Introduction

Fluorine is not only a common element present in the earth crust, but it is also found in the form of fluoride (F) in the soils, rocks, and water throughout the world. Higher concentrations are found in the areas where there have been recent/past pyroclastic activities or geologic uplift. In addition, fluoride is broadly employed in many industrial processes nowadays. The major sources of systemic fluoride exposure are the diet (food and water) and dental products, especially toothpastes [1].

Mild doses of fluoride have therapeutic action against dental caries while elevated levels, through intake of water, toothpastes, and diets containing high F levels will increase the body burden. The therapeutic window is very narrow, and there has been a lot of discussion concerning the appropriate levels of F intake to provide the maximum benefit (caries prevention) with minimum risk to cause disbenefits [2].

Once absorbed by the gastric intestinal system, F is distributed to all soft and mineralized tissues via the bloodstream [3]. Despite the most common adverse effects of excessive F intake are seen in the mineralized tissues (teeth and bones; dental and skeletal fluorosis, respectively) [4, 5], several studies that report the negative effects of F in soft tissues, such as the testis [6, 7], thyroid gland [8, 9], spleen [10, 11], liver [12], kidney [13, 14], and brain [15, 16], have been published.

Interestingly, the susceptibility to the toxic effects of F appears to be genetically determined. There are reports in the literature of populations that tend to develop higher levels of dental fluorosis than it would be expected from their background exposure to F [17,18,19]. Moreover, it was reported that inbred mice strains have different susceptibilities to dental fluorosis. The A/J strain is susceptible, with rapid onset and more severe development of dental fluorosis, while the 129P3/J strain is resistant, developing minimal dental fluorosis even under high exposure to F [20]. Despite the 129P3/J mice are resistant to the development of dental fluorosis, they excrete less F in urine and have higher circulating levels of F [21], which is very intriguing. Moreover, A/J mice drink higher volume of water than 129P3/J mice [21], which might be explained by the higher expression of α-AASA dehydrogenase (alpha-amino adipic semialdehyde dehydrogenase) in the latter [14], since this protein is an effective osmoprotectant.

Since the liver is a central organ in the metabolism, the metabolic differences shown by these animals in F handling prompted us recently to investigate the differential pattern of protein expression in the liver of these mice. We observed that A/J mice had many proteins increased when compared with their 129P3/J counterparts, most of them related to energy flux and oxidative stress, which could be implicated in their higher susceptibility to the development of dental fluorosis [22]. Thus, it is of great interest to evaluate the effect of F in the pattern of protein expression in the liver of A/J mice in comparison with 129P3/J mice, which was the aim of the present study.

Experimental section

Animal and Sample Collection

Weanling mice from two different inbred strains, A/J and 129P3/J (3-week-old; n = 12 from each strain), were kept in pairs in metabolic cages with ad libitum access to low-F food (AIN76A, 0.95 mg/kg F; PMI Nutrition, Richmond, IN, USA) and water containing 15 or 50 ppm F (as NaF) for a period of 42 days (n = 6 for each strain and level of F exposure). During the course of treatment, the humidity and temperature in the climate-controlled room were 23 ± 1 °C and 40–80%, respectively.

The experimental protocols were approved by the Ethics Committee for Animal Experiments of Bauru Dental School, University of São Paulo (license no. 031/2013). At the end of the study, the mice were anesthetized with ketamine/xylazine and the livers were collected for the study. Samples to be used for proteomic analysis were stored at − 80 °C, while those designated for F analysis were stored at − 20 °C.

Fluoride Analysis in the Liver

Fluoride analysis data was obtained with the ion-sensitive electrode, after hexamethyldisiloxane-facilitated diffusion [23], as previously described [12].

Statistical Analysis

The software GraphPad Prism version 6.0 for Windows (GraphPad Software, Inc., La Jolla, USA) was used to analyze differences in liver F concentration. Data passed normality (Kolmogorov-Smirnov test) and homogeneity (Bartlett test) and were then analyzed by two-way ANOVA followed by Sidak’s multiple comparison test. The significance level was set at 5%.

