Gut Commensal Barnesiella Intestinihominis Ameliorates Hyperglycemia and Liver Metabolic Disorders

Abstract

Recent studies have highlighted the role of the gut microbiota in type 2 diabetes (T2D). Improving gut microbiota dysbiosis can be a potential strategy for the prevention and management of T2D. Here, this work finds that the abundance of Barnesiella intestinihominis is significantly decreased in the fecal of T2D patients from 2-independent centers. Oral treatment of live B. intestinihominis (LBI) considerably ameliorates hyperglycemia and liver metabolic disorders in HFD/STZ-induced T2D models and db/db mice. LBI-derived acetate has similar protective effects against T2D. Mechanistically, acetate enhances fibroblast growth factor 21 (FGF21) through inhibition of histone deacetylase 9 (HDAC9) to increase H3K27 acetylation at the FGF21 promoter. The screening puerarin from Gegen Qinlian decoction in a gut microbiota-dependent manner improved hyperglycemia and liver metabolic disorders by promoting the growth of B. intestinihominis. This study suggests that gut commensal B. intestinihominis and puerarin, respectively have the potential as a probiotic and prebiotic in the treatment of T2D.

Keywords: acetate; fibroblast growth factor 21; gut microbiota; histone deacetylase 9; puerarin; type 2 diabetes.

Figure 1

The reduced abundance of B. intestinihominis in T2D mice and patients. A) Schematic diagram of the study design. B) Weighted UniFrac PCoA analysis based on the OTU data of chow and HFD/STZ groups (n = 7). C) The abundance of bacteria at the phylum level (percentage of total bacteria). D) Firmicutes‐to‐Bacteroidetes ratio (n = 7). E) The abundance of bacteria at the genus level (percentage of total bacteria). F) LDA score represents the taxonomic data with significant differences between chow and HFD/STZ groups (p < 0.05, LDA scores > 4). G) Pearson's analysis of the correlations between the fecal Barnesiella abundance and HOMA‐IR index. H) The abundance of B. intestinihominis, B. viscericola and B. propionica was assessed by qPCR (n = 7). I) Pearson's analysis of the correlations between the fecal B. intestinihominis abundance and HOMA‐IR index. J) Chao index from shotgun metagenomics. K) PCoA analysis for the human fecal from T2D patients (n = 31) and healthy subjects (n = 30). L) LDA score represents the taxonomic data with significant differences between T2D patients (n = 31) and healthy subjects (n = 30) in genus and species (p < 0.05, LDA scores > 3). M) Pearson's analysis of the correlations between the fecal Barnesiella abundance from shotgun metagenomics and 2hPBG and HbA1c, respectively. N) Pearson's analysis of the correlations between the fecal B. intestinihominis abundance from shotgun metagenomics and 2hPBG and HbA1c, respectively. O) The abundance of B. intestinihominis using T2D patients (n = 150) and healthy subjects (n = 56) from microbiota validation set 1, assessed by qPCR. P) The abundance of B. intestinihominis using T2D patients (n = 57) and healthy subjects (n = 47) from microbiota validation set 2, assessed by qPCR. Q) The abundance of Barnesiella of T2D patients (n = 233) and healthy subjects (n = 8697) from GMrepo Database. R) The abundance of B. intestinihominis of T2D patients (n = 112) and healthy subjects (n = 3356) from GMrepo Database. Data were shown as mean ± SEM. Statistical analysis was performed by a two‐tailed Student's t‐test (D and H) or Mann‐Whitney U test (O‐R). ***p < 0.001; ns, no significance.

