Biomimetic Mineralization of Woven Bone-Like Nanocomposites

(1) This can be attributed to the organic matrix (primarily a collagen fibrillar .... tests were performed on nanoindentation using software SPSS v16...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/Biomac

Biomimetic Mineralization of Woven Bone-Like Nanocomposites: Role of Collagen Cross-Links Yuping Li,†,‡ Taili T. Thula,† Sangsoo Jee,†,§ Sasha L. Perkins,† Conrado Aparicio,‡ Elliot P. Douglas,† and Laurie B. Gower*,† †

Department of Materials Science and Engineering, University of Florida, Gainesville, Florida 32611-6400, United States Minnesota Dental Research Center for Biomaterials and Biomechanics, School of Dentistry, University of Minnesota, Minneapolis, Minnesota 55455, United States

‡

S Supporting Information *

ABSTRACT: Ideal biomaterials for bone grafts must be biocompatible, osteoconductive, osteoinductive and have appropriate mechanical properties. For this, the development of synthetic bone substitutes mimicking natural bone is desirable, but this requires controllable mineralization of the collagen matrix. In this study, densified collagen films (up to 100 μm thick) were fabricated by a plastic compression technique and cross-linked using carbodiimide. Then, collagen-hydroxyapatite composites were prepared by using a polymer-induced liquid-precursor (PILP) mineralization process. Compared to traditional methods that produce only extrafibrillar hydroxyapatite (HA) clusters on the surface of collagen scaffolds, by using the PILP mineralization process, homogeneous intra- and extrafibrillar minerals were achieved on densified collagen films, leading to a similar nanostructure as bone, and a woven microstructure analogous to woven bone. The role of collagen cross-links on mineralization was examined and it was found that the cross-linked collagen films stimulated the mineralization reaction, which in turn enhanced the mechanical properties (hardness and modulus). The highest value of hardness and elastic modulus was 0.7 ± 0.1 and 9.1 ± 1.4 GPa in the dry state, respectively, which is comparable to that of woven bone. In the wet state, the values were much lower (177 ± 31 and 8 ± 3 MPa) due to inherent microporosity in the films, but still comparable to those of woven bone in the same conditions. Mineralization of collagen films with controllable mineral content and good mechanical properties provide a biomimetic route toward the development of bone substitutes for the next generation of biomaterials. This work also provides insight into understanding the role of collagen fibrils on mineralization.

1. INTRODUCTION Natural bone material is a organic-inorganic nanocomposite, well-known for its remarkable mechanical performance, including high strength and fracture toughness.1 This can be attributed to the organic matrix (primarily a collagen fibrillar network) and the inorganic mineral component (nanocrystals of hydroxyapatite), as well as their unique hierarchical geometrical arrangement.2 At the nanostructural level, bone is composed of mineralized collagen fibrils where plate-like hydroxyapatite (HA) crystals are preferentially oriented with their c-axes parallel to the longitudinal axis of the fibrils.3 Development of synthetic materials that resemble natural bone is a tremendous challenge because it involves a combination of two dissimilar phases, where the organic and inorganic phases have a specific hierarchical organization over various length scales.4,5 The development of synthetic bone substitutes in terms of mineralized fibrils with a nanostructure similar to bone has been attempted.6−9 Conventional methods consisting of nucleation and growth of HA on collagen matrices have been unable to duplicate the specific organic/inorganic nanostructure of bone. Bone mineralization is thought to initiate at the © 2011 American Chemical Society

hole zones of the collagen fibrils with preferentially orientated HA crystals, followed by further deposition at the overlap zones.10 There is evidence that intrafibrillar mineralization is a dominant contributor to the elasticity and hardness of the hard tissues such as bone and dentin.11 Thus, synthesis of mineralized collagen fibrils with intrafibrillar mineral is a key factor for mimicking natural bone. Olszta et al. reported successful intrafibrillar mineralization of collagen, resulting in [001] oriented HA nanocrystals embedded within the fibrils, similar to bone.12 This was accomplished by using poly-Laspartate sodium salt (polyAsp) to stabilize a precursor phase of amorphous calcium phosphate (ACP), which is highly hydrated, facilitating its infiltration into the interstices of the fibrils. This process has been called a polymer-induced liquidprecursor (PILP) process.12−15 Through further optimization of the process, a high mineral content, as much as 60−70 wt % mineral, was achieved, matching that of bone.14 Received: August 1, 2011 Revised: November 14, 2011 Published: December 1, 2011 49

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

(TGA). Furthermore, determination of the elastic modulus and hardness was conducted by a nanoindentation technique.

Bone exhibits different microstructural characteristics during aging and maturation. In fetuses, the human long bone consists entirely of a woven-fibrillar structure containing primary osteons (referred to as woven bone).16 After birth, woven bone is gradually replaced by secondary lamellar bone.17 As the bone structure evolves, the composition and mechanical properties also consequently change.18,19 Recently, selfassembled collagen-apatite matrices with bone-like hierarchy (lamellar organization) were developed through coprecipitating HA with concentrated collagen.20 However, the mineralization degree was only about 6 wt %, much less than fully developed bone. In general, reconstituted collagen sponges have weak mechanical properties due to their high porosity,21 soft nature, or lack of the native fibrillar structure. Even when such reconstituted collagen sponges are well mineralized (our group can achieve 70 wt % mineral), the composites are relatively weak due to the porosity of the underlying scaffold. As described herein, a densified collagen scaffold was prepared using a plastic compression technique,22,23 which leads to randomly distributed collagen fibrils with a microstructure reminiscent of woven bone.24 The densified scaffold was then mineralized with the PILP process, leading to a high mineral content and significantly enhanced mechanical properties. In biological systems, the assembly of type I collagen into fibrils is accompanied by formation of inter- and intramolecular covalent cross-links between α-chains,25 which confer the high tensile and mechanical strength needed for integrity of tissues. The cross-links are based on aldehyde formation and condensation involving specific peptidyl lysine and hydroxylysine residues.26 For reconstituted collagen scaffolds, several physical and chemical cross-linking methods have been reported.27−30 In comparison of all methods, the cross-linking of collagen with a carbodiimide is considered to be efficient and less harmful since the combination of carbodiimide with Nhydroxysuccinimide (NHS) results in the formation of amides between collagen molecules, and all excess residues can be removed.31 Collagen scaffolds cross-linked with carbodiimide can support the growth of fibroblasts, smooth muscle cells and periodontal ligament cells.32−35 Through studies on an in vitro model system, the role of collagen in directing apatite formation has been recognized recently.36 However, it is not known how collagen cross-links affect the mineralization process. During bone maturation, not only does the mineral content and bone macrostructure change, but covalent cross-links within the collagen matrix turn to mature trivalent cross-links consisting of hydroxylysyl-pyridinoline, lysyl-pyridinoline and pyrroles.37 In this report, there are two main goals: (1) to fabricate collagen scaffolds with a fibrillar interwoven and densely packed structure, emulating the microstructure of woven bone, and (2) to evaluate the effect of collagen cross-links on mineralization. Thus, densified collagen films were prepared by a plastic compression (PC) technique,22,23 which leads to randomly distributed collagen fibrils with a microstructure reminiscent of fetal woven bone. The scaffolds were then cross-linked using carbodiimide chemistry to generate various cross-linking densities, followed by biomimetic mineralization via the PILP process. The crystal phase and the morphologies of the mineralized collagen films were characterized by scanning electron microscopy (SEM), atomic force microscopy (AFM), transmission electron microscopy (TEM), and X-ray diffraction (XRD). The mineral contents were determined by thermogravimetric analysis

