LDL from PSu films exposed to binary protein solutions (LDL = 10 lg ml�1 and HSA = 1.0 mg ml�1) after washing with heparin (A), NaCl (B) and urea (C)solutions. LDL can be eluted from heparin-modified PSu, but not from plain surface, with heparin, NaCl or urea solution.

X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106 1105

some elution beginning at 10 lg ml�1. As heparin has a high affin-ity for LDL [17–21], it can displace the adsorbed LDL on the hepa-rin-modified PSu surface by competitive interaction and result inLDL elution from the heparin-modified PSu in all cases of variousconcentrations of heparin.

However, the high cost of heparin and the large amount ofwashing solution required results in expensive regeneration and/or extensive purification steps. From this point of view, salt andurea are good alternatives to heparin for an effective and economicregeneration. As shown in Fig. 8B, the adsorbed LDL was elutedwhen NaCl (PBS, pH 7.4) was used for elution as well as heparinsolution. The amount of the desorbed LDL was increased byincreasing the concentration of NaCl in PBS (pH 7.4). It is reason-able that high NaCl concentrations can reduce the surface potentialby reduction of the Debye–Hückel length and hence the range ofelectrostatic interaction force [21,40]. Compared to heparin andNaCl, urea can more effectively remove adsorbed LDL from hepa-rin-modified PSu (Fig. 8C). Approximately 70% of the adsorbedLDL was eluted from the heparin-modified PSu surface by 8 M ureasolution, while there was no elution of the adsorbed LDL from theplain PSu. This can be ascribed to the fact that urea can interferewith stabilizing intramolecular interactions and disrupt thethree-dimensional structure of proteins mediated by non-covalentforces [41]. On the other hand, there was some irreversibly

adsorbed amount of LDL on both plain and heparin-modified PSufilms. Combined with the irreversibly adsorbed LDL, the heparin-modified PSu described here can participate in selective LDLadsorption and remains relatively inert to the non-specific adsorp-tion at the same time.

4. Conclusions

We have found that it is possible to covalently immobilize hep-arin onto the surface of a PSu film by a three-step synthesis meth-od. A heparin density up to 0.86 lg cm�2 could be achieved ondense PSu film. The hydrophilicity of the PSu surface was improvedgreatly by covalent immobilization of heparin. The water contactangle of PSu films was decreased from 89.6 ± 3.7� to 50.5 ± 3.2�after binding of 0.36 lg cm�2 heparin. The results of proteinadsorption show that the surface-bound heparin greatly enhancedthe adsorption of LDL both from single protein solutions and mix-tures with HSA that closely resemble the conditions in human plas-ma in both healthy and diseased subjects. In addition, compared toplain PSu, the adsorbed LDL could be easily desorbed from themodified PSu surface with heparin, NaCl or urea solution. Inconclusion, heparin-modified PSu membranes may have greatpotential for simultaneous application in hemodialysis and LDLapheresis.

1106 X.-J. Huang et al. / Acta Biomaterialia 6 (2010) 1099–1106

Acknowledgments

Support from Fresenius Medical Care Deutschland GmbH’sNephrocore Program and the National Natural Science Foundationof China (Grant No. 50703034) are gratefully acknowledged.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.actbio.2009.08.039.

References

[1] Yang CY, Chen SH, Gianturco SH, Bradley WA, Sparrow JT, Tanimura M, et al.Sequence, structure, receptor-binding domains and internal repeats of humanapolipoprotein-B-100. Nature 1986;323:738–42.

[2] Fox KM, Gandhi SK, Bullano MF, Ohsfeldt RL, Davidson MH. Low-densitylipoprotein levels and dyslipidemia treatment in patients diagnosed withatherosclerosis. Circulation 2008;117:E425–6.

[3] Chamberlain AM, Folsom AR, Schreiner PJ, Boerwinkle E, Ballantyne CM. Low-density lipoprotein and high-density lipoprotein cholesterol levels in relationto genetic polymorphisms and menopausal status: the atherosclerosis risk incommunities (ARIC) study. Atherosclerosis 2008;200:322–8.

