Acta Biomaterialia 83 (2019) 245–256

Contents lists available at ScienceDirect

Acta Biomaterialia

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

Full length article

Antibiofilm elastin-like polypeptide coatings: functionality, stability, andselectivity

https://doi.org/10.1016/j.actbio.2018.10.0391742-7061/� 2018 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: martin.andersson@chalmers.se (M. Andersson).

Saba Atefyekta a, Maria Pihl a, Chris Lindsay b, Sarah C. Heilshorn b, Martin Andersson a,⇑aDepartment of Chemistry and Chemical Engineering, Chalmers University of Technology, Gothenburg, SwedenbDepartment of Materials Science and Engineering, Stanford University, Stanford, CA, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 20 June 2018Received in revised form 22 October 2018Accepted 24 October 2018Available online 27 October 2018

Keywords:Antimicrobial peptideElastinRGDBacteriaMedical device

Antimicrobial peptides (AMPs) are currently receiving interest as an alternative to conventional antibi-otics to treat biomaterial-associated infection. However, the inherent instability of such peptides oftenlimits their efficacy in intended clinical applications. Covalent immobilization of AMPs to surfaces isone strategy to increase the long-term stability and minimize the toxicity. In this work, an antimicrobialpeptide, RRPRPRPRPWWWW-NH2 (RRP9W4N), was used to modify elastin-like polypeptide (ELP) surfacecoatings containing cell-adhesive peptide domains (RGD) using covalent chemistry. The AMP retained itsantibacterial activity against Staphylococcus epidermidis, Staphylococcus aureus, and Pseudomonas aerugi-nosa when covalently bonded to ELP surfaces. Simultaneously, the AMP functionalization had insignifi-cant effect on the viability, function, and differentiation of human osteosarcoma MG63 cells andhuman mesenchymal stem cells (hMSCs). Furthermore, stability of the immobilized AMP in human bloodserum was investigated, and the results suggested that the AMP preserved its antibacterial activity up to24 h. Combined, the results show that covalently attached AMPs onto RGD-containing ELP are an excel-lent candidate as an antimicrobial coating for medical devices.

Statement of Significance

Biomaterial associated infection, caused by adherent biofilm, is usually difficult to treat. There is a highdemand for new materials and treatments to decrease the infection rates, especially with increasingthreats concerning resistant bacteria. Formation of biofilms on medical devices lowers the bacteria sus-ceptibility towards traditional antibiotics and also circumvent our immune system often resulting in revi-sional surgery and extensive use of antibiotics. One promising strategy is to develop surfaces having lowbacterial attractiveness or bacterial killing properties, but still retaining the main function of the device.In this study, we have developed an implant coating that demonstrates a high antimicrobial effect and atthe same time showing no negative affect on human cells.

� 2018 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

As different classes of biomaterials with variable forms andfunctions are introduced into the healthcare system, infectionsassociated with such medical devices remain an unsolved problem.Biomaterial associated infection (BAI) is often caused by the devel-opment of an adherent bacterial biofilm on the surface of biomate-rials [1–3]. Formation of such biofilms is believed to be the originof chronic infections, which are difficult to diagnose and treat bymeans of conventional antibiotic treatments [4,5]. Moreover,

current antimicrobial therapy routines for treating infections haveresulted in a rapid increase of drug-resistant infections [6,7]. Thedevelopment of biomaterial surface coatings and modifications toprevent initial bacteria attachment is a promising strategy to pre-vent biofilm formation [8].

Antimicrobial peptides (AMPs) are currently receiving consider-able attention as a potential alternative to conventional antibioticsto prevent and treat BAI [9–13]. AMPs are part of the innateimmune system protecting us from invading pathogens with broadspectrum activity against many microbes ranging from viruses toparasites [12,14,15]. AMPs are often short peptides of <50 aminoacids with positive net charges of +2 to +9, and are amphiphilic,containing more than 30% hydrophobic amino acids [9]. The

246 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

mechanism of action for most AMPs is their disruption of the bac-terial membrane, resulting in rapid membrane rupture and bacte-rial death. There are AMPs that also act through other mechanismsincluding intracellular methods, such as inhibiting enzyme activity[16]. AMPs are also thought to be less susceptible to evolvedantibacterial resistance because of their fast mechanism of action[10,12,17,18].

AMPs, like many other peptide-based drugs, show low bioavail-ability due to rapid protease digestion and peptide aggregation[19–21]. Most AMPs show a half-life ranging from a few minutesup to a few hours under physiological electrolyte conditions inplasma or serum [22–24]. Covalent immobilization of AMPs ontosurfaces is one suggested strategy to increase their long-term sta-bility compared to release-based approaches [25–27]. Severalstudies have investigated the effects of different immobilizationstrategies on the antimicrobial activity of modified surfaces [28–32]. For example, the AMP Magainin was shown to retain antimi-crobial activity against several gram positive and gram negativebacteria when immobilized onto a polyamide resin [33]. Similarly,the AMP Melimine demonstrated good antimicrobial activity whencovalently bonded to commercial contact lenses [34].

For orthopedic implants, where the biomaterial is expected tointegrate with the host tissue, the selectivity of the coating to pre-vent microbial infection while simultaneously being permissive forhost cell adhesion is critical. We hypothesized that such surfaceselectivity could be achieved by combining mammalian cell-adhesive bioactive coatings with surface-tethered AMPs. Specifi-cally, we have developed a coating based on covalent immobiliza-tion of RRPRPRPRPWWWW-NH2 (RRP9W4N) to provideantimicrobial properties combined with a photo-crosslinkedelastin-like polypeptide (ELP) containing the cell-adhesive peptidemotif: arginine, glycine, aspartic acid (RGD), to promote mam-malian cell adhesion. ELP is a biocompatible, biodegradable mate-rial that has been explored in a number of different biomedicalsystems [35–37]. Photo-crosslinkable ELPs with fibronectin-derived, extended RGD sequences have been shown to retain theircell-adhesive bioactivity after crosslinking with UV light[36,38,39]. When used as thin films, it has proven to promote rapidadhesion of osteoblasts and stem cells in vitro and to significantlyincrease the bone-implant contact area in vivo [38]. RRP9W4N wascovalently attached on the ELP films using 1-ethyl-3-(-3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) chemistry. The anti-biofilm activitywas investigated using S. epidermidis, S. aureus, and P. aeruginosaassays, as evaluated by fluorescent microscopy and scanning elec-tron microscopy (SEM). The chemical stability of the attached AMPand its sustained anti-biofilm activity in human blood serum wereinvestigated by quartz crystal microbalance with dissipation mon-itoring (QCM-D) and bacterial assays. The effect of the AMP modi-fication on the viability and differentiation of hMSCs was assessedusing mineralization and alkaline phosphatase tests.

