2.1

Synthesis of NACP

A spray-drying technique was used to fabricate NACP, as previously described [ ]. Calcium carbonate (CaCO ₃ ) and dicalcium phosphate anhydrous (CaHPO ₄ ) were dissolved in an acetic acid solution to obtain final Ca and PO ₄ ionic concentrations of 8 and 5.333 mmol/L, respectively. Formation of NACP was initiated by atomizing a suspension of CaP micro-droplets followed by drying in a heated chamber to generate solid particles. The resulted NACP has a mean particle size of 116 nm, as reported previously [ ].

2.2

Orthodontic adhesives

Two experimental orthodontic adhesives were formulated. The first consists of 44.5 wt percent (wt%) pyromellitic glycerol dimethacrylate (PMGDM) (International Laboratory USA, South San Francisco, CA, USA) and 39.5 wt% ethoxylated bisphenol A dimethacrylate (EBPADMA; MilliporeSigma, St. Louis, MO, USA; Table 1 ). Being an acidic adhesive monomer [ ], PMGDM possesses recharging capability by chelating calcium ions from a recharging solution [ ]. Ten wt% 2-hydroxyethyl methacrylate (HEMA; MilliporeSigma) and 5 wt% bisphenol A glycidyl dimethacrylate (BisGMA) (MilliporeSigma) were added to the PMGDM-EBPADMA mixture. These weight fractions were selected because previous studies showed that addition of 10% HEMA and 5% BisGMA into the PMGDM-EBPADMA resin comonomer blend increased its bond strength to dentin [ ]. Benzoyl peroxide (BPO, 0.8 wt%; Alfa Aesar, Heysham, Lancashire, United Kingdom) was added to enable chemical-cure, and camphorquinone (CQ, 0.2 wt%; MilliporeSigma) was added as photoinitiator to provide the capacity for the resin blend to be light-cured. The resin was designated as PEHB. The polymerizable QAS, MAE-DB ( Fig. 1 ), was added to the PEHB at a weight fraction of 5% ( Table 1 ) according to the previous study [ ]. For the second experimental orthodontic adhesive, NACP fillers were added to PEHB at a weight fraction of 40% [ ]. The MAE-DB resin monomer was subsequently added to the filled adhesive mixture at a weight fraction of 5% ( Table 1 ). According to the results of preliminary experiment, NACP filler level > 40% and MAE-DB filler level >5% were not used to avoid the adverse effect on bracket-enamel bond strength.

Structure of 2-methacryloxylethyl dodecyl methyl ammonium bromide (MAE-DB).

Transbond XT Light Cure Adhesive (TB; 3 M Unitek, Monrovia, CA, USA) was used as the control. The adhesive/cement was used in accordance with the manufacturer’s instructions. According to the manufacturer, TB contains silane-treated quartz (70–80 wt%), bisphenol A diglycidyl ether dimethacrylate (10–20 wt%), bisphenol-A-bis (2-hydroxyethyl) dimethacrylate (5–10 wt%), silane-treated silica (< 2 wt%) and diphenyliodonium hexafluorophosphate (< 0.2 wt%).

The three experimental groups were: (1) TB; (2) PEHB + 5% MAE-DB [designated as PD]; and (3) PEHB + 5% MAE-DB + 40% NACP [designated as PND].

2.3

Preparation of enamel-bracket bonding specimens

Eighty non-carious human premolars were collected from the clinics of the School of Stomatology, Fourth Military Medical University (FMMU), Xi’an, China. The protocol was approved by the FMMU Institutional Review Board. All experiments were performed in accordance with approved guidelines and regulations. The teeth were disinfected in a 0.005% promodyne solution for 4 h and stored at 4 °C in deionized water. The premolars were randomly distributed among three adhesive groups, each consisting of 26 teeth. The addition two teeth were used as control during morphologic examination.

