lable at ScienceDirect

Biomaterials 65 (2015) 22e31

Contents lists avai

Biomaterials

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

Zn/Ag micro-galvanic couples formed on titanium andosseointegration effects in the presence of S. aureus

Guodong Jin a, 1, Hui Qin b, 1, Huiliang Cao a, Yuqin Qiao a, Yaochao Zhao b,Xiaochun Peng b, Xianlong Zhang b, **, Xuanyong Liu a, *, Paul K. Chu c

a State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai200050, Chinab Department of Orthopedics, Shanghai Sixth People's Hospital, Shanghai Jiao Tong University, Shanghai 200233, Chinac Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China

a r t i c l e i n f o

Article history:Received 18 February 2015Received in revised form18 June 2015Accepted 22 June 2015Available online 23 June 2015

Keywords:TitaniumZincSilverMicro-galvanic coupleOsseointegrationAntibacterial ability

* Corresponding author. State Key Laboratory of HigSuperfine Microstructure, Shanghai Institute of CerSciences, 1295 Dingxi Road, Shanghai 200050, China.** Corresponding author. Department of OrthopedHospital, Shanghai Jiao Tong University, Shanghai 200

E-mail addresses: zhangxianl197826@163.com (X(X. Liu).

1 The authors contributed equally to this work.

http://dx.doi.org/10.1016/j.biomaterials.2015.06.0400142-9612/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Titanium implants possessing simultaneous osseointegration and antibacterial ability are desirable. Inthis work, three types of Zn/Ag micro-galvanic couples are fabricated on titanium by plasma immersionion implantation to investigate the osseointegration and antibacterial effects as well as the involvedmechanisms. The in vitro findings disclose enhanced proliferation, osteogenic differentiation, and geneexpressions of the rat bone mesenchymal stem cells (rBMSCs), as well as good antibacterial ability on allthree micro-galvanic couples. Excellent antimicrobial ability is also observed in vivo and the micro-CTand histological results reveal notable osseointegration in vivo despite the presence of bacteria. TheZn/Ag micro-galvanic couple formed on Zn/Ag dual-ion co-implanted titanium shows the bestosseointegration as well as good antibacterial properties in vivo obtained from a rabbit tibia model. Thedifference among the three Zn/Ag micro-galvanic couples can be ascribed to the contact between the AgNPs and Zn film, which affects the corrosion process. Our results indicate that the biological behavior canbe controlled by the corrosion process of the Zn/Ag micro-galvanic couples.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Titanium and its alloys have many clinical applications becauseof their favorable properties [1,2] but implant-associated infectionscaused by the adhesion and colonization of bacteria [3,4] andinadequate osseointegration may lead to osteolysis, implant loos-ening, and eventual failure [5]. Therefore, new titanium implantswith the desirable osseointegration and antibacterial ability arehighly desirable [6]. Our previous studies demonstrate that zinc(Zn) ion implantation into titanium can significantly enhance theosteogenic activity in vitro and stimulate bone growth in vivo [7e9],

h Performance Ceramics andamics, Chinese Academy of

ics, Shanghai Sixth People's233, China.. Zhang), xyliu@mail.sic.ac.cn

whereas silver (Ag) ion-implanted titanium exhibits excellentantibacterial ability both in vitro and in vivo [3,10,11]. Moreover, Zn/Ag co-implantation enhances the initial adhesion, proliferation,differentiation, and gene expressions of rat bone mesenchymalstem cells (rBMSCs) on titanium while excellent antimicrobialproperties are also observed both in vitro and in vivo [12]. Themicro-galvanic couples formed by the Ag nanoparticles (Ag NPs)and Ti substrate or Zn and Ag NPs are believed to be the key factorscontrolling the biological behavior of titanium. However, the effectsof Zn and Ag ions implantation process on the formation and bio-logical properties of Zn/Ag micro-galvanic couples still need to beinvestigated further.

As a non-line-of-sight technique, plasma immersion ion im-plantation (PIII) offers many advantages in modifying the surfaceproperties while preserving the favorable bulk attributesof biomaterials and biomedical implants [2,13e15]. The propersurface structural and compositional changes caused bybombardment of high-energy ions have been demonstrated toenhance the wear resistance, corrosion resistance, bioactivity andantimicrobial properties [16e18]. In recent years, sequential PIII

G. Jin et al. / Biomaterials 65 (2015) 22e31 23

involving two or more different metallic ions has attracted highattention because of the increased versatility with regard to thesurface treatment of knee joints, dental implants, artificial hips,and so on [19,20]. For instance, Krupa et al. [19] have demon-strated that sequential Ca and P ion implantation improves thebiocompatibility as well as corrosion resistance of titanium andXie et al. [20] have found that titanium after water and hydrogenPIII exhibits better surface bioactivity and cytocompatibility.Different from single-metal ion implantation, sequential im-plantation and co-implantation offer excellent advantages incontrolling the individual concentrations and depth distributionsof multiple metal ions in the near-surface of titanium [21].Therefore, it is possible to control the Zn/Ag micro-galvaniccouples by adjusting the implantation sequence of Zn andAg ions.

In this work, to investigate the effects of ion implantationsequence on the corrosion behavior of micro-galvanic couples,three newly designedmicro-galvanic couples were fabricated usingZn and Ag simultaneously and sequentially implanted into titaniumby PIII. The corrosion potentials and rates of the three micro-galvanic couples are measured. Although it has been reportedthat the biological actions can be controlled by the micro-galvaniceffects [18], the osseointegration and antibacterial properties needto be investigated systematically both in vitro and in vivo. Therefore,a new animal model was applied in the present study to evaluatethe osseointegration and antibacterial properties in vivo. Theobjective is to investigate the influence of the three differentmicro-galvanic couples on the osteogenic activity and antibacterial abilityof titanium both in vitro and in vivo and to further elucidate theunderlying mechanisms.

