Communication Osteopromotive Coatings
www.advmat.de
Diatom-Inspired Silica Nanostructure Coatings with Controllable Microroughness Using an Engineered Mussel Protein Glue to Accelerate Bone Growth on Titanium-Based Implants Yun Kee Jo, Bong-Hyuk Choi, Chang Sup Kim,* and Hyung Joon Cha* involve surface topographic modifications of bulk titanium substrates to create microrough surfaces.[4] Microroughness has been shown to enable higher success rates of implants by enhancing the surface area in contact with bone and thus reinforcing biomechanical interlocking.[5] However, the loosening or failure of titanium implants still occurs especially in cases of compromised bone conditions, which might be caused mainly by the limited bioactivity of inert titanium substrates.[6] Thus, a more advanced design for further accelerating the osseointegration of microrough implants is required. The assembly of inorganic nanostructures has emerged as one of the next generation of surface modification strategies for simultaneously altering the topography and chemistry of implant surfaces.[7] In bone tissue engineering, silica has received considerable attention as a promising coating material due to its osteoinductive properties and biocompatibility.[8] In particular, silica nanoparticles (SiNPs) have been applied in several approaches towards generating nanostructures in order to impart surface topographic features and intrinsic bioactivity on the substrates,[9] and SiNP-based surface modification has been shown to introduce nanoscale roughness within a thin film on a titanium surface.[10] However, a cluster of SiNPs (i.e., agglomerate) is susceptible to catastrophic failure due to its low toughness, restricting its scalability for clinical use on load bearing sites such as the bone–implant interface.[11] In this regard, to the best of our knowledge, the assembly of SiNPs has never been exploited to form microrough surfaces on implants, despite the well-known enhancements of performance observed from microrough implants. Furthermore, the microscale surface topographic effects of silica coatings on osteogenic cell behaviors have not yet been systematically investigated. In the present work, we propose a novel, osteopromotive SiNP coating that is capable of tuning surface microroughness for titanium implant applications by employing two intriguing bioinspired strategies from diatoms and mussels. First, the skeletal architectures of marine diatoms are composed of hierarchical silica nanostructures containing organic components such as proteins and amines that facilitate the rapid formation
Silica nanoparticles (SiNPs) have been utilized to construct bioactive nanostructures comprising surface topographic features and bioactivity that enhances the activity of bone cells onto titanium-based implants. However, there have been no previous attempts to create microrough surfaces based on SiNP nanostructures even though microroughness is established as a characteristic that provides beneficial effects in improving the biomechanical interlocking of titanium implants. Herein, a protein-based SiNP coating is proposed as an osteopromotive surface functionalization approach to create microroughness on titanium implant surfaces. A bioengineered recombinant mussel adhesive protein fused with a silica-precipitating R5 peptide (R5MAP) enables direct control of the microroughness of the surface through the multilayer assembly of SiNP nanostructures under mild conditions. The assembled SiNP nanostructure significantly enhances the in vitro osteogenic cellular behaviors of preosteoblasts in a roughness-dependent manner and promotes the in vivo bone tissue formation on a titanium implant within a calvarial defect site. Thus, the R5-MAP-based SiNP nanostructure assembly could be practically applied to accelerate bone-tissue growth to improve the stability and prolong the lifetime of medical implantable devices.
Titanium implants have been commonly used for replacing missing teeth in dentistry due to their desirable mechanical properties and biocompatibility.[1] The clinical success of titanium implants directly relates to their integration with the surrounding bone (osseointegration).[2] In this regard, surface modifications have been widely exploited for the design of implants aimed toward improving interaction with bone cells at the bone–implant interface.[3] Most commercial implants Dr. Y. K. Jo, Dr. B.-H. Choi, Prof. H. J. Cha Department of Chemical Engineering Pohang University of Science and Technology Pohang 37673, Korea E-mail:
[email protected] Prof. C. S. Kim School of Chemistry and Biochemistry Yeungnam University Gyeongsan 38541, Korea E-mail:
[email protected] The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adma.201704906.