2. Materials and methods

2.1. Expression and purification of elastin-like polypeptide (ELP)

ELP containing RGD sequences was recombinantly expressed inan Escherichia coli host, as previously described [38]. Briefly, a plas-mid encoding the elastin-like and fibronectin-derived, extendedRGD sequences was transformed into an E. coli host, and isopropylb-D-1 thiogalactopyranocide (Sigma Aldrich Co. St Louis, MO, USA)was added to induce expression with the T7-lac promoter. Afterculturing, the bacteria were lysed, and the protein was purifiedusing repetitive centrifuge cycles above and below the lower crit-ical solution temperature of ELP (4 �C and 37 �C, respectively). The

purified ELP was dialyzed against deionized (DI) water, freezedried, and stored at 4 �C.

2.2. Photo crosslinked ELP thin film preparation

Heterobifunctional N-hydroxysuccidimide ester-diazirinecrosslinker (2 g, NHS-diazirine, succinimidyl 4,40-azipentanoate,Pierce Biotechnology, USA) was dissolved in 1 ml dimethyl sulfox-ide and mixed with a solution of ELP in phosphate buffered saline(PBS, pH 7.4, 2 mg/ml). The reaction was incubated for 2 h, and 1 Mtris buffer (pH 7.0) was added to the solution to stop the reaction.The diazirine-conjugated ELP was dialyzed against DI water,frozen, and lypholized, as previously reported [38,39]. AllELP-coated surfaces mentioned subsequently refer to thisphoto-crosslinkable form of ELP. Both glass coverslips and polishedtitanium (Ti) alloy discs were used as substrates. The glass cover-slips (VWR, 15 mm) were cleaned using basic piranha (3:1 mixtureif ammonium hydroxide with hydrogen peroxide). The Ti alloywere punched into 12 mm discs from a Ti6Al4V sheet(1000 mm � 600 mm � 1.6 mm, Titanium Industries, Inc., USA).They were polished with 400, 600, and 1000 grit sandpaper. ForXPS studies, pure Titanium discs were used. Polished discs werecleaned by sonication in successive 1% Triton X-100 in DI water,acetone, and 70% ethanol solutions before nitrogen drying andstorage. (A 50 mg/ml (5 wt%) solution of ELP in PBS (pH 7.4) wasprepared at 4 �C. The deposition of ELP onto the substrates wasperformed by spin coating. For the glass substrates, 50 ll of theELP solution and for the Ti alloy discs 14 ll of the ELP solutionwas applied to the center of the surface and spun at 4000 rpmfor 90 s. The spin coated substrates were treated by UV-light usinga 365 nm, 8 W light source for 1 h to induce ELP crosslinking. Thecrosslinked films were rinsed 3 times for 0.5 h with PBS (pH 7.4)prior to further experiments to remove any unbound ELP andcrosslinker.

2.3. AMP immobilization on ELP surfaces

A solution of antimicrobial peptide (AMP) RRPRPRPRPWWWW-NH2 (RRP9W4N, Red Glead Discovery AB, Lund, Sweden) was pre-pared in sterilized water to a concentration of 200 lM. For covalentattachment of AMP to the ELP surfaces, prior to AMP modification,the ELP-coated substrates were submerged into a solution of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride(EDC) and N-hydroxysuccinimide (NHS) mixed in MES buffer (pH6, Sigma Aldrich Co. St Louis, MO, USA) at a final concentration of2 mg/ml and were allowed to react for 30 min on slow shake atroom temperature. Surfaces were then washed 3 times in PBS(pH 7.4) and incubated in 1 ml of 200 lM AMP solution in steril-ized water for 2 h in room temperature. The surfaces were washed3 times with sterilized water to remove unreacted peptides andused for bacterial and human cell assays. The same procedurewas used to prepare surfaces on Ti-coated quartz crystal microbal-ance with dissipation monitoring (QCM-D) disks (QSX 310,Q-sense, Sweden) for QCM-D analysis and Ti alloy discs forscanning electron microscopy (SEM) analysis.

2.4. Bacteria culture and growth

S. epidermidis (ATCC 35984), S. aureus (CCUG 56489), and P.aeruginosa (CCUG 10778) were used to assess the biofilm forma-tion on the surfaces. One day prior to the experiments, a sterilized10 ml loop was used to withdraw a single colony from cultured agarplates of each bacterium to inoculate a tube of 5 ml tryptic soybroth (TSB). The inoculated cells were cultured in the incubatorfor 6 h at 37 �C, diluted in TSB, and cultured in an incubator over-night to reach the stationary phase for bacterial growth.

S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256 247

The optical density of the bacteria culture was adjusted to 0.7 at600 nm (estimated to give 108 cfu/mL) using a spectrometer. Thebacterial suspension was centrifuged for 10 min at 2500 rpm andthe bacteria pellet was suspended in fresh TSB medium. 2 ml ofthe suspension was seeded onto glass substrates, ELP-coated sub-strates, activated ELP substrates, and AMP-functionalized ELP sub-strates in a 12 well plate. Bacteria were then cultured for 24 h and48 h under standard culture conditions (ambient air at 37 �C) topromote biofilm formation on the surfaces. For the 48 h time point,the media was aspirated and replaced with fresh TSB after 24 h ofculture. At the end of each time point, surfaces were rinsed 3 timeswith fresh PBS to wash off the unattached planktonic bacteriabefore biofilm analysis.

2.5. Evaluation of AMP immobilization on ELP surfaces

2.5.1. Quartz crystal microbalance with dissipation monitoring(QCM-D)

The adsorption and tethered stability of AMP on EDC-NHS acti-vated surfaces were studied by QCM-D (Q-sense E4, Sweden). TiQCM-D discs (QSX 310, Q-sense, Sweden) were coated with ELPand activated with EDC and NHS as described earlier. Pure Ti discsand non-activated, ELP-coated Ti discs were used as controls. DIH2O was used as the solvent (200 lM AMP solution) and rinsingliquid. The mass of adsorbed AMP (ng/cm2) was calculated fromthe recorded frequency shifts using the Sauerbrey equation,

Dm ¼ �Df � Cn

ð1Þ

where Dm is the adsorbed mass, Df the frequency difference, Cis the mass sensitivity constant (17.7 ng Hz�1 cm�2), and n theovertone number.

