Silk-functionalized titanium surfaces for enhancing osteoblast functions and reducing bacterial adhesion

Biomaterials 29 (2008) 4751–4759

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Silk-functionalized titanium surfaces for enhancing osteoblast functions and reducing bacterial adhesion Fan Zhang a, Zhengbiao Zhang a, b, Xiulin Zhu b, En-Tang Kang a, Koon-Gee Neoh a, * a b

Department of Chemical and Biomolecular Engineering, National University of Singapore, Kent Ridge, Singapore 119260, Singapore School of Chemistry and Chemical Engineering, Suzhou University, Suzhou 215006, People’s Republic of China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 June 2008 Accepted 28 August 2008 Available online 1 October 2008

It would be ideal to have implants which can simultaneously inhibit bacterial adhesion and promote osteoblast functions. In this work, titanium surfaces were modified with poly(methacrylic acid) (P(MAA)) followed by immobilization of silk sericin. Firstly a trichlorosilane coupling agent, which is an atom transfer radical polymerization (ATRP) initiator, was immobilized on the oxidized titanium surface to facilitate the surface-initiated ATRP of methacrylic acid sodium salt (MAAS). The pendant carboxyl end groups of the grafted and partially protonated MAA chains were subsequently coupled with silk sericin via carbodiimide chemistry. The functionalized Ti surfaces were characterized by X-ray photoelectron spectroscopy, and assayed for osteoblast cell functions and bacterial adhesion. The covalently immobilized MAA brushes significantly reduce the adhesion of the two bacterial strains (Staphylococcus aureus and Staphylococcus epidermidis) tested. The silk sericin-immobilized surfaces, at the same time, promote osteoblast cells’ adhesion, proliferation, and alkaline phosphatase activity. Thus, the P(MAA) and silk sericin functionalized Ti surfaces have potential applications combating biomaterial-centered infection and promoting osseointegration. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Titanium Silk Surface-initiated ATRP Methacrylic acid sodium salt Osteoblast Bacterial adhesion

1. Introduction Knee and hip fracture fixation, replacements and other surgical implants are becoming common procedures in recent years to replace or restore the function of fractured bone and damaged joints [1]. Because of its good biocompatibility and mechanical properties [2], titanium has attracted a lot of attention and research, especially on the surface modification of titanium-based implants for acceleration of bone formation and reduction of infection. Slow and partial osseointegration of orthopedic implants will typically increase the risk of implant loosening or even failure. Improvement in mammalian cell adhesion and proliferation is critical for good integration of the host hard tissue with the implanted biomaterial [3]. On the other hand, while improvements in the control of sterility during operations have been achieved, surgical infections associated with orthopedic implants still represent one of the most serious complications [4]. Adhesion of Staphylococcus aureus (S. aureus) and Staphylococcus epidermidis (S. epidermidis) on the implant surface promotes implant colonization and is the main cause of implant infection [5]. The effectiveness of systemic and

* Corresponding author. Tel.: þ65 65162176; fax: þ65 67791936. E-mail address: [email protected] (K.-G. Neoh). 0142-9612/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2008.08.043

local antibiotic prophylaxis of infections of implanted devices is constrained by the continual evolution of antibiotic resistant bacteria [4]. In addition, adhesion-mediated infections are extremely resistant to antibiotics due to biofilms that protect the organisms from the activity of antimicrobial agents and from host defense mechanisms. Hence, such infections frequently persist until the biomaterial is removed [6]. Silks are fibrous proteins with remarkable mechanical properties produced in fiber form by silkworms and spiders [7]. Silk contains two natural macromolecular proteins, sericin and fibroin. Sericin comprises granular and high molecular, water soluble glycoproteins and it acts as a protein glue to fix fibroin fibers together in the cocoon [7]. The less water soluble components of silk, i.e. fibroin have been widely used in biomedical application, especially as scaffold [8], but research on the use of the more water soluble sericin for biomedical applications is limited. In this study, sericin is selected as the bioactive material of interest for surface modification because of its ease of processing compared to fibroin, as well as its relative high solubility in water. Among the different surface modification techniques, surfaceinitiated atom transfer radical polymerization (ATRP) has been increasingly used in biomaterial modifications [9]. The ATRP technique allows polymerization of functional monomers and hence provides high capacity of binding sites (compared to self-assembled monolayers, SAMs). Moreover, the versatility of the technique for

4752

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

functionalization of titanium surfaces has been demonstrated in terms of functionality, flexibility and stability [10]. In the present work, Ti was functionalized via surface-initiated ATRP of methacrylic acid sodium salt (MAAS), followed by the conjugation of sericin to the pendant carboxyl groups on poly(methacrylic acid) (P(MAA)). With this technique, drastically different functions, viz., antibacterial adhesion and enhanced osteoblast functions, were concomitantly imparted on the same functionalized Ti surface. 2. Materials and methods

was then introduced into the reaction mixture. The reaction tube was sealed and kept in a 30  C water bath for a predetermined period of time. After the reaction, the Ti substrate with surface-grafted and partially protonated MAA polymer (denoted as Ti-g-P(MAA) in the subsequent discussion) was removed from the solution and washed thoroughly with an excess amount of doubly distilled water. In the third step, silk sericin was immobilized on the Ti-g-P(MAA) surface via carbodiimide chemistry. The –COOH groups of the Ti-g-P(MAA) surface were preactivated for 1 h at room temperature in phosphate buffered saline (PBS) solution, containing 1 mg/mL of NHS and 10 mg/mL of EDAC. The substrates were then transferred to PBS solution containing 5 mg/mL of silk sericin. The reaction was allowed to proceed for 24 h at 4  C to give the Ti-g-P(MAA-Silk) surface. After the reaction, the reversibly bound silk sericin was desorbed in copious amounts of PBS for 24 h at 4  C.

