Fabrication and in vitro biological performance of a double-layered nanoparticles-microarc oxidation composite coating on titanium for dental implant application

Sufficient residual alveolar bone volume plays an important role in the success rate and service life of dental implants. However, alveolar bone deficiency is a common clinical phenomenon, and the alveolar bone would be further absorbed by peri-implant infection. Therefore, it is highly desirable to promote peri-implant alveolar bone regeneration and inhibit alveolar bone resorption when the alveolar bone mass is insufficient. For this purpose, a pH-sensitive double-layered nanoparticles-microarc oxidation (MAO) composite coating was fabricated on titanium for dental implant application in this study. The pH-sensitive double-layered nanoparticles were prepared by a poly(L-lactic acid) inner layer and a chitosan outer shell, containing stromal-cell derived factor-1, recombinant human bone morphogenetic protein 2 and osteoprotegerin. The composite coating was fabricated on MAO coating by cross-linking the pH-sensitive double-layered nanoparticles with gelatin. The surface morphology of the composite coating showed that the pH-sensitive double-layered nanoparticles were well distributed and tightly cross-linked in the pores of MAO coating. The composite coating could sustain release the three drugs for more than 30 days. With decreasing pH, the release of osteoprotegerin from the composite coating increased (p < 0.05). In vitro biological studies suggested that the composite coating exhibited no cytotoxicity, and can recruit bone marrow-derived mesenchymal stem cells (BMSCs), promote BMSC differentiation into osteoblasts, and inhibit osteoclast generation. Moreover, with decreasing pH, the inhibitory effect on osteoclast generation was enhanced (p < 0.05). It can be concluded that the fabricated composite coating, which can promote bone regeneration and inhibit bone resorption, has the potential to be applied on the surface of dental implant, especially when the residual alveolar bone is in poor condition.


