The incorporation of biological systems into the toolbox of chemists is a major trend in today's materials science.1 This interest was initially triggered by two main factors: enhanced concern for sustainable development and tremendous efforts toward biomimetic/bioinspired approaches. In the first case, organic matter of biological origin is considered as a renewable resource (i.e., of nonpetroleum origin) with low cost and limited environmental impact.2, 3 In the second case, condensed phases formed by living organisms, that is, soft and hard tissues, constitute a model to design novel functional materials due to their astonishing structures and properties.4, 5 From these considerations, a new class of materials, termed bionanocomposites, has recently emerged.6 A possible definition of a bionanocomposite can be established on the basis of the traditional acceptance of nanocomposite, as a multiphase material where one of the phases has at least one dimension of less than 100 nm, and adding one of the phases of biological origin. In many cases, a more restricted acceptance of the term considers that bionanocomposites contain at least one biological and one inorganic component, making them closely related to bio-hybrid systems.7 Under this assumption, two limit systems can be envisioned consisting either of inorganic colloids entrapped in a biomolecular network or of biological colloids incorporated within an inorganic matrix. In the following review, only the first type of bionanocomposites will be discussed and we will therefore focus our presentation on biopolymer-based matrices, that is, hydrogels.
One very key advantage of using hydrogels is to provide the system with a high degree of modularity, dynamics (e.g., reversibility) and responsiveness. In particular, using hydrogels of biological origin aims at targeting responsiveness toward biological stimuli for the development of materials with applications in biomedical fields (drug delivery, imaging, biosensing, and tissue engineering). Noticeably, although the use of biopolymers allows building materials that are intrinsically bioresponsive, it is worth mentioning that bio-sensitive materials can also be prepared by modification of inorganic materials, such as mesoporous oxides.8 For example, adequate surface functionalization with organic groups9 allows the improvement of the accessibility of external guests,10 or favor biosystems immobilization on nanoparticles surface.11, 12
The purpose of elaborating bionanocomposite materials is to improve the properties of one of the component, to add up the functionalities of the partner phases and, ideally, to create synergetic effects between the bioorganic and inorganic components. In Table 1, we have summarized some of the key advantages and shortcomings of biomacromolecules and inorganic colloids. On this basis, hydrogels made up of biopolymers are well-adapted to work invivo (food, pharmacy, and medicine)13, 14 or in the environment15 but they often lack chemical and physical stability. Hence, chemical modification by crosslinking or incorporation within composite structures is often required to improve their properties and reduce degradation rates.16 In parallel, inorganic colloids exhibit many functional properties to designconductive, optical, or magnetic devices but most of them exhibit unfavorable interactions with living systems17 and their processing at the macroscale is a real issue.18
Table 1. Key Advantages and Shortcomings of Biomacromolecules and Inorganic Colloids
The goal of this review is to illustrate the benefits of building bionanocomposites, which can be prepared either by the mechanical mixing of a polymer with nanoparticles or through the in situ synthesis of particles using biopolymer as templates. After a short description of the preparation methods, we will specifically address the influence of the bio-inorganic interface on the overall physical properties of the material. In particular, materials with optical, conductive, magnetic, mechanical, and bioactive properties will be presented, that associate a variety of biopolymers, including mainly polysaccharides (alginate and chitosan) and proteins (collagen and gelatin) with metal, metal oxide, carbon nanotubes (CNTs), clays, and calcium phosphate nanoparticles, respectively.
