Mesoporous Calcium Silicate Nanoparticles withDrug Delivery and Odontogenesis PropertiesChing-Yuang Huang, DDS, PhD,*† Tsui-Hsien Huang, DDS, PhD,‡§ Chia-Tze Kao, DDS, PhD,‡§

Yuan-Haw Wu, MD,k¶ Wan-Chen Chen, DDS, Ms,‡§ and Ming-You Shie, PhDk#**

AbstractIntroduction: Calcium silicate (CS) –based materialsplay an important role in the development of endodonticmaterials that induce bone/cementum tissue regenera-tion and inhibit bacterial viability. The aim of this studywas to prepare novel mesoporous CS (MesoCS) nano-particles that have osteogenic, drug delivery, and anti-bacterial characteristics for endodontic materials andalso have an excellent ability to develop apatite miner-alization. Methods: The MesoCS nanoparticles wereprepared using sol-gel methods. In addition, the meso-porous structure, specific surface area, pore volume,and morphology of the MesoCS nanoparticles wereanalyzed. The apatite mineralization ability, in vitroodontogenic differentiation, drug delivery, and antibac-terial properties of the MesoCS nanoparticles werefurther investigated. Results: The results indicate thatthe 200-nm–sized MesoCS nanoparticles synthesizedusing a facile template method exhibited a high specificsurface area and pore volume with internal mesopores(average pore size = 3.05 nm). Furthermore, the Mes-oCS nanoparticles can be used as drug carriers to main-tain sustained release of gentamicin and fibroblastgrowth factor-2 (FGF-2). The MesoCS-loaded FGF-2might stimulate more odontogenic-related proteinthan CS because of the FGF-2 release. Conclusions:Based on this work, it can be inferred that MesoCSnanoparticles are potentially useful endodontic mate-rials for biocompatible and osteogenic dental pulp tissueregenerative materials. (J Endod 2017;43:69–76)

Key WordsCalcium silicate, dental pulp cell, drug delivery, mesopo-rous, odontogenic

Since its introduction in1993, mineral trioxide

aggregate (MTA) hasbeen used in endodontictreatment, not only as aroot-end filling materialbut also for direct pulpcapping and regenerativeendodontic procedures(1–3). MTA is a calciumsilicate (CS)-based cement that contains tricalcium silicate, dicalcium silicate,tricalcium aluminate, tetracalcium aluminoferrite, gypsum, and bismuth oxide (4).The bioactivity of CS-based materials has led to its use in constructing scaffolds withvarious stem cells for the purpose of tissue regeneration (5, 6). In our previousstudies, we developed a CS-based material that has since been introduced as a usefuladditive in the medical field because of its alkalinity, apatite-forming abilities, and anti-bacterial properties (7). In addition, CS-based materials can promote hard tissueregeneration and have the ability to stimulate odontogenic and osteogenic differentia-tion in various types of cells, such as bone marrow stromal cells, adipose-derived stemcells, human dental pulp cells, and periodontal ligament cells (8–11). However, thismaterial has exhibited 2 weaknesses:

1. the size of the CS-based material particles is usually at the micrometer level, makingit difficult to inject, and

2. traditional CS-based materials have a minimal nanopore structure, which greatlylimits the potential for drug delivery (12).

Several previous studies have shown that mesoporous materials may be used asnovel drug delivery carriers with release kinetics that can be controlled by adjustinghollow internal microstructures (13, 14). By definition, a mesoporous material is astructure with pores that have diameters between 2 and 50 nm; it is intermediate insize between microporous (<2 nm) and macroporous (>50 nm) materials (15). Ithas been suggested that structurally well-ordered mesoporous materials could poten-tially act as carriers for loading biomolecules and harmonizing their release (16). Tar-geted drug delivery is an important tumor therapy method, and the bioactivity ofmaterials is of great importance for promoting tissue regeneration (17–20).Therefore, mesoporous CS (MesoCS) nanoparticles that offer more surface areahave been developed to combine with biomaterials for various bioengineering

From the *Department of Stomatology and †Division of Family Dentistry, Chang Bing Show Chwan Memorial Hospital, Changhua; ‡School of Dentistry, Chung ShanMedical University, Taichung; §Department of Stomatology, Chung Shan Medical University Hospital, Taichung; k3D Printing Medical Research Center, China MedicalUniversity Hospital, Taichung; ¶School of Medicine and #School of Dentistry, China Medical University, Taichung; and **Department of Bioinformatics and Medical En-gineering, Asia University, Taichung City, Taiwan.

Address requests for reprints to Dr Ming-You Shie, 3D Printing Medical Research Center, China Medical University Hospital, China Medical University, Taichung,Taiwan. E-mail address: eviltacasi@gmail.com0099-2399/$ - see front matter

Copyright ª 2016 American Association of Endodontists.http://dx.doi.org/10.1016/j.joen.2016.09.012

SignificanceThe MesoCS nanoparticles were prepared by sol-gel and can be used as a drug carrier to releasegentamicin. The gentamicin-loaded MesoCS hadbetter antibacterial abilities than gentamicin-loaded CS. In addition, the FGF-2 loaded MesoCSup-regulation of odontogenic-related protein ofhDPCs.

Regenerative Endodontics

JOE — Volume 43, Number 1, January 2017 Mesoporous Calcium Silicate Nanoparticles 69

applications, such as intracellular molecule labeling, drug delivery, andgene targeting (21). Of particular interest are the findings of Chen et al(9), which suggested that the stimulation of proliferation andosteogenic-related protein synthesis by culturing primary cells afterexposure to extracts of sol-gel CS particles are related to Si contactand not Ca because there was no increase in the observed osteoblastactivity in the absence of Si ions.

