The effect of immobilization of heparin and bone morphogenic

lable at ScienceDirect

Biomaterials xxx (2010) 1e8

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

The effect of immobilization of heparin and bone morphogenic protein-2 (BMP-2)to titanium surfaces on inflammation and osteoblast function

Sung Eun Kim a, Sang-Hun Song b, Young Pil Yun a, Byung-Joon Choi b, Il Keun Kwon c, Min Soo Bae c,Ho-Jin Moon c, Yong-Dae Kwon b,*

aDepartment of Orthopedic Surgery and Rare Diseases Institute, Korea University Medical College, Guro Hospital, #80, Guro-dong, Guro-gu, Seoul, 152-703 KoreabDepartment of Oral and Maxillofacial Surgery, School of Dentistry, Kyung Hee University, 1 Hoegi-dong, Dongdaemun-gu, Seoul, 130-701 KoreacDepartment of Maxillofacial Biomedical Engineering and Institute of Oral Biology, School of Dentistry, Kyung Hee University, 1 Hoegi-dong, Dongdaemun-gu, Seoul,130-701 Korea

a r t i c l e i n f o

Article history:Received 6 July 2010Accepted 3 September 2010Available online xxx

Keywords:TitaniumHeparinBone morphogenic protein-2 (BMP-2)OsteoblastsInflammationPeri-implantitis

* Corresponding author. Tel.: þ82 2 958 9440; fax:E-mail address: [email protected] (Y.-D. K

0142-9612/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.biomaterials.2010.09.008

Please cite this article in press as: Kim SEBiomaterials (2010), doi:10.1016/j.biomateri

a b s t r a c t

The aim of this study was to investigate biologic function of bone morphorgenic protein-2 (rhBMP-2)immobilized on the heparin-grafted Ti surface. Ti surfaces were first modified by 3-amino-propyltriethoxysilane (ATPES), followed by grafting of heparin. BMP-2 was then immobilized on theheparin-grafted Ti surfaces. Pristine Ti and functionalized Ti surfaces were characterized by X-rayphotoelectron spectroscopy (XPS), measurement of water contact angles, and protein adsorption. Thebiological activity of MG-63 cells on pristine and functionalized Ti surfaces was investigated by cellproliferation assays, measurement of alkaline phosphate (ALP) activity, and determination of calciumdeposition. Anti-inflammatory effects were assessed by RT-PCR to measure the transcript levels of IL-6and TNF-a. XPS revealed that heparin and BMP-2 were successfully grafted and immobilized on the Tisurfaces, respectively. In addition, Ti surfaces with BMP-2 immobilized were more hydrophilic thanpristine Ti. Furthermore, BMP-2 immobilized Ti promoted significantly higher ALP activity and calciumdeposition by MG-63 cells than pristine Ti. The inflammatory response was also decreased when cellswere grown on heparin-grafted, BMP-2-immobilized Ti surfaces. The results of this study suggest that bygrafting heparin and immobilizing BMP-2 on Ti surfaces, inflammation can be inhibited and osteoblastfunction promoted.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Titanium (Ti) is widely used in the field of orthopedic and dentalimplants because of its good biocompatibility, superior mechanicalproperties, and excellent corrosion resistance. Because titaniumprevents direct boneetoebone contact, a variety of surface modi-fications have been developed to facilitate bone tissue responsestoward titanium surfaces including hydroxyapatite coating,biomolecular immobilization, and control of the surface topog-raphy [1e3]. In recent years, another approach to stimulate boneetitanium interactions has been developed; this approach involvesimmobilizing peptides, such as ArgeGlyeAsp (RGD) and lysineearginineeserineearginine (KRSR), on the Ti surface [4e6].However, the adhesion layer between the titanium and thesepeptides is weak and tends to crack. To solve these shortcomings,

þ82 2 966 4572.won).

All rights reserved.

, et al., The effect of immobals.2010.09.008

another approach involving surface modification with bonemorphogenic proteins (BMPs) has been developed. BMPs playimportant roles in bone and cartilage regeneration. Among BMPs,BMP-2 has very strong osteoinductive activity. Previous studieshave shown that BMP-2 can induce the osteogenic differentiationof mesenchymal cells and de novo orthotopic or ectopic boneformation [7e10]. Although BMP-2 has successfully been used tostimulate bone regeneration, use of BMP-2 is expensive, a high doseis required (1 mg BMP-2/mL defect), and BMP-2 has a short half-lifein vivo [11]. To overcome these problems, BMP-2 delivery systemssuch as collagen gels, sponges, scaffolds, hyaluronic acid, and fibringels for prolonged, local release of BMP-2 have been studied[12e19]. However, these systems have problems such as uncon-trolled release rates, release of BMP-2 for only a short period, anda high initial burst of release [20]. In this study, we used heparin tocontrol the release of BMP-2. Heparin, a highly sulfated and linernatural polysaccharide, has been shown to have binding affinitiesto various growth factors such as vascular endothelial growth factor(VEGF), basic fibroblast growth factor (bFGF), and transforming

ilization of heparin and bone morphogenic protein-2 (BMP-2) to...,

mailto:[email protected]

www.sciencedirect.com/science/journal/01429612

http://www.elsevier.com/locate/biomaterials

http://dx.doi.org/10.1016/j.biomaterials.2010.09.008



S.E. Kim et al. / Biomaterials xxx (2010) 1e82

growth factor-b (TGF-b) [21]. Moreover, biomaterial systems withheparin have been shown to have advantages for the controlledrelease of these growth factors [22,23]. However, inflammatoryresponses from the patient’s own skin and/or mucosa and theimplant materials sometimes occur during surgical insertion. Withthe popularization of dental implants, the incidence of peri-implantdisease is now a growing problem causing local bone destructionresulting in failure of the implants. For the above reasons, implantsurfaces should have both anti-inflammatory activity and facilitatebiomolecular adhesion to enhance osteoblast function. Therefore,in this in vitro study, we modified the surface of titanium usingheparin. Heparin has well characterized anti-inflammatory andanticoagulant properties [24]. According to Heana et al., low-molecular-weight heparin (enoxaparin) reduced the high glucose-induced activation of endothelial cells by inhibiting monocyteadhesion [25]. In this study, the free amino groups of 3-amino-propyltriethoxysilane (ATPES) were first anchored to the titaniumsurface to create regions of high positive charge, and then heparinwas covalently grafted. Heparin was grafted to the titanium surfaceusing a 1-3thyl-3-dimethylaminopropyl carbodiimide (EDC)-mediated coupling reaction between the primary amine groups ofrhBMP-2 and the carboxyl groups of heparin.