The buccal enamel surface of each tooth was etched with 35% phosphoric acid (Scotchbond, 3 M ESPE, St. Paul, MN, USA) for 30 s, rinsed with water for 30 s, and dried thoroughly using oil-and-moisture-free air. Prior to bracket placement, a thin uniform coat of Transbond XT primer was applied to the etched surfaces and cured for 10 s. Each adhesive was applied to the base of the bracket (Shinye, Hangzhou, China), which was lightly placed on the tooth surface and pressed firmly into the final position. Excess adhesive around the bracket was removed without disturbing the bracket. The adhesive was polymerized from four sides (occlusal, gingival, mesial and distal) of the bracket for 10 s each, using a light-emission diode curing unit (Elipar S10, 3 M ESPE, St. Paul, MN, USA) operated in the standard curing mode. The light-curing unit was kept approximately 1 mm from the bracket. Bonded specimens were stored in water at 37 °C for 24 h. All the brackets were bonded by an experienced clinician who was blinded with respect to the kind of adhesive utilized during bracket placement.

2.4

Aging

The bonded premolars in each adhesive group were divided into 3 subgroups. The first subgroup was subjected to shear bond strength testing after the specimens were immersed in deionized water for 24 h (referred to as 1-day in water, n = 10). The second subgroup (n = 10) was subjected to shear bond strength testing after the specimens were aged in demineralization solution (pH 4) for 28 days. The third subgroup (n = 6) was used for microhardness evaluation after the specimens were aged in demineralization solution (pH 4) for 28 days. The demineralization solution consisted of 3.0 mmol/L CaCl ₂ , 1.8 mmol/L KH ₂ PO ₄ , 0.1 mol/L lactic acid, and 1% carboxymethylcellulose, with pH adjusted to 4 with KOH [ ]. This mimicked the biofilm acids that the orthodontic adhesive-enamel bonded areas would encounter in vivo , which could potentially degrade the resin-enamel bond strength [ ]. Aging in demineralization solution was used to simulate cariogenic challenge on the bonded interface, for evaluating the bond durability and remineralising ability of the experimental adhesives [ ]. This was accomplished by placing each bonded tooth flat in a tube with the buccal surface in contact with the bottom of the tube. The bonded interface was completely immersed in 1.5 mL of the demineralization solution. The solution was changed daily so that the pH of the solution was consistently maintained at 4.

2.5

Bracket shear bond testing and adhesive remnant index

Shear bond strengths of the orthodontic brackets to enamel were tested as previously described [ ]. Each tooth was embedded vertically in self-curing acrylic resin (Lang Dental Manufacturing, Wheeling, IL, USA). The buccal axis of the clinical crown was oriented so that the labial surface was parallel to the applied force during bond testing. To measure the shear bond strength, a chisel attached to a Universal Testing Machine (AGS-10kNG, Shimadzu, Japan) was positioned over the upper part of the bracket base and parallel to the bonded interface. An occlusogingival load was applied at a cross-head speed of 0.5 mm/min until the bracket detached [ ].

After the bracket was debonded, each tooth surface was observed under a stereomicroscope (SMZ1500, Nikon, Japan) to examine the failure mode. The adhesive remnant index (ARI) was determined based on the presence of remaining adhesive on enamel. Evaluation was performed using the following criteria [ ]: 0 = no adhesive remaining on enamel; 1 = less than half of the adhesive remaining on enamel; 2 = more than half of the adhesive remaining on enamel; 3 = all the adhesive remaining on enamel.

2.6

Antibacterial activity

Experimental adhesives and commercial control adhesive were placed individually into disk-shaped glass molds (inner diameter: 10 mm; depth: 1.5 mm) to prepare polymerized resin disks (n = 15). The top and bottom surfaces of each adhesive were covered with a Mylar strip and a microscope slide; the latter was pressed gently to remove excess material. The adhesives were light-polymerized for 60 s on each side with the light-curing light [ ]. After the adhesive disks were removed from the mold, they were immersed in distilled water with agitation for 1 h to remove any uncured monomer [ ]. The disks were subsequently sterilized with ethylene oxide and degassed in a fume hood for 48 h prior to testing [ ].

Streptococcus mutans (UA159; State Key Laboratory of Military Stomatology, School of Stomatology, FMMU, Xi’an, China) was cultured overnight at 37 °C in Brain Heart Infusion (BHI) broth (Difco, Detroit, MI, USA) in an anaerobic atmosphere enriched with 5% CO ₂ . The resulting bacterial suspension was adjusted to an optical density of 0.5 at 600 nm and diluted one hundred folds with fresh BHI for further use [ ].

