Clinical characteristics
The characteristics of 42 HC subjects, 48 No-NAs patients, and 61 TDF patients were presented in Table S1. Gender, age, and BMI were matched among the three groups (Table S1). After TDF treatment, CHB patients showed a significant reduction of HBV DNA load (p < 0.001) (Figure S1). In addition, compared with the No-NAs group, the levels of r-glutamyltransferase, alanine aminotransferase, and aspartate aminotransferase were significantly decreased, while the level of prealbumin was significantly increased in the TDF group (Table S2).
Compositional analysis of gut microbiota
The ACE and Simpson indices of each group did not differ significantly. Significant difference in bacterial richness (Chao1) and diversity (Shannon) were observed among the three groups. Chao1 and Shannon indices were significantly increased in the No-NAs group, but decreased after TDF treatment, while no significant difference was observed between the HC and TDF groups (Figure 1A).
The relative abundances of each bacterial phylum in patients and the healthy controls were showed in Figure 1B. Among the three groups, the main phyla were Firmicutes, Bacteroidota, Proteobacteria, Actinobacteria, Verrucomicrobiota, and Fusobacteriota, of which Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria accounted for more than 98% (Figure 1B). The relative abundance of Bacteroidota decreased gradually from HC group to No-NAs and TDF groups. Firmicutes and Actinobacteria were more abundant in No-NAs and TDF groups than in the HC group. The relative abundances of Fusobacteriota were significantly higher in the No-NAs group than those in HC group. Furthermore, the relative abundances of Proteobacteria were significantly lower in the No-NAs group than those in HC and TDF groups (Figure 1C).
Gut microbiota difference analysis at the genus levels among the three groups
PCoA based on the Bray-Curtis distance matrix showed obvious differentiation of bacterial communities among the HC, No-NAs and TDF groups (Figure 2A). PERMANOVA pair-wise interactions were used to identify significant differences between groups (p < 0.05, Table included in Figure 2A). The data of average relative abundances of each bacterial genus in patients and the healthy controls were showed respectively in the Table S3. We used LEfSe analysis to identify the genera responsible for the differences in the fecal microbiota of the three groups. The differences in intestinal microbiota among the three groups were shown in Figure 2B. Some beneficial bacteria such as Ruminococcus gnavus group and Fusobacterium were enriched in the HC group, while some opportunistic pathogens such as Sphingomonas and Escherichia/Shigella were increased in the No-NAs and TDF groups.
Identification of differential genera and key taxa
To identify the key bacteria responsible for the differences in the HC, No-NAs and the TDF groups, a taxonomy-based bacterial comparison was conducted (Figure 3). The abundances of Bacteroides were significantly decreased in TDF. The abundances of Fusobacterium were significantly decreased in No-NAs and TDF, but there was no significant difference between the No-NAs and TDF groups. Among the different predominant genera, Dialister, Eubacterium_hallii_group, Halomonas, Collinsella, Sphingomonas, Xanthomonadaceae_unclassified, and Rhizobiaceae_unclassified were found to be overrepresented in the No-NAs and TDF groups, but there was no significant difference between No-NAs and TDF groups. In addition, the abundance of Pelagibacterium and Hyphomonadaceae_uncultured gradually increased in HC, No-NAs and TDF groups. In particular, the abundance of Odoribacter was significantly reduced in TDF group, while there was no significant difference in HC and No-NAs groups (Figure 3).
Comparison of the inflammation profiles
The transfer of lipopolysaccharide (LPS) from the intestine to the blood can cause immune activation. Usually, serum inflammatory factors are elevated in CHB patients. Thus, our study focused on the effect of TDF treatment on serum inflammatory factors. Of the 21 inflammatory factors (IL-18, GM-CSF, IFNγ, IL-1b, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70, IL-13, IL-17, IL-17F, IL-21, IL-22, IL-23, IL-28A, MIP-3α, TGFβ1, TNFα, TNFβ) studied, the level of these markers shows no significant difference between the TDF and No-NAs groups (Table S4).
Correlation between gut microbiota and inflammatory factors
Spearman correlations between the relative abundance of bacterial genera and the levels of inflammatory factors were evaluated (Figure S2). Inflammatory factors were positively correlated with some “harmful” bacteria, such as Escherichia/Shigella, while negatively correlated with some beneficial bacteria, such as Stomatobaculum and Atopobium. We focused on the correlation between the three groups of significantly different genera (Figure 4A) and inflammatory factors. We found a positive correlation between Odoribacter and inflammatory factors. Alternatively, there was a significant correlation between microbial genera enriched in the No-NAs and TDF groups (Figure 4B).
HBV specific trends in functional profile
To investigate the HBV-specific functional characteristics of the current microbiome dataset, we sought to investigate HBV-specific trends in the functional composition of gut microbiota among the three groups. We observed an effect of HBV and TDF treatment on the functional profiles of the gut microbiota. As shown in Figure 5, KEGG pathway analysis revealed that 46 significantly enriched pathways related to metabolism and disease in the three groups. Based on the LDA selection, 30 predicted microbial functions including Metabolism of cofactors and vitamins, Secondary bile acid biosynthesis, Biotin metabolism, Folate biosynthesis, Vitamin B6 metabolism, and Zeatin biosynthesis were remarkably enriched in the HC group, 5 functions including drug metabolism other enzymes were remarkably enriched in the No-NAs group, and 11 predicted microbial functions including Butanoate metabolism and Styrene degradation were remarkably enriched in the TDF group (all p < 0.05).
Difference analysis between two Phases of the TDF group
On the day of specimen collection, we grouped patients treated with TDF according to their e-antigen status, namely: Phase 1 is the HBV e-antigen (HBeAg)-negative patient group, and Phase 2 is the HBeAg-positive patient group. Phase 1 and Phase 2 clinical baseline tables were shown in Table S5. The prealbumin of Phase 2 was significantly higher than that of Phase 1. In addition, there was no significant difference in inflammatory factors between Phase 1 and Phase 2, except IL4 (p = 0.047), which was significantly lower at Phase 2 (Table S6).
There was no significant difference in the richness of gut microbiota in the two subgroups, and it was found that the diversity of intestinal microbiota in the Phase 2 was significantly reduced (reflected in the Shannon index, but there was no significant difference in the Simpson index) (Figure 6A). In addition, at phylum level, Firmicutes and Bacteroidota were not significantly different. Proteobacteria were significantly increased in Phase 2 compared with Phase 1, while Actinobacteriota was significantly decreased (Figure 6B).
The difference between Phase 1 and Phase 2 was also manifested at the genus level. PCoA results showed that there was a significant difference between the two subgroups (adonis p value = 0.003) (Figure 7A). Some beneficial bacteria such as Roseburia and Bifidobacterium were found to be enriched in the Phase 1 group, while some opportunistic pathogens such as Escherichia/Shigella were enriched in the Phase2 group (Figure 7B). Further analysis revealed that Escherichia/Shigella was significantly increased in Phase 2, while the relative abundance of Roseburia, Oscillospiraceae_UCG.005 and Lachnospiraceae_CAG.56 was significantly decreased in the Phase2 (Figure 7C). No significant change in the relative abundance of Bifidobacterium was observed in these two subgroups.
Effect of duration of TDF administration on gut microbiota
To explore the effect of medication time on the gut microbiota, we divided the TDF group into two groups according to medication time less than 24 months (Short group) and more than 24 months (Long group). Beta diversity analysis revealed that there was a significant difference between the two groups (Figure S3). At the genus level, we found that the relative abundance of Faecalibacterium was increased with increasing medication time, while Dialister was decreased (Table S7).