2.5.2. Stability of AMP in human plasma serumFluorescently tagged AMP (5(6) carboxyfluorescein- RRP9W4N,

Red Glead Discovery AB, Lund, Sweden) was used to study the sta-bility of the AMP attachment and its distribution on surfaces overtime. The tests were performed by incubating the fluorescentlylabeled AMP functionalized surfaces in 20% human plasma serum(sterilized filter serum from human male AB, Sigma Aldrich Co.St Louis, MO, USA) (from 1 h up to 3 weeks) and imaged througha fluorescent microscope. Diluted human serum was used to beable to compare the serum stability with similar antimicrobial pep-tide analogs [21]. Bacterial assays using S. epidermidis were per-formed to investigate for how long the attached AMPs retainedtheir anti-biofilm activity.

2.6. Cell culture and growth

2.6.1. hMSC cell cultureNormal, human, bone marrow-derived mesenchymal stem cells

(ATCC� PCS-500-012TM) were received frozen, thawed, and sub-cultured according to supplier’s recommendations. Briefly, cellswere cultured and expanded in Dulbecco’s Modified Eagle Medium(DMEM, Invitrogen, USA) containing GlutaMAXTM, 4.5 g/L glucose,110 mg/mL sodium pyruvate, 10% fetal bovine serum, and 1% peni-cillin/streptomycin, at 37 �C and 5% atmospheric CO2. Cell studiesalso included the following osteogenic supplements: 100 nMdexamethasone, 50 lg/mL ascorbic acid, and 10 mM b-glycerophosphate. Cells were rinsed in DPBS before passagingusing 0.05% trypsin-EDTA. Cells were seeded onto 12-mm diameterTi substrates in 24 well plates at a density of 10,000 cells/disc.

2.6.2. Mineralization assaySamples were prepared and seeded as previously described.

Samples were rinsed twice in PBS (pH 7.4) and moved to fresh

wells. 1 ml of HCl (0.5 M) was added to each sample and placedon a rocker overnight at room temperature. Calcium content wasdetermined using the Calcium o-Cresolphthalein Complexone(CPC) Liquicolor Test (Stanbio Laboratory, USA). The HCl washingsolution was mixed with the working reagent and allowed to reactfor 10 min at room temperature before the absorbance was mea-sured at 550 nm. Calcium content was quantified by correlatingto the calcium standard curve and divided by the substrate areato convert to a calcium surface density.

2.6.3. Alkaline phosphatase assayCellular alkaline phosphatase (ALP) activity was quantified

using the SIGMAFAST p-nitrophenyl phosphate tablet kit (p-NPPkit, Sigma Aldrich Co. St Louis, MO, USA), and normalized to totalDNA content using the PicoGreen assay kit (Quant-iT PicoGreen,Molecular Probes). Samples were prepared and seeded with cellsas stated above. Samples were rinsed with PBS (pH 7.4), movedto clean wells, and lysed in a buffer containing 10 mM Tris (pH8), 1 mM MgCl2, 20 lM ZnCl2 and 0.02% Triton X-100 in deionizedH2O. Samples were frozen at �80 �C, thawed, and assayed. Double-stranded DNA was quantified from the samples according to theinstructions provided for the PicoGreen assay kit by measuringsample fluorescence and comparing to a known DNA standardcurve. ALP activity was assayed from the same lysate used forthe DNA quantification. Lysate was mixed with a p-NPP solution(5 mM), while an ALP solution was mixed with serial dilutions ofthis p-NPP solution. Samples were incubated for 1 h in a dark atroom temperature. A NaOH (3 M) stop solution was added to eachof the wells before measuring the absorbance at 405 nm and corre-lated with the standard curve. ALP activity reported here has beennormalized to total DNA content for each sample.

2.7. Material characterization

2.7.1. Scanning electron microscopy (SEM)The SEM (Leo Ultra 55, Carl Zeiss Meditec AG, Oberkochen, Ger-

many) was operated at 5 kV to analyze the morphology of the ELPthin films. ELP films were deposited on glass microscope slides fol-lowed by 1 min of gold sputtering at a rate of 3 nm/min and aplasma current of 10 mA and on Ti discs without any additionalgold sputtering.

The morphology of the bacteria in the biofilms was also imagedby SEM. After biofilm formation onto surfaces with and withoutAMP modification, the substrates were washed 3 times with PBS(pH 7.4) and kept in formaldehyde for 2 h for fixation. The sub-strates were then washed with PBS (pH 7.4) and treated in a seriesof ethanol concentrations in DI water (50%, 70%, 90%, and two 100%ethanol rinses for 15 min each) for dehydration. The samples weretransferred to a fresh solution of hexamethyldisilizane (HDMS) andleft covered in a fume hood overnight to completely dry. The driedsamples were sputter coated with gold for 1 min before imaging.

2.7.2. Fluorescent microscopyThe population of live and dead bacteria on biofilms formed on

ELP substrates with and without AMP functionalization wasobtained by fluorescent microscopy (Carl Zeiss GmbH, Jena-Germany equipped with a HBO 100 illuminating system). The bio-films were stained using LIVE/DEAD�BacLightTM Bacterial ViabilityKit 7007 (Molecular probes, Invitrogen). Specifically, live cellsappear green due to SYTO� 9 staining, while dead cells appearred due to propidium iodide nucleic acid staining.

Mammalian cell viability was tested on samples that were pre-pared and seeded with cells as reported above using LIVE/DEAD(Molecular Probes) staining. Adherent cells on substrates wererinsed twice with PBS and stained with 2 lM calcein AM and4 lM ethidium homodimer for 20 min. The staining solution was

248 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

rinsed with PBS, and samples were imaged using fluorescentmicroscopy (Zeiss AxioImager). Calcein (Ex/Em 494/517 nm) andethidium homodimer (Ex/Em 528/617 nm) were imaged usingstandard FITC and Texas-Red fluorescent filters, respectively.