We hypothesized that a Ti surface with grafted heparin andimmobilized rhBMP-2 would inhibit inflammation and enhanceosteoblast function.

2. Materials and methods

2.1. Amino-functionalized titanium

Titanium (Ti) discs were kindly supplied by Dio Implant Co., Ltd (Busan, Korea).Ti discs were washed in an ultra-sonicator containing absolute ethanol for 1 h priorto use. Amine-functionalized Ti discs were produced in anhydrous toluene con-taining 3-aminopropyltriethoxysilane (ATPES, SigmaeAldrich, MO, USA). In brief, 50Ti discs were immersed in 250 mL anhydrous toluene and 10 mL (v/v) of ATPES. Thereaction was performed at 120 �C under an N2 atmosphere with a reflux condenserfor 24 h. Ti discs were then washed with toluene for 20 min to remove unreactedsilane. This washing procedure was repeated five times. After washing, the amino-functionalized Ti discs were dried at 60 �C for 24 h.

2.2. Grafting of heparin and immobilization of recombinant human bonemorphogenic protein-2 (rhBMP-2) to titanium surfaces

Heparinwas grafted to the surfaces of the amine-treated Ti discs by the 1-ethyl-3-dimethylaminopropyl carbodiimide (EDC)-mediated reaction between theprimary amine groups of the Ti surface and the carboxyl groups of heparin. Briefly,1mg/mL heparinwas dissolved in 0.59mg EDC/0.18mgNHS in 0.1MMES buffer (pH5.6). The Ti discs were then immersed in the aforementioned solution for 24 h atroom temperature. After the reaction, the Ti discs were thoroughly washed with0.1 M Na2HPO4 (1 h), 4 M NaCl (three times in 24 h), and distilled water (three timein 24 h), respectively. After washing, the Ti discs were frozen at �80 �C for 24 h andlyophilized for 3 days. rhBMP-2 at a concentration of 10 or 50 ng/mL was immobi-lized on the surface of the heparin-grafted Ti discs. In brief, the heparin-grafted Tidiscs were immersed in 0.1 M MES buffer solution (pH 5.6). rhBMP-2 at a concen-tration of 10 or 50 ng/mL was then added to the 0.1 M MES buffer solution (pH 5.6)and the reaction was allowed to proceed for 24 h at room temperature.

2.3. Determination of the amount of heparin-grafted to the titanium surface

The amount of heparin-grafted on the surface of the titanium discs was deter-mined using the toluidine blue method. Briefly, Ti discs were placed into 1 mLphosphate buffer saline (PBS, pH 7.4) solution containing 1 mL of 0.005% toluidineblue solution. After 30 min under gentle shaking, 2 mL of hexane was added. Afterthe Ti discs were removed from solution, the absorbance of the aqueous phase wasmeasured at 620 nm. The amount of heparin coated on the surface of titanium wascalculated from a calibration curve that was constructed using various concentra-tions of heparin.

2.4. In vitro rhBMP-2 release study

To evaluate the release kinetics of rhBMP-2 immobilized on the surfaces ofheparin-grafted Ti discs, the Ti discs were soaked in a 15 mL conical tube (Falcon,USA) containing 3 mL PBS (pH 7.4) at 100 rpm at 37 �C. At predetermined time

Please cite this article in press as: Kim SE, et al., The effect of immobBiomaterials (2010), doi:10.1016/j.biomaterials.2010.09.008

intervals of 1 h, 2 h, 8 h, and 1, 3, 5, 7, 14, 21, and 28 days, the supernatant wascollected and replaced with fresh PBS solution. All samples were stored at �20 �Cuntil analysis. The absorbance of samples was determined with an enzyme-linkedimmunosorbent assay (ELISA) according to the manufacturer’s instructions usinga microplate reader (Bio-Rad, Hercules, CA, USA) at a wavelength of 495 nm.

2.5. Characterization of the Ti discs

To determine the surface morphologies of pristine and functionalized Ti discs(heparin-grafted Ti discs and BMP-2-immobilized Ti discs), a scanning electronmicroscope (SEM, S2300, Hitachi, Japan) was used. The substrates were coated withgold using a sputter-coater (Eiko IB, Japan). The SEM was operated at 15 kV. Thesurface composition of pristine Ti discs, heparin-grafted Ti discs, and rhBMP-2-immobilized Ti discs was analyzed by X-ray photoelectron spectroscopy (XPS) ona K-Alpha spectrometer (Thermo Electron, USA) with an Al Ka X-ray source (1486.6 eV photons). The C1s hydrocarbon peak at 284.84 eV was used as the reference forall binding energies. The area of each peak was normalized to the total peak area ofall atomic elements to calculate surface atomic percentages. To evaluate thehydrophilic properties of pristine Ti and the functionalized Ti surfaces, contactangles were measured using the sessile drop method and a video contact angleinstrument (Phoenix 150, SEO, Korea) at room temperature.

