Contents lists available at ScienceDirect

Archives of Oral Biology

journal homepage: www.elsevier.com/locate/archoralbio

TiF4 and NaF varnishes induce low levels of apoptosis in murine and humanfibroblasts through mitochondrial Bcl-2 family and death receptor signalling

Priscila Maria Aranda Salomão, Flávia Amadeu de Oliveira, Daiana Moreli Soares dos Santos,João Paulo Domezi, Thiago José Dionísio, Rodrigo Cardoso de Oliveira, Ana Carolina Magalhães⁎

Department of Biological Sciences, Bauru School of Dentistry, University of São Paulo, Al. Dr. Octávio Pinheiro Brisolla, 9-75, Bauru, São Paulo, Brazil

A R T I C L E I N F O

Keywords:ApoptosisFas-LHuman gingival fibroblastNIH/3T3 fibroblastSodium fluorideTitanium

A B S T R A C T

Objectives: This study evaluated the level and mechanism of apoptosis in human gingival fibroblasts (HGF) andmurine fibroblasts (NIH/3T3) treated with a titanium tetrafluoride (TiF4) varnish compared those treated with asodium fluoride (NaF) varnish.Methods: Cells were treated with a TiF4, NaF (both 2.45%F) or placebo varnish for 6 h and were then examinedusing the TUNEL method. The activities of caspase-3, -8 and -9 were assessed. cDNA for Bax, Bad, Bcl-2 and Fas-Lwas amplified by quantitative PCR. Bax, Bcl-2 and Fas-L were further detected by western blot analysis.Results: Both fluorides similarly increased the percentage of apoptosis, while they failed to activate caspases. TheBax/Bcl-2 gene expression ratio was not altered by either fluoride treatment regardless of the type of cell. NaFvarnish increased the amplification of the Fas-L gene in NIH/3T3 and HGF cells, while treatment with the TiF4varnish resulted in a lower Bad/Bcl-2 expression ratio compared to that of the control for NIH/3T3 cells, but notfor HGF cells. No effect of the fluorides was detected in the protein analysis.Conclusions: NaF and TiF4, at the studied conditions, similarly induce a low level of apoptosis, with consequentmodest activation of the Bcl-2 and Fas-L-dependent signalling pathways. Generally, HGF cells are more sus-ceptible to the fluoride effect than NIH/3T3 cells.

1. Introduction

Sodium fluoride (NaF) a popular fluoride salt that is applied toprevent tooth caries and erosion (Castilho, Salomão, Buzalaf, &Magalhães, 2015; Magalhães, Wiegand, Rios, Buzalaf, & Lussi, 2011;Pessan, Toumba, & Buzalaf, 2011). NaF is included in professional andhome-care dental products since it is considered safe and has no re-levant local side-effects; however, in vitro studies in fibroblasts haveshown that NaF can cause damage at the cellular level (Inkielewicz-Stepniak, Santos-Martinez, Medina, & Radomski, 2014; Jeng et al.,1998; Lee et al., 2008; Otsuki, Sugiyama, Amano, Yasui, & Sakagami,2011; Salomão et al., 2017). NaF can also induce apoptosis dependingon the concentration and frequency of application (Inkielewicz-Stepniak et al., 2014; Lee et al., 2008). Among professional products,NaF varnish is the first choice for young children, as it is only slightlyabsorbed by the gastrointestinal system, and, therefore, it presents alow risk of the undesired systemic effects on hard tissues that can becaused by excessive fluoride ingestion, such as dental or skeletalfluorosis (Ekstrand, Koch, & Petersson, 1980; Olympio et al., 2009;

Pessan et al., 2005). Soft tissues, such as the liver, heart and kidney, arealso susceptible to the toxic effects of fluoride in water (Zuo et al.,2018).

Since 2008, TiF4 varnish has been tested as an alternative to NaFvarnish, especially for patients with a high risk of tooth caries and/orerosion. Several in vitro and in situ studies have shown that TiF4 var-nish has a stronger protective effect against tooth demineralization thanNaF varnish (Comar et al., 2012, 2015; Comar et al., 2017; Magalhãeset al., 2008, 2016; Martines de Souza, Vertuan, Buzalaf, & Magalhães,2017). A recent in situ study demonstrated that TiF4 varnish was able toinduce remineralization of artificial carious lesions regardless of thecariogenic activity, while NaF varnish failed to prevent further demi-neralization under high cariogenic challenges (Comar et al., 2017).Therefore, current knowledge supports the execution of clinical trials totest the protective effect of TiF4 varnish on dental lesions. However, it isimportant to determine the cytotoxic potential of TiF4 varnish beforeapplying it in patients, due to its low pH. A decrease in the pH of themedium facilitates fluoride influx but depresses efflux, inducing in-tracellular accumulation and inhibiting cell proliferation (Kawase &

https://doi.org/10.1016/j.archoralbio.2018.10.039Received 22 August 2018; Received in revised form 19 October 2018; Accepted 31 October 2018

⁎ Corresponding author.E-mail addresses: priaranda@usp.br (P.M. Aranda Salomão), fla_amadeu@yahoo.com.br (F.A. de Oliveira), daianamoreli@gmail.com (D.M.S. dos Santos),

joaopaulodomezi@usp.br (J.P. Domezi), thiagoj@usp.br (T.J. Dionísio), rodrigocardoso@usp.br (R.C. de Oliveira), acm@usp.br (A.C. Magalhães).

