Targeted delivery system for juxtacrine signaling growth factor based on rhBMP-2-mediated carrier-protein conjugation

Bone 39 (2006) 825 – 836 www.elsevier.com/locate/bone

Targeted delivery system for juxtacrine signaling growth factor based on rhBMP-2-mediated carrier-protein conjugation Hsia-Wei Liu a,b,1 , Chih-Hwa Chen b,1 , Ching-Lin Tsai c , Ging-Ho Hsiue a,⁎ a

b

Department of Chemical Engineering, National Tsing Hua University, 101, Section 2, Kuang Fu Road, Hsinchu 30013, Taiwan Department of Orthopaedic Surgery, Chang Gung Memorial Hospital, Chang Gung University, College of Medicine, Taoyuan, Taiwan c Department of Orthopaedic Surgery, National Taiwan University Hospital, Medical College Taipei, Taiwan Received 10 December 2005; revised 23 February 2006; accepted 6 April 2006 Available online 16 June 2006

Abstract We propose a model of artificial juxtacrine signaling for the controlled release of recombinant human bone morphogenetic protein-2 (rhBMP-2) suitable for guided bone regeneration. A porous three-dimensional scaffold of poly-(lactide-co-glycolide) was fabricated by means of gel molding and particulate leaching. Collagen immobilization onto the scaffold surface was produced by performing photo-induced graft polymerization of acrylic acid, and rhBMP-2 was tethered to the collagenous surface by covalent conjugation. On pharmacokinetic analysis, in vitro enzyme-linked immunosorbent and alkaline phosphatase assays revealed sustained, slow release of rhBMP-2 over 28 days, with a cumulative release of one third of the initial load diffusing out of the scaffold. Conjugation of rhBMP-2 inhibited the free lateral diffusion and internalization of the activated complex of rhBMP-2 and the bone morphogenetic protein receptor. Osteoprogenitor cells were used as bone precursors to determine the expression of biosignaling growth factor in regulating cell proliferation and differentiation. To identify the phenotype of cells seeded on the rhBMP-2-conjugated scaffold, cellular activity was evaluated with scanning electron microscopy and with viability, histological, and immunohistochemical testing. The rhBMP-2-conjugated scaffold prolonged stimulation of intracellular signal proteins in cells. Enhancement of cell growth and differentiation was considered a consequence of juxtacrine signaling transduction. Animal studies of rhBMP-2-containing filling implants showed evidence of resorption and de novo bone formation. The present study revealed the potential of biomimetic constructs with co-immobilized adhesion and growth factors to induce osteoinduction and osteogenesis. Such constructs may be useful as synthetic bone-graft materials in orthopaedic tissue engineering. © 2006 Elsevier Inc. All rights reserved. Keywords: Conjugation; Growth factor; Juxtacrine; Osteoprogenitor; Tissue engineering

Introduction Tissue-engineering approaches for osteoinductive bonegraft extenders rely on the stimulation of biosignaling proteins, such as cytokines or growth factors, to induce host–cell chemotaxis, proliferation, differentiation, and new-tissue formation at the site of bone deficiencies. Although many intercellular signals function as soluble proteins, certain growth factors act as ligands for cell receptors in membrane-bound forms, and their binding triggers an intracellular cascade of biochemical reactions that change cell behavior or function. Therefore, various protein-delivery systems have been devel⁎ Corresponding author. Fax: +886 3 5726825. E-mail address: [email protected] (G.-H. Hsiue). 1 Authors contributed equally. 8756-3282/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2006.04.027

oped in the field of tissue engineering [1,8,48]. Hubbell et al. has pursued schemes of growth factor delivery from biopolymeric cell ingrowth matrices that prevent diffusion of growth factor from the matrix but permits its sustained release under the control of matrix-degrading enzymes locally activated by cells [52,61,62]. Polymeric scaffolds serve a central role in the field of tissue engineering by directing cellular processes based on the structural and biochemical properties of the scaffold. The ability of scaffold to regulate cell behavior should be designed by mimicking the native extra cellular matrix and require control over surface chemistry and microstructure. To overcome this drawback of the synthetic materials, many different biologically functional molecules may be either physical adsorption or covalent attachment for surface engineering. These surfaces can be designed to present specific cell adhesion sequences at

826

H.-W. Liu et al. / Bone 39 (2006) 825–836

controlled densities [11], or can be functionalized with growth factor [41,51,62]. Importantly, the scaffold must provide the appropriate signals to direct cellular cellular processes that lead to tissue formation. Previous investigators have generally believed that growth factors in solution stimulate cells only as diffusible proteins with autocrine, paracrine, or endocrine mechanisms to transduce signals for proliferation and differentiation. Massagué coined the term juxtacrine signaling to indicate the process of cellular communication in which both biosignaling molecules and receptors are anchored in the cell membrane, where they can deliver intercellular signals while supporting adhesive interactions between cells [43]. New stimulation mode, juxtacrine or matricrine stimulation has been reported to regulate cell functions. This mode is based on the discovery of membraneanchored growth factor proteins, such as heparin-binding epidermal growth factor, tumor necrosis factor, and transforming growth factor-β [44]. In addition, some researchers have found that immobilized growth factors work by means of a juxtacrine or matricrine mechanism [18,29,30,33]. Ito et al. demonstrated that artificial juxtacrine stimulation by insulin immobilized on a solid matrix enhanced cell growth and showed mitogenic effects [28]. To visualize the effect of immobilized growth factors, micropatterns of growth factors have been immobilized on matrices [26]. Subsequently, several biosignal molecules were immobilized on various matrices and their biological activities were reported. It was shown that insulin and epidermal growth factor stimulated cell growth even after immobilization [5,25]. In other words, this type of stimulation by non-diffusional growth factors enabled us to regulate tissue formation with artificial biomaterials. Bone morphogenetic protein (BMP)-induced signal transduction is an important positive regulator of osteoblastic growth and differentiation [7,31,58]. In experimental models, recombinant human BMP (rhBMP) and other members of the transforming growth factor-β superfamily possessed potent bone-forming activity [35,53,60]. Delivered in a suitable matrix, rhBMP-2 can potentially repair local skeletal defects by inducing new bone formation from undifferentiated pluripotent stem cells in host tissue [56]. Polymeric scaffolds fabricated by using various methods are suitable materials for bone repair because they act as temporary substrates for anchorage-dependent osteoblasts [24,39,42]. Collagen is natural candidates for the delivery of BMP since the major protein component of bone is type I collagen. Collagen has been used as a carrier for rhBMP-2 and BMP-7 in experimental systems [6,50]. For example, inactivated and demineralized bone matrix, mostly type I collagen, was used to deliver BMP-2-transfected bone-marrow cells [37,38]. Bonadio and Goldstein have reported success with collagen and DNA plasmids encoding human PTH1-34 and mouse BMP-4 [2,3,10]. Collagen sponges currently are being evaluated in both preclinical and preliminary clinical studies for rhBMP-2 delivery. Preliminary clinical results of rhBMP-2 delivered with absorbable collagen sponges for maxillary sinus floor augmentation have been presented [4]. Recently, biologically active rhBMP-2 has also been immobilized on succinylated type I atelocollagen, studies of alkaline phospha-