The simplest method to prepare bionanocomposites based on biopolymer hydrogels is to start from a physical mixture between the polymer and the particles both in a sol phase and then to induce gelation, either via a change in temperature (gelatin), pH (collagen and chitosan) or addition of crosslinking species (alginate and carragheenan) [Scheme 1(a)]. The main advantage of this approach is that the size, morphology, and surface properties of the inorganic colloids are initially defined and designed. Moreover, if required, particles preparation can be performed in conditions that would not be compatible with the preservation of the biopolymer integrity or solubility, such as extreme pH, use of organic solvents, high temperature, or pressure. The main limitation of this technique is related to the achievable bio-organic/inorganic contents and ratios that depend (i) on the viscosity of the polymer sol, that increases with biopolymer concentration but also upon addition of nanoparticles,19 and (ii) on the colloidal stability of the inorganic particles, both in the initial media and in presence of the biomacromolecule sol. This will dramatically impact on the particle dispersion in the hydrogel and hence on the final homogeneity of the bionanocomposite.20 In certain cases, the addition of nanoparticles may even hinder the gelation process.21
This illustrates that the formation of bionanocomposites is governed by a subtle balance between particle/biopolymer interfacial interactions and biopolymer interchain bonding involved in hydrogel formation. For example, it was shown that colloidal silica could inhibit the formation of collagen gels, due to protein adsorption onto the nanoparticles.22 Noticeably, this phenomenon depends on the collagen concentration, that is, increasing collagen concentration allows more silica particles to be added while preserving the hydrogel network formation.23, 24 This can be explained considering that silica-collagen interactions deplete the solution of self-assembling protein so that the hydrogel forms only if the remaining collagen content is above the gelation critical concentration.
An interesting situation was observed when vanadate nanoclusters were mixed with gelatin sols at 40 °C. A complex coacervation process occurred due to favorable electrostatic interactions between the two components, giving rise to neutral aggregates forming a sol exhibiting a nonzero elastic modulus [Fig. 1(a)].25 When cooled down, a gel was formed and the inorganic species had almost no impact on the rheological properties of the system, suggesting that the protein–protein interactions are prevailing over the biopolymer-inorganic interface.26 In contrast, at high concentration, the coacervate system exhibited rubber-like properties at room temperature, due to vanadate crosslinking of the protein network, which were unpreceded in gelatin-based hydrogels [Fig. 1(b)].27
In Situ Synthesis of Particles, the Biotemplating Approach
Biopolymers can be used as templates for the synthesis of nanoparticles.28 When compared with mechanical mixing, this strategy should limit diffusion issues met with preformed particles. This method starts from inorganic nanoparticle precursor species, ions or poly-ions, further converted into solids by a second reaction sequence [Scheme 1(b)]. The first step is usually performed by addition of the inorganic precursor to the polymer sol before inducing gelation, but it is also possible to perform precursor impregnation once the hydrogel is obtained. The mineralization process can then be induced by pH modification, addition of reducing agent, carbonation or supplementation by any other suitable reagent, in the liquid phase or in the gas phase: for example, NaOH solution or ammonia vapors for pH increase, NaBH4 solutions, or H2/N2 gas flow for metal reduction. The gas phase should provide a better diffusion of the added reagent, but some important mineral groups, such as phosphates, are not easily obtained in a gaseous form. Moreover, the solution route is more adapted when a precise amount of reagent must be introduced to reach a certain stoichiometry. In principle, such a strategy may be used for the purpose of preparing nanoparticles of certain size, morphology, and packing and then recover them by removing the biopolymer network. Such a “sacrificial” approach of biotemplating can be very useful as a way to replicate complex morphologies found in biological systems,29 but the resulting material is not a biocomposite anymore.
In particular, polymer networks have been widely used for the synthesis of metal particles, starting from metallic salts that can be readily reduced chemically, thermally, or photocatalytically. For instance, cellulose fibers were used for the templated growth of a broad variety of metals (Ag, Au, Cu, Pt, and Pd) [Fig. 2(a)].30, 32, 33 In these examples, the nanoparticles density and size could be tuned with the concentration of precursor salts, the pH, and the reduction time. A recent report showed that cellulose can also be used for the controlled growth of metal oxide particles, such as silica.34 Other biopolymers have been used, as reported by Brayner et al., where gelatin allows the growth of gold particles from HAuCl4. In this case, when reduction is induced by the addition of hydrazine below the gelation temperature threshold, closed-packed linear assemblies along gelatin filaments could be observed [Fig. 2(b)].31 Interestingly, this approach could be extended to bimetallic Au-Ni nanoparticles.