2.7.3. Statistical analysisThe numerical data for live/dead bacteria were obtained by cal-

culating the red and green fractions of fluorescent microscopyimages and are presented by their mean values and standard devi-ations. All the experiments were conducted three times with tworeplicates. 20 images from each surface were used to obtain theimage analysis data. Paired Student’s t-test was performed for dataanalysis.

The data from ALP activity and Ca deposition assays are pre-sented by their mean and standard error of the mean, averaged

Fig. 1. a, Proposed schematic model for conjugation of ELP onto Ti substrates u

Fig. 2. QCM-D recordings showing the amount of D2O when exchanged with ultrap

over 3 replicates. Statistical analysis was performed using one-way ANOVA (OriginPro, OriginLab) followed by Tukey’s multiplecomparisons test.

A p-value <0.05 was considered significant for all the tests.

2.7.4. Swelling of ELPThe swelling of ELP thin films was calculated from the mea-

sured mass uptake of water on the ELP-coated QCM-D discs by sol-vent exchange method. This method is used to determine thewater content of the films [40]. A baseline was obtained by a stableflow of ultrapure water (Milli-Q) and replaced with D2O. The D2Oreplaces the absorbed water in the film, and a sudden shift in fre-quency occurs. This shift is due to the difference in density and vis-cosity of D2O and H2O. As a control, the test was performedsimultaneously on a pure Ti disk with no ELP coating. The fre-

sing diarizine moiety. b, SEM of photocrosslinked ELP onto a Ti substrate.

ure water absorbed into ELP thin films. Pure Ti surfaces were used as control.

Fig. 3. Schematic for EDC-NHS activation and AMP attachment on ELP.

Table 1The chemical composition in atomic percentage (%) of the ELP thin films coated on Tidiscs before and after AMP conjugation by EDC-NHS coupling as measured by XPS.

C1s N1s O1s Ti2PTitanium disk 0.75 0.29 55.80 43.16ELP coated Ti 64.20 15.71 20.09 –ELP coated Ti + AMP 51.98 16.88 31.14

S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256 249

quency shift obtained from the blank Ti surface was subtractedfrom the one obtained on an ELP surface. The final Df was corre-lated to the absorbed mass of D2O on the ELP films using the Sauer-brey equation (Eq. (1)). The thickness of the ELP thin films afterwater uptake was calculated and compared to the thickness ofthe films in the dry state as measured by SEM.

2.7.5. X-ray photoelectron spectroscopy (XPS)Chemical analysis of the ELP films before and after EDC-NHS

activation and antimicrobial peptide attachment was performedusing XPS. The chemical analysis was performed to monitor themodifications and to investigate if they introduced any unwantedimpurities to the films coated on pure Ti discs. The equipment usedwas a Quantum 2000 scanning microscope (Physical Electronics,

Fig. 4. Images of biofilms formed on glass and ELP thin films with and without AMP cepidermidis b, S. aureus and c, P. aeruginosa. (For interpretation of the references to colo

Inc., Chanhassen, MN, USA) with a 100 mm point diameter at a5-nm analysis depth.

3. Results and discussion

3.1. Formation and characterization of ELP coatings

A promising approach to improve the integrative properties ofbiomaterials is to deposit bioactive surface coatings. ELP coatingshave previously been shown to be stable and strong enough toendure clinically relevant implantation forces, to be favorable forosteogenic cell interactions, and to improve osseointegration [38].

We synthesized and purified photo-crosslinked ELP with cell-adhesive RGD sequences in the form of thin films as a substratefor antimicrobial modification by immobilization of antimicrobialpeptides. The crosslinking was performed by conjugation of diari-zine to the ELP, which was activated by UV exposure after spincoating. A reactive carbene intermediate is generated upon UVexposure to covalently bond the ELP to the titanium surface toinduce covalent crosslinking between adjacent ELP molecules toform a thin-film coating (Fig. 1a) [38,41]. SEM images obtainedfrom dried ELP coatings demonstrate a uniform coating on themicroscale with porous structures on the nanoscale (Fig. 1B). The

onjugation. The live bacteria appear green, and the dead bacteria appear red. a, S.ur in this figure legend, the reader is referred to the web version of this article.)

250 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

morphology of the crosslinked films did not change after extensivewashing under physiological conditions (PBS, 37 �C), indicating arobust ELP coating.

Upon immersion in water, covalently crosslinked ELP films areknown to form swollen hydrogels. D2O/H2O exchange experimentsusing QCM-D were performed to measure the water content of thehydrated ELP thin films. As described earlier, the films were coatedon QCM-D crystals, and the mass of absorbed D2O was monitoreduntil saturation (Fig. 2). Calculations were performed to measurethe thickness of wet film knowing the density difference betweenD2O and H2O. It was shown that the ELP thin films can swell �17times in aqueous medium.

3.2. Antimicrobial peptide modification and bacterial response

Covalent immobilization of AMPs is suggested to be an effectivemethod to enhance their long-term stability, making it feasible tobring AMPs into clinical applications. In this work, RRP9W4N wasimmobilized on the ELP substrates. RRP9W4N has previously beenshown to provide antibacterial activity against several antibiotic-resistant strains, including vancomycin-resistant enterococci,multi-drug-resistant P. aeruginosa, and methicillin-resistant S. aur-eus [42].While the presence of arginine-proline, RP, sequences pro-vides a highly positively charged AMP, the hydrophobicmodification of the peptide by tryptophan, W, end-tagging maxi-mizes the peptide interaction with the phospholipid membraneand facilitates its rupture [42–46]. The presence of the large sizedtryptophan, W, in the sequence is thought to limit the toxicity toeukaryotic cells. The cholesterol-rich membrane of eukaryotic cellscan hinder the insertion of tryptophan groups into the membrane[16]. Moreover, it has been shown that adsorption of these pep-tides onto the zwitterionic membranes of mammalian cells ismuch lower compared to adsorption on the more negativelycharged membranes of bacteria [42,43].

Fig. 5. Quantification of the relative substrate surface area covered with live or dead biofibiofilm area was significant (p < 0.05), both for dead and alive cells, for the ELP + AMP cofor both 24 and 48 h.

EDC-NHS coupling chemistry was used to covalently attach theantimicrobial peptide onto ELP surfaces (Fig. 3). In this method,EDC activates the present carboxyl groups on ELP for direct reac-tion with primary amine groups present in the AMP structure.NHS is used together with EDC to improve efficiency and createdry-stable (amine-reactive) intermediates. XPS analysis of surfacesshowed that there were no impurities introduced into the film dur-ing ELP spin coating, ELP crosslinking, and AMP conjugation(Table 1).