Introduction
Adequate alveolar bone mass is a key factor for the long-term survival and success of dental implants.However, alveolar bone deficiency at implant sites due to periodontal disease, trauma, surgical resection of tumors and other systemic reasons is common clinically.Therefore, numerous procedures including sinus floor elevation, guided bone regeneration, block bone grafting, and ridge splitting are introduced under these conditions [1].In addition, dental implants are susceptible to infection which may result in the loss of supporting alveolar bone [2].To address these issues, attention has been closely paid to the approaches which can promote peri-implant bone regeneration and inhibit bone resorption.
Bone tissue in the body will constantly remodel itself.During this process, osteoblasts that coordinate bone formation and osteoclasts that coordinate bone resorption synergistically regulate the precise reconstruction of bone [3].The application of stem cells for bone regeneration has been proposed and studied.Among these stem cells, bone marrow mesenchymal stem cells (BMSCs) play a crucial role in new bone formation [4].Because BMSCs can be induced into osteoblasts and form bone, they have been successfully used for bone repair or regeneration in experimental and clinical studies [5].Therefore, recruiting enough endogenous BMSCs to the defect area and induce the differentiation of BMSCs into osteoblasts are the key steps of bone regeneration.
Increasing evidences have suggested that stromal-cell derived factor-1 (SDF-1)/C-X-C chemokine receptor 4 (CXCR4) axis plays an important role in BMSC recruitment to the lesion site [6].SDF-1 is a small chemotactic signaling protein, which dictates the migration and activation of other cells.As a specific receptor of SDF-1, CXCR4 is expressed on various stem cells, including BMSCs and could direct cell migration by binding SDF-1 [7].
Bone morphogenetic proteins (BMPs) are one kind of bone growth factors which could induce BMSC differentiation into osteoblasts.Among them, bone morphogenetic protein 2 (BMP-2) plays a crucial role in osteoblast differentiation, and it is considered as a key regulator in osteogenesis [8].In particular, recombinant human bone morphogenetic protein 2 (rhBMP-2) is a commonly used osteogenic material for the treatment of bone defects, fracture nonunion and spinal fusion [9].Gomes-Ferreira reports that rhBMP-2 could accelerate the process of guided bone regeneration in cases of peri-implant bone defects [10].
Based on the above effects of chemokine SDF-1 and growth factor BMP-2, it is reasonable to hypothesize that the dual application of SDF-1 and BMP-2 may synergistically improve bone regeneration by activating BMSC recruitment to the lesion site and enhancing osteoblast differentiation [11].Wang demonstrated that the slow release of SDF-1 and BMP-2 from the scaffold significantly promoted bone regeneration [12].And improved osteoblast differentiation was observed by Hwang with the combined use of SDF-1 and BMP-2 [13].
Receptor activator of nuclear factor NF-κB (RANK), receptor activator of nuclear factor NF-κB ligand (RANKL) and osteoprotegerin (OPG) have received more attention for their significant roles in bone remodeling in the past few decades [14].RANK, expressed by osteoclasts and their precursors, interacts with its ligand RANKL to regulate the bone resorption by controlling osteoclast differentiation, activation and proliferation [15].OPG, secreted by osteoblasts, has a strong affinity for RANKL.Therefore, as a decoy receptor for RANKL, OPG inhibits the interaction between RANKL and RANK, thereby preventing bone resorption [16].
In our previous study, we have prepared a kind of pH-sensitive double-layered nanoparticles containing SDF-1, rhBMP-2 and OPG.The release study showed that the nanoparticles could sustain release the three drugs for over 30 days, and the released OPG increased significantly with decreasing pH [17].If the pH-sensitive double-layered nanoparticles could be applied on the surface of dental implants, it may have the effect on promoting peri-implant bone regeneration and inhibiting bone resorption in the case of poor alveolar bone condition.
Microarc oxidation (MAO) is an effective and common used technique for depositing ceramic coatings on titanium implants.Previous studies have confirmed that the MAO coating usually exhibits good interfacial bonding with the substrate, and the biocompatibility of the coating are significantly enhanced [18].And multiple desired elements (such as P and Ca) could be introduced into the MAO coating, which would improve cell adhesion and proliferation [19].Moreover, this technique can form a porous structure and then produce various composite coatings.Minocycline-loaded nanomicelles were incorporated into the pores of MAO treated titanium (MAO-Ti), and the composite coating showed excellent antibacterial ability [20].Zhou fabricated a composite dopamine-nanoclay coating on MAO-Ti by immersion method, which could significantly stimulate osteoblast adhesion, proliferation, and osteogenic differentiation [21].A bioactive silicon/copper-MAO composite coating was fabricated by doping silicon and copper ions into MAO coating, and the coating presented good antibacterial and osteogenic properties [22].
We also have developed a method to cross-link the drug-loaded nanoparticles into the porous structure on MAO coating [23].Thus, the purpose of this study was to fabricate a novel composite coating with the prepared pH-sensitive double-layered nanoparticles on MAO coating and evaluate the biological performance of the composite coating in vitro, so as to provide a new method which could be applied on the surface of dental implants.
2.2.Fabrication and release behavior of the composite coating 2.2.1.Preparation and characterization of the pH-sensitive double-layered nanoparticles The method for preparing the pH-sensitive double-layered nanoparticles containing SDF-1, rhBMP-2 and OPG was described detailedly in our previous study [17].Briefly, rhBMP-2-loaded PLLA nanoparticles (the inner core) were prepared by the double emulsion method (W/O/W) with PLLA.OPG was conjugated to the CS backbone with a pH-sensitive hydrazone (Hz) bond, and then the CS-Hz-OPG was synthesized.20 mg CS-Hz-OPG, 0.2 g rhBMP-2-loaded PLLA nanoparticles and 1 μg SDF-1 were sonicated and dispersed in 10 ml deionized water.Subsequently, 1 ml of 1% sodium tripolyphosphate solution was dripped ultrasonically for 3 min.The mixture suspension was stirred at room temperature for 12 h.Finally, the pH-sensitive doublelayered nanoparticles were washed, dried under vacuum.
10 mg of the pH-sensitive double-layered nanoparticles was sonicated in 10 ml distilled water to form a nanoparticle suspension.A particle analyzer (Delsa TM Nano C, Beckman Coulter, USA) was used to assess the nanoparticle size and polydispersity index.An adequate amount of the nanoparticle suspension was dripped on a clean coverslip, and then the coverslip was dried and coated with a gold film.The surface morphology of the nanoparticles was observed by scanning electron microscopy (SEM, S-4800, Hitachi, Japan).In addition, an adequate amount of the nanoparticle suspension was dripped on the copper omentum.After drying, the internal structure of the double-layered nanoparticles was observed by transmission electron microscopy (TEM, FEI Tecnai G2 Spirit, USA).