An alternative approach is to use bifunctional ions that act both as mineral precursors and network crosslinkers for the gelation of ionotropic systems [Scheme 1(c)]. In contrast to the later synthetic procedure, the network crosslinkers can be distinguished from the particle precursors [Scheme 1(b)] by their key role in hydrogel formation before particle templating. In this respect, Co, Ni, and CoNi nanoparticles have been formed within alginate matrices, where metal cations act as crosslinkers of gel networks, which in turn act as a confined medium that limits the growth of metal particles.35 It was found that the crystalline structure of the Ni nanoparticles was dependent on the initial biopolymer concentration. This was attributed to both the change in viscosity that regulate the kinetics of particle growth through ion diffusion, and the density of metal-binding carboxylate groups of the alginate background that modify the growth environment.
OPTICAL PROPERTIES OF METAL-BASED BIONANOCOMPOSITES
Metal nanoparticles are known to have an unique optical properties due to their surface plasmon absorption attributed to the collective oscillation of free conduction electrons, together with their large surface areas.36 Hence, polymer bionanocomposites containing metal nanoparticles have attracted a great deal of attention, with resulting optical properties that strongly depend on the structure, size, and dispersion of the embedded particles. In this respect, the addition of metal nanoparticles within biopolymer hydrogels allows the tuning of their optical properties. This is observed when introducing different metal particles within cellulose hydrogels that caused color changes from light blue of cellulose to characteristic colors of the added metal [Fig. 3(a)].33 As reported by Cai et al., yellow and pink gels were obtained with embedded silver and gold particles, respectively, indicating the surface plasmon resonance of the particles. Gray gels resulted from the presence of platinum particles. Moreover, the light transmittance of these nanocomposites could be significantly modified when drying the gels in supercritical conditions. This difference is attributed to the light scattering of metal nanoparticles and porous cellulose matrix, and the different refraction index of media. This underlines the possibility to control the optical properties of bionanocomposites by drying treatment and materials processing. Importantly, this points out the effect of the light scattering properties of every component of the system that will contribute to the final optical properties of the resulting bionanocomposite. This aspect has also been illustrated in systems made up of gelatin hydrogels embedding silver particles prepared by in situ reduction of silver nitrate in the presence of hydrazine.38
In fact, the influence of the biopolymer matrix on the optical properties of the metal nanoparticles has been poorly explored and most analyses rely on the use of Mie theory. A more detailed modeling approach was performed on gold nanoparticles formed in situ within alginate films.39 The Maxwell-Garnett model was used and the dielectric function εeff of the effective medium was calculated using eq 1
considering metallic inclusions of component i (dielectric function εi and volume filling factor f) randomly distributed in a continuous matrix m (dielectric function εm). The dielectric function of the metal was determined from the bulk Au dielectric function modified by introducing a correction term A which reflects the intensity of the size effects and depends on the nature of the particle/matrix interface. These calculations indicated that the concentration of the alginate matrix did not modify the optical response of the gold colloids, whereas the alginate structure, that is, guluronic:mannuronic acid ratio, had a significant effect.
The conditions of preparation of the bionanocomposites can also strongly impact on their optical properties. In a study by Brayner et al.,31 sols of gelatin and HAuCl4 were prepared at 40 °C. Gold reduction was performed by addition of hydrazine at the same temperature and a gel with a red color (λmax = 538 nm) was obtained after cooling down. However, if the gel was formed before reduction, it exhibited a blue color (λmax= 618 nm). In both cases, gold colloids have a similar size (ca. 20 nm) with an expected maximum absorption at λmax ≈ 530 nm based on Mie theory. Electron micrographs indicate that particles formed in the sol are well-dispersed (no interparticle interaction). In contrast, colloids initially formed in the gel state are deposited along the gelatin fibers resulting from the gelation process, forming local aggregates accounting for the observed blue-shift of the plasmon band. These properties could be used for sensing applications, based on optical shift induced by particle aggregation. Indeed, Zhang and coworkers recently reported that the aggregation of silver nanoparticles within lentinan polysaccharide matrices was related to the conformation transition of the polysaccharide, from triple helix to random coil [Fig. 3(b)].37 As this work demonstrates that the variation of silver particles aggregation and resulting optical switch could serve as a structural probe, it opens up new route for the engineering of colorimetric detectors.