Several growth and differentiation factors have been proven to bepotential therapeutic agents for hard tissue regeneration (22, 23).Fibroblast growth factors (FGFs) play an important role in thecontrol of cell adhesion, proliferation, and differentiation in theregeneration of several kinds of tissues (2). Notably, basic FGF (FGF-2) has been shown to promote proliferation and differentiation (24).In addition to the important role of FGF signaling in the control of os-teoprogenitor cells, reduced stem cell osteogenesis differentiation andinhibited bone formation have been witnessed in mice with FGF-2 geneknockout (25). Receptor phosphorylation of intrinsic tyrosine residuesis induced after FGF binds to an FGF receptor, activating several signaltransduction pathways, including mitogen-activated protein kinases(MAPKs) and phosphatidylinositol 3-kinase (24). In hard tissue, phos-phorylation of extracellular-related kinase (ERK1/2)/MAPK and proteinkinase C has been found to enhance osteogenic and odontogenic geneexpression (26–28).

The desired features of root canal filling material shouldinclude broad-spectrum antibacterial behavior to suppress thegrowth of bacteria in the periapical area, excellent mineralizationto facilitate sealing of the apical root canal of a tooth and enhancethe integration between biomaterials and periapical tissues, andexcellent osteogenic properties to promote hard tissue formation.For these reasons, it is very important to develop MesoCS nanopar-ticles with excellent apatite mineralization as well as osteogenic,drug delivery, and antibacterial characteristics for endodontic mate-rials. Therefore, the main purpose of this study was to developnovel MesoCS-based nanoparticles and to analyze their physiochem-ical, drug delivery, odontogenic, and antibacterial properties for theapplication of filling apical dental root canals.

Materials and MethodsSynthesis and Characterization of MesoCSNanoparticles

MesoCS nanoparticles were prepared using a template method.Briefly, 3.3 g cetyltrimethylammonium bromide (CTAB) (Sigma-Al-drich, St Louis, MO) and 6mL NH3 $H2O were mixed in double distilledwater (ddH2O, 300 mL) and then stirred for 15 minutes at 60

!C. Next,15 mL tetraethyl orthosilicate (Sigma-Aldrich) and 15.6 g calcium ni-trate were added with vigorous stirring for 3 hours. The precipitateproducts were then collected by filtration and washed 3 times eachwith 1 N hydrochloric acid and ethanol. After this, the collected powderswere dried at 60!C overnight and sintered at 800!C for 2 hours to re-move the remaining traces of CTAB. As for the preparation of the solidCS powder, appropriate amounts of CaO (Showa, Tokyo, Japan), SiO2(High Pure Chemicals, Saitama, Japan), and 5% Al2O3 (Sigma-Aldrich)powders were mixed and sintered at 1400!C for 2 hours using a high-temperature furnace. After sintering, the powders were ball milled inethyl alcohol using a centrifugal ball mill (S 100; Retsch, Hann, Ger-many) for 6 hours. The CS cement has a liquid/powder ratio of0.3 mL/g. The nanoparticles and cement were characterized usingsmall-angle X-ray diffraction (XRD) (Bruker D8 SSS, Karlsruhe, Ger-many) and transmission electron microscopy. The nanopore size dis-tribution and nanopore volume were determined with Brunauer–Emmett–Teller analysis.

In Vitro SoakingThe cements were immersed in 10 mL simulated body fluid (SBF)

solution in 15-mL tubes at 37!C. The ionic composition of the SBF so-lution was similar to that of human blood plasma and consisted of7.9949 g NaCl, 0.2235 g KCl, 0.147 g K2HPO4, 0.3528 g NaHCO3,0.071 g Na2SO4, 0.2775 g CaCl2, and 0.305 g MgCl2 $ 6H2O in1000 mL distilled H2O. The pH was adjusted to 7.4 with hydrochloricacid and tris(hydroxymethyl)aminomethane (Tris, [CH2OH]3CNH2).The solution in the shaker water bath was not changed daily under staticconditions. After soaking for different time intervals, the specimenswere removed from the tube, and various physicochemical propertiesof the specimens were evaluated. After immersion for different time pe-riods, the Ca and Si ion concentrations released from the specimens onSBF were analyzed using an inductively coupled plasma atomic emissionspectrometer (OPT 1 MA 3000DV; Perkin-Elmer, Shelton, CT). Threesamples were measured for each data point. The results were obtainedin triplicate from 3 separate samples for each test. The specimens werecoated with gold, and their morphologies were investigated under ascanning electron microscope (JSM-6700F; JEOL, Tokyo, Japan) oper-ated in the lower secondary electron image mode at a 3-kV acceleratingvoltage.

Gentamicin Loading and Release KineticThe gentamicin (GENT)-loaded specimens were prepared by im-

mersion in 1 g as-prepared CS-based material and GENT sulfate (MPBiomedicals, Solon, OH) of distilled H2O. This compound was thenshaken for 24 hours at room temperature. The GENT-CS–based mate-rials were then separated using centrifugation at 13,000 rpm for 20 mi-nutes to remove unloaded GENT sulfate. The GENT-CS–based materialswere then washed with ddH2O several times and dried at 37!C for 24hours. The drug release kinetics of GENT-CS–basedmaterials were eval-uated by soaking 10 mg of specimens in 1 mL PBS (pH = 7.4) at 37!C;fresh replacement PBS was substituted for the old PBS at various timepoints. The release of GENT sulfate was determined using the o-phthal-dialdehyde method (29). The o-phthaldialdehyde reagent wasformulated by adding 2.5 g o-phthaldialdehyde (Fluka, Buchs,Switzerland), 62.5 mL methanol (Sigma-Aldrich), and 3 mL2-mercaptoethanol (Sigma-Aldrich) to 560 mL 0.04 mol/L sodium tet-raborate (Fluka) in distilled water. The reagent was stored in a brownbottle in a dark chamber and allowed to settle for over 24 hours beforeuse. The o-phthaldialdehyde reacted with the GENT amino groups,which, in turn, formed chromophoric products. Sequentially, the prod-ucts were determined using an ultraviolet-visible spectrophotometer(Infinite Pro M200; TECAN, M€annedorf, Switzerland). Each value rep-resented the average of 3 runs. The amount of antibiotic released wasassayed through a comparison with a calibration curve for the individ-ual antibiotic made in PBS.