Archives of Oral Biology 97 (2019) 245–252

0003-9969/ © 2018 Elsevier Ltd. All rights reserved.

T

Suzuki, 1989).Recently, our research group showed that both NaF and TiF4 varnish

similarly reduce NIH/3T3 viability and increase cell death (Salomãoet al., 2017). Therefore, there is a need to understand their mechanismsof toxicity. Apoptosis is one cell death mechanism. Some studies haveshown that NaF induced apoptosis in fibroblasts via the mitochondriamediated by both the Bcl-2 family and death receptor pathways(Inkielewicz-Stepniak et al., 2014; Lee et al., 2008). Therefore, the aimsof this study are to compare the level of apoptosis induced by bothfluorides, as well as to elucidate the apoptotic mechanisms involved.

It is important to use a primary human cell instead of an im-mortalized murine linage cell, as done in our previous study (Salomãoet al., 2017), to better simulate the in vivo conditions (Inkielewicz-Stepniak et al., 2014; Jeng et al., 1998; Lee et al., 2008; Otsuki et al.,2011). Therefore, this study also investigated the apoptotic mechanismsof both fluorides on NIH/3T3 cells and primary human gingival fibro-blasts (HGF).

Previous studies have allowed a long time for contact between thefluoride and the cells. This does not simulate the real oral conditions, inwhich saliva washes the product away and thereby reduces the con-centration in the mouth over time (Inkielewicz-Stepniak et al., 2014;Salomão et al., 2017). Twetman, Sköld-Larsson, and Modéer, (1999)showed that significant levels of fluoride in saliva do not persist 6 hafter treatment with fluoride varnishes in vivo. Accordingly, Comar,Souza, Grizzo, Buzalaf, and Magalhães, (2014) demonstrated that thepeak of fluoride release into saliva from the varnishes occurs in the first3–6 h in vitro.

Considering the current knowledge, this study tests the followingnull hypotheses: 1) There is no difference between NaF and TiF4varnishes with respect to the level of apoptosis in NIH/3T3 and HGFcells submitted to a 6 h treatment; 2) There is no difference betweenNaF and TiF4 varnishes with respect to the molecular pathways in-volved in apoptosis of NIH/3T3 and HGF cells submitted to a 6 htreatment.

2. Material and methods

2.1. Cell culture

NIH/3T3 fibroblasts (ATCC® CRL1658™) and primary human gin-gival fibroblasts (HGF), previously obtained from the healthy gingivaltissue of patients receiving treatment (approved by the Local EthicalCommittee, Process n° CAAE 71,642,517.6.0000.5417), were culturedin Dulbecco’s modified Eagle’s medium (DMEM; Sigma-Aldrich Co. LLC,St. Louis, USA) supplemented with antibiotics (100 IU ml−1 penicillinand 0.1 mgml−1 streptomycin) and 10% (v/v) fetal bovine serum (FBS;GIBCO Laboratories, Life Technologies, Inc., New York, USA) at 37 °C ina humidified atmosphere of 5% CO2. Enzymatic digestion with 0.25%trypsin (Sigma-Aldrich Co. LLC, St. Louis, USA) was used to harvest thecells for experimental analyses. All experiments were carried out inexperimental triplicate (n= 3 wells or pool of wells) and in three in-dependent experiments (biological triplicate).

2.2. In situ apoptosis detection by TUNEL staining

Cells were cultivated on a 13mm diameter plastic coverslip(Sarstedt, Nümbrecht, Germany) in 24-well microplates at a density of5× 104 cells per well (1.5 mL culture medium per well) for 24 h andwere then treated with experimental varnishes containing 4.00% TiF4or 5.42% NaF (both at 2.45% F, native pH 1.0 for TiF4 and pH 5.0 forNaF), experimental placebo varnish (similar composition of the othervarnishes but without F, pH 5.0) or non-treated cells (control) for 6 h.All varnishes were prepared by FGM-Denstcare (Joinvile, Brazil). Thevarnishes were applied as previously described for 6 h (Salomão et al.,2017). The pH of the medium was kept neutral through the experiment,regardless of the treatment.

After washing with PBS, the cells were examined by the TUNELmethod using ApopTag® Plus Fluorescein In Situ Apoptosis Detection Kit(EDM Millipore Corporation, Hayward, USA) according to the manu-facturer’s instructions. The samples were visualized using a fluorescentmicroscope (40x, Leica DM IRBE, Wetzlar, Germany). Five images ofeach coverslip were captured and a mean of the percentage of apoptosiswas calculated.

