Fluoride induced leaky gut and bloom of Erysipelatoclostridium ramosum mediate the exacerbation of obesity in high-fat-diet fed mice

Abstract

Fluoride is widely presented in drinking water and foods. A strong relation between fluoride exposure and obesity has been reported. However, the potential mechanisms on fluoride-induced obesity remain unexplored. Objectives and methods The effects of fluoride on the obesity were investigated using mice model. Furthermore, the role of gut homeostasis in exacerbation of the obesity induced by fluoride was evaluated. Results The results showed that fluoride alone did not induce obesity in normal diet (ND) fed mice, whereas, it could trigger exacerbation of obesity in high-fat diet (HFD) fed mice. Fluoride impaired intestinal barrier and activated Toll-like receptor 4 (TLR4) signaling to induce obesity, which was further verified in TLR4^-/- mice. Furthermore, fluoride could deteriorate the gut microbiota in HFD mice. The fecal microbiota transplantation from fluoride-induced mice was sufficient to induce obesity, while the exacerbation of obesity by fluoride was blocked upon gut microbiota depletion. The fluoride-induced bloom of Erysipelatoclostridium ramosum was responsible for exacerbation of obesity. In addition, a potential strategy for prevention of fluoride-induced obesity was proposed by intervention with polysaccharides from Fuzhuan brick tea. Conclusion Overall, these results provide the first evidence of a comprehensive cross-talk mechanism between fluoride and obesity in HFD fed mice, which is mediated by gut microbiota and intestinal barrier. E. ramosum was identified as a crucial mediator of fluoride induced obesity, which could be explored as potential target for prevention and treatment of obesity with exciting translational value.

Keywords: Erysipelatoclostridium ramosum; Exacerbation of obesity; Fluoride; Gut microbiota; Intestinal barrier permeability.

Graphical abstract

Fig. 1

Fluoride exacerbates the obesity in HFD mice. (A) Scheme of animal experiment 2 over the 10 weeks of dietary intervention. Mice were randomly divided into three groups including ND group, HFD group, and HFD plus 50 mg/L of fluoride in drinking water (HFD-F) group (n = 8). (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Perirenal fat. (E) Mesentery fat. (F) Epididymal fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I) Adipocytes (μm²) distribution in the epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT carried out at week 9, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 5 per group). (M) AUC for OGTT. (N) Liver weight. (O-Q) Representative morphology, H&E staining and Sirius Red histology of liver. (R) Plasma ALT. (S) Liver TG. The results were expressed as means ± SEM. Statistical significance was carried out by one-way analysis of variance (ANOVA) with Tukey test for multiple-group comparisons. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Fluoride drives the intestinal barrier permeability in HFD mice. (A) Representative morphology of colon. (B) Colon length (n = 8). (C-D) Representative H&E staining of colon (10 × and 40 ×, respectively). (E-H) Relative mRNA expression levels of Claudin-1, Muc1, Occludin and Zo-1 in colon (n = 8). (I-K) Representative immunofluorescence images of Claudin-1, MUC1 and ZO-1 in colon, 40 ×, scale bar = 100 μm. (L-N) Mean density of Claudin-1, MUC1 and ZO-1 in immunofluorescence images was evaluated by ImageJ software (n = 8). (O) Bacterial DNA/Whole DNA in blood was measured by RT-qPCR (n = 8). (P) Plasma LPS. (Q) Scheme of animal experiment 3. (R) Plasma FITC-dextran from mice in animal experiment 3 was measured after 4 h of intragastric administration of FITC-dextran (n = 8). The results were expressed as means ± SEM. Statistical significance was carried out by one-way ANOVA with Tukey test for multiple-group comparisons. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant.

Fig. 3

Fluoride exacerbates the obesity in HFD mice through a TLR4-dependent mechanism. (A) Scheme of animal experiment 4 over the 8 weeks of dietary intervention. Wild-type (WT) mice and Tlr4 knockout (Tlr4^−/−) mice were randomly divided into two groups, respectively, including HFD group, and HFD-F groups (n = 8 per group). (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Epididymal fat. (E) Mesentery fat. (F) Perirenal fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I)Adipocytes (μm²) distribution in the Epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT was carried out at week 7, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 5 per group). (M) AUC for OGTT. (N-P) Representative morphology, H&E staining and Sirius Red histology of liver. (Q) Liver weight. (R) Plasma ALT. (S) Liver TG. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Fluoride disturbs the gut microbiota in HFD mice. Gut microbiota was analyzed by 16S rRNA gene sequencing (n = 8 for each group). (A) PCA based on the relative abundance of Features of the gut microbiota, (B) PCoA of the gut microbiota based on the unweighted unifrac distance matrix, (C) Bacterial taxonomic profiling at the phylum level of gut microbiota, the relative abundances of (D) Firmicutes, (E) Bacteroidetes, and (F) the ratio of Firmicutes to Bacteroidetes. (G-H) LEfSe analyses of gut microbiota (value of Kruskal-Wallis: 0.01, value of Wilcoxon: 0.01, value of LDA score: 4). the relative abundances of (I) Erysipelotrichaceae, and (J)Erysipelatoclostridium, (K) Level of E. ramosum in fecal samples measured by RT-qPCR using species-specific primers. Higher CT values suggest lower levels of E. ramosum. The results were expressed as means ± SEM. Statistical significance was carried out by one-way ANOVA with Tukey test. Adjusted p-values (q-values) were used to evaluate differences in analysis of gut microbiota based on false discovery rate (FDR) for multiple testing according to the Benjamini and Hochberg procedure. A value of p or q < 0.05 was considered to be significant.A value of p < 0.05 was considered to be significant.

