lable at ScienceDirect

Biomaterials 56 (2015) 198e205

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

Multilayer hydrogel coatings to combine hemocompatibility andantimicrobial activity

Marion Fischer a, Maryam Vahdatzadeh a, Rupert Konradi b, Jens Friedrichs a,Manfred F. Maitz a, Uwe Freudenberg a, Carsten Werner a, *

a Max Bergmann Center of Biomaterials Dresden, Leibniz Institute of Polymer Research Dresden, Hohe Str. 6, 01069 Dresden, Germanyb BASF SE, Advanced Materials and Systems Research, D-67056 Ludwigshafen, Germany

a r t i c l e i n f o

Article history:Received 28 February 2015Received in revised form28 March 2015Accepted 29 March 2015Available online

Keywords:AntibacterialHemocompatibilitySilverDrug releaseHydrogel

* Corresponding author.E-mail addresses: werner@ipfdd.de, carsten.werne

http://dx.doi.org/10.1016/j.biomaterials.2015.03.0560142-9612/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

While silver-loaded catheters are widely used to prevent early-onset catheter-related infections [1], longterm antimicrobial protection of indwelling catheters remains to be achieved [2] and antiseptic func-tionalization of coatings often impairs their hemocompatibility characteristics. Therefore, this workaimed to capitalize on the antimicrobial properties of silver nanoparticles, incorporated in anticoagulantpoly(ethylene glycol) (PEG)-heparin hydrogel coatings [3] on thermoplastic polyurethane materials. Forprolonged antimicrobial activity, the silver-containing starPEG-heparin hydrogel layers were shieldedwith silver-free hydrogel layers of otherwise similar composition. The resulting multi-layered gel coat-ings showed long term antiseptic efficacy against Escherichia coli and Staphylococcus epidermidis strainsin vitro, and similarly performed well when incubated with freshly drawn human whole blood withrespect to hemolysis, platelet activation and plasmatic coagulation. The introduced hydrogel multilayersystem thus offers a promising combination of hemocompatibility and long-term antiseptic capacity tomeet an important clinical need.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Bacterial infections related to indwelling medical devices rankas the fifth-leading cause of hospital patients death in the US [4].Central venous catheter-associated bloodstream infections occurwith a frequency of 5 per 1000 catheter days and result in a lethalityrate of 20% [5]. Therefore, in addition to sterile handling and sys-temic administration of anticoagulants, antimicrobial prevention isof paramount importance for the safety of central venous catheterapplications [6]. Current strategies to locally prevent infectionsinclude bactericidal and bacteriostatic approaches using antisepti-cally impregnated coatings or antibiotic lock therapies. For thatpurpose, both the catheter bulk material and hydrophilic hydrogelcoatings, providing catheters with a “slippery-when-wet” lubri-cious surface, were customized to embed and release bioactiveswith silver being the most frequently applied bioactive component.In these approaches, silver is applied as silver bulk material, in ionicform or as silver nanoparticles (AgNPs) referring to both oxidation

r@tu-dresden.de (C. Werner).

states (Ag0/Agþ). Although the antimicrobial mechanism of silver isstill controversially discussed [7e13], there both oxidation statesare considered to be active against bacteria, most probably througha conversion of bacterial sulfhydryl (R-SH) groups [9]. For silverions, the biological activity is multifold including the inhibition ofcell respiration and the inactivation of enzymes to effects on DNAreplication and cell division [7,8].

A common strategy for sustained Agþ release is the reduction ofionic silver in solution to form AgNPs with a large surface-to-volume ratio [14]. Previous studies have shown that AgNPs caneither penetrate bacterial cells [10] or become deposited on thebacterial cell wall, affecting membrane permeability as well aselectrolyte and metabolite transport [11]. The cytotoxicity of silvernanoparticles results from interactions with membrane proteins,the activation of signaling pathways and the inhibition of cellproliferation [12,13]. The toxicity of silver for bacterial andeukaryotic cells has been extensively tested and compared inearlier studies, showing a significantly lower toxicity of silver forthe eukaryotic cells [15,16].

The use of silver-loaded hydrogel coatings for central venouscatheters dates back to 1998 [2] and was found suitable to preventearly-onset catheter-related infections but concluded to be limited

M. Fischer et al. / Biomaterials 56 (2015) 198e205 199

for long lasting protection [1,2]. Recently reported strategies haveexploited the in situ formation of AgNPs within swollen hydrogelnetworks of catheter coatings [17e19] which were developed tominimize protein adsorption and platelet adhesion [18]. However,the antiseptic functionalization of biomaterials with silver-basedcompounds often is associated with a loss of the beneficial hemo-compatibility characteristics since silver not only affects microor-ganisms, but also can exert undesired side effects on mammaliancells [20,21]. Thus, biomaterial coatings combining sustained, long-term delivery of antimicrobial active silver and non-thrombogenicproperties are particularly important targets of current research[22].