2.3. Caspase activation assay

Cells were cultivated in 24-well microplates at a density of 8× 104

cells per well (1.5 mL culture medium per well) for 24 h and were thentreated as previously described for 6 h (Salomão et al., 2017). Doxor-ubicin (15 μg/mL) was used as a positive control for apoptosis. Cellswere harvested, and eight wells were pooled for each treatment toobtain the required number of cells for the experiment. Caspase activitywas assessed using a caspase-3, -8 and -9 Colorimetric Activity Assay Kit(EDM Millipore Corporation, Hayward, USA) according to the manu-facturer’s instructions. The final product (pNA) was measured using amicroplate reader (Synergy H1, BioTek®, Winooski, USA) at 405 nm.The unit of enzyme activity/mg of protein was calculated. For this,protein concentrations were determined by the Bradford method (Bio-Rad, Hercules, USA).

2.4. Real time RT PCR

Cells were cultivated in 24-well microplates at a density of 8× 104

cells per well (1.5 mL culture medium per well) for 24 h and were thentreated as previously described for 6 h (Salomão et al., 2017). Cellswere harvested, and four wells were pooled. Then, 1× 105 cells fromeach treatment were submitted to mRNA extraction using the PureLink®RNA Mini Kit (Life Lechnologies, Carlsbad, USA), and the isolated RNAwas quantified using a NanoDrop™ 1000 (Thermo Fisher Scientific,Waltham, USA). The cDNA was obtained using a High Capacity cDNAReverse Transcription Kit (Applied Biosystems, Foster, USA) accordingto the manufacturer’s instructions. The cDNA samples were incubatedwith Taqman® Gene Expression Master Mix and Taqman® Gene Ex-pression assay (Applied Biosystems, Foster, USA) for Bad(Mm00432042_m1 and Hs00188930_m1), Bax (Mm00432051_m1 andHs00180269_m1), Bcl-2 (Mm00477631_m1 and Hs00608023_m1) andFas-L genes (Mm00438864_m1 and Hs00181225_m1).

The reading was performed using the RT-PCR System ViiATM7(Applied Biosystems, Foster, USA) under the following conditions: 95 °Cfor 10min plus 40 cycles of 95 °C for 15 s and 60 °C for 1min. Act-bexpression (Mm00607939_s1) was used as the control for NIH/3T3cells, and RPL-13 expression (Hs00744303_s1) was used as the controlfor HGF cells (Applied Biosystems, Foster, USA). The relative expressionwas obtained using the formula 2–(ΔΔCt).

2.5. Western blot

Cells were cultivated in 6-well microplates at a density of 5× 105

Table 1Mean and standard deviation values of the percentage-% of cell apoptosis as-sessed by TUNEL staining for NIH/3T3 and HGF cells after 6 h of treatment.

Groups/ cells NIH/3T3 HGF

Control 0.00 ± 0.00a 0.00 ± 0.00a

Placebo 0.13 ± 0.23b 0.21 ± 0.50a

TiF4 1.00 ± 0.73c 3.16 ± 1.89b

NaF 0.63 ± 0.27bc 4.41 ± 2.29b

Values in the same column with distinct superscript lowercase letters are sig-nificantly different (Statistical values were determined using the ANOVA fol-lowed by Tukey test. p= 0.0019 and p < 0.0001 for NIH/3T3 and HGF, re-spectively).

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

246

cells per well (7 mL culture medium per well) for 24 h and were thentreated as previously described for 6 h (Salomão et al., 2017). Afterwashing with PBS, cells were lysed with buffer solution containing50mM Tris-HCl (pH 7.4), 25mM KCl, 5 mMMgCl2 and 0.2% Nonidet P-40 supplemented with protease inhibitors (Sigma-Aldrich Co. LLC, St.Louis, USA) and phosphatase inhibitor (0.2M Sodium orthovanadate).The lysates were pooled (3 wells), sonicated (10 s) and centrifuged at10.000 rpm for 10min at 4 °C. Protein samples quantified by theBradford method (50 μg/treatment) were applied to electrophoresis in aTris-HCl 12.5% polyacrylamide gel and were subsequently transferredto a PVDF membrane. The membrane was immune-labelled with rabbitpolyclonal anti-Fas ligand, rabbit monoclonal anti-Bcl-2, or anti-Baxprimary antibodies (ABCAM INC., Cambridge, MA, USA) followed bysecondary anti-Rabbit IgG conjugated to HRP (Horseradish Peroxidase,ABCAM INC., Cambridge, MA, USA) and Amersham™ ECL™ PrimeWestern Blotting Detection Reagent (GE Healthcare Bio-Sciences AB,Uppsala, Sweden). The relative densities of the bands were determinedby densitometry analysis using Image J software (National Institute ofHealth, NIH image, Bethesda, USA). The arbitrary density values ob-tained were normalized using alpha-Tubulin (ABCAM INC., Cambridge,MA, USA), as an internal control.

2.6. Statistical analysis

The software GraphPad InStat version 2.0 for Windows (Graph PadSoftware, San Diego, USA) was used for statistical analysis. The datawere submitted to a normality test (Kolmogorov-Smirnov test) andequality of variances test (Bartlett test). Values obtained from the cas-pase activity, RT-PCR (Bax/Bcl-2 ratio and Fas-L) and western blotassays were compared using the Kruskal-Wallis test followed by a post-hoc Dunn’s test. Values obtained from the in situ apoptosis detection byTUNEL staining and RT-PCR (Bad/Bcl-2 ratio) assays were comparedusing an ANOVA followed by a post-hoc Tukey test. The level of sig-nificance was set at 5%.