Antibacterial PropertiesTo investigate the antibacterial effects of the GENT released from

CS-based particles, Enterococcus faecalis in an lysogeny broth culturemedia (4.0 " 104 bacteria per mL) was cultured with CS-based parti-cles for 24 hours. Aliquots of 0.1 mL from each group were then mixedwith 0.9 mL PrestoBlue (Invitrogen, Waltham, MA) for 10 minutes; afterwhich, the solution in each well was transferred to a new 96-well plate.Plates were then read in a multiwell spectrophotometer (Infinite ProM200) at 570 nm with a reference wavelength of 600 nm. Bacteriacultured on the tissue culture plate without the specimens were usedas a control (Ctl). The results were obtained in triplicate from 3 separateexperiments in terms of optical density.

Regenerative Endodontics

70 Huang et al. JOE — Volume 43, Number 1, January 2017

Human Dental Pulp Cell Isolation and CultureThe human dental pulp cells (hDPCs) used in this experiment

were freshly derived from caries-free, intact premolars that had beenextracted for orthodontic treatment purposes as described previously.The patient gave informed consent, and approval from the Ethics Com-mittee of the Chung Shan Medicine University Hospital was obtained(no. CS14117). After extraction, the tooth was split sagittally with achisel. The pulp tissue was then immersed in a PBS (Caisson, North Lo-gan, UT) solution and digested in 0.1% collagenase type I (Sigma-Al-drich) for 30 minutes. After being transferred to a new plate, the cellsuspension was cultured in Dulbecco modified Eagle medium(DMEM, Caisson) supplemented with 20% fetal bovine serum (Gene-DireX, Las Vegas, NV) and 1% penicillin (10,000 U/mL)/streptomycin(10,000 mg/mL) (Caisson) in a humidified atmosphere with 5% CO2at 37!C; the medium was changed every 3 days. The cells were subcul-tured through successive passaging at a 1:3 ratio until they were used forexperiments (passages 3–7). The odontogenic differentiation mediumwas DMEM supplemented with 10#8 mol/L dexamethasone (Sigma-Al-drich), 0.05 g/L L-Ascorbic acid (Sigma-Aldrich), and 2.16 g/L glycerol2-phosphate disodium salt hydrate (Sigma-Aldrich).

Cell ViabilityBefore the in vitro cell experiments, the specimens were sterilized

by soaking in 75% ethanol followed by exposure to ultraviolet light for 1hour. After being directly cultured for various time periods, cell viabilitywas evaluated using the PrestoBlue assay (Invitrogen), which is basedon the detection of mitochondrial activity. Thirty microliters of Pres-toBlue solution and 300 mL DMEM were added to each well followedby 30 minutes of incubation. After incubation, 100 mL of the solutionin each well was transferred to a 96-well enzyme-linked immunosor-bent assay plate. The plates were read in a multiwell spectrophotometer(TECAN Infinite Pro M200) at 570 nm with a reference wavelength of600 nm. The results were obtained in triplicate from 3 separate exper-iments in terms of optical density. Cells cultured on a tissue culture platewithout the cement were used as a Ctl.

Cell MorphologyWe observed the cell morphology of the hDPCs after they had been

seeded for 6 hours on specimens using F-actin cytoskeleton stains. Thecells were washed with PBS, fixed in 4% paraformaldehyde (Sigma-Al-drich) at room temperature for 20 minutes, and then permeabilizedwith PBS containing 0.1% Triton X-100 (Sigma-Aldrich). The F-actinfilaments were stained with phalloidin conjugated to Alexa Fluor 594(Invitrogen) for 1 hour. The nuclei were stained with 300 nmol/L40,6-diamidino-2-phenylindole, dilactate (Invitrogen) for 30 minutes.After washing, the morphology was obtained using the Zeiss Axioskop2microscope (Carl Zeiss, Thornwood, NY). The cells were also culturedfor both 1 and 3 days; after which, the specimens were washed 3 timeswith cold PBS, fixed in 1.5% glutaraldehyde (Sigma-Aldrich) for2 hours, and then dehydrated using a graded ethanol series for 20 mi-nutes at each concentration and dried with liquid CO2 using a criticalpoint dryer device (LADD 28000; LADD, Williston, VT). The dried spec-imens were then mounted on stubs, coated with gold, and viewed usingscanning electron microscopy (JSM-7401F, JEOL).

FGF-2 LoadingThe FGF-2–loaded specimens were prepared by immersion in 1 g

as-prepared CS-based materials and FGF-2 (5 mg/mL; ProSpec, Reho-vot, Israel) of distilled H2O, which was then shaken for 24 hours atroom temperature. The FGF-2/CS-based materials were then separatedusing a centrifuge at 13,000 rpm for 20 minutes to remove unloaded

FGF-2. The powders were then washed with ddH2O several times anddried at 37!C for 24 hours; after which, they were mixed with ddH2Oand fully covered in a 24-well plate (GeneDireX) to a thickness of2 mm. The samples were stored in an incubator at 37!C and100% rela-tive humidity for 1 day. Before the cell experiments, all specimens weresterilized by immersion in 75% ethanol followed by exposure to ultra-violet light for 1 hour.

