Direct ″Click″ Synthesis of Hybrid Bisphosphonate–Hyaluronic Acid

Apr 23, 2012 - Direct ″Click″ Synthesis of Hybrid Bisphosphonate–Hyaluronic Acid Hydrogel in Aqueous Solution for Biomineralization. Xia Yang†...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/cm

Direct ″Click″ Synthesis of Hybrid Bisphosphonate−Hyaluronic Acid Hydrogel in Aqueous Solution for Biomineralization Xia Yang,† Sultan Akhtar,‡ Stefano Rubino,‡ Klaus Leifer,‡ Jöns Hilborn,† and Dmitri Ossipov*,† †

Department of Materials Chemistry and ‡Department of Engineering Sciences, Uppsala University, Uppsala, SE 751 21, Sweden S Supporting Information *

ABSTRACT: We report the synthesis of injectable in situ forming hybrid hydrogel material and investigate its ability to support the mineralization process under mild conditions. To achieve this, we have prepared a hyaluronic acid (HA) derivative that is dually functionalized with cross-linkable hydrazide groups and bisphosphonate ligands (HA-hy-BP). The hybrid hydrogel can be formed by simple mixing of two solutions: the solution of HA-hy-BP and the Ca2+ ions containing solution of aldehyde-derivatized HA (HA-al). We found that the conjugation of BP, a P−C−P analogue of pyrophosphate, to the hydrogel matrix promotes an efficient and fast mineralization of the matrix. The mineralization is facilitated by the strong interaction between BP residues and Ca2+ ions that serve as nanometer-sized nucleation points for further calcium phosphate deposition within the HA hydrogel. Compared with previously reported hydrogel template-driven mineralization techniques, the present approach is maximally adapted for clinical settings since the formation of the hybrid takes place during quick mixing of the sterilized solutions. Moreover, the hybrid hydrogel is formed from in vivo degradable components of the extracellular matrix and therefore can be remodeled in vivo through concerted HA degradation and calcium phosphate mineralization. KEYWORDS: hybrid organic−inorganic materials, injectable materials, hydrogels, hyaluronic acid, mineralization



phosphate and phosphoester-containing polymers4−6 and hydrogels7,8 in the simulated body fluid9 or in an acidic hydroxyapatite solution followed by an increase of pH.10,11 Incorporation of the hydrogel interface groups with strong binding capacity to the growing inorganic NPs as well as adjusting the concentration of these groups have been demonstrated to provide specific control over the size, morphology, and long-term stability of the integrated NPs.12 In another approach, a hydrogel was grafted with groups that can act as reagents themselves in the formation of inorganic NPs which provided metal−hydrogel nanocomposite materials by single placing the hydrogel in contact with the metal ion precursor.13 To our knowledge, there were no reports on hybrid hydrogels with covalent linkages between polymer chains and the dispersed inorganic NPs that can be prepared from a solution phase just in one pot under the conditions that are compatible with biological molecules, cells, and tissues. Here, we define hybrid materials as those that contain a chemical linkage between the organic and inorganic parts of the material. Bisphosphonates (BPs) are the analogues of pyrophosphate in which the bridging oxygen is substituted by carbon.14 Structurally, the P−C−P bridge of BPs can be viewed as a smallest hybrid motif linking two inorganic phosphates with an

INTRODUCTION Calcium phosphate mineralization is a key process during bone regeneration that takes place in vivo within a three-dimensional scaffold of extracellular matrix (ECM) proteins that are rich with glutamate, aspartate, and phosphoserine amino acids.1 These anionic amino acid residues serve as nucleation points for gradual deposition of the calcium phosphate inorganic phase that becomes integrated into the developing bone on a nanoscale. In bone tissue engineering, the design of polymeric materials that support biomineralization is essential to encourage the ingrowth of surrounding bone.2 Recreating natural mineralization processes within a molecularly designed hydrogel template has the potential to control shape, size, and distribution of the inorganic crystals in the in situ formed hybrid.3 Biomimetic synthesis of organic−inorganic hybrid hydrogels utilizes the initial preparation of a hydrogel template with tailored chemical functionalities. These functionalities in a form of the hydrogel side groups have affinity to the growing inorganic nanoparticles (NPs) that are subsequently formed within the three-dimensional polymer network. To perform the hydrogel-templated inorganic reaction between the inorganic ionic precursors, the hydrogel has to be swollen in an aqueous solution of the first precursor followed by providing the transport of the second precursor ions into the same hydrogel. For example, hybridization of a hydrogel with a calcium phosphate nanophase was achieved through immersion of the © 2012 American Chemical Society

Received: January 27, 2012 Revised: April 17, 2012 Published: April 23, 2012 1690