3. Results

3.1. In situ apoptosis detection by TUNEL staining

In NIH/3T3 cells, all varnishes significantly increased the percen-tage of apoptosis compared to the control. The highest percentage ofapoptosis was observed for TiF4-treated cells compared to the controland the placebo varnish; however, they did not significantly differ fromthe NaF-treated cells. In the HGF cells, both fluoride varnishes similarlyincreased the percentage of apoptosis compared to both the placebovarnish and the control (Table 1). Representative images are shown inFig. 1.

3.2. Caspase activation assay

Both fluoride varnish treatments did not increase the activity ofcaspases-3, -8 and -9 compared to the control or the placebo varnish forboth types of cells (Fig. 2).

3.3. Real time RT PCR

The relative expression of the genes involved in the apoptosispathway is shown in Figs. 3 and 4. The Bax/Bcl-2 ratio was not alteredby the fluoride treatments compared to the control regardless of thetype of cell. However, the NaF varnish increased the Fas-L expressionlevel 5-fold and 14-fold compared to the control for the NIH/3T3 andHGF cells, respectively. No difference was detected between the TiF4and NaF-treated cells with respect to Fas-L gene expression (Fig. 3).

Cells treated with TiF4 varnish presented a significantly lower Bad/Bcl-2 expression ratio compared to the control but not compared to theplacebo and NaF varnishes in the case of the NIH/3T3 cells. For theHGF cells, TiF4 only caused a significant difference from the placebo(Fig. 4).

3.4. Western blot

The relative expression levels of apoptosis pathway proteins areshown in Figs. 5 and 6. The Bax/Bcl-2 ratio and Fas-L expression werenot altered by the fluoride treatments regardless of the type of cell.

Fig. 1. Representative images (TUNEL assay, 40x) of the cells after 6 h of treatment: (A) Control, (B) Placebo, (C) 4.00% TiF4, (D) 5.42% NaF for NIH/3T3 and (E)Control, (F) Placebo, (G) 4.00% TiF4, (H) 5.42% NaF for HGF. All cell nuclei are stained with DAPI. TUNEL-positive cells showed fluorescein staining (see arrow).Scale bar, 20 μm.

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

247

4. Discussion

Previous work from our group has shown that both NaF and TiF4 (at2.45%F) had similar cytotoxic effects on NIH/3T3 cells, decreasing cellviability by ∼ 80% (MTT assay) after 6 h of treatment (Salomão et al.,2017). The present study has shown that the level of cell death byapoptosis is very low (< 5%), suggesting that F might only induceviability loss (seen in the MTT assay) and/or that other mechanisms ofcell death might be involved in the cytotoxicity of fluorides such asnecrosis, a violent form of cell death (Barbier, Arreola-Mendoza, & DelRazo, 2010; Kerr, Wyllie, & Currie, 1972). Necrosis has been observedin different cells treated with high fluoride concentrations due to anincrease of oxidative stress (Barbier et al., 2010; Ghosh, Das, Manna, &Sil, 2008). Satoh et al. (2005) also suggest that NaF may induce non-apoptotic cell death, such as necrosis or autophagy, in fibroblasts.

The TUNEL assay was applied in the present study to detect apop-tosis. In this method, a DNA strand break is detected enzymaticallyusing modified nucleotides to label free 3′eOH termini that are gen-erated upon DNA fragmentation in apoptotic cells (Gavrieli, Sherman,& Ben-Sasson, 1992). However, the sensitivity of the TUNEL assay isquestionable since it is unclear how many DNA strand breaks are ne-cessary for detection. Another limitation of the method is its inability to

differentiate breaks caused by apoptosis from those that happen duringDNA repair and gene transcription (Kockx, Muhring, Knaapen, & deMeyer, 1998; Schaper, Elsässer, & Kostin, 1999; Watanabe et al., 2002).To overcome this limitation, a second method (caspase activity detec-tion) was applied to confirm the occurrence of apoptosis (Watanabeet al., 2002).

In our study, no differences were found between NaF and TiF4 withrespect to the level of apoptosis induction, which allowed us to acceptthe first null hypothesis. This result is in agreement with a previouspublication of our group (Salomão et al., 2017), which showed nodifference between both fluorides with respect to loss of viability inNIH/3T3 cells, despite both varnishes presenting different pH values.However, the medium pH was not changed by TiF4 varnish application,which may explain the different result presented here versus those fromthe study of Sen, Kazemi, and Spångberg, (1998).

Despite the low level of apoptosis found in the present study, wetried to understand the mechanism involved in its occurrence (mi-tochondrial signalling via the Bcl-2 family- and/or death receptor-pathways), following the idea of previous works on the cytotoxicitymechanism of NaF (Inkielewicz-Stepniak et al., 2014; Lee et al., 2008).

