tricalcium phosphate/collagen composites prepared ...

Porous b-tricalcium phosphate/collagen composites prepared in an alkaline condition Chao Zou,1 Wenjian Weng,1 Kui Cheng,1 Piyi Du,1 Ge Shen,1 Gaorong Han,1 Binggang Guan,2 Weiqi Yan2 1 Department of Materials Science and Engineering, Zhejiang University, Hangzhou 310027, China 2 Institute of Orthopaedic Research, The Second Affiliated Hospital, Zhejiang University, Hangzhou 310009, China Received 2 November 2006; revised 9 July 2007; accepted 1 August 2007 Published online 13 December 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.31686 Abstract: Bone substitute materials with natural bone-like structure are considered to be favorable for bone regeneration. In this work, porous b-tricalcium phosphate (b-TCP )/ collagen composite consisting of bone-like microstructural units was prepared using nanosized b-TCP particles and alkaline-disassembled collagen. The resulting composite showed a good interconnecting porous structure with 90% porosity and 100 300 lm pore size. The pore walls were dense, and the combination status of collagen and nanosized b-TCP particles demonstrated that nanosized bTCP particles tightly connected collagen microfibrils as a bone-like microstructural unit. MTT and alkaline phospha-

tase (ALP) assays showed that the porous composite had enhanced effects on cellular proliferation and activity of osteoblast compared with a control of pure collagen. It is suggested that the adoption of nanosized b-TCP particles is a main contribution to the formation of the composite with a bone-like microstructural unit, and the unique microstructure could be a main role for the composite to have the positive influence on osteoblast cell proliferation. Ó 2007 Wiley Periodicals, Inc. J Biomed Mater Res 87A: 38–44, 2008

INTRODUCTION

or responses.11,12 That calls for versatile materials to make correct response at different stages of bone generation. To meet such versatility, composite material is one of the approaches. Among them, calcium phosphate/ collagen composite is thought as one of the best choices because collagen has good cell attachment ability besides good biodegradability,13 while calcium phosphate has good biocompatibility and osteoconductivity.6,7 Hence, the composite between them might provide several necessary properties at different bone regeneration stages. Natural bone is a complex biomineralized system with an intricate hierarchical structure,14 which is assembled through the orderly deposition of apatite minerals within a type I collagenous matrix.15 It is considered that bone substitute materials with microstructure similar to natural bone will be helpful to promote bone regeneration for shortening cure time. Efforts have been made to prepare hydroxyapatite (HA)/collagen composite with bone-like microstructure. Cui and coworkers prepared HA/collagen composite by biomimetic method16 and Kikuchi et al. reported self-organized HA/collagen composite.17 These composites have showed good bioactivity. With similar chemical components, if b-tricalcium phosphate (b-TCP) and collagen are combined and a bone-like microstructure is constructed, the resulting

The effective repair of bone defects and damages needs bone substitute materials, which are necessarily to be biocompatible, osteoconductive, and osteoinductive in order to facilitate cell and tissue growth, as well as bone formation.1–3 Also, in nonload bearing conditions, biodegradability is a plus, so that the regenerated bone tissues could completely replace the material.4 Numerous bone substitute materials including ceramics5–7 and polymers8–10 have been developed for different types of defects repair. However, these materials still need improvements because of the complexity of bone regeneration, which involves complicated physicochemical and biological reactions Correspondence to: W. Weng; e-mail: [email protected] Contract grant sponsor: Research Fund of the Doctoral Program of Higher Education of China; contract grant number: 20050335040 Contract grant sponsor: Research Fund of Science and Technology Department of Zhejiang province of China; contract grant number: 2006C24009 Contract grant sponsor: Agency for Science Technology and Research of Singapore; contract grant number: 032 101 0005 Ó 2007 Wiley Periodicals, Inc.