CONDUCTIVE PROPERTIES OF CARBON NANOTUBE-BASED BIONANOCOMPOSITES
Biopolymers have been scarcely studied as conductive polymers. Electron conductivity was identified in DNA strands due to π-stacking interactions of the bases.40 Proteins exhibiting conjugated systems, such as porphyrins in cytochromes, can also exhibit electron transfer.41 However, to our knowledge, none of these systems has been used in the gel form for the design of conductive materials. Ionic conductivity is expected to be more frequent due to the numerous biopolymers exhibiting polycationic or polyanionic behaviors. However, ionic conductivity requires a high density of ionized groups on the polymer backbone to allow efficient hoping of the charge carriers. Moreover, protein or polysaccharide conformations may lead to charge clustering rather than continuous charge density. This probably explains why conductivity data for biopolymer gels are not easily found in the literature.42
In contrast, metal, metal oxide, and carbon nanoparticles have been widely studied as electron and ionic conductors. However, within a composite material, conductivity requires a continuous network of conductors, that is, particles should reach a concentration-related percolation threshold to ensure connectivity. The variation of conductivity σ with particle volume fraction ϕ can be written as eq 2:43
with σ0 being a pre-exponential factor, ϕc the critical volume fraction at the percolation threshold, t the critical exponent, and a the anisotropy factor. In the case of spherical colloids (a = 1), this can only be achieved by using high particle concentration (ϕc) that is not easily reached due to previously discussed constraints. On the contrary, using particles with high aspect ratio (a ≫ 1), such as fibers or tubes, ϕc and t decrease, that is, electric connectivity is established at much lower concentration and is less sensitive to composition variation near the threshold. This is one of the reasons for the large development of conductive composites based on CNTs.
In the case of bionanocomposites, the biopolymer chains were initially introduced as a hydrophilic coating agent allowing the dispersion of initially hydrophobic nanotubes in aqueous solutions.44 A typical example is the gum arabic-CNT systems where stable dispersions can be obtained using low biopolymer amount (<1 wt %) and percolation could be achieved for CNT concentration as low as 0.1 wt %.45 However, no evidence for gel formation was reported so far. A more relevant example is provided by chitosan-CNT materials that have been prepared in many forms, such as films and macroporous scaffolds.46, 47 In the latter systems, the electrical conductivity of chitosan can be increased by two orders of magnitude through the introduction of 2.5 wt % CNT (Fig. 4).46 Further improvement of the conductivity is obtained if the porous scaffolds are compressed. This evidences the influence of CNTs packing, and therefore, connectivity, on the electrical behavior of these materials.
The possibility to obtain a biocompatible material with good conductivity properties allows the design of electrochemical biosensors. For example, it was possible to deposit thick chitosan-CNT films on glassy electrodes and to perform the direct electrochemical detection of NADH, with a decrease of about 0.25 V of the overpotential of NADH oxidation in the presence of the nanotubes.48 In a step further, glucose deshydrogenase enzymes were covalently immobilized within the chitosan-CNT films and the resulting bioelectrode was used for the detection of glucose oxidation. To close this section, it must be noticed that chitosan-CNT composites were also studied for cell immobilization, although none of these reports took advantage of the conductivity properties of the material to interact with cellular activity.49
MAGNETIC PROPERTIES OF IRON OXIDE-BASED BIONANOCOMPOSITES
The potentialities of magnetic nanoparticles in the biomedical field are now well-established.50In vivo, these applications take advantage of three properties of these magnetic colloids: (i) their field-induced mobility, that can be used for targeted drug delivery;51 (ii) their ability to modify magnetic relaxation times of surrounding molecules, with major application in magnetic resonance imaging;52 and (iii) their heating under an alternative magnetic field, the basis for hyperthermia treatment of tumors,53 as well as for magnetically controlled drug release.54 Combination of several of these properties at the nanoscale to obtain so-called theranostic platforms is a major research topic.55
In most of these applications, incorporation of the magnetic particles within a solid matrix is not adapted because high mobility and/or easy access to the colloidal surface are required. Therefore, biopolymers are only used as coating layers to enhance stealthiness or provide specific recognition properties.50 The only domain where magnetic hydrogels are of large interest is the design of magnetically controlled drug release systems.56 These “smart” devices consist of magnetic colloids entrapped within thermoresponsive polymers containing drugs. Upon application of an alternative magnetic field, heat is dissipated that locally increases temperature, inducing a structural transition in the polymer network from a rigid to a more flexible state, allowing easier diffusion and release of the drug.