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

aldehyde derivatives of HA. We reasoned that by simultaneously exposing the HA-hy-BP derivative to both organic HAal and inorganic Ca2+ reactants one can directly obtain the hybrid organic−inorganic hydrogel since the Brownian motion of the HA-hy-BP chains will be sufficiently restricted by rapid cross-linking reaction. In this work, we showed the feasibility of this technique that represents simultaneous application of orthogonal organic and inorganic reactions for in situ preparation of the hybrid hydrogel material. Finally, we have shown that the initial Ca2+•BP−HA hybrid hydrogel can be further mineralized through recruitment of endogenous ionic precursors (calcium and phosphates). These results provide a framework for the development of injectable hybrid biomaterials that may be useful in tissue engineering and drug delivery applications.

organic molecule. Along with the hydrolytic stability of BPs, the tetravalence of the bridging carbon provides synthetic possibilities to link BPs to other low molecular weight molecules and macromolecular carriers. BPs have exceptionally high affinity to calcium ions as well as to the bone mineral hydroxyapatite (HAP).15 Due to the properties of BPs such as inhibitory effects on osteoclasts and chemisorptions to bone mineral, the most common clinical use of BPs is the treatment of osteoporosis and osteolytic bone diseases (Paget’s disease and hypercalcemia).16 BPs have also been suggested as an adjuvant to anticancer agents for treatment of bone metastasis.17 Despite the excellent binding capacity of BPs to divalent metal ions, such as Ca2+ and Mg2+, only a few works have explored grafting polymers with BPs for biomimetic mineralization11,18 or integration into the preformed inorganic phase.19 In this work, we found that interaction between BPs and Ca2+ leads to a precipitation reaction in the same way as insoluble calcium phosphate is formed upon mixing of the phosphate and calcium ions. We hypothesized that conjugation of BP residues to a hydrogel matrix followed by diffusion of the calcium ions into the matrix would prevent bulk precipitation of the calcium bisphosphonate salt. Instead, one can expect formation of the matrix-linked three-dimensional structures in which several BP ligands become clustered via coordination to the metal ions. Our second assumption was that by providing fast network formation one can combine the organic networking with an inorganic Ca2+•BP coprecipitation reaction in one pot. In this case, the in situ forming hydrogel matrix will act both as a template and as a reactant affording an injectable hybrid hydrogel material with homogeneously dispersed inorganic nanoparticles. Hyaluronic acid (HA) is the main nonsulfated glycosaminoglycan of the ECM that regulates diverse cellular responses. Our group has recently demonstrated the HA hydrogelmediated delivery of bone morphogenetic protein-2 (BMP-2) to induce local bone formation subcutaneously 20 and subperiosteally21 as well as in combination with HAP nanoparticles for healing of cranial bone defects.22 Using our recently developed derivatization procedure,23 we modified HA with both hydrazide groups and pamidronate moieties that should have higher binding affinity to Ca2+ ions than the aminobisphosphonate without hydroxyl group side chain.17,24 This dually functionalized HA-hydrazide-BP polymer (HA-hyBP) can be cross-linked with aldehyde-derivatized HA (HA-al) through a hydrazone chemistry. This is a “click” type reaction since it is chemoselective, can be performed in aqueous solution under physiologic conditions, and forms no toxic side products. We investigated the interaction of BP groups of the hydrazone matrix with Ca2+ ions and found that diffusion of Ca2+ ions into the BP-functionalized hydrogel provides formation of Ca2+•BP clusters within the matrix. In contrast, diffusion of Ca2+ ions into the same hydrazone hydrogel lacking the BP groups did not result in the formation of hybrid material as was judged from scanning electron microscopy (SEM) and cryo-SEM analysis as well as from dynamic light scattering (DLS) measurements of the enzymatically degraded hydrogels. This observation provided a new insight into the role of different anionic groups, such as carboxylates and phosphates (phosphonates), in the mineralization process. Another important feature of the hydrazone cross-linking reaction is its relatively fast rate permitting formation of a self-standing hydrogel network within 10−30 s after mixing of the fluid solutions of the complementary reactive hydrazide and



RESULTS AND DISCUSSION Synthesis of Conventional Hydrogels and Study of Their Interaction with Calcium Ions. For the preparation of a suitable HA derivative that is grafted with cross-linkable hydrazide groups and bisphosphonate ligands, we employed our previously designed simultaneous one-pot modification of HA with orthogonal hydrazide and thiol groups.23,25 Specifically, a carbodiimide-mediated amide coupling of HA with cross-linkers 1 and 2, followed by treatment of the reaction mixture with dithiothreitol (DTT), generated HA side chains terminated with thiol and hydrazide groups, respectively (Scheme 1A). Addition of the acryloylated pamidronate derivative 311 to the reaction mixture allowed in situ grafting of the BP ligands. Michael addition of HA thiol groups to the Scheme 1. (A) Functionalization of HA with Hydrazide (Red Color) and Bisphosphonate Groups and (B) Preparation of an Organic−Inorganic Hybrid Hydrogel via a Conventional Organic Matrix-Templated Approach and by Direct Simultaneous Formation of the Organic Network and the Nanosized Ca2+•BP Clusters