Herein, we present a new concept to address these challenges. Arecently introduced star-shaped poly(ethylene glycol) (PEG)-hep-arin hydrogel [23] was grafted as a thin film coating onto ther-moplastic polyurethane bulk materials and further modified withsilver by incubation in AgNO3 solution (see Scheme 1). The buildingblocks, star-shaped PEG and heparin, widely applied in the surfacemodification of biomaterials due to their protein-resisting andanticoagulant characteristics, were covalently linked to form ahydrogel network. In an effort to further modulate this system, amultilayer gel coating was developed, where a silver loadedhydrogel is coated onto the surface of the polyurethane bulk ma-terial with a second silver-free hydrogel layer on top to act as adiffusion barrier for prolonged silver release and for avoiding directcellular interactions with AgNPs at the coating-blood interface.These multilayer hydrogel coatings were thoroughly evaluated inwhole blood hemocompatibility assays for coagulation and in-flammatory activation. The antiseptic capacity of the coating wasverified with different relevant bacterial strains.

2. Materials and methods

2.1. Preparation of starPEG-heparin hydrogel coatings

Biohybrid starPEG-heparin gels were prepared using amino functionalized four-armed starPEG and EDC-s-NHS activated heparin as previously described [3]. Briefly,a solution of heparin (14 000 g/mol; Calbiochem, Darmstadt, Germany) was acti-vated using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxysulfosuccinimide (sulfo-NHS) (ratio 2:1) at a stoichiometric balanced con-centration of sulfo-NHS to the starPEGeNH2egroups. Subsequently, a starPEG so-lution (10 000 g/mol Polymer Source Inc., Dorval, Canada) was added to the

Scheme 1. Schematic depiction of the explored v

activated heparin to a final concentration of 133 mg/ml. The starPEG-heparinmixture was spread on thermoplastic polyurethane discs (Elastollan® 1190A, pro-vided by BASF Polyurethanes GmbH, Lemf€orde, Germany) and was subsequentlycovered with a hydrophobized glass coverslip (by dipping the coverslip into a Sig-macote® [Sigma Aldrich, St. Louis, USA]). The coverslip facilitates the formation of adefined gel disc and can be easily removed after gel formation. Alternatively, thehydrogel was applied to thermoplastic polyurethane (TPU) tubes by dip coating.After overnight polymerization and removal of the glass coverslip, gels were swollenfor 12 h in acetate buffer (160 mmol/l; 314 mOsm). During swelling the acetatebuffer was exchanged several times to ensured complete removal of EDC/s-NHS andunreacted hydrogel building blocks. If needed, unmodified heparin was mixed with1% Alexa-488 labeled heparin (Invitrogen, Darmstadt, Germany) to allow for fluo-rescence microscopic analysis.

Different TPU surface modification techniques were evaluated to ensure stablegrafting of starPEG-heparin hydrogels toTPU substrates. TPUwas either treatedwithammonia plasma (MicroSys apparatus by Roth&Rau, Wüstenbrand, Germany;treatment time: 10 s, power: 600 W, ammonia gas flow: 15 standard cubic centi-meter per minute, pressure: 7 � 10�3 mbar) or air plasma using a plasma cleaner(Harrick, Ithaca, NY, USA; treatment time: 10 min pressure: approx. 0.2 mbar).Alternatively samples were treated with a low energy electron beam (ADU,Advanced Electron Beams, Wilmington, USA 100 keV) with an absorbed dose of258 kGy under inert N2 atmosphere or in air conditions to induce reactive groups onthe TPU surface. Fluorescently labeled hydrogels grafted to TPU were analyzed byconfocal laser scanning microscopy (cLSM, SP-5, Leica Microsystems, Wetzlar, Ger-many) in xy- and z-scan for coating thickness and coating homogeneity,respectively.

The thickness of the swollen starPEG-heparin hydrogels on planar TPU sub-strates was approx. 50e100 mm and the heparin content was approx. 0.3 mg heparinper cm2 surface area. StarPEG-heparin gels coated to TPU tubes revealed a thicknessof 16e34 mm. For rheometry and AgNP analysis, freestanding gel discs were pre-pared by applying the gel mixture between two Sigmacote®-coated coverslips.

2.2. Antiseptic modification of hydrogels

Hydrogels were modified with AgNPs by immersion into acetate buffer con-taining AgNO3 (Sigma Aldrich, Munich, Germany) for 1 h at room temperature. Silvernitrate solutions were freshly prepared prior to use. Solutions and samples werestored in the dark. The applied AgNO3 concentrations were chosen according to theminimal inhibitory concentration (MIC) given by Copra et al. and Ip et al. [24,25] andranged from 80 to 4000 ppm (corresponding to 1.5 mg- 75 mg AgNO3/mm3 gel vol-ume). A volume of 5.0 ml AgNO3 solution was applied per gel scaffold of 265 mm3.

2.3. Preparation of multi-layered hydrogels

Multi-layered hydrogels were prepared by vacuum drying TPU-bound AgNP-loaded hydrogels at room temperature. Subsequently, a reaction mixture of starPEGand heparin was applied to the dried gel layer and covered with Sigmacote® cov-erslips as described above. After polymerization, coverslips were removed and gels

ariants of multifunctional hydrogel coatings.

M. Fischer et al. / Biomaterials 56 (2015) 198e205200

were transferred to acetate buffer that was repeatedly exchanged to removeunreacted EDC/s-NHS and hydrogel building blocks. For analysis of the thickness ofthe upper and lower gel layer, 1% of Alexa-488-heparin or Alexa-647-heparin wereadded to the respective reaction mixtures.