In a manner that was different from our previous study (Salomãoet al., 2017), we decided to test a unique concentration of fluoride

Fig. 2. Box plots of the activity of Caspases-3, -8 and -9 for both cells. Doxorubicin (15 μg/mL) was applied as the positive control for apoptosis. Distinct letters showsignificant differences among the treatments. Statistical values were determined using the Kruskal-Wallis test followed by Dunn’s test (Caspase-3: p=0.0002 forNIH/3T3 and p= 0.0012 for HGF. Caspase-8: p < 0.0001 for NIH/3T3 and p=0.0002 for HGF. Caspase-9: p < 0.0001 for both NIH/3T3 and HGF).

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

248

equivalent to the amount of fluoride found in commercial varnishes(such as Colgate Duraphat®). We also performed a 6 h application ac-cording to the manufacturer’s recommendations and the results ofprevious works (Comar et al., 2014; Twetman et al., 1999). The varn-ishes were not applied in direct contact with the cells, but they wereapplied on devices that allowed fluoride to be released to the mediumcontaining the cells at a distance of 5mm (Salomão et al., 2017). Fur-thermore, to attribute the observed effect to the fluorides, a placebovarnish group was included. All varnishes (fluoridated or placebo) hada similar composition to isolate the effect of the type of fluoride.

Another strong point of our study was the inclusion of 2 types ofcells, which showed slight differences in their responses to the testedfluorides. Immortalized cells (NIH/3T3) are genetically manipulated,which might influence their response to cytotoxic agents. To overcomethis issue, primary cells should be used, whenever possible, to confirmthe findings (Huang et al., 1999; Kaur & Dufour, 2012). In the presentstudy, HGF cells were more susceptible to the effect of the fluoridesthan NIH/3T3 cells, regardless of the type of salt. Generally, a higherpercentage of apoptosis was found for HGF cells after 6 h of treatment

compared to NIH/3T3 cells, which may be explained by the higher rateof proliferation of the murine linage and consequent higher resistanceto cytotoxic agents.

According to previous data, the amount of F released from bothvarnishes in DMEM is approximately 0.9mM in the 6-h period(Salomão et al., 2017). Lee et al. (2008) showed that 20mM of fluoride(NaF) induced apoptosis in ∼18% of HGF cells during the same lengthof treatment, whereas Satoh et al. (2005) did not see DNA fragmenta-tion in fibroblasts treated for 6 h with 0 to 20mM NaF. Therefore, ourresults are in accordance with the literature.

With respect to the mechanism involved in apoptosis, our resultsshow that the fluorides are able to change mRNA expression (for somegenes) but not the synthesis of proteins (based on our western blots) orenzymatic activity (based on our caspase activity assays). These resultsmight be due to the low fluoride content released by the varnishes intothe medium and the short duration of application. Another possibleexplanation is the induction of cell death by a non-apoptotic pathway,as previously mentioned. We speculate that higher fluoride concentra-tions and/or longer periods of treatment might induce changes in

Fig. 3. Box plots of the Bax/Bcl-2 ratio and Fas-L relative expressions for both cells. Distinct letters show significant differences among the treatments. Statisticalvalues were determined using the Kruskal-Wallis followed by Dunn’s test (Bax/Bcl-2 ratio: p > 0.05 for both cells; Fas-L: p < 0.0001 for both cells).

Fig. 4. Mean and standard deviation values for Bad/Bcl-2 ratio for both cells. Distinct letters show significant differences among the treatments (Statistical valueswere determined by ANOVA followed by Tukey test. Bad/Bcl-2 ratio: p < 0.05 for both cells).

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

249

protein synthesis and enzymatic activity. However, testing such con-ditions is not realistic when we consider what happens in vivo.

Caspases are cysteine proteases that are activated in apoptotic cells,and they are classified as either initiators or effectors. Their activitiesare initiated by stimulus generated either in the cell membrane or in-side the cell and is followed by a cascade of events that finally induceDNA fragmentation (Agalakova & Gusev, 2012; Hengartner, 2000).Despite the fact that F (especially NaF) has been shown to inhibitprotease activity, especially that of MMPs and cysteine catepsins indifferent tissues (DenBesten et al., 2002), our study showed no effect offluorides on the tested caspases.

The lack of effect of the fluorides on caspase activity are in ac-cordance with Satoh et al. (2005), who did not find an alteration in theactivities of caspases -3, -8 and -9 in HGF treated with 0–20mM NaF for

4 h. Otsuki et al. (2011) also failed to observe caspase-3 activation inHGF after 6 h of treatment with 10mM NaF. On the other hand, Leeet al. (2008) showed caspases activity after a 6 h-treatment of HGF with20mM NaF, a concentration that is much higher than those released byour varnishes into the medium.

We investigated the intrinsic pathway, also called the mitochondrialpathway mediated by the Bcl-2 family, and the extrinsic, or death-re-ceptor, pathway involved in the apoptotic effect of fluorides. The Bcl-2family controls the intrinsic apoptotic pathway, regulating the releaseof molecules from the mitochondria to the cytosol, followed by caspase-9 and -3 activation (Agalakova & Gusev, 2012; Shoshan-Barmatz, De, &Meir, 2017). Bax and Bad, both pro-apoptotic members of the Bcl-2family, and the anti-apoptotic Bcl-2 were analysed in the present studyby RT-PCR.