Alkaline Phosphatase Activity AssayThe secretion of alkaline phosphatase (ALP) was determined on

days 3 and 7 after cell seeding. Briefly, the hDPCs were lysed using0.2% NP-40 (Sigma-Aldrich) and centrifuged at 6000 rpm for 15 mi-nutes. ALP activity was determined using p-nitrophenyl phosphate(Sigma-Aldrich) as the substrate. Each sample was mixed with p-nitro-phenyl phosphate in 1 mol/L diethanolamine buffer for 15 minutes. Thereaction was stopped by the addition of 5 N NaOH and quantified byabsorbance at 405 nm. All experiments were performed in triplicate.

Enzyme-linked Immunosorbent AssayOsteocalcin (OC), dentin matrix protein-1 (DMP-1), and dentin

sialophosphoprotein (DSP) released from the pulp cells were culturedon different substrates for 7 and 14 days after cell seeding. An enzyme-linked immunosorbent assay kit (Invitrogen) was used to determineprotein content following the manufacturer’s instructions. The concen-tration of OC, DMP-1, and DSP were measured by correlation with astandard curve. Analysis of blank disks was treated as a Ctl. All exper-iments were performed in triplicate.

Alizarin Red S StainThe accumulated calcium deposition was analyzed using alizarin

red S staining following a method developed for a previous study(30). In brief, the specimens were fixed with 4% paraformaldehyde(Sigma-Aldrich) for 15 minutes and then incubated in 0.5% alizarinred S (Sigma-Aldrich) at a pH of 4.0 for 15 minutes at room tempera-ture. After this, the cells were washed with PBS and quantified using asolution of 20% methanol and 10% acetic acid in water. After 15 mi-nutes, the liquid was transferred to a 96-well plate, and the quantityof alizarin red was determined using a spectrophotometer at 450 nm.

Statistical AnalysisOne-way variance statistical analysis was used to evaluate the sig-

nificance of the differences between the means in the measured data.The Scheffe multiple comparison test was used to determine the signif-icance of the deviations in the data for each specimen. In all cases, theresults were considered statistically significant if the P value was <.05.

ResultsPhysicochemical Properties

CS nanoparticles with a size of about 200 nm were successfullysynthesized through base-catalyzed hydrolysis and condensation of tet-raethyl orthosilicate in the presence of calcium ions. As seen inFigure 1A, CS nanoparticles with a porous internal structure (MesoCS)were produced when CTAB was introduced in the precursor solution toserve as a template. Figure 1B and C reveal the small-angle and wide-range XRD patterns of the CS nanoparticles, respectively. The resultsshow a distinct diffraction peak at 2q = 2! in MesoCS, indicating thatMesoCS has an ordered mesostructure (Fig. 1B). Additionally, thediffraction peak of CS at 2q = 29.4! and 32!–34! corresponds to CShydrate and b-dicalcium silicate (b-Ca2SiO4). MesoCS shares a similarphase composition with CS, suggesting that the presence of a CTAB

Regenerative Endodontics

JOE — Volume 43, Number 1, January 2017 Mesoporous Calcium Silicate Nanoparticles 71

template would not affect the formation and crystallization of CS synthe-sized by the basic catalyzation route. N2 adsorption-desorption resultsindicate that the average size of the mesopores in MesoCS was 3.05 nmwith a specific surface area of 462 m2/g and a mesopore volume of0.62 cm3/g (Fig. 1D).

Apatite Formation of CS and MesoCS CementsTo investigate the bioactivity of CS and MesoCS, cements prepared

throughmixing nanoparticles with ddH2O were soaked in SBF. The vari-ation in the SBF ions and the morphology of the cements were examinedusing an inductively coupled plasma atomic emission spectrometer andfield emission scanning electron microscopy. Figure 1E shows the evo-lution of Ca and Si ions in SBF as a function of immersion time. The re-sults reveal that an initial rapid decrease of Ca ions in SBF occurred inthe first 12 hours of the soaking period and then decreased from1.40 mmol/L to 1.08 and 0.92 mmol/L for CS and MesoCS, respectively.In contrast, the change in the Si ions in SBF exhibited the opposite ten-dency. The concentration of Si ions increased from 0 to 1.35 and1.62 mmol/L in SBF for the CS and MesoCS cements, respectively,and the release profile followed the first-order kinetic model. Interest-ingly, the MesoCS cement showed higher potency in terms of Si releaseand Ca adsorption in SBF. Scanning electron microscopic micrographsof the cements before and after soaking in SBF for 1 and 3 days areshown in Figure 1F. It can be seen that MesoCS cement has a smoothersurface and better interconnected inorganic particles than CS after mix-ing with ddH2O and sitting at 37!C with 100% humidity for 1 day. Aftersoaking in SBF, the cement surface was covered by an apatite layer. Itcan be seen that the amount of precipitated apatite on MesoCS is higherthan that on CS. However, all cement surfaces are covered with a denseapatite layer after immersion in SBF for 3 days. In addition, broad,diffuse peaks at 2q= 25.9 and 31.8–32.9 clearly appear in the resultingXRD patterns of specimen immersion in SBF for 1 day, which may beascribed to the characteristic peaks of apatite.