2.1. Materials

2.4. Surface characterization

Titanium foils (99.6%, 0.05 mm in thickness) were purchased from Goodfellow of Cambridge, UK. Trichloro(4-(chloromethyl)-phenyl)silane (chlorosilane, 97%), sodium methacrylate (methacrylic acid sodium salt, MAAS, 99%), triethylamine (TEA, 98%), copper(I) chloride (99%), copper(II) chloride (97%), N,N,N0 ,N00 ,N00 -pentamethyldiethylenetriamine (PMDETA, 99%), N-hydroxysuccinimide (NHS, 98%), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDAC), p-nitrophenylphosphate (PNPP), p-nitrophenol, ascorbic acid, b-glycerophosphate were obtained from Sigma–Aldrich. All the solvents used were of chromatographic or analytical grade. Water in toluene and TEA was removed by reaction with sodium and calcium hydride, respectively, followed by distillation. All other chemical reagents and solvents were used as-received. Peptone, yeast extract, agar, and beef extract were purchased from Oxoid (Hampshire, UK). S. aureus 25923 (S. aureus), S. epidermidis 12228 (S. epidermidis) and osteoblast cells (MC3T3-E1 subclone 14) were obtained from the American Type Culture Collection (ATCC). Ultra-pure water (>8.2 MU cm, from Millipore Milli-Q system) was used in all experiments.

The chemical composition of the modified titanium surfaces was determined by X-ray photoelectron spectroscopy (XPS). XPS measurements and calculations were performed as described previously [10]. The static water contact angles of the pristine and functionalized titanium surfaces were measured at 20  C and 60% relative humidity, using the sessile drop method with a 3 mL water droplet, in a telescopic goniometer (Rame-Hart model 100-00-(230), manufactured by the Rame-Hart, Inc., Mountain Lakes, NJ). The telescope with a magnification power of 23 was equipped with a protractor of 1 graduation. For each angle reported, four measurements from different surface locations were averaged. The angle reported was accurate to within 3 .

2.2. Preparation of silk sericin The cocoon shells of the silkworm, Bombyx mori (B. mori) were a gift from School of Life Science, Suzhou University (Suzhou, Jiangsu Province, China). The silk sericin was obtained by degumming the cocoon shells with an aqueous solution of 0.3% Na2CO3 (w/v) at 98  C for 1 h. The sericin solution was then dialyzed in cellulose membranes (MW ¼ 3500 g/mol) against deionized water for 3 days by changing the water daily to remove the ions and other impurities. The sericin powder was collected via rotary evaporation and drying under vacuum at room temperature. 2.3. Surface functionalization of titanium The surface modification of titanium comprised 3 steps, as shown in Fig. 1. The first step is the immobilization of the ATRP initiator (chlorosilane in this study) on the Ti surface and the details have been described previously [10]. The resulting surface is denoted as Ti–Cl. The second step is the surface-initiated ATRP of MAAS from the Ti–Cl surface. Two grams of MAAS and 4 mL of doubly distilled water were added to a Pyrex tube. The PMDETA ligand, CuCl and CuCl2 were added in the ratio of [MAAS]:[CuCl]:[CuCl2]:[PMDETA] ¼ 50:1:0.1:1.1. The addition of deactivating Cu(II) complex (CuCl2) is important for the equilibrium between the dormant and the active chains to establish quickly at the initial stage of ATRP [11]. The mixture was stirred and degassed with argon for 20 min. The Ti–Cl substrate (1  3 cm2 piece)

2.5. Cell culture Mouse osteoblast cell line MC3T3-E1 cells were cultured in alpha minimum essential medium (Invitrogen, USA) supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 mg/mL streptomycin. The cells were incubated at 37  C in a humidified atmosphere of 5% CO2 and 95% air; with the growth medium changed every 48 h. The cultured cells were detached by trypsinization (0.25% trypsin–EDTA), suspended in fresh culture medium and used for the designed experiments described below. 2.6. Cell attachment Osteoblast cell attachment on substrates was evaluated by counting the number of attached cells. The substrates were placed into a 24-well plate and seeded with a density of 50,000 cells/cm2. After 4 h of incubation in growth medium, the substrates were rinsed with PBS to remove unattached cells. Adherent cells were then detached from the substrates by trypsin and measured via a hemacytometer. 2.7. Cell proliferation The proliferation of osteoblast cells on substrates was examined by counting the number of cells after 1, 4, and 7 days of culture. The substrates were placed into a 24-well plate and seeded with a density of 5000 cells/cm2. At each predetermined time, the substrates were rinsed with PBS to remove unattached cells. Adherent cells were then detached from the substrates by trypsin and measured via a hemacytometer.

Step 1 OH Ti

+

OH OH

Cl Cl Si Cl

H2 C Cl

Step 2 O

TEA O Si

Ti

C H2

O

Ti-OH

Cl

MAAS CuCl/CuCl2/ PMDETA

Ti-Cl Step 3 O

Ti

H2 H2 CH3 C C C Cl O Si O C n O O H Ti-g-P(MAA)

Silk Sericin NHS/EDAC

Ti

O O Si

H2 CH3 C C Cl O C n O Silk Ti-g-P(MAA-Silk) H2 C

Fig. 1. Schematic diagram illustrating the processes of silanization of the Ti–OH surface to prepare the Ti–Cl surface, surface-initiated ATRP of MAAS chains from the Ti–Cl surface, and immobilization of silk sericin on the Ti–polymer hybrid [TEA, triethylamine; PMDETA, N,N,N0 ,N00 ,N00 -pentamethyldiethylenetriamine; MAAS, methacrylic acid sodium salt; EDAC, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride; NHS, N-hydroxysuccinimide].

