3.1 Evaluation of PGF rat model
To find out whether the intestinal flora of rats in PGF group was depleted during the 4 weeks modeling periods, the total amount of intestinal bacteria was measured using qPCR in our previous work. The result manifested that total amount of intestinal bacteria in the intestine of rats was continuously consumed [4] during the antibiotic intervention. In the current study, we evaluated the PGF rat model from the perspective of gut to microbiota diversity. The alterations in body weight during antibiotic treatment exhibited in Fig. 1.
16S rRNA sequencing was performed on the feces of the PGF model and the control groups to analyze the alpha diversity of the intestinal flora. As shown in Fig. 2A, with the increase of tags, the observed species (Sob) curve tends to be flat, indicating that the sequencing depth of this study can cover the vast majority of the bacterial species in the samples, and the amount of data is reasonable. The alpha index comparison between the two groups was calculated by Welch's t-test using the R language Vegan package (version 2.5.3). Chao1 index and Shannon index were used as indicators to reflect species richness. The total number of species included in the community was measured by Chao1 index while Shannon index were used as indicators to reflect the total number of taxa included in the sample, reflecting the species richness and evenness. The results showed that the Chao1 and Shannon indexes in the PGF groups were significantly lower than those in the control group (Fig. 2B and 2C), indicating that the diversity of intestinal flora composition in the antibiotic-treated PGF group was significantly decreased (P < 0.01). However, there was no significant difference in alpha diversity between PGF groups with different modeling cycles.
To explore the beta diversity of intestinal microbiota in PGF model, UPGMA (Unweighted Pair Group Method with Arithmetic Mean) clustering tree multivariate statistical analysis (Fig. 3A) and PCoA principal coordinate analysis (Fig. 3B) based on bray-curtis distance were performed using the R language Vegan package. Welch's t test of beta distance (Fig. 3C) was performed to calculate the distance index (degree of dissimilarity) of different sample microbiota composition to reflect the beta diversity of fecal flora of the two groups. The value range of the Bray-curtis distance was from 0 to 1, that is, the types and abundances of the two groups were completely consistent to the two groups did not share any species. The results indicated that significant separation was observed between the control group and the PGF groups of different modeling cycles (Fig. 3A,B). The species composition of the control group and the PGF group was significantly different (Fig. 3C), while the composition of the intestinal flora of the rats in the PGF group at different periods was similar.
The results of qPCR and 16S rRNA analysis showed that the 16S rRNA gene abundance and species diversity of intestinal flora decreased significantly after 1 week of antibiotic intervention.
3.2 Gut Microbiota Conposition In Rats With Pgf Model
The feces of the PGF group and the control group were subjected to 16S rRNA sequencing and the composition of the intestinal flora in the control group and the PGF group at different periods was analyzed. Abundance statistics for each species taxonomy were displayed using Krona (version 2.6), and species Venn diagrams were plotted against each group of OTU abundance levels (Fig. 4A). According to species annotation information, count the number of tags sequences in each group at each taxonomic level (kingdom, phylum, class, order, family, genus, species). The top 10 species in abundance in each taxonomic level (phylum, class, order, family, genus) were selected in present study. In all samples, Bacteroidetes were mainly represented by the class Bacteroidia, the order Bacteroidales, the families Bacteroidaceae, Prevotellaceae and Muribaculaceae and the genus Bacteroides, Sediminibacterium, Prevotellaceae_NK3B31_group and Prevotellaceae_UCG-001. In parallel, the composition of Firmicutes in samples included diverse genera, such as Roseburia, Fusicatenibacter and Lachnospiraceae_NK4A136_group (Fig. 4). Compared with the control group, the proportions of species at all taxonomic levels changed in the PGF group (Fig. 4B-F), while the changes in the proportion of species between the PGF groups with different modeling periods were small, implying that the gut microbial richness and diversity of the all PGF model rats decreased in varying degrees.
3.3 Indicator Species Of Pgf Model
The dominant species at the generic level in the control group and the PGF groups are shown in Fig. 5A-D, which shows that the indicator species in the control group are significantly more than those in the PGF group. In addition, the LEfSe software (version 1.0) was used to screen the indicator species at each level (phylum, class, order, family, and genus) based on the phylogenetic tree. Figure 6A-B shows the difference species with LDA (Linear Discriminant Analysis, Linear Discriminant Analysis) value greater than 3.5, and the histogram shows the influence of different species at each level. Most bacterial genera of PGF were depleted compared to the control group (P < 0.01). However, the relative abundance of Escherichia-Shigella, as an indicator species in the pgf group, was significantly higher than that in the control group. In addition, the relative abundance of Ralstonia, Bradyrhizobium in the PGF-1, 2, 3W groups was significantly higher than that in the control group, while PGF-4W group showed no difference compared with the control group. It seems that the gut flora of PGF-4W was depleted on a wider scale.
3. Changes Of Cyp450s Enzyme Expression In Pgf Model Rats
As previously described, gut microbiota and liver microsomes 450 co-shape the individual metabolic landscape. Many researchers explore the impact of gut microbiota on the metabolism of foreign substances using PGF model, often ignoring changes in host liver metabolic enzymes during modeling period. Here we found changes in the expression of hepatic metabolic enzymes in PGF group after 1 week of modeling time. Compared with the control group, the expression levels of CYP1A2, CYP2C19 and CYP2E1 in the PGF group were significantly upregulated (Fig. 7A–G). Meanwhile, the differences of the expression levels of CYP2C9, CYP2D6 and CYP3A4 between the control group and the PGF model were not statistically significant. Results showed that the liver metabolic enzyme expression of PGF model rats induced by antibiotics were lower than those of the control group.
3.5 Intestinal Mucosal Barrier Function Of Pgf Model
As indicators of intestinal mucosal permeability and inflammatory level, the DAO, ET and LEP levels were compared between the control group and PGF groups at different modeling periods. As shown in Fig. 8, there was no significant difference in the contents of serum ET and LEP among the groups, while The intestinal permeability index DAO was significantly increased from the second week of PGF model with antibiotics treatment, suggesting the intestinal mucosa was not inflamed but the permeability was improved from week 2 on. These results provide a new perspective on the differences in blood drug concentrations in PGF model rats, that is, the effect of increased drug absorption levels on blood exposure.
3.6 Gut Microbial Function Of Pgf Model
To describe the overall metabolic landscape of intestinal microbiota metabolism, we annotated the KEGG functional pathway of sample bacteria using Tax4Fun (version 1.0). As can be seen from the river plot (Fig. 9A), the functional pathway abundance characteristics of control group was different from the PGF groups with different modeling cycles. In addition, the PGF groups at weeks 1, 2, 3, and 4 also exhibited similar functional pathway abundance profiles. Then, the Kruskal-Wallis rank sum test was used to analyze the functional pathways of the control group and the PGF groups under different modeling cycles using the R language Vegan software package (version 2.5.3). The P-value threshold of 0.001 was used as a filter to display the functional prediction results. The predicted results of intestinal flora function of rats in the control group and PGF group are shown in Fig. 9B-C, and the pathways listed in Figure C are all differential metabolic pathways. What is striking is that the pathway abundence of nucleotide metabolism, lipid metabolism, metabolism of other Amino Acids and Metabolism of Terpenoids and Polyketides showed significant differences between the control group and PGF groups, implying the metabolic enzymes involved in the metabolic pathways in PGF intestinal microbiota were changed.