The sterile resin disks in each adhesive group were equally divided into five subgroups and placed in five 24-well plates containing 2-mL of BHI broth per well. Twenty microliter of the diluted S. mutans suspension was then added to each well. Because of the difference in surface roughness of resin disks prepared from PD with no nanofillers and the nanofiller-containing PND, it was anticipated that the bacterial killing efficacy of the co-polymerized MAE-DB resin in the two adhesives would be masked by the difference in bacteria adherence between the two materials. According, different bacteria culture time-points were used to examine the percentage of live bacteria within biofilms grown on resin disks prepared from the PD and PND groups until no more difference between the two materials could be detected. Accordingly, biofilms were evaluated by confocal laser scanning microscopy after 6, 12, 24, 48, and 72 h of anaerobic culture.

At each designated time-point, the biofilm-coated disks from the 5 subgroups were rinsed with sterile phosphate-buffered saline (PBS) to remove loose bacteria. The resin disks were stained with Live/Dead BacLight Bacterial Viability Kit L13152 (Molecular Probes, Invitrogen, Eugene, OR, USA). With live/dead staining, live bacteria produce green fluorescence upon staining with Syto 9, and bacteria with compromised membranes produce red fluorescence upon staining with propidium iodide. After incubation with the fluorescent dyes, the stained resin disks were rinsed with distilled water and examined with a confocal laser scanning microscope (CLSM; FluoView FV1000, Olympus, Tokyo, Japan). Excitation with a 488-nm laser resulted in green fluorescence emission of live bacteria; excitation with a 543-nm laser resulted in red fluorescence emission of bacteria with damaged membranes. Five 316.8 μm x 316.8 μm images were taken from each disk at random locations (5 images for each adhesive at each culture time-point). The live/dead ratio of biofilm cells was quantified using the Imaris software (Bitplane AG, Zurich, Switzerland).

The remaining 2 resin disks from each subgroup was used to prepared 6-h or 72-h biofilms for scanning electron microscopy. The biofilm-coated disks were gently rinsed with sterile PBS, soaked in 3% glutaraldehyde at 4 °C overnight, washed twice with PBS, dehydrated in a graded series of ethanol solutions, and dried in a critical-point drier. The specimens with sputtered-coated with gold-palladium and examined with a field emission-scanning electron microscope (FE-SEM; S-4800, Hitachi, Tokyo, Japan) at 5 kV.

2.7

Ca and P ion release

Sodium chloride solution (133 mmol/L) was buffered to pH 4 with 50 mmol/L lactic acid to measure the ion release, simulating a low pH, cariogenic condition [ ]. Specimens measuring 12 mm × 2 mm × 2 mm were fabricated using the three orthodontic adhesives. Three specimens from each group were immersed in 50 mL of the pH-adjusted NaCl solution to yield a specimen volume/solution of 2.9 mm ³ /mL, based on previous studies [ ]. The concentrations of Ca and P released from the specimens were measured at 1, 3, 5, 7, 14, 21, 28, 35 and 42 days, using a previously-reported protocol [ ]. At each designated time-period, a 0.5 mL aliquot of solution was retrieved from each storage medium and replaced with fresh NaCl solution. The pH of the incubating solutions was monitored and adjusted to pH 4 with 50 mmol/L lactic acid using a combination pH electrode (Orion, Cambridge, MA, USA) [ ]. The retrieved aliquots were analyzed for Ca and P concentrations using an inductively-coupled plasma emission spectrometer (ICP-MS; Model iCAP-6300, Thermo Fisher Scientific, USA) using known standards and calibration curves. Experiments were conducted in triplicates (n = 3).

2.8

Microhardness

Microhardness was used an indirect indicator of the extent of enamel demineralization and remineralization. Six teeth from each adhesive group were used for the experiment. Each bonded tooth was placed flat inside a tube with the buccal surface in contact with the bottom of the tube. The bracket-buccal bonded interface was completely immersed in 1.5 mL of demineralization solution (pH 4), while the lingual surface of each bonded tooth was not exposed to the solution. The demineralization solution was changed daily for 28 days prior to retrieval of the specimens for microhardness determination.

