Baixe Self-assembly of mineralized collagen composites e outras Notas de estudo em PDF para Engenharia de Produção, somente na Docsity!
Self-assembly of mineralized collagen composites
Fu-Zhai Cui *, Yan Li, Jun Ge
Advanced Materials Laboratory, Department of Materials Science & Engineering, Tsinghua University, Beijing 100084, China Accepted 6 April 2007 Available online 16 May 2007
Abstract
This paper presents a review of the current understanding of the structure, self-assembly mechanisms, and properties of mineralized collagen fibril composites in connective tissues, such as in lamellar bones, woven bones, zebrafish skeletal bone, and ivory. Recent work involving biomimetic synthesis of new materials with the structure of mineralized collagen is described. The focus in the paper is mainly on materials containing type I collagen, with mineralization by Ca–P crystals although some other systems are also described. Investigation and simulation of naturally occurring fibril structures can offer some new ideas in the design and fabrication of new functional materials, for applications such as bone grafts or for use as scaffolds in tissue engineering and biomimetic engineering materials. The development of bone grafts based on the mineralization of self- assembled collagen fibrils in vivo and in vitro is an active area of research. This kind of bone graft composite has already shown great promise and success in clinical applications, on account of its compositional and structural similarity to autologous bone. It is suggested that future work in this should focus on both basic theoretical aspects as well as the development of applications. In particular issues including control of morphology, incorporation of foreign ions, interaction with biomolecules, and the assembly of organic and inorganic phases are all still not well understood. In the area of applications, the design of composite materials with a hierarchical structure closer to that of natural hard tissues, and the synthesis of bone grafts and tooth regenerative materials, as well as biomimetic functional materials, are areas currently being examined by many research groups.
2007 Elsevier B.V. All rights reserved.
Keywords: Self-assembly; Biomaterials; Collagen; Calcium phosphate
Contents
- Introduction......................................................................... 2
- Hierarchical assembly of mineralized collagen in natural tissue...................................... 3 2.1. Collagen and model of self-assembled fibrils.............................................. 3 2.2. Organization of mineralized collagen fibrils............................................... 4 2.2.1. Organization in lamellar bones.................................................. 4 2.2.2. Organization in woven bones................................................... 6 2.2.3. Organization in zebrafish skeletal bone............................................ 6 2.2.4. Organization in ivory........................................................ 9 2.2.5. Size and shape of crystals in the mineralized collagen fibrils............................. 9
www.elsevier.com/locate/mser
Materials Science and Engineering R 57 (2007) 1–
- Corresponding author. Tel.: +86 10 6277 2850; fax: +86 10 6277 2850. E-mail address: [email protected] (F.-Z. Cui).
0927-796X/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.mser.2007.04.
- Biomimetic fabrication with self-assembled collagen mineralization................................. 10 3.1. Nucleation sites for collagen fibrils and the initial stage of collagen mineralization.................. 10 3.2. Assembly of nano-fibrils of mineralized collagen.......................................... 12 3.3. Assembly and mineralization of peptide-amphiphilic nanofibers................................ 14
- Applications of mineralized collagen composites in bone regeneration................................ 16 4.1. Synthesis and properties of nano-HA/collagen composites.................................... 16 4.1.1. Direct blended nano-HA/collagen composites...................................... 16 4.1.2. SBF immersion technology................................................... 17 4.1.3. Co-precipitation self-assembly method........................................... 18 4.2. Applications of nano-HA/collagen composites............................................ 19 4.2.1. Gels and powders as a filler for orthopaedic surgery.................................. 19 4.2.2. Surface coatings for implant materials........................................... 19 4.2.3. Three-dimensional scaffolds for bone tissue engineering............................... 20 4.3. Related improvements of mineralized collagen composites................................... 20 4.3.1. Optimization of controllable ionic substitution in nano-HA crystals....................... 20 4.3.2. Improved mechanical properties for implantation at load-wearing sites..................... 21 4.3.3. Development of hierarchical biomimetic technology by combination with other advanced nano-micro synthesis methods................................................. 22 4.4. Animal models and clinical applications................................................ 22 4.5. Other applications of self-assembled collagen fibrils........................................ 23 4.5.1. Combination with cartilage regeneration.......................................... 23 4.5.2. Combination with vascular tissue regeneration...................................... 23 4.5.3. Combination with neural regeneration............................................ 23 Acknowledgements................................................................... 24 References......................................................................... 24
- Introduction
In this paper we review the current understanding of structure, formation and properties of mineralized collagen fibril composites in connective tissues, as well as recent work involving biomimetic synthesis of new materials with the structure of mineralized collagen. The focus in the paper is mainly on type I collagen, of which at least twenty collagens have so far been discovered. By mineralized collagen we refer mostly to calcium phosphate based crystals, which in bone are found to consist primarily of calcium and phosphate ions, with traces of magnesium, carbonate, hydroxyl, chloride, fluoride, and citrate ions [1]. Mineralized collagen is one kind of material that can be produced by self-assembly at ambient temperatures. In this paper we have used the concept of self-assembling defined by Whitesides and Grzybowski [2], i.e., self-assembling is the autonomous organization of components into patterns or structures without human intervention. It is considered that self-assembling processes are common throughout nature and technology. Connective tissues are among the most advanced structural composite materials known to be made of macromolecular building blocks [3–5]. A wide range of tissues, each possessing very different properties are successfully synthesized in natural environments with only the same basic macromolecular design [6]. Nevertheless, these tissues in many instances show some common features—they are assembled in numerous assembly ways that allow control of the formation of varying hierarchical structures, from the nanometer scale to the macroscale [7]. The concept of hierarchical assembly has been recognized and emphasized by more and more scientists over the last decades, as exemplified by the investigations of numerous biocomposite systems. The hierarchical levels of organization with highly specific interconnectivity and with unique architectures are designed to give the required spectrum of properties for each oriented composite system. Based on these lessons in biology, the laws for the formation of complex composite systems for functional macromolecular assemblies have been probed [8,9]. In addition to gaining knowledge of the fundamental mechanisms for assembly of such materials, the ability to build architectures as a direct consequences of the precision in assembly would certainly open the gate to some new areas of materials science. Examples could be the design and construction of inorganic materials with specified atomic structure, size, shape, crystal orientation, and number of defects and the integration of these architectures into
of other amino acids differentiates between the different collagen types. It is now accepted that amino acids play a major role in determining the three-dimensional assembly conformations. The representative hallmark of collagen molecules is the multiple repetition of Gly-X-Y sequences. This feature is dictated by the unique triple helical conformation, built of three polypeptide chains, and is a widespread structural element occurring in collagens. The basic three-dimensional structure of the collagen triple-helix was first inferred from fiber diffraction studies on collagen in tendon [30,31]. Each of the three chains in the molecule forms a left-handed polyproline-II-type helix, which has exactly three residues per turn. The chains are arranged in parallel, staggered by one residue relative to each other, and supercoiled along a common axis to form a right-handed triple helix of 290 nm in length. In contrast to right- handed a-helices, left-handed polyproline-II-helices occur relatively rarely as structural elements in proteins, with the striking exception of collagens. The fiber diffraction performed on tendon fibers has revealed a clear right-handed supercoiling of the three individual chains to form a triple helix, resulting in an increase to 3.33 residues per turn and to a reduction of the axial repeat distance to 0.286 nm per residue. A functionally important structural feature of the collagen triple helix is the orientation of the side chains. Side chains of residues in X- and Y-positions point out of the helix, and are freely accessible for binding interactions, which play an important role in fibril formation through intermolecular interactions between oppositely charged residues and through hydrophobic interactions between residues of different molecules [32]. The interactions between the collagen molecules result in the characteristic quarter repeat of 67 nm and in a complex cross-striation banding, which results in agglomeration of the collagen triple helices into microfibrils, forming the supramolecular structures [33–35]. The most widely accepted model for packing of collagen molecules is that five triple helices align longitudinally with an overlap of approximately a quarter of the molecular length to form a microfibril [34]. Considering that the diameter of each collagen molecule is 1.5 nm [36], the diameter of the collagen fibrils in the five-stranded packing model should be approximate 3.6 nm. This has been verified by transmission electron microscope (TEM) observations of self- assembled collagen fibrils. This so-called quarter stagger, combined with the gap between successive macromolecules, is responsible for the typical 67 nm periodical cross-striation patterns as observed by TEM, atomic force microscope (AFM), and X-ray diffraction (XRD) investigations [21,34,37]. The microfibrils are then assembled into collage fibrils that may vary in thickness from 35 to 500 nm. These are further combined, oriented and laid up to form ordered structures with particular morphologies for tissues. The overwhelming consideration in the arrangement of collagen fibrils to form connective tissues is the resulting tissue function. This will be illustrated and discussed in the following sections. Recently, the formation of synthetic collagen triple helices as long or even longer than natural collagen has been reported [38], although applications for these materials are yet to be developed.
2.2. Organization of mineralized collagen fibrils
2.2.1. Organization in lamellar bones Mineralized collagen fibrils are the basis for various connective tissues such as bone and cartilage. The structure and organization of mineralized collagen in such mineralized tissues are both fundamentally important for many biochemical, physicochemical, and biomechanical events that determine the normal function of these tissues. Bone refers to a family of materials that are constructed by mineralized collagen fibrils with complex hierarchically assembled structures [39], of which lamellar bone is the most abundant type in many mammals, including humans. Previous studies have demonstrated that bone tissues are primarily adapted to the variety of mechanical functions that they fulfill. Although the ultra-structural organizational patterns of various bone tissues differ from one another, they all appear to share many common properties at the molecular level of organization, e.g., the characteristics of collagen fibrils as mentioned in Section 2.1. Furthermore, the tissue all follow a common developmental course, namely that the major organic constituent collagen is synthesized, extruded from the cell, and then self-assembled in the extracellular space before mineralization begins. For this reason, bone is a good example of an ‘‘organic matrix-mediated’’ mineralization process [39]. By utilizing the normally calcifying leg tendons of the domestic turkey, Meleagris gallopavo, Landis et al. have demonstrated a wonderful model of mineralized collagen assembly and examined the structural relations between the collagen and mineral components in tendon and other vertebrate tissues [21,40,41]. Both the structure and assembly of collagen impart a specific character via the principal framework in which the mineral forms. The triple helical collagen molecules are first organized into linear arrays with the N-termini of one molecule not adjacent to the C-termini of the next one. In this manner the molecules can cross-link into two-dimensional quarter-staggered arrays