2.4. Uptake and release of antiseptic substances to/from starPEG-heparin hydrogels

After silver functionalization of starPEG-heparin hydrogels, the amount ofAgNO3 remaining in the loading solution was analyzed by atomic absorption spec-troscopy (AAS, ZEENIT 700, Analytik Jena, Germany). The amount of uploadedAgNO3 was compared against standards of defined AgNO3 concentration. For AAS,solutions were acidified to pH 1.

Silver-loaded hydrogels were analyzed by XPS (Kratos Analytical, England) andcompared to AgNO3 for evaluating the Ag 3d spectrum of silver in the hydrogelnetwork.

The release of the silver was monitored using acetate buffer as release medium.Gels were incubated in 5 ml acetate buffer and subsequently the whole releasemedium was collected at defined time points (2 and 6 h), replaced with the samevolume of fresh solution and analyzed by AAS.

2.5. Characterization of antiseptic hydrogels

The rheological properties of silver-functionalized gels were determined byplateeplate rheometry (ARES LN2, TA Instruments, Germany). Freestanding gel discswere functionalized with 0, 800, and 4000 ppm silver concentration. Measurementsof storage moduli were determined at a frequency range of 101e10�1 rad s�1. Threeindependent measurements were performed.

To visualize the AgNPs, hydrogel samples were subjected to dehydration inincreasing ethanol concentrations, air-dried, and analyzed by scanning electronmicroscopy (SEM) (NEON40, Carl Zeiss Microscopy GmbH, Oberkochen, Germany).

2.6. Zone of inhibition assay

The antimicrobial activity of the silver-modified hydrogels was determined by azone of inhibition (ZOI) assay. For the assay, Müller Hinton agar (Sigma Aldrich,Munich, Germany) plates were prepared following the manufacturer instructionsand stored at 4 �C until use. Overnight cultures of Escherichia coli K12 DH5 (DSMZ,Germany) and Staphylococcus epidermidis PCI 1200 (ATCC, USA) were diluted to anOD600 of 0.1. A sterile cotton swab was dipped into the adjusted suspension and thesurface of a Müller-Hinton agar plate was inoculated by streaking the swab over theentire surface. Subsequently, hydrogel samples were placed in the middle of an agar.The plates were incubated overnight at 37 �C. Then, the zones of bacterial growthinhibition were measured using a digital caliper. For long-term ZOI experiments,samples were transferred daily to freshly inoculated Müller Hinton agar plates.

2.7. Whole blood incubation assay

Hydrogel samples were incubated with whole blood using in house designedincubation chambers as described elsewhere [26]. Briefly, samples were incubatedwith a blood volume of 1.95 ml, resulting in a surface: volume ratio of 3.2 cm�1.Venous blood was obtained from two AB0 compatible healthy voluntary donors.Blood of the two donors was pooled and anticoagulated with heparin (1 IU/ml). Thechambers were incubated for 2 h at 37 �C under permanent rotation to avoidsedimentation of blood cells. All surfaces were tested in triplicates.

Material induced hemolysis of red blood cells was evaluated by measuring theformation of cyanmethemoglobin at 450 nm with reference wavelength 490 nmagainst the initial plasma as a blank. For this, 20 ml citrate plasma was mixed with200 ml Drabkin solution (Sigma Aldrich, Munich, Germany). A mixture of 100 mlwhole blood in 900 ml distilled water served as reference of 10% hemolysis and serialdilutions thereof were made as standards. Commercial enzyme-linked immuno-sorbent assays (ELISAs) were used to characterize blood reactions including for-mation of prothrombin fragment 1þ2 (F1þ2; Enzygnost F1þ2, Siemens Healthcare,Marburg, Germany) and complement activation (C5a; Enzygnost C5a micro, DRGDiagnostica, Marburg, Germany).

For flow cytometry, cells were stained with phycoerythrin labeled anti CD11b(4 mL per 20 mL blood; clone ICRF44, Biozol, Eching, Germany) and PerCPCy5-labeledCD41a (4 ml per 20 ml blood; BD Biosciences, California, USA). To avoid cell activationduring the 30 min staining time, 0.1% NaN3 was added [27]. Subsequently, cells werefixed and red blood cells were lysed with FacsLyse (BD, Heidelberg, Germany).Analysis was performed in a FACSCalibur (BD, Heidelberg, Germany). Granulocytesand monocytes were identified for their characteristic appearance in the forward vs.side scatter plot. In general, cells were collected until 10 000 granulocytes weremeasured. The median intensity of the CD11b signal of the granulocyte populationand the rate of CD41a positive events in the granulocyte or monocyte population,respectively, were quantified. Data was analysed with Summit v.3.1 (DakoCytoma-tion, Fort Collins, USA).

Sample surfaces were rinsed blood-free with PBS for SEM investigation, sampleswere extensively washed with PBS to remove traces of blood, fixed with 2%glutaraldehyde in PBS (Serva Electrophoresis GmbH, Heidelberg, Germany) for 1 h,dehydrated in increasing ethanol concentrations and air dried. Before SEM

inspection (SEM XL 30 FEG, FEI-Philips, Eindhoven, The Netherlands), samples weregold sputter coated (SCD 050, BAL-TEC, Schalksmühle, Germany).

The effect of soluble silver nitrate on hemocompatibility parameters was eval-uated by adding aqueous solutions of AgNO3 to 2 ml blood in an incubation chambermade of inert polytetrafluorethylene (PTFE). Final silver concentrations in wholeblood were 800 and 4000 ppm. Soluble silver samples were tested in parallel tosilver hydrogels to guarantee equal test conditions.

