An Electrophoresis-Aided Biomineralization System for Regenerating

Sep 12, 2014 - Synopsis. A novel electrophoresis-aided calcium and phosphate agarose hydrogel system was developed for quickly regenerating tooth-like...
1 downloads 11 Views 6MB Size
Download PDF
Article pubs.acs.org/crystal

An Electrophoresis-Aided Biomineralization System for Regenerating Dentin- and Enamel-Like Microstructures for the Self-Healing of Tooth Defects Xiao-Ting Wu,† Ying Cao,‡ May Lei Mei,‡ Jia-Long Chen,† Quan-Li Li,*,† and Chun Hung Chu*,‡ †

College & Hospital of Stomatology, Anhui Medical University, Hefei 230032, China Faculty of Dentistry, University of Hong Kong, Sai Ying Pun, Hong Kong, China

‡

ABSTRACT: It is the challenging and desirable aim of biomineralization studies to regenerate tooth tissue microstructures and to accelerate the speed of available biomimetic remineralization protocols for potential use in clinical dentistry. In this study, we developed a novel electrophoresis-aided calcium and phosphate agarose hydrogel system that can be used to quickly regenerate tooth tissue microstructure. After remineralization for 12 h, a demineralized dentin collagen matrix was remineralized with intrafibrillar and interfibrillar hydroxyapatites that mimicked the structure of the original calcified dentin collagen matrix. The precipitated hydroxyapatites were densely packed, occluded the exposed dentinal tubules, and regularly and homogeneously distributed along the collagen fibrils in a “string-of-beads” structure within the dentin collagen matrix. Needle-like hydroxyapatite crystals were densely packed, with their c-axis parallel to one another, to form the enamel-like tissue that precipitated onto the remineralized dentin surface. This study provides a potential protocol for inducing the self-healing of dentin defects.

■

INTRODUCTION Dentin is the main component of teeth. Its outer surface is covered by enamel, and its inner surface forms a pulp cavity, full of dental pulp soft tissue. Enamel, the exterior layer of a tooth, is the hardest and most highly mineralized human tissue. It loads the stresses associated with mastication and protects the dentin−pulp complex. Tooth caries, abrasions, attrition, and erosion result from enamel loss and exposed dentin. Dental hard tissue, once exposed to an oral cavity, cannot self-heal if damaged; this is true of both enamel and the underlying dentin. Recently, cell-free biomimetic mineralization strategies have proved capable of regenerating enamel-like or dentin-like tissue in vitro. This breakthrough has created the potential for repairing tooth defects by self-healing processes, a much more desirable strategy for use in dentistry clinics than current procedures. However, no study so far has reported the simultaneous regeneration of both enamel- and dentin-like tissue nor has there been a report describing the assembly of enamel-like tissue on the surface of dentin-like tissue to form a natural tooth-like structure. Approximately 96% of enamel is composed of crystallites of the inorganic mineral hydroxyapatite (HA). HA crystals © 2014 American Chemical Society

assemble into a prism structure, and these prisms are tightly packed in an organized pattern to form a special enamel microstructure. Each prism is approximately 4−8 μm in diameter and is made up of HA crystals bundled in parallel with one another, with each crystal having a cross-section of 25−100 nm and a length that can vary anywhere from 100 nm to 100 μm or more along the c-axis. Various methods have been proposed to synthetically generate an enamel-like prism structure and include hydrothermal methodologies,1 the application of extremely acidic calcium phosphate pastes containing hydrogen peroxide and phosphoric acid,2 and the use of surfactants,3,4 ethylenediaminetetraacetic acid (EDTA),5 nano-HA and glutamic acid,6 amelogenin,7−10 gelatin,11 poly(ethylene oxide), and polyacrylamide.12 All of these methods regenerated enamel-like tissue on a demineralized enamel surface but not on a dentin surface. Structurally, the basic microstructure of dentin comprises a calcified collagen matrix. It is composed of inorganic HA-based Received: May 26, 2014 Revised: August 28, 2014 Published: September 12, 2014 5537

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 1. Schematic diagram of the electrophoresis-aided calcium and phosphate hydrogel system.

To summarize, the remineralization of dentin collagen fibrils to duplicate the hierarchical structure of calcified collagen fibrils cannot be accomplished by simply remineralizing enamel. Moreover, the time scale of mineralization directly limits the usefulness of biomimetic methods, and for most previously reported methods, it is very long. Furthermore, no study so far has reported the regeneration of enamel-like tissue on a remineralized dentin surface in a fashion that is similar to the enamel−dentin structure observed in natural tooth. We recently developed a calcium and phosphate agarose hydrogel system to remineralize dentin and enamel in the absence of noncollagenous protein analogues.23,30 However, as mentioned above, the speed of remineralization was very slow, and it had little effectiveness in remineralizing collagen fibrils to duplicate the calcified collagen fibril hierarchical structure found in natural tooth. It has been reported that electrophoresis can transport ions more rapidly through a gel or solution than diffusion alone and can be used to accelerate HA formation in agarose hydrogels for the synthesis of HA−agarose hybrid materials.31,32 Furthermore, electrophoresis also enables ion migration in a specific one-dimensional direction. Thus, based on our previous biomimetic mineralization model, we designed a novel mineralizing system aided by electrophoresis to promote crystal growth. We hypothesized that an electric current can promote the diffusion of calcium and phosphate ions into the interior of the demineralized collagen matrix to induce collagen fibril remineralization and to accelerate the speed of mineralization. In the present study, we aimed to both regenerate the tooth microstructure and shorten the time required to do so by the application of electrophoresis to the remineralization process. We also sought to cover the remineralized surface of demineralized dentin with regenerated enamel-like tissue in an agarose hydrogel system.