2. EXPERIMENTAL SECTION 2.1. Materials. PureCol collagen solution (97% bovine dermal type I collagen) was purchased from Inamed Biomaterials (Fremont, CA). Sodium phosphate dibasic and N-hydroxysuccinimide (NHS) were purchased from Fischer Scientific (Pittsburgh, PA). 1-Ethyl-3-[3dimethylaminopropyl] carbodiimide hydrochloride (EDC), poly-Laspartic acid sodium salts, and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). 2.2. Preparation of Collagen Films. Collagen fibrils were prepared by addition of 96 mL of type I collagen (2.9 mg/mL) to 24 mL PBS (10x), followed by adding 16 mL of 0.1 N NaOH to pH 8. The mixture was ultrasonicated in a cooled water bath for 5 min and placed in an incubator at 30 °C for three days to form gels. Densified collagen films were prepared using the method developed by Brown et al. with some modification.23 Briefly, a stainless-steel mesh (mesh size approximately 300 μm) and a layer of nylon mesh (approximately 50 μm mesh size) were placed on an absorbent paper. The collagen gel was then placed on the nylon mesh and covered with another nylon mesh and a glass plate, loaded with a 100 g flat weight for 8 h. Once compressed, samples were rinsed with deionized water to remove excess salts. The cross-linking was conducted in a solution of 50 mM 2-(N-morpholino)ethanesulfonic acid (MES) hydrate (pH 7.0) with increased concentrations of EDC (10, 50, or 100 mM) and NHS at a constant molar ratio of 2:1 overnight. The reaction was quenched by immersing collagen films in a solution (0.1 M Na2HPO4 and 2 M NaCl) for 2 h to hydrolyze any remaining activated carboxyl groups and EDC.31,38 Then, the collagen films were thoroughly rinsed with deionized water and air-dried. 2.3. Mineralization of Collagen Films. Collagen films were mineralized with calcium phosphate (CaP) via the polymer-induced liquid-precursor (PILP) process. In this study, the mineralization solution was prepared by mixing equal volumes of 9 mM CaCl2·2H2O and 4.2 mM K2HPO4 solutions. To maintain the pH of the mineralization solution at 7.4, calcium and potassium phosphate solutions were made in Tris-buffered saline (TBS). Poly-L-aspartic acid (polyAsp) sodium salt (Mw: 10,500 Da) was used as the PILP processdirecting agent at a 50 μg/mL concentration. The polymer was added to 50 mL of calcium solution before mixing an equal volume of the phosphate counterion solution. Collagen films were incubated in the mineralization solution under vacuum conditions for 30 min to remove any air bubbles trapped in the pores of the films. After degassing, the mineralization reaction was kept in a 37 °C oven to emulate physiological conditions. The mineralized samples were removed from the solution after 14 days, thoroughly washed with deionized water, lyophilized, and stored at −20 °C until use. 2.4. Determination of Cross-Linking Density. The crosslinking density of the collagen films was estimated from the free amine groups remaining on the collagen.39 Briefly, collagen samples (approximately 4 mg) were placed in a solution containing 1 mL of 2,4,6-trinitrobenzene sulfonic acid (TNBS, 0.5% w/v, water solvent) and 2 mL of NaHCO3 buffer (4% w/v) and heated for 2 h at 40 °C. Samples were then hydrolyzed by adding 3 mL of HCl (6 N) and heating to 60 °C for 1.5 h. After cooling to room temperature, the solution was diluted to 60 mL. A blank was created in the same manner without collagen. The absorption peak near 345 nm upon reaction of TNBS with ε-amino groups of L-lysine was recorded using a UV−vis spectrophotometer (Perkin-Elmer Lambda 800). Three samples of each cross-linked collagen film were tested. The crosslinking density was calculated from the reduction of absorbance per weight of collagen compared to noncross-linked collagen. 2.5. Contact Angle Measurement. Static contact angle measurements on the air-dried collagen films were performed by contact angle analyzer (DM-CE1, Kyowa Interface Science). The contact angles were recorded every second after the water droplets were contacted with the films to examine the penetration times. Three tests on each film were analyzed and averaged. 50