2.6. In vitro cellular responses

MG-63 cells (human osteosarcoma cell line, Korean Cell Bank Line, Seoul, Korea)were used to characterize the biocompatibility of the pristine and functionalized Tidiscs (heparin-grafted Ti discs, rhBMP-2 (10 ng)-immobilized Ti discs, and rhBMP-2(50 ng)-immobilized Ti discs) by measuring cell proliferation, alkaline phosphataseactivity, and calcium deposition. Cells were cultured in F-100 culture plates at 37 �Cin a humidified atmosphere supplied with 5% CO2. Cells were maintained in Dul-becco’s modified eagle’s medium (DMEM) supplemented with 10% FBS, 50 mg/mLascorbic acid, 10 nM dexamethasone, and 10 mM b-glycerolphosphate in the pres-ence of 100 U/mL penicillin and 100 mg/mL streptomycin. Prior to cell-seeding,specimens were sterilized with 70% EtOH for 10 min and rinsed twice with phos-phate-buffered saline (PBS).

2.7. Cytotoxicity test

Cytotoxicity tests for pristine and functionalized Ti discs were carried outaccording to the ISO/EN 10993 Part 5 guidelines. To obtain extraction media, DMEMmedium was incubated with pristine, amine-treated Ti, and functionalized Ti discs,respectively, for 24 h at 37 �C. NIH3T3 Fibroblasts and MG-63 cells were seeded into96-well plates at a concentration of 5�104 cells/well and incubated for 24 h at 37 �Cwith DMEM supplemented with 10% FBS, and 1% 100 U/mL penicillin and 100 mg/mLstreptomycin. After 24 h of culture, the DMEM medium was removed from the 96-well plates, the cells were washed with PBS, and extraction media was added. Cellswere incubated for 24 and 48 h. At each time point, the extraction medium wasaspirated and CCK-8 proliferation kit (Dojindo, Japan) reagents were added to cells.Cells were then incubated for 1 h at 37 �C, and the optical density of live cells wasmeasured using a microplate reader at a wavelength of 450 nm.

2.8. Live/dead assay

The viability of cells on the surface of the pristine and functionalized Ti discs wasassessed by live/dead staining. In brief, MG-63 cells were seeded at a density of5 � 104 cells/mL on the surface of a series of Ti discs in 48-well plates. After a 48 hincubation, cells/specimens were rinsed three times with PBS and then incubatedwith live/dead stain (2 mM calcein AM and 4 mM ethidium homodimer-1) for 30 minat room temperature (RT). Viable cells (green) and dead cells (red) were countedunder a confocal laser scanning microscope (CLSM, EZ-C1, Nikon, Japan).

2.9. Protein adsorption assay

To evaluate protein adsorption on the surfaces of pristine and functionalized Tisurfaces, fibronectin and bovine serum albumin (BSA) were used as representativeproteins. Five hundredmicroliters of BSA (1mg/mL BSA/PBS) orfibronectin (1mg/mLfibronectin/PBS) were pipetted onto each surface, respectively. After 3 h, non-adherent proteinswere removed and collected. Bradford solution (Bio-Rad, Hercules,CA, USA) was added to each surface for 1 h at 37 �C. Protein concentrations weredetermined using the Bradford assay according to the manufacturer’s protocols, andabsorbance was measured using a microplate reader at a wavelength of 595 nm.

2.10. Cell proliferation

MG-63 cells were seeded on pristine and functionalized Ti discs at a density of1 � 105 cells and incubated for 7 days. At predetermined time intervals (1, 3, and 7days), specimens were rinsed with PBS and CCK-8 proliferation kit reagents wereadded to the specimens. After a 1 h incubation, reagents were carefully transferred


Fig. 1. SEM of various Ti surfaces after amine-treated, heparin-grafted, and rhBMP-2-immobilized: (a) pristine Ti, (b) heparin-grafted Ti, (c) rhBMP-2 (10 ng)-immobilized Ti, and(d) rhBMP-2 (50 ng)-immobilized Ti.

Table 1.

Condition Contact angle (�)

Titanium (Ti) 84.18 � 3.29Heparin-Ti 52.22 � 0.09BMP-2 (10 ng) immobilized-Ti 48.73 � 0.56BMP-2 (50 ng) immobilized-Ti 41.72 � 0.34

S.E. Kim et al. / Biomaterials xxx (2010) 1e8 3

to 96-well plates. The optical density was measured using a microplate reader ata wavelength of 450 nm.

2.11. Alkaline phosphate (ALP) activity assay

After 7, 14 and 21 days of culture, ALP activity was measured as describedpreviously with some modifications. In brief, cells were seeded at a density of1 �105 cells/ml on pristine and functionalized Ti discs. The cells were washed withPBS and then 1X RIPA buffer [50 mM TriseHCl, pH 7.4, 150 mM NaCl, 0.25% deox-ycholic acid, 1% NP-40, 1 mM EDTA including protease and phosphatase inhibitors(1 mM PMSF, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mg/mL apro-tinin, 1 mg/mL leupeptin, 1 mg/mL pepstatin)] was added to cells. The cells in RIPAbuffer were sonicated using a Vibra Cell� instrument (Sonics & Materials INC.,Danbury, CT, USA) for 1 min at 110 W on ice. After sonication, the cell lysates werecentrifuged at 13,500 rpm at 4 �C for 3 min to remove cell debris. The supernatantswere incubated with p-nitrophenyl phosphate solution for 30 min at 37 �C. Thereactionwas stopped by adding 500 ml of 1 N NaOH. ALP activity was determined bymeasuring the conversion of p-nitrophenyl phosphate to p-nitrophenol. Opticaldensity was determined by using a microplate reader at a wavelength of 405 nm.