[4] Shoji T, Hatsuda S, Tsuchikura S, Shinohara K, Kimoto E, Koyama H, et al. Smalldense low-density lipoprotein cholesterol concentration and carotidatherosclerosis. Atherosclerosis 2009;202:582–8.

[5] Shah PK. Low-density lipoprotein lowering and atherosclerosis progression—does more mean less? Circulation 2002;106:2039–40.

[6] Bambauer R, Schiel R, Latza R. Low-density lipoprotein apheresis: an overview.Ther Apher Dial 2003;7:382–90.

[7] Bosch T. Therapeutic apheresis—state of the art in the year 2005. Ther ApherDial 2005;9:459–68.

[8] Putz G, Eckes J, Schmah O, Winkler K, Wieland H. Elimination of liposomes bydifferent separation principles used in low-density lipoprotein apheresis. TherApher Dial 2008;12:2–12.

[9] Thompson GR, Lowenthal R, Myant NB. Plasma exchange in management ofhomozygous familial hypercholesterolemia. Lancet 1975;1:1208–11.

[10] Klingel R, Mausfeld P, Fassbender C, Goehlen B. Lipidfiltration—safe andeffective methodology to perform lipid-apheresis. Transfus Apher Sci 2004;30:245–54.

[11] Koll RT. LDL-therasorb immunoadsorption for the treatment of severehypercholesterolemia refractory to conventional therapy. Ther Apher Dial1998;2:142–6.

[12] Blessing F, Wang Y, Nagel D, Seidel D. The efficacy and safety of the newheparin-induced extracorporeal low-density lipoprotein precipitation system(Plasmat Futura) in comparison with the currently used system (PlasmatSecura). Ther Apher Dial 2004;8:33–8.

[13] Wang Y, Walli AK, Schulze A, Blessing F, Fraunberger P, Thaler C, et al. Heparin-mediated extracorporeal low density lipoprotein precipitation as a possibletherapeutic approach in preeclampsia. Transfus Apher Sci 2006;35:103–10.

[14] Asahi T, Yamamoto T, Kutsuki H. Blood purification therapies using dextransulfate cellulose columns liposorber and Selesorb. Ther Apher Dial 2003;7:73–7.

[15] Bosch T, Gahr S, Belschner U, Schaefer C, Lennertz A, Rammo J. Directadsorption of low-density lipoprotein by DALI-LDL-apheresis: results of aprospective long-term multicenter follow-up covering 12,291 sessions. TherApher Dial 2006;10:210–8.

[16] Bosch T, Thiery J, Gurland HJ, Seidel D. Long-term efficiency, biocompatibility,and clinical safety of combined simultaneous LdL-apheresis and hemodialysisin patients with hypercholesterolemia and end-stage renal-failure. NephrolDial Transplant 1993;8:1350–8.

[17] Cardin AD, Randall CJ, Hirose N, Jackson RL. Physical–chemical interaction ofheparin and human-plasma low-density lipoproteins. Biochemistry 1987;26:5513–8.

[18] Gigli M, Consonni A, Ghiselli G, Rizzo V, Naggi A, Torri G. Heparin binding tohuman plasma low-density lipoproteins—dependence on heparin sulfationdegree and chain-length. Biochemistry 1992;31:5996–6003.

[19] Ma KW, Ma L, Cai SX, Wang X, Liu B, Xu ZL, et al. Preparation of heparin-immobilized PVA and its adsorption for low-density lipoprotein fromhyperlipemia plasma. J Mater Sci Mater Med 2008;19:3255–61.

[20] Matharu Z, Arya SK, Sumana G, Gupta V, Malhotra BD. Self-assembledmonolayer for low density lipoprotein detection. J Mol Recognit 2008;21:419–24.

[21] Gaus K, Hall EAH. Surface plasmon resonance measurement of the binding oflow-density lipoprotein at a heparin surface. J Colloid Interface Sci1999;217:111–8.

[22] Rao M, Guo DQ, Jaber BL, Sundaram S, Cendoroglo M, King AJ, et al. Dialyzermembrane type and reuse practice influence polymorphonuclear leukocytefunction in hemodialysis patients. Kidney Int 2004;65:682–91.