2. Methods

2.1. Zn and Ag PIII

Pure Ti plates with dimensions of 10 mm � 10 mm � 1 mmwere carefully polished by abrasive paper to a mirror finish, ul-trasonically cleaned with ethanol and ultrapure water. Before Znand Ag PIII, the specimens were cleaned by a radio-frequencyargon plasma for 15 min (sample bias is �550 V). Dual Zn andAg PIII were conducted using pulsed zinc and silver cathodic arcsources. The implantation voltage, voltage pulse duration, andpulsing frequency used in PIII were �30 kV, 500 ms, and 5 Hz,respectively. Zn and Ag ions were implanted into titanium for90 min each. Zn/Ag dual-ion co-implanted titanium, Zn prior toAg implanted titanium, and Ag prior to Zn implanted titaniumwere represented as Zn/Ag-PIII, Zn-Ag-PIII, and Ag-Zn-PIII,respectively.

2.2. Surface characterization

Field-emission scanning electron microscopy (FE-SEM) wasused to examine the surface topography of the specimens. Theelemental depth profiles and their chemical stateswere detected byX-ray photoelectron spectroscopy (XPS). Besides, the dynamic po-tential polarization plots were measured in 0.9% NaCl solution(pH ¼ 7) using a CHI760c electrochemical workstation (CHIInstruments).

2.3. Zn and Ag release

The titanium samples before and after Zn/Ag dual-ion implan-tation were immersed in 10 ml of 0.9% NaCl solution for 7, 14, 21,and 28 days at 37 �C. The solutions were refreshed every 7 days. Theconcentrations of released Zn and Ag were detected by inductively-

coupled plasma atomic emission spectroscopy (ICP-AES).

2.4. In vitro osteogenic activity

2.4.1. Cell proliferation and viabilityCell proliferation and viability was determined by ala-

marBlue™ method. The rBMSCs were seeded on the modifiedsurfaces at a density of 2.0 � 104 cells/ml. At each incubationperiod, the specimens (with the cells) were rinsed twice with PBSand 0.5 ml of fresh medium with 5% alamarBlue™ was added.After incubation for another 4 h, the absorbance values of 100 mlof medium at 570, and 600 nm were recorded. Calculation ofcell proliferation and viability followed the instruction of ala-marBlue™ assay.

2.4.2. ALP activity assayThe rBMSCs were seeded on the specimens (three replicates) at

a cell density of 1.0 � 104 cells/ml (for 7 days) or 0.5 � 104 cell/ml(for 14 days). After 7 and 14 days' incubation, a commerciallyavailable kit was employed for ALP staining. After culturing withp-nitrophenyl phosphate at 37 �C for 30 min, the ALP activity wasdetermined by measuring the optical density at a wavelength of405 nm for the quantitative assay. Finally, the ALP activity wasnormalized to total protein contents determined by a BCA proteinassay.

2.4.3. Collagen secretionCollagen secretion on the samples (three replicates) was eval-

uated by Sirius Red staining. The rBMSCs were seeded on the sur-faces at a density of 1.0 � 104 cells/ml (for 7 days) or 0.5� 104 cells/ml (for 14 days). On day 7 and 14, the cells werewashed thrice withPBS, fixed in 4% PFA for 20 min, rinsed with PBS three times, andstained with 0.1% Sirius Red for 18 h. The cells were washed with0.1 M acetic acid until no color appeared. In the quantitative anal-ysis, the stain was dissolved in 0.5 ml of the solution (0.2 M NaOH:methanol ¼ 1:1) and OD values were determined based on theabsorbance at 492 nm.

2.4.4. Matrix mineralizationExtracellular matrix (ECM) mineralization was investigated by

Alizarin Red. The rBMSCs (the cell density was the same as thatused in collagen secretion) were cultured on the various specimensfor 7 and 14 days, washed thrice with PBS, then fixed in 75% ethanolfor 1 h, finally they were stained with 40 mM Alizarin Red for10 min. To take the images, the cells were gently rinsed with ul-trapure water until red color disappeared. The stain was dissolvedin 10% cetylpyridinium chloride, and OD value at wavelength of600 nm was measured.

2.4.5. Quantitative real-time PCR assayReal-time PCR was employed to analysis the expressions of

osteogenesis-related genes. The rBMSCs were seeded on the spec-imens at densities of 1.0� 105 cell/well (for 7 days) or 0.5� 105 cell/well (for 14 days). The total RNAwas extracted using TRIzol reagentand the cDNA was reverse transcribed from 1.0 mg of the RNA. RT-PCR was conducted on the Roche LightCycler480 system using aSYBR Green I matermix. The relative expressions of osteogenesis-related genes (ALP, Col-I, Runx2, and OCN) were normalized tothat of the reference gene F-actin, the primers for RT-PCR are listedin Table S1.

2.5. In vitro antibacterial tests

The in vitro antimicrobial properties of the specimens weredetermined by bacteria counting method and Live/Dead staining,

G. Jin et al. / Biomaterials 65 (2015) 22e3124

using both Staphylococcus aureus (S. aureus) and Escherichia coli(E. coli). Before the experiments, the specimens were sterilized in75% ethanol for 2 h. The bacterial suspension (concentration of107 cfu/ml) was dripped onto the surfaces to a density of 60 ml/cm2.After 24 h incubation at 37 �C, the dissociated bacteria werecollected, introduced to a standard agar culture plate and foranother 24 h incubation. The bacterial colonies were counted andthe antibacterial ratio was calculated using the formula below

ðA� BÞA

� 100%;

in which A is average number of the bacteria on Ti (CFU/specimen),B is average number of the bacteria on testing specimens (CFU/specimen). For SEM examination, the bacteria were fixed with 2.5%glutaraldehyde solution, dehydrated in gradient ethanol solutions(30, 50, 75, 90, 95 and 100 v/v %), and finally dried in hexamethyldisilazane ethanol solution series.

For Live/Dead staining, the specimens were rinsed with PBStwice after culturing for 24 h, and stained with 0.5 ml of the Live/Dead BacLight reagent in dark for 15 min.