Preparation of polished cp-Ti specimens and MAO treatment
In this study, disk-shaped (diameter = 14.5 mm, thickness = 2 mm) cp-Ti specimens were used.The polishing process involved mechanical grinding with various grades of SiC paper from 80 # to 1200 # , followed by being ultrasonic cleaned in ethanol, rinsed with water and air-dried.In an electrolytic bath containing 0.04 M βglycerophosphate sodium and 0.2 M calcium acetate, the cp-Ti specimens were used as anodes and stainless steel plates were used as cathodes, and then the MAO treatment was operated by a pulsed direct current power supply (Xi'an University of Technology, Xi'an, China).The parameters of MAO treatment were applied voltage 300 V, frequency 600 Hz, duty cycle 8.0% and oxidizing time 5 min, respectively.The MAO-Ti specimens were then rinsed with distilled water and air-dried.The surface morphologies of the polished cp-Ti and MAO-Ti specimens were observed by SEM.

Fabrication of the composite coating
The composite coating was fabricated by cross-linking the pH-sensitive double-layered nanoparticles with gelatin on MAO-Ti specimens [23].In brief, 60 mg pH-sensitive double-layered nanoparticles were sonicated and dispersed in 1 ml 0.1% gelatin solution (w/v), from which 0.15 ml nanoparticle solution was added dropwise on one MAO-Ti specimen.Then the specimens were oscillated on an oscillator for one hour to render the nanoparticles penetrate into the pores on the specimens.Subsequently, the specimens were dried at 4 °C, followed by being immersed in 2.5% (w/v) glutaraldehyde solution for 30 min for cross-linking the gelatin.The residual glutaraldehyde on the specimens was removed by washing with ethanol three times.Finally, they were freeze-dried and sterilized by exposure to 60 cobalt radiation of 20 kGy.The composite coating specimen was also coated with a gold film, then observed by SEM.

In vitro drug release study
The drug release study was performed by placing one composite coating specimen into a dialysis bag with 5 ml phosphate buffered saline (PBS).This bag was then placed into 100 ml PBS (pH values were adjusted to 7.4, 6.8, and 6.2 respectively) and shaken at 100 rpm at 37 °C.At the scheduled time, 1 ml of PBS was extracted from the outside of the dialysis bag, and the same volume of fresh buffer was replaced.The amounts of SDF-1, rhBMP-2 and OPG released from the composite coating were determined by ELISA according to the instructions of ELISA kits.Meanwhile, the composite coating specimens immersed in the PBS were rinsed with distilled water for three times, dried, coated with a gold film, and observed by SEM at scheduled time to investigate the process of degradation.

2.3.
In vitro biological studies of the composite coating 2.3.1.Test groups Each test in the biological studies was divided into three groups: A: composite coating, B: MAO-Ti, and C: polished cp-Ti.All the specimens were 14.5 mm in diameter and 2 mm in thickness.

