Effect of Trimetaphosphate and Fluoride Association on Hydroxyapatite Dissolution and Precipitation In Vitro

on hydroxyapatite. Hydroxyapatite powder was suspended in different solutions: deionized water, 500 μg F/mL, 1,100 μg F/mL, 1%TMP, 3%TMP, 500 μg F/mL plus 1%TMP and 500 μg F/mL plus 3%TMP. The pH value of the solutions was reduced to 4.0 and after 30 min, raised to 7.0 (three times). After pH-cycling, the samples were analyzed by X-ray diffraction and infrared spectroscopy. The concentrations of calcium fluoride, fluoride, calcium and phosphorus were also determined. Adding 1% or 3% TMP to the solution containing 500 μg F/mL produced a higher quantity of calcium fluoride compared to samples prepared in a 1,100 μg F/mL solution. Regarding the calcium concentration, samples prepared in solutions of 1,100 μg F/mL and 500 μg F/mL plus TMP were statistically similar and showed higher values. Using solutions of 1,100 μg F/mL and 500 μg F/mL plus TMP resulted in a calcium/phosphorus ratio close to that of hydroxyapatite. It is concluded that the association of TMP and fluoride favored the precipitation of a more stable hydroxyapatite. Effect of Trimetaphosphate and Fluoride Association on Hydroxyapatite Dissolution and Precipitation In Vitro


Introduction
In human tooth enamel, hydroxyapatite (HA -Ca 10 (PO 4 ) 6 (OH) 2 ) crystals are arranged into highly organized prisms to for the main unit.In the oral cavity, tooth enamel can be damaged by the local cariogenic bacteria in biofilm (caries) or non-bacterially derived erosive challenges (such as acidic beverages) (1).The process of dental caries is characterized by the dissolution of hydroxyapatite (HA) from tooth enamel.The maintenance of that component in dental structures by decreasing its dissolution can be achieved by using fluoride (F)-containing products (2).In order to become less dependent of F-based items, combinations of different non-fluoride agents and F have been studied (3).Some studies have shown that inorganic phosphates have anticariogenic action (4).Among them, sodium trimetaphosphate (TMP) seems to be the most active.The mechanism by which TMP reduces the dissolution of HA has not been completely clarified.The literature suggests that it is related to TMP adsorption by the enamel surface (4), thus causing a reduction in ion exchange between the enamel and the oral environment, thereby reducing demineralization during acidic challenge.
The addition of TMP to dentifrices with low F concentrations has shown the potential of TMP as a cariespreventing (5,6) and anti-erosion agent (7,8).The results of an in vitro study (9) showed that this combination avoids laminated caries lesion formation, as has already been observed in another study (10).However, to obtain new information on the TMP mechanism of caries prevention, an analysis of the effects of TMP on the dissolution behavior of HA is important in order to verify the changes that can occur during the demineralization and remineralization process in the simultaneous presence of the two components, TMP and F.Moreover, the studies assessing the effects of F and TMP in the dynamics of dental caries and erosion cited above does not considered the direct interactions between F and TMP with dental enamel.Thus, in the present study, we used HA powder and an in vitro model to simulate the dissolution and precipitation of HA.This allows us to investigate the mechanisms by which TMP and/or F affect the HA structure.

Synthesis of HA
HA powders were prepared based on the synthesis route published by Qu and Wei (11) and Rintoul et al. (12).Calcium nitrate [Ca(NO 3 ) 2 •4H 2 O, 1 mol/L, 300 mL, Sigma-Aldrich Co., St. Louis, MO, USA] and dibasic ammonium phosphate [(NH 4 ) H 2 PO 4 , 0.3 mol/L 600 mL, Sigma-Aldrich] solutions were prepared.The pH values of both solutions were adjusted to 10-12 by adding concentrated ammonium hydroxide (29.5%).The dibasic ammonium phosphate solution was slowly added to the calcium nitrate solution (2-5 mL/min) under constant agitation at 37 °C.The precipitate was subsequently aged for 7 days at 37 °C and ground into a fine powder (synthetic HA) (13).

Demineralization and Treatment
Pure TMP solutions (100 mL Na 3 P 3 O 9 ; Sigma-Aldrich) at concentrations of 1.0% and 3.0% and the corresponding TMP solutions with added sodium fluoride (500 µg F/ mL), were prepared (5).As a positive control, a solution of 1,100 µg F/mL (NaF, Merck, Darmstadt, Germany) was also prepared.The HA powder (0.5 g) (n=5) was suspended in each experimental solution of TMP and/or F for 30 min at a pH value of 7.0.Aliquots of HA powder were also suspended in deionized water and were used as negative controls (HA NC ).
After preparation of the solutions, the pH value of each solution was slowly reduced to 4.0 using 1 mol/L nitric acid (HNO 3 , Merck) under agitation.After 30 min of equilibration, the pH value of each solution was raised to 7.0 by the addition of 1 mol/L sodium hydroxide (NaOH; Merck).The pH cycling process was repeated three times, ending at pH=7.0 at 37 °C (11,12).The resulting precipitates were filtered and washed five times with deionized water.The solid phase was oven dried for 24 h at 70 °C and ground into a fine powder using an agate mortar and a pestle.Particles of diameter 53 µm were isolated and analyzed using X-ray diffraction (XRD), Fourier-transform infrared (FTIR) spectroscopy, and fluoride (F), calcium (Ca) and phosphorus (P) analysis.