An operator who was blinded to the group allocation performed the analysis. Each tooth was hemi-sectioned in a sagittal direction through the center of the bracket into mesial and distal halves, using a low-speed diamond saw (Kejing, Shenyang, China) under water cooling. The sectioned specimens were embedded in self-curing epoxy-resin (EpoKwick, Buehler, Lake Bluff, IL, USA), leaving the cut face exposed. The resin-embedded specimens were polished with 3 grades of abrasive papers (320, 600 and 1200 grit). Final polishing was achieved with 1-μm diameter diamond cream and polishing-cloth disk (Buehler).

Microhardness measurement was conducted using a hardness tester (HXD-1000 TM/LCD, Digital Micro Hardness Tester, Shanghai Taiming Optical Instrument Co. Ltd., Shanghai, China) equipped with a Vickers diamond indenter at a load of 50 gf and a dwell time of 15 s. Four indentations were made in each embedded tooth-half from 4 locations and 2 depths (12 tooth-halves; n = 12). On the buccal surface, indentations were made under the bracket and at the occlusal and cervical edges of the bracket. An additional indentation made at the lingual enamel surface served as the internal control. For each location, two indentations were made at 30 μm and 120 μm from the external enamel surface. The latter was used to represent the part of the tooth enamel that was not affected by surface demineralization.

2.9

Surface roughness

After shear bond strength testing, specimens that had been immersed in demineralization solution (pH 4) for 28 days prior to testing were examined by FE-SEM and atomic force microscopy (AFM, Series 5500, Agilent, Santa Clara, CA, USA). Four debonded specimens from the TB, PD and PND groups with less than half of the bonded enamel adhesive covered by adhesive (i.e. ARI score 1) were used for examination. Two additional non-carious premolars were used as control to record the original enamel surface morphology. Each tooth was cut with a low-speed diamond saw into two halves along the mesiodistal plane. The buccal half specimens were ultrasonicated for 15 min to remove debris and dried with nitrogen gas [ ]. Two buccal halves were sputter-coated with gold/palladium for SEM, using the S-4800 FE-SEM at 5 kV. The remaining 2 specimens were used for AFM examination and surface roughness examination. Surface profiling was performed using a tapping model etched silicon probe (Keysight, Böblingen, Germany). Four AFM images were collected from each location of a tooth-half. Surface roughness was determined from the statistical data derived from those images (n = 4 × 2 tooth-halves = 8) and expressed as root mean square height (Sq, in μm).

2.10

Statistical analyses

Interval data generated from the experiments were evaluated for their normality (Shapiro-Wilk test) and equal variance assumptions (modified Levene test) prior to the adoption of parametric statistical methods. Statistical significance for all analyses was pre-set set at α = 0.05.

Data derived from bonding of orthodontic brackets to enamel (n = 10) were analyzed with two-factor analysis of variance (ANOVA) to examine the effect of “adhesive” and “aging”, and the interaction of these two factors on shear bond strength. Post-hoc pairwise comparisons were performed using the Holm-Šidák statistic. For analysis of the ARI results (n = 10), the Kruskal-Wallis H test was used to evaluate the scores among the three adhesive subgroups at each time-point; the Wilcoxon signed-rank test was used to compare the scores of each adhesive between the two time-points.

The percentages of live bacteria in biofilms taken with CLSM (n = 5) were analyzed with a two-factor ANOVA design to examine the effect of adhesive (TB, PD or PND) and culture time (6, 12, 24, 48 or 72 h), and the interaction of those 2 factors on antibacterial efficacy. Post-hoc pairwise comparisons were conducted using the Holm-Šidák statistic.

For microhardness evaluation, a 3-factor ANOVA design was originally planned for examining the effect of adhesive (TB, PB or PND), indentation location (occlusal, under bracket, cervical or lingual), the depth from which indentation was made (30 or 120 μm from enamel surface), and the interaction of those 3 factors on microhardness. However, the normality and equal variance assumptions of the data sets were not fulfilled despite different forms of non-linear transformation of the data sets. Accordingly, data derived from the 24 groups (n = 12) were analyzed using Kruskal-Wallis H test and Dunn’s multiple comparisons test.

For surface roughness evaluation, data from the control, TB PB and PND groups (n = 8) were logarithmically-transformed to satisfy the normality and homoscedasticity assumptions for the use of parametric statistical methods. The transformed data sets were analyzed with one-factor ANOVA and the Holm-Šidák method for pairwise comparisons.