minerals (approximately 70% by weight), an organic matrix (20%), and water (10%). The main organic substance of dentin is type I collagen, which self-assembles into fibrils to form collagen matrix scaffolds. The inorganic HA crystals can be classified as either intrafibrillar or interfibrillar crystallites. Intrafibrillar crystallites are deposited with their c-axis aligned in parallel to the collagen fibrils and are thus oriented along the collagen fibrils. Interfibrillar crystallites are deposited between the collagen fibers, and hence they are also referred to as extrafibrillar crystallites. Calcified collagen fibrils, especially those produced by the intrafibrillar mineralization of collagen fibrils, are believed to be responsible for the biomechanical properties of dentin as well as for protecting collagen fibers from degradation.13,14 Thus, it is essential to form intrafibrillar HA when remineralizing the dentin collagen matrix to regenerate a dentin microstructure that contains calcified collagen fibrins. The remineralization of dentin collagen fibrils is more complicated and more difficult to do than the remineralization of enamel.15 The role of the collagen matrix in apatite mineralization still remains a topic of debate. However, an everincreasing base of evidence supports the notion that the dentinal collagen matrix is ineffective for intrafibrillar mineralization. Noncollagenous proteins play an important role in initiating and regulating HA nucleation and growth to induce collagen mineralization, especially intrafibrillar mineralization.16−18 Several methods that do not use noncollagenous protein, such as caseinphosphopeptide−amorphous calcium phosphate, colloidal nano-β-tricalcium phosphate and bioactive glass particles, polydopamine, and agarose gel, did not demonstrate success in reproducing the natural structural hierarchy of apatite deposited within collagen matrices.19−23 On the contrary, the biomimicry of noncollagenous proteins with substances such as poly(acrylic acid) and polyvinylphosphonic acid, sodium trimetaphosphate, tripolyphosphate, oligopeptides inspired by noncollagenous proteins, polymerinduced liquid mineral precursors (PILP), and poly(amido amine) dendrimers have successfully triggered intrafibrillar mineralization.24−29 However, no researchers have reported forming enamel-like tissue on a remineralized dentin surface. It is noteworthy that the rate of precipitate growth is very slow in the aforementioned methods, which limits their clinical applicability. Tay et al. reported that mineralization in a 5 μm demineralized region started 2 weeks after treatment, and complete remineralization was achieved after 8 weeks.15 Busch reported that the rate of enamel-like tissue regeneration was approximately 500 nm/day with a gelatin hydrogel model.11

■

EXPERIMENTAL SECTION

Dentin Slice Preparation. This study was approved by the University of Hong Kong/Hospital Authority Hong Kong West Cluster Institutional Review Board (IRB UW10-210). Patients who required the extraction of their sound molars were invited to donate their extracted molars. The soft tissue attached to the molars was removed with a scalpel. The molars were disinfected with 3% sodium hypochlorite and rinsed with phosphate-buffered saline. Dentin slices of 2 mm thickness were prepared perpendicular to the longitudinal axis of the tooth using a low-speed diamond saw (IsoMet Low Speed Saw, Buehler, Lake Bluff, Illinois, USA). Silicon carbide papers were used for polishing. The slices were ultrasonically cleaned and stored at 4 °C in deionized water. 5538

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Agarose Gel Preparation. A CaCl2−agarose gel was prepared by mixing agarose powder (Regular Agarose G-10, BIOWEST, Nuaillé, France) (1.0 g) into 100 mL of 0.13 M CaCl2 solution (CaCl2·2H2O, Sigma-Aldrich, St. Louis, MO, USA). A Na2HPO4−agarose gel containing 500 ppm fluoride was prepared by mixing agarose powder (1.0 g) into 100 mL of 0.26 M Na2HPO4 (Sigma-Aldrich, St. Louis, MO, USA) solution containing 500 ppm fluoride (Sigma-Aldrich, St. Louis, MO, USA). The pH values of the solutions were adjusted to 6.5 using 0.1 M NaOH and 0.1 M HCl. The mixtures were allowed to swell for 30 min and were then heated to 100 °C to completely dissolve the agarose. Remineralization in Agarose Gel Aided by Electrophoresis. The dentin slices were etched with a 37% H3PO4 solution for 30 s to create a demineralized dentin surface with an exposed collagen matrix. The electrophoresis device consists of a two-way horizontal polyether tube, two plastic cells, and two graphite electrodes (DYY-10C Electrophoresis, Liuyi Instrument Factory, Beijing, China) (Figure 1). The CaCl2 agarose gel and the Na2HPO4 agarose gel were put into the two sides of the tube separated by the etched dentin slice. The tube was then connected to the plastic cells. Electrodes were set into the bottom of the cells, which were filled with 0.9% NaCl solution to enhance the electrical conductivity. The electric current was maintained constant at 20 mA during electrophoresis. The gels and NaCl solution were refreshed every 2 h, and their exchange defined the completion of a cycle of mineralization. The dentin slice was cleaned ultrasonically for 2 min after each cycle. The sample was taken out after two, four, and six cycles for assessment and characterization. Characterizing the Dentin Slices after Remineralization. XRD and FTIR Evaluation of the Composition and Structure of the Remineralized Dentin Surfaces. The composition of the precipitates formed on an etched dentin slice after exposure to six cycles of mineralization was evaluated by X-ray diffraction (XRD; X’Pert Pro, Philips Almelo, Netherlands) and diffuse-reflectance Fourier transform infrared spectroscopy (DR-FTIR; Nicolet 8700, Thermo Scientific Instrument Co, Friars Drive Hudson, NH, USA). For comparison, dentin slices prepared without acid-etching and acid-etched dentin slices not subjected to electrophoresis were studied as controls. SEM and AFM Evaluation of the Morphology and Location of the Precipitated HA Crystals. The dentin slices were dehydrated with a series of ethanol treatments and were dried in a critical-point evaporator before being sputter-coated with gold for SEM analysis. The morphologies and locations of the precipitates were evaluated by field emission scanning electron microscopy (FE-SEM; Sirion 200, FEI Co, Hillsboro, OR, USA, or S4800, Hitachi High Technologies America, Inc., Dallas, USA). For atomic force microscope (AFM) analysis, fresh samples were used in a tapping model with etched silicon probe (Dimension Edge, Bruker, California, USA). TEM Evaluation of Intrafibrillar Mineralization Formation. The surfaces of the dentin slice were exposed to four cycles of mineralization; the acid-etched dentin slices that were not subjected to remineralization were scratched using a probe, and the crumbs collected were smeared onto a copper grid for analysis with transmission electron microscopy (TEM) (Tecnai G2 20, FEI Co., Hillsboro, OR, USA). Bright-field and selected area electron diffraction (SAED) modes were used without staining the samples prior to analysis.