After immobilization of RRP9W4N, the coatings were found tohave a high anti-biofilm activity against S. epidermidis, S. aureus,and P. aeruginosa (Fig. 4). Fluorescent microscopy images of bio-films with live/dead staining clearly demonstrate a high anti-biofilm effect on the surfaces modified with AMP compared toELP coated surfaces before and after EDC-NHS activation. The acti-vated surfaces were included as a control to investigate if theyalone had an effect on the bacteria. Quantitative image analysisconfirmed that bacteria on AMP-modified surfaces were signifi-cantly more likely to be dead than alive after both 24 and 48 h(Fig. 5). An increase in the number of dead cells at 48 h comparedto 24 h suggest that the tethered AMP was still active to inducekilling of cells during 24–48 h. In contrast, all non-AMP-conjugated controls had significantly more living bacteria com-pared to dead bacteria. Additionally, when comparing the totalarea of the biofilm, AMP-modified surfaces trended towards lessbiofilm surface coverage.

The most commonly proposed bactericidal mechanism forAMPs involves physical interactions of the peptides with the bac-terium membrane. Electrostatic forces combined with hydropho-bic interactions may create pores in the membrane, cause cellleakage, and eventually promote bacterial cell lysis [12,18,47–49]. In addition, tethered AMPs may give rise to an alternativemechanism of antimicrobial activity, where cationic substrates,such as bound AMPs, create a high local charge density that causes

lms on glass and ELP surfaces with and without AMP attachment. The difference inmpare to all other surfaces. The difference was observed for all types of bacteria and

S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256 251

ionic displacement in the bacterial membrane resulting in celldeath [19,50,51]. Since the used AMP is relatively short and theEDC-NHS coupling does not introduce any significant spacerlength, the potential depth of penetration into the bacterial mem-brane would be limited. It has previously been reported that outer

Fig. 6. SEM images of bacteria taken after 24 h a, S. epidermidis, b, S. aureus and c, P. aerugbottom row are micrographs of the bacteria cultured on AMP-functionalized ELP surfac

membrane interactions with tethered AMPs coupled to short spac-ers (up to 6 carbons) showed a significant bacterial mortality [33].Consistent with these earlier studies, bacterial adhesion on ELPsurfaces with AMP conjugation was found to result in fewer adher-ent cells and an altered cellular morphology, as evaluated by SEM

inosa. The top row are micrographs of biofilms formed on bare ELP surfaces and thees.

252 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

(Fig. 6). A clear difference in the shape and number of the bacteriacan be seen when they have been subjected to AMP-functionalizedsurfaces. These results in combination with fluorescent microscopywith live/dead staining suggest that the AMP has induced high bac-terial killing effect while attached to ELP coatings.

3.3. AMP stability on ELP coatings

QCM-D measurements were performed to investigate theamount and stability of covalently bonded, antimicrobial peptideonto ELP, and to compare it with purely physically adsorbedAMP. It was shown that a higher content of antimicrobial peptide(�200 ng/cm2

, calculated using Eq. (1)) was attached to the EDC-

Fig. 7. QCM-D results obtained from loading AMP onto ELP surfaces followed by washpresented.

Fig. 8. Fluorescent microscopy images of 5(6) carboxyfluorescein tagged RRP9W4N impoints. Ultrapure water treated surfaces was used as control.

NHS activated ELP surfaces compared to non-activated surfaces(Fig. 7). Upon rinsing with ultrapure water, most of the AMP stillremained on the activated ELP surface, while AMPs were completelywashed from the other surfaces after rinsing. These results confirmthe stable attachment of AMP onto ELP surfaces via the EDC-NHScoupling approach.

A major obstacle to bringing peptide-based pharmaceuticalsinto the clinic are their inherently low bioavailability towardsdegradation [19]. There are several AMPs that exhibit significantbactericidal effects in vitro, but their activity is rapidly lost underphysiological salt and serum conditions [23]. Such low stabilityoften necessitates using high concentration of AMPs which canlead to systemic toxicity in vivo [19]. We hypothesized that stable

ing with ultrapure water. ELP surfaces with and without EDC-NHS activation are

mobilized onto ELP surfaces treated in 20% human blood serum at different time

S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256 253

immobilization of AMPs onto substrates, compared to incorpora-tion followed by release approaches, can overcome the cytotoxicityassociated with the use of higher concentrations of soluble peptideto compensate for their short half-life [19,33–35].

To assess the in vivo bioavailability of peptides, it has beenreported that in vitro stability in serum or plasma is a representa-tive model [21]. In the present study, AMP-modified ELP surfaceswere subjected to human blood serum. Fluorescently labeled,immobilized peptides start to aggregate after 24 h when subjectedto blood serum, while in water no differences were observed up to3 weeks (Fig. 8). The anti-bacterial assay on AMP-functionalizedsurfaces at the same time points showed that the immobilizedAMPs had a significant effect up to 24 h in serum and that theanti-biofilm activity was lost after 48 h (Fig. 9).

Two main causes have been suggested to explain the lost effi-cacy of AMPs in serum: increased ionic strength, which can leadto a change in the peptide conformation and loss of activity, andthe presence of serum enzymes, which may cause degradation ofthe AMP. Therefore, most AMPs show a half-life stability of min-utes up to a few hours when they are present in a serum-containing environment [21,22,24]. Cationic AMPs in particularshow very fast degradation due to cleavage sites introduced bytheir high arginine and lysine content [52]. In a previous study,the half-life of 1 h in 25% aqueous mouse serum for an arginine-rich, short AMP (NH2RRWRIVVIRVRR-CONH2) is reported [52].Serum stability profiles of short tryptophan- and arginine-richAMPs analogs in diluted serum media showed half-lives of<0.5 h–6.5 h [21]. For covalently bonded RRP9W4N on ELP incu-bated in human serum, the stability is retained up to 24 h, whichis a promising indication that covalent immobilization can increasethe stability in serum compared to release-based approaches. Thelost anti-bacterial effect could also be due to protein adsorptionoriginating from the serum; however, the fluorescence microscopydata, showing the disappearance of the labeled AMPs, links verywell to the bacterial assays, indicating that the protein adsorptionis of less importance at the measured time points.