Sample Preparation for Proteomic Analysis

Samples were prepared for proteomic analysis as previously described [24]. Cryogenic mill (model 6770; Spex, Metuchen, NJ, EUA) was used for the homogenization of frozen tissue. In order to extract proteins, liver homogenate was incubated in lysis buffer containing 7 M urea, 2 M thiourea, 4% CHAPS, 1% IPG buffer (pH 3–10), and 40 mM dithiothreitol (DTT) for 1 h at 4 °C with infrequent shaking. Afterwards, the homogenate was centrifuged at 15,000 rpm for 30 min at 4 °C and the supernatant having soluble proteins was collected. For the precipitation of proteins, PlusOne 2D Cleanup (GE Healthcare, Uppsala, Sweden) kit was used as recommended by the manufacturer. The pellets thus obtained were resuspended in rehydration buffer (7 M urea, 2 M thiourea, 0.5% CHAPS, 0.5% IPG buffer (pH 3–10), 18 mM DTT, 0.002% bromophenol blue). Later on, a protein pool was constituted while having 25 μL of liver proteins from each animal of the same group that was centrifuged for clarification. To each pool, 50 mM ammonium bicarbonate (AMBIC) containing 3 M urea was added. Each sample was filtered twice in 3 kDa Amicon (Millipore, St. Charles, MO, USA). Bradford protein assay [25] was done in order to quantify the proteins present in the pooled samples. To each sample (50 μg of total protein for each pool in a volume of 50 μL), 10 μL of 50 mM AMBIC was added. In sequence, 25 μL of 0.2% RapiGest™ (Waters Co., Manchester, UK) was added and incubated at 80 °C for 15 min. Afterwards, 2.5 μL of 100 mM DTT was added and incubated at 60 °C for 30 min. 2.5 μL of 300 mM iodoacetamide (IAA) was added and incubated for 30 min at room temperature (in the dark). Then, 10 μL of trypsin (100 ng; Trypsin Gold Mass Spectrometry; Promega, Madison, USA) was added and digestion was allowed to occur for 14 h at 37 °C. After digestion, 10 μl of 5% TFA was added, and the sample was left in an incubation phase for 90 min at 37 °C. It was then centrifuged (14,000 rpm for 30 min). Finally, the supernatant was collected and 5 μL of ADH (1 pmol/μL) plus 85 μL of 3% ACN was added to it.

LC-MS/MS and Bioinformatics Analyses

The nanoAcquity UPLC-Xevo QTof MS system (Waters, Manchester, UK) was used for the separation and identification of peptides, exactly as previously described [26]. In order to notify the differences in the expression among the respective groups, Protein Lynx Global Service (PLGS) software was used, where p < 0.05 and 1 − p > 0.95, and thus showed the downregulation or upregulation of proteins, respectively (Table 1). Bioinformatics analysis was performed, as reported earlier [24, 26,27,28]. For the comparison of A/J vs. 129P3/J, UniProt protein ID accession numbers were mapped back to their associated encoding UniProt gene entries. Furthermore, Gene Ontology annotation of broad biological process was performed using Cluego v2.0.7 + Clupedia v1.0.8, a Cytoscape plug-in. UniProt IDs were uploaded from Table 1 and analyzed with default parameters, which specify an enrichment (right-sided hypergeometric test) correction method using the Bonferroni step down, analysis mode “Function” and load gene cluster list for Mus musculus, evidence codes “All,” set networking specificity “medium” (GO levels 3 to 8), and kappa score threshold 0.03. The protein-protein interaction networking was downloaded from PSICQUIC, built-in Cytoscape version 3.0.2, and constructed as proposed by Millan [29]. Ultimately, a network was constructed, providing a global view of potentially relevant interacting partners of proteins whose abundances change.

Table 1 Proteins with expression significantly altered in the liver of A/J and 129P3/J mice treated with 15 ppm F
Full size table

Results

Liver Fluoride Analysis

Liver F concentrations were significantly different between the strains (F = 12.68, p = .002) and F concentrations in the drinking water (F = 36.55, p < 0.0001), without interaction between these criteria (F = 1.79, p = 0.196). A dose-response pattern was observed for liver F concentrations. However, when the strains were compared, only the 129P3/J mice treated with 50 ppm F in the drinking water had liver F concentrations significantly higher than their A/J counterparts (Fig. 1).