tase activity confirmed the effectiveness of rhBMP-2 immobilized on succinylated atelocollagen in augmenting cellular activity [12]. A biodegradable and porous framework that can localize and protect the payload, release that payload in a predictable and temporally controlled fashion, and deliver the signaling protein to the target is essential to express activity and to control the spatial configuration of the signaling-induced bone mass [16,17,32,34,49,55]. To this end, we designed a bioactive complex by developing a poly-(lactide-co-glycolide) (PLGA) matrix to which collagen was immobilized by grafting it with poly-(acrylic acid) and to which rhBMP-2 was tethered by covalent conjugation, as previously reported [36,46,54,59]. This architecture provided a delivery system offering prolonged retention, along with relevant growth-factor induction of signaling pathways in bone regeneration (Fig. 1). Such bioactive surfaces can influence cells and tissues by chemotactic as well as juxtacrine mechanisms. The purpose of this study was to examine the ex vivo potential of artificial juxtacrine-signaling therapy to provide alternate control of cell function. Osteoprogenitor cells (OPCs), which are pluripotent progenitor cells, were used as bone precursors to determine the expression of biosignaling growth factor in regulating cell proliferation and differentiation. Acellular scaffolds were also implanted into animals in vivo to allow cells to migrate onto the surface of the material and form new tissue. Materials and methods Fabrication of the scaffold and surface modification A porous three-dimensional matrix was fabricated by mixing PLGA copolymer (75:25, inherent viscosity of 0.86 dl/g; Purac Biochem, Gorinchem,

Fig. 1. Schematic diagram of stimulation modes of cell growth-factor proteins. (a) Proposed model of matricrine growth factor contribution to bone regeneration. BMP receptor (BMPR) and matrix-anchored growth factors are localized to the lateral membrane to maintain biological activity and direct differentiation without receptor internalization. (b) Injury of the cell layer by scratching it breaks the juxtacrine association between the ligand and the receptor in the disrupted area so soluble growth factors can diffuse and act on surrounding cells in the paracrine manner. (c) When the injured area is filled with matrix-anchored growth factors, signals are transmitted to neighboring cells in the juxtacrine manner to induce cell proliferation and migration.

H.-W. Liu et al. / Bone 39 (2006) 825–836 the Netherlands) in a 1:9 ratio with NaCl by means of gel molding and particulate leaching designed in our laboratory. In brief, fine PLGA particles were dissolved in dichloromethane at concentrations of 20% w/v. Pore sizes was controlled by using sieved NaCl particulates with a diameter of 210–350 μm. Gel mixtures were put into a cylindrical mold and compressed to obtain a diskshaped scaffold. The polymeric scaffold was immersed in water for 48 h to leach out the salts and then freeze-dried overnight. The PLGA scaffold was placed in 10 ml of 30% H2O2 solution and irradiated under ultraviolet light (300 W) at 40°C for 40 min to introduce –OOH groups into the surface [40]. After the photo-oxidization reaction was completed, the scaffold was subjected to photo-induced graft polymerization of 5% acrylic acid with 0.1% Mohr's salt at 40°C for 1 h. The poly-(acrylic acid)-grafted scaffold was immersed in 10 mg/ml 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (Pierce Biotechnology, Rockford, IL, USA) solution and reacted for 4 h at 4°C. Type I collagen (1 mg/ml in 0.5 M acetic acid; Sigma-Aldrich Corporation, St. Louis, MO, USA) was covalently immobilized onto the surface of the activated scaffold for 24 h at 4°C (Fig. 2a). After reaction, the collagengrafted scaffold was washed in 0.1-M Na2HPO4 (pH 9.1) for 1 h, washed with deionized water, and then soaked in the collagen solution (2.5 mg/ml). The collagen-grafted and -coated (CGC) scaffold was frozen at − 80°C for 12 h and lyophilized under a vacuum.

Conjugation of rhBMP-2 to carrier protein Purified, lyophilized rhBMP-2 (R&D Systems, Abingdon, UK) was reconstituted with sterile 4-mM HCl containing 0.1% bovine serum albumin. To impregnate the solid scaffold with rhBMP-2, we used a process that allowed for sequential coupling of the PLGA–CGC scaffold and growth factor with

827

1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride and Nhydroxysuccinimide (Pierce Biotechnology) (Fig. 2b). The PLGA–CGC scaffold was added to activation buffer consisting of 0.1-M 2-(N-morpholino) ethanesulfonic acid, 0.5-M NaCl (pH 6.0) with 20-mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride and 5-mM N-hydroxysuccinimide. It was mixed and incubated for 2 h, and 250 or 500 ng/ml of rhBMP-2 was reacted with the activated CGC surface for 4 h at 4°C. Finally, the PLGA–CGC–BMP scaffold was washed with phosphate-buffered saline and then washed with deionized water and lyophilized under a vacuum. The surface morphology and the porous structures inside the scaffold were observed under a scanning electron microscope (S-5000; Hitachi, Tokyo, Japan). To assess the effectiveness of immobilization of rhBMP-2 onto the collagengrafted and -coated surface, rhBMP-2 immobilized samples were reacted with BMP-2 polyclonal antibody (1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) in phosphate-buffered saline at room temperature for 1 h. The content of rhBMP-2 immobilized on CGC surface was not quantitatively measured, but can be estimated indirectly from the initial concentration of BMP-2 antibody in the PBS solution and the postreaction solution contents, and the OD of the aqueous phase was then measured at 280 nm in a spectrophotometer.

Isolation of OPCs and osteogenic cultivation Primary OPCs were obtained by stripping the periosteum from the tibia of New Zealand white rabbits. In brief, tibial periosteum was isolated and cut into chips. The chips were placed in a solution of 0.25% trypsin and 0.1% ethylenediaminetetraacetic acid for 30 min at 37°C and shaken in 1 mg/ml of type I collagenase digestive solution for 90 min at 37°C. After washing and centrifugation, the pellets were resuspended in high-glucose Dulbecco's Modified

Fig. 2. Schematic illustration of (a) immobilization of collagen (Coll) to the surface of the PLGA scaffold by using photo-induced graft polymerization and (b) rhBMP-2 tethering to the collagen-grafted and -coated surface by means of covalent conjugation with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS).