Figure 2. XRD spectra of the precipitates formed on dentin after six cycles of mineralization. Precipitates formed (a, red line) with the aid of electrophoresis, (b, black line) on dentin not subjected to acid etching, and (c, blue line) on acid-etched dentin.

ratio of the diffraction intensity of the c-axis (002) reflection to the diffraction intensity of the a-axis (300) reflection for the electrophoresis-treated slice was considerably greater than that of the slice not exposed to acid-etching (spectrum b). This result suggests that the HA precipitates were oriented along the c-axis. Distinct diffraction peaks at approximately 2θ = 31.9°− 33.0° could be identified in the spectra, indicating the good crystallinity of the HA. These diffraction peaks were difficult to identify in the spectrum of acid-etched or untreated dentin specimens. These findings were consistent with the observations made with SEM analysis, as shown in Figure 4. To further study the crystallinity of the HA formed, we calculated the average crystallite size according to Scherrer’s formula: Dhkl = kλ/(β cos θ), where D is the crystallite size, k is the Scherrer constant (0.89), λ is the wavelength of X-ray beam (0.15406 nm), and β is defined as the diffraction full width at halfmaximum (fwhm), expressed in radians. The fwhm values of the peaks of the (002) plane are representative of the crystallites along the c-axis, the peak of the (300) plane is representative of the crystallites along the a-axis, and θ is the diffraction angle of the (002) and (300) reflections, with values of 12.9° and 16.5°, respectively. The crystallites were found to be approximately 45 nm in length and 30 nm in width. The size of these crystals (calculated using XRD spectra) was not comparable to the size of the crystals on the dentin surface, which were observed to be 125 nm in length and 70 nm in width by SEM imaging (Figure 4j). It is important to note that the results derived from Scherrer’s equation are the average crystallite sizes that correspond to diffractive planes. However, the sizes measured from SEM image analysis described crystals that were aggregates of many single crystals. FTIR analysis demonstrated a compositional change on the dentin surface after mineralization (Figure 3). The FTIR spectrum of the acid-etched dentin surface (c, blue line) was characterized by peaks corresponding to collagen; few peaks that signified the presence of HA were observed before mineralization. The peaks at approximately 1633 and 1528 cm−1 could be labeled amide I and amide II, respectively, and were contributed by the dentin collagen matrix. In contrast, the FTIR spectrum of the remineralized dentin surface was characterized predominantly by HA peaks (a, red line). Few collagen peaks were visible. A distinct −PO4 band in the FTIR spectra of the remineralized slices showed a remarkable splitting of the P−O asymmetric stretching bands (ν3) at 1050 (ν3‑1) cm−1 and 1110 (ν3‑2) cm−1 and of the P−O

■

RESULTS AND DISCUSSION Regeneration and Assembly of Dentin- and EnamelLike Tissues for Ideal Self-Healing of Dentin Defects Using the Biomimetic Mineralization Strategy. XRD spectra confirmed that the precipitates that formed on the remineralized dentin slices were HA. Spectrum (a) in Figure 2 shows the XRD patterns of the crystals that grew on the dentin surface after six cycles of electrophoresis-assisted mineralization in the agarose hydrogel microenvironment. The diffraction peaks (002) at 2θ = 25.8°, (211) at 2θ = 31.9°, and (300) at 2θ = 33.0° corresponded well to the expected peaks for HA. The 5539

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Perhaps the most probable charge compensation route has the following mechanism: Ca 2 + + PO4 3 − → Na 2 + + CO32 −

where Na+ substitutes for Ca2+ and CO32− substitutes for PO43−. However, other mechanisms may also occur to a much lesser extent; one of these is the loss of Ca2+ (creating a Ca deficiency) to compensate for CO32+ substitution. This may be represented as follows: Ca 2 + + 2PO4 3 − → VCa 2+ + 2CO32 −

where VCa2+ represents a vacancy in site occupied by Ca2+ ions.35 SEM evaluation of the surface of the acid-etched dentin slice showed that the dentin collagen fibers were demineralized and that the collagen matrix was exposed (Figure 4a). The dentinal tubules on the dentin surface became obvious after acid etching. Furthermore, the diameters of the dentinal tubules in the demineralized dentin were enlarged (Figure 4b). The zone of demineralization was approximately 5 μm in depth. In contrast, needle-shaped HA crystals were observed on the remineralized dentin slices. The needle-shaped HA crystals grew with time from the dentin surface and the wall of dentinal tubule and gradually covered the dentinal tubules. The HA crystals were densely packed together after six cycles of mineralization. At this stage, they completely covered the underlining dentin tissue and the dentinal tubules (Figures 4c−h). The densely packed crystals shared the same orientation, with their crystallographic c-axis parallel to each other, and were perpendicular to the dentin surface, forming an enamel-like structure (Figures 4i,j). AFM observations (Figure 5) corresponded well with those made based on SEM analysis. The surface of the acid-etched dentin slice showed open dentinal tubules (Figure 5a). Upon application of the mineralization process, the dentin surface was covered by HA particles, and the open dentinal tubules were gradually occluded (Figures 5b,c). Finally, the needle-shaped HA crystals formed, densely packed together and completely covering the dentin surface (Figures 5d). Viewing the transverse sections of the remineralized dentin slices, we observed that the dentinal tubules were gradually occluded by the precipitated HA crystals (panels d, f, and h in Figure 4 and a and b in Figure 6) and were completely occluded by the densely packed HA crystals after six cycles of mineralization. The precipitated HA crystals grew and became densely packed together to form bulk material. Some walls of the dentinal tubules were torn away during the process of preparing the transverse section samples. Dentin collagen fibers are normally embedded and protected by HA (Figure 7a). Typically, HA can be removed, and the dentin collagen fibers can be exposed after acid etching (Figure 7b). In this study, the demineralized dentin collagen matrix was remineralized, and the dentin microstructure of the calcified dentin collagen matrix was regenerated after only two cycles of mineralization (Figure 7c). Importantly, it resembled natural dentin. Furthermore, the demineralized dentin collagen matrix was remineralized with intrafibrillar and interfibrillar HA formation (Figure 7d,e,f). In some areas, the collagen fibers were remineralized with nano-HA particles precipitating along the collagen fibers to form a “string-of-beads” appearance, which corroborated intrafibrillar mineralization (black arrow in Figure 6d,f). Intrafibrillar mineralization was further confirmed