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759 2.8. Alkaline phosphatase (ALP) activity assay Cells were seeded onto substrates at a density of 50,000 cells/cm2 in a growth medium supplemented with 50 mg/mL ascorbic acid and 10 mM sodium b-glycerophosphate. At specified time intervals, the cell layers were washed with PBS and scraped off from the surfaces. Cell lysis buffer was added, and after sonication and centrifugation, aliquots of cell lysis solution were collected for the analysis of the ALP activity and the total protein level. ALP activity was determined with respect to the release of p-nitrophenol from PNPP as described previously [12], and expressed as micromoles of p-nitrophenol formation per minute per milligram of total proteins. 2.9. Calcium deposition To investigate whether the functionalization of Ti affects the precipitation of calcium from the growth medium, the substrates were incubated in growth medium supplemented with 50 mg/mL ascorbic acid and 10 mM sodium b-glycerophosphate. The amount of calcium deposited was measured after 3 weeks. Briefly, to dissolve the calcium mineral deposited, the substrates after washing twice with PBS were soaked in 1 mL of 0.5 M acetic acid overnight. The solutions were then centrifuged and the supernatants were collected and tested for calcium content by the

O 1s

a

4753

ortho-cresolphthalein complexone (OCPC) method as reported previously [13]. The calcium concentrations of the samples were determined with a standard curve obtained from known concentrations of CaCl2 solution. 2.10. Antibacterial assay The antibacterial assay comprised two experiments to investigate whether the functionalized surfaces (a) reduced bacterial adhesion and/or (b) killed the bacteria. 2.10.1. Effect on bacterial adhesion Bacterial adhesion characteristics of the functionalized titanium surfaces were assessed via fluorescence microscope imaging and the spread plate method. S. aureus and S. epidermidis were cultured in tryptic soy broth and nutrient broth respectively as recommended by ATCC. The bacteria were incubated overnight at 37  C with agitation in the broth. An aliquot of bacterial culture was then added to the broth and incubated for another 2 h at 37  C. The bacterial culture was centrifuged at 2700 rpm for 10 min. After the removal of the supernatant, the cells were washed twice with PBS and resuspended in PBS at a concentration of 107 cells/mL. The bacterial cell concentration was estimated by measuring the UV absorption of the cell suspension at 540 nm [14]. An optical density of 1.0 (default

b

Ti 2p

Ti-OH C-H C-O/C-Cl

C 1s

0

200

400

600

800

c

1000 282

284

d

286

288

290

C-H

Ti-Cl Intensity (Arb. Units)

C-O/C-Cl C-Si

Si 2p Cl 2p

0

200

400

600

800

e

1000 282

f

284

286

288

C-H

Ti-g-P(MAA) Na 1s C-O/C-Cl

0

g

200

400

600

800

1000 282

h

284

Ti-g-P(MAA-Silk) C-H

200

400

286

288

O=C-O

290

C-O/C-Cl O=C-N O=C-O

N 1s

0

290

C-N

600

800 1000 282 284 Binding Energy (eV)

286

288

290

Fig. 2. XPS wide scan and C 1s core-level spectra of (a, b) the pristine or Ti–OH surface, (c, d) the Ti–Cl surface, (e, f) the Ti-g-P(MAA) surface after 3 h of surface-initiated ATRP of MAAS and (g, h) the corresponding Ti-g-P(MAA-Silk) surface.

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

units) for the bacterial suspension at 540 nm is approximately equivalent to 109 cells/mL. 1 mL of the bacterial suspension was then added to each substrate in a 24-well plate and incubated for 5 h at 37  C. The surfaces were then washed with PBS solution to remove the loosely adsorbed bacteria. The viability of the bacteria adhered on the substrates was investigated by staining with 0.1 mL solution of a LIVE/DEAD Baclight bacterial viability kit (Molecular Probes, L13152, Oregon, US), and subsequently imaged with a Leica DMLM microscope equipped with a 100 W Hg lamp. The solution of the kit consists of propidium iodide (PI) and SYTO 9, which can diffuse into the bacterial cells and fluoresce upon binding to nucleic acids [15]. SYTO 9 is membrane permeable and therefore stains both viable and non-viable bacteria (green fluorescence), whereas PI which has a higher affinity for nucleic acids (red fluorescence) is rejected from viable bacterial cells by membrane pumps [15]. Thus, under the fluorescence microscope, viable bacteria (appear green) and dead bacteria (appear red) can be distinguished. Quantification of bacterial adhesion was carried out as follows: after incubation in bacterial suspension for 5 h (as described above), the substrates were removed with sterile forceps and gently washed with PBS. The substrates were then placed in broth and the bacteria retained on substrates were dislodged by mild ultrasonication for 2 min in a 100 W ultrasonic bath operating at a nominal frequency of 50 Hz, followed by rapid vortex mixing (10 s). Serial ten-fold dilutions were performed and viable counts were estimated using the spread plate method. The number of bacteria on each substrate surface was counted and expressed relative to the surface area of the substrate (number of bacteria/cm2). 2.10.2. Bactericidal properties To determine whether the functionalized Ti surface possessed bactericidal properties, the pristine and functionalized Ti substrates were immersed in S. aureus and S. epidermidis bacterial suspensions in PBS at a concentration of 107 cells/mL. After 5 h of incubation and shaking, bacteria retained on the substrates were dislodged by mild ultrasonication for 2 min in a 100 W ultrasonic bath operating at a nominal frequency of 50 Hz followed by rapid vortex mixing (10 s). The numbers of viable bacteria in the suspension were measured by the spread plate method, as described above. The numbers of bacteria were counted and normalized by the number obtained with the pristine Ti control surface. A decrease in the number of viable bacterial cells in the suspension would indicate that these cells were killed upon contact with the substrates. 2.11. Statistical analysis In each cell and bacterial experimental run, three samples per time point for each experimental condition were used. The results were expressed as mean values  standard deviation and were assessed statistically using one-way analysis of variance (ANOVA) post-hoc Tukey test. The difference observed between samples was considered statistically significant when p < 0.05.