2.12. Calcium deposition

To quantify calcium deposition, cells were seeded at a density of 1 �105 cells/mlon pristine and functionalized Ti discs. At predetermined culture time intervals (7,14and 21 days), cells were washed with PBS and gently scraped off from the surface ofthe substrates. Cells were harvested by centrifugation at 13,500 rpm for 1 min, andlysis buffer (0.1% Triton X-100) was then added to the cells. The cells were sonicatedon ice for 1 min to pulverize the cell membranes. The resulting supernatant wasused for calcium deposition measurements using a QuantiChrom� Calcium AssayKit (DICA-500, BioAssay Systems, USA) according to the manufacturer’s instructions.The amount of calcium produced was estimated by measuring the absorbance at612 nm using a microplate reader.

2.13. Anti-inflammatory effects

Murine macrophage cells from the cell line RAW 264.7 (Korean Cell Lines Bank,Seoul, Korea) were seeded on the surface of pristine Ti discs (negative control),heparin-grafted Ti discs, BMP-2 (10 ng/mL)-immobilized Ti discs, BMP-2 (50 ng/mL)-immobilized Ti discs, and LPS (100 ng/mL)-induced Ti discs (positive control) ata density of 5 � 104 cells. After a 24 h incubation, cells were harvested for total RNAisolation. Isolated total RNAwas extracted using the RNeasy PlusMini Kit (Qiagen, CA,USA) and 1 mg of total RNA was reverse transcribed into cDNA using AccuPower RTPreMix (Bioneer, Daejeon, Korea) according to the manufacturer’s protocols. All PCRamplifications were performed using AccuPower PCR PreMix (Bioneer, Daejeon,


Korea). The primers for TNF-a were 50- GGC AGG TCT ACT TTG GAG TCA TTG C �30

(sense) and 50- ACATTCGAGGCT CCAGTGAAT TCGG�30 (antisense), the primers forIL-6 were 50- CTG GTGACA ACC ACG GCC TTC CCT A�30 (sense) and 50- ATG CTT AGGCATAACGCACTAGGT T�30 (antisense), and the primers for GAPDHwere 50-ACT TTGTCA AGC TCA TTT CC -30 (sense) and 50-TGC AGC GAACTT TAT TGA TG -30 (antisense).The PCR cycling profilewas as follows: 30 s at 94 �C,1min at 65 �C, and 1min at 72 �Cfor 35 cycles after an initial denaturation step for 5 min at 94 �C. PCR products wereseparated on 1% agarose gels and visualized by ethidium bromide staining.

2.14. Statistical analysis

Data are presented as means � standard deviations. Statistical comparisonswere carried out using one-way ANOVA tests. Statistically significant values aredefined as *P < 0.001.

3. Results

3.1. Surface characterization

The surface morphologies of pristine and functionalized Ti discswere determined by scanning electron microscopy (Fig. 1aed). Thepristine and functionalized Ti surfaces had similar morphologies. Toassess the wettability of the Ti surfaces at various stages of surfacefunctionalization, contact angles weremeasured. As shown in Table1, the mean� standard deviation of the contact angle on pristine Ti,heparin-grafted Ti, rhBMP-2 (10 ng)-immobilized Ti, and rhBMP-2(50 ng)-immobilized Ti was 84.18 � 3.29, 52 � 0.09, 48.73 � 0.56,and 41.72 � 0.34, respectively. To determine the chemical compo-sition of the surfaces of pristine Ti and functionalized Ti, XPSanalyses were conducted. The XPS wide-scan spectra of pristine Tiand functionalized Ti (amine-treated Ti, heparin-grafted Ti, and


Fig. 2. XPS wide-scan spectra of (a) pristine Ti, (b) amine-treated Ti, (c) heparin-grafted Ti, and (d) rhBMP-2-immobilized Ti.

Fig. 3. Release kinetics of rhBMP-2 (10 ng) and rhBMP-2 (50 ng) fromheparin-grafted Ti.


rhBMP-2-immobilized Ti) and their corresponding surfaceelemental composition are shown in Fig. 2 and Table 2, respectively.Successful grafting of heparin to the Ti surface was indicated by anincrease in the C content and a decrease in the N content comparedto the amine-treated Ti surface, as shown in Table 2. Furthermore,successful immobilization of rhBMP-2 on the Ti surface wasdemonstrated by a decrease in the C content and an increase in theN content compared to the heparin-grafted Ti surface (Table 2). Toquantify the amount of heparin-grafted to the Ti surface, toluidineblue dye was utilized. The amount of surface-exposed heparin was0.06 mg/sample.

3.2. In vitro rhBMP-2 release study

The bioactivity of two different concentrations of rhBMP-2immobilized on heparin-grafted Ti was analyzed by using enzyme-linked immunosorbent assays (Fig. 3). The amount of rhBMP-2(10 ng) and rhBMP-2 (50 ng) released on the first day wasapproximately 69% and 26%, respectively. Over a period of 28 days,6.9 � 0.51 and 13.0 � 1.38 ng of the rhBMP-2 was released from the10 ng-rhBMP-2 immobilized titanium surfaces and 50 ng-rhBMP-2immobilized titanium surfaces, respectively. Examination of therelease curves revealed that the rhBMP-2 immobilized titaniumsurfaces exhibited sustained release of different concentrations ofrhBMP-2 according to the starting concentration of rhBMP-2.

3.3. Protein adsorption

The protein adsorption results for pristine and functionalized Tidiscs are shown in Fig. 4. Significant amounts of albumin andfibronectin adsorbed to Ti surfaces with rhBMP-2 immobilizedcompared to pristine Ti surfaces (*P<0.001). However, the amountsof albumin and fibronectin adsorbed on the heparin-grafted Tisurfaces were not significantly different to those adsorbed topristine Ti surfaces. Furthermore, significantly higher amounts of

Table 2.

substrate C% N% Ti%

Titanium (Ti) 77.45 � 0.98 1.68 � 0.37 20.86 � 0.61Amine-treated Ti 65.22 � 0.48 16.35 � 0.41 18.42 � 0.83Heparin-grafted Ti 74.64 � 0.53 8.26 � 0.27 17.28 � 0.52BMP-2 immobilized Ti 71.51 � 0.59 10.45 � 0.32 18.52 � 0.93


albumin and fibronectin were absorbed on the rhBMP-2 (50 ng)-immobilized Ti surfaces than the rhBMP-2 (10 ng)-immobilized Tisurfaces (*P < 0.001).