Fig. 1
figure 1

Liver fluoride concentrations in the A/J and 129P3/J mice treated with water containing 15 or 50 ppm fluoride in the drinking water for 42 days. For each fluoride concentration, distinct lowercase letters indicate significant differences between the strains. Distinct uppercase letters denote significant differences between the fluoride concentrations. Two-way ANOVA and Sidak’s test (p < 0.05). n = 6. Bars indicate SD

Full size image

Functional Distribution of Identified Proteins

Figure 2 shows, for each comparison, the functional classification according to the biological process with the most significant term. For the 15 ppm F (Fig. 2a) and 50 ppm F (Fig. 2b) groups, the identified proteins were divided into 14 and 12 functional categories, respectively. For both groups, the category with the highest percentage of the number of genes was the carboxylic acid metabolic process (24 and 31% for 15 and 50 ppm F, respectively), followed by the cellular amino acid metabolic process (13 and 22% for 15 and 50 ppm F, respectively).

Fig. 2
figure 2

Functional distribution of the proteins identified with differential expression in the liver of A/J compared to 129P3/J mice treated with fluoride (a 15 ppm and b 50 ppm). Categories of proteins were based on the GO annotation Biological Process. Significant terms (kappa = 0.03) and distribution were according to the percentage of the number of gene association

Full size image

Liver Proteome Profile and Identification of Differentially Expressed Proteins

Tables 1 and 2 show the proteins with changes in expression when A/J mice are compared with their 129P3/J counterparts for the groups treated with 15 and 50 ppm F, respectively. The treatment with the lower F concentration (15 ppm) provoked more pronounced alterations in fold change in comparison to the treatment with the higher F concentration (50 ppm). Remarkably, all proteins with fold change upon treatment with 15 ppm F were increased in the A/J mice compared with their 129P3/J counterparts. Among them are delta-amino levulinic acid dehydratase (P10518) and 3-ketoacyl-CoA thiolase A, peroxisomal (Q921H8) that were increased 5.58- and 4.18-fold in the A/J mice, in comparison with the 129P3/J mice. In addition, most of the unique proteins were identified in the A/J mice. Treatment with 50 ppm F caused less fold change in comparison with the treatment with 15 ppm F, with some proteins increased and others decreased in the A/J mice in comparison to their 129P3/J counterparts.

Table 2 Proteins with expression significantly altered in the liver of A/J and 129P3/J mice treated with 50 ppm F
Full size table

Protein Interaction Networks

A network was created for each of the comparisons displayed above, employing all the interactions found in the search conducted using PSICQUIC. After the global network was created, nodes and edges were filtered using the specification for Mus musculus taxonomy (10090). The value of fold change and also the p value were added in new columns. The ActiveModules 1.8 plug-in to Cytoscape was used to make active modules connected to subnetworks within the molecular interaction network whose genes presented significantly coordinated changes in fold changes and p value, as shown in the original proteomic analysis. Figure 3a, b shows the subnetwork generated by VizMapper for each comparison. Regardless of the F concentration, most of the proteins with fold change presented interaction with disks large homolog 4 (Q62108) and calcium-activated potassium channel subunit alpha-1 (Q08460) when the two strains were compared.