828

H.-W. Liu et al. / Bone 39 (2006) 825–836

Eagle Medium supplemented with 10% fetal bovine serum (HyClone, South Logan, UT, USA), 1% penicillin–streptomycin (Biological Industries, Kibbutz Beit Haemek, Israel), and 0.1 μM dexamethasone, with 50 μg/ml L-ascorbic acid and 10 mM β-glycerol phosphate added to promote the osteoblastic phenotype. The OPCs were plated on tissue-culture polystyrene dishes (Corning Incorporated, Acton, MA, USA) at a cell-seeding density of 1 × 104 cells/cm2 and incubated at 37°C in 5% CO2 in air at 95% humidity. After 7 days, the confluent monolayers were propagated with 0.25% trypsin and 0.1% ethylenediaminetetraacetic acid. They were then centrifuged, resuspended in the medium, counted, and plated on 75 T flask (Corning Incorporated). OPCs obtained after five passages were used in the current study.

Release pharmacokinetics of rhBMP-2 To determine the growth-factor retention characteristics of matricrine model, PLGA–CGC scaffolds containing different amounts of rhBMP-2 were evaluated in a payload-release study. The scaffolds were placed in microcentrifuge tubes with 1 ml of Hank's balanced salt solution (BIOCHROM AG, Berlin, Germany) and incubated at 37°C in 5% CO2 at 95% humidity. The supernatant medium was completely removed and replaced with fresh Hank's balanced salt solution every 2 days for up to 28 days. The supernatant samples were stored at −20°C until they were used for rhBMP-2 determination. The concentration in the collected supernatant and the conjugated efficiency of active rhBMP-2 retention (E%) on the surface of the scaffold were calculated by using an enzyme-linked immunosorbent assay (BMP-2 ELISA kit; R&D Systems) as follows: E% = (Ci − Cf ) / Ci, where Ci was the initial concentration, and Cf was the final concentration of conjugate. The light absorbance at 450 nm was used for calibration and quantification.

Bioactivity of released rhBMP-2 To assess the bioactivity of rhBMP-2 released from rhBMP-2 conjugated CGC scaffold and rhBMP-2 coated CGC scaffold, an assay for alkaline phosphatase (ALP) activity was performed on monolayer OPCs (2 × 104 cells/ml) incubated with the released rhBMP-2 supernatant collected from each time point for 4 days. After 4 days, cells were lysed with 200 μl of extraction reagent (CelLytic-M; Sigma-Aldrich), and 40-μl aliquots of the cell extract were incubated with 100 μl of p-nitrophenyl phosphate solution at 37°C for 30 min.

The reaction was stopped by using 100 μl of 1-N sodium hydroxide, and the rate of p-nitrophenol production was measured at 405 nm. ALP activity was expressed as nanomoles of converted p-nitrophenol per minute per milligrams of protein. The total protein content of each lysate was measured by using a commercial protein assay kit (BCA; Pierce Biotechnology) according to the manufacturer's instructions.

Cell-seeded scaffolds To evaluate cell behavior relative to the microstructure in the pores, the scaffolds were pre-wet by submerging them in filtered 70% ethanol for 30 min, rinsed in sterile phosphate-buffered saline for 1 h, and treated with ultraviolet light overnight. OPCs were subsequently loaded onto scaffolds at a concentration of 2×104 cells/ml on 24well plates and cultured in a humidified 5% CO2 incubator at 37°C for 2 weeks. The cell-seeded scaffolds were then characterized ex vivo by means of morphological, proliferation, viability, differentiation, histochemical, and immunohistochemical study.

Scanning electron microscopy After 7 and 14 days of incubation, the cells-seeded scaffolds were fixed with 2.5% glutaraldehyde in 0.1-M cacodylic acid buffer (pH 7.4) for 1 h at 4°C. They were then dehydrated in graded alcohols, critical-point dried, sputter coated with platinum in vacuum, and analyzed in scanning electron microscope at an accelerating voltage of 5 kV.

Cell proliferation and viability OPCs were counted by using 3-(4, 5-dimethylthiazol-2-yl)-2, 5diphenyltetrazolium bromide (MTT assay; Sigma-Aldrich) after they were cultured on the scaffolds for 2, 4, 6, 8, 10, 12, and 14 days to estimate cell proliferation. At each time point, the medium was refreshed, and 500 μl of 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (0.5 mg/ml) was added to each well and the cultures were continued for 4 h at 37°C. At the end of the assay, the purple formazan-reaction precipitate was dissolved by adding 0.5 ml dimethylsulphoxide at room temperature and transferred to a 96well plate. The colorimetric changes were quantified by using a

Fig. 3. Scanning electron micrographs of (a) the microporous PLGA scaffold, (b) the surface pattern of the poly-(acrylic acid)-grafted PLGA scaffold, and (c) the collagen immobilized onto the surface of the scaffold. In panel d, view of the rhBMP-2 tethering conjugate reveals flakelike shapes around the pore walls.

H.-W. Liu et al. / Bone 39 (2006) 825–836 spectrophotometric microplate reader (MRX II; Dynex Technologies, Chantilly, VA, USA) at a wavelength of 570 nm. The absorbance was proportional to the number of cells attached to the scaffolds. Viability was qualitatively viewed by using live-dead fluorescent stains (Molecular Probes, Eugene, OR, USA). Double-staining cells anchored on the scaffolds were incubated for 45 min in a solution of 4-mM Calcein AM and 2-mM ethidium homodimer-1. Cut surfaces of the scaffolds then evaluated by using a confocal laser scanning microscope (TCS SP2; Leica Microsystems, Wetzlar, Germany).

Cell differentiation ALP activity, a relative marker for the differentiation of osteoblastic cells, was assessed for OPCs seeded on the 500-ng/ml PLGA–CGC–BMP and control scaffolds over 14 days in culture. Unconjugated scaffolds (PLGA–CGC coated with 500 ng/ml rhBMP-2) were simultaneously evaluated for comparison. At the end of the prescribed interval, cells were lysed and analyzed for ALP activity and total protein, as described previously.

Histochemistry and immunohistochemistry OPCs seeded on the PLGA–CGC and PLGA–CGC–BMP scaffolds were prepared for histochemical and immunohistochemical analysis after 14 days in culture. Sirius red and Hoechst 33258 dye (Aldrich, Milwaukee, WI, USA) were used to detect collagen-fibril synthesis and to visualize cells in the constructs, respectively. Samples were fixed with 4% paraformaldehyde for 10 min, permeabilized in 0.1% Triton X-100 for 5 min, blocked with 5% donkey serum, and rinsed in phosphate-buffered saline containing 2% bovine serum albumin. Reactivity to the osteocalcin-specific polyclonal antibody V-19 (1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and fluorescein isothiocyanate-conjugated, affinity-purified secondary antibody (1:200; Chemicon International, Temecula, CA, USA) were assessed to detect protein expression and localization. Intact samples were viewed with the confocal laser scanning microscope. Fluorescent cells in the scaffolds indicated positive osteocalcin expression.

