Alterations and correlations of gut microbiota, fecal, and serum metabolome characteristics in a rat model of alcohol use disorder

Abstract

Background: Growing evidence suggests the gut microbiota and metabolites in serum or fecal may play a key role in the process of alcohol use disorder (AUD). However, the correlations of gut microbiota and metabolites in both feces and serum in AUD subjects are not well understood.

Methods: We established a rat model of AUD by a chronic intermittent ethanol voluntary drinking procedure, then the AUD syndromes, the gut microbiota, metabolomic profiling in feces and serum of the rats were examined, and correlations between gut microbiota and metabolites were analyzed.

Results: Ethanol intake preference increased and maintained at a high level in experimental rats. Anxiety-like behaviors was observed by open field test and elevated plus maze test after ethanol withdraw, indicating that the AUD rat model was successfully developed. The full length 16S rRNA gene sequencing showed AUD significantly changed the β-diversity of gut microbial communities, and significantly decreased the microbial diversity but did not distinctly impact the microbial richness. Microbiota composition significantly changed in AUD rats, such as the abundance of Romboutsia and Turicibacter were significantly increased, whereas uncultured_bacterium_o_Mollicutes_RF39 was decreased. In addition, the untargeted metabolome analysis revealed that many metabolites in both feces and serum were altered in the AUD rats, especially involved in sphingolipid metabolism and glycerophospholipid metabolism pathways. Finally, multiple correlations among AUD behavior, gut microbiota and co-changed metabolites were identified, and the metabolites were directly correlated with the gut microbiota and alcohol preference.

Conclusion: The altered metabolites in feces and serum are important links between the gut microbiota dysbiosis and alcohol preference in AUD rats, and the altered gut microbiota and metabolites can be potentially new targets for treating AUD.

Keywords: alcohol preference; alcohol use disorder; fecal metabolites; gut microbiota; serum metabolites.

Figure 1

The schedule timeline of the IA2BC procedure in the experiment design. OFT represent open field test, EPM represent elevated plus maze test.

Figure 2

IA2BC affected alcohol intake preference but did not affect body weight or liquid/food intake. (A) The total liquid intake showed no significant difference between the water day and EtOH day in EtOH group. (B) The ethanol consumption was increasing with the water consumption was decreasing in IA2BC rats. (C) Alcohol intake preference was increasing with chronic intermittent ethanol voluntary drinking time. The total liquid intake (D), food consumption (E) or body weight (F) did not altered by chronic intermittent ethanol voluntary drinking. Data are presented as Mean ± SEM, n = 8/group; compared with the first day, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3

IA2BC rats exhibited anxiety-like behaviors. Behavior test was performed at the 28th day of the procedure. In the open field test, EtOH group rats spent less time in the center zone (t = 2.47, p = 0.027), traveled significantly less distance in the center zone (t = 2.30, p = 0.038) and less total distances (t = 2.94, p = 0.011) than that of CON group (A–C). The resting time between EtOH and CON group had no significant difference (t = 1.04, p = 0.351, D). In the elevated plus maze, the EtOH group spent less time in open arms (t = 3.04, p = 0.009) and the number entries into open arms significantly reduced (t = 3.24, p = 0.006) compared with CON group (E,F). Results are displayed as means ± SD. Significant results were determined by unpaired t-tests, *p < 0.05, **p < 0.01.

Figure 4

Chronic intermittent ethanol voluntary drinking alters the composition of gut microbiota in rats. Histogram of top 10 relative abundance at different level. The relative abundance and the main bacterial genera between the two groups with Wilcoxon rank-sum test at the phylum (A–E), family (F–J), genus (K–N), and species (O–R) level in the fecal samples. Values are presented as the mean ± SD (n = 8 per group), *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 5

Linear discriminant analysis Effect Size (LEfSe) analysis between the EtOH and CON groups, with a LAD score > 4.

Figure 6

Chronic intermittent ethanol voluntary drinking alters the diversity of gut microbiota in rats. Difference in alpha diversity index between EtOH and CON group showed in the plot (A–D). (A) represent ACE index; (B) represent Chao1 index; (C) respresent Simpson index; (D) represent Shannon index. The width of each curve in the violin plot corresponds with the approximate frequency of data points in each region, dotted line indicating the median value and quartile positions. Significant results were determined by unpaired t-tests, ***p < 0.001; ns, p > 0.05. (E), ANOSIM analysis showed differences between EtOH and CON groups were significantly greater than those within groups. (F) PCoA plot based on Bray–Curtis distance, EtOH and CON groups could be effectively separated, showing that the composition of the gut microbiota in EtOH group was significantly different from that of CON group. (G) samples heat map based on the Bray–Curtis distance showed that the samples in the same group clustered together, indicating that EtOH changed the gut microbiota community.

Figure 7

Multivariate statistical analysis of fecal metabolites measured by untargeted metabolomics analysis at positive (A–C) and negative (D–F) ion mode. PCA analyses comparing metabolites of all samples between EtOH and CON group (A,D); OPLS-DA scores showed significant differences between EtOH and CON group (B,E); and the OPLS-DA permutation test confirmed the differences of fecal metabolites in EtOH and CON group (C,F).

Figure 8

The fecal metabolic profile changed in AUD rats. Differential metabolites in fecal identification between EtOH and CON group at positive (A,C) and negative (B,D) ion mode. Expression of differential metabolites in the two groups was represented by Volcano plot (A,B) and heat maps (C,D). The number of differential metabolites were annotated and classified based on the HMDB database (E). The enrichment pathway of fecal differential metabolites by Kyoto encyclopedia of genes and genomes (KEGG) analysis (F).

Figure 9

Correlation heatmap of differential microbiota and fecal metabolites. Data were calculated by Spearman’s correlation method after mean centering and unit variance scaling. *p < 0.05, ** p < 0.01.

Figure 10

Multivariate statistical analysis of serum metabolites measured by untargeted metabolomics analysis at positive (A–C) and negative (D–F) ion mode. PCA analyses comparing metabolites of all samples between EtOH and CON group (A,D); OPLS-DA scores showed significant differences between EtOH and CON group (B,E); and the OPLS-DA permutation test confirmed the differences of serum metabolites in EtOH and CON group (C,F).

Figure 11

Differential metabolites in serum identification between EtOH and CON group at positive (A,C) and negative (B,D) ion mode. Expression of differential metabolites in the two groups was represented by Volcano plot (A,B) and heat maps (C,D). The number of differential metabolites were annotated and classified based on the HMDB database (E). The enrichment pathway of serum differential metabolites by Kyoto encyclopedia of genes and genomes (KEGG) analysis (F).

Figure 12

Correlation heatmap of differential microbiota and serum metabolites. Data were calculated by Spearman’s correlation method after mean centering and unit variance scaling. *p < 0.05, **p < 0.01.

Figure 13

Correlations of gut microbiota with the co-regulated metabolites in feces and serum in AUD rats. The Venn diagram showed the co-regulated metabolites in feces and serum (A). The enrichment pathway of co-regulated metabolites by KEGG analysis (B). Network analysis among the AUD behavior, gut microbiome and co-regulated metabolites (C), the red line represented the positive correlation, the blue line represented negative correlation, |r| > 0.8, p < 0.05.