[23] Kodras K, Benesch T, Neumann I, Haas M. Comparison of two dialysers(AN69ST vs. FX100) for heparin-free dialysis in patients with oralanticoagulation. Blood Purif 2008;26:226–30.

[24] Higuchi A, Shirano K, Harashima M, Yoon BO, Hara M, Hattori M, et al.Chemically modified polysulfone hollow fibers with vinylpyrrolidone havingimproved blood compatibility. Biomaterials 2002;23:2659–66.

[25] Chen H, Chen Y, Sheardown H, Brook MA. Immobilization of heparin on asilicone surface through a heterobifunctional PEG spacer. Biomaterials 2005;26:7418–24.

[26] Kang IK, Seo EJ, Huh MW, Kim KH. Interaction of blood components withheparin-immobilized polyurethanes prepared by plasma glow discharge. JBiomater Sci Polym Ed 2001;12:1091–108.

[27] Tzoneva R, Faucheux N, Groth T. Wettability of substrata controlscell–substrate and cell–cell adhesions. Biochim Biophys Acta Gen Subj 2007;1770:1538–47.

[28] Groth T, Klosz K, Campbell EJ, New RRC, Hall B, Goering H. Protein adsorption,lymphocyte adhesion and platelet-adhesion activation on polyurethane ureasis related to hard segment content and composition. J Biomater Sci Polym Ed1994;6:497–510.

[29] Higuchi A, Hashiba H, Hayashi R, Yoon BO, Sakurai M, Hara M. Serum proteinadsorption and platelet adhesion on aspartic-acid-immobilized polysulfonemembranes. J Biomater Sci Polym Ed 2004;15:1051–63.

[30] Koehler JA, Ulbricht M, Belfort G. Intermolecular forces between a protein anda hydrophilic modified polysulfone film with relevance to filtration. Langmuir2000;16:10419–27.

[31] Huang XJ, Xu ZK, Wan LS, Wang ZG, Wang JL. Surface modification ofpolyacrylonitrile-based membranes by chemical reactions to generatephospholipid moieties. Langmuir 2005;21:2941–7.

[32] Gatti G, Casu B, Hamer GK, Perlin AS. Studies on the conformation of heparinby 1H and 13C NMR spectroscopy. Macromolecules 1979;12:1001–7.

[33] Ye P, Xu ZK, Che AF, Wu J, Seta P. Chitosan-tethered poly(acrylonitrile-co-maleic acid) hollow fiber membrane for lipase immobilization. Biomaterials2005;26:6394–403.

[34] Nakamura K, Matsumoto K. Protein adsorption properties on a microfiltrationmembrane: a comparison between static and dynamic adsorption methods. JMembr Sci 2006;285:126–36.

[35] Bayramoglu G, Yalcin E, Arica MY. Adsorption of serum albumin and gamma-globulin from single and binary mixture and characterization of pHEMA-basedaffinity membrane surface by contact angle measurements. Biochem Eng J2005;26:12–21.

[36] Ji J, Feng LX, Barbosa MA. Stearyl poly(ethylene oxide) grafted surfaces forpreferential adsorption of albumin. Biomaterials 2001;22:3015–23.

[37] Martins MCL, Wang D, Ji J, Feng L, Barbosa MA. Albumin and fibrinogenadsorption on PU-PHEMA surfaces. Biomaterials 2003;24:2067–76.

[38] Vroman L, Adams AL. Adsorption of proteins out of plasma and solutions innarrow spaces. J Colloid Interface Sci 1986;111:391–402.

[39] Li D, Chen H, McClung WG, Brash JL. Lysine-PEG-modified polyurethane as afibrinolytic surface: effect of PEG chain length on protein interactions, plateletinteractions and clot lysis. Acta Biomater 2009;5:1864–71.

[40] Ramsden JJ, Prenosil JE. Effect of ionic strength on protein adsorption kinetics. JPhys Chem 1994;98:5376–81.

[41] Bennion BJ, Daggett V. The molecular basis for the chemical denaturation ofproteins by urea. Proc Natl Acad Sci USA 2003;100:5142–7.