Key words: composite; bone substitute; microstructure; collagen; TCP

PREPARATION OF POROUS b-TCP/COLLAGEN COMPOSITES

composite should have the advantages of HA/collagen composite as well as characteristic biodegradation of b-TCP. However, b-TCP phase is impossible to form in situ during wet-chemical synthesis,18 biomimetic, and self-organizing methods are not available for preparation of b-TCP/collagen composite. Our previous work reported the preparation of a porous b-TCP/collagen composite with bone-like microstructure using acidic-disassembled collagen and nanosized b-TCP particles as starting materials.19 In addition, collagen could be disassembled under aqueous alkali condition13,20 and reveal another groups for bonding to form bone-like microstructural units. In this work, an alkaline treatment is attempted to adopt for well-disassembled collagen, the resulting composites were characterized in microstructure, and evaluated in cell culture. The experimental results and the difference from acidic treatment were discussed. MATERIALS AND METHODS Materials b-TCP was synthesized by calcining amorphous calcium phosphate precursor.21 Type I collagen from bovine tendon (Sigma-Aldrich) was used for the collagen matrix, and aqueous solution of glutaraldehyde (GA, AR, Shanghai Pharm, China) was used as a cross-linking agent. All other reagents and solvents are of analytical grade and used as received. Aqueous alkali solution (pH 12) was got by dissolving sodium hydroxide into deionized water, monitor by pH Meter (320-S, Mettler Toledo).

Preparation of porous b-TCP/collagen composites Collagen was disassembled in aqueous alkali solution (pH 12) at room temperature to become collagen suspension. Certain amount of b-TCP powders were slowly added into the collagen suspension under stirring, in a bTCP/collagen weight ratio of 2:1. After a homogenous suspension was formed, the cross-linking agent of glutaraldehyde solution was added. The mixture was poured into plastic tubes, and froze in a refrigerator for 2 h at 2308C. Porous composites were obtained after further lyophilization. The porous composites were soaked for 4 days in deionized water with semidiurnal refreshing to remove free glutaraldehyde. The soaked porous composites were lyophilized again to form the samples studied in this work. As comparisons, b-TCP disk with 5 mm in diameter and 2 mm in thickness, porous collagen prepared following the same procedure, and b-TCP/collagen composite prepared at acidic condition19 were used as controls.

Characterizations The crystalline phases of the powders and porous composites were determined in an X-ray diffractometer (XRD,

39

Rigaku D/max-rA), with a step of 0.028 at a speed of 0.28/ min. For scanning electron microscopic observation, the composites were quenched in liquid nitrogen and then broken. After gold sputter-coated, the morphology of the collagen and the porous composites was observed in field emission scanning electron microscopy (SEM, FEI SIRION) at an accelerating voltage of 5 kV. For transmission electron microscopic observation, the porous composites were embedded into aqueous epoxy resins and solidified, then the block were cut into ultrathin sections by ultrathin cutter, floating on the surface of water. Before the ultrathin section was observed by TEM (TEM, JEOL 1200) with accelerating voltage of 160 kV, it was collected by carbonfilm-supported copper grids. The zeta potential of collagen and composite suspensions at different conditions was measured by ZETASIZER (ZETASIZER 3000HSa, Malvern, UK). The porosity of the porous collagen and the porous composites was measured by liquid displacement.19

Cell culture Osteoblast-like MG63 cells, which are human osteosarcoma cell line, were obtained from Institute of Cell and Biology in Shanghai. MG63 cells were cultured in MEM culture medium (GIBCO, NY) at 378C in a humidified incubator with a 5% CO2 atmosphere. The culture medium was supplemented with 0.03% L-glutamine (GIBCO, NY) and 10% fetal bovine serum (FBS), penicillin (50U/mL), and streptomycin (50 g/mL). Cells were harvested with 0.25% trypsin solution in PBS.

MTT (3-dimethylthiazol-2,5-diphenyltetrazolium bromide) colorimetric assay The MTT assay was performed to evaluate the early osteoblast proliferation on materials, as previously described.22 For each sample, 1 3 105 cells/well was suspended in the culture medium and seeded onto the samples in 96-well plate (Coastar, Corning, NY). Cells were cultured with samples for different time intervals (2, 4, 6 days), culture medium refreshed every 48 h. Before the end of incubation, 20 lL of MTT solution (5 3 103 mg/L) was added to each well and incubated at 378C for further 4 h. The inserts were removed, and the cells were rinsed with PBS. The supernatant was removed and 150 lL DMSO was added to each well. After shaking for 10 min and cooling for 5 min, the absorbance of the contents of each well was measured at 570 nm with a microplate reader.

Cell morphology After the cells were cultured on the composites for 4 days, the attached cells were in situ fixed in 29% glutaraldehyde for 2 h at 48C and then dehydrated with a series of graded ethanol solutions. The cells then were subjected to drying and sputter-coated with gold. The cell-bearing surface of each composite was examined using field emission SEM. Journal of Biomedical Materials Research Part A

40

ZOU ET AL.

