Betaine ameliorated liver and small intestine injury in ALF mice
Pathological changes in mouse liver tissue and serum biochemical markers were assessed. The structure of liver lobules and the arrangement of liver cells were clear and orderly in the normal group. There was also no necrosis of hepatocytes or infiltration of inflammatory cells in the normal group. However, in the model group, the liver lobular structure was destroyed, and a larger amount of hepatocyte necrosis was observed, with massive inflammatory cell infiltration. The degrees of hepatocyte necrosis and inflammation was were lessened in the betaine group compared with those in the model group (Fig. 1A). Serum ALT, AST and TBIL levels in the model animals were significantly increased compared with those in the normal group animals (P < 0.05). Nevertheless, the ALT, AST and TBIL levels were much lower in the betaine groups than in the model group (P < 0.05) (Fig. 1C-E). Compared with those in the normal group, the serum levels of TNF-α, IL-1β and IL-18 were greatly increased in the model group. Betaine significantly decreased the levels of TNF-α, IL-1β and IL-18 in the model group (P <0.05) (Fig. 1F-H). As shown in Fig. 1B, histological analysis indicated that the lamina propria and villi of the model group were denuded, the epithelial layer of the small intestine was exfoliated, the height of the small intestine was reduced, and abundant inflammatory cell infiltration was observed. In contrast, betaine conserved almost normal architecture in the small intestine. Moreover, the intestinal permeability remarkably increased in the model group compared with that in the normal group (P < 0.05). Betaine dramatically improved intestinal permeability in the model animals (P < 0.05) (Fig. 3A).
Betaine inhibited the TLR4/MyD88 pathway and improved the mRNA and protein expression of (ZO)-1 and occludin in ALF mice
The effect of betaine by regulating the TLR4/MyD88 pathway in ALF model mice was assessed. Compared with those in the normal group, the levels of the TLR4, MyD88, TRAF6 and TNF-α proteins and the levels of TLR4 and MyD88 mRNA were obviously increased in the model group (P < 0.05). The protein levels of TLR4, MyD88, TRAF6 and TNF-α and the mRNA levels of TLR4 and MyD88 were significantly decreased in the betaine group compared with those in the model group (Fig. 2A-C and Fig. 3C). Furthermore, the results showed that the mRNA and protein levels of (ZO)-1 and occludin were both distinctly decreased in the model group compared with those in the normal group (P < 0.05). Betaine significantly elevated the expression of (ZO)-1 and occludin in the model animals (P < 0.05) (Fig. 2C-D and Fig. 3D).
Betaine suppressed the TLR4/MyD88 pathway and enhanced the levels of (ZO)-1 and occludin in IEC-18 cells stimulated by LPS
Cell experiments were performed to verify the protective effect of betaine in LPS-stimulated IEC-18 intestinal epithelial cells. The results showed that the protein levels of TLR4, MyD88, TRAF6 and TNF-a were prominently increased in the model group compared with those in the normal group (P < 0.05) (Fig. 2E-H). Compared with those for the model treatment, the protein levels of TLR4, MyD88, TRAF6 and TNF-a were dramatically suppressed by betaine administration (P < 0.05). The inhibitory effect of betaine on the protein expression of TLR4, MyD88, TRAF6 and TNF-a was dose dependent (P < 0.05). The TLR4 and MyD88 mRNA and protein levels were also significantly different among the high-, medium- and low-dose betaine groups (Fig. 3E, P < 0.05). As shown in Fig. 2G, 2H and Fig. 3F, compared with those in the normal group, the protein and mRNA expression levels of (ZO)-1 and occludin were markedly reduced in the model group (P < 0.05). The mRNA and protein expression of (ZO)-1 and occludin were significantly improved in the betaine group compared with those in the model group (P < 0.05). Moreover, the mRNA and protein levels of (ZO)-1 and occludin in the medium-dose betaine group were elevated compared with those in the low-dose betaine group and were the highest in the high-dose betaine group (P < 0.05). This increased expression also showed dose dependence.
Betaine elevated the TEER value in LPS-stimulated IEC-18 cells
As shown in Fig. 3B, the TEER value was greatly decreased in the model group compared with that in the normal group (P < 0.05). The TEER value was significantly elevated after betaine treatment in the model group (P < 0.05). In addition, there were significant differences among the high-, medium- and low-dose betaine groups (P < 0.05).
OTU analysis
In our experiments, 16S rRNA gene sequence analysis was used to explore the potential therapeutic microecological mechanisms of betaine in ALF mice. The OTU network analysis for fecal samples of mice provided certain core OTUs and a universal microbial composition among each group. A total of 509 OTUs were identified from 15 fecal samples (Supplementary files Table S1). As shown in Fig. 4A, the normal group contained the maximum OTUs, but the betaine and model groups possessed a similar number of OTUs. The core microbiome that was present in each fecal sample could be found based on the shared OTUs of each sample and the species represented by the OTUs. There were 156 core microbiome constituents in fecal samples (Supplementary files Table S2).
Alpha diversity analysis
Alpha diversity contains the observed species index, the Chao1 index, the Shannon index and the Simpson index. The observed species index indicates the actual number of OTUs observed, and the Chao1 index is performed to calculate the total number of OTUs contained in a sample. Both of them indicate the species richness of the sample. The Simpson index and Shannon index are applied to evaluate species diversity. As shown in Table 2, the species richness and diversity of the microbiota among the three groups did not have notable differences. However, the fecal bacteria in the normal group had relatively higher levels of species richness and diversity than those in the other groups.