829

Results Scaffold morphology On scanning electron microscopy, the PLGA scaffold had a uniformly distributed and irregular and interconnected pore structure, with heterogeneous pores with a mean inner diameter of 250 μm and a porosity of approximately 80–90% (Fig. 3a). The surface morphology of poly-(acrylic acid)-grafted PLGA scaffold exhibited a fine texture covering the surface, which was a result of the organic groups immobilized on the surface of the scaffold (Fig. 3b). Microsponges of collagen immobilized onto the poly-(acrylic acid)-grafted PLGA scaffold (Fig. 3c) and microsponges of rhBMP-2 tethered to the CGC layer with interconnected pore structures were formed on the surface of scaffold (Fig. 3d). View of the rhBMP-2 tethering conjugate clearly revealed flake-like shapes around the pore walls. The porosity decreased slightly after the immobilization of collagen microsponges.

Preliminary animal study Six mature, male New Zealand white rabbits weighing a mean of 3.5 kg were used to evaluate in vivo tissue for biocompatibility and bone ingrowth. The animals were kept in the facilities of the Committee of Experimental Animal Sciences, Chang Gung Memorial Hospital, in accordance with the institutional guidelines for the care and use of laboratory animals. The animals were anesthetized by using a 1:1 intravenous injection of ketamine-xylazine (Rompun; Bayer Healthcare, Leverkusen, Germany) around the surgical site. A cylindrical bone defect 5 mm in depth and 5 mm in diameter was created over the metaphysis of the bilateral medial femoral condyles with a drill bit. In three animals, PLGA control scaffold was implanted into the void. In the other three animals, PLGA–CGC–BMP scaffold was also implanted into the void. These rabbits were sacrificed at 4, 8, and 12 weeks after implantation for histological examination to assess bone healing in the defect. Sacrifice was performed by means of overdose anesthesia, and implant sites were harvested with the surrounding tissues.

Histology and mineralization Morphological characteristics of the bone healing were evaluated by means of conventional histological methods. Specimens were removed from the animal on its death and fixed in 10% formalin solution. They were decalcified, dehydrated through alcohol gradients, cleared, and embedded in paraffin blocks. Tissue blocks were cut into 5-μm sections and stained with hematoxylin–eosin to visualize the morphology of the new bone and with Masson trichrome methods to evaluate the synthesis of collagen fibrils. The formation of calcium deposits in mineralized nodules was estimated by means of von Kossa staining and 0.25% Safranin O for counterstained.

Statistical analysis Experiments were performed in triplicate per sample. Statistical analysis was performed by applying an unpaired Student t test with a confidence level of 0.05. All data were expressed as the mean and standard deviation.

Fig. 4. (a) Release pharmacokinetics of rhBMP-2 from 500-ng/ml PLGA– CGC–BMP scaffolds on enzyme-linked immunosorbent assay. Bar graph shows noncumulative release after each time point. Line graph shows percentage cumulative release. (b) Bioactivity of rhBMP-2 released from PLGA–CGC– BMP scaffolds on ALP activity assay. Values are the mean and standard deviation (n = 4).

830

H.-W. Liu et al. / Bone 39 (2006) 825–836

Fig. 5. SEM photographs of OPCs morphology observed on the scaffolds. Panel a shows the PLGA–CGC scaffold at 7 days; note the fully spread polygonal cells on the scaffold wall. Panel b shows the PLGA–CGC–BMP scaffold at 7 days; note the interconnected cell filopodia equally distributed throughout the surface. Panel c shows the PLGA–CGC scaffold at 14 days; note that the individual cells migrated and are surrounded by new synthesized pericellular matrix. Panel d shows the PLGA–CGC–BMP scaffold at 14 days; note that the vast majority of the cells proliferated and that osseous tissue almost completely covered the porous surface.

Pharmacokinetics and bioactivity of released rhBMP-2

Morphology of the cell-seeded scaffold

Activity of rhBMP-2 analysis indicated that the active protein percentage in the 500- and 250-ng/ml PLGA–CGC–BMP scaffolds were about 98.6% and 98.4%, respectively. Retention of different amounts of rhBMP-2 tethered to PLGA–CGC scaffolds (250 or 500 ng/ml) revealed that conjugated efficiency was about 79.8–83.4%. On quantitative analyses with enzyme-linked immunosorbent assay, the release of rhBMP-2 from the carrier-protein conjugate showed sustained release over the first 8 days then slow release for up to 28 days. After 2 days, cumulative release of the PLGA–CGC–BMP scaffold with 250 ng/ml of rhBMP-2 conjugated to collagen was 2.49% of the original amount and increased to 28.41% after 28 days (data not shown). After 2 days, the cumulative release of PLGA–CGC–BMP scaffold with 500 ng/ml of rhBMP-2 was 3.46% of the original amount and increased to 33.27% after 28 days (Fig. 4a). A total of 56.69 ng, or 28.41% of the initial load, was released from the 250-ng/ml PLGA–CGC–BMP scaffold, and 138.69 ng, or 33.27% of the initial load, was released from the 500-ng/ml PLGA–CGC–BMP scaffold. The conditioned media with rhBMP-2 released from the PLGA–CGC–BMP scaffold over 28 days were incubated with OPCs for 4 days and harvested to test for ALP activity. The rhBMP2 standard curve showed a dose-dependent increase in ALP activity (data not shown). ALP activity was proportional to the concentration intermixed. That is, the PLGA–CGC scaffold conjugated with 250 ng/ml of rhBMP-2 (low dose) released low levels of active rhBMP-2, and the scaffold conjugated with 500 ng/ml rhBMP-2 (high dose) released high levels of active rhBMP-2 over 28 days, as reflected in the level of ALP expression (Fig. 4b).

Fig. 5 shows the morphology of OPCs observed on the scaffold after 7 and 14 days in cell culture, as observed under scanning electron microscopy. Cells attached to the walls of the PLGA–CGC scaffold were polygonal (Fig. 5a). Cross-sections of the PLGA–CGC–BMP scaffold obtained after 7 days clearly revealed that the well-spread OPCs were equally distributed throughout the surface, with interconnected cell filopodia (Fig. 5b). In contrast, at 14 days, surfacing appeared to have occurred with newly synthesized matrix, and the cell layers draped over the surface features of PLGA–CGC scaffold, masking the underlying grainy texture (Fig. 5c). The rhBMP-2 scaffold contained the most cells per field on qualitative examination, which showed packing of multiple layers of cells that partially filled the pores and that spanned the underlying osseous surface structure (Fig. 5d). Cell proliferation and viability The viability of cells cultured for 2 weeks on the PLGA control, PLGA–CGC, and PLGA–CGC–BMP scaffolds (Figs. 6a–c). All cells-seeded constructs qualitatively maintained good cell viability. Live cells, which stained green, appeared to have formed a confluent monolayer and a normal polygonal morphology on all materials. Visual examination revealed that live cells adherent to PLGA–CGC–BMP scaffold was denser than those on the PLGA–CGC and PLGA scaffolds. This finding indicated that the cells had vitally proliferated. Few dead cells, which stained red, were observed on all three