2.1

Synthesis of NACP

A spray-drying technique was used to fabricate NACP, as previously described [ ]. Calcium carbonate (CaCO ₃ ) and dicalcium phosphate anhydrous (CaHPO ₄ ) were dissolved in an acetic acid solution to obtain final Ca and PO ₄ ionic concentrations of 8 and 5.333 mmol/L, respectively. Formation of NACP was initiated by atomizing a suspension of CaP micro-droplets followed by drying in a heated chamber to generate solid particles. The resulted NACP has a mean particle size of 116 nm, as reported previously [ ].

2.2

Orthodontic adhesives

Two experimental orthodontic adhesives were formulated. The first consists of 44.5 wt percent (wt%) pyromellitic glycerol dimethacrylate (PMGDM) (International Laboratory USA, South San Francisco, CA, USA) and 39.5 wt% ethoxylated bisphenol A dimethacrylate (EBPADMA; MilliporeSigma, St. Louis, MO, USA; Table 1 ). Being an acidic adhesive monomer [ ], PMGDM possesses recharging capability by chelating calcium ions from a recharging solution [ ]. Ten wt% 2-hydroxyethyl methacrylate (HEMA; MilliporeSigma) and 5 wt% bisphenol A glycidyl dimethacrylate (BisGMA) (MilliporeSigma) were added to the PMGDM-EBPADMA mixture. These weight fractions were selected because previous studies showed that addition of 10% HEMA and 5% BisGMA into the PMGDM-EBPADMA resin comonomer blend increased its bond strength to dentin [ ]. Benzoyl peroxide (BPO, 0.8 wt%; Alfa Aesar, Heysham, Lancashire, United Kingdom) was added to enable chemical-cure, and camphorquinone (CQ, 0.2 wt%; MilliporeSigma) was added as photoinitiator to provide the capacity for the resin blend to be light-cured. The resin was designated as PEHB. The polymerizable QAS, MAE-DB ( Fig. 1 ), was added to the PEHB at a weight fraction of 5% ( Table 1 ) according to the previous study [ ]. For the second experimental orthodontic adhesive, NACP fillers were added to PEHB at a weight fraction of 40% [ ]. The MAE-DB resin monomer was subsequently added to the filled adhesive mixture at a weight fraction of 5% ( Table 1 ). According to the results of preliminary experiment, NACP filler level > 40% and MAE-DB filler level >5% were not used to avoid the adverse effect on bracket-enamel bond strength.

Structure of 2-methacryloxylethyl dodecyl methyl ammonium bromide (MAE-DB).

Transbond XT Light Cure Adhesive (TB; 3 M Unitek, Monrovia, CA, USA) was used as the control. The adhesive/cement was used in accordance with the manufacturer’s instructions. According to the manufacturer, TB contains silane-treated quartz (70–80 wt%), bisphenol A diglycidyl ether dimethacrylate (10–20 wt%), bisphenol-A-bis (2-hydroxyethyl) dimethacrylate (5–10 wt%), silane-treated silica (< 2 wt%) and diphenyliodonium hexafluorophosphate (< 0.2 wt%).

The three experimental groups were: (1) TB; (2) PEHB + 5% MAE-DB [designated as PD]; and (3) PEHB + 5% MAE-DB + 40% NACP [designated as PND].

2.3

Preparation of enamel-bracket bonding specimens

Eighty non-carious human premolars were collected from the clinics of the School of Stomatology, Fourth Military Medical University (FMMU), Xi’an, China. The protocol was approved by the FMMU Institutional Review Board. All experiments were performed in accordance with approved guidelines and regulations. The teeth were disinfected in a 0.005% promodyne solution for 4 h and stored at 4 °C in deionized water. The premolars were randomly distributed among three adhesive groups, each consisting of 26 teeth. The addition two teeth were used as control during morphologic examination.