In such systems, the properties of both the magnetic particles and the polymer network have to be considered and matched. In the case of magnetic colloids, the relevant parameter is the specific loss power (SLP), defined as eq 3:55
with where Ms is the saturation magnetization, μ0 the vacuum permeability, H is the magnetic field strength, L(ξ) is the Langevin function, ω is the angular frequency, τ is the relaxation time, K is the magnetic anisotropy constant, k is the Boltzmann constant, V is the particle volume, and η is the viscosity of the medium.
As shown in Figure 5, the SLP has some maxima at intermediate particle size range (5–20 nm) and anisotropy, and increases with magnetization.55 In principle, many metal-based nanoparticles exhibit suitable magnetic properties. However, for toxicity reasons, iron oxide particles, either magnetite Fe3O4 or maghemite Fe2O3, are used in these systems. As far as the polymer is concerned, one should consider the gelation temperature (Tg gel-to-sol transition) or lower critical solution temperature (LCST) corresponding to the transition from a mainly hydrophilic to mainly hydrophobic polymer conformation, resulting in water and drug expulsion. The typical example of synthetic thermosensitive polymers is N-isopropylacrylamide (p-NIPAM);58 however, it exhibits an LCST of 32 °C so that it is in the fluid state at body temperature. On this basis, p-NIPAM-based co-polymers as well as alternative polymer backbones have been developed, allowing a fine-tuning of the LCST values.59
In contrast, the gelation temperature of biomacromolecular networks is an intrinsic characteristic that is more difficult to control. For instance, gelatin from mammalian origin has a Tg close to 37 °C that would make it very sensitive to even a slight increase in temperature in vivo. To overcome this problem, gelatin hydrogels were crosslinked with genipin57 or modified with polyacrylamide60 to enhance their thermal stability. The incorporation of iron oxide colloids was performed by an in situ precipitation method where the hydrogels are first impregnated with iron salts, followed by pH modification. In these two systems, application of a magnetic field significantly increases the amount of drug released [doxorubicin, a well-known anticancer molecule].57, 60 Magnetic biopolymer-pNiPAAm hydrogels were also developed using alginate and chitosan.61–63 The idea behind this work is twofold: (1) to tune the LCST and swelling/deswelling behavior of pNiPAAm and (2) to favor the homogeneous in situ growth of iron oxide particles in the biopolymer network. By doing so, the magnetic properties of the hydrogels, and therefore the SLP values, can also be tuned.
Interestingly, application of a static magnetic field on gelatin-iron oxide sponges (called ferrosponges) demonstrated the opposite effect, that is, decrease of drug release rate.64 In this example, it was found that the average pore size of the network decreased after field application, and that this effect was more pronounced with increasing iron oxide content. This was attributed to interparticle interactions that bring magnetic colloids closer in the presence of the magnetic field. As a result, drug diffusion is limited either due to the presence of iron oxide particle aggregates and/or due to the shrinkage of the gelatin network.
To end this section, it is worth noting that magnetic scaffolds may have some interest in tissue engineering. On the one hand, collagen-iron oxide scaffolds were recently described, exhibiting good cytocompatibility, that can be useful to study the influence of magnetic fields on cell behavior.65 On the other hand, it was shown that cells grown within magnetic hydrogels become themselves magnetic by membrane binding and/or internalization of the iron oxide particles.66 In these conditions, it is possible to induce cell levitation, providing a new and original environment to study cell behavior.
MECHANICAL PROPERTIES OF SILICATE-BASED BIONANOCOMPOSITES
First investigations of nanocomposites mainly focused on the design of materials with improved mechanical properties. As recently reviewed by Ariga et al.,67 interesting strategies have been developed to build up devices responsive to mechanical and chemomechanical stimuli, especially involving enzymatic-mediated processes.68 In the scope of the present review, we will focus on the properties of the resulting composites. Bionanocomposites involving biopolymers (polysaccharides and proteins) have found widespread applications as biomedical materials, notably in the field of regenerative medicine, and tissue engineering.69 In materials sciences, many groups attempt to prepare bone-like materials from bioorganic (collagen and gelatin) and mineral (hydroxyapatite) constituents.70 As reported by Liu et al., the fabrication of 3D gelatin/apatite particles scaffolds was found to be an effective method to build up composites exhibiting enhanced mechanical strength. Interestingly, the particle number and size in the scaffold could be controlled by the incubation time and ionic concentration to reach compressive modulus more than 75% higher than that of the starting gelatin scaffold [Fig. 6].71 Very importantly, osteoblastic cell differentiation could also be enhanced. This key aspect will be developed in “Bioactivity of Hydroxyapatite-Based Bionanocomposites” section.