3.4. Cytotoxicity tests and live/dead staining

We performed cytotoxicity tests against NIH3T3 fibroblasts andMG-63 cells for all groups the pristine Ti, amine-treated Ti, hep-arinegeTi, rhBMP-2 (10 ng)-Ti, and rhBMP-2 (50 ng)-Ti discs priorto osteoblastic cell inoculation. No significant cytotoxic effects wereobserved for culture periods of up to 48 h (data not shown). Inaddition, we assessed the cytotoxicity of pristine and functionalizedTi using a fluorescence staining method (live/dead assay). Live cellsand dead cells were fluorescently labeled green and red, respec-tively. As shown in Fig. 5, almost all cells were alive after a 48-hexposure to either pristine Ti or functionalized Ti.

3.5. Cell proliferation

The cell proliferation of MG-63 cells on pristine and function-alized Ti was investigated at days 1, 3 and 7 (Fig. 6). Proliferation ofMG-63 cells on both pristine and functionalized Ti increasedslightly throughout the incubation period for up to 7 days. Cellproliferation on heparin-grafted Ti and rhBMP-2-immobilized Tisurfaces was not significantly different to that on pristine Ti

Fig. 4. Protein adsorption of pristine Ti, heparin-grafted Ti, and BMP-2 immobilized Tiafter 3 h treatments of model proteins (*p < 0.001).


Fig. 5. Live/dead assay: (a) pristine Ti, (b) heparin-grafted Ti, (C) BMP-2 (10 ng)-immobilized Ti, and (d) BMP-2 (50 ng)-immobilized Ti after a 48-incubation time.


surfaces. There were no significant differences after 7 days ofculture between cells grown on functionalized Ti and those grownon pristine Ti in terms of cell proliferation.

3.6. ALP activity

ALP is widely used as a marker of early differentiation of oste-oblasts; we measured the ALP levels of MG-63 cells after 7, 14 and

Fig. 6. Cell proliferation of osteoblasts (MG-63 cells) cultured on pristine Ti, heparin-grafted Ti, and BMP-2-immobilized Ti after 1, 3 and 7 days.


21 days of culture on pristine Ti and functionalized Ti discs. Theresults in Fig. 7 clearly show that MG-63 cells cultured on rhBMP-2-immobilized Ti discs had significantly higher ALP activity thanthose cultured on pristine Ti discs (*P < 0.001). MG-63 cells grownfor 7 or 14 days on rhBMP-2-immobilized Ti discs had significantlyhigher ALP activity than those cultured on pristine Ti discs(*P < 0.001). However, by 21 days, the ALP activity of MG-63 cellscultured on all substrates was slightly higher than the ALP activityof cells grown on all substrates at 14 days. The ALP activity ofMG-63 cells grown on heparin-grafted Ti was not significantlydifferent compared to the ALP activity of MG-63 cells cultivated onpristine Ti discs for 7, 14 or 21 days.

3.7. Mineralization

Calcium mineral deposition is a marker of the late differentia-tion of osteoblasts. The amount of calcium deposition that occurredafter cultivating MG-63 cells for 7, 14 and 21 days on pristine andfunctionalized Ti discs is shown in Fig. 8. The amount of calciumdeposition on all substrates increased through culture time. Therewere no significant differences in calcium deposition between cellsgrown on pristine Ti discs and those grown on heparin-grafted Tidiscs. In contrast, calcium deposition on the rhBMP-2-immobilizedTi surfaces was significantly greater than that on pristine Ti surfaces(*P < 0.001). Furthermore, there were significant differences incalcium deposition between cells grown on rhBMP-2 (10 ng)-immobilized Ti surfaces and those grown on rhBMP-2 (50 ng)-immobilized Ti surfaces (*P < 0.001).


Fig. 7. ALP activity of osteoblasts (MG-63 cells) cultured on pristine Ti, heparin-graftedTi, and BMP-2-immobilized Ti after incubation of 7, 14 and 21 days. (*) denotessignificant differences (P < 0.001).

Fig. 9. RT-PCR analysis for anti-inflammatory protein expression of pristine Ti,heparin-grafted Ti, and BMP-2-immobilized Ti cultured murine macrophage cells(RAW 264.7 cell) after 24 h incubation.


3.8. Anti-inflammatory tests

To determine whether functionalized Ti surfaces had an anti-inflammatory effect compared to pristine Ti surfaces, we measuredthe transcript levels of TNF-a and IL-6, which are genetic markers ofinflammation. As the shown in Fig. 9, mRNA levels of TNF-a and IL-6were slightly up-regulated in cells grown on pristine Ti surfaces andstrongly over-expressed in cells grown on LPS-induced Ti surfaces(positive control). In contrast, transcripts for TNF-a and IL-6 werenot expressed in cells grown on heparin-grafted Ti, nor weretranscripts of these cytokines detectable in cells grown on rhBMP-2-immobilized Ti surfaces (Fig. 9). These data indicate that heparin-grafted and rhBMP-2 immobilized Ti surfaces have anti-inflam-matory effects compared to pristine Ti surfaces.