3. Results and discussion 3.1. Surface modification and characterization The wide scan and the corresponding C 1s core-level spectra of the pristine(hydroxide-covered) titanium (Ti–OH), the trichloro(4-(chloromethyl)-phenyl) silane-coupled titanium (Ti–Cl), the poly(methacrylic acid) (P(MAA))-grafted titanium (Ti-gP(MAA)) and silk-coupled titanium (Ti-g-P(MAA-Silk)) surfaces are shown in Fig. 2. The wide-scan spectrum of the pristine titanium surface (Fig. 2(a)) is dominated by signals of titanium (Ti 2p at binding energy (BE) of about 460 eV [16]) and oxygen (O 1s at BE of about 532 eV), attributable to the oxide film present on titanium surfaces exposed to air. A weak carbon signal (C 1s at about 285 eV) is also discernible, probably due to hydrocarbon contamination of the sample surface exposed to the ambient atmosphere prior to XPS measurements. For the preparation of polymer chains on the Ti surface, a uniform layer of initiator was immobilized on the titanium surface via the triethylamine (TEA)-catalyzed chlorosilanization. The appearance of the Si 2p (at about 100 eV) and Cl 2p (at about 200 eV) signals, the decrease in the Ti signal intensity, and the increase in the C signal intensity in the wide scan (Fig. 2(c)), and the nature of the C 1s core-level spectra (Fig. 2(d)) of the Ti–Cl surface are consistent with the presence of a coupled chlorosilane layer on the Ti surface. Thus, the benzyl chloride groups have been successfully immobilized on the Ti surface for the subsequent ATRP from the Ti–Cl surface.

3.2. Surface-initiated ATRP from the Ti–Cl surface Since the carboxylic acid functional groups may poison the Cu ATRP catalyst, controlled polymerization of (meth)acrylic acid directly by the ATRP technique is a challenging problem [17]. Ashford et al. reported that the pH value is critical for successful ATRP of MAAS, and found that aqueous ATRP of MAAS resulted in controlled molecular weights and narrow molecular weight distributions [18]. Osborne and coworkers have also prepared triblock copolymer brushes containing a P(MAA) block by aqueous ATRP of MAAS [19]. Thus, in this work, MAAS polymer was immobilized on Ti substrates via surface-initiated ATRP as shown schematically in Fig. 1. Fig. 2(e) shows the wide-scan spectrum of the Ti-g-P(MAA) surface from 3 h of ATRP. The Ti surface was fully covered by the polymer chains to a thickness greater than the sampling depth of the XPS technique (about 7.5 nm in an organic matrix), since the Ti 2p peak component at BE of about 460 eV was no longer discernible in the wide-scan spectrum. The C 1s core-level spectrum of the Ti-g-P(MAA) surface in Fig. 2(f) can be curvefitted into three peak components with BEs at 284.6, 286.2, and 288.5 eV, attributable to the C–H, C–O (and C–Cl at the chain ends), and O]C–O species, respectively. The appearance of the O]C–O species shows the successful surface-initiated ATRP of MAAS on Ti. 3.3. Immobilization of silk sericin on the Ti-g-P(MAA) surfaces The C 1s core-level spectrum of the Ti-g-P(MAA-Silk) surface can be curve-fitted into five peak components with BEs at 284.6, 285.6, 286.2, 287.6 and 288.5 eV, attributable to the C–H, C–N, C–O/C–Cl, O]C–N and O]C–O species, respectively (Fig. 2(h)). In comparison with the C 1s core-level spectrum of the starting Ti-g-P(MAA) surface (Fig. 2(f)), the decrease in O]C–O peak component and the appearance of O]C–N peak component confirm the reaction of the carboxyl groups of P(MAA) with the amine groups of silk sericin. The O]C–N peak component is associated with the linkage formed between the –COOH group of the P(MAA) and –NH2 of the silk sericin molecules, as well as the peptide linkage in the sericin protein itself. The covalent immobilization of silk sericin is further supported by the appearance of the N 1s signal in the wide-scan spectrum (Fig. 2(g)) of the Ti-g-P(MAA-Silk) surface.

Adhesion of Cells 160

*

140 120 % of Control

4754

100 80 60

*

40 20 0 Ti

Ti-g-P(MAA)

Ti-g-P(MAA-Silk)

Fig. 3. Comparison of osteoblast attachment on surfaces of pristine and functionalized Ti substrates seeded with 50,000 cells/cm2. The number of cells per cm2 was measured and expressed as percentage of the number obtained in the control experiment using polystyrene cell culture surface. (*) Denotes significant differences (p < 0.05) compared with the pristine Ti.

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

4755

3.4. Osteoblast cell attachment and growth

ALP Activity

6 5 µM/min/mg protein

The quantification of cell attachment on the pristine Ti, Ti-gP(MAA) and Ti-g-P(MAA-Silk) surfaces (expressed as a percentage of the number obtained in the control experiment using polystyrene cell culture surface) is shown in Fig. 3. The results revealed that while the number of cells on the Ti-g-P(MAA) substrates was significantly lower than that on the pristine Ti, the number of cells on the Ti-g-P(MAA-Silk) substrates was significantly higher than that on the pristine Ti. The proliferation properties of osteoblast on the pristine and silk sericin-immobilized Ti surfaces, after 1, 4, and 7 days culture were compared, as shown in Fig. 4. For Ti-g-P(MAASilk), enhanced osteoblast proliferation was observed throughout the test period, and by the 7th day, 50% increase in cells was observed on the Ti-g-P(MAA-Silk) surface over that on Ti surface. In contrast, the number of cells on the Ti-g-P(MAA) surface was less than that on Ti surface. The mean of the relative cell concentration on the Ti-g-P(MAA-Silk) surface decreased with time, indicating that the positive effect of the Ti-g-P(MAA-Silk) surface over the control surface decreases with time, which probably is a result of the cells reaching confluency. As shown from the XPS results, the Ti-g-P(MAA) surface is wellcovered with covalently immobilized P(MAA). The contact angle of the Ti-g-P(MAA) surface was measured to be 8 , which is significantly lower than that of the pristine Ti surface (53 ). The inhibitive effects of P(MAA) surface on mammalian cells have not received much attention. The hydrophilicity of the Ti-g-P(MAA) surface may not be the only factor contributing to the inhibition of cell adhesion. One study [20] found that implanted surfaces grafted with anionically charged poly(acrylic acid) P(AA) (with a water contact angle of 26 ) decreased monocyte and macrophage adhesion and fusion in vivo (relative to the unmodified poly(benzyl N,N-diethyldithiocarbamate-co-styrene) surface, which has a water contact angle of 83 ), therefore reducing the interfacial foreign body response. Rubner et al. [21] studied surface–cell interactions on hydrogenbonded multilayers comprising polyacrylamide (P(AAm)) and a weak polyelectrolyte, such as P(MAA) or P(AA). Both P(MAA)/ P(AAm) and P(AA)/P(AAm) multilayers exhibited no cytotoxicity, but a high adhesion resistance towards mammalian fibroblasts. Another group [22] has reported the synergic inhibitive effect of carboxylate and sulfonate groups on mammalian cells adhesion. They copolymerized different concentrations of MAA and sodium