Besides these examples, silica and clays have also been used in combination with biopolymers to provide the network with improved thermal, mechanical, and chemical stability. This is attributed to conformation changes taking place at the particle/biopolymer interface, where new interactions may occur, for example, formation of hydrogen bonds between hydroxyl groups of macromolecules and silanols. The synergetic effects in silicate/hydrogels systems have been investigated using different polysaccharides such as chitosan,73, 74 carrageenans,75 cyclodextrine,76 and alginate.77 In particular, since the end of the 1980s,78 silica/alginate composites have been designed with the aim of improving the stability and mechanical properties of crosslinked alginate networks. This has been notably achieved by the composite approach, with the incorporation of colloidal silica within networks. In this context, Boguń and coworkers reported that the addition of nanoadditives such as silica nanoparticles (with sizes ranging from 50 to 100 nm) within calcium- or sodium-alginate fibers, ended up with composite materials exhibiting improved thermal stability compared with pure alginate fibers. This was characterized by significantly higher T50 values (i.e., temperatures at which a 50 wt % loss is obtained).79, 80 These results show that the composite strategy successfully allows designing materials that associate the biocompatibility of alginate biopolymers and the mechanical and thermal stability of an inorganic phase.
This approach has also been used to improve the mechanical properties of protein- and peptide-based nanocomposites, by the incorporation of clays. In these systems, the process of clay exfoliation is a key step for the homogeneous dispersion of the mineral phase at the nanoscale in the biopolymer network.81 In particular, we will focus in the following section on gelatin-based systems.82 Recent examples have been reported by Fernandes et al., where the dispersion of sepiolite nanoclay within the gelatin matrix up to 50 wt % induces an increase in the Young's modulus by a factor of 2.5 as compared with the pure gelatin materials due to the outstanding homogeneity of the composite materials. This illustrates the good dispersion of the clay within the biopolymer matrix even at such high clay content, where the sepiolite fibers show preferential organization within the gelatin matrix. Although absolute knowledge of the interactions between sepiolite and gelatin could not be established, electrostatic interactions take place together with hydrogen bonding interactions between silanol hydroxyl at the surface of exfoliated sepiolite and CO and NH groups of gelatin. At higher pH values, the electrostatic interactions between negatively charged silicates and gelatin lead to the precipitation of the inorganic phase.83 Alternatively, porous gelatin-sepiolite nanocomposites were built up via a freeze drying process of a gelled hybrid.84 The resulting nanocomposite foams exhibited improved mechanical properties with increasing sepiolite nanoclay content up to 9.1 wt %. At this content, the average values for the Young's modulus and compressive collapse stress of the gelatin foam increased from 1554 to 6031 kPa and from 80 to 326 kPa, respectively. This enhancement was attributed to hydrogen bonds between gelatin and sepiolite as suggested by infrared analysis and differential scanning calorimetry.
Very interestingly, Frydrych et al. showed in this work that the mechanical properties of their gelatin-sepiolite nanocomposites could be well predicted using the Gibson-Ashby model.84 The mechanical and thermal properties of foams are directly related to two main parameters: the density, which also dictates the porosity of the composite, and the size of the cell, that is, the internal microstructure.85 The relative Young's modulus (Ef/Es) and the relative density (ρf/ρs) are related according to eq 4, where f and s refer to foam and solid, respectively:
In eq 4, C1 and n are geometric constants that depend on the microstructure of the solid (geometry, curvature, and porosity). According to Gibson and Ashby, the normalized coefficient C1 can be calculated by eq 5:
where the superscript p refers to the polymer.