4. Discussion

The clinical success of dental implants is highly dependent onthe cell adhesion, matrix production, anti-inflammatory, andmineralization properties of the implant materials. In the present

Fig. 8. Calcium content of osteoblast (MG-63 cells) cultured on pristine Ti, heparin-grafted Ti, and BMP-2-immobilized Ti surfaces after incubation of 7, 14 and 21 days(*P < 0.001).


study, we demonstrated that modification of Ti surfaces withheparin and rhBMP-2 had an anti-inflammatory effect on MG-63cells and resulted in an increase in osteogenic activity. The rhBMP-2(50 ng)-immobilized Ti discs showed a significantly smaller watercontact angle than pristine Ti discs, indicating that the immobilizedrhBMP-2 disc surfaces weremore hydrophilic than those of pristineTi. This is consistent with a study by Zhiong and co-workers [2],who reported that rhBMP-2 immobilized carboxymethyl chitosan(CMCS)-grafted Ti surfaces were more hydrophilic than pristine Ti.The chemical compositions of the surfaces after modification withheparin and rhBMP-2 were determined by XPS. Successful graftingof heparin to the Ti surface was indicated by both an increase in Cand N content and a decrease in Ti content compared to Ti surfacesafter ATPES processing. Moreover, successful anchoring of rhBMP-2to the heparin-modified Ti surfaces was demonstrated by theincrease in N content due to the N-containing groups in rhBMP-2compared to the heparin-grafted Ti surfaces. A previous studyreported successful immobilization of rhBMP-2 to CMCS-grafted Tisurfaces because of the formation of covalent bonds between thecarboxyl groups of CMCS and the amine groups of rhBMP-2 [2].Similarly, in our study, the carboxyl groups of heparin formedcovalent bondswith the amine groups of rhBMP-2. Asmentioned inthe introduction, rhBMP-2 is one of the most osteoinductivegrowth factors, but its clinical application has been limited by itsshort half-life, rapid diffusion by body fluids, and the large dosesrequired. Large doses of rhBMP-2 in particular are a problembecause they can result in side-effects such as immune responseand bone overgrowth from the defect site [26]. To create an effec-tive delivery system for rhBMP-2 and ensure its release for a suffi-cient time period and in appropriate amounts, we modified thesurfaces of Ti discs with heparin to which we immobilized rhBMP-2.We demonstrated that the rhBMP-2 immobilized on the heparin-grafted Ti surface was successfully released over an extendedperiod of time. The sustained release of growth factors, includingvarious types of matrices, from heparin has been well documented[23,27,28]. Lin et al. [28] demonstrated that heparin-crosslinkeddemineralized bone matrix (DBM) to which BMP-2 was boundincreased the ALP activity of attached cells and the percentage ofcalcified tissue in vivo. Moreover, a previous study demonstratedthat heparin-immobilized PLGA porous microspheres loaded withbFGF released bFGF in a more sustained manner in vitro andstimulated angiogenesis in vivo when compared to the sameamount of bFGF loaded into PLGA microspheres [29]. These resultsdemonstrate that heparin is a suitable material for the sustainedrelease of growth factors. We observed an increase in osteoblastproliferation when osteoblasts were cultured on all substrates.However, cell proliferation on rhBMP-2-immobilized Ti discs did



not show a significant increase compared to that on pristine Ti,consistent with the reports of Zhilong et al. [2,30] In contrast, Parket al. reported an increase in the proliferation of osteoblasts grownon nanofibrous chitosan membranes with immobilized BMP-2compared to their proliferation on nanofibrous chitosanmembranes without immobilized BMP-2 [31]. Thus, the effects ofrhBMP-2 on osteoblast proliferation are still unclear.

ALP activity and calcium deposition are widely used as markersfor early and late differentiation of osteoblast cells, respectively[32,33]. ALP activity was measured after a culture period of 7, 14 or21 days. There was no significant difference in osteoblast prolifer-ation between cells grown on heparin-grafted Ti versus pristine Tidiscs throughout the 21 days. In contrast, the ALP activity of oste-oblasts cultured on rhBMP-2-immobilized Ti surfaces was signifi-cantly higher than that of osteoblasts cultured on pristine Tisurfaces for the different culture periods. Thus, rhBMP-2 stimulatesosteoblast differentiation. However, the ALP activity of osteoblastson all substrates at 21 days was slightly lower than that on allsubstrates at 14 days. This means that ALP activity reacheda maximum before mineralization actually began [34]. Further-more, these results indicate that ALP activity may diminish slightlyand calcium deposition may increase slowly after a culture periodof 14 days. Materials like DBM, titanium, and chitosan-immobilizedBMP-2 have been shown to increase the ALP activity of osteoblast-like cells [2,28,31]. Significantly more calciumwas deposited on therhBMP-2-immobilized Ti surfaces than on pristine Ti surfaces overa culture period of 21 days. As was expected from the ALP activityprofile, the amount of calcium deposed on heparin-grafted Tisurfaces was similar to that deposited on pristine Ti surfacesthroughout the 21 day culture period. In contrast, there wasa significant difference in calcium deposition between cellsgrowing on rhBMP-2 (10 ng)-immobilized Ti surfaces and thosegrowing on rhBMP-2 (50 ng)-immobilized Ti surfaces at 7,14 and 21days. Together, these results indicate that rhBMP-2 immobilized Tisubstrates can stimulate matrix formation and enhance osteoblastcell function.

We examined the anti-inflammatory effects of pristine Ti andfunctionalized Ti by measuring transcript levels of the pro-inflammatory cytokines interleukin 6 (IL-6) and tumor necrosisfactor (TNF-a). Expression of these cytokines was not detected inosteoblasts cultured on heparin-grafted Ti surfaces. In contrast,slight up-regulation of pro-inflammatory cytokines was detected incells grown on pristine Ti surfaces. In an earlier paper, Thouraniet al. demonstrated that heparin and O-desulphated heparininhibited NFekB activation in a TNF-aestimulated human endo-thelial cell line and in ischemic-reperfused rat myocardium [35]. Afurther study showed that unfractionated heparin and low-molecular weight heparin downregulated pro-inflammatory cyto-kines and NFekB in LPS-stimulated human monocytes [36]. Theseresults suggest that heparin has anti-inflammatory activity,consistent with the results of our study.