2.8. Statistical analysis

Data are presented asmean ± standard deviation unless stated otherwise. Groupcomparison was performed by one way analysis of variance (ANOVA) using studentt-test.

3. Results

3.1. Characterization of hydrogels and hydrogel coatings on TPU

Hydrogel immobilization to TPU substrates requires modifica-tion of the TPU surface to increase the number of available reactivegroups for the substrate-gel coupling reaction. Different TPU sur-face modification techniques were evaluated, amongst them oxy-gen and ammonia plasma and electron beam treatment in nitrogenatmosphere and under air conditions. Electron beam treatment ofTPU induced an increase in O/C ratio as detected by XPS (data notshown). The C1s peak of the TPU can be deconvoluted into threedifferent binding energies at 285.0 eV (aliphatic compounds, CeH),286.5 eV (ether, eCeOeC) and 289.5 eV (carboxyl, eC]O). Thepeak containing carboxyl groups increased from 6 to 9% whilst adecrease in ether group peak occurred from 38 to 35%. The soproduced carboxyl groups provide binding sites for gel graftingprocesses. Activation of gel mixture with EDC/NHS was done toenable crosslinking between gel components heparin and PEG.Addition of the gel mixture containing EDC/NHS to the underlyingirradiated TPU leads to activation of eCOOH groups from the sub-strate. By that, a cohesive gel coupling to the TPU substrate throughlinkage to still available PEG-derived NH2 groups is realized.Hydrogel coatings on electron beam-irradiated TPU in air condi-tions were analyzed by confocal laser scanning microscopy andrevealed a homogeneous gel layer (see Fig. S1 in supplementalpart).

Rheometry measurements did not reveal a significant change ingel storage modulus between silver-functionalized hydrogels andnon-functionalized native gels (Fig. 1C). All gels showed a similarstiffness in a range of 2.6e3.6 kPa. Silver nanoparticle analysisperformed by SEM showed an equal particle distributionwithin thehydrogel and particle size was determined to be 25e50 nm(Fig. 1B). This correlates well with particle dimensions of 20e30 nmreported in literature for silver hydrogels by various authors[17,18,28].

Alexa-labeled single-layer hydrogels grafted to TPU wereanalyzed by cLSM in xy- and z-scans for coating thickness andcoating homogeneity, respectively. Coating thicknesses of16e34 mm were detected for single starPEG-heparin hydrogellayers on TPU substrates.

Silver hydrogel coatings were further equipped with an addi-tional gel layer to enable modulation of release profiles and thematerials biocompatibility. To this end, the silver hydrogel coatingswere shielded with a second silver-free layer that acts as a physi-ological barrier. In detail, still available COOH and NH2 groupswithin the first gel layer were used for coupling reactions to com-plementary functional groups of the second gel layer. The analysisof themultilayer system includes the analysis of layer thickness andhomogeneity as well as stability tests. The obtained data (presentedin the supplement) confirmed the presence of two distinct layers,with 24.6 ± 13.1 mm and 16.1 ± 7.0 mm for inner and outer layer,respectively (Fig. S1 of the supplement). The stability of the

Fig. 1. A: Photographs of single-layered starPEG-heparin hydrogels modified withdifferent AgNO3 concentrations; B: SEM images of silver loaded hydrogels showingnanoparticle size and distribution; C Rheological characterization of silver-function-alized hydrogels determined by plateeplate rheometer.

M. Fischer et al. / Biomaterials 56 (2015) 198e205 201

multilayer gel coatings on a TPU substrate was verified by exposureto PBS at 37 �C for time periods of 3 weeks (see Fig. S1C).

3.2. Silver uptake and release from single-layered PEG-heparinhydrogels

The silver uptake to hydrogels incubated with a solution of 1000and 1500 ppm AgNO3 (i.e. 19 and 28 mg AgNO3/mm3 scaffold)resulted in a 51% and 29% loading efficiency respectively (Fig. 2).Silver release from hydrogels within 10 days was found to be 8% ofthe silver amount in the loading solution for both loading con-centrations. Our photoelectron spectroscopy data uncover a shift of

Fig. 2. Uptake (left) and release (right) after 10 days of silver from single-layered silver-1500 ppm gel ¼ 28 mg AgNO3/mm3 scaffold) in mg/mm3 gel scaffold measured by AAS in acconcentration is indicated in front of the columns.

the Ag 3d signal from 368.6 eV for AgNO3 to 367.8 eV for silverloaded hydrogels. This indicates the formation of Ag2O nano-particles within the starPEG-heparin hydrogels upon incubationwith AgNO3 solution. We assume that the Agþ/Ag2O species isstabilized via the negatively charged sulfate groups of the heparinwithin the hydrogel network.

3.3. Evaluation of antiseptic efficacy

The inhibition zone assay was used for single-layered hydrogelcoatings loaded with 0, 800, and 4000 ppm AgNO3 (SL-0, SL-800,and SL-4000) and for the multi-layered 4000 ppm sample (ML-4000) to determine the inhibition of bacterial growth throughrelease of the antiseptic substance from the coating (Fig. 3). Thedata reveals a dose-dependent inhibition of E. coli and S. epidermidisgrowth by silver hydrogels and a prolonged antiseptic effect for theSL-4000 ppm sample compared to SL-800 gels. The multilayersample loaded with 4000 ppm silver did show a prolonged anti-microbial affect for the E. coli strain.