2.6. In vivo osteogenic and antibacterial activity

2.6.1. Surgical implantationThe experiments were approved by the Animal Care and

Experiment Committee of Sixth People's Hospital Affiliated withthe School of Medicine of Shanghai Jiaotong University. Usingaseptic techniques, 4 holes perpendicular to the centerline of thetibia were sequentially drilled using a Kirschner wire (2 mmdiameter) of the left tibia of male rabbits (pitches of 1 cm). Fourimplants were randomly implanted into each drilled hole. Forbacterial inoculation, 20 ml of the bacterial suspension with adensity of 104 cfu/ml S. aureus (ATCC 43300) was injected into themedullary cavity with a microsyringe. After bacterial inoculation,the fascia and skin were sutured. Following surgery, the rabbits

Fig. 1. Surface views: (a) Pure Ti, (b) Zn/Ag

were housed in the separate cages and allowed to eat and drinkadlibitum.

2.6.2. Radiographic and micro-computed tomography evaluationRadiography of days 3, 14, 28, and 42 was performed while

the rabbits were under 3% pentobarbital (1 ml/kg) anesthesia. TheX-ray results were assessed in line with the literature [22]. Theoperated tibia were dissected, harvested, and fixed in 10% bufferedformalin. After fixation for 48 h, the specimens were scanned byhigh-resolution micro-computed tomography (micro-CT; Skyscan1176, Skyscan, Belgium) at an image resolution of 18.0 mm (55kVpand 181 mA radiation source with 0.5 mm aluminum filter). The 2Dand 3D high-resolution reconstructed images were obtained by thesoftware provided by the manufacturer.

2.6.3. Histopathological analysisAfter 6 weeks and microCT scanning, the samples were decal-

cified using 10% EDTA solution for 28 days, followed by washingwith running tap water for about 4 h, and then transferred to a 75%ethanol solution and embedded in paraffin. 5 mm thick shaftsections were collected. Masson's Trichrome staining, hematoxylin-eosin (HE) staining, and Giemsa staining were used to assess themorphology and bacterial contamination [23].

2.7. Statistical analysis

The data were expressed as means ± standard deviations. Thestatistical analysis was performed using the two-way analysis usinga GraphPad Prism statistical software package, and the p values<0.05 were considered to be statistically significant.

3. Results

3.1. Characterization of the Zn/Ag dual-ion implanted titanium

The surface views of the specimens are shown in Fig. 1. The Ti

-PIII, (c) Zn-Ag-PIII, and (d) Ag-Zn-PIII.

G. Jin et al. / Biomaterials 65 (2015) 22e31 25

surface shows a flat and smooth topography. However, a largenumber of Ag nanoparticles (Ag NPs) with a wide size distributioncan be observed from Zn/Ag-PIII (Fig. 1b). It was also evident thatthe clarity of the Ag NPs is not so good and most of them appear tobe wrapped by a Zn film. This may be ascribed to simultaneousnucleation of the implanted Zn and Ag and the results are inagreement with our previous work [12]. Larger Ag NPs can be foundfromZn-Ag-PIII (Fig.1c) which looks clearer on the surface of the Znfilm produced by the previously implanted Zn ions. Ag NPs canhardly be detected from Ag-Zn-PIII (red arrows in Fig. 1d) possiblybecause some have been sputtered from the surface and some arecovered by the Zn film formed afterwards. The difference in theembedding of Ag NPs in the various specimens arises from thechange in the implantation sequence.

The depth profile of Zn (Fig. 2a) resembles a Gaussian distri-bution from 10 to 140 nm but significantly more Zn can be observedfrom the surface (0e10 nm) in the following order of Ag-Zn-PIII > Zn/Ag-PIII > Zn-Ag-PIII. The depth profiles of Ag (Fig. 2b)also resemble Gaussian distributions (from 20 to 140 nm) and thesurface contents of Ag arising from deposition follow the order ofZn-Ag-PIII > Zn/Ag-PIII > Ag-Zn-PIII, which is opposite to that of Zn.As mentioned earlier, few Ag NPs are observed from Ag-Zn-PIII.Considering that the surface Ag content in Ag-Zn-PIII (1.57%) isalmost half of that in Zn/Ag-PIII which is 2.89%, it can be concludedthat some of the Ag NPs are covered by the Zn film formed laterdespite the sputtered Ag NPs. The high-resolution spectra of Zn 2pand Ag 3d detected from the various surfaces (Fig. S1) indicate thatthe implanted Zn exist as ZnO (1021.9 eV) in the near surface andmetallic Zn (1021.0 eV) underneath [24,25], while the Ag NPs havethe metallic state in lieu of the oxidized one [26], consistent withour previous study [12].

To investigate the corrosion resistance of the Zn/Ag dual-ionimplanted titanium and corrosion rate of the micro-galvanic cou-ples formed between the implanted Zn and Ag, potentiodynamicpolarization is conducted in 0.9% NaCl solution. The tafel plots of

Fig. 2. Elemental depth profiles detected from Zn/Ag dual-ion im

Fig. 3. Potentiodynamic polarization plots (a), released Zn concentrations (b) and

the specimens are presented in Fig. 3a. The corrosion potentials ofall the specimens shift positively as opposed to pure Ti, especiallyZn/Ag-PIII and Zn-Ag-PIII, in agreement with our previous study[12]. The corresponding corrosion currents of Ti, Zn/Ag-PIII, Zn-Ag-PIII, and Ag-Zn-PIII are 8.98� 10�7, 6.71� 10�7, 4.99� 10�7, and1.81� 10�7, respectively. The larger corrosion potential and smallercorrosion current suggest better corrosion resistance on thesequentially implanted titanium [27].