of holes and overlap zones. It has been proposed that these arrays strictly and contiguously pack and assemble in three-dimensions to create hole zone channels or gaps, as a consequence of the exact registration of the hole and overlap sites among the constituent molecules [40,42–45]. This assembly and organization of collagens in mineralized tissues creates intramolecular space where the nucleation and growth of apatite crystals can occur [46]. The earliest detectable platelet-like crystals of apatite nucleate principally within the hole zone channels and thereby become develop a preferred orientation. Following independent growth and development of crystals within numerous gaps and overlap sites, the mineral-collagen interaction results in the formation of a wide variety of apatite platelets each characterized by preferential growth along their crystallographic c-axis. This results in a structure where the crystals are all generally parallel to each other and to the long axis of the collagen fibrils accommodating them [40,47]. However, the individual holes are about 1.5–2.0 nm in diameter and some 36 nm in length. In contrast, bone crystals are on the average of 50 nm long, 28 nm wide, and around 2 nm thick. The notion that a single crystal can fit into an individual hole cannot therefore be correct. There must instead be some special structural arrangement of the triple helical molecules in type I collagen that allows accommodation of the crystals. The platelets are coplanar as they grow, a result presumably dictated both by the specific stereochemistry, including the nature of the flexible regions, comprising the hole channels, and by the cross-linking character of the collagen assemblage. The factors that dictate nucleation of apatite in association with the gap and overlap regions of collagen fibrils are not clear. It is most probable that there are critical stereochemical arrangements comprising the sites that provide requisite bonds for the individual nucleation events and subsequent parallel orientation and alignment of crystals [48,49]. The major cross-linking character of collagen described above is also important [50]. Models have been proposed that relate the cross-linking character to the molecular location of both the hole and overlap zones and their fine structure [51]. The models have identified both hydrophobic and hydrophilic sites as potential locations for mediating mineral formation in the tissue. Moreover, it has been suggested that the continuously flexible regions within the collagen hole zone may play a role in calcium and phosphate localization and also assist in binding non- collagenous proteins at these sites [51]. In addition to crystal growth within the channels and gaps, nucleation and progressive mineralization on the surface of the collagen fibrils may also occur [48]. This process was first suggested by Nylen et al. in a study of tendon calcification, and was subsequently visually supported by a few in vivo investigations of natural calcified tissues [5,52], as well as by in vitro biomimetic experiments of collagen mineralization [53,54]. As the crystals are intimately associated with the collagen framework in which they form, resulting in a highly complex but ordered mineral–organic composite material, the composite itself is organized into layers or lamellae with the thickness of a few microns. These layers in turn are arranged into higher order structures in a variety of ways depending on the different bone types [20]. The collagen structure, organization, and mineral formation process, are, however, highly complex, and are not completely understood, particularly as the structural hierarchical order increases [20]. Over a given hierarchical sequence, changes in the assembly and organization of mineralized collagen occur to yield more and more complicated arrangements of crystal platelets, plates, and lamellae associated with the collagen fibrils and fibers, both at the macroscopic and anatomic levels. At higher hierarchical assembly levels the mineralized collagen fibrils in bone are almost always present as bundles or aligned arrays, although these can be arranged in a variety of different motifs. One such motif is the lamellar structure [39]. A basic structural motif of lamellar bone is the presence of arrays of parallel collagen fibrils, with successive arrays having different orientations to form a plywood-like structure. Weiner et al. have measured the plywood angles from many micrographs taken from lamellar bones and found a bimodal distribution, with a major peak centered at 30 8 and a minor peak at around 70 8 [55]. They have suggested therefore that lamellar bone should be viewed as a series of lamellar units. The model for collagen organization in a single lamellar unit is characterized by units made up of five sublayers containing successive arrays of parallel fibrils, with their orientation to the lamellar boundary plane increasing in four increments of about 30 8. The first sublayer of fibrils, adjacent to one side of the lamellar boundary, is aligned perpendicular to the long axis of the bone, as observed by Weiner et al. [56,57]. Based on the fact that in vitrified transverse sections one layer is almost inevitably in the plane of the section, they proposed that this first sublayer should arbitrarily be assigned a value of 0 8 for the plywood angle. After four increments of 30 8 the final array is oriented at 120 8. Support to this model has been presented from scanning electron microscopy (SEM) micrographs that show that the fifth sublayer is indeed oriented in a different direction from the other layers [39]. This model is consistent with the twisted plywood model [58], but defines both
along the long axes. The hydroxyapatite (HA) crystals are mineralized initially in the holes zones of collagens. They then grow along the surface of the collagen fibrils, which results in a cross-striation periodicity of nearly 60–70 nm of each mineralized collagen fibril. These fibrils are always present in bundles or arrays aligned along their length. The fibril arrays organize into two common patterns: arrays of parallel fibrils or a plywood-like structure, as shown in the level 4 of Fig. 1. At higher levels, the initially deposited bone undergoes internal remodeling to form circular lamellar structured bone, with a central canal for notochord and two arches for the neural tube and blood vessels from the dorsal and ventral sides of the centra. The characteristics of biomineralization and of the microstructure of zebrafish bone are similar to those of human Haversian bone. Recent AFM and TEM investigations of the zebrafish skeletal bone have shown that as the degree of mineralization increases, the collagen fibrils became thicker and more ordered, with increased cross-linking from the outermost layer toward the center. In association with these structural variations, the nanomechanical properties as measured by nanoindentation significantly improve [66]. Recent investigation of wild
Fig. 1. The seven hierarchical levels of organization of the zebrafish skeleton bone [28]. Level 1: Isolated crystals and part of a collagen fibril with the triple helix structure. Level 2: Mineralized collagen fibrils. Level 3: The array of mineralized collagen fibrils with a cross-striation periodicity of nearly 60–70 nm. Level 4: Two fibril array patterns of organization as found in the zebrafish skeleton bone. Level 5: The lamellar structure in one vertebra. Level 6: A vertebra. Level 7: Skeleton bone.