Antibacterial Activity of GENT-loaded CementsGENT-loaded CS and MesoCS were prepared by soaking CS and

MesoCS powders in GENT solution to allow the drug molecules todiffuse and adsorb onto the surface of the inner pores of the inorganicparticles. After that, GENT-CS and GENT-MesoCS cements were fabri-cated using the method described previously. The release profiles ofthe cements were conducted bymeasuring the amount of GENT releasedin PBS (pH = 7.4) as a function of time (Fig. 2A). The results indicatethat the burst release of drugmolecules was visible on the first day of theimmersion period for all groups; 0.93 and 0.83 mg drug moleculeswere released by GENT CS and GENT MesoCS, respectively. However,the release rate of the drug molecules decreased markedly for GENTCS, and the total release amount of GENT over an immersion periodof 5 days was 1.0933 mg. In contrast, the GENT released from GENTMesoCS on the fifth day of immersion was approximately twice thaton the first day, implying that the sustained-release profile could be ob-tained in MesoCS acquired in the presence of mesopores, which appearto provide a higher surface area for GENT adsorption.

The antibacterial activity of GENT-load cements against gram-positive bacteria strain E. faecalis is shown in Figure 2B. For this exper-iment, bacteria cultured in a tissue culture plate and cultured onCa(OH)2 served as the negative and positive Ctls, respectively. The re-sults indicate that the amount of S. aureus increased with time forthe Ctl, CS, andMesoCS cements and that there was a significantly higheramount of E. faecalis on the Ctl than on other groups for all culture pe-riods. As a positive Ctl for antibacterial analysis, Ca(OH)2 exerted thehighest level of antibacterial activity against E. faecalis in comparisonwith the Ctl, CS, and MesoCS cements (P < .05). The presence ofGENT, an antibiotic agent, in the CS and MesoCS cements seemed toinhibit the adhesion and growth of bacteria, which was validated bycomparing the amount of bacteria cultured on cements with andwithout GENT loading (P< .05). It is worth mentioning that the antibac-terial activity of MesoCS was as high as that of Ca(OH)2. More

Figure 1. (A) Transmission electron microscopic micrographs and (B) small-angle and (C) wide-range XRD patterns of CS and MesoCS. (D) Pore size distri-bution in MesoCS calculated from N2 adsorption-desorption data using the Brunauer–Emmett–Teller method. V, volume of adsorbed N2; D, pore diameter. (E) Theevolution of Ca and Si ion concentrations in SBF with CS and MesoCS cements as a function of immersion time. (F) Scanning electron microscopic micrographs andXRD of cements before and after soaking in SBF.

Regenerative Endodontics

72 Huang et al. JOE — Volume 43, Number 1, January 2017

importantly, MesoCS displayed significantly more antibacterial activitythan CS when the culture period was 12 hours or longer (P < .05).

Cell Adhesion and ProliferationThe PrestoBlue assay was conducted to evaluate cell adhesion and

proliferation on the CS and MesoCS cements. Figure 1A and B indicatethe absorbance of PrestoBlue treatment with hDPCs cultured from 3 to12 hours and from 1 to 7 days. The results reveal no significant differ-ences between hDPCs cultured on CS and MesoCS cements. It is note-worthy that the absorbance of both cements was statistically higher(P < .05) than that of Ctl for all culture periods for hDPC adhesion(Fig. 3A) and proliferation (Fig. 3B). The absorbance value of cementsfor the hDPC culture periods up to 1 day was approximately 40% higherthan that of the Ctl. Interestingly, the difference in absorbance value be-tween Ctl and cements went down to 25% over the 3- to 7-day cultureperiod, implying that the cell proliferation of hDPCs cultured on ce-ments gradually plateaued and shifted to a differentiation stage. Fluores-cent staining of actin cytoskeleton and scanning electron microscopywere used to better monitor cell adhesion and spreading and the inter-action between the cement surface and hDPCs. As can be seen inFigure 3C, during the initial 6 hours of the culture period, hDPCscultured on MesoCS adhere better through a biological cell adhesionprocess, including substrate attachment, spreading, and cytoskeletondevelopment. A similar trend occurred in the scanning electron micro-scopic micrographs of hDPCs cultured on cements for 1 day. However,the hDPCs becamemore flattened, and the difference in themorphologyof cells was negligible after 3 days of culture.

Odontogenesis DifferentiationAnalysis of quantitative examination data shows the ALP activity of

cells cultured on the different specimens to increase over time, andthere is no significant difference (P < .05) between CS and MesoCS(Fig. 4A). The ALP activity of the cells cultured on MesoCS FGF-2 washigher than that of the cells cultured on other groups; there was a sta-tistically significant difference (P< .05) for all culture periods. The pro-tein expression of OC, DMP-1, and DSP on the hDPCs was measuredusing enzyme-linked immunosorbent assay (Fig. 4B–D). On days 7and 14, a significant difference (P < .05) was found between the rawmaterials and the FGF-2–load materials. In addition, all protein expres-sions of hDPCs cultured with MesoCS/FGF-2 were higher than thosecultured with CS/FGF-2 on the 14th day. When containing FGF-2, the

color of the alizarin red S staining presented from light to deep purple,and the amounts of calcium mineral deposits increased on day 7(Fig. 4E). There are significant differences (P < .05) between theamounts of FGF-2 loaded in CS and MesoCS. These results indicatedthat MesoCS-loaded FGF-2 might stimulate mineralized nodule forma-tion and calcium deposition through FGF-2 release.