Fig. 9. S. epidermidis response to AMP-functionalized ELP surfa

3.4. Mammalian cell responses to AMP-modified ELP surfaces

The viability of human bone marrow-derived mesenchymalstem cells (hMSCs) on ELP surfaces modified by AMP immobiliza-tion was studied by live/dead imaging using fluorescent micro-scopy. Interestingly, there were no visible changes in the viabilityof the cells adhered to ELP surfaces with or without AMP modifica-tion (Fig. 10). These data indicate that covalent immobilization ofAMPs onto ELP surfaces did not affect the cell-adhesivity of thesubstrates and showed no toxicity to the mammalian cells usedin this study. One argument for such selective behavior can be thatthe cell membrane of eukaryotic and prokaryotic cells differs sig-nificantly in composition [52]. The eukaryotic cell membrane ismainly composed of phospholipids such as cholesterol and othersterols that are neutrally charged. Such phospholipids are seldomfound in prokaryotic membranes. Instead, their membranes aredominantly composed of more hydroxylated phospholipids withnegative net charge. Therefore the overall cell membrane netcharge of bacteria cells is much higher than mammalian cells[52]. AMPs interact with microorganisms through electrostaticinteraction between their positively charged amino acids and neg-atively charged membrane of microorganisms. Thus, the differentcomposition and transmembrane potential of bacteria and mam-malian cell surfaces may explain the preferential binding of AMPsto bacteria cells and their selective behavior [53,54].

3.5. Effect of AMP modification on hMSC osteogenic differentiation

It has previously been shown that ELP coatings can promote sig-nificantly early mineralization on Ti surfaces [38]. To assess if AMPmodification can affect the early-stage osteogenic differentiation ofhMSCs on ELP-coated substrates, alkaline phosphatase (ALP) activ-ity was quantified. 200 mM AMP solution was used to prepare sub-strates to investigate if AMPs bonded onto surfaces affected theALP activity. We observed an increase in ALP activity betweenday 7 and day 14 for all samples with ELP coatings, regardless of

ces treated in 20% human serum at different time points.

Fig. 10. Live/Dead imaging of hMSCs cells on ELP surfaces coated on Ti discs with and without AMP modification. Cells are stained with calcein AM (Live stain, green) andethidium homodimer (Dead stain, red). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 11. a, ALP activity and b, total calcium deposition of hMSCs seeded onto bare Tiand ELP substrates treated in 0 and 200 lM AMP solution. ALP activity isnormalized to DNA concentration. Values are mean ± SEM, *p < 0.05.

254 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

whether they were unmodified or modified by AMPs (Fig. 11). Thisincrease in ALP activity was not observed on bare Ti substratesuntil significantly later time points (days 21 and 28). These resultsimply that the presence of AMPs on ELP surfaces does not interferewith the early onset of ELP-induced differentiation of hMSCs.

Calcium deposition, which is known to occur after ALP activa-tion, showed no statistically significant differences across all sur-faces tested. Both AMP-treated surfaces showed similar calciumdeposition to non-treated ELP and bare titanium surfaces. Giventhis observation, the mineralization of surfaces was not affectedby tethering of AMPs. The overall conclusion is that AMP modifica-tion does not interfere with the osseointegrative properties of ELP-coated surfaces over 28 days in vitro.

To date there have been few reports on in vitro osteogenicdifferentiation and in vivo bone growth onto antimicrobialpeptide-loaded implants. A relevant report has shown thatcovalent immobilization of AMP KR-12 on Ti surfaces can promoteosteogenic differentiation of bone cells [55]. In vivo release of AMPpeptide hLF1-11 from calcium phosphate cements has shown thatthey can be considered as a prophylactic agent for osteomyelitistreatment [56]. There are also studies showing that the presenceof some AMPs does not impair in vivo bone growth onto implants.For example AMP-loaded Ti dental implants did not affect the rateof in vivo implant osseointegration in rabbit femurs compared togold-standard, non-coated dental implants [57]. Similarly, it wasshown that loading of HHC36 into calcium phosphate cementsdid not impair in vivo bone growth onto the implants [58].Consistent with these previous studies, our results also suggestthat some AMPs can be well tolerated by mammalian osteogeniccells.

4. Conclusion

We have developed an antimicrobial surface coating intendedfor implants by combining a recombinant ELP coating with cell-adhesive RGD sequences with a covalently attached AMP,RRP9W4N. The coating showed selective behavior, killing bothgram-positive and gram-negative bacteria and at the same timeenabling mammalian osteogenic cell adhesion and mineral deposi-tion. Moreover, the stability of covalently attached AMP in humanserum was highly increased showing high activity up to 24 h. TheAMP functionalization approach presented in this paper introducesa versatile method to increase the life time and stability of AMPswhile retaining their antimicrobial performance.

S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256 255

Acknowledgements

We acknowledge the Wallenberg Foundation through their Wal-lenberg Academy Fellow Program for financial support.

We acknowledge Dr. Lydia-Marie Joubert and the Beckman CellSciences Imaging Facility for training and assistance with the SEMimaging. We acknowledge Dr. Cedric Espenel and the Shriram CellSciences Imaging Facility for training and assistance with the fluo-rescent microscopy imaging. We acknolwedge funding supportfrom the National Science Foundation DMR-1508006 (S.C.H.).

Appendix A. Supplementary data

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.actbio.2018.10.039.

References

[1] Zaborowska, K. Welch, R. Branemark, P. Khalipour, H. Engqvist, P. Thomsen, M.Trobos, Bacteria-material surface interactions: methodological developmentfor the assessment of implant surface induced antibacterial effects, J. Biomed.Mater. Res. Part B Appl. Biomater 103 (2015) 179–187.

[2] K. Bazaka, M.V. Jacob, R.J. Crawford, E.P. Ivanova, Efficient surface modificationof biomaterial to prevent biofilm formation and the attachment ofmicroorganisms, Appl. Microbiol. Biotechnol. 95 (2012) 299–311.

[3] D. Campoccia, L. Montanaro, C.R. Arciola, The significance of infection relatedto orthopedic devices and issues of antibiotic resistance, Biomaterials 27(2006) 2331–2339.

[4] C.R. Arciola, D. Campoccia, P. Speziale, L. Montanaro, J.W. Costerton, Biofilmformation in Staphylococcus implant infections. a review of molecularmechanisms and implications for biofilm-resistant materials, Biomaterials 33(2012) 5967–5982.