3.4. Hemocompatibility of silver hydrogels

Whole blood incubation assay was performed for single layeredhydrogel coatings loaded with 0, 800, and 4000 ppmAgNO3 and forthe multilayer sample loaded with 4000 ppm AgNO3.

3.5. Hemolysis and coagulation

Single-layered hydrogel coatings loaded with 800 ppm AgNO3revealed values of 1.1% hemolysis events (Fig. 4). The SL-4000 gelsreached values of 2.7% being clearly in critical range with respect toits hemocompatibility. Only the 4000 ppm-loaded multilayer gelsprevented lysis of red blood cells obtaining hemolysis levels thatare in the range of unloaded gels. For the blood incubation withsoluble silver a concentration up to 800 ppm did not induce he-molysis (0.6%), whereas for 4000 ppm silver hemolysis was trig-gered (5.6%).

The activation of plasmatic coagulation was significantly lowerfor theML-4000 gel compared to both SL-800 and SL-4000 being inthe range of unloaded SL-0 gels (Fig. 5 left). The procoagulant effectof soluble silver nitrate was found to be elevated for both concen-trations tested (800 and 4000 ppm AgNO3).

We likewise observed increased platelet activation measured asplateletegranulocyte conjugate formation [29,30] on silver loaded

functionalized hydrogels (1000 ppm loading solution ¼ 19 mg AgNO3/mm3 scaffold;etate buffer solution. Loading efficiency and release rate in % of initial AgNO3 loading

Fig. 3. Inhibition zones of E. coli (left) and S. epidermidis (right) on single- and multi-layered hydrogels loaded with different amounts of silver (800 and 4000 ppm) determined byinhibition zone assay with daily transfer of gel coatings to fresh bacterial agar plates.

Fig. 4. Hemolysis of whole blood after 2 h incubation at 37 �C determined by cyan-methemoglobin detection; statistical significances: * is significantly different from0 ppm and ML- 4000 ppm sample (P < 0.05). Hemolysis of soluble silver in wholeblood under same conditions was 0.6 ± 0.2 and 5.6 ± 3.0% for 800 ppm and 4000 ppmrespectively.

Fig. 5. left: Prothrombin fragment F1þ2 measured in plasma after whole blood incubation aand 441 ± 219 nmol/l for 800 ppm and 4000 ppm respectively; right: Plateletegranulocytsilver in whole blood under same conditions was 96 ± 4 and 99 ± 1% for 800 ppm and 400ML- 4000 ppm sample (P < 0.05).

M. Fischer et al. / Biomaterials 56 (2015) 198e205202

single-layer hydrogels (SL-800 and SL-4000) (Fig. 5 right). Thevalues for unloaded gel coatings (SL-0) and ML-4000 samples (allbelow 33.3%) were significantly reduced and in the range of initialvalues of the experiment (i.e. control, 24.5% ± 8.7). Soluble silversamples of both concentrations (800 and 4000 ppm) revealednearly 100% plateletegranulocyte conjugate formation.

3.6. Inflammation

The inflammatory response measured as complement fragmentC5a in blood after material incubation was low for all hydrogelcoatings tested, increased inflammatory response in critical rangewas only found for soluble 4000 ppm AgNO3 (Fig. 6). Although apossible AgNP derived pro-inflammatory and toxic effect is dis-cussed as a result of induced ROS formation [31] and increasedmacrophage IL-6 and IL-10 for 5e10 mg/ml silver NP levels [28]were found before, the in here presented materials did not showany inflammatory response related to silver functionalization forthe concentration used.

Granulocyte activation was not increased for silver loaded gelscompared to unloaded gels. Adhesion of granulocytes to the sam-ples (characterized by SEM-pictures) is outlined in the inset ofFig. 6. Elevated number of adherent granulocytes with highest cell

ssay by ELISA; F1þ2 of soluble silver in whole blood under same conditions was 23 ± 18e conjugates measured in whole blood after surface incubation; Conjugates of soluble0 ppm respectively; statistical significances: * is significantly different from 0 ppm and

Fig. 6. Complement fragment C5a detected in plasma after whole blood incubation for2 h at 37 �C detected by ELISA; inset: Cell adhesion to hydrogel coatings after in vitrowhole blood incubation assay (2 h, 37 �C) done by ESEM; C5a of soluble silver in wholeunder same conditions was 2.6 ± 1.4 and 12.7 ± 5.3 mg/l for 800 ppm and 4000 ppmrespectively.

M. Fischer et al. / Biomaterials 56 (2015) 198e205 203

spreading was found on SL-4000, whereas for all other gel coatingsno or low numbers of cells were found. Adherent cells on ML-4000are mainly erythrocytes, having minor influence on inflammatoryresponses.

4. Discussion

Combining antimicrobial and anticoagulant properties withinbiomaterial coatings defines a practically important challenge ofcurrent research. With the reported approach, we were able tomerge two previously established principles, namely the antimi-crobial properties of AgNPs and the anticoagulant potential ofheparin-containing biohybrid hydrogels within a multifunctionalbiomaterial.

