Sirt1 overexpression suppresses fluoride-induced p53 acetylation to alleviate fluoride toxicity in ameloblasts responsible for enamel formation

Abstract

Low-dose fluoride is an effective caries prophylactic, but high-dose fluoride is an environmental health hazard that causes skeletal and dental fluorosis. Treatments to prevent fluorosis and the molecular pathways responsive to fluoride exposure remain to be elucidated. Previously we showed that fluoride activates SIRT1 as an adaptive response to protect cells. Here, we demonstrate that fluoride induced p53 acetylation (Ac-p53) [Lys379], which is a SIRT1 deacetylation target, in ameloblast-derived LS8 cells in vitro and in enamel organ in vivo. Here we assessed SIRT1 function on fluoride-induced Ac-p53 formation using CRISPR/Cas9-mediated Sirt1 knockout (LS8^Sirt/KO) cells or CRISPR/dCas9/SAM-mediated Sirt1 overexpressing (LS8^Sirt1/over) cells. NaF (5 mM) induced Ac-p53 formation and increased cell cycle arrest via Cdkn1a/p21 expression in Wild-type (WT) cells. However, fluoride-induced Ac-p53 was suppressed by the SIRT1 activator resveratrol (50 µM). Without fluoride, Ac-p53 persisted in LS8^Sirt/KO cells, whereas it decreased in LS8^Sirt1/over. Fluoride-induced Ac-p53 formation was also suppressed in LS8^Sirt1/over cells. Compared to WT cells, fluoride-induced Cdkn1a/p21 expression was elevated in LS8^Sirt/KO and these cells were more susceptible to fluoride-induced growth inhibition. In contrast, LS8^Sirt1/over cells were significantly more resistant. In addition, fluoride-induced cytochrome-c release and caspase-3 activation were suppressed in LS8^Sirt1/over cells. Fluoride induced expression of the DNA double strand break marker γH2AX in WT cells and this was augmented in LS8^Sirt1/KO cells, but was attenuated in LS8^Sirt1/over cells. Our results suggest that SIRT1 deacetylates Ac-p53 to mitigate fluoride-induced cell growth inhibition, mitochondrial damage, DNA damage and apoptosis. This is the first report implicating Ac-p53 in fluoride toxicity.

Keywords: Ameloblast; Apoptosis; CRISPR; Fluoride; P53; Sirtuin.

Fig.1. CRISPR/Cas9-targeted disruption of SIRT1 in LS8 cells

(a) Whole cell lysates were subjected to western blot analysis for Cas9 (160 kDa) expression in WT, negative control (LacZ), and SIRT1 Knockout (KO) clones (321 and 325). β-actin (44 kDa) was used as a loading control. Numbers show the Cas9/β-actinloading control ratio after scanning densitometry.(b) Unique bi-allelic indels were detected by sequence analysis in each knockout clone. The sgRNA targeting site was PCR-amplified from genomic DNA, cloned, multiple clones sequenced, and compared to the reference sequence for mouse WT Sirt1 (RefSeq Accession NC_000076.6). Nucleotide insertions are marked in blue, mutations are marked in green and deletions are marked by dashes. (c) Sirt1 mRNA in WT, LacZ and selected KO clones (321 and 325) was quantified by qPCR. Gapdh was the internal reference control gene. Data were presented as the mean ± SD (NS; not significant, *P < 0.05, **P < 0.01 vs LacZ control). (d) Whole cell lysates were subjected to western blot analysis for SIRT1 (120 kDa) expression in WT, LacZ and KO clones (321 and 325). β-actin (44kDa) was the loading control. Numbers show the SIRT1/β-actinloading control ratio.

Fig.2. CRISPR/dCas9/SAM-mediated overexpression of SIRT1 in LS8 cells

(a) Whole cell lysates were subjected to western blot analysis for dCas9 (160 kDa) expression in WT, LS8^Sirt1/over and negative control (LacZ) cells. β-actin (44 kDa) was the loading control. Numbers show the dCas9/β-actin loading control ratio.(b) The schema shows dCas9-SAM target site (green) in promoter region at -125 bp to -106 bp from the transcription start site (TSS). (c) Sirt1 mRNA in WT, LS8^Sirt1/over and LacZ control cells were quantified by qPCR. B2m was the internal reference control gene. Data were presented as the mean ± SD (*P < 0.05, **P < 0.01). (d) Whole cell lysates were subjected to western blot analysis for SIRT1 (120 kDa) expression in WT, LS8^Sirt1/over and LacZ control cells. α-tubulin (52 kDa) was the loading control. Numbers show the SIRT1/α-tubulin loading control ratio.

Fig. 3. Fluoride induced acetylation of p53 in WT cells and in rat enamel organ

Whole cell lysates were subjected to western blot analysis. Acetylated p53 (Ac-p53) and total p53 were detected. β-actin (44 kDa) was the loading control. Numbers show the Ac-p53 or Total p53/β-actin loading control ratio. (a) WT cells were treated with 5 mM fluoride for indicated times. Ac-p53 formation increased from 2 h to 18 h after fluoride treatment. (b) WT cells were treated with fluoride at indicated concentrations for 6 h. Fluoride increased Ac-p53 in a dose-dependent manner. (c) Rats were treated with 0 or 100 ppm fluoride in their drinking water for 6 weeks. Hematoxylin and Eosin (H&E) staining (upper panel) and Immunohistochemistry (lower panel) were performed on maturation stage incisor sections. More Ac-p53 (Lys370) was formed in rat ameloblasts treated with 100 ppm fluoride compared to control ameloblasts (0 ppm). Shown are representative images from three rats. Scale bar represents 20 μm. Brackets denote ameloblasts (Am).