Figure 3. FTIR spectra of the precipitates formed on dentin after six cycles of mineralization. Precipitates (a, red line) arising from electrophoresis, (b, black line) formed on dentin not exposed to acid etching, and (c, blue line) formed on acid-etched dentin.

splitting bending bands (ν4) at 604 (ν4‑1) cm−1 and 569 (ν4‑2) cm−1. The intensities of the amide peaks corresponding to amide I and amide II were significantly reduced. However, in addition to the “regular” apatite spectral features characteristic of HA, the nanocrystalline apatites also exhibited supplementary bands that were due to the presence of nonapatitic ionic environments within the surface layer of the nanocrystals. In particular, a band due to nonapatitic, labile CO32− ions has been identified in the spectra, which appear at 872 cm−1. This corresponds to the presence of CO32− groups that have been substituted for phosphate groups in the apatite structure.33 This substitution is named a B-type, similar to what is observed in biological carbonate-HAP.34 In the FTIR spectrum of the dentin surface not subjected to acid etching (b, black line), the −PO4 band peaks present were relatively weak and vague in appearance compared with the remineralized samples. This suggested that the HA crystals precipitated on this sample were smaller in size and lower in crystallinity than the HA precipitated onto the remineralized samples. The observation also corresponded well with the XRD spectra. The intensities of the amide peaks were reduced in the spectrum of the dentin exposed to the mineralization system. Additionally, the intensity of the PO4 band was enhanced in this spectrum when compared with the spectrum of acid-etched dentin. These differences indicated that more HA was deposited on the collagen fiber surface after electrophoresis. SAED spectra further confirmed that demineralized collagen fibrins inside the dentin surface were calcified with nano-HA formed along the collagen fibrils after electrophoresis (Figure 9). Biological apatites are usually described as substituted calcium hydroxyapatite (HAP, Ca10(PO4)6(OH)2). The most common ion substitutions consist in the replacement of the original ions in the structure, namely, Ca2+, PO43−, and OH−. In our study, FTIR spectroscopy showed that carbonate ions tended to substitute for phosphate ions (B-type HAP). This particular substitution requires the replacement of the original ions by ions with different charges. This type of substitution proceeds by a charge compensation mechanism that seems to be especially adaptable in apatites due to the possibility of ion insertions and ion vacancies. In this experiment, we provided ample opportunity for such ion transfers by employing disodium hydrogen phosphate as a source of phosphate ions; in doing so, we also introduced many Na+ ions to our system. 5540

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 4. SEM micrographs of the remineralized dentin slices. (a, b) Dentin structures after acid-etching with 37% H3PO4 for 30 s. Remineralized samples after (c, d) two, (e, f) four, and (g, h) six mineralizing cycles. Panels b, d, f, and h are cross-sectional views of the samples shown in panels a, c, e, and g, respectively. Panel I shows the enamel after acid etching with 37% H3PO4 for 30 s and reveals HA crystals with their c-axis parallel to one another. Panel j provides a magnified view of panel g showing the precipitated HA crystals with orientations similar to those in panel i. The micrographs show that needle-shaped HA crystals grew from the dentin surface and the walls of the dentinal tubules, and gradually covered the dentinal surface and tubules to form enamel-like tissue.

Figure 5. AFM micrographs of the surfaces of the remineralized dentin slices. (a) Dentin slice after acid etching with 37% H3PO4 for 30 s. (b, c, d) Remineralized samples after two, four, and six mineralizing cycles, respectively. The left-hand images show two-dimisional photographs, and the ones on the right show the corresponding three-dimisional photographs.

by subsequent TEM examination. Interfibrillar HA crystals were also formed with substantial mineralization. The spaces between the collagen fibrils were occupied by nanocrystals embedded between the collagen fibrils. A “corn-on-the-cob” appearance was found in some regions and eventually made it difficult to identify collagen fibrils covered with interfibrillar HA formed (white arrow in Figure 6d,e,f). A tight junction was observed at the interface between the precipitates in the dentinal tubules and the wall of the dentinal tubules (Figure 6b and Figure 8). In the dentinal tubules, the precipitated HA particles were densely packed and adhered to the surface of the dentinal tubules. Thus, in some areas, the

walls of the dentinal tubules were torn away during the process of preparing the transverse sections of the samples (Figure 6b,d). The precipitates on the remineralized dentin surface and the substrate of the remineralized dentin intergraded together, such that the interface was hardly distinguishable (Figures 8b,d). Thus, we confirmed that the binding between the precipitated HA crystals and the dentin tissue was reliable. TEM bright-field micrographs showing remineralized dentin collagen fibrils after six cycles of mineralization are shown in Figure 9a. The fibrils showed a dark contrast, discrete electron5541

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 6. Dentin remineralization with occlusion of dentinal tubules. (a, b) Cross-sections of the dentin slices after two and six mineralization cycles, respectively. Panel c is a magnified area of panel b. Panel d shows the natural dentinal tubule structure for comparison. Part of the wall of the dentinal tubules was torn away during preparation of the transverse section samples, suggesting that the precipitated HA adhered firmly to the dentin structure (Figure 5b).