To elucidate the relationship between the release of Zn and Agfrom the Zn/Ag dual-ion implanted titanium and the micro-galvanic couples, the amounts of released Zn and Ag aremeasured by ICP-AES. The concentration of Zn2þ released fromZn/Ag-PIII is larger than that from Zn-Ag-PIII (Fig. 3b). However,the Zn2þ concentration of Ag-Zn-PIII is significantly larger thanthose of Zn/Ag-PIII and Zn-Ag-PIII possibly due to the bigger Znsurface content (Fig. 2a). The concentration of released Ag is quitesmall (Fig. 3c) and not more than 0.0695 mg/ml after soaking in0.9% NaCl solution for 28 days. The positively-shifted corrosionpotential and difference in the corrosion currents as well as Znand Ag released from the Zn/Ag dual-ion implanted titaniumlikely stems from the difference in the corrosion rates of themicro-galvanic couples and this will be discussed in more detailsin the discussion section.

3.2. In vitro osteogenic activity

Fig. 4a shows the proliferation of rBMSCs cultured on the surfacefor various times. There is no obvious difference among the fourgroups at the first day. And cell proliferation on Zn/Ag-PIII and Ag-Zn-PIII is more substantial than that on pure Ti and Zn-Ag-PIII atday 4. The cells proliferate better on the implanted samplescompared to Ti at day 7, suggesting that Zn/Ag dual-ion implanta-tion spurs proliferation of rBMSCs.

The ALP activity of rBMSCs is presented in Fig. 4b. The ALP ex-pressions on the Zn/Ag dual-ion implanted titanium are notably

planted samples: (a) Zn depth profiles, (b) Ag depth profiles.

released Ag concentrations (c) of all samples soaked in 0.9% NaCl solution.

Fig. 4. Cell proliferation (a), Quantitative ALP activity (b), Collagen secretion (c), ECM mineralization (d) of rBMSCs cultured on the various surfaces, and osteogenisis-related gene:(e) ALP, (f) Col-I, (g) Runx2 and (h) OCN expressions by rBMSCs on various surfaces; *P < 0.05, **P < 0.01, ***P < 0.001.

G. Jin et al. / Biomaterials 65 (2015) 22e3126

improved at day 7 and 14, especially on Zn/Ag-PIII specimen. Theresults are further confirmed by the ALP staining assay, as shown inFig. S2, indicating the enhancement of Zn/Ag dual-ion implantationon rBMSCs differentiation.

Collagen secretion from rBMSCsmeasured by Sirius Red stainingis displayed in Fig. 4c. More collagen is secreted at day 7, especiallyZn/Ag-PIII and again, the trend becomes more significant at day 14,further confirming the results by staining analysis (Fig. S3). ECMmineralization assayed by Alizarin Red staining is presented inFig. 4d. Matrix mineralization on the Zn/Ag dual-ion implanted ti-tanium was up-regulated as opposed to Ti at day 7, especially Zn/Ag-PIII. At day 14, matrix mineralization on Zn/Ag-PIII is stillhigher than that on Ti, which was consistence with the stainingresults (Fig. S4).

The osteogenesis-related genes expressions on the specimensare displayed in Fig. 4e, f, g, and h. Although the expressions of ALPon the Zn/Ag dual-ion implanted titanium are only slightly up-regulated at day 7, the expressions at day 14 is more significantcompared to Ti. As an early marker of osteoblastic differentiation,

Col-I is statistically up-regulated on the modified specimens at day7, especially Zn/Ag-PIII and Ag-Zn-PIII. The expressions at day 14are still higher than that on Ti. The Runx2 (a key bone-specifictranscription factor) expression shows statistically difference onZn/Ag-PIII at day 7 compared to the other three groups, whereasthose on the Zn/Ag dual-ion implanted titanium are all up-regulated at day 14 compared to Ti. The up-regulated expressionof OCN (a later marker of osteoblastic differentiation) indicates thatosteoblastic differentiation occurs in the later stage on themodifiedspecimens.

3.3. In vitro antibacterial tests

Biomaterials often face a wide variety of Gram positive andGram negative bacteria in the human body [28]. In the presentwork, E. coli and S. aureus are utilized to evaluate the antimicrobialproperties by employing the bacteria counting method and Live/Dead staining. It can be easily seen from the bacteria countingresults (Fig. S5), the reduction rate of E. coli on Zn/Ag-PIII is

G. Jin et al. / Biomaterials 65 (2015) 22e31 27

approximately 98%, whereas these on Zn-Ag-PIII and Ag-Zn-PIII are87% and 52%, respectively. The larger reduction rate on Zn/Ag-PIIIand Zn-Ag-PIII is likely due to the embedded Ag NPs which arebelieved to disrupt the bacteria membrane through short-rangeinteractions rather than release of Ag ions as reported earlier[10,11,18]. The reduction rates of S. aureus on Zn/Ag-PIII, Zn-Ag-PIII,and Ag-Zn-PIII are approximately 99%, 98%, and 68%, respectively,presenting a similar tendency.

Fig. 5a shows the fluorescent images of E. coli and S. aureus onthe various surfaces. Red spots can hardly be detected on Tiwhereas there are many green spots, indicating the existence of alarge number of live E. coli cells. However, little live bacteria and afew dead bacteria can be observed from the Zn/Ag dual-ionimplanted titanium, suggesting that the implanted samples, espe-cially Zn/Ag-PIII and Zn-Ag-PIII, are more resistant against E. coli.Similar results are observed from S. aureus.

The morphology of E. coli and S. aureus are displayed in Fig. 5b.The E. coli cells on Ti have a rod shape, while complete lysis andcytosolic content leakage are observed from Zn/Ag-PIII and Zn-Ag-PIII, very few intact cells are detected from Ag-Zn-PIII. This isin agreement with the Live/Dead staining (Fig. 5a) and bacteriacounting results (Fig. S5). S. aureus on Ti have the normal sphericalshape without apparent leakage (Fig. 5b), but complete lysis isobserved on Zn/Ag-PIII and Zn-Ag-PIII and it occurs occasionally onAg-Zn-PIII.