DiscussionIdeally, endodontic materials should be biocompatible and have

satisfactory physicochemical properties. Although CS-based cement isa biocompatible material, it is generally not a very effective drug carrier.In this study, newMesoCS nanoparticles were successfully prepared us-ing a synthesis strategy. The MesoCS nanoparticles were shown to haveuniform spherical morphology, ordered mesoporous channels, andsustained release of Ca and Si ions. The as-prepared MesoCS nanopar-ticles have distinct mesopores with a size distribution between 150 and200 nm. Small-angle XRD, wide-angle XRD, and N2 adsorption-desorption analyses confirmed the MesoCS nanoparticles to have mes-oporous structures with CS composition and high surface area (462m2/g) and pore volume (0.62 cm3/g) as well as uniformmesopore size dis-tribution (3.05 nm). The MesoCS nanoparticles have 2 important qual-ities. First, the nanostructure of MesoCS allows it to be prepared as aninjectable paste to fill root canals. Second, the MesoCS nanoparticleshave high surface areas and pore volumes, which should make themuseful antibiotic carriers.

The formation of the apatite that precipitates on the surface of ma-terials has proven useful in predicting the hard tissue bonding ability ofmaterial in vitro (5, 11). After soaking in SBF for 1 day, the MesoCSformed distinct apatite precipitates, indicating that MesoCS has asimilar ability to that of conventional cement in developing apatitemineralization. The excellent apatite formation of the MesoCSnanoparticles may play an important role in maintaining theirbioactivity for potential hard tissue regeneration (31). The Si-OH func-tional groups of CS-based materials have been shown to act as nucle-ation centers for apatite precipitation (32, 33). The Si ions releasedfromMesoCS nanoparticles possibly originate from less-ordered hydra-tion products and can significantly increase apatite growth by promot-ing local Si ion supersaturating, as reported in a previous study (2).

Multifunctional properties of biomaterials, such as the ability topromote osteogenic and angiogenic activity as well as antibacterialbehavior, are essential in developing bioactive hard tissue regeneration

Figure 2. (A) The release amount of GENT from CS and MesoCS cements in PBS (pH = 7.4) at 37!C. (B) The antibacterial activity of GENT-loaded CS and MesoCScements against S. aureus. #Statistically significant difference (P < .05) from the cements without loading GENT. *Statistically significant difference (P < .05) fromthe GENT-CS cement.

Regenerative Endodontics

JOE — Volume 43, Number 1, January 2017 Mesoporous Calcium Silicate Nanoparticles 73

materials (12, 34). This should allow the incorporation of antibiotics orfunctional proteins into bioactive materials and should also enableantibiotics or functional proteins to exert a synergistic effect on suchmaterials. The loading efficiency of proteins and antibiotics inmesoporous bioactive glass nanoparticles is significantly better thanthat of conventional bioactive glass because most drugs can be easilyloaded into the mesopores of the inner structure (35). Our data clearlyindicate that the released GENT from MesoCS nanoparticles has distinctantibacterial activity compared with that of conventional CS cement. TheGENT-loaded MesoCS nanoparticles showed a better antibacterial effectagainst a S. aureus strain than CS cement. The reasons for the antibac-terial effect of the MesoCS nanoparticle is that it releases calcium ionsthat lead to an alkaline microenvironment and an increased pH value,which may inhibit bacterial growth. In addition to its own antibacterialeffect, MesoCS nanoparticles could also serve to deliver antibiotics tofurther inhibit bacteria growth. Hence, the MesoCS nanoparticles arepromising biomaterials that can be developed into excellent antibacte-rial agents.

Several studies have shown that FGF-2 promotes alveolar bone for-mation and cementum regeneration in dogs (36). Moreover, FGF-2 isknown to be involved in all stages of bone formation, including osteo-blast cell proliferation, differentiation, maturation, and apoptosis.Recently, FGF-2 was incorporated into bioglass microparticles and ex-

hibited a slow-release pattern of FGF-2 from microparticles (37). Theobservation of the proliferative potential of stem cells was consideredto reflect the biological efficacy of FGF-2 release. On the other hand, aproblem with this kind of administration is that a large amount ofFGF-2 protein was used because of its rapid in vivo release (38). Toovercome this disadvantage when attempting to treat more severebone defects with FGF-2, substitution might be the best approach.Many studies have shown positive results from developing compositesfor the delivery of FGF-2 based on sulfated or natural materials (39).The authors of the present study have previously shown that CS promotesvarious types of cell proliferation, differentiation into odontoblasts, andmineralization (40–42). Interestingly, the phosphorylated ERK and p38protein of cells have been shown to be involved in the CS-inducedsignaling pathway but not Jun kinases (26). These data prove that fibro-blast growth factor receptor stimulates cell proliferation and differenti-ation through the ERK/MAPK pathway in hDPCs. It seems that the CS-basedmaterials control the lineage allocation of cells from the fibroblastto the osteoblast lineage via mechanisms involving activation of the fibro-blast growth factor receptor/ERK/MAPK pathway. Therefore, we mixedCS with FGF-2 for local delivery in order to evaluate the capacity of thesesystems for inducing cell proliferation and differentiation.

In summary, bioactive MesoCS nanoparticles with a size of200 nm were synthesized using a facile template method. These

Figure 3. (A) Adhesion and (B) proliferation of hDPCs cultured on different cements. *Statistically significant difference (P < .05) from the Ctl group. (C) Fluo-rescent images of (blue) nuclei and (red) cytoskeleton staining in hDPCs cultured on cements for 6 hours; scanning electron microscopic micrographs of hDPCscultured on cements for 1 and 3 days.