H.-W. Liu et al. / Bone 39 (2006) 825–836

831

Fig. 6. Confocal laser scanning microscopic projection images of fluorescence cell-viability stains show (a) the control PLGA scaffold, (b) the PLGA–CGC scaffold, and (c) the PLGA–CGC–BMP scaffold. OPCs growth and attachment expressed as the number of cells adhered to the different scaffolds through a 14-day period in culture (PLGA control, n = 3) (d). Data are the mean ± standard deviation.

constructs. These results clearly demonstrated that the PLGA– CGC–BMP scaffold was more favorable than the control for the proliferation of osteoblasts. Cells adhesion and proliferation were examined on the control and synthesized scaffolds for 14 days (Fig. 6d). The cells cultured in the synthesized scaffolds all suggested relatively improved cytocompatibility after a period of seeding, but only the PLGA–CGC–BMP scaffold showed a significant difference during 8 days of culturing compared with the control. Long-term culturing on the PLGA–CGC–BMP scaffold also showed a rhBMP-2 dose-dependent increase in cell density. By 14 days, cells numbered (4.17 ± 0.25) × 105 and (3.53 ± 0.09) × 105 for the 500- and 250-ng/ml PLGA– CGC–BMP scaffolds, respectively.

Immunofluorescence Positive staining on confocal microscopy, as assessed by means of collagen production, was evident in the PLGA–CGC matrix (Fig. 8a). Collagen synthesis was significantly increased with PLGA–CGC–BMP compared with PLGA–CGC alone, and it appeared organized, indicating uniform layering of fibrils on most of the matrix (Fig. 8b). A similar increase in the cell density of a confluent layer on the PLGA–CGC–BMP matrix was observed on Hoechst 33258 labeling of nuclei. Immunohistochemical staining for osteocalcin was positive, with a dose-dependent response, after 14 days of culture. OPCs

Cell differentiation Expression of ALP activity was distinctly higher in OPCs grown on the conjugated scaffold expressed than in OPCs on the control scaffold. Peak values for ALP expression on the conjugated scaffold were 46.46 ± 3.19 and 71.56 ± 5.07 nmol/min/mg of total cellular protein, at days 8 and 14, respectively. Compared with the conjugated scaffold, cells seeded on the unconjugated scaffold had relatively low expression of ALP activity, which peaked at 37.22 ± 3.72 and 44.45 ± 3.85 nmol/min/mg of total protein, at days 8 and 14, respectively. The level of ALP expression on the PLGA– CGC and PLGA scaffolds slightly increased over 14 days in culture (Fig. 7).

Fig. 7. Expression of ALP activity by OPCs cultured on the different scaffolds for 14 days (PLGA control, n = 3). Data are the mean ± standard deviation.

832

H.-W. Liu et al. / Bone 39 (2006) 825–836

Fig. 8. Confocal micrographs of collagen-fibril formation (Sirius red fluorescence) on (a) the PLGA–CGC scaffold and (b) the PLGA–CGC–BMP scaffold labeled with Hoechst 33258 (blue fluorescence) for nuclear staining. Fluorescence photographs of (c) unconjugated rhBMP-2 scaffold (faint staining) and (d) conjugated rhBMP-2 scaffold (intense fluorescence). Note the autofluorescence of the modified-PLGA scaffold. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

with spindle and polygonal morphologies were observed on the surface. The unconjugated rhBMP-2 scaffold showed faint staining (Fig. 8c), and the conjugated rhBMP-2 scaffold showed intense fluorescence (Fig. 8d).

histochemical staining after 8, and 12 weeks of induction (Figs. 9h, i). The mineralized nodule formation was observed mainly at peripheral sites surrounding the new trabecular bone in all animals.

De novo bone formation

Discussion

On histological hematoxylin–eosin staining analysis, PLGA control (Figs. 9a–c) and PLGA–CGC–BMP scaffolds induced de novo bone at 4, 8, and 12 weeks. At 4 weeks after surgery, almost all pores in the central sites were evenly filled with trabecular bone and hematopoietic marrow (Fig. 9d). At 8 weeks, the junctions between the implant and the host bone were fused, with new bone forming outside the implant and in the marrow cavity of the bone adjacent to the implant (Fig. 9e). At 12 weeks, prominent bone formation covered almost the entire implant (Fig. 9f). In contract, in the control group, little bone tissue was found in the pores at central sites of the implants, and most peripheries around implant were filled with fibrous tissue at 4 weeks after implantation (Fig. 9a). Masson trichrome stains showed collagen–fibril formation, indicating uniform layering of fibrils on most of the matrix of PLGA– CGC–BMP implant (Fig. 9g). Osteogenic differentiation assays showed that most of the mineralized tissue of experiment group had calcium deposits, with positive von Kossa/ Safranin-O

BMPs are multifunctional cytokines and members of the transforming growth factor-β superfamily on the basis of their homology in amino acid sequences. BMP binds to type I and type II serine–threonine kinase receptors. Two BMP type I receptors and one BMP type II receptor have been identified in mammals [31]. The refolded rhBMP-2 is a disulfide-linked homodimer. Each 113 amino acid residue monomer has a calculated molecular mass of 12.8 kDa. The rhBMP-2 is ∼26 kDa, contains numerous lysine/arginines (12/6 per molecule). The rhBMP-2 is one of the most potent osteoinductive signal proteins and can regulate various cellular functions, such as growth, differentiation, secretion, and apoptosis. Its combination with a biomaterial to deliver it to a target site is essential to express rhBMP-2 activity and to control the three-dimensional configuration of the BMPinduced bone mass [45]. Biomaterial–cellular interactions are based on the binding of cell surface receptors with ligands, which promote specific cellular responses [47]. Growth factor protein interacts with the