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

Figure 1. Densification of collagen scaffolds using the plastic compression technique. (a) Photograph of the collagen gel (2.0 mg/mL) placed in a 5 cm Petri dish. (b) Plastic compressed collagen film (approximately 80 mg/mL) in the same size dish. Note − the compressed film was folded. For the mineralization, the densified films were flat. TEM image (c) and SEM image (d) of a collagen film cross-linked at an EDC concentration of 50 mM. The characteristic periodicity of type-I collagen fibrils can be observed in the TEM image. 2.6. Scanning Electron Microscopy (SEM) Analysis. The lyophilized samples were mounted on aluminum stubs with doublesided copper tape and sputter-coated with amorphous carbon. The surface morphologies of samples were analyzed using field-emission scanning electron microscopy (JEOL 6335F and 6500 FEG-SEM) at 10 kV. For elemental analysis of mineralized samples, energy dispersive X-ray spectroscopy (EDS) analysis was performed during SEM examination. 2.7. Atomic Force Microscopy (AFM) Analysis. The surface topography of mineralized collagen films was examined in an Atomic Force Microscope (Veeco Dimension 3100) using a standard tapping mode equipped with a silicon nitride probe at a scan rate of 1 Hz. 2.8. Thermogravimetric and Differential Thermal Analysis (TG/DTA). The mineralization degree of the collagen films was determined by thermogravimetric and differential thermal analysis (TG/DTA; Seiko TG/DTA 320). The temperature was raised from room temperature to 800 °C at a heating rate of 5 °C min−1 under air. The lyophilized films were cut into small pieces and about 10 mg of samples were used for examination. Alumina powder was used as the standard. The mineral content was determined by the remaining at 600 °C after the organic portion of the samples was combusted. 2.9. X-ray Microdiffraction (XRD) Analysis. The crystal structure of mineralized samples was characterized with a microdiffractometric system with a two-dimensional area detector (Bruker AXS) operated at 45 kV and 40 mA. The incident angle was 15° and the detector position was fixed at 30°, which covered the angular range from 15 to 45° in 2θ. The data collection time was 2000 s and the results were analyzed using JADE8 software (Materials Data Inc., JADE, Livermore, CA). A trabecular rabbit bone was used as a control. 2.10. Transmission Electron Microscopy (TEM) Analysis. TEM samples of pure collagen fibrils were prepared by placing the selfassembled collagen gel on copper grids with 200 mesh size, and then removing excess water by placing filter paper at the edge of the grids. The fibrils were negatively stained with 1% phosphotungstic acid solution at pH 7.4 for 15 s and rinsed with distilled water and air-dried. Preparation of mineralized samples for TEM analysis followed the protocols performed on bone and naturally mineralized tendon.40 Briefly, the mineralized collagen films cross-linked at an EDC concentration of 50 mM were crushed into a fine-grained powder in liquid nitrogen, dispersed in ethanol and dropped on a carbon/ Formvar coated copper TEM grid. The pure collagen fibrils were analyzed by TEM (JEOL 200CX) at an accelerating voltage of 80 kV, and the mineralized collagen fibrils was analyzed by FEI Tecnai 12 TEM at 120 kV in bright-field (BF), dark-field (DF), and selected-area electron diffraction (SAED) modes. 2.11. Nanoindentation. Nanoindentation measurements were performed on a nanomechanical indenter system (Nano Indenter XP,

MTS Systems Corporation) with a diamond Berkovich indentation tip. The system was calibrated using a fused quartz standard and its nanoindentation modulus and hardness was calculated as 74.0 ± 3.2 and 10.0 ± 0.5 GPa, respectively. For the dehydrated specimens, the maximum indentation depth was 1000 nm, a 10 s hold time at maximum loading and 50 s at 10% of maximum load during unloading was used to minimize thermal drift and creep effects. The lyophilized collagen films (non-cross-linked and cross-linked at an EDC concentration of 100 mM), and mineralized collagen films were glued on aluminum stubs directly and 30 indents were run for each sample with an indent spacing of 10 μm. Nanoindentation on cortical bone from the bovine tibia was performed as a control. For bone sample preparation, bovine bone was embedded in PMMA, sectioned to reveal an indentation surface, and then polished with polishing paper (SiC, P1200), micropolish alumina suspension (0.1 and 0.05 μm, respectively). A total of 10 indents were performed on the polished bovine bone surface. For the rehydrated specimens, the maximum indentation depth was 2000 nm. The sectioned bovine tibia bone, non-cross-linked collagen film, and mineralized collagen film cross-linked at an EDC concentration of 100 mM were rehydrated overnight. During nanoindentation tests, the specimens were kept in water and 20 indents were run for each sample. The elastic modulus and hardness were obtained from the curves using the Oliver-Pharr method.41 ANOVA tables with Tukey’s or Tamahane’s T2 posthoc multiplecomparison tests were performed on nanoindentation using software SPSS v16.0.1 (IBM, NY) to assess statistically significant differences (p-value ≤ 0.01) between groups. Tamahane’s test was performed when equality of variances between the tested groups could not be assumed.

3. RESULTS 3.1. Fabrication and Mineralization of Densely Packed Cross-Linked Collagen Films. When the plastic compression technique was used, a large amount of fluid was squeezed out and nonfibrous collagen was removed as well, leading to densified collagen films (Figure 1a,b). The cross-linking reaction applied to the collagen films using EDC and NHS leads to the formation of peptide-like cross-links between amino groups and carboxylic acid groups of collagen molecules.31 After cross-linking, the densified collagen films displayed similar characteristics to the collagen matrix in woven bone in terms of exhibiting a native banding pattern in the collagen nanofibrils, and a homogeneous interwoven micro51

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

Figure 2. SEM and TEM micrographs of collagen fibrils cross-linked with 50 mM EDC and mineralized by the PILP process for 14 days. (a) SEM image of the mineralized collagen film showing randomly distributed fibrils; (b) EDS of the mineralized collagen film from (a); (c) high magnification of mineralized collagen films with both intrafibrillar mineral and an extrafibrillar mineral coating on the surface of individual fibrils; (d) BF-TEM image of a few unstained PILP-mineralized collagen fibrils; (e) SAED pattern of the mineralized fibrils obtained from the region circled in (d), showing the crystal planes of hydroxyapatite as broad arcs for the (002) plane and overlapping arcs for the (112), (211), and (300) planes forming a ring shape; (f) DF-TEM image of the mineralized fibrils constructed by selecting the beam from the (002) arc with the objective aperture.

Table 1. Cross-Linking Density and Width of Collagen Fibrils as a Function of EDC Concentrationa width of fibrils (nm) [EDC] (mM)

EDC/NH2collagen (mol/mol)

NHS/EDC (mol/mol)

0 10 50 100

0 50 50 50

0 0.5 0.5 0.5

b

cross-linking density (%) 0 26 ± 6 59 ± 5 82 ± 4

before mineralization 150 145 146 142

± ± ± ±

37 45 28 28

after mineralization 311 337 318 312

± ± ± ±

54 53 57 49

Values of cross-linking density are mean ± standard deviation (n = 3). The width of fibrils before and after mineralization was measured from SEM images. Values are mean ± standard deviation (n > 100). bPercentage of free amine groups that have been reacted.

a

structure with randomly distributed fibrils (Figure 1c and d, respectively). Mineralization of a densified collagen film cross-linked with 50 mM EDC was conducted in a PILP solution for 14 days. Figure 2a,c shows the morphology of the collagen films after mineralization, where randomly distributed collagen fibrils were preserved. The EDS spectrum identified the elements of Ca and P at high levels (Figure 2b), yet there are no HA spherulitic clusters on the surface, suggesting that the mineral phase is primarily embedded within the collagen fibrils (further corroborated by the following TEM and DTA data). The arrangement of HA nanocrystals within the collagen fibrils was examined by TEM using bright field (BF), dark field (DF), and selected area electron diffraction (SAED) modes. The mineral orientation with respect to the collagen fibrils is shown in Figure 2d, e and f. As can be seen, the periodic banding pattern of the fibril was obscured in the mineralized fibrils. The SAED image from the circled fibril in Figure 2d displayed (002), (300), (112), and (211) arcs of hydroxyapatites. This result indicates that the intrafibrillar hydroxyapatite crystals were oriented in the [0 0 1] direction parallel to the long axis of the collagen fibril, matching that of the intrafibrillar mineral in bone, with tilting and rotational disorder of nanocrystals creating arcs in the diffraction spots for most of the planes. When choosing the (002) arcs by the objective