Culture of mouse BMSCs
According to the established protocol, mouse BMSCs were isolated and cultured [24].Typically, the femurs and tibias of 6-week-old C57BL6 mice (Animal testing center, School of Stomatology, Fourth Military Medical University) were dissected, and the bone marrow was rinsed with DPBS.Bone marrow-derived cells were filtered with a 70 μm cell filter (BD Falcon, USA), centrifuged at 300× g for 10 min, and suspended in 1 ml DMEM containing 1.5 mg ml −1 D-glucose supplemented with 1% penicillin, 1% glutamine and 10% FBS (BMSC medium).The cells were inoculated in a 25 cm 2 flask at a density of 1.6 × 10 6 cells/cm 2 and cultured at 37 °C in a humidified incubator with 5% CO 2 .Then the non-adherent cells were removed by washing with DPBS twice, and 5 ml fresh BMSC medium was added after 1 day or 4 days.The cells were seeded in BMSC medium at a density of 1 × 10 4 cells/cm 2 for subculture, and the medium was refreshed every 4 days.

Cytotoxicity assay
The possible cytotoxicities of the specimens were carried out by CCK-8 assay.The leaching solutions of each group were prepared using DMEM media as the extraction medium with the ratio of 1.0 ml cm −2 at 37 °C in a humidified incubator with 5% CO 2 for 72 h.Fresh DMEM media was applied as negative control and DMEM media containing 0.64% phenol was set up as positive control.BMSCs were inoculated in a 96-well plate at a density of 2 × 10 4 cells/well and cultured in BMSC medium at 37 °C in a humidified atmosphere of 5% CO 2 for 24 h to allow attachment.The medium was replaced with different leaching solutions and cultured for 1 day, 4 days and 7 days.Then 10 μl CCK-8 reagent was added to each well, and the plate was incubated for 4 h to induce nasal production.The optical density (OD) was determined by enzyme immunoassay analyzer (BioTek Instruments, Winooski, VT, USA) at 450 nm.The relative growth rate (RGR) of BMSCs were calculated using the following equation: RGR (%) = Optical density of experimental group / Optical density of negative control × 100%.

BMSC recruitment assay
Transwell method was used to assess BMSC recruitment.Briefly, the upper and lower compartments of the Transwell compartments were separated with a microporous membrane of 8.0 μm (BD Falcon, USA), in which BMSC medium (pH values were adjusted to 7.4, 6.8, and 6.2 respectively) was added.BMSCs in a logarithmic growth phase were seeded in the upper compartment at a density of 2 × 10 4 cells/ml.One specimen from each group was added to the lower compartment.Then the Transwell compartments were incubated in a humidified atmosphere of 5% CO 2 at 37 °C for 8 h.The BMSCs on the underside of the microporous membrane in the lower compartment were fixed, stained and finally counted under inverted microscope (Olympus, Japan).

BMSC differentiation assay
The BMSC differentiation into osteoblasts was evaluated by detecting the activity of ALP.BMSCs were inoculated in a 96-well plate at a density of 2 × 10 4 cells/well and cultured in BMSC medium (pH values were adjusted to 7.4, 6.8, and 6.2 respectively) at 37 °C in a humidified atmosphere of 5% CO 2 .One specimen from each group was added for co-culturing.After 7-day culture, the medium was discarded and the specimens were lightly rinsed with PBS for three times.The cells on the specimens were collected, and 100 μl cell lysate was transferred to a 96-well plate after 4 freeze-thaw cycles.150 μl ALP determination solution was added and incubated for 30 min at 37 °C.Next, the reaction was terminated by adding 50 μl NaOH (1 mol/l).And the optical density was measured at 405 nm.Meanwhile, after 7-day culture, the cells on the specimens were fixed, dyed with an ALP dye kit and observed under stereo microscope (Olympus, Japan).