The buccal enamel surface of each tooth was etched with 35% phosphoric acid (Scotchbond, 3 M ESPE, St. Paul, MN, USA) for 30 s, rinsed with water for 30 s, and dried thoroughly using oil-and-moisture-free air. Prior to bracket placement, a thin uniform coat of Transbond XT primer was applied to the etched surfaces and cured for 10 s. Each adhesive was applied to the base of the bracket (Shinye, Hangzhou, China), which was lightly placed on the tooth surface and pressed firmly into the final position. Excess adhesive around the bracket was removed without disturbing the bracket. The adhesive was polymerized from four sides (occlusal, gingival, mesial and distal) of the bracket for 10 s each, using a light-emission diode curing unit (Elipar S10, 3 M ESPE, St. Paul, MN, USA) operated in the standard curing mode. The light-curing unit was kept approximately 1 mm from the bracket. Bonded specimens were stored in water at 37 °C for 24 h. All the brackets were bonded by an experienced clinician who was blinded with respect to the kind of adhesive utilized during bracket placement.

2.4

Aging

The bonded premolars in each adhesive group were divided into 3 subgroups. The first subgroup was subjected to shear bond strength testing after the specimens were immersed in deionized water for 24 h (referred to as 1-day in water, n = 10). The second subgroup (n = 10) was subjected to shear bond strength testing after the specimens were aged in demineralization solution (pH 4) for 28 days. The third subgroup (n = 6) was used for microhardness evaluation after the specimens were aged in demineralization solution (pH 4) for 28 days. The demineralization solution consisted of 3.0 mmol/L CaCl ₂ , 1.8 mmol/L KH ₂ PO ₄ , 0.1 mol/L lactic acid, and 1% carboxymethylcellulose, with pH adjusted to 4 with KOH [ ]. This mimicked the biofilm acids that the orthodontic adhesive-enamel bonded areas would encounter in vivo , which could potentially degrade the resin-enamel bond strength [ ]. Aging in demineralization solution was used to simulate cariogenic challenge on the bonded interface, for evaluating the bond durability and remineralising ability of the experimental adhesives [ ]. This was accomplished by placing each bonded tooth flat in a tube with the buccal surface in contact with the bottom of the tube. The bonded interface was completely immersed in 1.5 mL of the demineralization solution. The solution was changed daily so that the pH of the solution was consistently maintained at 4.

2.5

Bracket shear bond testing and adhesive remnant index

Shear bond strengths of the orthodontic brackets to enamel were tested as previously described [ ]. Each tooth was embedded vertically in self-curing acrylic resin (Lang Dental Manufacturing, Wheeling, IL, USA). The buccal axis of the clinical crown was oriented so that the labial surface was parallel to the applied force during bond testing. To measure the shear bond strength, a chisel attached to a Universal Testing Machine (AGS-10kNG, Shimadzu, Japan) was positioned over the upper part of the bracket base and parallel to the bonded interface. An occlusogingival load was applied at a cross-head speed of 0.5 mm/min until the bracket detached [ ].

After the bracket was debonded, each tooth surface was observed under a stereomicroscope (SMZ1500, Nikon, Japan) to examine the failure mode. The adhesive remnant index (ARI) was determined based on the presence of remaining adhesive on enamel. Evaluation was performed using the following criteria [ ]: 0 = no adhesive remaining on enamel; 1 = less than half of the adhesive remaining on enamel; 2 = more than half of the adhesive remaining on enamel; 3 = all the adhesive remaining on enamel.

2.6

Antibacterial activity

Experimental adhesives and commercial control adhesive were placed individually into disk-shaped glass molds (inner diameter: 10 mm; depth: 1.5 mm) to prepare polymerized resin disks (n = 15). The top and bottom surfaces of each adhesive were covered with a Mylar strip and a microscope slide; the latter was pressed gently to remove excess material. The adhesives were light-polymerized for 60 s on each side with the light-curing light [ ]. After the adhesive disks were removed from the mold, they were immersed in distilled water with agitation for 1 h to remove any uncured monomer [ ]. The disks were subsequently sterilized with ethylene oxide and degassed in a fume hood for 48 h prior to testing [ ].

Streptococcus mutans (UA159; State Key Laboratory of Military Stomatology, School of Stomatology, FMMU, Xi’an, China) was cultured overnight at 37 °C in Brain Heart Infusion (BHI) broth (Difco, Detroit, MI, USA) in an anaerobic atmosphere enriched with 5% CO ₂ . The resulting bacterial suspension was adjusted to an optical density of 0.5 at 600 nm and diluted one hundred folds with fresh BHI for further use [ ].