Fig. 3
figure 3

a Subnetwork generated by VizMapper for the comparison of A/J vs. 129P3/J treated with 15 ppm fluoride. Color of node and single asterisk indicate the differential expression of the respective protein, for each comparison. Red and green nodes indicate protein downregulation and upregulation, respectively, while single asterisk and double asterisks indicate the presence and absence of protein, respectively, in the first group of each comparison. Gray node indicates proteins presenting interaction that were not identified in the present study. The access numbers in nodes correspond to the following: Q811D0 (Dlg1) disks large homolog 1; Q9DB20 (Atp5o) ATP synthase subunit O, mitochondrial; P08228 (Sod1) superoxide dismutase [Cu-Zn]; P16858 (Gapdh) glyceraldehyde 3-phosphate dehydrogenase; P27773 (Pdia3) protein disulfide isomerase A3; P56480 (Atp5b) ATP synthase subunit beta, mitochondrial; P26443 (Glud1) glutamate dehydrogenase 1, mitochondrial; P21550 (Eno3) beta-enolase; Q9CPU0 (Glo1) lactoylglutathione lyase; P10518 (Alad) delta-aminolevulinic acid dehydratase; P60710 (Actb) actin, cytoplasmic 1; P63260 (Actg1) actin, cytoplasmic 2; P16627 (Hspa1l) heat shock 70 kDa protein 1-like; P06151 (Ldha) l-lactate dehydrogenase A chain; P07724 (Alb) serum albumin; P63101 (Ywhaz) 14-3-3 protein zeta/delta; Q62108 (Dlg4) disks large homolog 4; Q08460 (Kcnma1) calcium-activated potassium channel subunit alpha-1. b Subnetwork generated by VizMapper for the comparison of A/J vs. 129P3/J treated with 50 ppm fluoride. Color of node and single asterisk indicate the differential expression of the respective protein, for each comparison. Red and green nodes indicate protein downregulation and upregulation, respectively, while single asterisk and double asterisks indicate the presence and absence of protein, respectively, in the first group of each comparison. Gray node indicates proteins presenting interaction that were not identified in the present study. The access numbers in nodes correspond to the following: Q8CGP2 (Hist1h2bp) histone H2B type 1-P; P11499 (Hsp90ab1) heat shock protein HSP 90-beta; P19157 (Gstp1) glutathione S-transferase P 1; P52760 (Hrsp12) ribonuclease UK114; P56480 (Atp5b) ATP synthase subunit beta, mitochondrial; P26443 (Glud1) glutamate dehydrogenase 1, mitochondrial; P38647 (Hspa9) stress-70 protein, mitochondrial; P27773 (Pdia3) protein disulfide isomerase A3; P16858 (Gapdh) glyceraldehyde 3-phosphate dehydrogenase; P10649 (Gstm1) glutathione S-transferase Mu 1; P05064 (Aldoa) fructose-bisphosphate aldolase A; P10518 (Alad) delta-aminolevulinic acid dehydratase; P06151 (Ldha) l-lactate dehydrogenase A chain; P70296 (Pebp1) phosphatidylethanolamine-binding protein 1; P05201 (Got1) aspartate aminotransferase, cytoplasmic; P17182 (Eno1) alpha-enolase; P17742 (Ppia) peptidyl-prolyl cis-trans isomerase A; Q00623 (Apoa1) apolipoprotein A-I; Q9CPU0 (Glo1) lactoylglutathione lyase; P07724 (Alb) serum albumin; Q08460 (Kcnma1) calcium-activated potassium channel subunit alpha-1; Q62108 (Dlg4) disks large homolog 4

Full size image

When the animals were treated with 15 ppm F, 7 proteins with fold change between the two strains interacted with disks large homolog 4 (Q62108), while 14 proteins with fold change presented interaction with calcium-activated potassium channel subunit alpha-1 (Q08460). Moreover, 11 proteins were upregulated and 5 proteins were downregulated in A/J mice, while 5 proteins were present only in A/J mice and another 5 proteins were present only in 129P3/J mice (Fig. 3a).

For the animals treated with 50 ppm F, 9 proteins with fold change interacted with disks large homolog 4 (Q62108), while 16 proteins with fold change interacted with calcium-activated potassium channel subunit alpha-1 (Q08460). Moreover, 7 proteins were upregulated and 13 proteins were downregulated in A/J mice, while 5 proteins were present only in A/J mice and another 7 proteins were present only in 129P3/J mice (Fig. 3b).

Discussion

Despite there are many reports highlighting the toxic effects of excessive F ingestion in different organs [14, 30,31,32,33], information related to the liver, especially to proteomics, is limited to two studies conducted with rats [34, 35]. The liver is an important organ in the body, which secretes bile and efficiently processes nutrients [36]. Moreover, the majority of toxicants are biotransformed into metabolites by the liver through various enzyme systems. Consequently, liver undergoes different levels of damage and alterations. These damage/alterations are often associated with a degenerative-necrotic condition [37]. Various reports have shown that drugs and other chemical substances can damage hepatocytes, thus leading to hepatic dysfunction [38,39,40]. In addition, excessive F intake can damage the liver and the degree of damage is related to the quantity of F ingested [41,42,43]. However, the mechanism of F-induced liver dysfunction remains unclear.

A/J and 129P3/J mice strains have been widely studied over the last few years because they respond quite differently to F exposure [14, 20, 44,45,46]. Thus, it is interesting to know if the distinct responses to F by these two strains are related to events occurring in the liver and what are the mechanisms involved. Recently, we observed that even without exposure to F, A/J mice had an increase in proteins related to energy flux and oxidative stress when compared to their 129P3/J counterparts [22], which seems to be a good explanation for the high susceptibility of these mice to the effects of F, since the exposure to this ion also induces oxidative stress [47,48,49,50]. Thus, it is of great interest to investigate the effects of both genetic background and F exposure on the liver proteomic profile, which was the main aim of the present study.