collagen fibers from degradation.13,14 Thus, by remineralizing the collagen fibrils to regenerate the dentin microstructure, one expects to also halt or reverse the initial disease processes that originally caused demineralization. In the present study, a demineralized collagen matrix was remineralized with the formation of intrafibrillar and interfibrillar HA crystallites, resulting in the duplication of a dentin microstructure containing calcified collagen fibrils (Figure 7). Second, exposing dentin to the oral environment often causes numerous pores, which are actually the openings of dentinal tubules exposed on the dentin surface. Local environmental changes in temperature, pressure, osmolality, and chemical agents cause the movement of fluids in the dentinal tubules. This stimulates the underlying nerves surrounding odontoblasts in the dental pulp and gives rise to pain (termed as dentin hypersensitivity), as postulated by the hydrodynamic theory. A common strategy for managing dentin hypersensitivity is to occlude the openings of the dentinal tubules using dentin bonding agents, desensitizing toothpastes, laser therapy, and even dental restoration.36 However, these treatments have considerable limitations.37 With this novel electrophoresis-aided biomineralization system, the HA precipitates almost completely occluded the dentinal tubules. In addition, the HA precipitates grew from the walls of the dentinal tubules and most likely firmly adhered to the dentinal tubules. The precipitates also have a crystal structure similar to that of the natural tooth and conceivably have a similar coefficient of expansion. This is a favorable factor that contributes to the longevity of the adhesion of the HA precipitate to dentin (Figure 6 and Figure 8).

dense islands, and areas where the collagen fibrils were difficult to distinguish. The samples were not stained with phosphotungstic acid or any other electron-dense substance, and thus, these observations suggest the collagen fibrils were highly mineralized. Intrafibrillar mineralization of the collagen fibrils, as well as interfibrillar mineralization of the matrix adjacent to the fibrils, could be found in the electron-dense images. The SAED pattern of the precipitates revealed discrete string-like patterns that were characteristic of HA (Figure 9c). The EDS spectrum shows the presence of calcium, phosphate, and oxygen in the remineralized collagen fibrils (Figure 9d). In contrast, the TEM micrograph of dentin collagen fibrin in demineralized dentin was hazy and indistinct (Figure 9e). There is no arch or ring image in the SAED pattern (Figure 9f). This study first provides a self-healing strategy using biomimetic mineralization to form a dentin- and enamel-like microstructure that can repair dentin defects and be used clinically as an alternative for managing dentin-related diseases (Figure 10). Our self-healing system remineralizes dentin collagen fibrils to duplicate the natural dentin microstructure, occludes dentinal tubules, and regenerates enamel-like tissue to cover the remineralized dentin surface. The significance of the self-healing system from the point of clinical use is stressed in the schematic of Figure 10. First, dentin caries, erosion, abrasion, and attrition contribute to tooth surface loss and cause dentin demineralization that exposes the dentin organic matrix to the oral environment, resulting in the breakdown of collagen fibrils. As mentioned above, the calcification of collagen fibrils and especially the intrafibrillar mineralization of collagen fibrils are responsible for the biomechanical properties of dentin and for protecting 5542

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 7. Cross-sectional images of dentin slices. (a) Natural dentin and (b) demineralized dentin collagen matrix after acid etching. (c) Remineralized dentin collagen matrix after two cycles of remineralization. (d, e, f) Remineralized collagen matrices, showing the formation of intrafibrillar and interfibrillar HA at high magnification. The demineralized dentin collagen matrix was remineralized with the formation of intrafibrillar (black arrow) and interfibrillar HA (white arrow) to regenerate the original dentin microstructure.

much attention has been paid to noncollagenous proteins (NCPs), which have a high affinity for calcium ions and collagen fibrils.40 Moreover, NCPs are responsible for regulating the nucleation and growth of the mineral phase in mineralized hard tissues. Many studies have focused on investigating the promoting effects of biomimetic analogs of NCPs on crystal nucleation and crystal growth.15,17,25,41,42 Several in vitro studies have demonstrated that ordered mineralization of apatite inside collagen fibrils is impossible without such biomimetic analog additives.43,44 On the other hand, some in vitro studies showed that type I collagen can initiate and orient the growth of carbonated apatite mineral in the absence of any other extracellular matrix molecules from vertebrate calcified tissues. Furthermore, the collagen matrix can influence the structural characteristics of the apatite at the atomic scale and control the size and the three-dimensional distribution of the apatite at larger length scales.44,45 In this

Third, natural dentin is covered by a layer of enamel that provides protection for the dentin−pulp complex against external stimuli or assault. In this study, we successfully regenerated a layer of enamel-like tissue that grew from the remineralized dentin collagen matrix. There are few studies in the literature reporting the regeneration of enamel-like tissue from the remineralized dentin collagen matrix (Figure 4g,h,j). The Contribution of Electric Field to the Regeneration of Tooth-Like Tissue Structure in the Agarose Hydrogel Model. Collagen fibrils are the basic building blocks of mineralized dentin. It is therefore essential to understand their role in biomineralization. However, the function of collagen in the mechanism of hard tissue biomineralization remains unclear. Studies have reported that the collagen matrix can serve as a scaffold for crystal deposition, although it does not have the capacity to induce matrix-specific mineral formation from metastable calcium phosphate solutions.38,39 As a result, 5543

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 8. SEM micrographs showing the interface between precipitates and the walls of the dentinal tubules of a remineralized dentin slice. (a) Remineralized dentin in cross-section. (b) A magnified region of panel a. A magnified view of (c) the rectangular area and (d) the oval area in panel b. The five-point star indicates remineralized dentin collagen matrix, the three-point star indicates precipitates in the dentinal tubules, and the fourpoint star indicates precipitates on the remineralized dentin collagen matrix surface. The interface of the HA precipitates and the underlying remineralized dentin was hardly distinguishable (b, d), suggesting tight and strong binding of the precipitated HA crystals to dentin.