1691

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

double bond of 3 was confirmed by 1H and 31P NMR spectroscopy (see Figure S1 in Supporting Information). An analogue of HA-hy-BP lacking the BP groups, named HA-hy, has been prepared as well with the amount of hydrazide groups corresponding to their amount in HA-hy-BP (10% of HA disaccharide repeating units). The synthesis of HA-hy was described by us previously.25 The ability of both derivatives, HA-hy and HA-hy-BP, to form conventional hydrogels that consist of randomly crosslinked macromolecules was first examined by mixing pure aqueous solution of the hydrazide-derivatized HA with the aqueous solution of aldehyde-derivatized HA (HA-al in Scheme 1B). HA-al has been synthesized according to our previous protocol.25 The control hydrogel and the one with covalently immobilized BP groups (Scheme 1B, step 1) were formed at 2 wt % concentration within 30 s as was judged from a test tube inversion assay. The presence of covalently linked BP groups in the network of randomly cross-linked HA chains should make this hydrogel interactive with calcium ions since the mixing of an aqueous solution of 3 with a CaCl2 solution led to the formation of a white precipitate. The same type of inorganic reaction should principally take place upon transport of Ca2+ ions into the HA−BP hydrogel with the only difference being that the molecular mobility of the BP groups is substantially restricted in the hydrogel by covalent linkages to the 3D network. Nevertheless, molecular segmental mobility of the polymer chains in the hydrogel may still be enough to allow clustering of several BP groups through interaction with Ca2+ ions. As a result, structural reorganization of the network with the formation of nanosized substructures composed of the condensed HA chains is expected in the case of the HA−BP hydrogel (Scheme 1B, step 2). Immersion of the hydrogels in a 0.2 M CaCl2 solution (pH 7.2) results in drastically different reactions that can be clearly seen by the naked eye already during the first minutes of diffusion of the Ca2+ ions into the hydrogels (see video in Supporting Information). The BP-immobilized hydrogel becomes gradually opaque, while the control hydrogel stays completely transparent after 24 h of incubation in the CaCl2 solution (Supporting Information, Figure S2A). The Ca2+-containing hydrogels are characterized by a different swelling ratio and elastic modulus. Thus, the BP-immobilized hydrogel has a lower swelling ratio and higher G′ values as compared with the control hydrogel after swelling in the CaCl2 solution (Supporting Information, Table S1). This observation reflects higher cross-linking density of the former hydrogel as a result of strong Ca2+•BP interactions. These interactions prevent water to be accommodated in such stiffened hydrogel material. To prove this assumption, both hydrogels were thoroughly washed with water by replacing the swelling medium with a fresh one at least three times. It has been done to remove all unbound Ca2+ ions from the hydrogel samples. After washing, the hydrogels became softer but retained the same differences in mechanical properties that were observed for the hydrogels before washing (Supporting Information, Figure S2B). To prove the formation of nanoparticles in the HA−BP hydrogel after interaction with Ca2+ ions, the washed hydrogel samples were freeze-dried and analyzed by SEM. The obtained images of the BP-linked hydrogel showed white particular inclusions homogeneously distributed throughout the sample volume, whereas the surface of the control gel was plain (Figure 1). The size of the integrated particles was in the range of 200− 500 nm. It should be noted that no integrated nanoparticles

Figure 1. SEM images of (A) (HA-hy-BP + HA-al) and (B) (HA-hy + HA-al) hydrogels after immersion in a 0.2 M CaCl2 solution followed by repeated immersions in pure water.

were also detected in the case of BP-immobilized hydrogel before interaction with Ca2+ ions (Supporting Information, Figure S3), which indicates that the observed nanoparticulate structures can only be formed when both BP and Ca2+ are engaged in the synthesis procedure. These results were further supported by energy-dispersive X-ray spectroscopy (EDS) which revealed the presence of both calcium and phosphorus only in the hydrogel derived from the HA-hy-BP derivative (Supporting Information, Figure S4). We also performed cryoSEM analyses of the samples to confirm that the hybrid hydrogel nanostructure is not perturbed during the freezedrying step (Supporting Information, Figure S4). In parallel, the hydrogels were subjected to degradation with a HA specific enzyme, hyaluronidase (Hase). Hase cleaves HA macromolecules along the glycosidic bonds and should bypass the modified segments of the HA hydrogel. The degradation products were analyzed by DLS. As expected, the control hydrogels were degraded down to the ideal solution of HA oligomers that did not scatter the incident light. On the contrary, degradation of the hybrid Ca2+•BP−HA hydrogel resulted in the formation of an aqueous dispersion of submicrometer particles with an average size of 475 ± 93 nm (Figure 2).