H.-W. Liu et al. / Bone 39 (2006) 825–836

833

Fig. 9. Histological cross-sections of the controlled PLGA implant at (a) 4 weeks, (b) 8 weeks, (c) 12 weeks after surgery. rhBMP-2 containing implant at (d) 4 weeks, almost all pores in the central sites were evenly filled with trabecular bone and hematopoietic marrow, and newly formed bone islands were observed; at (e) 8 weeks, the junctions between the implant and host bone were bony fused, involving new bone formation outside the implant and in the marrow cavity of the bone adjacent to the implant; at (f) 12 weeks group, prominent bone formation covered almost the entire implant. In the experiment groups at (g) 4 weeks, note blue areas in the synthetic implant and new trabecular bone with positive, organized collagen-fibril formation; at (h) 8, and (i) 12 weeks after surgery, bone marrow stromal cells were able to differentiate into osteolineages, with black calcium deposits inside and outside of the cells. (hematoxylin–eosin in a–f; Masson Trichrome stains in g; von Kossa and Safranin-O stains in h, i) BM indicates bone marrow; HB, host bone; NB, new bone; SF, scaffold. Scale bars = 200 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

cognate receptor on the cell surface to form a complex, and the complex aggregates on the cell surface before it is internalized into the cell. The effect of growth-factor binding can be attenuated by endocytosis and subsequent degradation of the receptor–ligand complex, so, by presenting immobilized ligands in a conformation that maintains biological activity but inhibits receptor internalization, the intracellular signaling cascade may be sustained for longer times [1]. The receptor–ligand interaction can be controlled by modifying the surface of the implant. Many solid-surface immobilization techniques have been used [13,23]. We previously demonstrated that the introduction of an acrylic-acid segment as spacer arm onto the functional surface was useful to further graft various bioactive macromolecules [19–22,36,57]. To eventually immobilize large biomolecules, a two-step procedure consisting of photo-induced graft polymerization and bioconjugation is needed to prepare a PLGA-grafted carrier protein for matricrine signaling of growth factor. Because collagen is one of the major biological components of the extracellular matrix in bone tissues, it has received increasing attention in recent years because of its excellent biocompatibility and its suitable interaction with cells and other macromolecules. Therefore, we selected a collagen-grafted surface. Collagen contains a large number of reactive terminal and side chain residues such as amino and carboxylic acid groups are ideal for rhBMP-2 coupling since –NH2 groups (α-NH2

group at N-terminus and ε-NH2 groups in lysines and arginines) are abundant in proteins, and conjugation with amide bond is effective at physiological buffers and temperatures. Our project entailed the design of a polymeric implant system that directly influences cellular behavior by means of covalent binding reaction of the collagen-grafted surface with rhBMP-2. Direct covalent conjugation has great potential because of the stability of the resultant bond. The immobilized growth factor is a good model for explaining the mechanism of juxtacrine stimulation [27]. Tethered rhBMP-2 provides sustained stimulation along relevant growth factor-induced Smad signal transduction pathways and thus ultimately dictates cell behavior in the context of bone regeneration. Profiles for rhBMP-2 release from the PLGA–CGC–BMP scaffolds demonstrated that approximately one third of the initial load of rhBMP-2 diffused from the scaffolds, unlike the rapid release observed in collagen sponges [9] and poly-(αhydroxyacids) carriers [15]. The high retention of rhBMP-2 that we observed – combined with its slow release from a degradable, interconnected, porous, space-maintaining, and moldable scaffold – suggests that the introduction of collagen microsponges into the poly-(acrylic acid)-grafted scaffold results in nondiffusion-based release. The work described here demonstrates that rhBMP-2 can be introduced on the PLGA scaffold by covalent attachment onto the collagen-grafted and -coated surface is feasible. To control exposure and concentration,

834

H.-W. Liu et al. / Bone 39 (2006) 825–836

retention and/or release of growth factor rhBMP-2 from scaffold surfaces, direct covalent attachment has the greatest potential for the prolonged retention of rhBMP-2 due to the stability of the resultant bond. In the reaction 1-ethyl-3-(3dimethylaminopropyl) carbodiimide and N-hydroxysuccinimide are used as they convert the carboxylic acid into a more stable reactive intermediate which is susceptible to attack by amines. The synergistic effect of co-immobilized adhesion and growth factors might be explained as follows: The basic, aminerich N-terminus of rhBMP-2 may covalently interact with the acidic, carboxylic acid residues of collagen and create tight binding to maintain rhBMP-2 bioactivity over the long term. It is possible that the initial rhBMP-2 release is controlled by desorption of the non-covalent binding growth factor from the collagen-grafted and -coated surface, which subsequently becomes depleted of rhBMP-2. Non-binding rhBMP-2 is initially retained on the surfaces by weak physiosorption forces, then, the adsorbed chains would be readily desorbed from the surface. The amine-rich N-terminus chains of rhBMP-2 chemically tethered to the collagen-grafted and -coated surface are in solid state and produce no novel surface unless they are in contact with their solvent or swelling agents. The tethered polymer chains become semi-soluble only when a solvent or swelling agent is present [23,24]. In this study, after immersing in the Hank's balanced salt solution (pH = 7.4) for 8 d there was very little decrease in the collagen amounts in the PLGA scaffold, for collagen is only soluble in acidic solution. The later release phase is strongly dependent on the collagen properties, such as hydrophilicity and swelling rate. Collagen swells to a certain extent by exposure to water. But due to its special sterical arrangement, native collagen can only be digested completely when collagenases participate. Degradation of the collagenrhBMP-2 conjugates requires water and enzyme penetration and digestion of linkages. Afterwards, growth factor release is governed by diffusion through the intact and water swollen collagen barriers. If degradation of PLGA occurs, the release rate becomes additionally erosion dependent [14]. Thus, possible growth factor rhBMP-2 release from monolithic scaffolds is based on desorption, swelling, diffusion, matrix erosion or combinations. Cellular affinity on the scaffold is an important factor in evaluating the biocompatibility of a biomaterial and in cell attachment, whereas spreading is crucial for cell growth and differentiation. Our results clearly demonstrated that the pores of scaffolds were highly interconnected, a feature important for cell penetration and growth. Although cell attachment occurred on different types of scaffolds, scanning electron microscopy showed that cells on the scaffolds containing rhBMP-2 were able to form a dense, extracellular matrix-like layer at the periphery. Consistent stimulation by prolonged rhBMP-2 retention enhanced the migration and proliferation of OPCs in a dose-dependent manner. Moreover, cell proliferation on these scaffolds was significantly improved over that of the rhBMP-2conjugated PLGA–CGC scaffold after 7 days of incubation. The results showed high cell-attachment levels due to the chemotactic effect of immobilized rhBMP-2. In addition, in result similar to that of the MTT assay, the rhBMP-2 conjugated