3.4. In vivo osseointegration and antibacterial ability

3.4.1. X-ray and micro-CT analysisImplant-related infection and new bone formation after surgery

are evaluated by X-ray and micro-CT. Fig. 6a shows the X-rayradiographical images of the implants at prescribed time points. No

Fig. 5. Fluorescent images of live and dead bacteria cultured on the various surfaces

evident signs of osseous destruction are observed within 3 days.However, small-density areas are evident in the tibia with the Tiimplant after 2 weeks' implantation. The presence of cortical bonethinning and disruption as well as osteolysis indicates bacterialinfection and inflammation and the observation is in accordancewith septic osteomyelitis [22,29]. The small-density areas, whichoften lead to osteolysis, become much more obvious after 4 weeks'implantation. After 6 weeks, significant osseous destruction andsoft tissue swelling can be detected. It is also evident that Ti implantloses contact with the host bone, indicating that bacterial infectionloosens the implant and contributes to bone resorption in thecortical bone region around the implant. In contrast, osseousdestruction is not observed by X ray from the tibias in the Zn/Agdual-ion implanted titanium groups within 6 weeks, despite thepresence of sparse signs of tissue swelling in Ag-Zn-PIII. The resultsdemonstrate that Zn/Ag dual-ion implantation produces goodantibacterial effects in vivo.

Micro-CT scans are obtained from the tibia 6 weeks aftersurgery and the results are presented in Fig. 6b. The formation ofabscess, osteolysis, and resorption of cortical bone around the Tiimplant are consistent with the radiographic results in Fig. 6a.Moreover, necrotic bone sequestra (red arrow) can be found fromthe Ti implant indicating bacterial infection and inflammation.Sequestra is a complication of osteomyelitis and present whenbacterial infection occurs around the implant due to the scarcevascularity of bones [30,31]. On the contrary, no sign of inflam-mation can be observed from Zn/Ag-PIII and Zn-Ag-PIII and thereis also evidence of new bone formation around the implant nearthe cortical bone, indicating that these two groups exhibitexcellent osseointegration at the presence of S. aureus in vivo.New bone formation is also observed from Ag-Zn-PIII, but thepresence of osteolysis suggests that Ag-Zn-PIII only has partial

(a) and SEM morphology (b) of E. coli and S. aureus seeded on various surfaces.

Fig. 6. X-ray images (a) and micro-CT results (b) of rabbit tibia implanted with the rod implants at 3 days, 2, 4, and 6 weeks post surgery. (a) No radiographic signs of osseousdestruction are observed within 3 days, small-density areas can be evident in tibia with Ti implant at 2 weeks, and they become much more obvious after 4 weeks, significantosseous destruction and soft tissue swelling can be detected at week 6. In contrast, osseous destruction does not occur in any of the X-ray-monitored tibias in the Zn/Ag dual-ionimplanted titanium groups within 6 weeks. (b) Formation of abscess, osteolysis, resorption of cortical bone and necrotic bone sequestra (red arrow) can be observed around the Tiimplant, no signs of inflammation but new bone formation are observed from Zn/Ag-PIII and Zn-Ag-PIII implants. New bone formation and presence of osteolysis can be detected onAg-Zn-PIII implant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

G. Jin et al. / Biomaterials 65 (2015) 22e3128

antibacterial ability. The trend is consistent with the antibacterialtests in vitro.

3.4.2. Histological evaluationTo evaluate bacterial infection and osseointegration on the

rabbit tibia, Masson's trichrome staining, HE, and Giemsa stainingare employed. The histological slice after Masson's trichromestaining shows the typical signs of bone infection (Fig. S6) asillustrated by the formation of abscess and destruction of corticalbone around the Ti implant [32]. It is consistent with the radio-graphic and micro-CT results in Fig. 6. Bone resorption is accom-panied by a large number of inflammatory cells in the medullarycavity. There is no sign of bacterial infection on Zn/Ag-PIII and Zn-Ag-PIII and evidence of new bone formation on the surface in themedullary cavity indicative of osseointegration and antibacterialability in vivo. Relatively slight abscess, minor bone destruction, andless new bone formation are detected from Ag-Zn-PIII and the re-sults are confirmed by HE and Giemsa staining.

The high-magnification histological images are displayed inFig. 7. Tissues infiltrated by inflammatory cells can easily beobserved from Ti after Masson's trichrome staining and multinu-cleated osteoclasts (yellow arrows) are detected near the trabecularbone. As the principal resorptive cells of bone, osteoclasts appear atsites of active bone resorption and develop specialized cell mem-branes further dissolving bone minerals by active Hþ secretion inthe microenvironment in between them and bone [33,34]. There-fore, the observation of osteoclasts suggests the presence of bac-terial infection and bone resorption. No sign of bacterial infection orbone resorption can be detected from Zn/Ag-PIII and Zn-Ag-PIII.

Formation of new bone extending to the medullary cavity isobserved from the implant surface and it is consistent with themicro-CT analysis in Fig. 6b. The newly formed bone grows verti-cally to the cortical bone at the interface between the implant andhost bone. Similar results are obtained by HE and Giemsa stainingas shown in Fig. 7, indicating that Zn/Ag dual-ion implantationproduces good osseointegration and antibacterial effects in vivo.

4. Discussion

Three different micro-galvanic couples are fabricated on tita-nium using PIII in our experiments and they exhibit different sur-face morphology. Specifically, the Ag NPs on Zn/Ag-PIII are coveredby a simultaneously formed Zn film, leaving only the top to bevisible on the titanium surface. In comparison, they are on thesurface of the Zn film on Zn-Ag-PIII and the majority of the Ag NPsare underneath the Zn film on Ag-Zn-PIII (Fig. 1). According to thethree different structures, the 3D Max software is employed tocreate the surface models of the Zn/Ag dual-ion implanted titaniumas shown in Fig. 8.

