Preparation of a porous hydroxyapatite/collagen nanocomposite using glutaraldehyde as a crosslinkage agent

Abstract

Hydroxyapatite (HAp) has been intensively investigated as one of suitable bone substitutes in order to give the biocompatible, bioactive, biodegradable and osteoconductive properties of the natural bone [1, 2]. HAp has been frequently modified by the organics of PMMA, polylactide, chondroitin sulfates, chitosan and collagen (COL) [3–11]. It is known that the human bone is an extracellular matrix mainly composed of HAp nanocrystals and COL fibers. HAp nanocrystals are aligned their c-axes along COL fibers [6]. This HAp embedded collagen nanostructure could have been reproduced using a biomimetic coprecipitation method [6]. In this study we tried to prepare a porous structure of HAp/COL nanocomposite using glutaraldehyde (GA) as a cross-linkage agent. The pure Ca(OH)2 was obtained through the hydration of CaO calcined at 1150 ◦C for 3 h [8]. The slurry of HAp/COL nanocomposites was prepared by the simultaneous titration method using the pH controller as described in detail by Kikuchi et al. [6]. Homogeneous suspension including 0.1994 mol of Ca(OH)2 dispersed in 2 1 of H2O and 59.7 mM of H3PO4 aqueous solution with 5 g of COL was gradually added into a reaction vessel through tube pumps. The weight ratio of HAp/Col was fixed at 80/20. The temperature and pH of reaction solution in vessel was set as 38 ◦C and 8.4, respectively. After the coprecipitation process, the slurry obtained was aged at 38 ◦C for 12 h and pH has gradually lowered to 7.0. Then an aqueous solution of glutaraldehyde (0.2%) was slowly dropped into the slurry solution 38 ◦C; The total amount of GA was regulated to be 30, 90, 300 and 600 molecules per collagen molecule. The samples obtained were hereafter denoted by HCG30, HCG90, HCG300 and HCG600, respectively; the pure collagen and the HAP/collagen composite without cross-linkage were abbreviated as COL and HAP/COL, respectively. Crosslinkage of COL with GA involves the reaction of the free amine groups of lysine or hydroxlysine amino acid residues of the polypeptide chains with the GA aldehyde groups [14]. The HAp/COL slurry was filtered using glass filter and gently washed five times with ion-exchanged water. The precipitates obtained were dried in a freeze dryer (Advantage, Vir Tis, USA) at −30 ◦C under vacuum, or were naturally dried in the air at 25 ◦C. The samples were dipped in ion-exchanged water or in simulated body-fluid in order to check dissolution and water absorption. The precursors used were CaCO3 (alkaline analysis grade, Wako, Japan), H3PO4 (AP grade, Wako, Japan), collagen (MW 300,000, Nitta Gelatin, Japan) and GA (25% aqueous solution, MW 100, AP grade, Wako, Japan). Collagen was extracted from porcine dermis and telopeptide was removed by treating with an enzyme. A length of COL fiber is 300 nm, which is the length of α chain of COL. The collagen was diluted in a phosphoric water solvent by the supplier (Nitta Gelatin) and the typical property was pH 2.23, conc. 10 mg/l. and 20 mM (phosphoric acid). The microstructures of the composites were observed by scanning electron microscopy (SEM, TOPCON, Japan). A chemical reaction between HAP nanocrystals and functional groups of collagen was evaluated using a diffuse reflection FT-IR (Spectrum 2000, Perkin-Elmer, UK). Three-point bending strength and Young’s modulus were measured by a universal testing machine (AGS-H, Shimadzu, Japan) at a cross-head speed of 500 μm/min with a span of 15 mm and the typical samples size was 5 × 3 × 20 mm3. Fig. 1 shows the SEM photographs of the samples prepared. Peculiar pore structures could be observed for the cross-linked samples (HCG30, HCG90, HCG300 and HCG600). Figs 1B and D show the microstructure of naturally dried HCG30, and Figs 1A and E those of a freeze-dried HCG30 sample and a naturally dried HCG300 sample, respectively. One of HCG30 samples (Fig. 1C) was semi-dried at room temperature in air and then completely dried in a freeze dryer. From Fig. 1C we can confirm the strong development of COL bundles. The fractured COL bundles (a white arrow) due to the sectioning process show an evidence for the selfalignment of COL bundles. The COL bundles are well aligned and shows woven networks (a black arrow). From the freeze dried sample (A) we can also observe the randomly distributed pore channels. From the naturally dried samples (B, E) we can observe the columnar open pore channels in matrices and many small pores in the channels. Fig. 1D shows a microstructure for a perpendicular section for Fig. 1C and we can observe many open pores (white arrows). Therefore the open channels probably formed a 3-dimensional network. When glutaraldehyde was added into the composite slurry during the preparation process, the size of the