FTIR Spectroscopy and XRD
For each solution subjected to pH cycling, two HA samples were analyzed by FTIR spectroscopy.The IR absorption spectra of the samples were recorded using a diffuse reflectance method on an FTIR spectrophotometer (Nexus 670; Nicolet, Ramsey, MN, USA), performing 128 scans at a resolution of 2 cm -1 in a spectral range of 4000-400 cm -1 .The intensities of the absorption bands were measured relative to a baseline by joining the points of lowest absorbance of a given peak.The intensity values obtained in this way were all expressed as percentages (%) and compared to those of HA NC (13).XRD data were obtained using Cu-Kα radiation (Ultima IV X-ray diffractometer; Rigaku Corp., Osaka, Japan) generated at a voltage of 40 kV and a current of 40 mA.The XRD patterns were obtained in a scanning range (2θ) from 10° to 60°, with a step size of 0.02°.The CRYSTMET database (Toth Information Systems Inc., Ottawa, Canada) was used for phase identification.

Analysis of Loosely bound (CaF 2 ) and Firmly Bound Fluoride (FA-like) in HA
Assessment of the uptake of loosely bound fluoride (alkali-soluble fluoride, CaF 2 ) by HA was performed following the method of Caslavska et al. (14).For this purpose, 5 mg of cycled HA powder were weighed in microtubes and 0.5 mL of 1.0 mol/L potassium hydroxide were added.After 24 h of stirring (Orbital Shaker TE-141; Tecnal, Piracicaba, SP, Brazil), the suspension was centrifuged at 2900 g for 20 min; 0.3 mL of the supernatant were neutralized by 0.3 mL of TISAB II (ionic strength adjuster buffer, pH 5.0) modified with 1.0 mol/L HCl.The fluoride concentration was determined using an F-selective electrode (Orion 9409, Orion Research Inc., Beverly, MA, USA), which had previously been calibrated.The firmly bound fluoride (fluorapatite-like, FA-like) was measured in the precipitates, which were washed three times in distilled/deionized water and once in methanol.After methanol evaporation, 0.25 mL of 1.0 mol/L HCl were added to the microtube.After stirring for 1 h, 0.25 mL of TISAB II modified by 20 g of NaOH/L were added.Duplicate samples were analyzed as described previously.

HA Calcium and Phosphorus Analysis
For the determination of the Ca and P content, 0.5 mL of 1.0 mol/L hydrochloric acid were added to 5 mg of each of the pH-cycled HA samples in microtubes.After stirring for 1 h, the samples were neutralized with 0.5 mL of 1.0 mol/L NaOH.The Ca concentration was analyzed by the Arsenazo III colorimetric method (15) using a CaCl 2 standard (40-200 µg Ca/mL).The P content was determined by the molybdate method using a KH 2 PO 4 (1.5-24 µg P/mL) standard (16).All reactions were performed in 96-well plates and the absorbance data were measured in a microplate spectrophotometer (PowerWave 340; BioteK, Winooski, VT, USA) at 650 nm (Ca) and 660 nm (P).

Data Analysis
The FTIR spectroscopic data were analyzed as a function of the presence and intensities (%) of the spectral bands of the HA samples subjected to pH-cycling in comparison to the negative control (HA NC ).The results of the concentrations, expressed as mg/g, of F (CaF 2 and FA-like), Ca, and P in HA were subsequently tested for normality (Shapiro-Wilk test) and homogeneity (Cochran test).P and F (log 10 transformation) data were subjected to variance analysis (one-way), followed by the Student-Newman-Keuls test.The Ca data showed a heterogeneous distribution and were submitted to the Kruskal-Wallis test, followed by the Student-Newman-Keuls test.Statistical analysis was performed using the program SigmaPlot version 12.0 (Systat Software Inc., USA), and the significance level was set at 5%.