Beta diversity analysis
Different from alpha diversity analysis, beta diversity analysis is appropriate to distinguish the differences in species diversity among a pair of samples. UniFrac compares species community differences using phylogenetic evolution information. The results can serve as an index to detect beta diversity, which has taken into account the evolutionary distance between species. UniFrac results are classified into weighted UniFrac and unweighted UniFrac. Weighted UniFrac considers the abundance of sequences, and unweighted UniFrac does not. Beta diversity analyses include ANOSIM and principal coordinates analysis (PCoA). ANOSIM is a nonparametric test used to detect whether the difference between two or more groups is significantly greater than the intragroup difference to judge whether the grouping is reasonable. As shown in Fig. 5A-B, ANOSIM using weighted UniFrac distances and unweighted UniFrac clustered samples. The data revealed that intermouse variations in the fecal microbiota were lower than the intragroup variations, which showed that the fecal microbiota of each group had better individual similarity. PCoA showed that the distances between two samples were close, which indicated that the species compositions of the two samples were similar. The results of PCoA showed that the microbial composition between the normal group and betaine group was more similar than that between the normal and model groups (Fig. 5C-D).
LDA Effect Size (LEfSe)
LEfSe emphasizes statistical significance and biological relevance. As shown in Fig. 5E-F, a cluster tree displayed different colors representing different groups, and nodes of different colors represented corresponding important microorganisms in each group. However, the yellow node represents nonsignificant microbes. The 11 bacterial species in fecal samples, such as g-Bacteroide, f-Bacteroidaceae, o-Campylobacterales and c-Epsilonproteobacteria, had a significant effect on the normal group. F-Lachnospiraceae, g-Enterorhabdus, o-Coriobacteriales and f-Coriobacteriaceae contributed greatly in the model group. G-Parabacteroides, g-Odoribacter, g-Prevotella, f-Oxalobacteraceae, g-Anaerovorax and f-Clostridiales-IncertaeSedisXIII were of crucial importance in the betaine group. As shown in Fig. 6A and Supplementary files Table S3, there were a total of 143 OTUs that had a significant difference between groups (P < 0.05). As shown in Fig. 6B-C and Supplementary files Table S4, a total of 24 species, which represented 11 genera, were significantly different between groups (P < 0.05). At the genus level, they are g-Anaeroplasma, g-Anaerovorax, g-Bacteroides, g-Desulfovibrio, g-Dorea, g-Enterorhabdus, g-Helicobacter, g-Odoribacter, g-Parabacteroides, g-Paraprevotella and g-Prevotella. Among them, an increased relative abundance of g-Enterorhabdus was detected in the model group compared with that in the normal group (P < 0.05). Betaine downregulated the relative abundance of g-Enterorhabdus (P < 0.05). The relative abundance of g-Bacteroides was the highest in the normal group and the lowest in the model group (P < 0.05). The relative abundance of g-Prevotella was almost the same in the normal and betaine groups and was reduced in the betaine group (P < 0.05).
Species classification and abundance analysis
A sequence with the highest abundance was selected from each OTU as a representative sequence of the OTU. Using the RDP method, the representative sequence was aligned with the 16S database to classify each OTU to a corresponding species. Forming a relative abundance histogram of species, we visually observed the proportion of different species abundances in each sample and group. As shown in Fig. 7A-E and Fig. 8A-E, at the classification level of phylum, class, order, family, and genus, the corresponding histograms of microbiome species profiling were produced for each sample and group. The majority of the microbiome at the phylum level among each group belonged to Firmicutes, Bacteroidetes, Proteobacteria, Deferribacteres, Actinobacteria, Tenericutes, Verrucomicrobia and Candidatus Saccharibacteria. Surprisingly, the majority of the microbiome phyla occupied similar proportions between the normal group and the model group. In addition, Firmicutes was the most abundant phylum in each group. The relative abundance of Firmicutes in model mouse feces (58.8%) was significantly higher than that in the normal (42.5%) and betaine group mouse feces (42.3%) (P < 0.05). A previous study confirmed that the levels of Firmicutes were positively correlated with ACLF severity and that the abundance of Bacteroidetes was inversely correlated. It has been suggested that betaine plays a role in potentially modifying the gut bacterial community. The relative abundance of Bacteroidetes in model mouse feces (37.8%) was apparently lower than that in the normal (50.2%) and betaine group mouse feces (49.5%) (P < 0.05) (Fig. 8A and Supplementary files Table S5). The relative abundance of Alistipes (belonging to Bacteroidetes) was enriched in the normal group (22.9%) and decreased in the betaine (18.8%) and model groups (19.1%) (P < 0.05). Moreover, compared with that in the normal group, the relative abundance of Clostridium XlVa (belonging to Bacteroidetes) was significantly higher in the model group. Betaine pretreatment reduced the relative abundance of Clostridium XlVa in the model group (P < 0.05) (Fig. 8E and Supplementary files Table S6). In general, the abundances of the gut microbiota taxa Coriobacteriaceae, Lachnospiraceae, Enterorhabdus and Coriobacteriales were remarkably increased in the model group, contrary to those of the gut microbiota taxa Bacteroidaceae, Bacteroides, Parabacteroides and Prevotella. Betaine significantly altered the microbial communities, depleted the gut microbiota constituents Coriobacteriaceae, Lachnospiraceae, Enterorhabdus and Coriobacteriales and markedly enriched the taxa Bacteroidaceae, Bacteroides,Parabacteroides and Prevotella (P < 0.05) (Fig. 8A-E). These results indicated that alteration of the gut microbiome might be a crucial therapeutic target in ALF.