4.1. Silver functionalization of starPEG-heparin hydrogels

Silver functionalization of hydrogels was successfully performedresulting in a coating that shows a dose-dependent uptake andrelease of ionic silver to act antimicrobially. By this, a coating sys-tem of tunable silver supply is established that enables localizedand surface-mediated delivery of bioactive molecules. Since theuptake process consists of an exchange of the counter ions of theheparin sulfate groups (Naþ or Hþ) by Agþ ions from the solution,the sulfate groups of heparin govern the Agþ entrapment withinthe polymer network [17]. The loading efficiency depends on thesilver concentration of the loading solution, on the incubation time,incubation temperature and on the solvent of loading solution [2],and resulted in our case in a maximum of 40e51% loading effi-ciency. SEM images outlined in this work revealed nanoparticleformation in silver-functionalized gels resulting from a redox re-action of Agþ to AgNPs, that serve as a depot for the sustained Agþ

release at the coating-tissue interface. Stevens et al. described anAgNPs composition of metallic Ag (Ag0) and about 12% Agþ, withthe ions residing at the periphery of the particles [18]. A stabiliza-tion of NPs in hydrogel network for a period of 6 month throughCOOH-groups of the co-polymeric chains was further described byother authors [17]. Alternatively the formation of Ag2O-NPs aredescribed [32]. Our XPS data suggest the formation of Ag2O

nanoparticles within starPEG-heparin hydrogels upon AgNO3 in-cubation. We assume that the Agþ/Ag2O species is stabilized via thenegatively charged sulfate groups of the heparin within thehydrogel network.

4.2. Single layer hydrogel coatings

The antimicrobial effect of the silver loaded gel coatings iscaused by Agþ at the outer AgNP surface or through released Agþ.The antimicrobial action of AgNPs and Ag ions was discussed before(see introduction and [9,12,13]). AgNPs incorporated in a hydrogelmatrix can serve as a reservoir for sustained Agþ release withrelease patterns modulated by hydrogel characteristics. Silver-containing hydrogels allow AgNPs located at the outer surface toattack cellular membranes without being taken up [18] with ahigher antibacterial activity for smaller NPs (200 nm vs 100 nm)and an optimal Agþ concentration for bactericidal action of660 ppm [17]. Lower bacterial adhesion to Ag-hydrogel-catheterwas found compared to Ag-free hydrogel catheter (72 h and 7days incubation) [2].

We herein report on starPEG-heparin hydrogel coatings loadedwith different silver concentrations. Silver hydrogel coatings with astiffness of 2.6 kPa have a calculated mesh size of 12 nm that en-ables entrapment of AgNPs inside the gel network. In long-terminhibition assays we observed a dose-dependent coating derivedantiseptic activity correlating to the uploaded silver amount. Theobserved inhibition curves follow a steep decrease with abrupttermination after 2 days for low rate functionalized silver gels and5e6 days for 4000 ppm gels.

Adverse effects of silver nanoparticles on in vitro hemo-compatibility were observed by Stevens et al. who found athrombogenic effect for 1 mM released Agþ, that causes a thrombingeneration time of 4 min (instead of 15 min) and an immediatethrombin formation for a concentration of 3 mM Agþ. Coagulationactivation was described to originate from increased platelet acti-vation occurring at Ag-coatings [18]. The authors found thatplatelet collision with surface-bound Ag induced an Agþ transferfrom the particle to the outer phospholipid-membrane of theplatelet with subsequent Agþ desorption leading to platelet acti-vation without adhering to the surface.

We likewise observed increased platelet activation on single-layered silver hydrogels (SL-800, SL-4000) compared to both theunloaded gel coatings (SL-0) and initial values. SL-4000 gels alsoshowed highest prothrombin levels.

We calculated the amount of released AgNO3 that would beexpected in our whole blood incubation setting based on our up-take and release studies done in buffer systems. According to this,the SL-4000 and SL-800 sample would release 320 ppm and64 ppm AgNO3, respectively (considering the two gel coatings forbottom and top side of the incubation chambers and a 2 ml bloodvolume). Based on our studies with soluble silver, this low level ofreleased Ag does neither induce hemolysis nor prothrombin acti-vation, thus we conclude that the hemolytic and procoagulant ef-fect originates from hydrogel-bound silver as the released amountis too low to increase both parameters to that extent.

Micrographs of SL-4000 showed a high spreading degree ofadherent granulocytes with concomitant platelet adhesion andfibrin deposition. Single layered samples induced coagulationactivation and hemolysis of 2e3%, most probably through surface-bound silver that is associated with AgNPs.

4.3. Multilayer hydrogel coatings

Considering the poor hemocompatibility given by the single-layer coating, an advanced system was developed to provide both

Scheme 2. Proposed mechanism of hydrogel mediated reactions in whole blood, adapted from Ref. [13].

M. Fischer et al. / Biomaterials 56 (2015) 198e205204

antimicrobial and hemocompatible capacities. For this, a secondsilver-free layer was deposited on top of the silver hydrogels. Thiswas achieved by carbodiimide chemistry, linking carboxyl andamine groups of multilayer gel components. Multilayer coatingstrategies are widely used to fine-tune the supply of bioactivesthrough the multilayer architecture. Phospholipid polymers andpolyelectrolytes rank among multilayer systems to adjust releasekinetics of drugs such as antibiotics [33,34] and anticancer drugs[35] in biomedical devices or to modulate the provision of boneminerals [36] to support bone mineralization processes.