[5] A. Engelsman, F.I. Saldarriaga, M. Nejadnik, G. van Dam, K. Francis, R. Ploeg, H.Busscher, H. Van der Mei, The risk of biomaterial-associated infection afterrevision surgery due to an experimental primary implant infection, Biofouling26 (2010) 761–767.

[6] C. Faber, H. Stallmann, U. Joosten, E.C. Von, A. Amerongen, P. Wuisman,Comparable, efficacies of the antimicrobial peptide human lactoferrin 1-11and gentamicin in a chronic methicillin-resistant Staphylococcus aureusosteomyelitis model, Antimicrob. Agents Chemother. 49 (2005) 2438–2444.

[7] N. Woodford, D.M. Livermore, Infections caused by gram-positive bacteria: areview of the global challenge, J. Infect. 59 (2009) S4–S16.

[8] M.L.W. Knetsch, L.H. Koole, New strategies in the development of antimicrobialcoatings: the example of increasing usage of silver and silver nanoparticles,Polymers (Basel) 3 (2011) 340–366.

[9] R.E.W. Hancock, H.G. Sahl, Antimicrobial and host-defense peptides as newanti-infective therapeutic strategies, Nat. Biotechnol. 24 (2006) 1551–1557.

[10] O. Toke, Antimicrobial peptides: new candidates in the fight against bacterialinfections, Biopolym. Pept. Sci. Sect. 80 (2005) 717–735.

[11] H.W. Huang, Molecular mechanism of antimicrobial peptides: the origin ofcooperativity, Biochim. Biophys. Acta Biomembr. 1758 (2006) 1292–1302.

[12] M. Zasloff, Antimicrobial peptides of multicellular organisms, Nature 415(2002) 389–395.

[13] V. Nizet, Antimicrobial peptide resistance mechanisms of human bacterialpathogens, Curr. Issues Mol. Biol. 8 (2006) 11–26.

[14] K.L. Brown, R.E.W. Hancock, Cationic host defense (antimicrobial) peptides,Curr. Opin. Immunol. 18 (2006) 24–30.

[15] A. Mor, Antimicrobial peptides, Kirk-Othmer Encycl. Chem. Technol. 1–24(2001), https://doi.org/10.1016/j.cub.2015.11.017.

[16] A. Schmidtchen, M. Pasupuleti, M. Davoudi, J. Alenfall, A. Chalupka, M.Malmsten, Boosting antimicrobial peptides by hydrophobic oligopeptide endtags, J. Biol. Chem. 284 (2009) 17584–17594.

[17] N. Papo, Y. Shai, Can we predict biological activity of antimicrobial peptidesfrom their interactions with model phospholipid membranes?, Peptides 24(2003) 1693–1703

[18] M.R. Yeaman, N.Y. Yount, Mechanisms of antimicrobial peptide action andresistance, Pharmacol. Rev. 55 (2003) 27–55.

[19] F. Costa, I.F. Carvalho, R.C. Montelaro, P. Gomes, M.C.L. Martins, Covalentimmobilization of antimicrobial peptides (AMPs) onto biomaterial surfaces,Acta Biomater. 7 (2011) 1431–1440.

[20] J. Svenson, W. Stensen, B. Brandsdal, B. Haug, J. Monrad, J. Svendsen,Antimicrobial peptides with stability oward tryptic degradation,Biochemistry 47 (2008) 3777–3788.

[21] L. Nguyen, J. Chau, N. Perry, L. de Boer, S. Zaat, H. Vogel, Serum stabilities ofshort tryptophan- and arginine-rich antimicrobial peptide analogs, PLoS ONE 5(2010) 1–8.

[22] G. Maisetta, M. Di Luca, S. Esin, W. Florio, F. Brancatisano, D. Bottai, M. Campa,G. Batoni, Evaluation of the inhibitory effects of human serum components onbactericidal activity of human beta defensin 3, Peptides 29 (2008) 1–6.

[23] M. Werle, A. Bernkop-Schnürch, Strategies to improve plasma half life time ofpeptide and protein drugs, Amino Acids 30 (2006) 351–367.

[24] M.J. Goldman, G. Anderson, E. Stolzenberg, U. Kari, M. Zasloff, J. Wilson, Humanb-defensin-1 is a salt-sensitive antibiotic in lung that is inactivated in cysticfibrosis, Cell 88 (1997) 553–560.

[25] Y. Endo, T. Tani, M. Kodama, Antimicrobial activity of tertiary aminecovalently bonded to a polystyrene fiber, Appl. Environ. Microbiol. 53 (1987)2050–2055.

[26] M. Bagheri, M. Beyermann, M. Dathe, Immobilization reduces theactivity of surface-bound cationic antimicrobial peptides with no influenceupon the activity spectrum, Antimicrob. Agents Chemother. 53 (2009)1132–1141.

[27] M.K.S. Batista, M. Gallemi, A. Adeva, C.A.R. Gomes, P. Gomes, Facileregioselective synthesis of a novel chitosan-pexiganan conjugate withpotential interest for the treatment of infected skin lesions, Synth. Commun.39 (2009) 1228–1240.

[28] A. Peschel, R. Jack, M. Otto, L. Collins, P. Staubitz, G. Nicholson, H. Kalbacher, W.Nieuwenhuizen, G. Jung, A. Tarkowski, K. van Kessel, J. van Strijp,Staphylococcus aureus resistance to human defensins and evasion ofneutrophil killing via the novel virulence factor MprF is based onmodification of membrane lipids with l-lysine, J. Exp. Med. 193 (2001)1067–1076.

[29] D. Andreu, L. Rivas, Animal antimicrobial peptides: an overview, Inc Bipoly. 47(1998) 415–433.

[30] A. Gallardo-Moreno, M. Pacha-Olivenza, L. Saldaña, C. Pérez-Giraldo, J. Bruque,N. Vilaboa, M. González-Martín, In vitro biocompatibility and bacterialadhesion of physico-chemically modified Ti6Al4V surface by means of UVirradiation, Acta Biomater. 5 (2009) 181–192.

[31] A. Peschel, L. Vincent Collins, Staphylococcal resistance to antimicrobialpeptides of mammalian and bacterial origin, Peptides 22 (2001)1651–1659.

[32] B. Jose, V. Antoci, A.R. Zeiger, E. Wickstrom, N.J. Hickok, Vancomycin covalentlybonded to titanium beads kills Staphylococcus aureus, Chem. Biol. 12 (2005)1041–1048.