Figure 2

Oral administration of LBI attenuates metabolic disorder and gut barrier dysfunction in HFD/STZ mice. A) A schematic diagram showing the procedure of HFD/STZ mice treated with LBI and KBI. B) The abundance of B. intestinihominis assessed by qPCR (n = 7). C) Fasting blood glucose levels at 2, 4 and 6 weeks (n = 7). D) Insulin levels (n = 7). E) HOMA‐IR index (n = 7). F‐G) OGTT F) and ITT G) with AUC (n = 7). H) Liver weight (n = 7). I) Representative photomicrographs of liver H&E staining and histological scores (scale bar, 100 µm, n = 4–5). J) The serum LPS levels (n = 7). K) Representative photomicrographs of colon H&E staining and histological scores (scale bar, 100 µm, n = 4). L‐N) Colon expression of ZO‐1 and Occludin assayed by western blot and quantitation using Image J software (n = 3). O‐Q) Colon expression of ZO‐1 and Occludin assayed by immunohistochemistry and calculation of integrated option density using Image Pro Plus software (scale bar, 50 µm, n = 3). Data were shown as mean ± SEM. Statistical analysis was performed by Kruskal‐Wallis test (B) or one‐way ANOVA with Dunnett's post‐test (C‐K, M‐N, and P‐Q). *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 3

Metabolomic identified the B. intestinihominis‐derived metabolites. A) PCA score plots for discriminating the fecal metabolome from chow, HFD/STZ, and LBI groups (n = 7). B) OPLS‐DA analysis of metabolic profiles in HFD/STZ and LBI groups (n = 7). C) Heatmaps of the differential metabolites that were altered by HFD/STZ compared with LBI‐treated mice (n = 7). D) Metabolic pathways in the HFD/STZ versus LBI groups. E) Acetate levels in fecal (n = 7). F) Acetate levels in the liver (n = 7). G) OPLS‐DA analysis of metabolic profiles in healthy subjects (n = 30) and T2D patients (n = 32). H) Heatmaps of the differential metabolites that were altered by healthy subjects (n = 30) compared with T2D patients (n = 32). I) The levels of acetate from T2D patients (n = 71) and healthy subjects (n = 30). J) Pearson's analysis of the correlations between the abundance of B. intestinihominis and acetate levels. K‐L) Pearson's analysis of the correlations between the levels of acetate and 2hPBG (K) and HbA1c (L), respectively. M) Acetate levels in the supernatant after 48 h of incubation with B. intestinihominis (n = 3). N) Wood‐Ljundahl pathway. O) The viability of ackA in fecal (n = 7). P) Relative expression of ackA in fecal (n = 4). Data were shown as mean ± SEM. Statistical analysis was performed by one‐way ANOVA with Dunnett's post‐test (E‐F and O‐P), two‐tailed Student's t‐test (M), or Mann‐Whitney U test (I). *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 4

Acetate attenuated metabolic disorder and gut barrier dysfunction in HFD/STZ mice. A‐C) The consumption A), gluconeogenesis B), and the glycogen content C)were detected (n = 3) in primary hepatocytes. D) A schematic diagram showing the procedure of HFD/STZ mice treated with acetate. E) Fasting blood glucose levels at 2, 4, and 6 weeks (n = 7). F) Insulin levels (n = 7). G‐H) OGTT G) and ITT H) with AUC (n = 7). I) HOMA‐IR index (n = 7). J) Liver weight (n = 7). K) Liver weight/body weight ratio (n = 7). L‐M) AST and ALT levels in the blood (n = 7). N) Representative photomicrographs of liver H&E staining and histological scores (scale bar, 100 µm, n = 5). O‐P) TG and TC levels in the blood (n = 7). Q) The serum LPS levels (n = 7). R) Representative photomicrographs of colon H&E staining and histological scores (scale bar, 100 µm, n = 4–5). S) Colon expression of ZO‐1 and Occludin assayed by western blot and quantitation using Image J software (n = 3). T) Colon expression of ZO‐1 and Occludin assayed by immunohistochemistry and calculation of integrated option density using Image Pro Plus software (scale bar, 50 µm, n = 3–4). Data were shown as mean ± SEM. Statistical analysis was performed by one‐way ANOVA with Dunnett's post‐test. *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 5