The multilayer system presented here meets two indispensablerequirements given by blood contacting antiseptic surfaces. Theintroduction of a second gel layer to act as a diffusion barrier servedto adapt silver release kinetics and additionally enabled to over-come observed adverse procoagulant effects of single-layer gels.Themultilayer sample (ML-4000) achieved similar inhibition zonesto SL-4000 over the whole time period for both bacterial strainstested, with an even prolonged effect for the E. coli strain. By addinga silver-free gel layer of variable layer thickness, a release gradientis created inside the inert gel layer that influences the silver ionconcentration in the surrounding medium enabling the tuning ofrelease dynamics for the coating system.

Hemocompatibility parameters were significantly diminishedfor the ML-4000 gels regarding hemolysis and coagulation pa-rameters being in the range of initials and silver free gel coat-ings. The multi-layered hydrogels not only serve to create agradient of bioactive substances, but also provides a physicalbarrier at the cell-coating boundary to prevent unfavorable cell

and enzyme activation and subsequent undesired host responses(Scheme 2).

5. Conclusion

Multilayered silver-functionalized starPEG-heparin hydrogelscombining antiseptic and hemocompatible properties have beendeveloped as a versatile surface coating to fight the infection ofblood contacting medical products such as central venous cathe-ters. The introduced multilayer architecture was demonstrated tobalance hemocompatibility and antiseptic efficacy. Specifically, asilver-free starPEG-heparin layer on top of an AgNP containingstarPEG-heparin layer resulted in a hemocompatible material withdurable antimicrobial functionality. Addressing a persisting prob-lem of medical devices in cardiology, the explored starPEG-heparinmultilayer concept can obviously be used in a variety of additionalapplications that require the long-term release of bioactives to becombined with excellent hemocompatibility.

Acknowledgment

We thank Monique Marx (preparation of surfaces and hemo-compatibility assays) and Martina Franke (gold evaporation) fortechnical assistance. We furthermore gratefully acknowledge con-tributions by Dr. Heike Hund (ITM, TU Dresden) and Dr. Petr For-manek (IPF Dresden) for performing AAS and SEM analyses,respectively. We thank in particular Dr. Mirko Nitschke for surfaceplasma treatments, XPS measurements and the related scientific

M. Fischer et al. / Biomaterials 56 (2015) 198e205 205

discussions. We are thankful for the contributions of Dr. Elke Feese,Catharina Hippius and Stephan Saum (all BASF-SE) for scientificdiscussions.

Appendix A. Supplementary data

Supplementary data related to this article can be found online athttp://dx.doi.org/10.1016/j.biomaterials.2015.03.056.

References

[1] Boswald M, Mende K, Bernschneider W, Bonakdar S, Ruder H, Kissler H, et al.Biocompatibility testing of a new silver-impregnated catheter in vivo. Infec-tion 1999;27(Suppl. 1):S38e42.

[2] Gatter N, Kohnen W, Jansen B. In vitro efficacy of a hydrophilic central venouscatheter loaded with silver to prevent microbial colonization. Zentralblatt fürBakteriol 1998;287:157e69.

[3] Freudenberg U, Hermann A, Welzel PB, Stirl K, Schwarz SC, Grimmer M, et al.A star-PEG-heparin hydrogel platform to aid cell replacement therapies forneurodegenerative diseases. Biomaterials 2009;30:5049e60.

[4] Wenzel RP, Edmond MB. The impact of hospital-acquired bloodstream in-fections. Emerg Infect Dis 2001;7:174e7.

[5] CDC. National Nosocomial Infections Surveillance (NNIS) system report, datasummary from October 1986. Am J Infect Control 1998:522e33.

[6] Brun-Buisson C, Doyon F, Sollet JP, Cochard JF, Cohen Y, Nitenberg G. Pre-vention of intravascular catheter-related infection with newer chlorhexidine-silver sulfadiazine-coated catheters: a randomized controlled trial. Intens CareMed 2004;30:837e43.

[7] Castellano JJ, Shafii SM, Ko F, Donate G, Wright TE, Mannari RJ, et al.Comparative evaluation of silver-containing antimicrobial dressings anddrugs. Int Wound J 2007;4:114e22.

[8] Jung WK, Koo HC, Kim KW, Shin S, Kim SH, Park YH. Antibacterial activity andmechanism of action of the silver ion in Staphylococcus aureus and Escherichiacoli. Appl Environ Microbiol 2008;74:2171e8.

[9] Chmielewska DK, Sartowska B, Starosta W, Walo M. Radiation synthesis ofsilver nano- and microparticles in cellulose fibres. Nukleonia 2010;55:345e9.

[10] Morones JR, Elechiguerra JL, Camacho A, Holt K, Kouri JB, Ramirez JT, et al. Thebactericidal effect of silver nanoparticles. Nanotechnology 2005;16:2346e53.

[11] Sondi I, Salopek-Sondi B. Silver nanoparticles as antimicrobial agent: a casestudy on E. coli as a model for gram-negative bacteria. J Colloid Interface Sci2004;275:177e82.

[12] McShan D, Ray PC, Yu H. Molecular toxicity mechanism of nanosilver. J FoodDrug Anal 2014;22:116e27.