It has been demonstrated that micro-galvanic couples areformed by Zn and Ag PIII with Zn serving as the anode and Ag as thecathode [12]. In the present study, we further investigate thecorrosion process of the three different micro-galvanic couples todemonstrate the relationship between the enhanced biologicalbehaviors and the micro-galvanic couples. The corrosion potentialsof the Zn/Ag dual-ion implanted titanium shift positively comparedto pure Ti (shown in Fig. 3a) and the corrosion currents are espe-cially important when corrosion occurs. The corrosion currents

Fig. 7. High magnification histological images of masson's trichrome staining, HE staining, and giemsa staining of transverse sections at 6 weeks after surgery. The images below arehigher magnifications of the areas within the yellow boxes. Multinucleated osteoclasts (yellow arrows) are detected near the trabecular bone around Ti implant, indicating thepresence of bacterial infection and bone resorption. No signs of bacterial infection or bone resorption but new bone formation are detected from the Zn/Ag-PIII, Zn- Ag-PIII, and Ag-PIII groups. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

G. Jin et al. / Biomaterials 65 (2015) 22e31 29

increase in the following order: Ag-Zn-PIII < Zn-Ag-PIII < Zn/Ag-PIII. This may be because the Ag NPs on Ag-Zn-PIII are almostcovered by the Zn film and it is relatively difficult for the micro-galvanic couples to reach the conducting state thereby resultingin the smaller corrosion current and less consumption of Hþ ac-cording to the following cathodic reaction:

2Hþþ 2e�/H2: (1)

Compared to Ag-Zn-PIII, it is easier for Zn/Ag-PIII and Zn-Ag-PIIIto reach the conducting state as the embedded Ag NPs are in directcontact with the Zn film. Hence, the micro-galvanic couples aretriggered when the samples are immersed in a physiological me-dium. Nevertheless, the difference in the corrosion currents is likelydue to the difference in the contact between the Ag NPs and Zn filmin Zn/Ag-PIII and Zn-Ag-PIII. According to the surface morphologyand model (Fig. 8), the Ag NPs on Zn/Ag-PIII are embedded on thetitanium surface and wrapped by a Zn film whereas those on Zn-Ag-PIII are immobilized on the top of Zn film. The contact areabetween the Ag NPs and Zn on Zn/Ag-PIII is larger than that on Zn-

Ag-PIII. The micro-galvanic couples on Zn/Ag-PIII are still triggereduntil Zn, which is in contact with the Ag NPs, is consumed therebyexposing the Ag NPs in the microenvironment between thetitanium and cell (or bacteria). However, triggering of the micro-galvanic couples on Zn-Ag-PIII ends when Ag NPs leave thetitanium surface and are released into the microenvironment.Therefore, the corrosion rate of Zn/Ag-PIII is faster than that ofZn-Ag-PIII as confirmed by the corrosion currents.

A larger corrosion current implies a faster reaction and moresubstantial release of Zn2þ and consumption of Hþ. According tothe 3D Max model in Fig. 8, the amount of Zn2þ released from Zn/Ag-PIII is larger than that from Zn-Ag-PIII (confirmed by the ionrelease results in Fig. 3b). The larger amount of released Zn2þ ions isdue to the significantly larger surface Zn content on Ag-Zn-PIII(Fig. 3b). Zn2þ in the microenvironment between the titaniumand cells can be adjusted by Zn transporters [35,36] further exert-ing a stimulatory effect on the metabolism of bones [37]. Moreover,the synergistic effects of the long-range imposed by Zn2þ andshort-range interactions produced by the Ag NPs enhance theosteogenic activity and antibacterial properties both in vitro and

Fig. 8. Surface model and corrosion process of three types of Zn/Ag micro-galvanic couples formed on the Zn/Ag dual-ion implanted titanium mimicked by 3D Max software.

G. Jin et al. / Biomaterials 65 (2015) 22e3130

in vivo. The pH of the microenvironment between the osteoclastsand bone surface is crucial to cellular processes and should bemaintained at 4.5 [33,38]. Consumption of Hþ by the cathodichydrogen evolution reaction of the micro-galvanic couples inhibitsbone resorption by osteoclasts. According to the corrosion currents,the consumption of Hþ increases in the following order: Ag-Zn-PIII < Zn-Ag-PIII < Zn/Ag-PIII and therefore, bone resorption onAg-Zn-PIII is more serious and Zn/Ag-PIII exhibits the bestosseointegration in vivo (Fig. S6). The present work demonstratesthat the biological properties can be controlled by the corrosionprocess on the designed micro-galvanic couples.

5. Conclusion

Three types of Zn/Ag micro-galvanic couples are fabricated ontitanium by plasma immersion ion implantation to investigate theosseointegration and antibacterial effects as well as the involvedmechanisms. The micro-galvanic couples exhibit excellent osteo-genic activity and antibacterial ability in vitro without producingcytotoxicity. The Zn/Ag micro-galvanic couple formed on Zn/Agdual-ion co-implanted titanium shows the best osseointegration aswell as good antibacterial properties in vivo obtained from a rabbittibia model. The difference among the three structures in vitro andin vivo can be explained by the contact between the Ag NPs and Znfilm, which further affects the corrosion rates of the micro-galvanic

couples. The findings provide insights into the design of new or-thopedic and dental implant implants with simultaneousosseointegration and antibacterial properties.

Acknowledgments

Financial support from the National Basic Research Program ofChina (973 Program, 2012CB933600), National Natural ScienceFoundation of China (31200721 and 31370962), Shanghai Com-mittee of Science and Technology, China (13441902400 and14XD1403900), City University of Hong Kong Strategic ResearchGrant (SRG) No. 7004188, and Hong Kong Research Grants Council(RGC) General Research Funds (GRF) No. CityU 112212 areacknowledged.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2015.06.040.

References

[1] X. Liu, P.K. Chu, C. Ding, Surface modification of titanium, titanium alloys, andrelated materials for biomedical applications, Mater. Sci. Eng. R. 47 (2004)49e121.

[2] S. Mei, H. Wang, W. Wang, L. Tong, H. Pan, C. Ruan, et al., Antibacterial effects

G. Jin et al. / Biomaterials 65 (2015) 22e31 31

and biocompatibility of titanium surfaces with graded silver incorporation intitania nanotubes, Biomaterials 35 (2014) 4255e4265.