Inhibiting osteoclast generation
Osteoclasts were differentiated from bone marrow-derived monocytes (BMMs) as described [25].Briefly, BMMs were isolated from tibiae and femurs from 6-week-old C57BL6 mice.The cells were seeded at a density of 2 × 10 4 cells/well in a 96-well plate.Then they were cultured for 6 days in α-MEM containing 10% FBS, 10 ng ml −1 RANKL and 10 ng ml −1 M −1 -CSF (osteoclast medium) in a humid atmosphere with 5% CO 2 at 37 °C.
The pH values of the osteoclast medium were adjusted to 7.4, 6.8 and 6.2 respectively, and one specimen from each group was added for co-culturing.The medium was discarded after 6-day culture and the specimens were lightly rinsed with PBS for three times.The cells on the specimens were fixed and stained with a leukocyte acid phosphatase kit for evaluating the TRAP level.TRAP-positive cells (more than three nuclei) appeared dark red and were counted by stereo microscope.

Statistical analysis
Each assay was repeated 6 times, and the data were expressed as mean ± standard deviation.The normality test was performed with SPSS software with one-way or two-way analysis of variance (ANOVA) to determine the levels of statistical significance between different groups.Statistical significance was considered for p < 0.05.

Results
3.1.Fabrication and release behavior of the composite coating 3.1.1.Fabrication and characterization of the composite coating Measured by the particle analyzer, the mean nanoparticle size of the pH-sensitive double-layered nanoparticles was 422.7 ± 17.7 nm, and the polydispersity index was 0.144 ± 0.007.Observed under SEM, the nanoparticles were spherical and smooth, with no pores and cracks, as shown in figure 1(A).A double-layered structure was observed under TEM: a thin outer layer with a low image density and a thick inner layer with a high image density (figure 1(B)).
Figures 1(C) and (D) show the SEM images of the surface morphologies of polished cp-Ti before and after MAO treatment.Before MAO treatment, the polished cp-Ti specimen exhibited a smooth surface with just a few grooves parallel to the polishing direction (figure 1(C)), while a porous coating with pores approximate 1-3 μm in diameter shaped like craters was observed under SEM after MAO treatment (figure 1(D)).
Figures 1(E) and (F) show the surface morphology of the composite coating.The pores of the MAO coating were almost filled with the pH-sensitive double-layered nanoparticles, and just a few nanoparticles adhered to the plane area among the pores (figure 1(E)).The pH-sensitive double-layered nanoparticles were well distributed and tightly cross-linked with each other and the wall of pores on the MAO coating (figure 1(F)).

In vitro drug release study
Figure 2 shows the release profiles of SDF-1, rhBMP-2 and OPG from the composite coating at different pH values in PBS in vitro.All the three drugs were observed to have an initial burst release (the first several hours) and a slower continuous release (move than 30 days) at different pH values.At the same pH value, the cumulative SDF-1 and OPG release from the composite coating were higher than the cumulative rhBMP-2 release.The release study also showed that the cumulative OPG release from the composite coating increased significantly with decreasing pH (p < 0.05).Compared with OPG, the cumulative SDF-1 and rhBMP-2 release from the composite coatings increased slightly with decreasing pH, but the difference was not significant (p > 0.05).
The surface morphologies of the composite coating specimens immersed in PBS at different pH values are shown in figure 3.After 10 days, the composite coating immersed in PBS (pH 7.4 and 6.8) did not show significant changes (figures 3(A) and (D)), while the nanoparticles deformed slightly in PBS (pH 6.2, figure 3(G)).Subsequently, the nanoparticles were deformed obviously and the number of the nanoparticles in the pores of the composite coating decreased after 20-day immersion (figures 3(B), (E) and (H)).After 30 days, the decrease in the numbers of nanoparticles was even more pronounced (figures 3(C), (F) and (I)).

In vitro biological studies of the composite coating 3.2.1. Cytotoxicity assay
The result of the BMSC RGRs calculated by CCK-8 assay is shown in figure 4. The RGRs of all the specimens were much greater than 70%, demonstrating that none of the groups exhibited cytotoxicity (ISO 10993-5: 2009).