We observed a dose-response relationship in liver F concentrations, which confirms that the treatment with F was effective. Regarding the differences between the strains, despite 129P3/J mice had higher liver F concentrations when compared with their A/J counterparts, this difference was only significant for the highest F dose. This could be explained by the F uptake into the mineralized tissues, while higher doses of F could saturate the uptake [51]. The higher F concentrations found in the liver of 129P3/J mice is in-line with previous findings showing that these mice excrete less F and thus have higher circulating F levels than A/J mice, which is reflected in the F concentrations found in the bone, enamel, and different organs of these mice [44, 46, 52].

One remarkable finding of the present study was the fact that treatment with the lower F concentration (15 ppm) provoked more pronounced alterations in fold change in liver proteins in comparison to the treatment with the higher F concentration (50 ppm), which suggests low-dose hormesis of fluoride exposure. This is in-line with previous studies showing that treatment with 15 ppm F causes more alterations in the liver and kidney than treatment with 50 ppm F [34, 53, 54]. The reason for this finding is attributed to adaptive mechanisms of the organism to F [55] that are triggered by higher doses of this ion, but not by lower ones [54]. It is possible that if the treatment with 50 ppm F had been conducted for a shorter time, it could have led to more pronounced alterations, similar to the ones seen in the 15 ppm F group in the present study. In fact, F can have a dual effect (protective or toxic), depending on the dose and time of exposure [35, 53, 55,56,57]. Interestingly, all proteins with fold change upon treatment with 15 ppm F were increased in the A/J mice compared with their 129P3/J counterparts (Table 1), which suggests an attempt of the organism to fight the deleterious effects of F, since they seem to be dependent on the F dose and duration of the treatment [55]. Treatment with the lower F concentration also led to an increase in protein in the duodenum of rats after chronic exposure [58]. Among the increased proteins, the ones with the highest fold changes are related to metabolism (delta-aminolevulinic acid dehydratase (δ-ALA-D) and 3-ketoacyl-CoA thiolase A, peroxisomal with fold changes of 5.58 and 4.18, respectively. δ-ALA-D is a sulfhydryl-containing enzyme very sensitive to oxidizing agents [59]. It is essential in the heme biosynthesis [60], since it catalyzes the asymmetric condensation of two molecules of δ-ALA to porphobilinogen, which, in subsequent steps, is assembled into tetrapyrrole molecules that constitute the prosthetic groups of proteins and enzymes, such as catalase (CAT) [61]. CAT is an important antioxidant enzyme that is inhibited by F [62]. In fact, treatment with water containing 15 ppm F led to the lowest CAT activity in the liver of rats after treatment for 60 days [53], which is similar to the protocol employed in the present study. Thus, the increase in δ-ALA-D in our study might be an attempt to increase the activity of CAT, which is expected to be decreased upon treatment with F [53]. It is also important to highlight that in our previous study, where the animals were not treated with F, this enzyme was present only in the liver of A/J animals but not in their 129P3/J counterparts [22], which could be an attempt to fight against the oxidative stress in the first. Another protein, 3-ketoacyl-CoA thiolase A, peroxisomal, had an increase of 4.18 times in the A/J mice treated with 15 ppm F, in comparison to their 129P3/J counterparts. This protein is involved in the beta-oxidation of fatty acids, which is a multistep process by which fatty acids are broken down to produce energy. These events take place in mitochondria and peroxisomes, by mechanisms involving dehydrogenation, hydration, redehydrogenation, and thiolytic cleavage [63]. An increase in enzymes related to fatty acids beta-oxidation in the presence of F has been previously reported in the kidney [33, 64] and liver [34], suggesting a vigorous state of fatty acid metabolism, in attempt to counteract the inhibitory effect of F in the glycolytic pathway that has been known for a long time [65].