zebrafish skeleton utilizing AFM and TEM has provided some new evidence for surface mineralization of collagen fibrils [37]. As shown in Fig. 2, observations of unstained zebrafish skeleton bone without decalcification treatments provide direct evidence for the deposition of the HA crystals on the surface of the fibrils. Further supportive evidence will depend on the results of investigations of mineralized collagens across the whole thickness of the bone wall. As a consequence of these processes, it is conjectured that as the mineralization becomes heavier, more minerals deposit on the surface, and therefore the fibrils become thicker. Results reported in regard to the structural characteristics of zebrafish bone walls generally concern the formation of the hierarchical assembly of mineralized collagen fibrils. Collagen filaments are secreted by osteoblasts in the outmost layer, and then self-assemble at the collagen fibrils. The newly formed collagen fibrils are therefore thin and lie with a random distribution. Only a few minerals deposit in such thin fibrils. As the age of bone increases from the outmost bone layer to the inner layer, the fibrils increasingly undergo stronger interactions and intercrossings with each other. Accordingly the fibrils assemble into thicker mineralized fibers that are aligned preferentially along the circumferential orientation of the bone wall and which in turn become more ordered and increase in diameter tending to form densely packed collagen fibril bundles. In addition to this research on wild zebrafish, studies have also demonstrated the predominance of systematic mutagenesis of zebrafish with regard to bone diseases. Using various materials analysis technologies, including SEM, AFM, TEM and polarized light microscopy, significant variations of the hierarchical organization [28] have been reported in gene-mutated zebrafish skeleton bone, all of which are associated with abnormal bone mineralization and bone disease. Light microscopy examinations reveal that gene mutant bone has asymmetric mineralization and much thinner bone wall. The SEM studies show the presence of many microcracks in such bone wall. Additionally, the plywood-like structure of normal bone is not observed in mutant bone, as confirmed by polarized light microscopy measurements. Furthermore, TEM investigations reveal two typical diameters for the collagen fibrils. The diameter of the thinner collagen fibrils in mutant bone is about twice that that of natural bone. For the thicker fibrils, there is only a small increase in diameter after mutation, consistent with the result of AFM studies. Study of the morphologies of the minerals revealed that the minerals in mutated bone were in bigger size and that their shape was irregular and not plate- shaped [67]. These hierarchically structural abnormalities, together with the studies of the nanomechanical properties
Fig. 2. Observation in the TEM of collagen fibrils from zebrafish skeleton reveals the existence of surface minerals [36]. The TEM micrograph shows the relationship between the collagen fibrils and the surface minerals at high resolution. The regions adjacent to the fibrils show a high electron density, indicating that these regions are filled with crystals. Furthermore, crystals are also observed in the middle areas of the fibrils as well, although the electron density in these regions is lower. The black arrows point to collagen fibrils of diameter 85.8 5.3 nm, while the white arrows point to surface minerals that deposit along the fibrils of 60.6 5.6 nm in thickness. The measurements of the characteristic sizes of the fibrils and the minerals illustrate that the sum of the diameter of the fibrils and the thickness of the surface mineral layer matches the diameter of the mineralized fibrils. The SAD pattern (insert) taken from domain A shows that the crystals are HA in nature and have a preferred (0 0 2) crystallographic orientation over the small region examined. The diagram to the right of the TEM image shows the proposed model for the deposition of minerals on the surface of the collagen fibrils.
bone have average maximum lengths and widths of about 50 and 25 nm, respectively, just as Robinson reported in 1952 [79]. Another indirect approach for estimating average bone crystal sizes is to measure X-ray reflection line widths and then analyze the data using the Scherrer equation. The resulting parameter is directly related to the coherence length, which can be assumed to correspond to the particle size dimensions if the crystals are small and sufficiently perfect. Finean and Engstrom [91] were the first to use this method. Subsequently it has been widely used [92–94], yielding the result that human bone crystals tend to be between 10 and 35 nm long. It is clear that compared to estimation of the crystal size using the XRD method, direct measurement of dispersed crystals using the TEM is a more reliable method. However values obtained from TEM studies are probably underestimates of the in vivo crystal sizes, because of the potential dissolution and breakage of crystals during sample preparation. Another problem in TEM investigations is the subjective choice of the crystals to be measured [95]. To such problems the use of other methods, such as atomic force microscopy (AFM), has been investigated for the measurement of the crystal size in natural tissue. AFM in particular has many advantages compared to the other methods listed above for the direct observation of crystals in natural tissue [96–99].
- Biomimetic fabrication with self-assembled collagen mineralization
3.1. Nucleation sites for collagen fibrils and the initial stage of collagen mineralization
Previous investigations into the nucleation sites of HA crystals on collagen fibers have suggested that the binding of calcium ions on the negatively charged carboxylate groups of collagen is one of the key factors for the first-step nucleation of HA crystals [100]. Other partially negatively charged functional groups on collagen have previously been considered to be possible nucleation sites for HA crystals. It has been demonstrated, however, that carboxyl groups (–COOH) are the major nucleation sites for collagen fibrils, and these are present in about 11% of the amino acid residues of collagen molecules. In a neutral solution, more than 99% of the carboxyl groups of aspartyl and glutamyl ionize to COO, which favors chelation of calcium ions. The carboxyl groups on the outside of the collagen threefold spiral are one kind of site for collagen mineralization. The molecular modeling method has used by Yang et al. [101] to investigate the interactions between calcium ions and the collagen-like peptide CH3CO-(Gly-Pro-Pro)10-NHCH 3 , which was employed as a simplified model of collagen. Seven C O groups along the longitudinal axis of the collagen-like peptide were specified, as is shown in detail in Fig. 3(a). These seven groups were then classified into three triplets: (1)(2)(3), (3)(4)(5), and (5)(6)(7) (see Fig. 4 for the numbering scheme). The calcium ions were placed step by step, such that the distance to each of the three oxygen atoms in the C O groups of every triplet was 6 A˚ , as shown in Fig. 3(b, c and d). The molecular mechanics simulations were then carried out for each step.
Table 1 Size of crystals found in some natural tissues
Source Length, nm Width, nm Preparation procedure
Human cortical bone 50(20–150) 25(10–80) Autoclaving [79] 40 40 Embedded thin section [90]
Human cortical femur 53.5(22.5–110) 28.5(15–60) NaOCL [89] Whale tympanic bulla 48.5(15–100) 26(7.5–45) NaOCL [89] Calf cortical bone 49(15–90) 21(7.5–45) NaOCL [89] Rat cortical bone 40.5(15–90) 21.5(7.5–45) NaOCL [89] Tuna vertebra 35(15–67.5) 20.5(7.5–37.5) NaOCL [89] Turkey calcified tendon 35.5(15–75) 20.5(7.5–45) NaOCL [89] Turkey calcified tendon 29(13–97.5) 16.5(5–50) Ethanol [89] Human fetal femur (16 weeks) 18.1 10.4 11.5 5.4 NaOCL [5] Human fetal femur (19 weeks) 27.4 12.6 16.7 7.5 NaOCL [5] Human fetal femur (21 weeks) 28.8 14.5 19.3 8.6 NaOCL [5] Human fetal femur (26 weeks) 31.6 17.5 19.2 9.2 NaOCL [5] Ivory 31 1.7 20 1.2 NaOCL [73]
Compiled and adapted from Refs. [5,73,79,89,90].