BMSC recruitment assay
After 8 h of culture, the numbers of BMSCs on the underside of the microporous membrane in each visual field observed by inverted microscope are shown in figure 5.The number of BMSCs in composite coating group was significantly higher than those in MAO-Ti and polished cp-Ti groups (p < 0.05).Moreover, at different pH values, there was no statistically significant difference in the numbers of BMSCs on the underside of the microporous membrane (p > 0.05), which was consistent with the SDF-1 release study from the composite coating.Figure 6 shows the BMSCs on the underside of the microporous membrane under inverted microscope after culturing for 8 h at the pH value of 6.8.

BMSC differentiation assay
The ALP optical density in the cells on the three groups after culturing for 7 days is shown in figure 7. The ALP optical density of the composite coating was significantly higher than those of MAO-Ti and polished cp-Ti (p < 0.05).At different pH values, there was no statistically significant difference in the ALP optical density (p > 0.05).The result was also consistent with the rhbmp-2 release study from the composite coating.After 7-day co-culture, the cells on the specimens were dyed by ALP dye kit and observed under stereo microscope.Figure 8 shows the dyed cells on the specimens after 7-day co-culture at the pH of 6.8.

Inhibiting osteoclast generation
Figure 9 shows the numbers of TRAP-positive cells on the specimens after 6-day culture in each visual field under stereo microscope.At different pH values, the number of TRAP-positive cells on composite coatings was significantly lower than those on MAO-Ti and polished cp-Ti (p < 0.05).Meanwhile, as the pH decreased, the number of TRAP-positive cells on the composite coating decreased (p < 0.05), because the cumulative OPG release from the composite coating increased significantly with decreasing pH (p < 0.05).Figure 10 shows the cells on the specimens observed by stereo microscope at the pH of 6.8.