Figure 2. DLS measurements of the digest mixtures obtained after enzymatic degradation of the calcified and washed (A) (HA-hy-BP + HA-al) and (B) (HA-hy + HA-al) hydrogels. (C) Schematic representations of the hyaluronidase-mediated degradation processes of the hydrogels corresponding to each DLS plot (A and B) are shown on the right side (top panel and bottom panel of C), respectively.

In Situ Formation of the Hybrid Ca2+•BP−HA Hydrogel. Hybridization of the HA−BP hydrogel with Ca2+ ions follows the steps 1→2 in Scheme 1B. Step 1 of this pathway is necessary to obtain a 3D reactive framework for the subsequent reaction with metal ions (step 2). The kinetics of hydrazone network formation is, however, sufficiently fast (≈30 s) and may be accomplished prior to the Ca2+•BP coprecipitation reaction even when the transport of Ca2+ ions is provided simultaneously with the transport of the HA−aldehyde 1692

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

reactant. Such one-pot synthesis can be envisioned if the solution of HA-hy-BP is admixed to the solution containing both HA-al and Ca2+ ions (Scheme 1B, reaction 3). To test this assumption, we prepared the solutions of hydrazide components, HA-hy-BP or HA-hy, in the PBS buffer lacking Ca2+ ions, while the HA-al derivative was dissolved in the commercial PBS buffer containing Ca2+ and Mg2+ ions. The solutions were vortexed giving hydrogels almost instantaneously when using HA-hy-BP and after around half a minute in the case of the HAhy component. Analogously to the immersion of the control hydrogel (i.e., the hydrogel lacking BP groups) in the CaCl2 solution, no visible mineral precipitation took place after mixing the buffered solution of HA-hy with the Ca2+-containing buffered solution of HA-al. On the contrary, in situ formation of the BP-linked hydrogel was simultaneously accompanied by the appearance of a white precipitate phase that was homogeneously dispersed within the formed organic matrix. The nanostructure of the in situ formed hybrid Ca2+•BP−HA hydrogel was confirmed by cryo-SEM of the wet hydrogel samples as well as by conventional SEM analysis of the same samples after their subsequent freeze-drying (discussed in the following section). Biomimetic Mineralization of the Hybrid Ca2+•BP−HA Hydrogel. In the field of bone tissue engineering, there is significant evidence for the importance of the synthetic polymer scaffolds that support biomineralization and ultimately osteoinduction and osseointegration.26 Compared to the previous studies performed with phosphorus-containing materials,8,13 our approach of ex vivo in situ formation of the hybrid premineralized Ca2+•BP−HA hydrogel scaffold is advantageous since it is less dependent on contact with body fluid. It may, nevertheless, support further biomineralization after implantation in vivo. We simulated this process in vitro by placing the in situ formed hybrid Ca2+•BP−HA and the control hydrogels in Ca2+-containing PBS buffer and examining the samples over time. After mineralization, an opaque hydrogel was obtained with BP-immobilized HA, while the control hydrogel without BP residues was transparent (Figure 3).

Figure 4. Swelling ratios (A) and elastic moduli (B) of the in situ formed HA−BP hybrid (black bars) and control hydrogels (gray bars) before mineralization (day 0), after 1 day of mineralization (day 1), and after 7 days of mineralization in a Ca2+/Mg2+-containing PBS buffer (day 7). The hydrogel samples were repeatedly washed with water before the measurements.