Fig. 4. Fluoride increased Cdkn1a/p21 expression in WT cells

(a) Cdkn1a/p21 expression was quantified by qPCR. B2m was the internal reference control gene. WT cells were treated with 5 mM fluoride for indicated times. Fluoride treatment significantly increased Cdkn1a/p21 expression at 24 h. (b) WT cells were treated with fluoride at indicated concentration for 24 h. Cdkn1a/p21 expression increased in a dose-dependent manner. Data is presented as the mean ± SD (*P < 0.05). A regression analysis was performed and the results demonstrated that fluoride dose dependently induced Cdkn1a/p21 expression (*P < 0.0001).

Fig. 5. Acetylation of p53 in LS8^Sirt1/over and LS8^Sirt1/KO cells

Whole cell lysates were subjected to western blot. (a) WT cells were treated with 5 mM fluoride with or without 50 μM resveratrol (RES) for 6 h. RES suppressed fluoride-induced Ac-p53 formation. (b) Cultured cell lysates from WT, LS8^Sirt1/KO clones (321 and 325) and LS8^Sirt1/Over cells were subjected to western blotting. Compared to WT cells, Ac-p53 formation was increased in LS8^Sirt1/KO clones (321 and 325), but was diminished in LS8^Sirt1/Over cells. (c) WT, LS8^Sirt1/Over and control LS8^LacZ/Over cells were treated with 5 mM fluoride for 6 h. Fluoride-induced Ac-p53 formation was decreased in LS8^Sirt1/Over cells compared to control LS8^LacZ/Over cells. (d) WT, LS8^Sirt1/KO and control LS8^LacZ/KO cells were treated with 5 mM fluoride for 6 h. Ac-p53 formation was elevated in control LS8^LacZ/KO cells, but Ac-p53 was not significantly altered by fluoride treatment. α-tubulin (52 kDa) or β-actin (44 kDa) were the loading controls. The numbers show the Ac-p53 or Total p53/loading control ratio.

Fig. 6. Cdkn1a/p21 expression and cell growth inhibition in WT, LS8^Sirt1/Over and LS8^Sirt1/KO

(a) WT, LS8^Sirt1/Over, LS8^Sirt1/KO, LACZ control LS8^LacZ/Over and LS8^LacZ/KO cells were treated with 5 mM fluoride for 18 h. Cdkn1a/p21 expression was quantified by qPCR. Fluoride significantly increased Cdkn1a/p21 expression in LS8^Sirt1/KO, but not in WT, LS8^Sirt1/Over and LacZ control cells. Geometric means of B2m and Gapdh were used for normalization. (b) Net ΔΔCt increase was calculated [(ΔΔCt fluoride treated) minus (ΔΔCt non-fluoride treated)]. Cdkn1a/p21 expression was significantly elevated in LS8^Sirt1/KO cells compared to WT or control cells, but there was no significant difference among WT, LS8^Sirt1/Over and control cells. Data were presented as the mean ± SD. Multiple group comparison was performed by one-way ANOVA with Bonferroni/Dunn post-hoc test. *P < 0.05. NS = Not significant. (c) WT (□) vs LS8^Sirt1/KO (■) cells and (d) and WT (□) vs LS8^Sirt1/over (■) cells were seeded into 96-well plates and treated with NaF at the indicated concentrations for 24 h. Cell proliferation percentage was measured by MTT assays. Fluoride-induced cell growth inhibition was enhanced in LS8^Sirt1/KO cells, but was mitigated in LS8^Sirt1/over cells. Four wells were assayed for each experimental treatment and six separate experiments were performed. Data are expressed as mean ± SE. Differences between WT vs LS8^Sirt1/KO, or WT vs LS8^Sirt1/over were analyzed by Student’s t-test.*P < 0.05, **P < 0.01.

Fig.7. Fluoride-induced mitochondrial damage, apoptosis and DNA damage were ameliorated in LS8^Sirt1/over cells

WT, LS8^Sirt1/over, LS8^Sirt1/KO or control LacZ cells were treated with NaF (0 or 5 mM) for 18 h. Whole cell lysates were subjected to Western blot analysis. (a) Cytochrome-c (Cyto-c, 12 kDa) from the cytosol fraction or from the mitochondrial fraction were detected by Western blots. Fluoride-induced Cyto-c release to cytosol was suppressed in LS8^Sirt1/over cells. β-actin (44 kDa) was the cytosol loading control and VDAC/porin (31 kDa) was the mitochondrial loading control protein. The numbers show the Cyto-c/loading control ratio.(b) Cleaved caspase-3 (19kDa) was assessed. Fluoride-induced caspase-3 (Casp3) cleavage was suppressed in LS8^Sirt1/over cells compared to WT and LacZ control cells. The numbers show the cleaved Casp3/α-tubulin loading control ratio. (c) DNA double strand break marker γH2AX (17 kDa) was assessed. Fluoride significantly elevated γH2AX formation in WT, but γH2AX induction was attenuated in LS8^Sirt1/over cells and was substantially augmented in LS8^Sirt1/KO cells. α-tubulin (52 kDa) was the loading control. The numbers show the γH2AX/α-tubulin loading control ratio.