Discussion
In the past few decades, various delivery systems loaded with many kinds of drugs have been used to achieve their controlled release for the therapeutic purpose.Among them, drug-loaded polymer nanoparticles (PNPs) are widely used to deliver certain drug to target cells [26].The mechanism by which a drug pulls away from its carrier is summarized as diffusion, swelling, erosion and degradation, and several factors influence drug release from PNPs, including drug composition (drug, polymer, and adjuvants), preparation method, internal or external stimuli (temperature, pH value, electromagnetic and magnetic fields, or ultrasonic waves) [27].Recently, double-layered microparticles/nanoparticles are considered an effective structural design technique to obtain co-delivery.They can realize multi-stage drug release through the time difference between the core layer and the shell layer [28].CS is a copolymer of glucosamine and N-acetylglucosamine, on which the -OH and -NH 2 active groups are prone to chemical reactions [29].Various modifications of CS have been performed in order to improve the pH sensitivity, targeting and water solubility of CS derivatives [30].Therefore, in this study CS was used to conjugate with a pH-sensitive hydrazone bond in the preparation of the pH-sensitive double-layered nanoparticles.The mean nanoparticle size of the prepared pH-sensitive double-layered nanoparticles (422.7 ± 17.7 nm) was much smaller than the pore size on the MAO coating (1-3 μm), which made it possible for the nanoparticles to penetrate into the pores.The pH-sensitive double-layered nanoparticles also exhibited a low polydispersity index (0.144 ± 0.007).It was reported that a low polydispersity index would benefit to the stability of     nanoparticles [31].Moreover, a smooth surface without any pore or crack was observed on surface of the nanoparticles under SEM (figure 1(A)).The surface morphology of microspheres/nanoparticles was a crucial property, since the initial burst release from smooth surfaces was less than that from rough surfaces in vitro release experiments [32].The composite coating was fabricated successfully by cross-linking the pH-sensitive double-layered nanoparticles into the pores of MAO coating with gelatin.Gelatin is derived from collagen, which is the most abundant protein in the body.Due to its good biocompatibility, biodegradability and anti-inflammatory effect, gelatin has been widely used in the food, cosmetic, engineering and pharmaceutical industry [33].Gelatin has also been widely used as a cross-linking agent to cross-link nanoparticles to the surface of MAO coating to fabricate composite coating [20,23].SEM images showed that the pH-sensitive double-layered nanoparticles were well distributed and tightly cross-linked with each other and the wall of pores by gelatin (figure 1(F)).This design would effectively prevent the pH-sensitive double-layered nanoparticles from falling off during implantation.
The in vitro release profiles of all the three drugs from the composite coating showed an initial burst release (the first several hours) and a slower continuous release (move than 30 days) at different pH values (figure 2).The drugs being adsorbed or near the surface of the microspheres/nanoparticles usually result in the burst release [34].In the second stage, the drugs were released gradually and slowly with the continuous degradation of the polymer.The first 28 days after dental implant implantation are the most active period for various growth factors, proteins, and cells to participate in osseointegration [35].During this period, the release of the three drugs from the composite coating could help the process of osseointegration.As shown in the SEM images of the process of degradation of the composite coating (figure 3), the number of the nanoparticles in the pores decreased after 20-day immersion.This was mainly resulted from the external degradation of the nanoparticles, so they fell off from the pores under stirring of PBS.After 30-day immersion in stirring of PBS, most of the nanoparticles were fell off from the pores.So the composite coating fabricated in this study would not affect the final osseointegration of the new bone with the MAO coating.
The cumulative SDF-1 and OPG released from the composite coating were higher than the cumulative rhBMP-2 released at the same pH value, which resulted from rhBMP-2 being loaded in the inner layer of the pHsensitive double-layered nanoparticles.The cumulative OPG released from the composite coating increased significantly with decreasing pH (p < 0.05), because OPG was conjugated to the CS backbone by pH-sensitive hydrazone bonds during the preparation.The acidic environment would self-catalyze the polymer degradation and then increase the drug release [36].So the cumulative SDF-1 and rhBMP-2 released from the composite coatings increased slightly with decreasing pH, but the difference was not significant (p > 0.05).
The pH varies in different tissues and cells, especially in inflammatory or tumor conditions [37].Local acidosis could be detected in the inflamed area, with a pH between 5.5 and 7 [38].For peri-implantitis, the pH in the peri-implant tissue is also lower than the normal physiological value [39].Therefore, three pH values were arranged in this study, among which the pH 6.8 and 6.2 were simulated the pH values in infected environment.The pH-sensitive drug-loaded microspheres/nanoparticles are designed to control the drug release rate in response to different pH values under pathological environment [40].In the body, weak base groups (such as amines) and weak acid groups (such as phosphoric and acids carboxylic acids) exhibit corresponding ionized states, therefore the conformation of drug carriers changes, which then induces the controlled release of drugs [41].Yamamoto indicated that the pH-sensitive nanodevices loaded siRNA drugs may be a promising approach to treat hepatitis B virus [42].
Studies have indicated that SDF-1 is essential for the recruitment and survival of BMSCs, thereby participating in the regeneration of bone [43].So, the released SDF-1 from the composite coating promoted the migration of BMSCs to the lower compartment in this assay.BMSCs exhibit multidirectional differentiation potential to differentiate into chondrocytes, osteoblasts and myoblasts [5].Producing ALP is considered to be an important indicator of the differentiation of BMSCs into mature osteoblasts [44].Therefore, BMSC differentiation into mature osteoblasts was usually evaluated by detecting ALP activity.RhBMP-2 has become a commonly used growth factor for its efficient ability to induce osteoblast differentiation [45].Hence, the released rhbmp-2 from the composite coating induced BMSC differentiation into mature osteoblasts, and then the mature osteoblasts produced ALP.
Contrary to the biological effects of RANKL, OPG can inhibit the osteoclastogenesis and accelerate the apoptosis of osteoclast [16].So, the differentiation and maturation of osteoclasts were inhibited due to the released OPG from the composite coating.TRAP is mainly distributed in mature osteoclasts and alveolar macrophages in human body, so the TRAP level can reflect the function of the related cells [46].In this study, TRAP dye was used to evaluate the numbers of mature osteoclasts generated from BMMs.The more mature osteoclasts generated, the more TRAP-positive cells would be observed on the specimens under stereo microscope.On the contrary, when the generation of osteoclasts was inhibited by OPG, the TRAP-positive cells also reduced.More OPG was released from the coating as the pH decreased, therefore the osteoclast generation was further inhibited.
When bacterial infection occurred around the implant, the pH value around the implant was significantly lower than that of normal tissue [47].In this study, OPG was designed to conjugate to the CS backbone by pHsensitive hydrazone bonds during the preparation.pH-sensitive hydrazone bonds would break under acidic conditions, and then release OPG.The lower the ambient pH, the faster the pH-sensitive hydrazone bond breaks, thus more OPG is released.The RANKL/OPG ratio determines the physiological balance of bone formation or bone resorption [14].Therefore, the accelerated release of OPG from the composite coating in acidic can bind with RANKL, competitively inhibit the activity of RANK, and then effectively inhibit the further bone resorption in inflammatory environment.
If the fabricated composite coating could be applied on the surface of dental implants, it may have the effect on promoting peri-implant bone regeneration and inhibiting bone resorption.On one hand, the released SDF-1 would induce the recruitment of endogenous BMSCs to the planting area, and the released rhbmp-2 would induce BMSC differentiation into mature osteoblasts.Under the synergistic action of SDF-1 and rhbmp-2, endogenous bone regeneration was achieved around the implant.On the other hand, the released OPG would effectively inhibit bone resorption.In particular, more OPG would be released during peri-implant infection (pH value decreasing).The bone mass around implants is very important, so the composite coating fabricated in this study would benefit for the long-term stability and success rate of dental implants, especially when the alveolar bone condition is poor.