Results
The measured XRD pattern of synthetic HA was in agreement with the reference diffractogram of HA, according to the CRYSMET database.No qualitative difference was observed between XRD pattern of the HA NC and pure HA specimens (Fig. 1A) as well as between the pure HA and 1%TMP and 3%TMP (Fig. 1B).Traces of CaF2 were found in the HA and HA NC patterns.This phase (◊) was easily identified in the XRD pattern of the HA dispersed in the solution at a concentration of 1,100 µg F/ mL (F1100 pattern).The F1100 showed sharper peaks than those of the samples of pure HA and HA NC (to see inset of the Fig. 1A).This narrowing of the peaks is an indicative of the increased crystallinity of HA.Based on the Scherrer equation, D=0.9λ/βcosθ (λ is the X-ray wavelength, β is the line broadening at half the maximum intensity (FWHM) in radians and θ is the Bragg angle), the size (D) of crystallites estimated by us was 20 nm for HA and 34 nm for F1100 samples.Treatment with 500 µg F/mL and 1% or 3% TMP also produced sharper peaks than those of pure HA (Fig. 1b).
treatments always led to lower intensities.
Chemical analysis showed higher Ca concentrations in HA when treated in solutions of 1,100µg F/mL, 500 µg F/mL and 1%TMP, and 500 µg F/mL and 3%TMP (Table 1) compared with the other samples (p<0.001).The combination of 500 µg F/mL and TMP resulted in Ca/P ratios similar to that of pure HA.Fluoride-treated samples had higher values of CaF 2 and FA-like than the samples without F (p<0.05).The combination of F and TMP enabled CaF 2 formation at rates two and four times higher than for HA treated in pure fluoride solutions of 1,100 and 500 µg F/mL (p<0.001),respectively (Table 1).Increasing the F concentration in solution (from 500 µg F/mL to 1,100 µg F/mL) also increased the FA-like content in HA by a factor of two.The combination of F and TMP did not increase the values of FA-like, but reduced it for a solution of 3% TMP.

Discussion
According to Rintoul et al. (12) and Sundaram et al. (17), the presence of OH bands in the frequency range 642-663 cm -1 indicates HA precipitation, as observed in the present study.Despite lower Ca concentrations and Ca/P ratios, mainly in the HA NC samples, calcium-deficient HA can show the same crystalline arrangement as stoichiometric HA (13,18).Fluoride incorporation in apatite structures increases the crystallinity and the calcium phosphate amount (19).This was observed when hydroxyapatite was pH-cycled in the presence of 1,100 µg F/mL.The increase in the F concentration and the intensity decrease of the OH bands (at 642-663 cm -1 and 3570 cm -1 , respectively) in the IR spectra indicate the presence of F ions in HA (17,20).The reduction in the number of OH groups may suggest that F-is linked to OH, without replacing it, indicating the precipitation of fluoridated HA with a high Ca/P ratio and good crystallinity.Two data support this assertion: (a) the correlation between the IR intensity of the OH group in the 642-663 cm -1 region and the FA-like concentration (r=-0.836,p=0.038) and (b) the absence of a peak (745-747 cm -1 ) indicating F replacement in any spectrum (Fig. 3) (12,17).On the other hand, calcium fluoride formation did not show a correlation with the OH bands.The large calcium fluoride amount found in HA when treated with a combined solution of F and TMP could be one of the factors that may explain the better results obtained by Takeshita et al. (5,9).The narrow OH stretching band is associated with free OH-ions (21,22) and the broad band with the existence of lattice or adsorbed water in the samples (24).As in the case of the F group, the intensity of the narrow OH band also decreased in the presence of TMP.However, a correlation was only observed for the intensity of the OH bands in the regions 642-663 cm -1 and 3570 cm -1 (r=0.746;p=0.033) and fluoride.The intensity reduction of the narrow OH band upon TMP treatments indicates the hydroxyl binding site of TMP on HA.Two data support this hypothesis: (a) there is a reduction in the concentration of FA-like with increasing TMP and its associated F concentration and (b) the intensity of the narrow OH band is smaller in the presence of TMP.This indicates competition for the same binding site between TMP and F in HA.
Although the IR spectra indicate an intensity increase of most phosphate bands, the chemical analysis showed a reduction in the P concentration in F groups, mainly for the samples treated in a solution of 1,100 µg F/mL.This decrease could be related to P in the form of HPO 4 2-replacing PO 4 3- sites ( 23), whose band shows a greater intensity reduction in fluoride-treated HA compared with samples prepared in solutions of 1% and 3% TMP.The HPO 4 2-band is related to the formation of calcium-deficient HA (Ca/P 1.5-1.6)synthesized in this experiment (21).The corresponding peak intensity is higher for a lower Ca/P ratio or Ca concentration (21,22).These data are in agreement with the results obtained by FTIR spectroscopy and chemical analysis.
A reduction of calcium loss by adsorption of TMP and/or fluoride on HA results in a concomitant intensity decrease of the carbonate ν 3 band.A higher concentration of Ca in HA can produce a positive charge deficit, which is counterbalanced by the release of carbonate from the HA crystal (24).This effect varies with the concentration of TMP and associated F in the solutions, resulting in different carbonate structures.The reduction of P and the increase of Ca in HA can explain the best Ca/P ratio being observed in the presence of TMP and/or F, and the increase in crystallinity, as observed by XRD.
It is important to mention that in this study an in vitro model and a HA powder was used to simulate cariogenic challenge (process of de-remineralization) for the evaluation of the effect of TMP and/or F on HA.However, this is a chemical model and therefore it presents limitations, especially related to the inability to reproduce the complex intraoral conditions.Such as, the saliva and the acquired pellicle are extremely important in the de-and remineralization process as well for adsorption of ions and molecules to the HA structure (25).
Our study shows that TMP associated with 500 µg F/g reduces the demineralization of HA and decreases Ca loss, improving the HA crystallinity.Furthermore, the adsorption of calcium fluoride on the crystal structure is increased.TMP and fluoride compete for the same binding sites in the hydroxyapatite.