The sterile resin disks in each adhesive group were equally divided into five subgroups and placed in five 24-well plates containing 2-mL of BHI broth per well. Twenty microliter of the diluted S. mutans suspension was then added to each well. Because of the difference in surface roughness of resin disks prepared from PD with no nanofillers and the nanofiller-containing PND, it was anticipated that the bacterial killing efficacy of the co-polymerized MAE-DB resin in the two adhesives would be masked by the difference in bacteria adherence between the two materials. According, different bacteria culture time-points were used to examine the percentage of live bacteria within biofilms grown on resin disks prepared from the PD and PND groups until no more difference between the two materials could be detected. Accordingly, biofilms were evaluated by confocal laser scanning microscopy after 6, 12, 24, 48, and 72 h of anaerobic culture.

At each designated time-point, the biofilm-coated disks from the 5 subgroups were rinsed with sterile phosphate-buffered saline (PBS) to remove loose bacteria. The resin disks were stained with Live/Dead BacLight Bacterial Viability Kit L13152 (Molecular Probes, Invitrogen, Eugene, OR, USA). With live/dead staining, live bacteria produce green fluorescence upon staining with Syto 9, and bacteria with compromised membranes produce red fluorescence upon staining with propidium iodide. After incubation with the fluorescent dyes, the stained resin disks were rinsed with distilled water and examined with a confocal laser scanning microscope (CLSM; FluoView FV1000, Olympus, Tokyo, Japan). Excitation with a 488-nm laser resulted in green fluorescence emission of live bacteria; excitation with a 543-nm laser resulted in red fluorescence emission of bacteria with damaged membranes. Five 316.8 μm x 316.8 μm images were taken from each disk at random locations (5 images for each adhesive at each culture time-point). The live/dead ratio of biofilm cells was quantified using the Imaris software (Bitplane AG, Zurich, Switzerland).

The remaining 2 resin disks from each subgroup was used to prepared 6-h or 72-h biofilms for scanning electron microscopy. The biofilm-coated disks were gently rinsed with sterile PBS, soaked in 3% glutaraldehyde at 4 °C overnight, washed twice with PBS, dehydrated in a graded series of ethanol solutions, and dried in a critical-point drier. The specimens with sputtered-coated with gold-palladium and examined with a field emission-scanning electron microscope (FE-SEM; S-4800, Hitachi, Tokyo, Japan) at 5 kV.

2.7

Ca and P ion release

Sodium chloride solution (133 mmol/L) was buffered to pH 4 with 50 mmol/L lactic acid to measure the ion release, simulating a low pH, cariogenic condition [ ]. Specimens measuring 12 mm × 2 mm × 2 mm were fabricated using the three orthodontic adhesives. Three specimens from each group were immersed in 50 mL of the pH-adjusted NaCl solution to yield a specimen volume/solution of 2.9 mm ³ /mL, based on previous studies [ ]. The concentrations of Ca and P released from the specimens were measured at 1, 3, 5, 7, 14, 21, 28, 35 and 42 days, using a previously-reported protocol [ ]. At each designated time-period, a 0.5 mL aliquot of solution was retrieved from each storage medium and replaced with fresh NaCl solution. The pH of the incubating solutions was monitored and adjusted to pH 4 with 50 mmol/L lactic acid using a combination pH electrode (Orion, Cambridge, MA, USA) [ ]. The retrieved aliquots were analyzed for Ca and P concentrations using an inductively-coupled plasma emission spectrometer (ICP-MS; Model iCAP-6300, Thermo Fisher Scientific, USA) using known standards and calibration curves. Experiments were conducted in triplicates (n = 3).

2.8

Microhardness

Microhardness was used an indirect indicator of the extent of enamel demineralization and remineralization. Six teeth from each adhesive group were used for the experiment. Each bonded tooth was placed flat inside a tube with the buccal surface in contact with the bottom of the tube. The bracket-buccal bonded interface was completely immersed in 1.5 mL of demineralization solution (pH 4), while the lingual surface of each bonded tooth was not exposed to the solution. The demineralization solution was changed daily for 28 days prior to retrieval of the specimens for microhardness determination.