scaffold had higher ALP activity than that of unconjugated scaffold, regardless of when the culture was examined. Of note, rhBMP-2 conjugated scaffolds transduced a signal longer than did soluble growth factor, and they sustained excellent osteogenic expression ex vivo. Osteocalcin transiently appeared in embryonic bone at the time of mineral deposition. This effect accelerated mineralization followed by death of the osteoblasts as they underwent terminal differentiation. The interconnected and highly porous structure of the scaffold allowed for efficient migration of bone-producing cells from pore to pore and for vascular invasion, which is essential for new bone formation. The type I collagen was biocompatible and biodegradable, with no immunogenic potentiality; therefore, it did not interfere with new bone formation in the pores. The animal study demonstrated that the PLGA–CGC– BMP scaffold induced new bone formation in pores and around the implant, leading to complete repair of cancellous bone defects in the rabbit femoral condyle. The implant containing rhBMP-2 achieved histological union at the junctional sites, indicating the initiation of bone ingrowth. Extensive cellular activity on the interface between the peripheral bone and the implant suggested active bone ingrowth at 8 weeks. De novo bone ingrowths almost completely replaced the bone void by 12 weeks. Mineralization in the structures observed in this preliminary animal study indicated that rhBMP-2 could elicit osteogenesis, as detected with von Kossa staining. Therefore, it was reasonable to think that a carrier tethers or slowly releases rhBMP-2 at the site of implantation, biologically stimulating cells surrounding the carrier to induce bone formation. In summary, rhBMP-2 chemically tethered onto an interconnected and porous scaffold promoted osteoblastic adhesion, spread, growth, and differentiation. Matrix-bound growth factors can transmit signals to only neighboring cells in the juxtacrine manner, whereas soluble growth factors can diffuse and act on surrounding cells in the paracrine manner. In the case of juxtacrine signal transmission, bidirectional signaling is also possible. Differentiation was due to rhBMP-2 retained on the scaffold, to rhBMP-2 released from the scaffold in solution, or to both mechanisms. However, the expression of the osteoblastic phenotype of OPCs seeded on the PLGA–CGC– BMP scaffold demonstrated that rhBMP-2 was bioactive and could direct the differentiation of pluripotent stem cells. Our research was to develop the matricrine signaling growth factor transduction to prolong its retention for use in bone healing that would provide localized release in a manner that is triggered by cellular activity. The bound factors were competent in inducing cell proliferation, the matrix-bound forms being more effective than native growth factor. Our results indicate that BMP-2 can be bound to a carrier-protein, in this case collagen, and retain its ability to increase matrix production. Not only did the results demonstrate the growth factor immobilized within this delivery system could be released in active form; moreover, an enhancement of osteopromotion was observed in animal study. Our findings revealed the potential for biomimetic constructs with coimmobilized adhesion and growth factors to induce

H.-W. Liu et al. / Bone 39 (2006) 825–836

osteoinduction and osteogenesis. These constructs may provide a novel platform for tissue-engineered orthopaedic therapies. Acknowledgments This study was supported by grant DOH94-TD-1128 from the Department of Health, Taiwan. References [1] Baldwin SP, Saltzman WM. Materials for protein delivery in tissue engineering. Adv Drug Deliv Rev 1998;33:71–86. [2] Bonadio J, Goldstein SA, Levy RJ. Gene therapy for tissue repair and regeneration. Adv Drug Deliv Rev 1998;33:53–69. [3] Bonadio J, Smiley E, Patil P, Goldstein SA. Localized, direct plasmid gene therapy in vivo: prolonged therapy results in reproducible tissue regeneration. Nat Med 1999;5:753–9. [4] Boyne PJ, Marx RE, Nevins M, Triplett G, Lazaro E, Lilly LC, et al. A feasibility study evaluation rhBMP-2/ absorbable collagen sponge for maxillary sinus augmentation. Int J Periodont Restor Dent 1997;17:11–25. [5] Chen G, Ito Y, Imanishi Y. Mitogenic activity of water-soluble and waterinsoluble insulin conjugates. Bioconj Chem 1997;8:106–10. [6] Cook SD. Preliminary report: new technology: preclinical and clinical evaluation of osteogenic protein-1 (BNMP-7) in bony sites. Orthopedics 1999;22:669–71. [7] de John DS, Steegenga WT, Hendriks JMA, van Zoelen EJJ, Olijve W, Dechering KJ. Regulation of notch signaling genes during BMP2-induced differentiation of osteoblast precursor cells. Biochem Biophys Res Commun 2004;320:100–7. [8] Fu K, Kibanov AM, Langer R. Protein stability in controlled-release systems. Nat Biotechnol 2000;18:24–5. [9] Geiger M, Li RH, Friess W. Collagen sponges for bone regeneration with rhBMP-2. Adv Drug Deliv Rev 2003;55:1613–29. [10] Goldstein SA, Bonadio J. Potential role for direct gene transfer in the enchancement of fracture healing. Clin Orthop 1998;355:S154–62. [11] Hersel U, Dahmen C, Kessler H. RGD modified polymers: biomaterials for stimulated cell adhesion and beyond. Biomaterials 2003;24:4385–415. [12] Hidetsugu T, Hitoshi N, Mehmet G, Andrea R, Rosario SR, Racquel ZL, et al. Effects of immobilized recombinant human bone morphogenetic protein-2/succinylated type I atelocollagen on cellular activity of ST2 cells. J Biomed Mater Res Part A 2005;75:210–5. [13] Hoffman AS. Immobilization of biomaolecules and cells on and within polymeric biomaterials. Clin Mater 1992;11:61–6. [14] Holland SJ, Tighe BJ. Polymers for biodegradable medical devices. 1. The potential of polyesters as controlled macromolecular release systems. J Control Release 1986;4:155–80. [15] Hollinger JO, Leong K. Poly (α-hydroxy acids): carriers for bone morphogenetic proteins. Biomaterials 1996;17:187–94. [16] Hollinger JO, Schmitz JP. Macrophysiologic roles of a delivery system for vulnerary factors needed for bone regeneration. Ann NY Acad Sci 1997;831:427–37. [17] Hollinger JO, Winn SR. Tissue engineering of bone in the craniofacial complex. Ann NY Acad Sci 1999;875:379–85. [18] Horwitz JI, Toner M, Tompkins RG, Yarmush ML. Immobilized IL-2 preserves the viability of an IL-2 dependent cell line. Mol Immunol 1993;30:1041–8. [19] Hsiue GH, Lee SD, Chang PCT. Platelet adhesion and cellular interaction with poly (ethylene oxide) immobilized onto silicone rubber membrane surfaces. J Biomater Sci Polym Ed 1996;7:839–55. [20] Hsiue GH, Lu PL, Chen JC. Multienzyme-immobilized modified polypropylene membrane for an amperometric creatinine biosensor. J Appl Polym Sci 2004;92:3126–34. [21] Hsiue GH, Wang CC. Glucose oxidase immobilized polyethylene-gacrylic acid membrane for glucose oxidase sensor. Biotechnol Bioeng 1990;36:811–5.