Fig. 5

The association between Erysipelotrichaceae and BMI based on the database of American Gut Project (AGP). (A) Spearman correlation between the relative abundance of Erysipelotrichaceae and BMI (n = 10376). (B) The relative abundance of Erysipelotrichaceae in three BMI categories, namely Normal (18.5 ≤ BMI < 25, n = 6546), Overweight (25 ≤ BMI < 30, n = 2869), and Obesity (BMI ≥ 30, n = 961). The results were expressed as means ± SEM. Statistical significance was carried out by the nonparametric Mann–Whitney test. Adjusted p-values (q-values) were used to evaluate differences in analysis of gut microbiota based on false discovery rate (FDR) for multiple testing according to the Benjamini and Hochberg procedure. A value of q < 0.05 was considered to be significant.

Fig. 6

FMT is sufficient to induce some phenotypical changes caused by fluoride in a HFD context. (A) Scheme of animal experiment 5 over the 10 weeks. Pseudo-germfree mice, induced by Abx for two weeks, were colonized by fecal slurry from either HFD group or HFD-F group mice (n = 8) for 8 weeks. (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Perirenal fat. (E) Epididymal fat. (F) Mesentery fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I) Adipocytes (μm²) distribution in the Epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT was carried out at week 9, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 4 per group). (M) AUC for OGTT. (N) Liver weight. (O) Liver TG. (P) Plasma ALT. (Q-S) Representative morphology, H&E staining and Sirius Red histology of liver. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7

The fluoride-induced key gut microbiota could colonize in pseudo-germfree mice by FMT. Gut microbiota was analyzed by 16S rRNA gene sequencing (n = 8 for each group). (A) PCA based on the relative abundance of Features of the gut microbiota, (B) PCoA of the gut microbiota based on the unweighted unifrac distance matrix, (C) Bacterial taxonomic profiling at the phylum level of gut microbiota, the relative abundances of (D) Firmicutes, (E) Bacteroidetes, and (F) the ratio of Firmicutes to Bacteroidetes. (G-H) LEfSe analyses of gut microbiota (value of Kruskal-Wallis: 0.05, value of Wilcoxon: 0.05, value of LDA score: 3.5). the relative abundances of (I) Erysipelotrichaceae, and (J)Erysipelatoclostridium. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant.

Fig. 8

The exacerbation of the obesity by fluoride is blocked after depletion of the gut microbiota by Abx. (A) Scheme of animal experiment 6 over the 10 weeks. Mice were randomly divided into two groups including HFD fed with Abx (HFD-Abx) group, and HFD fed with Abx plus 50 mg/L of fluoride in drinking water (HFD-Abx-F) group (n = 8 per group). Abx was fed by intragastric gavage. (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Epididymal fat. (E) Mesentery fat. (F) Perirenal fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I) Adipocytes (μm²) distribution in the Epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT was carried out at week 9, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 5 per group). (M) AUC for OGTT. (N) Liver weight. (O) Liver TG. (P) Plasma ALT. (Q-S) Representative morphology, H&E staining and Sirius Red histology of liver. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 9

Fluoride fails to induce the obesity in ND mice. (A) Scheme of animal experiment 7 over the 10 weeks of dietary intervention. Mice were randomly divided into two groups including ND group, and ND plus 50 mg/L of fluoride in drinking water (ND-F) group (n = 8 per group). (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Mesentery fat. (E) Epididymal fat. (F) Perirenal fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I) Adipocytes (μm²) distribution in the Epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT was carried out at week 9, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 5 per group). (M) Area under the curve (AUC) for OGTT. (N) Liver weight. (O) Liver TG. (P) Plasma ALT. (Q-S) Representative morphology, H&E staining and Sirius Red histology of liver. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 10

E. ramosum aggravates the obesity in HFD mice. (A) Scheme of animal experiment 8 over the 8 weeks. Pseudo-germfree mice were obtained by treatment of Abx for two weeks. Then, pseudo-germfree mice were randomly divided into two groups including HFD fed with PBS and HFD fed with E. ramosum (HFD-ER) by intragastric gavage for two weeks. (B) Dynamic changes in body weight in mice. (C) Body weight gain. (D) Epididymal fat. (E) Mesentery fat. (F) Perirenal fat. (G-H) Representative morphology and H&E staining of Epididymal fat. (I) Adipocytes (μm²) distribution in the Epididymal fat tissue. (J-K) Plasma levels of TC and LDL-C. (L) OGTT was carried out at week 7, mice were fasted overnight and gavaged with a dosage of glucose with 1.5 mg/g body weight (n = 5 per group). (M) AUC for OGTT. (N) Liver weight. (O) Liver TG. (P) Plasma ALT. (Q-S) Representative morphology, H&E staining and Sirius Red histology of liver. The results were expressed as means ± SEM. Difference in two groups was calculated using the Mann-Whitney test or Kruskal-Wallis test. A value of p < 0.05 was considered to be significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)