Optimized structures for systems including collagen-like peptide and one calcium ion, two calcium ions, and three calcium ions in order were obtained. The distances between the calcium ion and the three oxygen atoms of the C O groups were 5.6, 3.7, 3.4 A˚ for the case of just one calcium ion placed near triplet (1)(2)(3). The distances increased to 7.7, 6.3, 6.4 A˚ when a second calcium ion was placed near triplet (3)(4)(5). This increase is due to the repulsion between the calcium ions. When a third calcium ion was placed near triplet (5)(6)(7), the final predicted configuration had a similar structure to that of HA crystal. The predicted calcium–oxygen spacings were 10.3, and 9.1 A˚ , with an angle from calcium to two oxygen of 79.8 8 , in comparison to HA crystal with values of 10.7, 9.3 A˚ with an angle of 83.8 8. The results showed that collagen-like peptide can effectively attract calcium ions, and that the binding energies for the first, second, and third precipitated calcium ions were 145.7, 69.8, and 47.1 kCal/mol, respectively. These results suggest that the collagen-like peptide can directly guide the arrangement of calcium ions during the initial stage of the collagen-mineralization process. This initial control may also be one of the underlying mechanisms controlling the formation and regular orientation of HA crystals in the collagen-mineralization process. Zhang et al. have reported some investigations into the nucleation sites of calcium phosphate crystals on collagen fibers by using FTIR [53], which has a broader range of spectrometry than that used in previous investigations [100]. The amide arrangement in collagen was analyzed by comparative measurements of the positions and intensities of three representative peaks from the resulting spectra. It was concluded that the amide I peak at 1657 cm^1 predominantly corresponds to the C O stretch, the amide II peak at 1528 cm^1 corresponds to a combination of the
Fig. 3. (a) The optimized (lowest energy) structure of the collagen-like peptide CH 3 CO-(Gly-Pro-Pro)10-NHCH 3. White, red, grey and blue spheres in the picture represent hydrogen, oxygen, carbon, and nitrogen atoms respectively. The three green and purple spheres represent the hydrogen atoms in CH 3 CO– and –NHCH 3 , respectively. Numbers in the picture represent the seven specified C O groups. (b–d) The three calcium ions placed near certain C O groups. All the solid green lines in the three pictures represent a distance of 6 A˚ in length. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
gap zones. Pederson and Ruberti [108] reported a strategy for exploiting temperature driven self-assembly of collagen and thermally triggered liposome mineralization to form a mineralized collagen composite. Our team have now synthetically prepared nano-fibrils of mineralized collagen as a self-assembly model system, with the objective of evaluating the possibility of synthesizing materials with hierarchical structures similar to those found in nature [54]. For this we used different compositions of monomeric collagen and solutions containing calcium and phosphate ions, and then used either pH or temperature to induce the formation of collagen fibrils. TEM investigations (Fig. 4(a)) of unstained samples at low magnification revealed that the composites formed consist of an intertwined assembly of collagen fibrils bundles more than 1 mm long. Each collagen fibril is surrounded by a layer of HA nanocrystals grown on the surface of the collagen fibrils. Each mineralized bundle of collagen fibrils is much
Fig. 4. TEM image of mineralized collagen fibrils. The insert is a selected area electron diffraction pattern of the mineralized collagen fibrils. HRTEM image of mineralized collagen fibrils. The two long arrows indicate the longitudinal direction of the collagen fibrils. The two short arrows indicate two HA crystals.
thicker than the self-assembled collagen fibrils, implying that the self-assembled collagen nano-fibrils act as the template for HA precipitation. Additionally, in order to discern the relative orientation of the HA crystals with respect to collagen fibrils, electron diffraction investigation have also been carried out. The results finally demonstrate the preferential alignment of the HA crystallographic c-axis with the collagen fibril longitudinal axis. To study the relationship between the newly formed crystals and the collagen fibrils directly, high resolution TEM (HRTEM) analysis of the lattice plane arrangements were performed (Fig. 4(b)). The HRTEM analysis of the parallel-aligned mineralized collagen fibrils revealed that crystal lattice is seen not only on the side area of the collagen fibrils, but also in the middle area, and that the electron density on the surface of the collagen fibrils is higher than in the interior area. These findings indicate that HA crystals grown on the surface of the collagen surround the fibrils, giving the first direct evidence to support previous theories that this occurs. A schematic picture of self-assembled HA/collagen composites comprising multiple levels of hierarchical organization has been depicted in this study. As shown in Fig. 5, the lowest level of this hierarchy is the organization of collagen molecules with some particulates of nano-HA crystals. The collagen fibrils are formed by self-assembly of collagen triple helices, and the HA crystals are formed initially in the gap zones between the collagens. Considering that the diameter of the collagen molecule is 1.5 nm, the diameter of the collagen fibrils in the five-stranded packing model should be approximately 4.0 nm, which is in agreement with TEM observations of self-assembled collagen fibrils. The second level of the hierarchy is the organization of these collagen fibrils with respect to the growth of the HA crystals. HA crystals, which have a sheet-like shape, grow on the surface of these fibrils in such a way that their c- axes are oriented along the longitudinal axes of the fibrils and surround the fibrils. This arrangement implies that the nucleation and growth of HA crystals are not random but rather are controlled by the fibrils themselves. The third level of the hierarchy is the organization of these mineralized collagen fibrils, which are aligned parallel to one another to form mineralized collagen fibers. The epitaxial growth mechanism may possibly be used to explain the assembly. As mentioned previously, it has been shown that the negatively charged groups of the collagen molecules are the nucleation sites for HA crystals. The positions of the oxygen atoms in the hydroxyl groups of HA crystals have epitaxial relationships with those in the carboxylate groups of collagen fibrils. The nanoscale organization of the composites resembles that of HA crystals in mineralized tissue in which the HA crystals also align their c-axes with the longitudinal axes of the collagen fibrils. Such alignment is the most impressive characteristic of bone minerals. Development of novel self-assembled structures should therefore improve our understanding of collagen-mediated mineralization in other calcified tissues, and point the way to the development of new functional materials for biomimetic engineering. Moreover, these fundamental studies provide the basic theoretical support for the fabrication of HA/collagen composites and their application in bone regeneration [109,110,102], both of which will be introduced in the next sections.