835

[22] Hsiue GH, Wang CC. Functionalization of polyethylene surface using plasma-induced graft copolymerization of acrylic acid. J Polym Sci Polym Chem 1993;31:3327–37. [23] Ikada Y. Surface modification of polymers for medical applications. Biomaterials 1994;15:725–36. [24] Ishaug SL, Crane GM, Miller MJ, et al. Bone formation by threedimensional stromal osteoblast culture in biodegradable polymer scaffold. J Biomed Mater Res 1997;36:17–28. [25] Ito Y, Li J-S, Takahashi T, Imanishi Y, Okabayashi Y, Kido Y, et al. Enhancement of the mitogenic effect by artificial juxtacrine stimulation using immobilized EGF. J Biochem 1997;121:514–20. [26] Ito Y. Surface micropatterning to regulate cell functions. Biomaterials 1999;20:2333–42. [27] Ito Y. Tissue engineering by immobilized growth factors. Mater Sci Eng 1998;C6:267–74. [28] Ito Y, Kondo S, Chen G, Imanshi Y. Patterned artificial juxtacrine stimulation of the cells by covalently immobilized insulin. FEBS Lett 1997;403:159–62. [29] Ito Y, Liu SQ, Imanishi Y. Enhancement of cell growth on growth factorimmobilized polymer film. Biomaterials 1991;12:449–53. [30] Ito Y, Zheng J, Imanishi Y, Yonezawa K, Kasuga M. Protein-free cell culture on artificial substrate immobilized with insulin. Proc Natl Acad Sci U S A 1996;93:3598–601. [31] Kawabata M, Imamura T, Miyazono K. Signal transduction by bone mophogenetic proteins. Cytokine Growth Factor Rev 1998;9:49–61. [32] Kim HD, Valentini RF. Retention and activity of BMP-2 in hyaluronic acid-based scaffolds in vitro. J Biomed Mater Res 2002;59:573–84. [33] Kuhl PR, Griffith-Cima LG. Tethered epidermal growth factor as a paradigm for growth factor-induced stimulation from the solid phase. Nat Med 1996;2:1022–7. [34] Lee KY, Peters MC, Anderson KW, Mooney DJ. Controlled growth factor release from synthetic extracellular matrices. Nature 2000;408:998–1000. [35] Lee SC, Shea M, Battle MA, et al. Healing of large segmental defects in rat femurs is aided by rhBMP-2 in PLGA matrix. J Biomed Mater Res 1994;28:1149–56. [36] Lee SD, Hsiue GH, Chang PCT, Kao CY. Plasma-induced grafted polymerization of acrylic acid and subsequent grafting of collagen onto polymer film as biomaterials. Biomaterials 1996;17:1599–608. [37] Lieberman JR, Daluiski A, Stevenson S, et al. The effect of regional gene therapy with bone morphogenetic protein-2 producing bone-marrow cells on the repair of segmental femoral defects in rats. J Bone Joint Surg Am 1999;81:905–17. [38] Lieberman JR, Le LQ, Wu L, et al. Regional gene therapy with a BMP-2 producing murine stromal cell line induces heterotopic and orthotopic bone formation in rodents. J Orthop Res 1998;16:330–9. [39] Lo H, Ponticello MS, Leong KW. Fabrication of controlled release biodegradable foams by phase separation. Tissue Eng 1995;1:15–28. [40] Ma Z, Gao C, Gong Y, Shen J. Cartilage tissue engineering PLLA scaffold with surface immobilized collagen and basic fibroblast growth factor. Biomaterials 2005;26:1253–9. [41] Mann BK, Schmedlen RH, West JL. Tethered-TGF-beta increases extracellular matrix production of vascular smooth muscle cells. Biomaterials 2001;22:439–44. [42] Mikos AG, Sarakinos G, Leite SM, et al. Laminated three-dimensional biodegradable foams for use in tissue engineering. Biomaterials 1993;14:323–30. [43] Massagué J. Transforming growth factor-α: a model for membraneanchored growth factor. J Biol Chem 1990;265:21393–6. [44] Massagué J, Pandiella A. Membrane-anchored growth factors. Annu Rev Biochem 1993;62:515–41. [45] Nimni ME. Polypeptide growth factors: targeted delivery systems. Biomaterials 1997;18:1201–25. [46] Okada T, Ikada Y. Surface modification of silicone for percutaneous implantation. J Biomater Sci Polym Ed 1995;7:171–80. [47] Puleo DA, Nanci A. Understanding and controlling the bone-implant interface. Biomaterials 1999;20:2311–21. [48] Putney SD, Burke PA. Improving protein therapeutics with sustainedrelease formulation. Nat Biotechnol 1998;16:153–7.

836

H.-W. Liu et al. / Bone 39 (2006) 825–836

[49] Richardson TP, Peters MC, Ennett AB, Mooney DJ. Polymeric system for dual growth factor delivery. Nat Biotech 2001;19:1029–34. [50] Riedel GE, Valentin-Opran A. Preliminary report: new technology. Clinical evaluation of rhBMP-2/ACS in orthopedic trauma: a progress report. Orthopedics 1999;22:663–5. [51] Sakiyama-Elbert SE, Hubbell JA. Development of fibrin derivatives for controlled release of heparin-binding growth factors. J Control Release 2000;65:389–402. [52] Sakiyama-Elbert SE, Panitch A, Hubbell JA. Development of growth factor fusion proteins for cell-triggered drug delivery. FASEB J 2001;15:1300–2. [53] Sampath TK, Maliakal JC, Hauschka PV, et al. Recombinant human osteogenic protein-1 (hOP-1) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates osteoblast proliferation and differentiation in vitro. J Biol Chem 1992;267:20352–62. [54] Tretinnikov ON, Kato K, Ikada Y. In vitro hydroxyapatide deposition onto a film surface-grafted with organophosphate polymer. J Biomed Mater Res 1994;28:1365–73. [55] Uludag H, Norrie B, Kousinioris N, Gao T. Engineering temperaturesensitive poly (N-isopropylacrylamide) polymers as carriers of therapeutic proteins. Biotech Bioeng 2001;73:510–21.

[56] Vukiević S, Stavljenić A, Pećina M. Discovery and clinical applications of bone morphogenetic proteins. Eur J Clin Chem Clin Biochem 1995;33:661–71. [57] Wang CC, Hsiue GH. Immobilization of glucose oxidase on polyethylene film using a plasma induced graft copolymerization process. J Biomater Sci Polym Ed 1993;4:357–67. [58] Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, et al. Novel regulators of bone formation: molecular clones and activities. Science 1988;242:1528–34. [59] Yamamoto M, Kato K, Ikada Y. Ultrastructure of the interface between cultured osteoblasts and surface-modified polymer substrates. J Biomed Mater Res 1997;37:29–36. [60] Yasko AW, Lane JM, Fellinger EJ, et al. The healing of segmental bone defects, induced by recombinant human bone morphogenetic protein (rhBMP-2). J Bone Joint Surg Am 1992;74:659–70. [61] Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Cell-demanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth. FASEB J 2003;17:2260–2. [62] Zisch AH, Schenk U, Schense JC, Sakiyama-Elbert SE, Hubbell JA. Covalently conjugated VEGF-fibrin matrices for endothelialization. J Control Release 2001;72:101–13.