Conclusion
In this study, a composite coating was successfully fabricated on Ti.The composite coating could achieve sustained release SDF-1, rhBMP-2 and OPG for over 30 days.With decreasing pH, the release of OPG from the composite coating increased (p < 0.05).In vitro biological studies suggested that the composite coating exhibited no cytotoxicity, and can recruit BMSCs, promote BMSC differentiation into osteoblasts and inhibit osteoclast generation.Moreover, with decreasing pH, the inhibitory effect on osteoclast generation was enhanced (p < 0.05).It can be concluded that the fabricated composite coating, which can promote peri-implant bone regeneration and inhibit bone resorption, has the potential to be applied on the surface of dental implant, especially when the residual alveolar bone is in poor condition.

Figure 1 .
Figure 1.(A) SEM image of the pH-sensitive double-layered nanoparticles.(B) TEM image of the pH-sensitive double-layered nanoparticles.(C) SEM image of the polished cp-Ti.(D) SEM image of the MAO coating.(E) SEM image of the composite coating at low magnification (×3,000).(F) SEM image of the composite coating at high magnification (×20,000).

Figure 5 .
Figure 5. Numbers of BMSCs on the underside of the microporous membrane in each visual field under inverted microscope ( * p < 0.05, mean ± SD, n = 6).

Figure 6 .
Figure 6.BMSCs on the underside of the microporous membrane observed by inverted microscope at the pH value of 6.8.(A) Composite coating.(B) MAO-Ti.(C) Polished cp-Ti.

Figure 7 .
Figure 7. Optical density of ALP in the cells on the three groups ( * p < 0.05, mean ± SD, n = 6).

Figure 8 .
Figure 8. Dyed cells on the specimens observed by stereo microscope at the pH of 6.8.(A) Composite coating.(B) MAO-Ti.(C) Polished cp-Ti.

Figure 9 .
Figure 9. Numbers of TRAP-positive cells on the specimens in each visual field under stereo microscope ( * p < 0.05, mean ± SD, n = 6).

Figure 10 .
Figure 10.Cells on the specimens observed by stereo microscope at the pH of 6.8.(A) Composite coating.(B) MAO-Ti.(C) Polished cp-Ti.