Alkaline phosphatase activity Alkaline phosphatase (ALP) activity was used as an early marker for osteoblast differentiation and measured in cell lysate by determining the release of p-nitrophenol (PNP) from a p-nitrophenol phosphate solution (PNPP). At each time point (7, 14 days), the seeded samples were rinsed twice by PBS. After Triton X-100 (200 lL of 0.05%) was added to each well, the solutions were transferred to a micro centrifuge tube and frozen at 2808C for 2 h. After 3 times freezing-thawing cycles to homogenize the solutions, aliquots of the solutions were used for estimation of protein content in samples by Coomassie brilliant blue protein reagent procedure. ALP activity was quantified by reaction with 2.5 mM PNPP in 0.1M 2-amino-1-methyl-1propanol buffers (pH 9.8), supplemented with 2 mM magnesium chloride. After 30 min, 350 lL of 1M sodium hydroxide was added to stop the reaction. The absorbance of each solution was measured at the wavelength of 405 nm by spectrophotometer.

Statistical analyses All values were expressed as mean values 6 standard deviation (SD). Numerous data were analyzed using oneway analysis of variance (ANOVA) techniques. Statistical significance was considered at p < 0.05.

RESULTS Characterizations of the porous composites After the formation process of b-TCP/collagen composite, the XRD pattern of porous b-TCP/collagen composite is identified with that of the b-TCP powders in Figure 1. Pore structure of the collagen prepared under alkaline condition as control is shown in Figure 2(a), the

Figure 1. XRD patterns of TCP powders and particles in composite. Journal of Biomedical Materials Research Part A

Figure 2. SEM micrographs of collagen after treated by (a) aqueous alkali, macropores with 200 lm in diameters; (b) aqueous acidic, micropores with 30 lm in diameters.

thickness of pore walls is about 5 lm and the size of pores is 200 lm. As comparison, pore structure of collagen prepared under acidic condition is shown in Figure 2(b). The size of pores is about 30 lm. The pore (Fig. 3) of the b-TCP/collagen composite prepared under alkaline condition has pore size of 100 300 lm, similar to that of pure collagen. And the pore structure skeleton is constructed by dense pore walls with about 10 lm in thickness. The surface morphology of the pore shows that b-TCP particles are homogenously embedded into the pore walls. The morphology (Fig. 4) of the cross-section of pore walls shows ordered collagen fibrils confine b-TCP particles at intervals, which act as compact pore walls in composite. It is noteworthy that there is no void between collagen fibrils and b-TCP particles in TEM micrograph, which reveals firmly bonding between them (Fig. 5). For the pure collagen and composites prepared under aqueous alkali conditions, the porosities of the pure collagen and composites are no less than 90% (Table I). It can be deduced that the pores in them are interconnected.


41

Figure 3. SEM micrographs of porous b-TCP/collagen composite. (a) Composite with 100 300 lm in pore size; (b) Pores skeleton constructed by dense pore walls with 10 lm in thickness. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 4. SEM micrographs of pore walls of composite (cross section). (a) Dense pore walls in composite; (b) Ordered collagen fibrils confined b-TCP particles at interval. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Zeta potential of collagen and composite suspension

days cultured. After 4 days, all of the values increased. Furthermore, the values of composites were significant higher than that of b-TCP disk control. Although 6 days cultured, the values of compo-

The Zeta potential of collagen and composite suspensions at different conditions are shown in Table II. Absolutely values of collagen and composite suspensions prepared at pH 2 were larger than that of pH 12, respectively. Furthermore, the difference of values between collagen suspensions was larger than those of composite suspensions.

In vitro Cell morphologies cultured in composites were shown as Figure 6. Cells expressed a high affinity for the composites after 4 days culture in vitro. The attached cells sprawled and flattened on the bottom of pores in composites, anchored on the pores by filopodia (blue arrows). The degree of cell proliferation was quantified using MTT assay after culturing for up to 6 days, as shown in Figure 7. There were no difference of the OD values between composites and controls after 2

Figure 5. TEM micrograph of pore wall (ultrathin section) shows tight contact between collagen fibrils and bTCP particles. White arrow: b-TCP particles; black arrow: collagen; asterisk: epoxy resins. Journal of Biomedical Materials Research Part A

42

ZOU ET AL.