DOI: 10.1002/adma.201704906
Adv. Mater. 2017, 1704906
1704906 (1 of 9)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advmat.de
of SiNPs under physiologically mild conditions.[12] The use of diatom-inspired SiNP coatings could be a particularly attractive approach for biomedical applications due to mild processing in ambient temperature and pH without the presence of caustic chemicals.[13] Second, mussels secrete specialized glues, mussel adhesive proteins (MAPs) that exhibit superior adhesion strength on both organic and inorganic surfaces even under seawater, as well as biocompatibility and biodegradability.[14] To overcome the availability limitation of natural MAPs, recombinant MAP has been developed as a practical bioglue that can be mass-produced with a simple purification process and shown to be efficiently applicable for adhesion, coating, immobilization, and binding.[15–17] Importantly, the highly positive charge of recombinant MAP enables the fabrication of multilayer composites with negatively charged molecules, such as SiNPs, using the layer-by-layer (LbL) assembly technique.[15,16] In fact, the LbL assembly of nanoparticles has the advantages of improving mechanical properties and fine-tuning surface topography.[18] Moreover, recombinant MAP can be genetically fused with a short functional peptide, which enables it to provide desirable bioactivity on a target surface without additional modification steps.[17] Here, we engineered a silica-forming protein glue, R5-MAP, through the genetic fusion of recombinant MAP with the R5 peptide derived from a marine diatom Cylindrotheca fusiformis and evaluated the potential applicability of the resulting SiNP nanostructure composite (R5-MAP@SiNPs) for accelerating bone growth on titanium implants via in vitro and in vivo studies. In particular, we examined for the first time the microscale topographic effects of SiNP nanostructures on osteogenic cell behaviors by tailoring the microroughness of R5-MAP@ SiNPs multilayer assembled surfaces. The engineered R5-MAP was successfully produced and purified from a bacterial system (Figure S1, Supporting Information). We prepared the R5-MAP@SiNPs multilayer assemblies on titanium surfaces by repeating the sequential process of R5-MAP coating and silicification for up to six layers (Figure 1A,B). The resulting multilayered surfaces were denoted as nSi, where n is the number of R5-MAP@SiNPsassembled layers. The SiNP nanostructure was successfully generated on the R5-MAP-coated titanium surface with a diameter of 111.5 ± 31.0 nm, as determined by scanning electron microscopy (SEM), transmission electron microscopy (TEM), and energy-dispersive X-ray spectroscopy (EDS) (Figure S2A–D and Table S1, Supporting Information). The ability of R5-MAP to facilitate the generation of the SiNP nanostructure on titanium was confirmed by comparing it with negative controls; however, the coating of original MAP without the silica-precipitating moiety also yielded a few granular silica precipitates, possibly due to its high isoelectric point (pI ≈ 10).[15,19] Thus, we concluded that the adhesive properties of MAP and the silica-precipitating activity of the R5 peptide in R5-MAP enabled the formation of a SiNP nanostructure efficiently on the target surface under mild conditions. From Fourier-transform infrared (FT-IR) spectroscopic analysis, we observed the distinctive silicon peaks at ≈800 cm−1 (SiOSi bending), ≈1050 cm−1 (SiOSi stretching), and ≈3380 cm−1 (SiOH stretching), as well as the characteristic peaks for protein amide bonds at ≈1550 cm−1 (NH deformation) and ≈1635 cm−1 (CO Adv. Mater. 2017, 1704906