First, hard tissue mineralization occurs naturally under a unique gel-like organic matrix. To mimic the gel-like microenvironment, a hydrogel-based biomimetic mineralization model using agarose gel loaded with calcium and phosphate ions was developed. The mode of crystal growth is different than that in aqueous solutions in this gel-like microenvironment. The physicochemical nature of the gel-like microenvironment more realistically mimics the unique mineralized tissue matrix environment than do aqueous solutions. Second, the applied electric field plays a critical role in promoting bioremineralization. The generated current accelerates the speed of calcium and phosphate ion migration in the agarose hydrogel. Moreover, it can impose a specific directionality to the ion migration. Calcium and phosphate can thus be directed to flow easily in hydrogel.31,32,46 In this electrophoresis-aided biomineralization system, the rate of mineral deposition on a dentin slice is faster than it is in a traditional diffusion system. A “cloud” of white suspension was formed in the agarose hydrogel after the current was applied (Figure 11) that became larger and denser with time, and reached its maximum size after 2 h. In this study, the hydrogel was refreshed every 2 h, and this recharge defined the end of a single cycle of mineralization. Watanabe and Akashi found that electrophoresis could be used to precipitate HA in an agarose gel to obtain an HA/ agarose composite and that complete mineral formation was

study, a novel electrophoresis-aided biomineralization system was successfully developed to induce dentin matrix remineralization without the addition of biomimetic NCP analogs. The results of SEM analysis showed that the precipitated particles were regularly and homogeneously distributed along the collagen fibrils in a “string-of-beads” structure throughout the dentin collagen matrix. This observation suggested the intrafibrillar mineralization of the dentin collagen matrix. Intrafibrillar mineralization could also be identified by TEM (Figure 9). Furthermore, substantial mineral growth was observed, exhibiting a “‘corn-on-the-cob”’ appearance due to the connections between the remineralized collagen fibrils. This pattern suggested interfibrillar mineralization. It is suggested that the products of intrafibrillar mineralization might act as apatite seed crystallites that facilitate the growth of nanocrystals along the collagen fibril, resulting in ongoing mineral buildup on the exterior and the formation of interfibrillar mineralization. The HA nanocrystals then grew and connected to the mineralized intrafibrillar collagen fibrils. Our previous study showed that agarose gel loaded with calcium and phosphate ions alone was not able to induce interfibrillar mineralization.7 Therefore, the function of the electric current in inducing the formation of interfibrillar mineralization is important in this electrophoresis-aided biomineralization system. There are several advantages of the electrophoresis-aided biomineralization system. 5544

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 9. TEM micrographs of unstained remineralized dentin collagen fibrils. (a) TEM bright-field micrograph of remineralized dentin collagen fibrils, and (b) a magnified view of panel a. (c) SAED pattern of remineralized dentin collagen fibrils and (d) the EDS spectra of the remineralized fibrils. (e) TEM bright-field micrograph of demineralized dentin collagen fibrils of acid-etched dentin. (f) SAED pattern of the demineralized dentin collagen fibrils.

achieved in only 30 min when the electrophoresis was performed at 100 V. They also demonstrated that the precipitation of HA was accelerated by electrophoresis. The linear velocity in the mineralizing areas was shown to be approximately 15 times greater than that caused by simple diffusion.31 In the present study, the remineralizing speed was much improved compared with diffusion. In a previous report, remineralization of 5 μm demineralized dentin collagen matrix required several weeks, whereas in our protocol, it only

required 12 h. The electric current in the hydrogel enhanced the movement of calcium and phosphate ions or mineral precursors to “get in touch with” the collagen matrix. Therefore, the electric current facilitates not only intrafibrillar and interfibrillar mineralization but also the remineralization of the peritubular and intertubular dentin collagen matrix. This may be the reason for that intrafibrillar and interfibrillar mineralization can successfully be induced without the addition of NCPs or NCP analogs. In this study, our results corroborate 5545

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 10. A schematic diagram of the natural tooth structure and the regeneration of a prismatic enamel-like tissue on demineralized dentin.

allows the diffusion of the crystallization solution in and out of the scaffold, and the pores act as compartmentalized areas that regulate the nucleation and growth of the mineral phase in the mineralization process. The properties of agarose hydrogel also make it favorable as a biomimetic mineralization template.47 The restricted space in the gel network may confer a uniform and controllable size and structure to the precipitated HA. Kamitakahara et al. reported that the size of HA granules precipitated in this context could be controlled by the concentration of the hydrogel present.46 Our previous study demonstrated that nanoscale complex amorphous calcium phosphate precursors could be formed in the agarose hydrogel.48 In the present study, the unique phenomena of HA crystal growth, transformation, and assembly were investigated inside the demineralized dentin, at the dentin− gel interface and inside the agarose hydrogel. Figure 12 shows the mineral morphology at the interface and inside the hydrogel. The mineral morphology and crystallinity transformed gradually, moving from the interface between the dentin slice and the hydrogel to the inside of the hydrogel. At the interface, the mineral formed ball-like particles with sizes of more than 20 nm. However, the minerals inside the hydrogel were rod-like and needle-like in shape, and their sizes were larger than the particles at the interface. These findings were different from those of our previous study.44 These transformations, then, might be considered a result of the applied electric field. The minerals formed at the interface, with a smaller size and lower crystallinity, additionally, might facilitate the phase transformation to promote dentin collagen remineralization. After the demineralized collagen fibrils were remineralized, HA crystals subsequently nucleated and gradually grew on the calcified collagen fibrils with their c-axis aligned parallel to each other and perpendicular to the surface of the dentin slices. With crystal growth and assembly, a prism-like enamel structure formed (Figures 4g,h,j). On the other hand, it is predicted by a universal rule that crystalline interfacial water layers played a fundamental role in the formation of early planetary life forms.49 Specifically, this theory suggests that life could have started with crystalline water layers inducing order in prebiotic molecules present on solid surfaces. One way of imposing order into the biomineralization processes can be explained by assuming the interplay of ordered water molecules that use the polymer chain as a template for inducing the hard tissue structure.49 The Novelty and Advantage of the Biomimetic Mineralizing System. Using the mineralization system presented here, we are able to regenerate a dentin microstructure containing a calcified collagen matrix, occlude exposed

Figure 11. Deposits formed on the dentin slice with time.

that electrophoresis fosters the induction of intrafibrillar and interfibrillar mineralization and accelerates the speed of remineralization. It should also be noted that the precipitation of HA was uneven on the dentin surface and that the rate of precipitation varied at different areas of the dentin slice. This might be due to the variation of the electric resistance in different areas of the dentin slice and, thus, the variation in electric current at different points on the slice. As a result, the concentrations of calcium and phosphate were not uniform over the entire slice. A pilot study was performed in this electrophoresis-aided calcium and phosphate hydrogel system, and the results showed that 20 mA is adequate for electrophoresis to occur. Increasing the amount of electric current applied in the system had no significant effect on either the rate of HA precipitation or the morphology of the precipitates. The Mechanism of HA Crystal Growth and Assembly on Dentin Surface in the Agarose Hydrogel Environment Aided by Electrophoresis. Agarose hydrogel mimics the gel-like microenvironment of natural tooth formation and also acts as a biomimetic mineralization template. Numerous pores with diameters in the nanometer to micrometer range are present in the hydrogel. The porous structure of the hydrogel 5546