aperture in DF-TEM images (Figure 2f), the HA crystals oriented in the [0 0 1] direction were visualized. 3.2. Cross-Linking Effect on Mineralization of Collagen Films via the PILP Process. The cross-linking density as a function of EDC concentration was calculated as a percentage based on the maximum number of free amino groups available and is shown in Table 1. For calculation, the molecular weight of collagen was assumed to be 300,000 g/mol with a lysine content of 25/1000 residues.42 A range of crosslinking values between 26 and 82% was achieved using various EDC concentrations while under a constant ratio of EDC/free amine group of collagen and NHS/EDC. The cross-linked collagen films became white in the aqueous solution as they were cross-linked. The higher the cross-linking density, the fewer free amino and carboxylate groups that should remain; thus, the collagen films presumably became less hydrophilic. Although it is difficult to measure contact angles on a porous material, the change in the contact angle with time was determined as the water drops penetrated into the material (Figure S1). The water penetration rate on the air-dried collagen films was faster when they were cross-linked. As the water droplet made contact with the films, the contact angle was markedly reduced on the cross-linked films even though they had a higher initial contact angle. This phenomenon seemed to be enhanced with higher cross-linking density. 52

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

Figure 3. SEM images and AFM amplitude images of mineralized dense collagen films with various cross-linking densities: X0 is a mineralized pure collagen film, and X26, X59, and X82 are the mineralized films with cross-linking densities of 26, 59, and 82%, respectively. Scale bar of SEM images is 2 μm. The areas marked with arrows show collagen fibrils with D-periodicity, suggesting these regions were not as well mineralized. Such regions were also imaged in AFM, as seen in the inset images. Scale bar of AFM images is 500 nm.

Figure 4. TG/DTA of the mineralized collagen films. (a) TGA of collagen films mineralized via the PILP process for 14 days. The mineral contents (wt %) of the films were measured as the ash content at 600 °C. The small weight loss beyond this temperature is attributed to CO2 loss from carbonated apatite.14 (b) DTA curves of mineralized collagen films with various cross-linking degrees showing thermal decomposition at a temperature of around 330 °C, which is comparable to native bone.

that the width of fibrils doubled after mineralization, regardless of cross-linking density (Table 1). Such an increase in the width is due to the infiltration of the mineral into the hydrated collagen fibrils during mineralization, where the water becomes displaced by mineral. Swelling of collagen fibrils upon mineralization has been reported in a cryo-TEM study using reaction conditions similar to ours.36 On the non-cross-linked film (Figure 3, X0), the SEM image shows randomly distributed fibrils along with some filmy, web-

Densified collagen films with various cross-linking densities were mineralized with the PILP process using 50 μg/mL of polyAsp for 14 days. Some disintegration of the non-crosslinked collagen films occurred, while the cross-linked films remained intact due to the improved mechanical strength and stability provided by cross-links. The SEM and AFM images in Figure 3 illustrate the changes that occurred when the collagen fibrils became mineralized. As can be seen, the fibrils were uniform in thickness and cylindrical in shape, but it was found 53

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

Figure 5. SEM images and EDS line-scans spectra of the cross-section of densely packed mineralized collagen films. (a) Cross-section of a mineralized film using densified collagen with a cross-linking density of 59%. (b) Magnified SEM image of a region of the mineralized film outlined by the rectangle in (a). (c) EDS line scan of Ca (bottom) and P (top) across the cross-section marked with the line on the mineralized collagen scaffold (X59) shown in (a).

collagen changes the decomposition behavior of collagen (pure collagen (not mineralized) has a high temperature peak at ∼500 °C).14,45,46 In this study, a strong exothermic peak at about 324−331 °C for the mineralized films was observed. This data corresponding to collagen decomposition is comparable to native bone and dentin and other PILP-mineralized samples,14,45−47 suggesting that the majority of hydroxyapatite crystals are embedded within the collagen fibrils. The X26 sample shows a small hump at higher temperature, suggesting that there may be some lesser mineralized regions in this sample. Cross-sectional morphological analysis on the mineralized film with a cross-linking density of 59% demonstrated that uniform mineralization occurred across the sample with a thickness of 100 μm. As seen in Figure 5, the collagen fibrils at the interior were thick and densely packed and the texture was uniform. Elemental analysis of Ca and P by EDS line scan showed that Ca and P were evenly distributed along the cross section. 3.3. Crystallographic Structure and Orientation of Minerals. X-ray diffraction patterns of the mineralized collagen films with different cross-linking densities are shown in Figure 6, including the XRD pattern of a rabbit trabecular bone. The peaks of HA from the (002), (210), (211), (300), (301), (310), and (311) planes were observed and matched well with the diffraction pattern of trabecular bone. The sample with 26% cross-linking had a pattern more similar to the native bone pattern. On other samples, the intensity of the peak at 32.9°, representing the (300) plane of HA, as well as the (210) and (310) planes, was stronger than in the trabecular bone. 3.4. Nanoindentation. The nanoindentation technique was employed to study the variation in mechanical properties of the mineralized collagen films. Figure 7a, b, and c shows a representative nanoindentation load−displacement curve and the dependency of the elastic modulus versus the penetration depth of indenter, as well as the hardness versus displacement on the mineralized collagen film with cross-linking density of 82%, at dry conditions. The indentation load−displacement curve was remarkably similar to that of natural bone (data not

like collagen spanning some of the fibrils. The mineralized fibrils show little to no banding pattern in the SEM image. However, an apparent banding pattern could be found in the nonfibrillar regions (both in the marked regions of the SEM image, and at high magnification in the AFM image), suggesting that those regions were not well mineralized. In the films that were cross-linked, it was found that the amount of nonfibrillar collagen decreased with increasing cross-linking density (X26, X59, X82). Likewise, the amount of collagen with an apparent banding pattern also decreased, as can be seen by the diminished banding texture shown in the corresponding AFM images. The banding pattern of the collagen fibrils became more distorted, which in our system typically indicates the formation of intrafibrillar hydroxyapatite. This result is consistent with the TEM observations, where the periodicity of the collagen fibrils was obscured after mineralization. At the higher cross-link densities, it appears that a thin layer of mineral formed on the surface of fibrils, as is seen by the roughened surface texture in the SEM images. This extrafibrillar mineral is different than that produced in the control reaction, which consists of large spherulitic clusters. With the PILP reaction, the extrafibrillar mineral is in the form of a thin coating.12,14,15 The nanostructure of the native bone also consists of intra- and extrafibrillar hydroxyapatite of this nature, where the extrafibrillar mineral is of a nondescript nature (i.e., not spherulites).43,44 TGA analysis showed that more than 48 wt % mineral content was achieved in the non-cross-linked collagen film and the film with a cross-linking density of 26% (Figure 4a). This result further confirmed that the crystals were primarily deposited within collagen fibrils because that high of a mineral content would be obvious if it was on the surface. Interestingly, when the cross-linking density was increased to 59 and 82%, the mineral content reached up to 53 and 64 wt %, respectively. Figure 4b shows the DTA curves for mineralized collagen films with different cross-linking density. As we demonstrated in our previous studies, the change in thermal stability can be attributed to intrafibrillar mineralization because the intimate structural relationship between hydroxyapatite crystals and the 54