3.3. Assembly and mineralization of peptide-amphiphilic nanofibers
There is no doubt that the preparation of any material with multiple levels of hierarchical organization on the nanoscale is a challenging problem. Fabrication of materials that resemble bone, even at the lowest level of
Fig. 5. The hierarchical structure of a self-assembled HA-collagen composite [53]. (I) The first level of the hierarchy is the organization of collagen molecules with the nano-HA crystals formed initially in the gap zones between the collagen fibrils. (II) The second level of the hierarchy is the organization of collagen fibrils with respect to HA crystals. The HA crystals are sheet like and grow on the surface of these fibrils in such a way that their c- axes are oriented along the longitudinal axes of the fibrils, as indicated by the white arrows in the figure. (III) The third level of the hierarchy is the organization of the mineralized collagen fibrils. A number of mineralized collagen fibrils align parallel to each other to form mineralized collagen fibers.
viable for commercial manufacture. Work has therefore been started to find new expression hosts by use of multigene expression technology [129,131,134,150]. Purified recombinant collagens are capable of undergoing spontaneous alignment to form collagen fibrils and defined features that are characteristic of collagen. Recent studies [151,152] have revealed that recombinant human- like collagen has the same characteristics in the initial minerlization stage as natural collagen. Additionally it also can induce the deposition and direct the growth of HA nanocrystals in vitro, in the form of self-assembly of nano-fibrils of mineralized collagen resembling an extracellular matrix. Molecular self-assembly is a powerful approach for the synthesis of novel supramolecular architectures. Zhang and colleagues have focused on the fabrication of several self-assembling peptides and proteins for a variety of studies of biomaterials [153]. Their studies have shown that a broad range of peptides and proteins have the ability to produce very stable nanofibers, which are very well ordered and possess remarkable regularity and helical periodicity [154,155]. Moreover, these nanofibers are similar in scale to the extracellular matrices that are crucial in manufacturing artificial functional tissues. Furthermore, work in their group has demonstrated that a variety of cells, including neuronal cells, encapsulated and grown in three-dimensional peptide scaffolds, show interesting functional cellular behaviors, including proliferation, functional differentiation, active migration, and extensive production of their own extracellular matrices [156–158].
- Applications of mineralized collagen composites in bone regeneration
4.1. Synthesis and properties of nano-HA/collagen composites
Although bone has good mechanical properties, it often undergoes damage or suffers defects resulting from tumor reconstruction, chronic infection or traumatic bone loss. There is a great need therefore for bone graft materials for various bone applications. Autologous bone grafting is widely accepted as the gold standard for the treatment of bone defects and nonunions [159]. However, the supply of autograft is limited and donor site morbidity is also a concern along with the required prolonged operation times [160]. The alternative of allogeneic bone has potential risks of disease transmission and infection [161]. In order to avoid the problems associated with either autologous or allogeneic bone grafts, there has been a continuing interest in the use of synthetic bone graft materials during the past decades. Table 2 lists some example commercial products, including matrix material derived from natural bone as well as synthetic substitutes [163–169]. However, these substitute matrices each have specific disadvantages in biocompatibility, degradability, osteogenic capability and histochemical responses by the host tissue. The ideal bone scaffold should promote early mineralization and support new bone formation while at the same time allowing for replacement by new bone. Based on this principle, nano-HA/collagen based composites, inspired from research on natural bone, have received great attention. The composites are prepared by directly mixing the nano- HA and collagen. Nano-HA is sourced from the modern ceramic technology, while collagen is purified from animal tissue as fixing agent for HA. The weak binding between HA and collagen make them no cooperation effect in vivo for bone defects repair. The collagen degrades fast, but HA ceramic remains in the original form which do not attend the remodeling progress of bone. Researchers have also tried to develop mimetic methods to prepare nano-HA/collagen composite. One such method involves immersion in a simulated body fluid (SBF), used to improve the biocompatibility of conventional metal, alloy and polymer implants. Another method involves co-precipitation of collagen fibrils and nano-HA spontaneously, and is a promising route for achieving the same hierarchical structure in synthetic materials as in bone.