7 days 14 days 21 days *

3 2 1 0 Ti

Ca Deposition *

10 *

100

7

80

6

*

*

40

*

9 8

*

Ti-g-P(MAA-Silk)

styrene sulfonate (NaSS) with methyl methacrylate (MMA) (MMA content more than 75% for all the studies, in order to keep the required physical and mechanical properties for intraocular lens application), and investigated the adhesion and proliferation of fibroblast cells. They concluded that the observed inhibitory effect is due to the chemical composition of the surface exposed to binding proteins and cells, more than to its hydrophilic properties. The water contact angle of the surfaces was not reported, but the authors assumed that carboxylate and sulfonate groups in the copolymers contribute equally to the hydrophilicity of the surfaces. Since P(MAA) inhibits cell adhesion, the promotion of osteoblast adhesion and proliferation by the Ti-g-P(MAA-Silk) surface are due to silk sericin. The contact angle of the Ti-g-P(MAA-Silk) surface was measured to be 35 , which is still lower than that of the pristine Ti surface. Hence, the positive effect of immobilized silk sericin on osteoblast adhesion and proliferation is attributable more to its bioactive components than its hydrophobicity. Sericin protein is made up of 18 amino acids, with aspartic acid and serine being the dominant components [23]. Sericin protein associated with the native silk fibers has been earlier implicated in immune response

120

60

Ti-g-P(MAA)

Fig. 5. ALP activity of osteoblasts seeded at a density of 50,000 cells/cm2 on pristine and functionalized Ti substrates after 7, 14 and 21 days. (*) Denotes significant differences (p < 0.05) compared with the pristine Ti.

g/cm2

% of Control

140

*

1 day 4 days 7 days

*

4

Proliferation of Cells 160

*

*

5 4 3

20

2

0 Ti

Ti-g-P(MAA) Ti-g-P(MAA-Silk)

Fig. 4. Comparison of osteoblast proliferation with 5000 cells/cm2 seeded on surfaces of pristine and functionalized Ti substrates after 1, 4 and 7 days. The number of cells per cm2 was measured and expressed as percentage of the number obtained in the control experiment using polystyrene cell culture surface. (*) Denotes significant differences (p < 0.05) compared with the pristine Ti.

1 0 Ti

Ti-g-P(MAA)

Ti-g-P(MAA-Silk)

Fig. 6. Calcium deposition on pristine and functionalized Ti substrates after incubation in culture medium for 21 days. (*) Denotes significant differences (p < 0.05) compared with the pristine Ti.

4756

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

a

Pristine Ti

b

c

Ti-g-P(MAA)

d

e

Ti-g-P(MAA-Silk)

f

50 µm

Fig. 7. Fluorescence microscopic images of (a) and (b) pristine Ti, (c) and (d) Ti-g-P(MAA), (e) and (f) Ti-g-P(MAA-Silk) surfaces after exposure to a suspension of S. aureus (107 cells/ mL) for 5 h.

against the fiber. However, a subsequent study has shown that there is a lack of macrophage response to soluble sericin, and sericin-coated fibroin fibers induced a stronger macrophage response to lipopolysaccharide than sericin-coated tissue culture plate [24]. The authors concluded that sericin-mediated activation of the macrophages is dependent on its physical association with the fibroin resulting in a conformational change or better adhesion by the macrophages. In vivo experiments have also shown that sericin does not have significant immunogenicity [25,26]. It has also been reported that wounds of rats treated with 8% sericin-containing cream (formulated using petrolatum ointment as base) had much smaller inflammatory reactions, and woundsize reduction was much greater than in the controls which were treated with the cream base without sericin, throughout the inspection period [27]. Tsubouchi et al. [28] showed that sericin enhances the attachment of primary cultured human fibroblasts and the acceleration of proliferation of several mammalian cells by sericin was also reported by Terada et al. [29]. As proliferation of osteoblasts is critical for the formation of hard tissue, the above results suggest that sericin immobilized on MAA brushes would be helpful for improving osteoblasts’ growth and tissue formation. A

study on the growth and attachment of fibroblasts on silk suggested that the improved cell attachment on sericin may not be due to biospecific interaction of a receptor, but could have resulted from the electrostatic interactions between the cells and the substrate [30]. 3.5. Cell differentiation ALP is widely used as an assay for osteoblastic activity, especially osteoblastic differentiation. ALP activities after 7, 14 and 21 days of cell culture on the different substrates are shown in Fig. 5. From this figure, it can be seen that osteoblasts on Ti-g-P(MAA-Silk) showed significantly higher ALP activity than those on pristine Ti. In particular, ALP activity of osteoblasts cultivated after 14 days reached a maximum on Ti-g-P(MAA-Silk), and is twofold higher than those on the pristine Ti. Since the ALP activity is normalized by the total intracellular proteins synthesized, the values have taken account of the difference in cell numbers among the samples. The results show that osteoblasts on Ti-g-P(MAA) surface have no significant ALP activity increase over the pristine Ti surface. Hence, these results reveal that the silk sericin immobilized on the

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

a

4757

Pristine Ti

b

c

Ti-g-P(MAA)

d

e

Ti-g-P(MAA-Silk)

f

50 µm

Fig. 8. Fluorescence microscopic images of (a) and (b) pristine Ti, (c) and (d) Ti-g-P(MAA), (e) and (f) Ti-g-P(MAA-Silk) surfaces after exposure to a suspension of S. epidermidis (107 cells/mL) for 5 h. (The full colour images can be found in the on-line version.)