[13] Brz�oska K, Meczy�nska-Wielgosz S, Stepkowski TM, Kruszewski M. Adaptationof HepG2 cells to silver nanoparticles-induced stress is based on the pro-proliferative and anti-apoptotic changes in gene expression. Mutagenesis2015.

[14] Abou El-Nour KMM, Aa Eftaiha, Al-Warthan A, Ammar RAA. Synthesis andapplications of silver nanoparticles. Arabian J Chem 2010;3:135e40.

[15] Lansdown AB. Silver in health care: antimicrobial effects and safety in use.Curr Prob Dermatol 2006;33:17e34.

[16] Kvitek L, Panacek A, Prucek R, Soukupova J, Vanickova M, Kolar M, et al.Antibacterial activity and toxicity of silver e nanosilver versus ionic silver.J Phys Conf Ser 2011;304:12029.

[17] Thomas V, Yallapu MM, Sreedhar B, Bajpai SK. A versatile strategy to fabricatehydrogelesilver nanocomposites and investigation of their antimicrobial ac-tivity. J Colloid Interface Sci 2007;315:389e95.

[18] Stevens KN, Crespo-Biel O, van den Bosch EE, Dias AA, Knetsch ML,Aldenhoff YB, et al. The relationship between the antimicrobial effect ofcatheter coatings containing silver nanoparticles and the coagulation ofcontacting blood. Biomaterials 2009;30:3682e90.

[19] Stevens KN, Croes S, Boersma RS, Stobberingh EE, van der Marel C, van derVeen FH, et al. Hydrophilic surface coatings with embedded biocidal silvernanoparticles and sodium heparin for central venous catheters. Biomaterials2011;32:1264e9.

[20] Beer C, Foldbjerg R, Hayashi Y, Sutherland DS, Autrup H. Toxicity of silvernanoparticlesdnanoparticle or silver ion? Toxicol Lett 2012;208:286e92.

[21] Kim S, Choi JE, Choi J, Chung KH, Park K, Yi J, et al. Oxidative stress-dependenttoxicity of silver nanoparticles in human hepatoma cells. Toxicol In Vitro2009;23:1076e84.

[22] Danese PN. Antibiofilm approaches: prevention of catheter colonization.Chem Biol 2002;9:873e80.

[23] Maitz MF, Freudenberg U, Tsurkan MV, Fischer M, Beyrich T, Werner C. Bio-responsive polymer hydrogels homeostatically regulate blood coagulation.Nat Commun 2013;4:2168.

[24] Chopra I. The increasing use of silver-based products as antimicrobial agents:a useful development or a cause for concern? J Antimicrob Chemother2007;59:587e90.

[25] Ip M, Lui SL, Chau SSL, Lung I, Burd A. The prevalence of resistance to silver ina burns unit. J Hosp Infect 2006;63:342e4.

[26] Streller U, Sperling C, Hübner J, Hanke R, Werner C. Design and evaluation ofnovel blood incubation systems for in vitro hemocompatibility assessment ofplanar solid surfaces. J Biomed Mater Res 2003;66B:379e90.

[27] Yeo EL, Sheppard JA, Feuerstein IA. Role of P-selectin and leukocyte activationin polymorphonuclear cell adhesion to surface adherent activated plateletsunder physiologic shear conditions (an injury vessel wall model). Blood1994;83:2498e507.

[28] Martínez-Gutierrez F, Thi EP, Silverman JM, de Oliveira CC, Svensson SL,Hoek AV, et al. Antibacterial activity, inflammatory response, coagulation andcytotoxicity effects of silver nanoparticles. Nanomed Nanotechnol Biol Med2012;8:328e36.

[29] Pidard D, Si-Tahar M, Chignard M. Platelets, thrombosis and the vessel wall.In: Berndt C, editor. Neutrophil-platelet interactions in medicine. HarwoodAcademic Publishers Amsterdam; 2000. p. 189e207.

[30] McEver RP. In: Berndt C, editor. Platelets, thrombosis and the vessel wall.Harwood Academic Publishers Amsterdam; 2000. p. 209e30.

[31] Kang JL, Moon C, Lee HS, Lee HW, Park EM, Kim HS, et al. Comparison of thebiological activity between ultrafine and fine titanium dioxide particles inRAW 264.7 cells associated with oxidative stress. J Toxicol Environ Health PartA 2008;71:478e85.

[32] Hosseinpour-Mashkani SM, Ramezani M. Silver and silver oxide nano-particles: synthesis and characterization by thermal decomposition. MaterLett 2014;130:259e62.

[33] Guillaume O, Garric X, Lavigne J-P, Van Den Berghe H, Coudane J. Multilayer,degradable coating as a carrier for the sustained release of antibiotics: prep-aration and antimicrobial efficacy in vitro. J Control Release 2012;162:492e501.

[34] Delig€oz H, Tieke B. QCM-D study of layer-by-layer assembly of polyelectrolyteblend films and their drug loading-release behavior. Colloid Surf A: Phys-icochem Eng Asp 2014;441:725e36.

[35] Choi J, Konno T, Takai M, Ishihara K. Regulation of cell proliferation by multi-layered phospholipid polymer hydrogel coatings through controlled release ofpaclitaxel. Biomaterials 2012;33:954e61.

[36] Salama A, El-Sakhawy M. Preparation of polyelectrolyte/calcium phosphatehybrids for drug delivery application. Carbohydr Polym 2014;113:500e6.