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

elastic modulus and hardness for various films. The non-crosslinked collagen film (without mineral) had very low values of elastic modulus (148 ± 28 MPa) and hardness (31 ± 13 MPa) in the dry state (Table 2). Cross-linking reactions can improve the mechanical properties of collagen.33,35 We found that the films became stiffer after cross-linking as we lifted them up by tweezers. However, the cross-linked collagen film (82%) exhibited a similar modulus (133 ± 26 MPa) and hardness (16 ± 5 MPa) as the non-cross-linked collagen film when measured by nanoindentation in dry state (Table 2). In contrast, the mineralized films show more than 10 times greater values in elastic modulus and hardness. This was particularly pronounced for the cross-linked collagen films, which had moduli between 3 and 9 GPa, and hardness values between 0.2 and 0.7 GPa. Compared to adult bovine lamellae bone (elastic modulus 25.2 GPa, hardness 1.2 GPa), a third of the elastic modulus and half of the hardness of bovine bone were achieved in the mineralized collagen films with a cross-linking density of 82%. On the other hand, the highest value of elastic modulus in the mineralized films is quite similar to the value reported for woven (fetal) bone.48 In addition, a direct relationship between mineral content and mechanical properties (hardness) was found. Such a trend is in agreement with measurements of human fetal bone, as the collagenous tissue becomes more mineralized with fetal age, bone hardness increases.16 It is important to evaluate the mechanical properties of bone in the fully hydrated condition because bone in its mature form is a hydrated hard tissue composed of approximately 65 wt % of mineral, 25 wt % of collagenous and noncollagenous proteins, and 10 wt % of water.49 In this study, the collagen films (pure and cross-linked X82) suffered a severe reduction in moduli

Figure 6. XRD of the mineralized collagen films with different mineralization degrees prepared from collagen scaffolds with crosslinking densities denoted by X0, X26, X59 and X82. XRD of trabecular bone of rabbit was used for comparison. The characteristic peaks of HA, from the (002), (210), (300), (202), (310) and (311) planes, were observed. The X26 sample has the pattern most similar to bone. Stronger peaks from the (300) planes were found for the other samples, indicating orientational effects of mineral.

shown), including a long plastic deformation phase. The maximum value of elastic modulus occurred at penetration depths of more than 50 nm. With regard to the hardness, the maximum hardness was found at a penetration depth from 100 to 200 nm. Figure 7d,e summarizes the maximum values of the

Figure 7. Nanomechanical testing of mineralized collagen films. (a) A representative nanoindentation load−displacement curve on a mineralized collagen film (X82). (b) The elastic modulus of a mineralized collagen film (X82) vs the penetration depth of the indenter. (c) The hardness of a mineralized collagen film (X82) vs displacement. Elastic modulus (d) and hardness (e) values of the mineralized collagen films with cross-linking densities of X0, X26, X59, and X82. A large increase in elastic modulus and hardness was found for the mineralized collagen films. Significant differences in the hardness of mineralized collagen films with different cross-linking density were found. (f) A comparison of elastic modulus on films (X82) before and after mineralization in dry and rehydration condition. The values indicated for the elastic modulus of woven bone were obtained from literature reports.48,50 The pure collagen film (X82) has a very low elastic modulus. A severe reduction of modulus of the mineralized film (X82) occurred upon rehydration, which is also shown in the literature for woven bone. Values with the same symbol had no statistically significant differences. 55

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

Table 2. Moduli and Hardness Values Obtained from Nanoindentation Measurements films

conditions

effective tests/tests

moduli (GPa)

hardness (GPa)

non-cross-linked collagen films

dry rehydrated dry rehydrated dry rehydrated dry rehydrated

22/30 4/15 18/20 4/20 10/30 10/20 10/10 15/20

0.145 ± 0.028 0.019 ± 0.011 0.133 ± 0.026 0.031 ± 0.029 9.11 ± 1.35 0.177 ± 0.031 25.2 ± 1.7 14.5 ± 0.7

0.031 ± 0.013

cross-linked film (X82) mineralized collagen film (X82) bovine cortical bone a

a

0.016 ± 0.005 a

0.73 ± 0.14 0.008 ± 0.003 1.21 ± 0.08 0.35 ± 0.03

Values were below measurement sensitivity.

acid residues of aspartic and glutamic acid, while nucleophilic groups, such as free amine groups of lysine and hydroxylysine residues, attack O-acylisourea to form cross-links. As the crosslinking density increases, more free amine groups and carboxylic acid groups are reacted.56 These cross-links forming within the fibrils (intermolecular cross-links) or between the nearby fibrils (interfibrillar cross-links) can improve mechanical strength and stability. Traditional methods of cross-linking collagen scaffolds involving acidic solution, organic solvent, or physical irradiation compromised the structure of collagen due to denaturation.57,58 The cross-linking process we chose minimized the side effects of denaturation and preserved the periodic banding of the fibrils. Providing that the periodic banding in collagen fibrils plays an important role in intrafibrillar mineralization (the hole zones that enable infiltration of the mineral precursor), the process used here to fabricate densely packed collagen films with variable crosslinking densities provided suitable matrices for mineralization. Collagen fibrils serve as the primary template for HA nanocrystal deposition and orientation. The mineralization rate depends on the capability of infiltration of amorphous calcium phosphate (ACP) into the fibrils. In vivo, noncollagenous proteins (NCPs) composed of a high density of acidic amino acids, including aspartic acid, glutamic acid, or phosphorylated serine, are thought to be involved with mineral formation due to their high affinity to Ca2+ and collagen.59 Recently, the role of collagen on early mineral formation within a collagen fibril, using reaction conditions similar to ours, was reported.36 That study demonstrated that ACP-polyAsp formed negatively charged complexes that preferentially infiltrated in the gap region of the fibrils. It was suggested that interactions between negatively charged pAsp-ACP complexes and positive domains in the collagen fibrils are the driving force for mineralization.36 In our study, cross-linking of the collagen films promoted the mineralization, presumably resulting from some type of facilitation of the infiltration of ACP-polyAsp particles/ droplets. It should be noted that the cross-linking reaction was conducted on insoluble collagen films. There was little change in their morphology after cross-linking, indicating the preservation of available interstitial space within the fibrils. Biomimetic mineralization was enhanced by increasing the cross-linking density, which cannot be explained by charge− charge interactions between pAsp-ACP clusters and collagen fibrils because the cross-linking reaction should have consumed charged amino and carboxylate groups. Perhaps the consumption of the charge groups led to more extrafibrillar mineral, but this appeared to be minor as judging from the morphology of the fibrils. Plus, the low temperature peak in the DTA data is consistent with a high degree of intrafibrillar mineral, so even if there is a small amount of extrafibrillar