Ti-g-P(MAA-Silk) surface can improve ALP activity of osteoblasts. A similar effect has been reported with silk fibroin [31,32], although the mechanism is still unclear. 3.6. Ca deposition The deposition of calcium phosphate based materials on titanium implant devices is now generally regarded to be advantageous to enhance bone formation around the implants, to contribute to bone fixation and integration, and thus to increase clinical success of the implantation [33]. The extent of Ca deposition on the functionalized surfaces, after incubation in culture medium for 21 days, is shown in Fig. 6. It has been widely reported that hydroxyapatite is formed on Ti, especially its hydrophilic derivative surfaces, when they are immersed in simulated body fluid (SBF) at 37  C [33]. The results shown in Fig. 6 reveal that upon immersion in culture medium, the surface chemistry of Ti-g-P(MAA) and Ti-g-P(MAA-Silk) layers significantly increases the rate of the spontaneous Ca mineralization over that of the pristine Ti surface. Because of the high rate of spontaneous Ca mineralization by the functionalized surfaces, it

would not be possible to accurately quantify the production of Ca by osteoblasts seeded on these surfaces. This is consistent with previous similar findings of Taguchi et al. [34], who compared the Ca deposition by osteoblast-encapsulated type II collagen gels and the cell-free gels. They found that Ca deposition did not occur in Cafree media, despite cell encapsulation. However, when the gels were soaked with 5 mM of Ca2þ, Ca deposition occurred regardless of cell encapsulation, although the rate of Ca deposition was slightly higher in cell-encapsulated gels than that in cell-free gels. The Ca deposition results (Fig. 6) were obtained without cells to show that P(MAA) enhances spontaneous Ca deposition from the culture medium. Thus, while the Ti-g-P(MAA) and Ti-g-P(MAA-Silk) surfaces have similar Ca deposition, the enhanced osteoblast adhesion, proliferation and ALP activity of the latter would indicate higher potential for osseointegration in vivo. 3.7. Antibacterial properties The distribution of viable and dead bacteria on the surface of the pristine and functionalized Ti substrates, after immersion in a S. aureus and S. epidermidis suspension of 107 cells/mL for 5 h at

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

37  C, was observed via staining with a combination dye (shown in Figs. 7 and 8, respectively). The images reveal that there is a clear distinction between the number of S. aureus or S. epidermidis cells on the pristine Ti and that on the functionalized substrates. Substantially more viable bacteria (stained green) were attached on the pristine Ti surface (Figs. 7(a) and 8(a)) in comparison with Ti-gP(MAA) and Ti-g-P(MAA-Silk) (Figs. 7(c) and 8(c), 7(e) and 8(e), respectively). (For interpretation of the references to colour in this text, the reader is referred to the web version of this article.) Since the fluorescence images can only provide qualitative information about the effect of the functionalized surfaces on bacterial adhesion, the viable bacterial cells on different substrates were also counted after detachment from the surfaces. The quantitatively determined densities of bacteria on the different substrates after 5 h cultivation (Fig. 9) showed good agreement with the qualitative trend observed in the fluorescence microscopic investigations (Figs. 7 and 8). The results show that the number of viable S. aureus cells on the Ti-g-P(MAA) and Ti-g-P(MAA-Silk) substrates has decreased by about threefold from those on the pristine Ti. The decrease in the number of S. epidermidis cells on the Ti-g-P(MAA) and Ti-g-P(MAA-Silk) substrates is also significant (Fig. 9). The decrease in bacterial cells on the Ti-g-P(MAA) and Ti-gP(MAA-Silk) surfaces can be attributed to the anti-adhesive nature of P(MAA) and sericin, as discussed below. As there is a growing need for alternative approaches to circumvent the limitations of antibiotic therapies, two approaches have been developed: materials with antimicrobial properties and materials with anti-adhesive properties. The first approach involves materials containing bactericidal substances that are incorporated into or bound to the biomaterial surfaces [35,36]. The other approach is the modification of substrate materials with the anti-adhesive polymers to prevent the adhesion of bacteria, thus, inhibiting biofilm formation and infection [37]. To determine whether the decrease in bacteria on Ti-g-P(MAA) and Ti-g-P(MAA-Silk) surfaces resulted from the bactericidal or anti-adhesive effect of the surfaces, the number of viable bacteria in suspension after contacting with the surfaces was counted. After 5 h of shaking and incubation, the number of viable bacteria in the suspension was determined by the spread plate method, and expressed relative to that obtained with the pristine Ti surface (see Fig. 10). It can be seen from Fig. 10 that there is no statistically significant difference in the number of both S. aureus and S. epidermidis cells when the different surfaces were used (p > 0.05). Therefore, the reduction in bacteria count on the Ti-g-P(MAA) and Ti-g-P(MAA-Silk) surfaces reported in Fig. 9 is due to the

Number of bacteria per cm2

4x105 S. aureus S. epidermidis 3x105

2x105 * 1x105

*

*

*

0 Ti

Ti-g-P(MAA) Ti-g-P(MAA-Silk)

Fig. 9. Number of adherent of S. aureus and S. epidermidis cells per cm2 on surfaces of pristine and functionalized Ti substrates. (*) Denotes significant differences (p < 0.05) compared with the pristine Ti.