On the other hand, different to what was seen for the animals treated with 15 ppm F, most of the proteins with fold change upon treatment with 50 ppm F were decreased in the A/J mice compared with their 129P3/J counterparts. The protein that presented the highest level of downregulation was mitochondrial stress-70 protein, that is a heat shock protein related to protein folding (UniProt). The downregulation means a lower degree of protein folding with consequently reduced toleration to F-induced stress in A/J mice [26, 34]. The other two proteins with the highest degree of downregulation in the A/J mice in the group treated with 50 ppm F were two isoforms of glutathione S-transferase (GST). GSTs are a multigene family of isozymes responsible for the detoxification of electrophiles by conjugation with the nucleophilic thiol-reduced GSH [66]. These enzymes play a crucial role in cellular detoxification of endogenous and xenobiotic substrates and in protection against oxidative stress [67]. Their downregulation means a difficulty of the A/J mice treated with 50 ppm F to fight oxidative stress. Despite most of the proteins with fold change in the A/J mice treated with 50 ppm F were downregulated in comparison with their 129P3/J counterparts, one of the upregulated proteins should be highlighted. This is the case of retinal dehydrogenase 1 (ALDH1A1). These enzymes are known to have an antioxidant role by producing NAD(P)H [68, 69], and the upregulation of ALDH1A1 might have been induced by the high dosage of F.

The two proteins in the center of the protein-protein interaction networks (calcium-activated potassium channel subunit alpha-1 and disks large homolog 4), regardless the F concentration, are related to potassium channels. Curiously, the same proteins were found in the center of the interaction networks in a previous study where A/J mice were compared with their 129P3/J counterparts without exposure to F [22]. This suggests that the pattern of the networks is driven mainly by the type of strain than by the exposure to F. In our previous study, the presence of these two interacting proteins in the center of the network was suggested to be related to alteration in the brain function of A/J mice [22] induced by accumulation of ammonia due to liver failure [70], but this needs to be investigated in further studies. Among the proteins identified with altered expression in the present study in the interaction networks, some of them were upregulated in the A/J mice, regardless the F concentration in the drinking water. One of them was mitochondrial glutamate dehydrogenase 1 (GLUD1), a sensitive marker of hepatotoxicity. This enzyme is highly expressed in the hepatic mitochondria, and its upregulation supports the occurrence of mitochondrial dysfunction [71, 72], which is described to be induced upon exposure to F [34, 73]. Another upregulated protein in A/J mice, regardless exposure to F, was serum albumin. The main function of this protein is the regulation of the colloidal osmotic pressure of blood, since it has a good binding capacity for water, Ca2+, Na+, and K+ (UniProt). Its upregulation in A/J mice might be related to the higher volume of water ingested by this strain, regardless exposure to F [44].

In the network generated for the groups treated with 15 ppm F, most of the proteins with altered expression are associated with energy flux. The increase of the glycolytic enzyme beta-enolase, the reduction of l-lactate dehydrogenase, as well as the increase in subunits of ATP synthase involved in the oxidative phosphorylation indicate an increased energy flux in the A/J strain. This was also observed in our previous study that did not include exposure to F [22]. The increased energy flux might produce oxidative stress, which is reflected in the increased levels of SOD and ALAD.

Despite the network generated for the groups treated with 50 ppm F had the same proteins in the center as the one generated for the groups treated with 15 ppm F, some of the interacting partners were different. One of the downregulated proteins was histone H2B type 1-P. Despite not appearing in the network, many other isoforms of histones were downregulated in the A/J mice treated with 50 ppm F, as compared with their 129P3/J counterparts (Table 2). Histones are quite known as core components of nucleosome, and their reduction leads to disorganized and ineffectively structured genomic DNA, which might impair gene transcription [74]. This might help to explain the fact that most of the proteins with fold change upon treatment with 50 ppm F were decreased in the A/J mice compared with their 129P3/J counterparts.

In conclusion, treatment with the lower F concentration (15 ppm) provoked more pronounced alterations in fold change in liver proteins in comparison to the treatment with the higher F concentration (50 ppm), which is in-line with previous findings [34, 54, 75] and is possibly related to the duration of the treatment with F [55]. Interestingly, most of the proteins with fold change upon treatment with 15 ppm F were decreased in the A/J mice compared with their 129P3/J counterparts, suggesting an attempt of the latter to fight the deleterious effects of F. However, upon treatment with 50 ppm F, most proteins with fold change were decreased in the A/J mice compared with their 129P3/J counterparts, especially proteins related to oxidative stress and protein folding, which might be related to the higher susceptibility of the A/J animals to the deleterious effects of F. Our findings add light into the mechanisms underlying genetic susceptibility to fluorosis.