and an undetectable nanohardness after rehydration. When the mineralized collagen film (X82) was rehydrated, it also showed much lower modulus and hardness than in dry state, but significantly higher mechanical properties were assessed in comparison with the nonmineralized film (Table 2). Notably, the elastic modulus of the mineralized collagen film (X82) in wet condition was 177 ± 031 MPa, which matches well with the one reported for wet woven bone (rat fracture callus), where a median of 132 MPa elastic modulus was measured by nanoindentation, with values heterogeneously ranging from 26.92 to 1010.00 MPa.50 This significant reduction in modulus and hardness in the hydrated state of both the mineralized collagen films (X82) and natural woven bone is probably triggered by the high quantities of incorporated water (49% for our rehydrated mineralized collagen films (X82)). That is in contrast with the moderate reduction in modulus and hardness of bovine cortical bone, which still has a compact structure after rehydration, that was also studied here (Table 2).

4. DISCUSSION Biomimetic bone substitutes that are able to emulate the mechanical properties of bone are anticipated to involve in the development of inorganic−organic composites with bone-like nanostructure. Traditional mineralization approaches, such as the alternate soaking of collagen scaffolds in simulated body fluid (SBF) solutions only yield composites with extrafibrillar clusters of HA on the scaffold,51−53 which does not mimic the natural bone structure. In the current study, we fabricated collagen films with randomly distributed collagen fibrils using a plastic compression process and mineralized them via PILP process. The collagen matrix developed by plastic compression exhibits a woven nanofibrillar structure with periodic banding pattern, as well as a densified microstructure. The biomimetic mineralization process yielded intra- and extrafibrillar mineralization of the collagen matrices. Although the mechanical properties still fall short of those found in secondary bone, reasonable hardness, and stiffness values were obtained, similar to those found in woven bone. A cross-linking chemical reaction was used to improve the mechanical stability of the collagen films, which otherwise partially disintegrate during the mineralization reaction. Crosslinking can improve a collagen-based skin substitute’s strength and stability.33 It is unclear how much of an impact the crosslinking has on bone’s mechanical properties. Some studies found that bone’s strength and stiffness is independent of the number of cross-links,54 while others found the compromise of bone’s strength and intrinsic toughness during aging is correlated with the inter- and intrafibrillar cross-links.55 Cross-linking on reconstituted collagen molecules is a process involving the water-soluble carbodiimide to activate carboxylic 56

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

mineralized structure with composition, structure, and mechanical properties very close to the ones in the natural primary tissue. Although the properties of the composites in the hydrated state could not be used for full load-bearing applications, the mineralized collagen films (X82) have enough mechanical integrity to retain growth factors and to be used as placeholders for space filling applications that might induce a fast and appropriate remodeling of the structure in the site of implantation, as the human body does with woven bone, once colonized by cells. Due to the intricate hierarchical structure of bone, ranging from the molecular to the macroscopic scale, the outstanding mechanical properties of bone are difficult to understand. Our mineralized films still have much lower elastic modulus and hardness than secondary (lamellar) bone, particularly when in the hydrated condition. Unlike secondary bone where mineralized collagen fibrils are well aligned and always present in bundles or arrays, the fibrils in our samples and in woven bone are more randomly distributed and loosely packed, leaving in microporosity that degrade the mechanical properties. Future studies targeting the hierarchical structure of the matrix in order to emulate that of secondary bone are anticipated to provide further improvements to the mechanical properties.

mineral, the reduction in charge groups did not seem to prohibit the initial intrafibrillar mineralization. Therefore, we propose that the cross-linking made the collagen films more rigid by limiting the motion of collagen molecules within the collagen fibrils, perhaps holding open the intermolecular spaces. As the polyAsp-ACP complexes infiltrate the fibrils, they would encounter less perturbation from nearby collagen molecules, and this might facilitate mineralization. Indeed, the contact angle measurements showed that the rate of entry of water into the collagen scaffolds was enhanced with the higher crosslinking densities. Thus, it might be that the rate of infiltration of the PILP droplets could also be enhanced. One might have anticipated that cross-linking would hinder the infiltration of the mineral precursor. In our prior paper,46 where we remineralized a demineralized bone sample, the mineralization was quite heterogeneous and was markedly enhanced in the osteonal regions, as opposed to the interosteonal regions. We had hypothesized that the more mature regions (that surround the osteons) might be more difficult to mineralize due to cross-linking of the collagen, which is known to increase with the maturity of the bone. Therefore, we were expecting the cross-linking in this study to hinder the mineralization, which seemed not to be the case here. However, the chemical cross-links in bone collagen are substantially different than the synthetic cross-links created by the EDC reaction used here, so it is difficult to draw any conclusions regarding the cross-linking-mineralization capability in bone. Nevertheless, we were pleased to find that the synthetic cross-linking did not prohibit the mineralization reaction and could thus be used to prepare composites for biomedical applications. Intrafibrillar mineralization is a key ingredient to ensuring that collagen fibrils have high modulus and hardness, as is the case for bone.11 Our nanomorphological analysis revealed the infiltration and formation of HA crystals associated with the fibrils. The [001] direction of the HA crystals is aligned parallel along the longitudinal axis of the collagen fibril, similar to bone. The XRD pattern showed peaks that were enlarged for the (300), (210), and (310) planes, relative to bone. This is likely due to the preferential orientation of the collagen fibrils created during the compression process. One might expect the fibrils to be aligned more parallel to the substrate, thus leading to enhancement of those planes that are parallel to the c-axis of the crystals (the (hk0) planes). In fact, the X0, X57, and X82 mineralized films were flat, while the X26 film was somewhat crumpled after mineralization. In regions of low mineral content, the periodic banding of the fibrils was maintained. At the higher degree of mineralization, the banding pattern was lost and a small amount of surface mineral was observed, indicating that extrafibrillar mineralization had also occurred to some degree. These results are consistent with the mechanism of bone formation, in that the apatite crystals are formed initially within the fibrils, followed by extrafibrillar mineralization of the interstitial positions between the fibrils.60 Our approach of developing densified collagen films with intra- and extrafibrillar mineral not only led to composites which roughly resembled the microstructure of woven bone, but resulted in comparable mechanical properties of elastic modulus and hardness in both the dry and wet state. Biomimetic incorporation of HA minerals dramatically increased the mechanical properties of the collagen films. All these findings strongly suggest that the biomimetic mineralization of the collagen structure with the PILP process delivers a