observed throughout the 7 days of mineralization. However, mineralization of the control hydrogel increased its strength after the first day of incubation in the medium, followed by the decrease of the G′ when the hydrogel was incubated for 7 days. This can be explained by taking into account possible hydrolytic degradation of the hydrazone cross-linked gels upon incubation in the mineralization medium. Figure 4 presents data for the hydrogel samples that have been washed with water three times. We have also performed determination of the G′ value for the mineralized hydrogels before washing with water. Interestingly, the washing procedure led to an increase of the G′ for the HA−BP hydrogels, whereas the control hydrogels became softer after the washing. For example, the HA−BP and the control hydrogels had elastic moduli of 865 ± 87 and 703 ± 71 Pa, respectively, when measured just after 7 days of mineralization. However, after the subsequent washing, the G′ values have changed to 1920 ± 127 Pa for the HA−BP hydrogel and to 538 ± 40 Pa for the control one. This is expected since phosphate ions (HPO42− and H2PO4−) of the PBS buffer can compete for the BP-bound calcium ions, shifting the equilibrium in the hybrid HA− BP•Ca2+ hydrogel to the unbound HA−BP: HA−BP•Ca2+ + HPO42−/H2PO4− ↔ HA−BP + calcium phosphate↓. This may weaken the interactions on the organic−inorganic interfaces of the hybrid hydrogel. The subsequent hydrogel washing eliminates phosphate ions from the equilibrium and restores BP•Ca2+ interactions, thus making the network stronger. When the control hydrogel is, however, swollen in the Ca2+containing PBS buffer, calcium ions can only bind to the carboxylate groups of the HA network. These interactions are much weaker than the Ca2+•phosphate/phosphonate interactions, and therefore aqueous washing of the control hydrogel removes most of the counter-cations from the network. This leads to the repulsion of the carboxylate groups and, as a result, to higher swelling in pure water. These data corroborate with the fact that the immersion of the BP-linked and the control hydrogels in 0.2 M CaCl2 solution followed by their washing with water resulted in the softening of both types of gels (Supporting Information, Table S1). In that case, there was no competitive interaction between phosphate and calcium ions, and the Ca2+ ions were bound to either grafted BP groups or to the HA carboxylate groups of the corresponding hydrogels before their washing. SEM analysis of the freeze-dried hydrogels confirmed our hypothesis that mixing the HA-hy-BP/PBS and HA-al/(Ca2+/ Mg2+-PBS) solutions results in the formation of a hybrid hydrogel (Figure 5A) with evenly distributed particles that are

Figure 3. Bottom panel: top view of the in situ formed control (HA-hy + HA-al, left) and hybrid (HA-hy-BP + HA-al, right) hydrogels after 7 days of mineralization in a Ca2+/Mg2+-containing PBS buffer. Top panel: side view of the same hydrogel samples.

The mechanical properties of the hydrogels were very much affected by the presence of covalently bound BP groups (Figure 4). Generally, the HA−BP hydrogels were mechanically stronger than the control hydrogels lacking the BP groups. Incubation in the Ca2+/Mg2+-containing PBS buffer led to a strengthening of the BP-linked hydrogel, for which the increase of the elastic moduli and the decrease of the swelling ratios was 1693

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

Figure 5. SEM images of the hybrid HA−BP hydrogel before mineralization (A), after 1 day of mineralization (B), and after 7 days of mineralization (C). SEM images of the control hydrogel after 7 days of mineralization in a Ca2+/Mg2+-containing PBS buffer (D). Cryo-SEM images of thin slices of the corresponding wet hydrogel samples are shown above the SEM images. (E) Concentration of calcium in the nonmineralized hydrogels (day 0) as well as in the hydrogels that have been incubated in the mineralization medium for different intervals of time (day 1 and day 7). All the hydrogel samples were repeatedly washed with water before analysis.

around 52 ± 17 nm in diameter. These particles can act as nucleation points during the subsequent mineralization procedure when the in situ formed hybrid hydrogel is incubated in the Ca2+-containing PBS buffer. The particles grew steadily during incubation in the buffer, and their mean diameter reached 361 ± 59 nm and 476 ± 93 nm after 1 day (Figure 5B) and 7 days (Figure 5C) of mineralization, respectively, as was judged from the corresponding SEM images. On the contrary, no particulate structures were observed for the control hydrogel (HA-hy + HA-al) even after its incubation for a week in the mineralization medium (Figure 5D). To confirm that during the freeze-drying procedure we do not damage the initial structure of the wet hydrogels, we have also performed low-temperature SEM on the cryogenically fixed samples. The cross-sections of the hydrogels are presented by upper images in Figure 5. The cross-section of the control hydrogel after 7 days of incubation in the mineralization medium appeared smooth with no indication of particles (Figure 5D). Almost the same structure was observed for the in situ formed Ca2+•BP−HA hydrogel that has been washed with water after its preparation (Figure 5A). However, when this hydrogel was subjected to mineralization, its structure changed, showing very clearly contrast of nanoparticles with a mean size of 348 ± 14 nm and 358 ± 26 nm for 1 day and 7 days of mineralization correspondingly (see arrows in Figure 5B,C). The same trend was observed when examining the surfaces of the wet hydrogels by cryo-SEM (Supporting Information, Figure S5). The sizes of the particles found on the surfaces of the mineralized BP-linked hydrogels were larger as compared with the particles in the interior of the same samples which can be explained by the diffusion gradient of the mineral ions in the direction from the surface to the interior of the hydrogel