TABLE I Porosities of Collagen and Composite Sample

Porosity (%)

Collagen b-TCP/collagen composite

95.2 6 1.39 93.2 6 1.87

sites were significantly higher than both b-TCP disk and collagen. The functional activity of the cells on the composite was assessed by measuring the ALP expressed by the cells after culturing for up to 14 days, as shown in Figure 8. At day 7, the cells on both composite exhibited significantly higher ALP levels than those on the controls, b-TCP disk and collagen. The similar took place at day 14. It is worth noting that there is no significant difference between two composites prepared, respectively, at alkaline and acidic condition.

DISCUSSION Bone mainly consists of collagen and nanosized carbonated apatite crystallite with a hierarchical organization. In the hierarchical microstructure, the basic building unit is collagen fibrils with adherence of the crystallites.14,15 In view of microstructure, such naturally created microstructure in a biological system is considered to be most favorable for osteoblast proliferation and differentiation, as well as new bone growth. To mimic the microstructure and attain better biological performance, porous b-TCP/collagen composite with a bone-like microstructure is attempted to be prepared in this work. The porous composite is prepared using well-disassembled collagen and nanosized b-TCP particles to guarantee the formation of bone-like microstructure. The well-disassembled collagen is obtained by a treatment in an aqueous alkali solution with pH 12. Nanosized b-TCP particles are added into the well-disassembled collagen suspension to mix, and then lyophilized to become a porous product. The resulting product shows to have still b-TCP phase without other Ca phosphates (Fig. 1), 100 300 lm interconnected pores (Fig. 3), 90% porosity (Table I). These indicate the alkaline treatment has no influence on TABLE II Zeta Potential of Collagen and Composite Suspensions at Different Conditions Sample Pure collagen suspension at pH 2 Pure collagen suspension at pH 12 Composite suspension at pH 2 Composite suspension at pH 12

Journal of Biomedical Materials Research Part A

Zeta Potential (mV) 99.7 228.7 22.4 213.8

6 6 6 6

1.0 2.9 2.2 1.4

Figure 6. SEM photographs of MG63 cells cultured on bTCP/collagen composites after 4 days incubation. (a) Cells sprawled in the bottom of pores; (b, c) Cells anchored in the pores by filopodia, indicated by blue arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

the phase of the nanosized particles and the obtaining of a pore structure required in bone substitutes. When the microstructures of the porous composite are observed, it is found that the pore wall is dense, nanosized b-TCP particles are homogeneously distributed in collagen matrix of pore walls (Fig. 3). The cross section of the pore wall depicts that the wall is constituted of ordered collagen fibrils with adherence of b-TCP particles (Fig. 4). Furthermore, the microstructure of the wall inside shows that the contact


Figure 7. MTT assays of b-TCP/collagen composites and controls. Values are mean 6 SD (n 5 8). yp < 0.05; significant against b-TCP disk control at the corresponding day. { p < 0.05; significant against collagen control at the corresponding day.

between b-TCP particles and collagen fibrils (Fig. 5) is very tight, reflecting a strong bonding. The both demonstrate that the microstructure in the porous composite has a characteristic of the bone building unit. However, the organization of the bone-like units in the present work is different from that in acidic treatment, in which the units are organized as pore walls with micropores rather than a dense wall.19 Since collagen fibrils are made up of ordered collagen molecules that connected by intermolecular covalent bonds and intramolecular hydrogen bonds,13 the collagen fibrils can be disassembled in aqueous acidic or alkaline solutions because the chemical bonds are weakened or broken by solvation or hydrolysis (Fig. 2). The disassembled collagen fibrils have carboxyl groups which are able to be charged and to react with other species. The collagens treated in aqueous solutions with pH 2 and 12 shows to have zeta potential of 99.7 and 228.7 mV, respectively (Table II). This implies that disassembled collagen could have a different status in aqueous solutions with pH 2 and 12, which could be attributed to that protonation for collagen fibrils are easier than de-protonation. When nanosized b-TCP particles are added into the collagen suspension, the particles can interact with collagen fibrils because of the active carbonyl groups on the fibrils.23,24 In both conditions, such interactions decrease the surface charge of fibril-particle composites. As a result, the measured zeta potential increased from 228.7 to 213.8 mV in basic condition. While in acidic condition, a thin layer of calcium phosphate even reprecipitated on the fibril-particle composite, leading to a drastic zeta potential decreased from 99.7 to 22.4 mV. After freeze-drying, no matter that the suspension is treated under acidic or alkaline condition, a charac-