5. CONCLUSIONS Densified reconstituted collagen films were successfully prepared by a plastic compression technique, and then mineralized via the PILP process, producing collagenhydroxyapatite nanocomposites with a nanostructure and high mineral content similar to woven bone. The plastic compression technique combined with cross-linking resulted in the production of collagen films with high packing and mechanical integrity. Upon mineralization, the crystals were uniformly distributed throughout the collagen fibrils, where both intra- and extrafibrillar mineralization was accomplished. The HA nanocrystals embedded within the collagen fibrils were preferentially oriented parallel to long axis of the fibril. Surprisingly, the collagen cross-linking was found to enhance the mineralization process. We believe that cross-links in the collagen films promote mineralization through propping open the intermolecular spaces and minimizing the exclusive volume of collagen fibrils, which in turn facilitates the infiltration of polyAsp-ACP complexes. The intra/extramineralization and high mineral content of the densified collagen films enhanced their mechanical properties with respect to elastic moduli and hardness. Current efforts are directed at building hierarchically structured collagen matrices for mimicking secondary bone structure and properties.

■

ASSOCIATED CONTENT

* Supporting Information Contact angle measurements on the collagen films with various cross-linking densities were measured at a series of time points from the initial placement of the drop of water. A plot of contact angle versus time is provided as Figure S1. This material is available free of charge via the Internet at http:// pubs.acs.org. S

57

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

■

Article

(24) Su, X.; Sun, K.; Cui, F. Z.; Landis, W. J. Bone 2003, 32 (2), 150−162. (25) Akeson, W. H.; Amiel, D.; Mechanic, G. L.; Woo, S. L. Y.; Harwood, F. L.; Hamer, M. L. Connect Tissue Res. 1977, 5 (1), 15−19. (26) Yamauchi, M. G. M. Crosslinking of Collagen; CRC Press: Boca Raton, FL, 1988. (27) Naimark, W. A.; Pereira, C. A.; Tsang, K.; Lee, J. M. J. Mater. Sci.: Mater. Med. 1995, 6 (4), 235−241. (28) Narayani, R.; Rao, K. P. Int. J. Pharm. 1996, 142 (1), 25−32. (29) Welz, M. M.; Ofner, C. M. J. Pharm. Sci. 1992, 81 (1), 85−90. (30) Tu, R.; Shen, S. H.; Lin, D.; Hata, C.; Thyagarajan, K.; Noishiki, Y.; Quijano, R. C. J. Biomed. Mater. Res. 1994, 28 (6), 677−684. (31) Kuijpers, A. J.; Engbers, G. H. M.; Krijgsveld, J.; Zaat, S. A. J.; Dankert, J.; Feijen, J. J. Biomater. Sci., Polym. Ed. 2000, 11 (3), 225− 243. (32) Suh, H.; Park, S. N.; Park, J. C.; Kim, H. O.; Song, M. J. Biomaterials 2002, 23 (4), 1205−1212. (33) Powell, H. M.; Boyce, S. T. Biomaterials 2006, 27 (34), 5821− 5827. (34) Feijen, J.; Buijtenhuijs, P.; Buttafoco, L.; Poot, A. A.; Daamen, W. F.; van Kuppevelt, T. H.; Dijkstra, P. J.; de Vos, R. A. I.; Sterk, L. M. T.; Geelkerken, B. R. H.; Vermes, I. Biotechnol. Appl. Bioc. 2004, 39, 141−149. (35) Yunoki, S.; Nagai, N.; Suzuki, T.; Munekata, M. J. Biosci. Bioeng. 2004, 98 (1), 40−47. (36) Nudelman, F.; Pieterse, K.; George, A.; Bomans, P. H. H.; Friedrich, H.; Brylka, L. J.; Hilbers, P. A. J.; de With, G.; Sommerdijk, N. A. J. M. Nat. Mater. 2010, 9 (12), 1004−1009. (37) Knott, L.; Bailey, A. J. Bone 1998, 22 (3), 181−187. (38) Lee, J. M.; Edwards, H. H. L.; Pereira, C. A.; Samii, S. I. J. Mater. Sci.: Mater. Med. 1996, 7 (9), 531−541. (39) Antoni, G.; Presentini, R.; Neri, P. Anal. Biochem. 1983, 129 (1), 60−63. (40) Weiner, S.; Traub, W. Organization of Crystals in Bone; SpringerVerlag: NY, 1991; Vol. Biomineralization. (41) Oliver, W. C.; Pharr, G. M. J. Mater. Res. 1992, 7 (6), 1564− 1583. (42) Ikoma, T.; Kobayashi, H.; Tanaka, J.; Walsh, D.; Mann, S. Int. J. Biol. Macromol. 2003, 32 (3−5), 199−204. (43) Bagambisa, F. B.; Joos, U.; Schilli, W. Cells Mater. 1993, 3 (1), 93−102. (44) Ziv, V.; Sabanay, I.; Arad, T.; Traub, W.; Weiner, S. Microsc. Res. Tech. 1996, 33 (2), 203−213. (45) Thula, T. T.; Svedlund, F.; Rodriguez, D. E.; Podschun, J.; Pendi, L.; Gower, L. B. Polymers 2011, 3, 10−35. (46) Gower, L. B.; Thula, T. T.; Rodriguez, D. E.; Lee, M. H.; Pendi, L.; Podschun, J. Acta Biomater. 2011, 7 (8), 3158−3169. (47) Bucio, L.; Lozano, L. F.; Pena-Rico, M. A.; Heredia, A.; OcotlanFlores, J.; Gomez-Cortes, A.; Velazquez, R.; Belio, I. A. J. Mater. Sci. 2003, 38 (23), 4777−4782. (48) Mulder, L.; Koolstra, J. H.; den Toonder, J. M. J.; van Eijden, T. M. G. J. J. Biomed. Mater. Res. A 2008, 84A (2), 508−515. (49) Blitz, R. M.; Pellegrino, E. D. J. Bone Joint Surg. 1969, 51, 456− 466. (50) Leong, P. L.; Morgan, E. F. Acta Biomater. 2008, 4 (5), 1569− 1575. (51) Jalota, S.; Bhaduri, S. B.; Tas, A. C. Mater. Sci. Eng. C 2008, 28 (1), 129−140. (52) Li, X. K.; Chang, J. J. Biomed. Mater. Res., Part A 2008, 85A (2), 293−300. (53) Marelli, B.; Ghezzi, C. E.; Barralet, J. E.; Boccaccini, A. R.; Nazhat, S. N. Biomacromolecules 2010, 11 (6), 1470−1479. (54) Willems, N. M. B. K.; Mulder, L.; Bank, R. A.; Grunheid, T.; den Toonder, J. M. J.; Zentner, A.; Langenbach, G. E. J. J. Biomech. 2011, 44 (6), 1132−1136. (55) Ritchie, R. O.; Zimmermann, E. A.; Schaible, E.; Bale, H.; Barth, H. D.; Tang, S. Y.; Reichert, P.; Busse, B.; Alliston, T.; Ager, J. W. Proc. Natl. Acad. Sci. U.S.A. 2011, 108 (35), 14416−14421.

