Remineralization of enamel subsurface lesions using toothpaste containing tricalcium phosphate and fluoride: an in vitro µCT analysis

A procedural flow chart of the present study is shown in Fig. 1. Thirty-six extracted, permanent bovine incisors without damage were stored under freezing conditions until further use. All adherent soft tissues on the teeth were removed by thoroughly cleaning and washing the teeth under running water. A low-speed diamond saw (Isomet, Buehler, Lake Bluff, IL, USA) was used under water to remove the root of the tooth with extra care to keep only the crown. Specimens were prepared by cutting the teeth into 4 × 4 × 3 mm enamel-dentin blocks using the low-speed diamond saw under water. The enamel surfaces were ground flat using 600- to 2000-grit silicon carbide papers (Fuji Star, Sankyo Rikagaku, Saitama, Japan) under water. To set a reference landmark for the μCT scans, a hole (1.0 mm in diameter, 0.5 mm in depth) was made at the side of the tooth using a diamond bur (440SS ISO # 010, Shofu, Kyoto, Japan). Then, nail polish (680, Revlon, New York, NY, USA) was used to cover the surfaces, leaving a window (2 × 2 mm) to expose the specimen’s polished enamel surface.

Subsurface enamel lesions were produced in the enamel using the method described by Margolis et al.[34]. Each specimen was then separately immersed in 10 mL of a demineralization solution (17.8 mM CaCl₂, 8.8 mM KH₂PO₄, 100 mM lactic acid, and 1.0 mM NaN₃; pH was adjusted to 4.3 using 10 mM KOH) at 37 °C for 7 days [34]. The demineralization solution was refreshed daily. To align the demineralization conditions, 36 specimens were selected with initial demineralization within the standard deviation (SD) of the mean demineralization.

After lesion formation, the specimens were coded and randomly divided into four experimental groups based on the type of paste used (n = 9 for each group): (a) fTCP-free, fluoride-free toothpaste (fTCP − F −); (b) fTCP-containing, fluoride-free toothpaste (fTCP + F −); (c) fTCP-free, fluoride-containing toothpaste (fTCP − F +); and (d) fTCP-containing, fluoride-containing toothpaste (fTCP + F +) (Experimental paste, 3M, Tokyo, Japan). Table 1 shows the details of the provided materials. After demineralization, the specimens were removed from the solution and subjected to μCT scanning.

All specimens were treated to a standard regime of pH-cycling [35]. The enamel blocks were alternately immersed in remineralization solution (1.5 mM CaCl₂, 0.9 mM KH₂PO₄, 130 mM KCl, and 20 mM HEPES; the pH was adjusted to 7.0 with 10 mM KOH and 1.0 mM NaN₃) and demineralization solution (17.8 mM CaCl₂, 8.8 mM KH₂PO₄, and 100 mM lactic acid, 1.0 mM NaN₃; pH was adjusted to 4.3 with 10 mM KOH). During each 24-h period, the specimens were immersed in the remineralization solution (4 mL per block) for 23 h and then in the demineralization solution (4 mL per block) for 1 h. The blocks were rinsed with pure water between solution changes. The blocks were treated with the four types of experimental pastes during pH-cycling. Treatments were conducted once each day and one time after demineralization and before remineralization. During each treatment, the specimens were immersed in each treatment solution kept at 37 °C for 6 min, once a day, followed by rinsing with pure water for 1 min, and then finally incubated at 37 °C in pure water. Experimental paste suspensions were prepared at 1:3 dilutions (paste: deionized water) to minimize the influence of other ingredients (e.g., thickeners or ora-base), thoroughly mixed, and subjected to mechanical agitation for 1 min using a vortex mixer (MF-71, TGK, Tokyo, Japan), as described previously [31]. In this treatment, the toothpastes were applied passively by immersion into paste suspensions. pH-cycling and the treatment cycle were continued for 12 days using freshly prepared suspensions each day [36]. After completing pH-cycling and the treatment cycle, the specimens were subjected to μCT scanning for evaluation.

A μCT system (inspeXio SMX-100CT; Shimadzu, Kyoto, Japan) was used to assess the changes in MD in the specimens after demineralization and remineralization. Each specimen was mounted on to a computer-controlled turntable, with the treated dentin surface perpendicular to the X-ray beam. A wet wiper roll was placed on top of the specimen to prevent it from drying out during scanning. A 0.2-mm-thick brass filter was used in the beam path to reduce the beam-hardening effect [31]. The tube voltage was set to 100 kV, and a current of 70 µA was applied. The distance between the X-ray source and the specimen was 68.2 mm and that between the X-ray source and the detector was 300 mm. The specimen was rotated 360° in increments of 0.3°. Air calibration of the detector was performed before each scanning to minimize the number of ring artifacts. Additionally, eight-frame averaging was applied during the acquisition phase to improve the signal-to-noise ratio [37]. The 360° rotation of each specimen was performed at an integration time of 6 min. Data were acquired as 250 TIFF files to reconstruct the three-dimensional (3D) images of the coronal aspect at a resolution of 1024 × 1024 pixels and an isotropic voxel size of 7.0 μm. A series of mineral reference phantoms were scanned for MD calibration; these included three hydroxyapatite (HAp) disks (Phantoms; Ratoc System Engineering, Tokyo, Japan) with different concentrations (0.20, 0.40, 0.50, and 0.70 gHAp cm⁻³) of hydroxyapatite crystals embedded in an epoxy resin as well as a pure HAp disk (concentration, 3.14 gHAp cm⁻³; Cellyard; HOYA Technosurgical Corporation, Tokyo, Japan) [33]. Each specimen was subjected to μCT scanning three times within the experimental period as follows: before demineralization (baseline), after demineralization, and after remineralization.