samples and also by the fact that the cross-section does not cut through the center of each particle. To quantitatively prove the progress of mineralization, we performed a calcium assay for both types of hydrogel before mineralization as well as after 1 and 7 days of mineralization. Figure 5E shows that the amount of the deposited calcium steadily grows in the course of mineralization of the BP-linked hydrogel reaching after 7 days of incubation 174 ± 24 μg/mg of dry hydrogel, i.e., almost 20% of the whole mass of the hydrogel. The amount of calcium in the control hydrogel was significantly less and has reached its leveling already after 1 day of incubation (54 ± 8 and 59 ± 8 μg/mg of dry hydrogel for 1 and 7 days of mineralization, respectively). The increase in calcium content of the HA−BP hydrogel with only a moderate increase in size of the particles deposited in the hydrogel between the first and seventh day of mineralization (compare Figure 5B with Figure 5C) can be explained by recruitment of a higher number of nucleation points into the deposition of the mineral phase during the time. It should be noted that the conditions of mineralization in the calcium/magnesiumcontaining PBS buffer ([Ca2+] = 0.9 mM, [Mg2+] = 0.5 mM, [phosphate] = 9.7 mM) are very mild and may resemble incubation in the simulated body fluid (SBF, [Ca2+] = 2.5 mM, [Mg2+] = 1.5 mM, [phosphate] = 1 mM),27 a type of medium with ion concentration approximating those of human plasma. It has been shown that mineralization in 1 × SBF correlates well to the in vivo bioactivity of a material.9 Therefore, it may be possible that our new in situ formed hybrid hydrogel material could enhance mineralization in vivo. We used EDS to determine the elemental content of the observed particles. Four points for each sample were selected and examined. For mineralized hydrogels exhibiting integrated 1694

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

preparation of the hybrid hydrogel was done in PBS buffer where phosphate anions compete with the matrix-immobilized BP ligands for the Ca2+ ions, thus further decreasing the amount of the BP-bound calcium. Nevertheless, the observation of the very small particles in the hydrogel obtained by mixing of HA-BP-hy/PBS and HA-al/(Ca2+/Mg2+-PBS) solutions can only be attributed to the formation of bisphosphonate−metal ion complexes since the BP-linked hydrogel prepared in pure water did not show any appearance of nanoparticles (Supporting Information, Figure S3). The appearance of the calcium signal as well as the increase in intensity of the phosphorus signal after hydrogel mineralization indicate that the initial bisphosphonate−metal ion complexes acted as nucleation cores for further deposition of the calcium/ magnesium phosphates on them. Finally, we have examined the mineralized injectable hybrid hydrogels using transmission electron microscopy (TEM, Figure 7). We detected well-dispersed, almost round particles

nanoparticles, two points were chosen to coincide with NP areas, and two points were selected away from the NP areas. Figure 6A,B compares EDS spectra for the control and the

Figure 6. EDS point analysis of different areas of (A) the control and (B) the hybrid HA−BP hydrogels after in situ preparation and mineralization for 7 days. (C) EDS point analysis of different areas of the hybrid HA−BP hydrogel before soaking in a Ca2+/Mg2+-PBS buffer. Different examined areas of the hydrogels are shown as red spots on the corresponding SEM images.

hybrid HA−BP hydrogels after 7 days of mineralization measured at different positions. As expected, no phosphorus and calcium signals were detected for the control hydrogel at any position after 7 days of mineralization followed by repeated washing with water (Figure 6A). Oppositely, the elemental composition of the mineralized BP-linked hydrogel was finely structured on a nanoscale. Specifically, only the observed NP areas comprised both phosphorus and bound calcium and magnesium ions (Figure 6B). Examination of the interparticulate space of the hybrid hydrogel did not reveal the presence of these elements (Figure 6B). This suggests that the observed particles are composed of calcium and magnesium phosphate salts since the mineralization medium contained the corresponding inorganic ion precursors. Interestingly, the EDS analysis of the hybrid hydrogel before mineralization did not show calcium and magnesium ions; furthermore, the intensity of the phosphorus signal was much lower than that for the mineralized samples (Figure 6C). On the other hand, we did observe Ca2+ ions for the BP-linked hydrogel after its incubation in the 0.2 M CaCl2 solution followed by the washing step (Supporting Information, Figure S4B). This fact can be explained by the much lower amount of Ca2+ for the in situ preparation of the hybrid hydrogel compared to the amount of Ca2+ that could diffuse into the preprepared hydrogel sample during its incubation in the 0.2 M CaCl2 solution. Another reason was that the injectable (in situ)

Figure 7. TEM images of the hybrid HA−BP hydrogels mineralized for 7 days (A, B) and 1 day (C, D). The high magnification images of compound particles are shown in panels B and D taken from the corresponding images A and C as indicated by red arrows. Two individual grains are highlighted by black arrows in panel B. SAED patterns are shown in B and D as insets.