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

Article

Figure 12. Mineral composites formed at the interface between the dentin slice and the agarose hydrogel. (a) Cross-section of the calcium phosphate-loaded hydrogel at the interface between the dentin slice and agarose hydrogel. (b) Mineral morphology at the interface between dentin slice and agarose hydrogel (the inset shows a magnified portion). (c) Magnified image of the rectangle-marked area in panel a. (d) Magnification of the oval-marked area in panel a.

dentinal tubules, and create a layer of enamel-like tissue to cap the remineralized dentin surface, and furthermore, we are able to induce assembly into a natural tooth structure. Although other studies have reported the regeneration of enamel-like tissue or a calcified collagen matrix that mimics dentin microstructure, few have reported the ability to regenerate both types of tissue, much less their spontaneous assembly into a tooth-like structure, using only one mineralization system. Therefore, this methodology provides a valuable potential pathway for the self-healing of tooth defects. Moreover, the electric current used in this novel electrophoresis-aided biomineralization system shortened the time required for tooth-like tissue regeneration. This system accelerated the mineralization speed by many times over the previously reported system. In future clinical use, we can conceive of assembling the hydrogel layers on the tooth surface, then connecting the electrodes through an iontophoresis device.50 Further studies are being carried out to probe the utilization of this system for the clinical management of dentin-related disease. It is true that from a fundamental perspective, electrophoresis techniques have been used in mineralization studies in gel matrices to form HA-poly hybrid materials. However, the subtleties of our study make it different from others presenting such research. In these previous cases, electrophoresis was employed to promote calcium and phosphate ion diffusion in the agarose/gelatin hydrogel to synthesize the agarose/gelatin-HA composites. In our study, we used the hydrogel as a polymer template to induce mineralization on a third material (dentin slice) and aided this process by electrophoresis.

■

CONCLUSION

■

AUTHOR INFORMATION

A novel system utilizing electrophoresis to facilitate a calciumand phosphate-loaded agarose hydrogel biomineralization system was successfully developed to remineralize dentin defects much more quickly than previously reported remineralization systems. Thus, this work represents a self-healing system for biomimetically reforming a tooth-like structure in damaged teeth. The system can induce the remineralization of the dentin collagen matrix by the formation of intrafibrillar and interfibrillar HA crystallites to generate a dentin microstructure of calcified collagen fibrils that occlude dentinal tubules and eventually regenerate enamel-like tissue on the dentin surface. Therefore, the system may be a good candidate for developing a self-healing treatment system for dentin defects for use in the future.

Corresponding Authors

*Q.-L. Li. E-mail: [email protected]. Tel: (+)86-551-65118677. Fax: (+)86-551-65111538. *C. H. Chu. E-mail: [email protected]. Tel: (+)852-28590287. Fax: (+)852-28587874. Author Contributions

Xiao-Ting Wu and Ying Cao contributed equally. Notes

The authors declare no competing financial interest. 5547

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548

Crystal Growth & Design

■

ACKNOWLEDGMENTS

■

REFERENCES

Article

(32) Watanabe, J.; Kashii, M.; Hirao, M.; Oka, K.; Sugamoto, K.; Yoshikawa, H.; Akashi, M. J. Biomed. Mater. Res., Part A 2007, 83, 845−52. (33) Nelson, D.; Featherstone, J. Calcif. Tissue Int. 1981, 34, S69−81. (34) Zapanta LeGeros, R. Prog. Cryst. Growth Charact. 1981, 4, 1−45. (35) Mostafa, N. Y.; Hassan, H. M.; Abd Elkader, O. H. J. Am. Ceram. Soc. 2011, 94, 1584−1590. (36) Chu, C.; Lam, A.; Lo, E. Gen. Dent. 2011, 59, 115. (37) Bakry, A. S.; Tamura, Y.; Otsuki, M.; Kasugai, S.; Ohya, K.; Tagami, J. J. Dent. 2011, 39, 599−603. (38) Saito, T.; Arsenault, A.; Yamauchi, M.; Kuboki, Y.; Crenshaw, M. Bone 1997, 21, 305−311. (39) Saito, T.; Yamauchi, M.; Abiko, Y.; Matsuda, K.; Crenshaw, M. A. J. Bone Miner. Res. 2000, 15, 1615−1619. (40) Zhang, S.; Gangal, G.; Uludag, H. Chem. Soc. Rev. 2007, 36, 507−531. (41) Kim, J.; Arola, D. D.; Gu, L.; Kim, Y. K.; Mai, S.; Liu, Y.; Pashley, D. H.; Tay, F. R. Acta Biomater. 2010, 6, 2740−2750. (42) Kim, Y. K.; Gu, L. S.; Bryan, T. E.; Kim, J. R.; Chen, L.; Liu, Y.; Yoon, J. C.; Breschi, L.; Pashley, D. H.; Tay, F. R. Biomaterials 2010, 31, 6618−6627. (43) Nudelman, F.; Pieterse, K.; George, A.; Bomans, P. H.; Friedrich, H.; Brylka, L. J.; Hilbers, P. A.; Sommerdijk, N. A. Nat. Mater. 2010, 9, 1004−1009. (44) Silver, F. H.; Landis, W. J. Connect. Tissue Res. 2011, 52, 242− 254. (45) Wang, Y.; Azaïs, T.; Robin, M.; Vallée, A.; Catania, C.; Legriel, P.; Pehau-Arnaudet, G.; Babonneau, F.; Giraud-Guille, M.-M.; Nassif, N. Nat. Mater. 2012, 11, 724−733. (46) Kamitakahara, M.; Kimura, K.; Ioku, K. Colloids Surf., B 2012, 97, 236−239. (47) Munro, N. H.; McGrath, K. M. Bioinspired, Biomimetic Nanobiomater. 2012, 1, 26−37. (48) Guentsch, A.; Busch, S.; Seidler, K.; Kraft, U.; Nietzsche, S.; Preshaw, P. M.; Chromik, J. N.; Glockmann, E.; Jandt, K. D.; Sigusch, B. W. Adv. Eng. Mater. 2010, 12, B571−B576. (49) Sommer, A. P.; Zhu, D.; Fecht, H.-J. Cryst. Growth Des. 2008, 8, 2628−2629. (50) Kim, H.; Kwon, H.; Kim, B. J. Oral Rehabil. 2009, 36, 770−775.