Relative bacteria number %

4758

140

S. aureus S. epidermidis

120 100 80 60 40 20 0 Ti

Ti-g-P(MAA) Ti-g-P(MAA-Silk)

Fig. 10. Number of S. aureus and S. epidermidis viable cells in suspension after contacting with pristine and functionalized Ti substrates for 5 h. The number of cells was expressed relative to that after contacting with the pristine Ti. No significant statistical difference among the surfaces was observed.

anti-adhesive properties of the surfaces, rather than bactericidal properties. In the present study, the P(MAA) brushes behave as the antiadhesive polymer, and play an important part in reduction of bacterial adhesion on the Ti-g-P(MAA). As early as 1991, Harkes et al. [38] found that polymers from copolymerization of methyl methacrylate (MMA) and methacrylic acid (MAA, 15% molar ratio) can significantly inhibit bacterial adhesion compared with those of the MMA homopolymers. They also found that higher numbers of bacteria adhered to positively charged surfaces than on uncharged (hydrophilic) and negatively charged substrates. Thus, they concluded that electrostatic interactions played an important role in the adhesion process. Yang et al. [39] later studied the number of viable bacterial cells after coming into contact with P(AA) grafted polypropylene nonwoven fabric. They found that the viable cell number decreased with increasing acrylic acid grafting content; hence they believed that the increased hydrophilicity of the modified nonwoven fabric enhanced the antibacterial activity. Recently, Anagnostou et al. investigated the effect of MAA and sodium styrene sulfonate (NaSS) on S. aureus adhesion and osteoblast functions of P(MMA)-based polymers [40]. They reported significant synergic effect of P(MAA) and P(NaSS) on inhibition of S. aureus adhesion. It should be mentioned that the difference between the number of bacteria on Ti-g-P(MAA) and that on Ti-gP(MAA-Silk) is not significant (p > 0.05) (Fig. 9) although the contact angle of the Ti-g-P(MAA-Silk) surface (35 ) is higher than that of Ti-g-P(MAA) (8 ). The results thus show that with the immobilized sericin on the P(MAA) chains, the anti-adhesive properties of P(MAA) brushes towards bacteria are still retained. It should be noted that, on one hand, the P(MAA) brushes act as a linker for immobilization of silk sericin, and on the other hand, the brushes can effectively inhibit the bacterial adhesion. It is also interesting that the Ti-g-P(MAA-Silk) substrates interact differently with osteoblast cells and bacteria, i.e. adhesive to osteoblast cells but inhibiting bacterial adhesion, a property which is considered to be advantageous for surfaces of implant materials. 4. Conclusions Titanium substrate surfaces were modified via a simple route: immobilization of ATRP initiator on the titanium surface, surfaceinitiated ATRP of MAAS, followed by the attachment of silk sericin on the poly(MAA) brushes. Significant inhibition of S. aureus and S. epidermidis adhesion on the P(MAA) and Ti-g-P(MAA-Silk) surfaces was achieved. On the other hand, enhanced proliferation

F. Zhang et al. / Biomaterials 29 (2008) 4751–4759

and growth of osteoblasts were achieved simultaneously on the Ti-g-P(MAA-Silk) surfaces. The results showed that the desirable properties of both the P(MAA) brushes and the immobilized silk sericin protein could be combined, preserved, and imparted on the Ti surfaces, that is, the P(MAA) brushes serve to inhibit bacterial adhesion while the silk sericin protein enhances osteoblast attachment, proliferation, and ALP activity. Thus, such P(MAA) and silk sericin functionalized surfaces prepared via surface-initiated ATRP possess the two important functions simultaneously, and are useful for the fabrication of titanium-based biomedical devices for prevention of biofilm-related infections and enhancement of tissue integration. Acknowledgements The authors gratefully acknowledge the financial support from the Ministry of Education under the National University of Singapore Grant No.: R279000202112. References [1] Davies JE. Bone engineering. Toronto: Em Squared Incorporated; 2000. [2] Brunette DM, Tengvall P, Textor M, Thomsen P. Titanium in medicine: material science, surface science, engineering, biological responses, and medical applications. Berlin; New York: Springer; 2001. [3] Song WH, Jun YK, Han Y, Hong SH. Biomimetic apatite coatings on micro-arc oxidized titania. Biomaterials 2004;25:3341–9. [4] Montanaro L, Campoccia D, Arciola CR. Advancements in molecular epidemiology of implant infections and future perspectives. Biomaterials 2007;28:5155–68. [5] Campoccia D, Montanaro L, Arciola CR. The significance of infection related to orthopedic devices and issues of antibiotic resistance. Biomaterials 2006;27:2331–9. [6] Donlan RM, Costerton JW. Biofilms: survival mechanisms of clinically relevant microorganisms. Clin Microbiol Rev 2002;15:167–93. [7] Vepari C, Kaplan DL. Silk as a biomaterial. Prog Polym Sci 2007;32:991–1007. [8] Wang YZ, Kim HJ, Vunjak-Novakovic G, Kaplan DL. Stem cell-based tissue engineering with silk biomaterials. Biomaterials 2006;27:6064–82. [9] Dalsin JL, Hu BH, Lee BP, Messersmith PB. Mussel adhesive protein mimetic polymers for the preparation of nonfouling surfaces. J Am Chem Soc 2003;125:4253–8. [10] Zhang F, Shi ZL, Chua PH, Kang ET, Neoh KG. Functionalization of titanium surfaces via controlled living radical polymerization: from antibacterial surface to surface for osteoblast adhesion. Ind Eng Chem Res 2007;46:9077–86. [11] Dong R, Krishnan S, Baird BA, Lindau M, Ober CK. Patterned biofunctional poly(acrylic acid) brushes on silicon surfaces. Biomacromolecules 2007;8: 3082–92. [12] Faghihi S, Azari F, Li H, Bateni MR, Szpunar JA, Vali H, et al. The significance of crystallographic texture of titanium alloy substrates on pre-osteoblast responses. Biomaterials 2006;27:3532–9. [13] van den Dolder J, de Ruijter AJ, Spauwen PH, Jansen JA. Observations on the effect of BMP-2 on rat bone marrow cells cultured on titanium substrates of different roughness. Biomaterials 2003;24:1853–60. [14] Hogt AH, Dankert J, Feijen J. Adhesion of coagulase-negative Staphylococci to methacrylate polymers and copolymers. J Biomed Mater Res 1986;20:533–45. [15] Gray JE, Norton PR, Alnouno R, Marolda CL, Valvano MA, Griffiths K. Biological efficacy of electroless-deposited silver on plasma activated polyurethane. Biomaterials 2003;24:2759–65. [16] Moulder JF, Chastain J. Handbook of X-ray photoelectron spectroscopy: a reference book of standard spectra for identification and interpretation of XPS data. Eden Prairie, Minn: Perkin–Elmer Corp.; 1992.