with an average size of 325 ± 10 nm and 375 ± 12 nm for the gels mineralized for 1 day and 7 days, respectively. It is noteworthy that individual grains could be easily identified within the particles, although the grains were more compact after 7 days of mineralization than after 1 day of the gel exposure to the mineralization medium. Each compound particle comprises about 30 small grains of ≈50 nm in diameter. Moreover, single grains that are not part of the larger aggregates can be seen well dispersed in the space between the aggregates with a density of 21 ± 1 grains/μm2. Smaller compound particles together with larger ones as well as single grains were detected in the sample mineralized for 1 day, shedding light on the mechanism of the mineralization process. It is most probable that this process begins with mineral deposition on the small nucleation points of the chemically 1695

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials

Article

be examined as well to ensure the biological safety of the novel material. We have shown above that the hybrid Ca2+•BP−HA hydrogel is degraded with hyaluronidase with the formation of the hybrid nanogel particles (Figure 2). These particles were therefore added to the NIH 3T3 cells at concentrations varying from 0.025 to 0.2 wt %. The degradation products of the hydrogel lacking BP groups were also used for the control experiment. Figure 8B shows that the hybrid Ca2+•BP−HA nanoparticles and the HA fragments from the control hydrogel are nontoxic to the cells within the chosen concentration range. More than 85% of the cells survived after incubation with the hydrogel degradation products for 48 h. The obtained results imply that the novel hybrid hydrogel material and its potential degradation products will not be toxic in vivo as well.

immobilized bisphosphonate−metal ion complexes until enough grown grains meet each other and aggregate through the ionic interactions. Selected area electron diffraction (SAED) patterns of the particles were acquired (Figure 7B,D). From the width of the diffraction rings, it was possible to determine that they are amorphous or nanocrystalline with a crystallite size of less than 1−2 nm. Cytotoxicity. Figure 8A shows the relative viability of the NIH 3T3 cells after incubation with the HA-hy polymer and



CONCLUSIONS In conclusion, we have developed an injectable hyaluronic acid hydrogel system hybridized with the nanosized Ca−bisphosphonate clusters. Here we combine organic and inorganic functional groups along the biodegradable polysaccharide macromolecule to induce formation of the inorganic clusters within the simultaneously forming hydrogel template. The capacity of the in situ generated hybrid for further mineralization has been demonstrated in vitro. The herein presented approach permits a direct transition from the nonordered solution phase of the randomly moving macromolecules and ions to the hybrid hydrogel matrix. This transition is accomplished within a very short time and under physiologic conditions which allows easy encapsulation of complex biological molecules, such as proteins and nucleic acids, as well as cells. The new hybrid hydrogel showed no apparent toxic effect on cells and therefore can be considered as a potential candidate in various biomedical applications.



ASSOCIATED CONTENT

S Supporting Information *

Experimental section and the following figures and table: (Figure S1) 1H and 31P NMR of the HA-hy-BP derivative; (Figure S2) Photographs of HA hydrogel with or without BP groups after immersion in CaCl2 solution; (Figure S3) SEM images of BP-linked hydrogel before immersion in CaCl2; (Figure S4) EDS spectra of HA hydrogel with or without BP groups after immersion in CaCl2; (Figure S5) Cryo-SEM images of the surfaces of wet hydrogel before mineralization, after 1 day of mineralization, and after 7 days of mineralization; and (Table S1) Mechanical properties of HA hydrogels after interaction with Ca2+ ions (PDF). This material is available free of charge via the Internet at http://pubs.acs.org.

Figure 8. (A) Cell viability of NIH 3T3 cells after incubation for 48 h with various concentrations of the HA-hy polymer (black bars) and the HA-hy-BP polymer (gray bars). (B) Cell viability of NIH 3T3 cells after incubation for 48 h with various concentrations of the Hase digests derived from the control (black bars) and the Ca2+•BP−HA (gray bars) hydrogels.

HA-hy-BP polymer. The viabilities of cells were more than 90% after 48 h incubation with HA-hy polymer at concentrations varying from 0.025 to 0.2 wt %. In the case of the highest concentration of HA-hy-BP polymer (0.2 wt %), still 85.6 ± 5.1% of cells survived. The cell viability increased to 99.9 ± 4.7% when the concentration of HA-hy-BP decreased to 0.1 wt %. In our previous study, we have also shown that due to the higher reactivity of aldehyde groups the HA-al derivative has a higher toxicity on cells. In the same work, we have also presented that after cross-linking the formed hydrogels become completely nontoxic to the cells.25 Since we inject the mixture of the gel-forming components already after reaching the gel point, the BP-immobilized hydrogel presented in this work should be nontoxic to cells as the control one that has been examined in our previous study.25 It has been demonstrated in our previous animal studies that HA-based hydrogels are gradually degraded in vivo.20−22 Due to chemical modification of native HA with BP groups, the degradation products of the BP-immobilized hydrogel have to