Acetate activated FGF21 gene transcription through down‐regulation of HDAC9. A) Volcano plot shows the number of differential genes downregulated or upregulated by acetate (n = 3). B) Gene ontology analysis in the HFD/STZ versus LBI groups (n = 3). C) Heatmaps of the differential gene related to histone deacetylation terms (n = 3). D) Liver expression of HDAC9 assayed by western blot and quantitation using Image J software (n = 3). E) Liver expression of HDAC9 assayed by immunohistochemistry and calculation of average option density using Image Pro Plus software (n = 5). F) Western blotting analysis of HDAC9 expression in HepG2 cells (n =  3). G–I) Detection of the gluconeogenesis G), glucose consumption H), and glycogen content I) after overexpression of HDAC9 (n = 3). J‐L) Detection of the gluconeogenesis (J), glucose consumption (K), and the glycogen content L) after knocking down HDAC9 (n = 3). M) Liver expression of FGF21 assayed by western blot and quantitation using Image J software (n = 3). N) Relative mRNA expression of FGF21 in the liver (n = 5). O) Overexpression of HDAC9 to detect the expression of FGF21 in HepG2 cells (n = 3). P) Detection of FGF21 expression in HepG2 cells after knocking down HDAC9 (n = 3). Q) Detection of H3K27ac enrichment in the FGF21 promoter region using ChIP‐qPCR (n = 4). Data were shown as mean ± SEM. Statistical analysis was performed by one‐way ANOVA with Dunnett's post‐test (D‐P) or two‐tailed Student's t‐test (Q). *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 6

Puerarin attenuated hyperglycemia and liver metabolic disorder in HFD/STZ mice in a gut microbiota‐dependent manner. A) Schematic diagram for antibiotic treatment. B) Fasting blood glucose levels at 2, 4, and 6 weeks for antibiotic treatment (n = 7). C) Insulin levels for antibiotic treatment (n = 7). D) HOMA‐IR index for antibiotic treatment (n = 7). E‐F) OGTT (E) and ITT (F) with AUC (n = 7). G) Liver expression of HDAC9 and FGF21 in antibiotic‐treated HFD/STZ mice assayed by western blot and quantitation using Image J software (n = 3). H) Experimental design diagram for fecal microbiota transplantation. I) Fasting blood glucose levels at 2, 4, and 6 weeks after fecal microbiota transplantation (n = 7). J) Insulin levels for fecal microbiota transplantation (n = 7). K‐L) OGTT (K) and ITT (L) with AUC (n = 7). M) HOMA‐IR index for fecal microbiota transplantation (n = 7). N) Representative photomicrographs of liver H&E staining and histological scores after fecal microbiota transplantation (scale bar, 100 µm, n = 5). O) Liver expression of HDAC9 and FGF21 after fecal microbiota transplantation assayed by western blot and quantitation using Image J software (n = 3). Data were shown as mean ± SEM. Statistical analysis was performed by one‐way ANOVA with Dunnett's post‐test (B‐G) or two‐tailed Student's t‐test (I‐O). *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 7

LBI, acetate, and puerarin attenuated hyperglycemia and liver metabolic disorder in db/db mice. A) A schematic diagram showing the procedure of db/db mice treated with LBI, acetate, and puerarin. B) Body weight curve (n = 6). C) Body weight gain (n = 6). D) Fasting blood glucose levels at 2, 4, and 6 weeks (n = 6). E‐F) OGTT (E) and ITT (F) with AUC (n = 6). G) Insulin levels (n = 6). H) HOMA‐IR index (n = 6). I) Liver expression of HDAC9 and FGF21 assayed by western blot and quantitation using Image J software (n = 3). J) The serum LPS levels (n = 6). K) Representative photomicrographs of colon H&E staining and histological scores (scale bar, 100 µm, n = 5). L) Colon expression of ZO‐1 and Occludin assayed by western blot and quantitation using Image J software (n = 3). Data were shown as mean ± SEM. Statistical analysis was performed by one‐way ANOVA with Dunnett's post‐test (C‐L). *p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance.

Figure 8

Schematic illustration of the proposed underlying mechanism by how B. intestinihominis ameliorated hyperglycemia and liver metabolic disorders in T2D.