4.1.1. Direct blended nano-HA/collagen composites Composites can be made by blending or mixing a heterogeneous combination of two or more materials, each which differ in morphology or composition. Although blending is not a new concept, it has gained a considerable interest in the past few years. It is well known that blending of multiple materials with different characteristics leads to composites with tailor-made properties, but it is quite difficult to control the homogeneity and uniformity of the secondary or reinforcing phases. Based on earlier research into the advantages of using HA and collagen separately, a combination of HA and collagen should provide an advantage over other materials. About ten years ago this point was realized by several researchers who then began to fabricated such composites. Injectable HA/collagen composite gels have also been evaluated. The bone bonding property of HA ceramic was used to sustain new bone formation after
collagen degradation. Current commercialized HA/collagen composites are all a direct mixture of these two components [169]. In blended HA/collagen composites the crystallite sizes of HA are not uniform, and the HA is often aggregated and randomly distributed into the fibrous matrix. There is, therefore, no structural similarity to natural bone and only a compositional similarity to that of natural bone. Furthermore there is no sign of chemical bonding between the HA and collagen phases, probably due to the lack of suitable interfacial-bonding during mixing. It has also been observed that synthetic HA crystals remain unaltered, even after long-term implantation. This may hinder the subsequent bone formation and remodeling. Implant materials with much a higher modulus than bone can impose a stress-shielding on the surrounding bone bed, leading to bone resorption and to a loosening of the implant over the longer term. Since cell processing is essential for the construction of the complex and hierarchical structures of bone, the interaction of a material with bone cells is an important issue in the development of bone substitute materials. This type of processing method is thus not recommended for making high performance composites with both the structural and compositional characteristics of natural bone. It has been shown however that a modified blending method is able to retain the self-assembly collagen fibrils after blending with the nano-HA [170]. Such composites are made up of bundles of collagen fibers embedded with nanocrystals of HA in a sponge-like structure by freeze-drying. This procedure represents an important improvement for the development of nano-HA/collagen composites, as the self-assembly biomimetic concept in this nanocomposite is a key factor for its preparation.
4.1.2. SBF immersion technology SBF has been the basis of a large amount of research on biomimetic apatite coating techniques. This solution has a nearby equal ionic concentration to that of human blood plasma but does not contain an organic component. The simplest, and first tried, strategy to provide enough ions for apatite formation is to refresh the SBF every day or every week. However, apatite formation is very slow and uncontrollable. The difficulty in using such solution reflects the fact that our body fluid provides sufficient ionic resources for bone formation but that mineral precipitation is delicately
Table 2 Some commercial bone graft products. Compiled and adapted from [162–169]
Company Product name Composition and form
Synthes DBX DBM available in putty or paste form [163]
GenSci Regeneration Sciences DynaGraft DBM available as an injectable gel, matrix or putty [169] Orthoblast DBM and allograft cancellous bone chips, injectable paste or putty [169]
Osteotech Grafton DBM with Glycerol. Available in gel, flex, and putty forms [163] Sofamor Danek Osteofil DBM (24%) with gelatin carrier (17%) and water [164] Exactech Opteform Compacted corticocancellous bone chips mixed with the same materials as Osteofil [163] Geistlich Biomaterials BioOss Hydroxyapatite particulate [169] Ceremed Dental OsteoGraf Hydroxyapatite, particulates or blocks [169]
Wright Medical Technology Allomatrix Demineralized human bone matrix in an Osteoset medium (calcium sulfate powder), injectable or malleable putty [165,169] CELLPLEX TCP synthetic cancellous bone [165] Osteoset Calcium sulfate pellets [165,169]
Zimmer Collagraft Bovine collagen, HA, and TCP available in granular and strip configurations [163,164] DePuy CONDUIT Synthetic Porous Ceramic (TCP) granules [166] Merck KGaA Endobon Sintered bovine cancellous bone blocks [164] US Biomaterials NovaBore Bioactive glass (SiO 2 and minerals) [164] Interpore Cross ProOsteon Coralline HA granules and blocks: 200, 500, and R forms. Harvested from marine coral exoskeleton which is hydrothermally converted to HA [168,169] Norian SRS Calcium phosphate (carbonated apatite) injectable cement [163,164] Orthovita Vitoss Ultraporous b-TCP [168,169] Orquest Healos Hydroxyapatite-coated collagen microfibers [167,169] Allgens Relive Mineralized collagen type I [162]
DBM: demineralized bone matrix.
solution. Simple immersion in SBF cannot support mineralization of collagen. One reason is that the ion concentration is insufficient for mineralization; another reason is that there is no trigger for self-assembly of collagen. Accordingly, after completion of the collagen self-assembly, the mineralization on the collagen is too restricted to form into an apatite assembly with a pattern as in natural bone. Guided by the above observations, we have designed a manufacturing route for assembly of a nano-HA/collagen composite, and the repeatability of this method has been demonstrated [162,176]. Moreover, three-dimensional scaffolds have also been prepared with the addition a small amount of poly(lactic acid) (PLA) polymer, to support cell ingrowth [162]. It has been proven that these materials are nanocomposites in which nano-sized bone-like apatite is induced and uniformly embedded in collagen matrix. The self-organization of calcified collagen fibrils into fibers is also possible under certain conditions. These materials are good examples that mimic natural bone to some extent at the ultra-structural level. Their subsequent biological evaluation suggests that such bone-resembling composites are readily incorporated into the bone metabolism in a way similar to bone remodeling, instead of acting as permanent implant [110,162].
4.2. Applications of nano-HA/collagen composites
4.2.1. Gels and powders as a filler for orthopaedic surgery Pederson et al. have described a strategy for exploiting temperature driven self-assembly of collagen and thermally triggered liposome mineralization to form a mineralized collagen composite from an injectable precursor fluid [108]. They combined thermally triggered liposome mineralization with collagen gel formation in an attempt to develop an injectable mineral/collagen composite gel. The combination of an acid-soluble collagen solution with calcium- and phosphate-loaded liposomes composed of dipalmitoylphosphatidylcholine and dimyristoylphosphatidylcholine resulted in a liposome/collagen precursor fluid, which when heated from room temperature to 37 8 C formed a mineralized collagen gel. They also found that the dynamic storage modulus of the thus prepared scaffold increased upon mineralization. This finding suggests that in situ liposomal mineralization of collagen gels might provide a route for formation of mineral/collagen composites with significant interaction between the mineral and collagen phases. However, for this approach to be practical a significant increase of the modulus into a range more suitable for repair of skeletal hard tissue will be required. In addition to conventional directly mixed HA/collagen composites, the freeze drying method is a common route for preparation, storage and transportation of self-assembly nano-HA/collagen composites [100,102,110,162,176]. With this technique the assembled structure of collagen with hydroxyapatite can be kept until its in-patient use. Freeze-dried nano-HA/collagen composites can be implanted into bone defects as bioactive materials with or without autogenous bone. In most orthopedic surgery there is not enough autogenous bone to fill the entire defect space. Removal of bone from another part of a patient will result in increased pain in the possibility of inflammation, as well as prolonging the surgery time. Biomimetic bone materials can instead be used in conjunction with natural bone, to induce new bone tissue formation and promote bone remodeling. At present this is the most promising route for the repair of defects in natural bone.