4759

[17] Matyjaszewski K, Xia JH. Atom transfer radical polymerization. Chem Rev 2001;101:2921–90. [18] Ashford EJ, Naldi V, O’Dell R, Billingham NC, Armes SP. First example of the atom transfer radical polymerisation of an acidic monomer: direct synthesis of methacrylic acid copolymers in aqueous media. Chem Commun 1999:1285–6. [19] Osborne VL, Jones DM, Huck WTS. Controlled growth of triblock polyelectrolyte brushes. Chem Commun 2002:1838–9. [20] Brodbeck WG, Patel J, Voskerician G, Christenson E, Shive MS, Nakayama Y, et al. Biomaterial adherent macrophage apoptosis is increased by hydrophilic and anionic substrates in vivo. Proc Natl Acad Sci U S A 2002;99:10287–92. [21] Yang SY, Mendelsohn JD, Rubner MF. New class of ultrathin, highly celladhesion-resistant polyelectrolyte multilayers with micropatterning capabilities. Biomacromolecules 2003;4:987–94. [22] El Khadali F, Helary G, Pavon-Djavid G, Migonney V. Modulating fibroblast cell proliferation with functionalized poly(methyl methacrylate) based copolymers: chemical composition and monomer distribution effect. Biomacromolecules 2002;3:51–6. [23] Zhang YQ. Applications of natural silk protein sericin in biomaterials. Biotechnol Adv 2002;20:91–100. [24] Panilaitis B, Altman GH, Chen JS, Jin HJ, Karageorgiou V, Kaplan DL. Macrophage responses to silk. Biomaterials 2003;24:3079–85. [25] Zhang YQ, Ma Y, Xia YY, Shen WD, Mao JP, Xue RY. Silk sericin–insulin bioconjugates: synthesis, characterization and biological activity. J Controlled Release 2006;115:307–15. [26] Zhang YQ, Tao ML, Shen WD, Mao JP, Chen YH. Synthesis of silk sericin peptides-L-asparaginase bioconjugates and their characterization. J Chem Technol Biotechnol 2006;81:136–45. [27] Aramwit P, Sangcakul A. The effects of sericin cream on wound healing in rats. Biosci Biotechnol Biochem 2007;71:2473–7. [28] Tsubouchi K, Igarashi Y, Takasu Y, Yamada H. Sericin enhances attachment of cultured human skin fibroblasts. Biosci Biotechnol Biochem 2005;69:403–5. [29] Terada S, Nishimura T, Sasaki M, Yamada H, Miki M. Sericin, a protein derived from silkworms, accelerates the proliferation of several mammalian cell lines including a hybridoma. Cytotechnology 2002;40:3–12. [30] Minoura N, Aiba S, Gotoh Y, Tsukada M, Imai Y. Attachment and growth of cultured fibroblast cells on silk protein matrices. J Biomed Mater Res 1995;29:1215–21. [31] Cai K, Hu Y, Jandt KD. Surface engineering of titanium thin films with silk fibroin via layer-by-layer technique and its effects on osteoblast growth behavior. J Biomed Mater Res Part A 2007;82A:927–35. [32] Cai KY, Yao KD, Lin SB, Yang ZM, Li XQ, Xie HQ, et al. Poly(D,L-lactic acid) surfaces modified by silk fibroin: effects on the culture of osteoblast in vitro. Biomaterials 2002;23:1153–60. [33] Narayanan R, Seshadri SK, Kwon TY, Kim KH. Calcium phosphate-based coatings on titanium and its alloys. J Biomed Mater Res Part B 2008;85:279–99. [34] Maeno S, Niki Y, Matsumoto H, Morioka H, Yatabe T, Funayama A, et al. The effect of calcium ion concentration on osteoblast viability, proliferation and differentiation in monolayer and 3D culture. Biomaterials 2005;26:4847–55. [35] Chua PH, Neoh KG, Kang ET, Wang W. Surface functionalization of titanium with hyaluronic acid/chitosan polyelectrolyte multilayers and RGD for promoting osteoblast functions and inhibiting bacterial adhesion. Biomaterials 2008;29:1412–21. [36] Shi ZL, Neoh KG, Kang ET. Antibacterial and adsorption characteristics of activated carbon functionalized with quaternary ammonium moieties. Ind Eng Chem Res 2007;46:439–45. [37] Kingshott P, Wei J, Bagge-Ravn D, Gadegaard N, Gram L. Covalent attachment of poly(ethylene glycol) to surfaces, critical for reducing bacterial adhesion. Langmuir 2003;19:6912–21. [38] Harkes G, Feijen J, Dankert J. Adhesion of Escherichia coli on to a series of poly(methacrylates) differing in charge and hydrophobicity. Biomaterials 1991;12:853–60. [39] Yang JM, Lin HT, Wu TH, Chen CC. Wettability and antibacterial assessment of chitosan containing radiation-induced graft nonwoven fabric of polypropylene-g-acrylic acid. J Appl Polym Sci 2003;90:1331–6. [40] Anagnostou F, Debet A, Pavon-Djavid G, Goudaby Z, Helary G, Migonney V. Osteoblast functions on functionalized PMMA-based polymers exhibiting Staphylococcus aureus adhesion inhibition. Biomaterials 2006;27:3912–9.