A 3D analysis software (TRI/3D-BON; Ratoc System Engineering) was used to reconstruct 3D images from two-dimensional (2D) images. Grayscale values were converted to MD values (gHAp cm⁻³) using a linear calibration curve based on the grayscale values received from the phantoms (in linear regression, R² > 0.9997). The 3D data images of each treatment from the same specimen were aligned and registered into one coordinate system to compare the changes in the enamel lesions. The rendered 3D volumes were translated and rotated in an optional software (TRI/3D-DIF; Ratoc System Engineering) to visually match the baseline image, which served as the reference. The features used for the corresponding process were sound surface, specimen edges, and the reference landmark on the side of the specimen (the small hole made by a high-speed round diamond bur).

Mean MD values were calculated and plotted against the depth in a volume of interest measuring 425 × 425 × 900 μm³ at the center of the test window using an optional software (TRI/TMD; Ratoc System Engineering) (Fig. 2). The MD profile was converted to a relative MD profile by assuming sound enamel with a maximum MD of 100 vol%. Mineral loss (ΔZ; vol% in micrometers) was calculated from the relative MD profiles; the reference point of the depth axis (0 µm) was set at the axial position of the apparent surface of the enamel lesion. To calculate the ΔZ value for each specimen from the profiles, the area under the curve was subtracted from the assumed area of the sound enamel before demineralization. The mean percentage of remineralization (%R) was calculated via trapezoidal integration, according to the following formula:

where ΔZd is the difference between the profiles of the area under the sound enamel and the demineralized enamel and ΔZr is the difference between the profiles of area under the sound enamel and the remineralized enamel [38]. These calculations were conducted by importing the MD data into a spreadsheet (Microsoft Excel; Microsoft, Redmond, WA, USA).

The specimens were prepared as described for μCT scanning. The conditioned enamel surfaces from each group were observed under a scanning electron microscope (JSM-5310LV, JEOL, Tokyo, Japan) after pH-cycling. The specimens were placed on a filter paper placed in a covered glass vial for 24 h at room temperature for desiccation. After gold-sputter coating (SC-701AT, Elionix, Tokyo, Japan), the conditioned specimens were longitudinally fractured at the center using a cutting plier, and the enamel structure was observed cross-sectionally using a scanning electron microscope (JSM-5310LV) at 2,000 magnification (n = 2).

The typical 2D images of the treatment groups after demineralization and pH-cycling are shown in Fig. 3. After remineralization, the specimens in the treatment groups showed mineral recovery in the enamel subsurface lesions. The parts beneath the enamel surface were demineralized in the fTCP − F − and fTCP + F − groups compared with in the fTCP − F + and fTCP + F + groups. The fTCP + F + group showed mineral recovery in most enamel subsurface lesions.

The MD profiles of the enamel in each treatment group are summarized in Fig. 4. After demineralization, all test groups demonstrated a demineralized enamel thickness of approximately 180 µm. After pH-cycling, the fTCP − F − group had the lowest amount of mineral gain. In contrast, the fTCP + F + group showed the highest degree of remineralization within all lesion parts. The surface MD values (around 25 µm) were higher in the fTCP − F + and fTCP + F + groups than in the other two groups. Alternatively, the MD values at the bottom of the specimens (around 150 µm) were higher in the fTCP + F − and fTCP + F + groups than in the other two groups.

The mean %R in each group is summarized in Fig. 5. The mean %R ± SD values were 8.9% ± 5.5%, 19.2% ± 7.9%, 25.3% ± 6.9%, and 38.2% ± 7.8% for the fTCP − F − , fTCP + F − , fTCP − F + , and fTCP + F + groups, respectively, in ascending order. Statistical analysis conducted using one-way ANOVA revealed significant differences between the treatment groups (P < 0.001). In addition, multiple comparisons using the Tukey’s post-hoc test revealed significant differences in %R within the treatment groups. There were significant differences in all groups (P < 0.05, for all comparisons) except between the fTCP + F − and fTCP − F + groups (P = 0.284).

Representative SEM micrographs of the enamel surfaces after pH-cycling are shown in Fig. 6. In all the groups, the trabecular structures of the enamel appeared as irregular and black areas beneath the surface. SEM analysis revealed that the fTCP − F − group had the narrowest enamel trabeculae among all the tested groups, with dark areas between the trabeculae (Fig. 6a). On the contrary, SEM analysis revealed that the dark area present between the enamel trabeculae was not clear in the fTCP + F − and fTCP + F + groups compared with in the fTCP − F − and fTCP + F − group (Fig. 6b, c). Moreover, SEM analysis revealed the presence of ~ 1-µm-diameter crystal-like structures in the fTCP + F − group (Fig. 6b) and the presence of crystal-like structures with a slightly larger size (diameter of 1–1.5 µm) in the fTCP + F + group (Fig. 6d).