Fig. 5. A. Representative image of the western blot for NIH/3T3 cells. B. Box plots of the Bax/Bcl-2 ratio and Fas-L normalized densitometry values for NIH/3T3 cells.Similar letters show no significant differences among the treatments. Statistical values were determined using the Kruskal-Wallis test (p > 0.05).

Fig. 6. A. Representative image of the western blot for HGF cells. B. Box plots of the Bax/Bcl-2 ratio and Fas-L normalized densitometry values for HGF cells. Similarletters show no significant differences among the treatments. Statistical values were determined using the Kruskal-Wallis test (p > 0.05).

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

250

The extrinsic or death-receptor pathway involves the action of Fas-Lon the cell membrane, a death receptor ligand that is able to induce asequence of apoptotic events such as caspase-8 and -3 activation(Elmore, 2007; Hengartner, 2000; Lee et al., 2008). In the presentstudy, we analysed Fas-L gene expression.

With respect to RT-PCR results, no differences were found betweenNaF and TiF4, allowing us to accept the 2nd null hypothesis; in orderwords, we could not find any differences in gene expression involved inthe apoptosis process (except for the Fas-L gene). It is important to keepin mind that post-translational modification was not analysed in thepresent study, such as (de)phosphorylation or proteolytic cleavage ofBad, which has been discussed as a possible death signal of the coreapoptotic machinery at the mitochondria (Danial & Korsmeyer, 2004;Danial, 2009; Youle & Strasser, 2008). Correia et al. (2015) also showthat the anti-apoptotic functions of Bcl-2 are governed not only bychanges in expression but also by phosphorylation. We did not studythis process deeply, since our results (Bax/Bcl-2 and caspases activity)do not support our previous hypothesis that apoptosis would be themain mechanism involved in the toxicity of fluoride varnishes.

An interesting result was found when fluorides were compared tothe control. NaF induced a higher expression of Fas-L for both types ofcells compared to the control, suggesting that its mechanism of apop-tosis could be extrinsic. Lee et al. (2008) also observed an upregulationof Fas-L expression in HGF after 6 h of treatment with 5–20mM NaF.However, we could not detect higher protein synthesis in NaF-treatedcells by western blot. It is likely that NaF was able to increase the mRNAexpression; however, the 6-h treatment might not be enough time tosignificantly change protein synthesis.

Our results are also in agreement with Inkielewicz-Stepniak et al.(2014), who found no difference in Bax and Bcl-2 expression for cellstreated with 0.5 and 1mM NaF compared to the cells from the control.However, TiF4 induced a lower Bad/Bcl-2 gene expression ratio in NIH/3T3 cells, which could be considered to be an anti-apoptotic effect.However, this finding was not confirmed in HGF cells or by westernblot, so it may not be clinically relevant.

Lee et al. (2008) detected a down-regulation of Bcl-2 (≥ 10mMNaF), an anti-apoptotic protein, but they could not find an alteration inBax, a pro-apoptotic factor, regardless of the fluoride concentration.The only possible explanation for the lack of such effects in the presentstudy is the low amount of fluoride released by the varnishes into themedium during the 6 h of contact.

On the other hand, the present results are very promising con-sidering the clinical application of the experimental TiF4 varnish toprevent tooth caries and erosion in the future. Based on the results ofthe present study, we expect a low incidence of local side effects (suchas mucosa desquamation) by its application. Taken together, the resultsof the present study and of all works in the field of tooth caries anderosion protection (Comar et al., 2012, 2015; Comar et al., 2017;Magalhães et al., 2008, 2016; Martines de Souza et al., 2017) supportthe conduction of clinical trials.

5. Conclusion

NaF and TiF4, in the studied conditions, similarly induce a low levelof apoptosis, with consequent modest activation of Bcl-2 and Fas-L-dependent signalling pathways. Generally, HGF cells are more suscep-tible to the fluoride effects than NIH/3T3 cells.

Conflict of interest

None.

Ethical approval

Ethics Committee from Bauru School of Dentistry/USP – Brazil(CAAE 71,642,517.6.0000.5417).

Acknowledgments

We thank FAPESP for granting a scholarship to the first author(Proc. 2013/13484-6), and a research grant to the last author (Proc.2013/20372-0). The authors are grateful to Centro Integrado dePesquisa (CIP), Bauru School of Dentistry, USP for providing researchfacilities and to Marcia Sirlene Zardin Graeff for the microscopy images.We thank Instituto de Pesquisas Energéticas e Nucleares (IPEN) forkindly sterilizing the modified plates using gamma radiation. We thankFGM Brazilian Company (Joinville-SC, Brazil) for manufacturing allexperimental varnishes, and Prof. Dr. Carla Damante from the BauruSchool of Dentistry/USP for the HGF donation.

References

Agalakova, N. I., & Gusev, G. P. (2012). Molecular mechanisms of cytotoxicity andapoptosis induced by inorganic fluoride. ISRN Cell Biology, 2012, 16.