[3] H. Qin, H. Cao, Y. Zhao, C. Zhu, T. Cheng, Q. Wang, et al., In vitro and in vivoanti-biofilm effects of silver nanoparticles immobilized on titanium, Bio-materials 35 (2014) 9114e9125.

[4] R.O. Darouiche, Treatment of infections associated with surgical implants,New. Engl. J. Med. 350 (2004) 1422e1429.

[5] L. Zhao, P.K. Chu, Y. Zhang, Z. Wu, Antibacterial coatings on titanium implants,J. Biomed. Mater. Res. B 91B (2009) 470e480.

[6] S. Qian, Y. Qiao, X. Liu, Selective biofunctional modification of titaniumimplants for osteogenic and antibacterial applications, J. Mater. Chem. B 2(2014) 7475e7487.

[7] Y. Qiao, W. Zhang, P. Tian, F. Meng, H. Zhu, X. Jiang, et al., Stimulation of bonegrowth following zinc incorporation into biomaterials, Biomaterials 35 (2014)6882e6897.

[8] G. Jin, H. Cao, Y. Qiao, F. Meng, H. Zhu, X. Liu, Osteogenic activity and anti-bacterial effect of zinc ion implanted titanium, Colloid Surf. B 117 (2014)158e165.

[9] H. Hu, W. Zhang, Y. Qiao, X. Jiang, X. Liu, C. Ding, Antibacterial activity andincreased bone marrow stem cell functions of Zn-incorporated TiO2 coatingson titanium, Acta Biomater. 8 (2012) 904e915.

[10] H. Cao, Y. Qiao, F. Meng, X. Liu, Spacing-dependent antimicrobial efficacy ofimmobilized silver nanoparticles, J. Phys. Chem. Lett. 5 (2014) 743e748.

[11] H. Cao, Y. Qiao, X. Liu, T. Lu, T. Cui, F. Meng, et al., Electron storage mediateddark antibacterial action of bound silver nanoparticles: smaller is not alwaysbetter, Acta Biomater. 9 (2013) 5100e5110.

[12] G. Jin, H. Qin, H. Cao, S. Qian, Y. Zhao, X. Peng, et al., Synergistic effects of dualZn/Ag ion implantation in osteogenic activity and antibacterial ability oftitanium, Biomaterials 35 (2014) 7699e7713.

[13] S. Kajita, D. Kitaoka, N. Ohno, R. Yoshihara, N. Yoshida, T. Yoshida, Surfacemodification of titanium using He plasma, Appl. Surf. Sci. 303 (2014) 438e445.

[14] E.J.D.M. Pillaca, M. Ueda, H. Reuther, C.M. Lepienski, Study of the effects ofplasma immersion ion implantation on austenitic stainless steel using E�Bfields, Surf. Coat. Tech. 246 (2014) 1e5.

[15] G. Wang, J. Li, W. Zhang, L. Xu, H. Pan, J. Wen, et al., Magnesium ionimplantation on a micro/nanostructured titanium surface promotes itsbioactivity and osteogenic differentiation function, Int. J. Nanomed 9 (2014)2387e2398.

[16] R. Xu, X. Yang, P. Li, K.W. Suen, G. Wu, P.K. Chu, Electrochemical propertiesand corrosion resistance of carbon-ion-implanted magnesium, Corros. Sci. 82(2014) 173e179.

[17] P. Budzynski, A.A. Youssef, J. Sielanko, Surface modification of Tie6Ale4Valloy by nitrogen ion implantation, Wear 261 (2006) 1271e1276.

[18] H. Cao, X. Liu, F. Meng, P.K. Chu, Biological actions of silver nanoparticlesembedded in titanium controlled by micro-galvanic effects, Biomaterials 32(2011) 693e705.

[19] D. Krupa, J. Baszkiewicz, J.A. Kozubowski, A. Barcz, J.W. Sobczak, A. Bili�nski, etal., Effect of dual ion implantation of calcium and phosphorus on the prop-erties of titanium, Biomaterials 26 (2005) 2847e2856.

[20] Y. Xie, X. Liu, A. Huang, C. Ding, P.K. Chu, Improvement of surface bioactivity

on titanium by water and hydrogen plasma immersion ion implantation,Biomaterials 26 (2005) 6129e6135.

[21] F. Ren, C.Z. Jiang, L. Zhang, Y. Shi, J.B. Wang, R.H. Wang, Formation andmicrostructural investigation of AgeCu alloy nanoclusters embedded in SiO2formed by sequential ion implantation, Micron 35 (2004) 489e493.

[22] Y.H. An, R.J. Friedman, Animal models of orthopedic implant infection, J. InvestSurg. 11 (1998) 139e146.

[23] V. Brinkmann, U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D.S. Weiss, etal., Neutrophil extracellular traps kill bacteria, Science 303 (2004) 1532e1535.

[24] C.T. Campbell, K.A. Daube, J. White, Cu/ZnO (0001) and ZnOx/Cu (111): modelcatalysts for methanol synthesis, Surf. Sci. Lett. 182 (1987) 458e476.

[25] B.R. Strohmeier, D.M. Hercules, Surface spectroscopic characterization of theinteraction between zinc ions and g-alumina, J. Catal. 86 (1984) 266e279.

[26] J. Li, Y. Qiao, Z. Ding, X. Liu, Microstructure and properties of Ag/N dual ionsimplanted titanium, Surf. Coat. Tech. 205 (2011) 5430e5436.

[27] M.I. Jamesh, G. Wu, Y. Zhao, W. Jin, D.R. McKenzie, M.M.M. Bilek, et al., Effectsof zirconium and nitrogen plasma immersion ion implantation on the elec-trochemical corrosion behavior of MgeYeRE alloy in simulated body fluid andcell culture medium, Corros. Sci. 86 (2014) 239e251.

[28] D. Campoccia, L. Montanaro, C.R. Arciola, A review of the biomaterials tech-nologies for infection-resistant surfaces, Biomaterials 34 (2013) 8533e8554.