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The research leading to these results has received funding from the European Community’s Seventh Framework Programme (MultiTERM, Grant no: 238551) and the Higher Education Commission (HEC) of Pakistan. We are also thankful to Mr. Kootala S. for assisting in cell toxicity studies. 1696

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697

Chemistry of Materials



Article

REFERENCES

(1) Hunter, G. K.; Goldberg, H. A. Biochem. J. 1994, 302, 175−179. (2) Leeuwenburgh, S. C.; Jansen, J. A.; Mikos, A. G. J. Biomater. Sci. Polym., Ed. 2007, 18, 1547−1564. (3) Mann, S. Biomineralization: principles and concepts in bioinorganic materials chemistry; Oxford University Press: Oxford, NY, 2001. (4) Stancu, I. C.; Filmon, R.; Cincu, C.; Marculescu, B.; Zaharia, C.; Tourmen, Y.; Basle, M. F.; Chappard, D. Biomaterials 2004, 25, 205− 213. (5) Suzuki, S.; Whittaker, M. R.; Grøndal, L.; Monteiro, M. J.; Wentrup-Byrne, E. Biomacromolecules 2006, 7, 3178−3187. (6) Wan, A. C. A.; Mao, H. −Q.; Wang, S.; Phua, S. H.; Lee, G. P.; Pan, J.; Lu, S.; Wang, J.; Leong, K. W. J. Biomed. Mater. Res. 2004, 70B, 91−102. (7) Nuttelman, C. R.; Benoit, D. S. W.; Tripodi, M. C.; Anseth, K. S. Biomaterials 2006, 27, 1377−1386. (8) Wang, D. −A.; Williams, C. G.; Yang, F.; Cher, N.; Lee, H.; Elisseeff, J. H. Tissue Eng. 2005, 11, 201−203. (9) Kokubo, T.; Takadama, H. Biomaterials 2006, 27, 2907−2915. (10) Song, J.; Malathong, V.; Bertozzi, C. R. J. Am. Chem. Soc. 2005, 127, 3366−3372. (11) Wang, L.; Zhang, M.; Yang, Z.; Xu, B. Chem. Commun. 2006, 2795−2797. (12) Wang, C.; Flynn, N. T.; Langer, R. Adv. Mater. 2004, 16, 1074− 1079. (13) Tarnavchyk, I.; Voronov, A.; Kohut, A.; Nosova, N.; Varvarenko, S.; Samaryk, V.; Voronov, S. Macromol. Rapid Commun. 2009, 30, 1564−1569. (14) Fleisch, H.; Russell, R. G. G.; Straumann, F. Nature 1966, 212, 901−903. (15) Yoshinary, M.; Oda, Y.; Ueki, H.; Yokose, S. Biomaterials 2001, 22, 709−715. (16) Vasikaran, S. D. Ann. Clin. Biochem. 2001, 38, 608−623. (17) Heymann, D.; Ory, B.; Gouin, F.; Green, J. R.; Redini, F. Trends Mol. Med. 2004, 10, 337−343. (18) Yin, Y. J.; Luo, X. Y.; Cui, J. F.; Wang, C. Y.; Guo, X. M.; Yao, K. D. Macromol. Biosci. 2004, 4, 971−977. (19) Greish, Y. E.; Brown, P. W. Biomaterials 2001, 22, 807−816. (20) Bergman, K.; Engstrand, T.; Piskounova, S.; Ossipov, D.; Hilborn, J.; Bowden, T. J. Biomed. Mater. Res., Part A 2009, 91, 1111− 1118. (21) Martinez-Sanz, E.; Ossipov, D. A.; Hilborn, J.; Larsson, S.; Jonsson, K. B.; Varghese, O. P. J. Controlled Release 2011, 152, 232− 240. (22) Docherty Skogh, A. −C.; Bergman, K.; Jensen Waern, M.; Ekman, S.; Hultenby, K.; Ossipov, D.; Hilborn, J.; Bowden, T.; Engstrand, T. Plast. Reconstr. Surg. 2010, 125, 1383−1392. (23) Varghese, O. P.; Sun, W.; Hilborn, J.; Ossipov, D. J. Am. Chem. Soc. 2009, 131, 8781−8783. (24) Russell, R. G. G.; Watts, N. B.; Ebetino, F. H.; Rogers, M. J. Osteoporosis Int. 2008, 19, 733−759. (25) Ossipov, D. A.; Piskounova, S.; Varghese, O. P.; Hilborn, J. Biomacromolecules 2010, 11, 2247−2254. (26) Kretlow, J. D.; Mikos, A. G. Tissue Eng. 2007, 13, 927−938. (27) Kokubo, T.; Takadama, H. Biomaterials 2006, 27, 2907−2915.

1697

dx.doi.org/10.1021/cm300298n | Chem. Mater. 2012, 24, 1690−1697