Barbier, O., Arreola-Mendoza, L., & Del Razo, L. M. (2010). Molecular mechanisms offluoride toxicity. Chemico-biological Interactions, 188(2), 319–333.

Castilho, A. R., Salomão, P. M., Buzalaf, M. A., & Magalhães, A. C. (2015). Protectiveeffect of experimental mouthrinses containing NaF and TiF4 on dentin erosive loss invitro. Journal of Applied Oral Science, 23(5), 486–490.

Comar, L. P., Cardoso, C. A., Charone, S., Grizzo, L. T., Buzalaf, M. A., & Magalhães, A. C.(2015). TiF4 and NaF varnishes as anti-erosive agents on enamel and dentin erosionprogression in vitro. Journal of Applied Oral Science, 23(1), 14–18.

Comar, L. P., Souza, B. M., Grizzo, L. T., Buzalaf, M. A., & Magalhães, A. C. (2014).Evaluation of fluoride release from experimental TiF4 and NaF varnishes in vitro.Journal of Applied Oral Science, 22(2), 138–143.

Comar, L. P., Souza, B. M., Martins, J., Santos, M. G., Buzalaf, M. A. R., & Magalhães, A. C.(2017). Response of carious enamel to TiF4 varnish treatment under diverse cario-genic activities in situ. Journal of Dentistry, 63, 81–84.

Comar, L. P., Wiegand, A., Moron, B. M., Rios, D., Buzalaf, M. A., Buchalla, W., et al.(2012). In situ effect of sodium fluoride or titanium tetrafluoride varnish and solutionon carious demineralization of enamel. European Journal of Oral Sciences, 120(4),342–348.

Correia, C., Lee, S. H., Meng, X. W., Vincelette, N. D., Knorr, K. L., Ding, H., et al. (2015).Emerging understanding of Bcl-2 biology: Implications for neoplastic progression andtreatment. Biochimestry Biophysics Acta, 1853(7), 1658–1671.

Danial, N. N. (2009). BAD: Undertaker by night, candyman by day. Oncogene, 27(suppl 1),S53–S70.

Danial, N. N., & Korsmeyer, S. J. (2004). Cell death: critical control points. Cell, 116(2),205–219.

DenBesten, P. K., Yan, Y., Featherstone, J. D., Hilton, J. F., Smith, C. E., & Li, W. (2002).Effects of fluoride on rat dental enamel matrix proteinases. Archives of Oral Biology,47(11), 763–770.

Ekstrand, J., Koch, G., & Petersson, L. G. (1980). Plasma fluoride concentration and ur-inary fluoride excretion in children following application of the fluoride-containingvarnish Duraphat. Caries Research, 14(4), 185–189.

Elmore, S. (2007). Apoptosis: A review of programmed cell death. Toxicology Pathology,35(4), 495–516.

Gavrieli, Y., Sherman, Y., & Ben-Sasson, S. A. (1992). Identification of programmed celldeath in situ via specific labeling of nuclear DNA fragmentation. The Journal of CellBiology, 119(3), 493–501.

Ghosh, J., Das, J., Manna, P., & Sil, P. C. (2008). Cytoprotective effect of arjunolic acid inresponse to sodium fluoride mediated oxidative stress and cell death via necroticpathway. Toxicology in Vitro, 22(8), 1918–1926.

Hengartner, M. O. (2000). The biochemistry of apoptosis. Nature, 407(6805), 770–776.Huang, D. C., Hahne, M., Schroeter, M., Frei, K., Fontana, A., Villunger, A., et al. (1999).

Activation of Fas by FasL induces apoptosis by a mechanism that cannot be blockedby Bcl-2 or Bcl-x(L). Proceedings of the National Academy of Sciences of the United Statesof America, 96(26), 14871–14876.

Inkielewicz-Stepniak, I., Santos-Martinez, M. J., Medina, C., & Radomski, M. W. (2014).Pharmacological and toxicological effects of co-exposure of human gingival fibro-blasts to silver nanoparticles and sodium fluoride. International Journal ofNanomedicine, 9, 1677–1687.

Jeng, J. H., Hsieh, C. C., Lan, W. H., Chang, M. C., Lin, S. K., Hahn, L. J., et al. (1998).Cytotoxicity of sodium fluoride on human oral mucosal fibroblasts and its mechan-isms. Cell Biology and Toxicology, 14(6), 383–389.

Kaur, G., & Dufour, J. M. (2012). Cell lines: Valuable tools or useless artifacts.Spermatogenesis, 2(1), 1–5.

Kawase, T., & Suzuki, A. (1989). Studies on the transmembrane migration of fluoride andits effects on proliferation of L-929 fibroblasts (L cells) in vitro. Archives of OralBiology, 34(2), 103–107.

Kerr, J. F., Wyllie, A. H., & Currie, A. R. (1972). Apoptosis: A basic biological phenom-enon with wide-ranging implications in tissue kinetics. British Journal of Cancer,26(4), 239–257.