[29] T.P. Schaer, S. Stewart, B.B. Hsu, A.M. Klibanov, Hydrophobic polycationiccoatings that inhibit biofilms and support bone healing during infection,Biomaterials 33 (2012) 1245e1254.

[30] A. Piattelli, A. Scarano, M. Piattelli, E. Vaia, S. Matarasso, A microscopicalevaluation of 24 retrieved failed hollow implants, Biomaterials 20 (1999)485e489.

[31] F. Jennin, V. Bousson, C. Parlier, N. Jomaah, V. Khanine, J.-D. Laredo, Bonysequestrum: a radiologic review, Skelet. Radiol. 40 (2011) 963e975.

[32] M. Lucke, G. Schmidmaier, S. Sadoni, B. Wildemann, R. Schiller, A. Stemberger,et al., A new model of implant-related osteomyelitis in rats, J. Biomed. Mater.Res. B 67B (2003) 593e602.

[33] S.L. Teitelbaum, Bone resorption by osteoclasts, Science 289 (2000)1504e1508.

[34] J. Salo, P. Lehenkari, M. Mulari, K. Metsikk€o, H.K. V€a€an€anen, Removal ofosteoclast bone resorption products by transcytosis, Science 276 (1997)270e273.

[35] Z. Tang, S.N. Sahu, M.A. Khadeer, G. Bai, R.B. Franklin, A. Gupta, Over-expression of the ZIP1 zinc transporter induces an osteogenic phenotype inmesenchymal stem cells, Bone 38 (2006) 181e198.

[36] J.P. Liuzzi, R.K. Blanchard, R.J. Cousins, Differential regulation of zinc trans-porter 1, 2, and 4 mRNA expression by dietary zinc in rats, J. Nutr. 131 (2001)46e52.

[37] S. Gomez, R. Rizzo, M. Pozzi-Mucelli, E. Bonucci, F. Vittur, Zinc mapping inbone tissues by histochemistry and synchrotron radiationeinduced x-rayemission: correlation with the distribution of alkaline phosphatase, Bone 25(1999) 33e38.

[38] F.L. Yuan, X. Li, W.G. Lu, C.W. Li, J.P. Li, Y. Wang, The vacuolar ATPase in bonecells: a potential therapeutic target in osteoporosis, Mol. Biol. Rep. 37 (2010)3561e3566.

Supplementary materials for

Zn/Ag micro-galvanic couples formed on titanium and

osseointegration effects in the presence of S. aureus

Guodong Jina,1, Hui Qinb,1, Huiliang Caoa, Yuqin Qiaoa, Yaochao Zhaob, Xiaochun

Pengb, Xianlong Zhang b,*, Xuanyong Liua,*, Paul K Chuc

a State Key Laboratory of High Performance Ceramics and Superfine Microstructure,

Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050,

China

b Department of Orthopedics, Shanghai Sixth People’s Hospital, Shanghai Jiao Tong

University, Shanghai 200233, China

c Department of Physics and Materials Science, City University of Hong Kong, Tat

Chee Avenue, Kowloon, Hong Kong, China

1 The authors contributed equally to this work.

Corresponding Authors:

X.Y. Liu - Tel: + 86 21 5241 2409; Fax: + 86 21 5241 2409; Electronic mail:

xyliu@mail.sic.ac.cn

X.L. Zhang - Tel: + 86 21 6436 9183; Fax: +86 21 6470 1363; Electronic mail:

zhangxianl197826@163.com

Supplemental Content:

1. Supplemental table,

2. Captions and descriptions of the supplemental figures.

1. Supplementary Table

Table S1. Primers for RT-PCR

2.

Gene

Prime sequence

(F, forward; R, reverse; 5′ to 3′)

Product

Size (bp)

F-actin

F: CACCCGCGAGTACAACCTTC

R: CCCATACCCACCATCACACC

207

ALP F: CGTCTCCATGGTGGATTATGCT

R: CCCAGGCACAGTGGTCAAG

209

Col-I

F:CTGCCCAGAAGAATATGTATCACC

R: GAAGCAAAGTTTCCTCCAAGACC

198

OCN F: GCCCTGACTGCATTCTGCCTCT

R: TCACCACCTTACTGCCCTCCTG

103

Runx2

F: TCTTCCCAAAGCCAGAGCG

R: TGCCATTCGAGGTGGTCG

154

2. Supplementary Figures

Supplementary Fig. S1. High resolution spectra detected from Zn/Ag dual-ion

implanted titanium surfaces: (a) Zn2p, (b) Ag 3d from Zn/Ag-PIII; (c) Zn2p, (d) Ag

3d from Zn-Ag-PIII; (e) Zn2p, (f) Ag 3d from Ag-Zn-PIII

Supplementary Fig. S2. ALP positive areas of rBMSCs cultured on various surfaces

for 7 and 14 days

Supplementary Fig. S3. Collagen secretion of rMSCs cultured on various surfaces

for 7 and 14 days

Supplementary Fig. S4. Matrix mineralization of rBMSCs cultured on various

surfaces for 7 and 14 days

Supplementary Fig. S5. Re-cultivated bacterial colonies on agar: (a) E. coli and (b) S.

aureus seeded on various surfaces with the bacteria concentration being 107 cfu/ml;

Percentage reduction: (c) E. coli and (d) S. aureus re-cultured on agar after

dissociation from the various surfaces with the re-cultivated bacteria concentration

being 107 cfu/ml; ***P < 0.001.

Supplementary Fig. S6. Low magnification histological images of masson’s

trichrome staining, HE staining, and giemsa staining of transverse sections at 6 weeks

after surgery. Formation of abscess, destruction of cortical bone accompanied by large

number of inflammatory cells can be observed around Ti implant. Slight abscess,

minor bone destruction and less new bone formation are detected from Ag-Zn-PIII.

However, no sign of bacterial infection is detected from Zn/Ag-PIII and Zn-Ag-PIII

groups.