Regenerative Endodontics

74 Huang et al. JOE — Volume 43, Number 1, January 2017

MesoCS nanoparticles were shown to have a high specific surfacearea and pore volume with internal mesopores (average poresize = 3.05 nm). In addition, the MesoCS nanoparticles induceda significant amount of apatite precipitate with excellent mineraliza-tion behavior. Furthermore, the MesoCS nanoparticles can be usedas a drug carrier to maintain the sustained release of GENT. Boththe MesoCS nanoparticles themselves and their ability to deliverdrugs efficiently inhibited the growth of bacteria. Differentiationinto odontoblasts was confirmed by ALP activity and protein levels

of odontoblastic protein, including OC, DMP-1, and DSP. Based onthis work, the MesoCS nanoparticles produced in this work arepotentially useful endodontic materials for dental pulp tissue regen-erative materials with biocompatibility and odontogenesis.

AcknowledgmentsChing-Yuang Huang and Tsui-Hsien Huang contributed

equally to this work.

Figure 4. (A) ALP activity of hDPCs cultured on specimens for 3 and 7 days. (B) OC, (C) DMP-1, and (D) DSP concentrations of hDPCs cultured on specimens for7 and 14 days. (E) Alizarin red S staining and quantification of calcium mineral deposits of hDPCs cultured on specimens for 14 days. #Significant difference(P < .05) compared with raw material. *Significant difference (P < .05) compared with CS FGF-2.

Regenerative Endodontics

JOE — Volume 43, Number 1, January 2017 Mesoporous Calcium Silicate Nanoparticles 75

Supported by a grant from the Chung Shan Medical Universityand Show Chwan Memorial Hospital under the project (CSMU-SHOW-104-02) and the China Medical University Hospital (DMR-106-002) of Taiwan.

The authors deny any conflicts of interest related to this study.

References1. Huang SC, Wu BC, Kao CT, et al. Role of the p38 pathway in mineral trioxide

aggregate-induced cell viability and angiogenesis-related proteins of dental pulpcell in vitro. Int Endod J 2015;48:236–45.

2. Liu CH, Huang TH, Hung C Jr, et al. The synergistic effects of fibroblast growthfactor-2 and mineral trioxide aggregate on an osteogenic accelerator in vitro.Int Endod J 2014;47:843–53.

3. Hung C Jr, Kao CT, Shie MY, Huang TH. Comparison of host inflammatory responsesbetween calcium-silicate base material and intermediate restorative material. J DentSci 2014;9:158–64.

4. Hsu TT, Yeh CH, Kao CT, et al. Antibacterial and odontogenesis efficacy of mineraltrioxide aggregate combined with CO2 laser treatment. J Endod 2015;41:1073–80.

5. Huang MH, Shen YF, Hsu TT, et al. Physical characteristics, antimicrobial and odon-togenesis potentials of calcium silicate cement containing hinokitiol. Mater Sci Eng CMater Biol Appl 2016;65:1–8.

6. Huang MH, Kao CT, Chen YW, et al. The synergistic effects of chinese herb andinjectable calcium silicate/b-tricalcium phosphate composite on an osteogenicaccelerator in vitro. J Mater Sci Mater Med 2015;26:161.

7. Hsu TT, Kao CT, Chen YW, et al. The synergistic effects of CO2 laser treatment withcalcium silicate cement of antibacterial, osteogenesis and cementogenesis efficacy.Laser Phys Lett 2015;12:055602.

8. Lai WY, Chen YW, Kao CT, et al. Human dental pulp cells responses to apatite pre-cipitation from dicalcium silicates. Materials 2015;8:4491–504.

9. Chen YW, Yeh CH, Shie MY. Stimulatory effects of the fast setting and degradable Ca–Si–Mg cement on both cementogenesis and angiogenesis differentiation of humanperiodontal ligament cells. J Mater Chem B 2015;3:7099–108.

10. Tu MG, Chen YW, Shie MY. Macrophage-mediated osteogenesis activation in co-culture with osteoblast on calcium silicate cement. J Mater Sci Mater Med 2015;26:276.

11. Chen YW, Hsu TT, Wang K, Shie MY. Preparation of the fast setting and degrading Ca-Si-Mg cement with both odontogenesis and angiogenesis differentiation of humanperiodontal ligament cells. Mater Sci Eng C Mater Biol Appl 2016;60:374–83.

12. Chang NJ, Chen YW, Shieh DE, et al. The effects of injectable calcium silicate-basedcomposites with the Chinese herb on an osteogenic accelerator in vitro. BiomedMater 2015;10:055004.

13. Islam MS, Choi HN, Choi WS, Lee HJ. Polyelectrolyte-mediated hierarchical mesopo-rous calcium silicates: a platform for drug delivery carrier with ultrahigh loadingcapacity and controlle release behavior. J Mater Chem B 2015;3:1001–9.

14. Qi G, Jiaoyang Z, Yuan Y, et al. A dual-delivery system of pH-responsive chitosan-functionalized mesoporous silica nanoparticles bearing BMP-2 and dexamethasonefor enhanced bone regeneration. J Mater Chem B 2014;3:2056–66.

15. Kresge CT, Leonowicz ME, Roth WJ, Vartuli JC. Ordered mesoporous molecularsieves synthesized by a liquid-crystal template mechanism. Nature 1992;359:710–2.

16. Xiong L, Du X, Shi B, et al. Tunable stellate mesoporous silica nanoparticles forintracellular drug delivery. J Mater Chem B 2015;3:1712–21.

17. Su CC, Kao CT, Hung C Jr, et al. Regulation of physicochemical properties, osteo-genesis activity, and fibroblast growth factor-2 release ability of b-tricalcium phos-phate for bone cement by calcium silicate. Mater Sci Eng C Mater Biol Appl 2014;37:156–63.