Over all the models tested in this work, the Gibson-Ashby model showed the best approximations in the whole range of relative densities studied, supporting recent results on the use of the open-cell model for the analysis of low-density polymer-clay nanocomposite foams.86
From these studies, it is clear that the incorporation of nanoclay in biopolymers successfully improve the mechanical properties of nanocomposites, raising the question of their bioactivity. Although this key aspect will be developed in details in the coming section, recent works have highlighted the good biocompatibility and low-toxicity of sepiolite clay, opening the way for the design of biomedical devices with tunable mechanical properties.87
BIOACTIVITY OF HYDROXYAPATITE-BASED BIONANOCOMPOSITES
Bones exhibit a very complex composite structure whose formation is driven by the interplay between osteogenic cells, collagen, and hydroxyapatite.88 Briefly, osteoblast cells produce a collagen network with a very high degree of organization, similar to liquid crystal phases, that serves as a matrix for hydroxyapatite nanocrystal formation. Going into more details, it is well-known that cells also control the mineralization process, whereas, in parallel, the mechanical properties of the collagen/apatite composite influence the cellular activity.
It was soon hypothesized that synthetic collagen/hydroxyapatite composites that can be used as bone repair materials must mimic bone tissue as close as possible.70 Considering the collagen matrix, early studies involved either gelatin,89 that is, hydrolyzed collagen, or collagen solutions at low concentration.90 More recently, concentrated collagen hydrogels have also been used to engineer biomimetic matrices.91 In parallel, the mineralization step can be performed by addition of preformed hydroxyapatite nanoparticles,92 or, more often, by in situ precipitation methods, either by addition of phosphate salts to calcium-containing collagen hydrogels90 or calcium salts addition to phosphate-containing hydrogels.93 However, these approaches do not replicate the in vivo mineralization process where calcium and phosphate are present in the body fluid at concentrations above supersaturation, instantaneous precipitation being prevented by organic stabilizing agents. On this basis, impregnation techniques were recently developed, using undersaturated solutions (known as simulated body fluid)94 or supersaturated solutions in the presence of poly-aspartic acid as a calcium stabilizer.95
Not surprisingly, such a variety of synthetic processes leads to many different structures (level of organization, particle size and mineralization rate) making the various studies difficult to compare. This is particularly true when mechanical properties are considered, especially because measurement techniques may also vary. Typically, elastic moduli or compressive stiffness were in the 1 kPa to 100 MPa range with hydroxyapatite content in the 20–40 wt %.70, 92, 93 The highest reported modulus for hydroxyapatite-collagen composite is about 6 GPa (to be compared with ca. 20 GPa for lamellar bone tissue), as measured by nanoidentation techniques, with only 6 wt % of mineral phase.96 However, in this work, a highly concentrated and organized collagen matrix was used with an initial modulus of about 4 GPa, demonstrating the primary importance of the matrix itself. In parallel, recent calculations suggested that the optimal hydroxyapatite particle size for collagen gel strengthening is about 2 nm, in agreement with observations of the mineral phase of the bone tissue.97
Several in vitro evaluation of hydroxyapatite-collagen bionanocomposites were conducted via the seeding of osteoblastic cells.70, 93, 94 In most cases, the presence of the mineral phase does not significantly influence the cell adhesion and proliferation processes. The cell-mediated induction of hydroxyapatite formation at the material surface (bioactivity) is also reported but, in many cases, no control experiment with nonmineralized collagen materials is available.70 This remark also applies to in vivo data that demonstrate suitable integration and new bone formation of these materials in bone defects (Fig. 7),70, 92, 98, 99 but information about the pure collagen materials in similar conditions are often lacking. Nevertheless, it is now admitted that hydroxyapatite-collagen bionanocomposites are optimal intermediate systems between hydroxyapatite materials that exhibit high bioactivity but too slow degradation/remodeling rates and collagen networks with fast degradation but low bioactivity. Noticeably, the presence of the protein network also allows the easy shaping of the materials as films, particles or fibers in mild conditions, whereas hydroxyapatite, as a ceramic material, requires more complex processing.