AUTHOR INFORMATION

Corresponding Author *Tel.: +1-352-846-3336. Fax: +1-352-846-3355. E-mail: [email protected]. Present Address § Samsung Electronics, Samsung Advanced Institute of Technology, Nongseo-dong, Giheung-gu, Yongin-si Gyeonggi-do, 446−712, South Korea.

■

ACKNOWLEDGMENTS This work is supported by the National Science Foundation Grant BES-0404000. The authors thank the Major Analytical Instrumentation Center, Department of Materials Science and Engineering, University of Florida, for the use of TEM, SEM, and AFM. Parts of this work were carried out in the University of Minnesota I.T. Characterization Facility, which receives partial support from NSF through the NNIN program.

■

REFERENCES

(1) Currey, J. D. J. Bone Miner. Res. 2003, 18 (4), 591−598. (2) Fratzl, P.; Gupta, H. S.; Paschalis, E. P.; Roschger, P. J. Mater. Chem. 2004, 14 (14), 2115−2123. (3) Traub, W.; Arad, T.; Weiner, S. Proc. Natl. Acad. Sci. U.S.A. 1989, 86 (24), 9822−9826. (4) Rho, J. Y.; Kuhn-Spearing, L.; Zioupos, P. Med. Eng. Phys. 1998, 20 (2), 92−102. (5) Weiner, S.; Traub, W. FASEB J. 1992, 6 (3), 879−885. (6) Bradt, J. H.; Mertig, M.; Teresiak, A.; Pompe, W. Chem. Mater. 1999, 11 (10), 2694−2701. (7) Itoh, S.; Kikuchi, M.; Takakuda, K.; Koyama, Y.; Matsumoto, H. N.; Ichinose, S.; Tanaka, J.; Kawauchi, T.; Shinomiya, K. J. Biomed. Mater. Res. 2001, 54 (3), 445−453. (8) Tampieri, A.; Celotti, G.; Landi, E.; Sandri, M.; Roveri, N.; Falini, G. J. Biomed. Mater. Res., Part A 2003, 67 (2), 618−625. (9) Zare, R. N.; Maas, M.; Guo, P.; Keeney, M.; Yang, F.; Hsu, T. M.; Fuller, G. G.; Martin, C. R. Nano Lett. 2011, 11 (3), 1383−1388. (10) Arsenault, A. L. Calcif. Tissue Int. 1988, 43 (4), 202−212. (11) Kinney, J. H.; Habelitz, S.; Marshall, S. J.; Marshall, G. W. J. Dent. Res. 2003, 82 (12), 957−961. (12) Olszta, M. J.; Cheng, X. G.; Jee, S. S.; Kumar, R.; Kim, Y. Y.; Kaufman, M. J.; Douglas, E. P.; Gower, L. B. Mater. Sci. Eng. R 2007, 58 (3−5), 77−116. (13) Olszta, M. J.; Douglas, E. P.; Gower, L. B. Calcif. Tissue Int. 2003, 72 (5), 583−591. (14) Jee, S. S.; Thula, T. T.; Gower, L. B. Acta Biomater. 2010, 6 (9), 3676−3686. (15) Jee, S. S.; Culver, L.; Li, Y. P.; Douglas, E. P.; Gower, L. B. J. Cryst. Growth 2010, 312 (8), 1249−1256. (16) Su, X. W.; Feng, Q. L.; Cui, F. Z.; Zhu, X. D. Connect Tissue Res. 1997, 36 (3), 271−286. (17) Weiner, S.; Wagner, H. D. Annu. Rev. Mater. Sci. 1998, 28, 271− 298. (18) Mulder, L.; van Ruijven, L. J.; Koolstra, J. H.; van Eijden, T. M. G. J. J. Biomech. 2007, 40 (7), 1575−1582. (19) Mulder, L.; Koolstra, J. H.; Weijs, W. A.; Van Eijden, T. M. G. J. Anat. Rec., Part A 2005, 285 (1), 659−667. (20) Nassif, N.; Gobeaux, F.; Seto, J.; Belamie, E.; Davidson, P.; Panine, P.; Mosser, G.; Fratzl, P.; Guille, M. M. G. Chem. Mater. 2010, 22 (11), 3307−3309. (21) Lee, C. H.; Singla, A.; Lee, Y. Int. J. Pharm. 2001, 221 (1−2), 1− 22. (22) Abou Neel, E. A.; Cheema, U.; Knowles, J. C.; Brown, R. A.; Nazhat, S. N. Soft Matter 2006, 2 (11), 986−992. (23) Brown, R. A.; Wiseman, M.; Chuo, C. B.; Cheema, U.; Nazhat, S. N. Adv. Funct. Mater. 2005, 15 (11), 1762−1770. 58

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59

Biomacromolecules

Article

(56) Lloyd, D. R.; Burns, C. M. J. Polym. Sci., Polym. Chem. 1979, 17 (11), 3473−3483. (57) Powell, H. M.; Boyce, S. T. Biomaterials 2007, 28 (6), 1084− 1092. (58) Yamauchi, K.; Goda, T.; Takeuchi, N.; Einaga, H.; Tanabe, T. Biomaterials 2004, 25 (24), 5481−5489. (59) Glimcher, M. J. Anat. Rec. 1989, 224 (2), 139−153. (60) Landis, W. J.; Song, M. J.; Leith, A.; Mcewen, L.; Mcewen, B. F. J. Struct. Biol. 1993, 110 (1), 39−54.

59

dx.doi.org/10.1021/bm201070g | Biomacromolecules 2012, 13, 49−59