18. Ho CC, Ding SJ. The pH-controlled nanoparticles size of polydopamine for anti-cancer drug delivery. J Mater Sci Mater Med 2013;24:2381–90.

19. Shadjou N, Hasanzadeh M. Bone tissue engineering using silica-based mesoporousnanobiomaterials: recent progress. Mater Sci Eng C Mater Biol Appl 2015;55:401–9.

20. Hsu FY, Lu MR, Weng RC, Lin HM. Hierarchically biomimetic scaffold of a collagen-mesoporous bioactive glass nanofiber composite for bone tissue engineering. Bio-med Mater 2015;10:025007.

21. Shen YF, Ho CC, Shie MY, et al. Hinokitiol-loaded mesoporous calcium silicatenanoparticle induce apoptotic cell death through regulation of the function ofMDR1 in lung adenocarcinoma cells. Materials 2016;9:306.

22. Marie PJ, Miraoui H, S"ev#ere N. FGF/FGFR signaling in bone formation: progress andperspectives. Growth Factors 2012;30:117–23.

23. Marie PJ. Fibroblast growth factor signaling controlling bone formation: an update.Gene 2012;498:1–4.

24. Liu CH, Hung C Jr, Huang TH, et al. Odontogenic differentiation of human dentalpulp cells by calcium silicate materials stimulating via FGFR/ERK signaling pathway.Mater Sci Eng C Mater Biol Appl 2014;43:359–66.

25. Montero A, Okada Y, Tomita M, et al. Disruption of the fibroblast growth factor-2gene results in decreased bone mass and bone formation. J Clin Invest 2000;105:1085–93.

26. Shie MY, Ding SJ. Integrin binding and MAPK signal pathways in primary cell re-sponses to surface chemistry of calcium silicate cements. Biomaterials 2013;34:6589–606.

27. Kim HJ, Kim JH, Bae SC, et al. The protein kinase C pathway plays a central role in thefibroblast growth factor-stimulated expression and transactivation activity of Runx2.J Biol Chem 2003;278:319–26.

28. Xiao G, Jiang D, Gopalakrishnan R, Franceschi RT. Fibroblast growth factor 2 induc-tion of the osteocalcin gene requires MAPK activity and phosphorylation of the oste-oblast transcription factor, Cbfa1/Runx2. J Biol Chem 2002;277:36181–7.

29. David Chen CH, Chen CC, Shie MY, et al. Controlled release of gentamicin from cal-cium phosphate/alginate bone cement. Mater Sci Eng C Mater Biol Appl 2011;31:334–41.

30. Yeh CH, Chen YW, Shie MY, Fang HY. Poly(dopamine)-assisted immobilization of XuDuan on 3D printed poly(lactic acid) scaffolds to up-regulate osteogenic and angio-genic markers of bone marrow stem cells. Materials 2015;8:4299–315.

31. Chen YW, Ho CC, Huang TH, et al. The ionic products from mineral trioxide aggre-gate–induced odontogenic differentiation of dental pulp cells via activation of theWnt/b-catenin signaling pathway. J Endod 2016;42:1062–9.

32. Kao CT, Huang TH, Chen YJ, et al. Using calcium silicate to regulate the physico-chemical and biological properties when using b-tricalcium phosphate as bonecement. Mater Sci Eng C Mater Biol Appl 2014;43:126–34.

33. Zhou Y, Wu C, Xiao Y. The stimulation of proliferation and differentiation of peri-odontal ligament cells by the ionic products from Ca7Si2P2O16 bioceramics.Acta Biomater 2012;8:2307–16.

34. Cheng YL, Chen YW, Wang K, Shie MY. Enhanced adhesion and differentiation ofhuman mesenchymal stem cell inside apatite-mineralized/poly(dopamine)-coatedpoly(ε-caprolactone) scaffolds by stereolithography. J Mater Chem B 2016;4:6307–15.

35. Wu C, Chang J. Mesoporous bioactive glasses: structure characteristics, drug/growthfactor delivery and bone regeneration application. Interface Focus 2012;2:292–306.

36. Anzai J, Kitamura M, Nozaki T, et al. Effects of concomitant use of fibroblastgrowth factor (FGF)-2 with beta-tricalcium phosphate (b-TCP) on the beagledog 1-wall periodontal defect model. Biochem Biophys Res Commun 2010;403:345–50.

37. Perez RA, El-Fiqi A, Park JH, et al. Therapeutic bioactive microcarriers: co-deliveryof growth factors and stem cells for bone tissue engineering. Acta Biomater 2014;10:520–30.

38. Nakamura T, Hara Y, Tagawa M, et al. Recombinant human basic fibroblast growthfactor accelerates fracture healing by enhancing callus remodeling in experimentaldog tibial fracture. J Bone Miner Res 1998;13:942–9.

39. Wu BC, Youn SC, Kao CT, et al. The effects of calcium silicate cement/fibroblastgrowth factor-2 composite on osteogenesis accelerator in human dental pulp cells.J Dent Sci 2015;10:145–53.

40. Hung C Jr, Hsu HI, Lin CC, et al. The role of integrin av in proliferation and differ-entiation of human dental pulp cell response to calcium silicate cement. J Endod2014;40:1802–9.

41. Huang SH, Chen YJ, Kao CT, et al. Physicochemical properties and biocompatibilityof silica doped b-tricalcium phosphate for bone cement. J Dent Sci 2014;10:282–90.

42. Huang SC, Wu BC, Ding SJ. Stem cell differentiation-induced calcium silicate cementwith bacteriostatic activity. J Mater Chem B 2015;3:570–80.

Regenerative Endodontics

76 Huang et al. JOE — Volume 43, Number 1, January 2017