stretching) on each multilayered surface, confirming the successful assembly of the R5-MAP@SiNPs multilayers on titanium surfaces (Figure 1C). The silicon peaks grew gradually with the increase in the number of assembled layers, whereas the protein amide bond peaks declined. One possible explanation for this observation could be the denser exposed SiNPs on the exterior surface of a high number of assembled layers, considering the occlusion of R5-MAP inside the SiNP nanostructure (Figure S2C, Supporting Information). In topographic images obtained by white light interferometry (WLI), all the R5-MAP@SiNPs multilayered surfaces were markedly rough compared to a bare titanium surface (Figure 1D, and Figure S3A, Supporting Information). These findings could be attributed to the undulating surface features of the R5-MAP coating with randomly distributed agglomerations as well as the formation of highly dense SiNPs within the 3D nanostructure. As observed by the gradual increase in average surface roughness (Sa), R5-MAP@SiNPs multilayered surfaces became rougher when the number of assembled layers increased, which might be attributed to the assembly of opposite charges between R5-MAP and SiNPs that leads to their aggregation resulting in rougher surfaces (Figure 1E).[20] Once the surface roughness reached a maximum value of ≈4.4 µm, there were no more significant increases in the surface roughness upon further assembly of the R5-MAP@SiNPs nanostructures, probably due to the saturation of the surface with broadening ridges and the filling of grooves over a certain level of roughness. We classified the R5-MAP@SiNPs multilayer-assembled surfaces into three groups according to average surface roughness: moderately rough silica (MR-Si, 1.0 µm 4.0 µm). Moreover, the R5-MAP@SiNPs multilayer assembly led to a linear increase in thickness of ≈3.9 µm per assembled layer, as measured by WLI height profiling along the two-point straight line, which implied that the R5-MAP-based SiNP nanostructure assembly could be properly designed for practical applications with reliable thickness control (Figure S3B,C, Supporting Information). From load–displacement curves obtained by nanoindentation analyses (Figure S4A, Supporting Information), all the R5-MAP@SiNPs multilayered surfaces exhibited significantly higher elastic modulus (E) modulus (≈1.1–2.3 GPa) than that (≈10 kPa) of SiNPs agglomerates (Table 1).[21] We surmise that the R5-MAP bridges opening gap between SiNPs, increasing flexibility of network in nanocomposites and thus efficiently dissipating energy, in great contrast with the SiNPs agglomerates that exhibit local stress concentration due to their large void fraction.[21] One interesting observation is that elastic modulus, hardness (H), and elastic strain (H/E) of the R5-MAP@SiNPs multilayered surfaces from 2Si to 6Si were all increased compared to 1Si surface. These results indicate that the multilayer assembly of R5-MAP@SiNPs nanostructure imposes higher resistance to local elastic and plastic deformation than the monolayer assembly, while exhibiting somewhat regular level of nanostructural deformability around top layers. A gradual increase of variability with increasing the number of assembled layers might be because surface roughness causes uncertainties in nanoindentation tests.[22]
1704906 (2 of 9)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advmat.de
Figure 1. Surface characterization of the R5-MAP@SiNPs multilayer-assembled titanium surfaces. A) Schematic representation of the R5-MAP-based SiNP nanostructure multilayer assembly on a titanium surface. B) SEM analyses for 6Si and C) FT-IR spectra. D) Topographic images (456.6 × 608.6 µm2) for a) top view, b) oblique view, and c) side view obtained by WLI analysis. Different colors indicate the relative altitude for peaks (red), flat regions (green), and valleys (blue). E) Average roughness (Sa). Abbreviations: NC, bare titanium surface (negative control); R5-MAP, nonsilicified R5-MAPcoated titanium surface; nSi, multilayered titanium surface assembled with n layers of R5-MAP@SiNPs nanostructure; MR-Si, moderately rough R5-MAP@SiNPs-assembled titanium surface (1.0 µm < Sa < 2.5 µm); R-Si, rough R5-MAP@SiNPs-assembled titanium surface (2.5 µm 4.0 µm).