Our present study has the following clinical implications: (i)sustained release of rhBMP-2 from rhBMP-2 immobilized, heparin-grafted Ti surfaces can be achieved over long periods of time; (ii)suitable doses of rhBMP-2 can enhance osteoblast function; and(iii) heparin-grafted Ti can decrease the inflammatory response.Regarding the current concerns on peri-implant diseases [37,38],the anti-inflammatory action of this functionalized Ti can give uspossibility to reduce the incidence of peri-implant disease.Considering the prevention of peri-implantitis is one of majorconcerns in implant dentistry, the possible anti-inflammatoryeffect by this functionalization process may be promising to reducethe tissue destruction by peri-implantitis. Further studies arerequired to determine the appropriate doses of rhBMP-2 for in vivoapplication of this delivery system and validate our results.


5. Conclusions

Ti substrates were successfully functionalized via a chemicalmethod involving grafting of heparin and subsequent immobili-zation of rhBMP-2. Heparin reduced inflammation effectively andsustained release of rhBMP-2 was achieved. The rhBMP-2-immo-bilized Ti substrates stimulated osteoblast functions successfully.Furthermore, cells grown on rhBMP-2 (50 ng)-immobilized Tisurfaces had significantly greater ALP activity and calcium contentsthan cells grown on pristine Ti surfaces. Thus, rhBMP-2-immobi-lized, heparin-grafted Ti discs are a suitable delivery system toenhance osteoblast function and decrease inflammatory reactionand may therefore be suitable for incorporation in implant mate-rials for orthopedic and dental applications.

Acknowledgments

This research was supported by Basic Science Research Programthrough the National Research Foundation of Korea(NRF) funded bythe Ministry of Education, Science and Technology(2010-0005144).This workwas supported by grand No. 20100002088 from the BasicResearch Program of the Korea Science & Engineering Foundation.

Appendix

Figures with essential color discrimination. Fig. 5 in this articleare difficult to interpret in black and white. The full color imagescan be found in the on-line version, at 10.1016/j.biomaterials.2010.09.008

References

[1] Piskounova S, Forsgren J, Brohede U, Engqvist H, Stromme M. In vitro char-acterization of bioactive titanium dioxide/hydroxyapatite surfaces function-alized with BMP-2. J Biomed Mater Res B Appl Biomater 2009;91:780e7.

[2] Shi Z, Neoh KG, Kang ET, Poh CK, Wang W. Surface functionalization of tita-nium with carboxymethyl chitosan and immobilized bone morphogeneticprotein-2 for enhanced osseointegration. Biomacromolecules 2009;10:1603e11.

[3] Miron RJ, Oates CJ, Molenberg A, Dard M, Hamilton DW. The effect of enamelmatrix proteins on the spreading, proliferation and differentiation of osteo-blasts cultured on titanium surfaces. Biomaterials 2010;31:449e60.

[4] Michael J, Schonzart L, Israel I, Beutner R, Scharnweber D, Worch H, et al.Oligonucleotide-RGD peptide conjugates for surface modification of titaniumimplants and improvement of osteoblast adhesion. Bioconjug Chem2009;20:710e8.

[5] Oya K, Tanaka Y, Saito H, Kurashima K, Nogi K, Tsutsumi H, et al. Calcificationby MC3T3-E1 cells on RGD peptide immobilized on titanium through elec-trodeposited PEG. Biomaterials 2009;30:1281e6.

[6] Balasundaram G, Webster TJ. Increased osteoblast adhesion on nanograined Timodified with KRSR. J Biomed Mater Res A 2007;80:602e11.

[7] Wang EA, Rosen V, D’Alessandro JS, Bauduy M, Cordes P, Harada T, et al.Recombinant human bone morphogenetic protein induces bone formation.Proc Natl Acad Sci U S A 1990;87:2220e4.

[8] Cook SD, Wolfe MW, Salkeld SL, Rueger DC. Effect of recombinant humanosteogenic protein-1 on healing of segmental defects in non-human primates.J Bone Jt Surg Am 1995;77:734e50.

[9] Kim SE, Jeon O, Lee JB, Bae MS, Chun HJ, Moon SH, et al. Enhancement ofectopic bone formation by bone morphogenetic protein-2 delivery usingheparin-conjugated PLGA nanoparticles with transplantation of bonemarrow-derived mesenchymal stem cells. J Biomed Sci 2008;15:771e7.

[10] Govender S, Csimma C, Genant HK, Valentin-Opran A, Amit Y, Arbel R, et al.Recombinant human bone morphogenetic protein-2 for treatment of opentibial fractures: a prospective, controlled, randomized study of four hundredand fifty patients. J Bone Jt Surg Am 2002;84-A:2123e34.

[11] Sellers RS, Zhang R, Glasson SS, Kim HD, Peluso D, D’Augusta DA, et al. Repairof articular cartilage defects one year after treatment with recombinanthuman bone morphogenetic protein-2 (rhBMP-2). J Bone Jt Surg Am2000;82:151e60.

[12] Bax BE, Wozney JM, Ashhurst DE. Bone morphogenetic protein-2 increasesthe rate of callus formation after fracture of the rabbit tibia. Calcif Tissue Int1999;65:83e9.

[13] Sheehan JP, Kallmes DF, Sheehan JM, Jane Jr JA, Fergus AH, diPierro CG, et al.Molecular methods of enhancing lumbar spine fusion. Neurosurgery1996;39:548e54.





[14] King GN, King N, Cruchley AT, Wozney JM, Hughes FJ. Recombinant humanbone morphogenetic protein-2 promotes wound healing in rat periodontalfenestration defects. J Dent Res 1997;76:1460e70.

[15] Sellers RS, Peluso D, Morris EA. The effect of recombinant human bonemorphogenetic protein-2 (rhBMP-2) on the healing of full-thickness defects ofarticular cartilage. J Bone Jt Surg Am 1997;79:1452e63.