[33] S.L. Haynie, G.A. Crum, B.A. Doele, Antimicrobial activities of amphiphilicpeptides covalently bonded to a water-insoluble resin, Antimicrob. AgentsChemother. 39 (1995) 301–307.

[34] M.D.P. Willcox, E.B.H. Hume, Y. Aliwarga, N. Kumar, N. Cole, A novel cationic-peptide coating for the prevention of microbial colonization on contact lenses,J. Appl. Microbiol. 105 (2008) 1817–1825.

[35] K. Melissa, M.S. McHale, L.A. Setton, A. Chilkott, Synthesis and in Vitroevaluation of enzymatically cross-linked elastin-like polypeptide gels forcartilaginous tissue repair, Tissue Eng. 11 (2005) 1767–1779.

[36] S.L. Bellis, Advantages of RGD peptides for directing cell association withbiomaterials, Biomaterials 32 (2012) 4205–4210.

[37] L. Cai, S.C. Heilshorn, Designing ECM-mimetic materials using proteinengineering, Acta Biomater. 10 (2014) 1751–1760.

[38] J. Raphel et al., Engineered protein coatings to improve the osseointegration ofdental and orthopaedic implants, Biomaterials 83 (2016) 269–282.

[39] J. Raphel, A. Parisi-Amon, S. Heilshorn, Photoreactive elastin-like proteins foruse as versatile bioactive materials and surface coatings, J. Mater. Chem. 22(37) (2013) 19429–19437.

[40] J. Kittle, F. Jiang, C. Qian, T. Heinze, M. Roman, A. Esker, Equilibrium watercontents of cellulose films determined via solvent exchange and quartz crystalmicrobalance with dissipation monitoring, Biomacromolecules 12 (2011)2881–2887.

[41] A. Sinz, Chemical cross-linking and mass spectrometry to map three-dimensional protein structures and protein-protein interactions, MassSpectrom. Rev. 25 (2006) 663–682.

[42] M. Malmsten, G. Kasetty, M. Pasupuleti, J. Alenfall, A. Schmidtchen, Highlyselective end-tagged antimicrobial peptides derived from prelp, PLoS ONE 6(2011).

[43] M. Pasupuleti, A. Schmidtchen, A. Chalupka, L. Ringstad, M. Malmsten, End-tagging of ultra-short antimicrobial peptides by W/F stretches to facilitatebacterial killing, PLoS ONE 4 (2009).

[44] D.J. Schibli, P.M. Hwang, H.J. Vogel, The structure of the antimicrobial activecenter of lactoferricin B bound to sodium dodecyl sulfate micelles, FEBS Lett.446 (1999) 213–217.

[45] W. Jing, H.N. Hunter, J. Hagel, H.J. Vogel, The structure of the antimicrobialpeptide Ac-RRWWRF-NH2 bound to micelles and its interactions withphospholipid bilayers, J. Pept. Res. 61 (2003) 219–229.

[46] D.I. Chan, E.J. Prenner, H.J. Vogel, Tryptophan- and arginine-rich antimicrobialpeptides: structures and mechanisms of action, Biochim. Biophys. Acta –Biomembr. 1758 (2006) 1184–1202.

[47] J.P.S. Powers, R.E.W. Hancock, The relationship between peptide structure andantibacterial activity, Peptides 24 (2003) 1681–1691.

[48] H. Steiner, D. Andreu, R.B. Merrifield, Binding and action of cecropin andcecropin analogues: antibacterial peptides from insects, BBA – Biomembr. 939(1988) 260–266.

[49] Y. Pouny, D. Rapaport, A. Mor, P. Nicolas, Y. Shai, Interaction of antimicrobialdermaseptin and its fluorescently labeled analogues with phospholipidmembranes, Biochemistry 31 (1992) 12416–12423.

[50] K. Lienkamp, A.E. Madkour, K.N. Kumar, K. Nüsslein, G.N. Tew, Antimicrobialpolymers prepared by ring-opening metathesis polymerization: manipulatingantimicrobial properties by organic counterion and charge density variation,Chem. - A Eur. J. 15 (2009) 11715–11722.

256 S. Atefyekta et al. / Acta Biomaterialia 83 (2019) 245–256

[51] R. Kügler, O. Bouloussa, F. Rondelez, Evidence of a charge-density threshold foroptimum efficiency of biocidal cationic surfaces, Microbiology 151 (2005)1341–1348.

[52] D. Knappe, P. Henklein, R. Hoffmann, K. Hilpert, Easy strategy to protectantimicrobial peptides from fast degradation in serum, Antimicrob. AgentsChemother. 54 (2010) 4003–4005.

[53] F. Guilhelmelli, N. Vilela, P. Albuquerque, L. Derengowski, I. Silva-Pereira, C.Kyaw, Antibiotic development challenges: the various mechanisms of action ofantimicrobial peptides and of bacterial resistance, Front. Microbiol. 4 (2013)1–12.

[54] T. Ebenhan, O. Gheysens, H.G. Kruger, J.R. Zeevaart, M.M. Sathekge,Antimicrobial Peptides: their Role as Infection-Selective Tracers forMolecular Imaging, Biomed Res. Int. 2014 (2014).

[55] B.A. Nie, H. Chen, C. Xie, K. Zhou, J. Tang, B. Yue, Covalent immobilization ofKR-12 peptide onto titanium surface for decreasing infection and promotingosteogenic differentiation, Acta Biomater. 6 (2016) 46733–46743.

[56] H.P. Stallmann, R.De Roo, C. Faber, A.V.N. Amerongen, P.I.J.M. Wuisman, In vivorelease of the antimicrobial peptide hLF1-11 from calcium phosphate cement,J. Orthop. Res. 531–538 (2008).

[57] X.C.X.C. Zhou, S.L.R.F. Wu, C.A.J.Y. Wu, In vivo osseointegration of dentalimplants with an antimicrobial peptide coating, J. Mater. Sci. 28 (2017).

[58] M. Kazemzadeh-Narbat, S. Noordin, B. Masri, D. Garbuz, C. Duncan, R. Hancock,R. Wang, Drug release and bone growth studies of antimicrobial peptide-loaded calcium phosphate coating on titanium, J. Biomed. Mater. Res. – Part BAppl. Biomater. 100 (B) (2012) 1344–1352.