To assess the adhesion strength of multilayer coating on titanium substrate, the R5-MAP@SiNPs multilayered surfaces were scratched by applying a progressive load (Figure S4B, Supporting Information). The critical load for chippings or spallation (Lc2) increased with the number of assembled layers
Adv. Mater. 2017, 1704906
(Table 1). From the external appearance of scratches, large-scale spallation already appeared on 1Si surface at the load of ≈2.0 N, while relatively small regions of chippings were obvious on 3Si surface at the load of ≈4.0 N (Figure S4C, Supporting Information). Similarly, the critical load for complete delamination
1704906 (3 of 9)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advmat.de
Table 1. Mechanical properties of the R5-MAP@SiNPs multilayer-assembled titanium surfaces. Lc2c) [N]
Lc3d) [N]
Ea) [GPa]
Hb) [GPa]
1Sie)
1.09 ± 0.35
0.040 ± 0.016
0.0374
0.81 ± 0.07
2.22 ± 0.17
2Sie)
2.30 ± 0.48
0.155 ± 0.031
0.0673
1.49 ± 0.09
3.37 ± 0.18
3Sie)
2.33 ± 0.64
0.152 ± 0.054
0.0665
1.92 ± 0.08
4.57 ± 0.20
4Sie)
2.31 ± 0.74
0.158 ± 0.079
0.0683
2.27 ± 0.09
5.63 ± 0.38
5Sie)
2.30 ± 0.87
0.162 ± 0.076
0.0704
2.53 ± 0.09
7.27 ± 0.32
e)
2.32 ± 1.13
0.157 ± 0.085
0.0676
2.90 ± 0.11
8.38 ± 0.38
Surface
6Si
H/E
a) E, elastic modulus; b)H, hardness; c)Lc2, critical load for chipping failure or local spallation; d)Lc3, critical load for complete delamination; e)nSi, multilayered titanium surface assembled with n layers of R5-MAP@SiNPs nanostructure.
(Lc3) also increased with the number of assembled layers (Table 1). Considering that mechanical property of hybrid composite strongly depends on interfacial bridging between organic polymer and inorganic particles,[23] the reinforcement effects might be attributed to the increase of adhesion number between R5-MAP and SiNPs by the addition of layer as well as the intrinsic adhesion/cohesion properties of R5-MAP molecules toward SiNPs, titanium substrate, and themselves. Particularly, it is noteworthy that the maximum of adhesion strength at ≈8.4 N for R5-MAP@SiNPs multilayers (i.e., Lc3 for 6Si surface) is about an order of magnitude comparable to those of conventional hydroxyapatite coatings, the most widely used mineral-based methods for titanium implants.[24] Therefore, the R5-MAP@SiNPs multilayer assembly can be expected to have sufficient damage tolerance and adhesion strength, making it suitable coating system for titanium implant applications. To verify the osteopromotive ability of the R5-MAP@SiNPs multilayer and to investigate the cell responses to the distinct surface microroughness of the silica coating, the osteogenic cell behaviors of preosteoblasts on the R5-MAP@SiNPs multilayer-assembled titanium surfaces were systematically studied in vitro. A bare titanium surface and a nonsilicified R5-MAPcoated titanium surface were used as negative controls. We first measured the cell counts of preosteoblasts on the surfaces both 1 h (for attachment) and 24–72 h (for proliferation) post-cell seeding. Cells on all the R5-MAP@SiNPs multilayered surfaces showed significantly increased attachment and proliferation compared with those on the bare surface (Figure 2A, and Figure S5, Supporting Information). Cell attachment was enhanced gradually with the increase in the roughness of the multilayered surfaces, which was consistent with previous studies demonstrating that rougher surfaces exhibited improved cell attachment by providing osteoblasts with more anchoring sites.[25] By contrast, the proliferation level of cells on multilayered surfaces after 72 h of incubation showed a maximum value on R-Si surfaces. Cells on ER-Si surfaces exhibited lower cell counts than those of the cells on R-Si surfaces at this time point, possibly because of the excessive irregularity and limited flat regions of ER-Si surfaces, which could disrupt the stable adhesion and active growth of cells. Cell morphology, an important regulating factor for cell cycle progression,[26] was also analyzed under serum-free conditions (Figure S6, Supporting Information). Cells on MR-Si surfaces displayed a flattened appearance with the largest spreading