An operator who was blinded to the group allocation performed the analysis. Each tooth was hemi-sectioned in a sagittal direction through the center of the bracket into mesial and distal halves, using a low-speed diamond saw (Kejing, Shenyang, China) under water cooling. The sectioned specimens were embedded in self-curing epoxy-resin (EpoKwick, Buehler, Lake Bluff, IL, USA), leaving the cut face exposed. The resin-embedded specimens were polished with 3 grades of abrasive papers (320, 600 and 1200 grit). Final polishing was achieved with 1-μm diameter diamond cream and polishing-cloth disk (Buehler).

Microhardness measurement was conducted using a hardness tester (HXD-1000 TM/LCD, Digital Micro Hardness Tester, Shanghai Taiming Optical Instrument Co. Ltd., Shanghai, China) equipped with a Vickers diamond indenter at a load of 50 gf and a dwell time of 15 s. Four indentations were made in each embedded tooth-half from 4 locations and 2 depths (12 tooth-halves; n = 12). On the buccal surface, indentations were made under the bracket and at the occlusal and cervical edges of the bracket. An additional indentation made at the lingual enamel surface served as the internal control. For each location, two indentations were made at 30 μm and 120 μm from the external enamel surface. The latter was used to represent the part of the tooth enamel that was not affected by surface demineralization.

2.9

Surface roughness

After shear bond strength testing, specimens that had been immersed in demineralization solution (pH 4) for 28 days prior to testing were examined by FE-SEM and atomic force microscopy (AFM, Series 5500, Agilent, Santa Clara, CA, USA). Four debonded specimens from the TB, PD and PND groups with less than half of the bonded enamel adhesive covered by adhesive (i.e. ARI score 1) were used for examination. Two additional non-carious premolars were used as control to record the original enamel surface morphology. Each tooth was cut with a low-speed diamond saw into two halves along the mesiodistal plane. The buccal half specimens were ultrasonicated for 15 min to remove debris and dried with nitrogen gas [ ]. Two buccal halves were sputter-coated with gold/palladium for SEM, using the S-4800 FE-SEM at 5 kV. The remaining 2 specimens were used for AFM examination and surface roughness examination. Surface profiling was performed using a tapping model etched silicon probe (Keysight, Böblingen, Germany). Four AFM images were collected from each location of a tooth-half. Surface roughness was determined from the statistical data derived from those images (n = 4 × 2 tooth-halves = 8) and expressed as root mean square height (Sq, in μm).

2.10

Statistical analyses

Interval data generated from the experiments were evaluated for their normality (Shapiro-Wilk test) and equal variance assumptions (modified Levene test) prior to the adoption of parametric statistical methods. Statistical significance for all analyses was pre-set set at α = 0.05.

Data derived from bonding of orthodontic brackets to enamel (n = 10) were analyzed with two-factor analysis of variance (ANOVA) to examine the effect of “adhesive” and “aging”, and the interaction of these two factors on shear bond strength. Post-hoc pairwise comparisons were performed using the Holm-Šidák statistic. For analysis of the ARI results (n = 10), the Kruskal-Wallis H test was used to evaluate the scores among the three adhesive subgroups at each time-point; the Wilcoxon signed-rank test was used to compare the scores of each adhesive between the two time-points.

The percentages of live bacteria in biofilms taken with CLSM (n = 5) were analyzed with a two-factor ANOVA design to examine the effect of adhesive (TB, PD or PND) and culture time (6, 12, 24, 48 or 72 h), and the interaction of those 2 factors on antibacterial efficacy. Post-hoc pairwise comparisons were conducted using the Holm-Šidák statistic.

For microhardness evaluation, a 3-factor ANOVA design was originally planned for examining the effect of adhesive (TB, PB or PND), indentation location (occlusal, under bracket, cervical or lingual), the depth from which indentation was made (30 or 120 μm from enamel surface), and the interaction of those 3 factors on microhardness. However, the normality and equal variance assumptions of the data sets were not fulfilled despite different forms of non-linear transformation of the data sets. Accordingly, data derived from the 24 groups (n = 12) were analyzed using Kruskal-Wallis H test and Dunn’s multiple comparisons test.

For surface roughness evaluation, data from the control, TB PB and PND groups (n = 8) were logarithmically-transformed to satisfy the normality and homoscedasticity assumptions for the use of parametric statistical methods. The transformed data sets were analyzed with one-factor ANOVA and the Holm-Šidák method for pairwise comparisons.