4.2.2. Surface coatings for implant materials In this section, we will discuss coating of HA/collagen composites on various substrates. In combination with the currently used implantation materials, nano-HA/collagen composites have shown superior properties compared to other conventional coating methods. In a study by Fan et al., a uniform collagen fibril/OCP composite coating on silicon substrate was prepared by electrolytic deposition [177]. The coating process involved self-assembly of collagen fibrils and the subsequent deposition of calcium phosphate minerals as a result of the cathode (Si) reaction and a local pH increase. The porous composite layer thus formed consists of a collagen fibril network on which clusters of octacalcium phosphate crystals nucleate and grow. Preliminary results by nanoindentation testing showed that properly prepared composite coatings may have a higher elastic modulus and improved scratch resistance compared to monolithic porous calcium phosphate coatings. Poorly crystallized HA/collagen composite coatings formed on the surface of NiTi shape memory alloy have been studied by introducing collagen to a simulated body fluid [172]. The morphology of the composite coatings produced is uniform and the micro-morphology is lamellar. This coating may be very useful for enhancing the bioactivity of NiTi shape memory alloys.
4.2.3. Three-dimensional scaffolds for bone tissue engineering The principal concept of tissue engineering is to isolate a small biopsy of specific cells from a patient, to culture these cells on a scaffold material, and then transplant the cell-engineered scaffold into the defect site of the patient’s body in order to direct new tissue formation into the scaffold, which will biodegrade over time. An ideal material for bone tissue engineering should therefore be nonimmunogenic, biodegradable, highly effective in osteoinduction at relatively low doses of inducing signals, ready for vascular and mesenchymal cell ingrowth, and amenable to contouring for optimal adaptation to various shapes of bone defects, thereby providing mechanical support when needed. Tissue-engineered HA/collagen nanocomposites seem to be a very promising system for bone reconstructive or regenerative surgery. We have already developed osteogenic cells/nanocomposite scaffold structures using both advanced culture techniques as well as by conventional static culture methods, and in vitro cellular functions of these materials have been investigated [110]. It was found that the scaffolds support well cellular growth and related functions, and lead to new bone formation. More recently, a three-dimensional bone-resembling nanocomposite matrix prepared using nano-HA/collagen/osteoblasts has been developed in conjunction with PLA [178]. This system supports cellular adhesion, proliferation, and migration. Interestingly, cells grown on this material were observed to penetrate deep into the matrix, to a depth of about 200–400 mm, within a short period, probably due to both the compositional and structural similarity of this material with natural bone, thereby providing a promising cell/scaffold for bone tissue engineering. Scaffolds loaded with growth factor have been shown to regulate cellular growth and related functions in a better way [179]. Growth factors can be effectively delivered to a bone defect through nanocomposites, and the in vivo efficacy of such methods have been evaluated [162]. The mechanical reliability of this system corroborates well with the strength of autogeneous cancellous bone and the in vivo performance of the nanocomposite with recombinant human bone morphogenetic protein-2 (rhBMP-2) is better than that of the nanocomposite without rhBMP-2. In use, early bone formation occurred for the rhBMP-2 treated composite, which implies its efficacy as a good bone graft. The efficacy in bone regeneration of such a scaffold combined with bone marrow mesenchymal stem cells was also evaluated [180]. It was found that the implanted scaffold could enhance and accelerate bone formation in segmental defects in rabbits. These experimental results indicate that an effective bone graft should consist of an osteoconductive matrix in conjunction with osteogenic cells and osteoinductive growth factors with a structure, composition, physicochemical, mechanical, and biological features analogous to natural bone. Bone tissue engineering using nano-HA/collagen composites is still however in its infancy although our knowledge of this area is expanding. Although in vitro and in vivo evidence strongly supports the effective use of this biomimetic nanocomposites as bone graft materials, further clinical studies are needed to confirm their promise as effective graft materials for bone regeneration.
4.3. Related improvements of mineralized collagen composites
4.3.1. Optimization of controllable ionic substitution in nano-HA crystals Nano-HA is generally accepted as a prototype for the apatite minerals in calcified tissues. Among these minerals the lattice ions of HA are substituted, to different extents, by other ions, including F, Cl, CO 32 , Na +, K +, Mg 2+^ , and Sr2+^. Such substitutions have a large influence on the physical and chemical properties, as well as the physiochemical properties, of apatite. Current research into ionic substitution in HA is focused on carbonated silicon, magnesium, and fluoride, as described in following section. However, the number of studies into the co-reaction of collagen with ion- substituted nano-HA still remains extremely limited. Many experiments have shown the high solubility of carbonated hydroxyapatite (CHA) in vitro and in vivo. More highly carbonated minerals are associated with higher osteoconductivity and earlier bioresorption. On the other hand, the differences in carbonated content in natural bone reflect differences in the ages and other physiological conditions between individuals. Scientists have tried by control of the morphology and chemical composition, to prepare nano- sized CHA with low crystallinity and of a nanometer size suitable for the processing of bone-resembling materials [181,182] by various methods. In order to produce an ideal bone material that mimics natural bone tissue, we have prepared nano-CHA/collagen composites by a self-assembly method [176]. Moreover, the CHA is combined with collagen fibrils by a self-assembly method, exactly mimicing the components and microstructure found in natural bone. Based on this research, it is now possible to achieve the proper type of nCHA when this material is implanted in different defect sites in hard tissue by controlling the reaction condition.