Adv. Mater. 2017, 1704906
area, as observed by SEM and immunofluorescence images (Figure 2B, and Figure S7A,B, Supporting Information). We found that the morphologies of the cells on multilayered surfaces tended to be less spread out and more elongated as the surface roughness increased (Figure S7C,D, Supporting Information). Cells on ER-Si surfaces clearly and prominently exhibited a spindle-like shape with long filopodia extensions for clinging to the grooves and ridges of the underlying nanostructure. The excessive irregularity of ER-Si surfaces perhaps exerted stronger constraints on cell spreading, similar to the cell proliferation results. In general, osteoconductivity refers to a situation in which the attachment of new osteoblasts can be supported for subsequent bone growth.[27] We suggest that the R5-MAP@SiNPs nanostructures served as a highly osteoconductive scaffold for encouraging the ingrowth of bone cells in the early phase and that the initial cell–substrate interactions could be tailored by altering the roughness of the multilayered surface. Immediately after attachment to substrate, cells adapt to the surrounding microenvironment through an integrin-mediated focal adhesion process.[28] Upon the triggering of the signal transduction pathway with integrin engagement, focal adhesion kinase (FAK) becomes activated by undergoing a phosphorylation of its tyrosine-397 residue and thereby organizes clusters with cytoskeletal proteins such as vinculin to connect the actin cytoskeleton to transmembrane integrins.[29] From immunoblotting results, it was found that, after culturing for 24 h, the cells on all the R5-MAP@SiNPs multilayered surfaces showed superior expression of phosphorylated FAK (pY397-FAK) and vinculin compared with the expression of the two proteins in cells on the bare surface (Figure S8A, Supporting Information). The levels of FAK phosphorylation and vinculin expression on the multilayered surfaces increased with the increasing roughness until the surface roughness exceeded a critical value of ≈2.5 µm; after that, they dropped steadily (Figure 2C,D). Considering that FAK phosphorylation directly regulates the strength of cellular adhesion,[29] the higher FAK phosphorylation on MR-Si surfaces could thus be coupled with the strong focal adhesion of the widespread cells. Similarly, the severe irregularity of ER-Si surfaces could be associated with the lower FAK phosphorylation levels of the spindle-shaped cells. Moreover, vinculin-positive focal adhesions were widely distributed over the entire area of cells on the R5-MAP@SiNPs nanostructured surface, whereas vinculin appeared narrowly around the
1704906 (4 of 9)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advmat.de
Figure 2. In vitro osteogenic cell behaviors of preosteoblasts on the R5-MAP@SiNPs multilayer-assembled titanium surfaces. A) Cell counts for the initial 3 d. B) Morphologies of the cells at day 1. The white triangles indicate the filopodia extensions of the cells into the SiNP nanostructure. C) FAK phosphorylation and D) vinculin expression of the cells at day 1. FAK activation was quantified through the relative band intensity of phosphorylated focal adhesion kinase (pY397-FAK) to focal adhesion kinase (FAK). E) ALP activity of the cells at day 15. F) Calcium deposition levels and G) optical microscopic images of Alizarin Red S-stained cells at day 21. The data represent the mean ± standard deviation with statistical significance (*, #p < 0.05, **, ##p < 0.01, ***, ###p < 0.005; unpaired t-test). Abbreviations: NC, bare titanium surface (negative control); R5-MAP, nonsilicified R5-MAPcoated titanium surface; nSi, multilayered titanium surface assembled with n layers of R5-MAP@SiNPs nanostructure; MR-Si, moderately rough R5-MAP@SiNPs-assembled titanium surface (1.0 µm