[16] Zellin G, Linde A. Importance of delivery systems for growth-stimulatoryfactors in combination with osteopromotive membranes. An experimentalstudy using rhBMP-2 in rat mandibular defects. J Biomed Mater Res1997;35:181e90.

[17] Kim J, Park Y, Tae G, Lee KB, Hwang CM, Hwang SJ, et al. Characterization oflow-molecular-weight hyaluronic acid-based hydrogel and differential stemcell responses in the hydrogel microenvironments. J Biomed Mater Res A2009;88:967e75.

[18] Kim J, Kim IS, Cho TH, Lee KB, Hwang SJ, Tae G, et al. Bone regeneration usinghyaluronic acid-based hydrogel with bone morphogenic protein-2 and humanmesenchymal stem cells. Biomaterials 2007;28:1830e7.

[19] Lee TC, Ho JT, Hung KS, Chen WF, Chung YH, Yang YL. Bone morphogeneticprotein gene therapy using a fibrin scaffold for a rabbit spinal-fusion exper-iment. Neurosurgery 2006;58:373e80.

[20] Jeon O, Song SJ, Yang HS, Bhang SH, Kang SW, Sung MA, et al. Long-termdelivery enhances in vivo osteogenic efficacy of bone morphogenetic protein-2 compared to short-term delivery. Biochem Biophys Res Commun2008;369:774e80.

[21] Sasisekharan R, Ernst S, Venkataraman G. On the regulation of fibroblast growthfactor activity by heparin-like glycosaminoglycans. Angiogenesis 1997;1:45e54.

[22] Perets A, Baruch Y, Weisbuch F, Shoshany G, Neufeld G, Cohen S. Enhancingthe vascularization of three-dimensional porous alginate scaffolds by incor-porating controlled release basic fibroblast growth factor microspheres.J Biomed Mater Res A 2003;65:489e97.

[23] Ishibe T, Goto T, Kodama T, Miyazaki T, Kobayashi S, Takahashi T. Boneformation on apatite-coated titanium with incorporated BMP-2/heparin invivo. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;108:867e75.

[24] Young E. The anti-inflammatory effects of heparin and related compounds.Thromb Res 2008;122:743e52.

[25] Manduteanu I, Voinea M, Antohe F, Dragomir E, Capraru M, Radulescu L, et al.Effect of enoxaparin on high glucose-induced activation of endothelial cells.Eur J Pharmacol 2003;477:269e76.

[26] Shields LB, Raque GH, Glassman SD, Campbell M, Vitaz T, Harpring J, et al.Adverse effects associated with high-dose recombinant human bonemorphogenetic protein-2 use in anterior cervical spine fusion. Spine (Phila Pa1976) 2006;31:542e7.


[27] Ho YC, Mi FL, Sung HW, Kuo PL. Heparin-functionalized chitosanealginatescaffolds for controlled release of growth factor. Int J Pharm2009;376:69e75.

[28] Lin H, Zhao Y, Sun W, Chen B, Zhang J, Zhao W, et al. The effect of crosslinkingheparin to demineralized bone matrix on mechanical strength and specificbinding to human bone morphogenetic protein-2. Biomaterials2008;29:1189e97.

[29] Chung HJ, Kim HK, Yoon JJ, Park TG. Heparin immobilized porous PLGAmicrospheres for angiogenic growth factor delivery. Pharm Res2006;23:1835e41.

[30] Shi Z, Neoh KG, Kang ET, Poh C, Wang W. Titanium with surface-grafteddextran and immobilized bone morphogenetic protein-2 for inhibition ofbacterial adhesion and enhancement of osteoblast functions. Tissue Eng Part A2009;15:417e26.

[31] Park YJ, Kim KH, Lee JY, Ku Y, Lee SJ, Min BM, et al. Immobilization of bonemorphogenetic protein-2 on a nanofibrous chitosan membrane for enhancedguided bone regeneration. Biotechnol Appl Biochem 2006;43:17e24.

[32] Turksen K, Bhargava U, Moe HK, Aubin JE. Isolation of monoclonal antibodiesrecognizing rat bone-associated molecules in vitro and in vivo. J HistochemCytochem 1992;40:1339e52.

[33] van den Beucken JJ, Walboomers XF, Boerman OC, Vos MR, Sommerdijk NA,Hayakawa T, et al. Functionalization of multilayered DNA-coatings with bonemorphogenetic protein 2. J Control Release 2006;113:63e72.

[34] Bancroft GN, Sikavitsas VI, van den Dolder J, Sheffield TL, Ambrose CG,Jansen JA, et al. Fluid flow increases mineralized matrix deposition in 3Dperfusion culture of marrow stromal osteoblasts in a dose-dependent manner.Proc Natl Acad Sci U S A 2002;99:12600e5.

[35] Thourani VH, Brar SS, Kennedy TP, Thornton LR, Watts JA, Ronson RS, et al.Nonanticoagulant heparin inhibits NF-kappaB activation and attenuatesmyocardial reperfusion injury. Am J Physiol Heart Circ Physiol2000;278:H2084e93.

[36] Hochart H, Jenkins PV, Smith OP, White B. Low-molecular weight andunfractionated heparins induce a downregulation of inflammation: decreasedlevels of proinflammatory cytokines and nuclear factor-kappaB in LPS-stim-ulated human monocytes. Br J Haematol 2006;133:62e7.

[37] Berglundh T, Persson L, Klinge B. A systematic review of the incidence ofbiological and technical complications in implant dentistry reported inprospective longitudinal studies of at least 5 years. J Clin Periodontol 2002;29(Suppl. 3):197e212. discussion 232e193.

[38] Kwon YD, Karbach J, Wagner W, Al-Nawas B. Peri-implant parameters in headand neck reconstruction: influence of extraoral skin or intraoral mucosa. ClinOral Implants Res 2010;21:316e20.


Documents

The effect of immobilization of heparin and bone morphogenic