Despite these achievements, it must be noticed that, at this time, a large majority of clinical interventions for bone repair involve bone auto- or allo-graft rather than synthetic materials.100 One of the many advantages of bone grafts is that these systems contain biological molecules (in allografts) or even osteogenic cells (in autografts) that favor integration and remodeling. On this basis, recent efforts have been made to incorporate cellular growth factors and/or cells within the bionanocomposites before implantation.101, 102 This induces further constraints on the materials structure, especially in terms of porosity, with significant impact on their mechanical properties. In this context, it was recently proposed to add silica to the hydroxyapatite-collagen matrices leading to the improvement of the mechanical properties without negatively impacting on its bioactivity, at least in vitro.103, 104
This overview enlightens the diversity of compositions, structures, and properties that can be achieved by combining biopolymers and inorganic colloids within hydrogels. Very interestingly, these associations may be beneficial for both partners interacting in a synergetic manner. On one hand, the biopolymer network controls the size and organization of the functional nanoparticles and provides the resulting composite with ability to interact with living systems thanks to their biocompatibility and specific molecular recognition properties. On the other hand, the incorporation of inorganic colloids improve the (bio)-chemical, thermal, and mechanical properties of the hydrogel and add up tunable functions. Indeed, the control of the bio-mineral interface, from the influence of nanoparticles on the aggregation/self-assembly process of biological molecules to the role of hydrogel composition and density on particle size, packing, and mobility, is a key factor to optimize the composite properties and relies on a fine tuning of the chemical interactions (electrostatic forces and hydrogen bonding) between the two phases.
In many occasions, the benefits of the bionanocomposite approach were illustrated in the context of biomedical applications. This is because biocompatibility, in the accepted sense of successful in vivo integration by the human body, is the key criteria on which natural polymers are currently preferred to synthetic ones. It is, therefore, very likely that biomaterials and pharmaceutical devices will remain major fields of application of bionanocomposites. However, the recent concern for the environmental impact of materials, from design to recycling, also motivated a huge interest for “green” plastics (or bioplastics) that could be used, for instance, in packaging.105 Here, the targeted property is biodegradability or, at least, bioresorption. In these applications, artificial biopolymers formed from natural monomers, such as polylactic acid, are commonly used because their properties can be controlled during the synthetic polymerization process. As this is not possible for natural “preformed” biopolymers, additional chemical, biochemical, or physical methods are required to reach suitable properties. We believe that the nanocomposite approach will be particularly fruitful for this purpose.
Finally, it is also possible to take advantage of other specific properties of biomacromolecules such as self-assembly, specific recognition, or biocatalysis to build bio-responsive hydrogels, for instance for medical diagnostic applications.106 However, a key challenge in this area is to build-up three-dimensional networks from polymers that do not exhibit gel formation ability. Several recent examples suggest that this may be achieved by a bionanocomposite approach where inorganic colloids act as nucleation centers for the growth of biopolymer networks.107, 108 Indeed, such developments must be supported by more fundamental studies aiming at understanding the specificities of the bio-mineral interfaces at the nanoscale.
The authors thank the ANR-09-RPDOC-023-01 program and the CNRS for supporting their past and current research in the field of bionanocomposites. T. Coradin acknowledges the contribution of R. Brayner, F. Carn (Université Paris VII, France), and M. F. Desimone (Universidad de Buenos Aires, Argentina) to some of the works presented in this review.
Carole Aimé is a CNRS researcher working in Thibaud Coradin's group in the Laboratoire de Chimie de la Matière Condensée de Paris. After a PhD in Reiko Oda's group in Bordeaux University-France, working on self-assembling amphiphilic systems, she joined Pr. Nobuo Kimizuka's group in Kyushu University-Japan, where she designed functional coordination nanoparticles from nucleotides and lanthanide ions. She is now developing bio-inspired systems made up of inorganic nanoparticles and biopolymers.
Thibaud Coradin, born in 1970, is Directeur de Recherche at the CNRS since 2007. He is currently leading the «Materials and Biology» group in the Laboratoire de Chimie de la Matière Condensée de Paris (UPMC-Paris 06). His research topics include biomineralization, bionanocomposites, biomaterials, bioencapsulation, and green materials chemistry. He co-authored over 110 publications and 14 book chapters. He is a member of the Advisory Editorial Board of Current Medicinal Chemistry and Silicon.