43

teristic of the bone building unit can appear in the resulting porous composites in regards to microstructural considerations. The pore walls in the porous composite have micropores in acidic condition, while the pore walls are dense in alkaline condition. The different organization of bone-like microstructural units could be attributed to the difference in zeta potential for collagen treated in acidic and alkaline solutions (Table II). When a mixture of nanosized bTCP particles and collagen suspension is frozen, water ices up. The crystallization leads to a separation between ice and b-TCP particle adhered collagen fibrils, the ice space become macropores in the resulting composite after lyophilization. During the separation, collagen fibrils undergo getting together, the collagen fibrils could restore the originally undisassembly status like the case in alkaline condition (Fig. 4). If the fibrils are charged a lot, the restoring becomes difficult because of static repelling, as a result, micropores remain in the macropore walls like the case in acidic condition.19 As b-TCP/collagen composites prepared in acidic condition, the b-TCP/collagen composites prepared in alkaline condition also have good biocompatibility. The composites support the cells growth, and the cells migrate inside the pores (Fig. 6), which implies that the composite is suitable for the cell to grow. Cells are inherently sensitive to environmental features at scales from the macrodown to the molecular.25 The bone-like microstructure in porous b-TCP/ collagen composite could make significant contributions to cell adhesion. Affinity of cells to pore walls in 4 days culture demonstrated good interactions between them, which could be attributed to bone-like structure of pore walls.

Figure 8. ALP activity of b-TCP/collagen composites and controls. Values are mean 6 SD (n 5 8). yp < 0.05; significant against the values of b-TCP disk control at the corresponding day. {p < 0.05; significant against the values of collagen control at the corresponding day. Journal of Biomedical Materials Research Part A

44

ZOU ET AL.

Proliferation of cells in vitro can be demonstrated by the MTT assay.22 On the basis of MTT assay, the porous composites with bone-like structure have good effect on cellular proliferation (Fig. 7). Moreover, the expression of ALP has been recognized as an index for detecting the activity of osteoblast functions in the bone forming process.26,27 The results of increased ALP activity with time (Fig. 8) are in accordance with cellular proliferation, which indicate that the normal functions of osteoblast are maintained and enhanced during cellular proliferation in culture. The porous composite with bone-like structure significantly promotes the cellular proliferation, with osteoblast function maintained and enhanced, which can provide potential candidates for bone substitute material. Porous b-TCP/collagen composite with bone-like microstructural units is expected to be biodegradable because of its components, b-TCP and collagen. Also, the composite with insufficient mechanical properties results from its porous structure could be used when only low mechanical strength is required, such as in the repair of jaw or head.28 CONCLUSION A porous b-TCP/collagen composite with bonelike microstructural units could be obtained using well-disassembled collagen and nanosized b-TCP particles in aqueous alkali solution. In the porous composite, the bone-like microstructure units are organized into dense pore walls. In addition, the composites have suitable pore structure with 90% porosity and 100 300 lm pore size, as well as good cytocompatibility that could be attributed to the microstructure suitable for cell proliferation and differentiation. References 1. Finkemeier CG. Current concepts review: Bone-grafting and bone-grafting substitutes. J Bone Joint Surg 2002;84:454–464. 2. Hollinger JO, Brekke J, Gruskin E, Lee D. Role of bone substitutes. Clin Orthop Relat R 1996;324:55–65. 3. El-Ghannam A. Bone reconstruction: From bioceramics to tissue engineering. Expert Rev Med Devices 2005;2:87–101. 4. Lewandrowski KU, Gresser JD, Wise DL, Trantolo DJ. Bioresorbable bone graft substitutes of different osteoconductivities: A histologic evaluation of osteointegration of poly(propylene glycol-co-fumaric acid)-based cement implants in rats. Biomaterials 2000;21:757–774. 5. Landi E, Celotti G, Logroscino G, Tampieri A. Carbonated hydroxyapatite as bone substitute. J Eur Ceram Soc 2003; 23:2931–2937. 6. Galois L, Mainard D, Delagoutte JP. Beta-tricalcium phosphate ceramic as a bone substitute in orthopaedic surgery. Int Orthop 2002;26:109–115.