Kockx, M. M., Muhring, J., Knaapen, M. W., & de Meyer, G. R. (1998). RNA synthesis andsplicing interferes with DNA in situ end labeling techniques used to detect apoptosis.The American Journal of Pathology, 152(4), 885–888.

Lee, J. H., Jung, J. Y., Jeong, Y. J., Park, J. H., Yang, K. H., Choi, N. K., et al. (2008).Involvement of both mitochondrial- and death receptor-dependent apoptotic

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

251

pathways regulated by Bcl-2 family in sodium fluoride-induced apoptosis of thehuman gingival fibroblasts. Toxicology, 243(3), 340–347.

Magalhães, A. C., Dos Santos, M. G., Comar, L. P., Buzalaf, M. A., Ganss, C., & Schlueter,N. (2016). Effect of a single application of TiF4 varnish versus daily use of a low-concentrated TiF4/NaF solution on tooth erosion prevention in vitro. Caries Research,50(5), 462–470.

Magalhães, A. C., Kato, M. T., Rios, D., Wiegand, A., Attin, T., & Buzalaf, M. A. (2008).The effect of an experimental 4% TiF4 varnish compared to NaF varnishes and 4%TiF4 solution on dental erosion in vitro. Caries Research, 42(4), 269–274.

Magalhães, A. C., Wiegand, A., Rios, D., Buzalaf, M. A., & Lussi, A. (2011). Fluoride indental erosion. Monographs in Oral Science, 22, 158–170.

Martines de Souza, B., Vertuan, M., Buzalaf, M. A. R., & Magalhães, A. C. (2017). Theimpact of the demineralized organic matrix on the effect of TiF4 varnish on theprogression of dentin erosive loss. Caries Research, 51(3), 264–270.

Olympio, K. P., Cardoso, V. E., Bijella, M. F., Pessan, J. P., Delbem, A. C., & Buzalaf, M. A.(2009). Urinary fluoride output in children following the use of a dual-fluoridevarnish formulation. Journal of Applied Oral Science, 17(3), 179–183.

Otsuki, S., Sugiyama, K., Amano, O., Yasui, T., & Sakagami, H. (2011). Negative reg-ulation of NaF-induced apoptosis by Bad-CAII complex. Toxicology, 287(1-3),131–136.

Pessan, J. P., Pin, M. L., Martinhon, C. C., de Silva, S. M., Granjeiro, J. M., & Buzalaf, M. A.(2005). Analysis of fingernails and urine as biomarkers of fluoride exposure fromdentifrice and varnish in 4- to 7-year-old children. Caries Research, 39(5), 363–370.

Pessan, J. P., Toumba, K. J., & Buzalaf, M. A. (2011). Topical use of fluorides for cariescontrol. Monographs in Oral Science, 22, 115–132.

Salomão, P. M. A., Oliveira, F. A., Rodrigues, P. D., Al-Ahj, L. P., Gasque, K. C. D. S.,

Jeggle, P., et al. (2017). The cytotoxic effect of TiF4 and NaF on fibroblasts is in-fluenced by the experimental model, fluoride concentration and exposure time. PloSOne, 12(6), e0179471.

Satoh, R., Kishino, K., Morshed, S. R., Takayama, F., Otsuki, S., Suzuki, F., et al. (2005).Changes in fluoride sensitivity during in vitro senescence of normal human oral cells.Anticancer Research, 25(3B), 2085–2090.

Schaper, J., Elsässer, A., & Kostin, S. (1999). The role of cell death in heart failure.Circulation Research, 85(9), 867–869.

Shoshan-Barmatz, V., De, S., & Meir, A. (2017). The Mitochondrial Voltage-DependentAnion Channel 1, Ca(2+) Transport, Apoptosis, and Their Regulation. Frontiers inOncology, 7, 60.

Sen, B. H., Kazemi, R. B., & Spångberg, L. S. (1998). Morphologic effects on L929 fi-broblasts of titanium tetrafluoride application. Oral Surgery Oral Medicine OralPathology Oral Radiology Endodontic, 86(3), 341–346.

Twetman, S., Sköld-Larsson, K., & Modéer, T. (1999). Fluoride concentration in wholesaliva and separate gland secretions after topical treatment with three differentfluoride varnishes. Acta Odontologica Scandinavica, 57(5), 263–266.

Youle, R. J., & Strasser, A. (2008). The BCL-2 protein family: Opposing activities thatmediate cell death. Nature Revision Molecular Cell Biology, 9(1), 47–59.

Watanabe, M., Hitomi, M., van der Wee, K., Rothenberg, F., Fisher, S. A., Zucker, R., et al.(2002). The pros and cons of apoptosis assays for use in the study of cells, tissues, andorgans. Microscopy and Microanalysis: The Official Journal of Microscopy Society ofAmerica, Microbeam Analysis Society, Microscopical Society of Canada, 8(5), 375–391.

Zuo, H., Chen, L., Kong, M., Qiu, L., Lü, P., Wu, P., et al. (2018). Toxic effects of fluorideon organisms. Life Science, 198, 18–24.

P.M. Aranda Salomão et al. Archives of Oral Biology 97 (2019) 245–252

252