This project was supported by NSFC/RGC Joint Research Scheme (No. 81061160511 and N_HKU 776/10) and NSFC Grant (No. 30973352).

(1) Chen, H.; Tang, Z.; Liu, J.; Sun, K.; Chang, S. R.; Peters, M. C.; Mansfield, J. F.; Czajka-Jakubowska, A.; Clarkson, B. H. Adv. Mater. 2006, 18, 1846−1851. (2) Yamagishi, K.; Onuma, K.; Suzuki, T.; Okada, F.; Tagami, J.; Otsuki, M.; Senawangse, P. Nature 2005, 433, 819. (3) Chen, H.; Clarkson, B. H.; Sun, K.; Mansfield, J. F. J. Colloid Interface Sci. 2005, 288, 97−103. (4) Fowler, C. E.; Li, M.; Mann, S.; Margolis, H. C. J. Mater. Chem. 2005, 15, 3317−3325. (5) Xie, R.; Feng, Z.; Li, S.; Xu, B. Cryst. Growth. Des. 2011, 11, 5206−5214. (6) Li, L.; Mao, C.; Wang, J.; Xu, X.; Pan, H.; Deng, Y.; Gu, X.; Tang, R. Adv. Mater. 2011, 23, 4695−4701. (7) Fan, Y.; Sun, Z.; Moradian-Oldak, J. Biomaterials 2009, 30, 478− 483. (8) Fan, Y.; Wen, Z. T.; Liao, S.; Lallier, T.; Hagan, J. L.; Twomley, J. T.; Zhang, J.-F.; Sun, Z.; Xu, X. J. Bioact. Compat. Polym. 2012, 27, 585−603. (9) Ruan, Q.; Zhang, Y.; Yang, X.; Nutt, S.; Moradian-Oldak, J. Acta. Biomater. 2013, 9, 7289−7297. (10) Wang, L.; Guan, X.; Yin, H.; Moradian-Oldak, J.; Nancollas, G. H. J. Phys. Chem. C 2008, 112, 5892−5899. (11) Busch, S. Angew. Chem., Int. Ed. 2004, 43, 1428−1431. (12) Liu, S.; Yin, Y.; Chen, H. CrystEngComm 2013, 15, 5853−5859. (13) Balooch, M.; Habelitz, S.; Kinney, J.; Marshall, S.; Marshall, G. J. Struct. Biol. 2008, 162, 404−410. (14) Carrilho, M. R. d. O.; Tay, F. R.; Pashley, D. H.; Tjäderhane, L.; Marins Carvalho, R. Dent. Mater. 2005, 21, 232−241. (15) Tay, F. R.; Pashley, D. H. Biomaterials 2008, 29, 1127−1137. (16) He, G.; Dahl, T.; Veis, A.; George, A. Connect. Tissue Res. 2003, 44, 240−245. (17) He, G.; George, A. J. Biol. Chem. 2004, 279, 11649−11656. (18) He, G.; Dahl, T.; Veis, A.; George, A. Nat. Mater. 2003, 2, 552− 558. (19) Rahiotis, C.; Vougiouklakis, G. J. Dent. 2007, 35, 695−698. (20) Rahiotis, C.; Vougiouklakis, G.; Eliades, G. J. Dent. 2008, 36, 272−280. (21) Mitchell, J. C.; Musanje, L.; Ferracane, J. L. Dent. Mater. 2011, 27, 386−393. (22) Zhou, Y. Z.; Cao, Y.; Liu, W.; Chu, C. H.; Li, Q. L. ACS Appl.Mater. Interfaces 2012, 4, 6901−6910. (23) Ning, T. Y.; Xu, X. H.; Zhu, L. F.; Zhu, X. P.; Chu, C. H.; Liu, L. K.; Li, Q. L. J. Biomed. Mater. Res., Part B 2012, 100, 138−144. (24) Niu, L. N.; Zhang, W.; Pashley, D. H.; Breschi, L.; Mao, J.; Chen, J. H.; Tay, F. R. Dent. Mater. 2014, 30, 77−96. (25) Cao, Y.; Mei, M. L.; Xu, J.; Lo, E. C.; Li, Q.; Chu, C. H. J. Dent. 2013, 41, 818−825. (26) Cao, Y.; Liu, W.; Ning, T.; Mei, M. L.; Li, Q.-L.; Lo, E. C.; Chu, C. Clin. Oral Invest. 2013, 1−9. (27) Wang, Q.; Wang, X.; Tian, L.; Cheng, Z.; Cui, F. Soft Matter 2011, 7, 9673−9680. (28) Burwell, A. K.; Thula-Mata, T.; Gower, L. B.; Habelitz, S.; Kurylo, M.; Ho, S. P.; Chien, Y. C.; Cheng, J.; Cheng, N. F.; Gansky, S. A.; Marshall, S. J.; Marshall, G. W. PloS One 2012, 7, No. e38852. (29) Li, J.; Yang, J.; Li, J.; Chen, L.; Liang, K.; Wu, W.; Chen, X.; Li, J. Biomaterials 2013, 34, 6738−6747. (30) Cao, Y.; Mei, M. L.; Li, Q. L.; Lo, E. C.; Chu, C. H. ACS Appl.Mater. Interfaces 2014, 6, 410−420. (31) Watanabe, J.; Akashi, M. Biomacromolecules 2006, 7, 3008− 3011. 5548

dx.doi.org/10.1021/cg5007633 | Cryst. Growth Des. 2014, 14, 5537−5548