Journal of Biomedical Materials Research Part A

7. Walsh WR, Chapman-Sheath PJ, Cain S, Debes J, Bruce WJM, Svehla MJ, Gillies RM. A resorbable porous ceramic composite bone graft substitute in a rabbit metaphyseal defect model. J Orthop Res 2003;21:655–661. 8. Deng Y, Lin XS, Zheng Z, Deng JG, Chen JC, Ma H, Chen GQ. Poly(hydroxybutyrate-co-hydroxyhexanoate) promoted production of extracellular matrix of articular cartilage chondrocytes in vitro. Biomaterials 2003;24:4273–4281. 9. Weiss P, Gauthier O, Bouler JM, Grimandi G, Daculsi G. Injectable bone substitute using a hydrophilic polymer. Bone 1999;25:67S–70S. 10. Bruens ML, Pieterman H, de Wijn JR, Vaandrager JM. Porous polymethylmethacrylate as bone substitute in the craniofacial area. J Cranio Surg 2003;14:63–68. 11. Marks SC, Hermey DC. The structure and development of bone. In: Bilezikian JP, Raisz LG, Rodan GA, editors. Principles of Bone Biology. San Diego: Academic; 1996. p 3–14. 12. Anderson JM. Biological responses to materials. Annu Rev Mater Res 2001;31:81–110. 13. Beckman MJ, Shields KJ, Diegelmann RF. Collagen. In: Wnek GE, Bowlin GL, editors. Encyclopedia of Biomaterials and Biomedical Engineering. New York: Marcel Dekker; 2004. p 324–334. 14. Weiner S, Traub W. Bone structure: From angstroms to microns. FASEB J 1992;6:879–885. 15. Robinson RA. An electron microscope study of the crystalline inorganic component of bone and its relationship to the organic matrix. J Bone Joint Surg 1952;34A:389–434. 16. Du C, Cui FZ, Zhang W, Feng QL, Zhu XD, De Groot K. Formation of calcium phosphate/collagen composites through mineralization of collagen matrix. J Biomed Mater Res 2000;50: 518–527. 17. Kikuchi M, Itoh S, Ichinose S, Shinomiya K, Tanaka J. Selforganization mechanism in a bone-like hydroxyapatite/collagen nanocomposite synthesized in vitro and its biological reaction in vivo. Biomaterials 2001;22:1705–1711. 18. Pan Y, Huang JL, Shao CY. Preparation of b-TCP with high thermal stability by solid reaction route. J Mater Sci 2003;38: 1049–1056. 19. Zou C, Weng WJ, Deng XL, Cheng K, Liu XG, Du PY, Shen G, Han GR. Preparation and characterization of porous betatricalcium phosphate collagen composites with an integrated structure. Biomaterials 2005;26:5276–5284. 20. Friess W. Collagen—Biomaterial for drug delivery. Eur J Pharm Biopharm 1998;45:113–136. 21. Li YB, Weng WJ, Cheng K, Du PY, Shen G, Wang JX, Han GR. Preparation of amorphous calcium phosphate in the presence of poly(ethylene glycol). J Mater Sci Lett 2003;22: 1015–1016. 22. Mosmann T. Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays. J Immunol Methods 1983;65:55–63. 23. Mann S. Molecular recognition in biomineralization. Nature 1988;332:119–124. 24. Zhang WZ, Huang ZL, Liao SS, Cui FZ. Nucleation sites of calcium phosphate crystals during collagen mineralization. J Am Ceram Soc 2003;86:1052–1054. 25. Stevens MM, George JH. Exploring and engineering the cell surface interface. Science 2005;310:1135–1138. 26. Peter SJ, Lu LC, Kim DJ, Mikos AG. Marrow stromal osteoblast function on a poly(propylene fumarate)/beta-tricalcium phosphate biodegradable orthopaedic composite. Biomaterials 2000;21:1207–1213. 27. Stringa E, Filanti C, Giunciuglio D, Albini A, Manduca P. Osteoblastic cells from rat long bone. I. Characterization of their differentiation in culture. Bone 1995;16:663–670. 28. Hench LL. Bioceramics. J Am Ceram Soc 1998;81:1705–1728.