Consecutive patients with LC were recruited for this prospective observational study, which was conducted between January 2022 and December 2022. The diagnosis of LC was established by evaluating clinical, laboratory, and radiological data, and/or conducting liver biopsies. The diagnosis of PVT followed the consensus guidelines for managing PVT in LC patients (2020, Shanghai), in brief the diagnosis of PVT relies mainly on imaging (liver vascular ultrasound, CT or MRI; Hepatobiliary Disease Study Group et al., 2021). The inclusion criteria required patients to be between 18 and 70 years of age, with well-characterized cirrhosis that was confirmed by compatible clinical, imaging, liver transient elastography, and laboratory values. Patients were excluded from the study if they were on anticoagulation therapy at the time of enrollment, had hepatocellular carcinoma (HCC), pregnant, prior trans jugular intrahepatic portosystemic shunt (TIPS), prior liver transplant, or systemic neoplasia.

This research was conducted in compliance with the tenets of the Declaration of Helsinki. The Ethics Committee of Sichuan Provincial People’s Hospital approved the study (No. SCCT2021-525), and all patients provided informed consent prior to participation.

A standardized method was used to collect fecal samples. Samples were transported to the laboratory within 2 h of collection along with an ice pack. The samples were promptly frozen and reserved at −80°C until the time of analysis.

The QIAamp Fast DNA Stool Mini Kit (Qiagen, Germany) was used to extract microbial DNA from fecal samples, following the manufacturer’s instructions. For PCR amplification of the bacterial 16S ribosomal RNA genes, primers 341F (5′-CCTACGGGRSGCAGCAG-3′) and 806R (5′-GGACTACVVGGGTATCTAATC-3′) were used to target the V3-V4 regions of the genes. Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, United States) according to the manufacturer’s instructions and quantified using Qubit®2.0 (Invitrogen, United States). All quantified amplicons were pooled to equalize concentrations for sequencing using Illumina NovaSeq (Illumina, Inc., CA, United States). Using cutadapt in Qiime2 (v2022.2) to remove primer sequences from the data. DNA extraction, Library construction, and sequencing were conducted at Realbio Genomics Institute (Shanghai, China).

The DADA2 plug-in in Qiime2 (v2022.2) software was used to filter, denoise, merged and nonchimeric the sequence after the removal of primes, and the Amplicon Sequence Variant or Feature sequence (ASV) was formed. The Feature representative sequences were compared with the Silva database (v138) to obtain the species classification of each feature, and then subsequent analysis was conducted according to the species abundance table.

Alpha-diversity metrics, such as Chao1 richness estimator, observed species, Faith’s PD, and Good’s coverage were calculated using the ASV table in QIIME2, and visualized as box plots. Beta diversity analysis was performed to investigate the structural variation of microbial communities across samples using Jaccard metrics, Bray–Curtis metrics, and UniFrac distance metrics, and visualized via principal coordinate analysis, non-metric multidimensional scaling, and unweighted pair-group method with arithmetic means hierarchical clustering. We employed linear discriminant analysis (LDA) of effect size (LEfSe) to identify significantly different taxa between the two groups (the default screening condition was LDA ≥ 2). Microbial functions were predicted by PICRUSt2 upon MetaCyc and KEGG databases.

Male Wistar rats were purchased from Yangzhou University Animal Medicine Center (SCXK 2022-0009). Rats were randomly divided into three groups (five rats per group): Control, CCl₄, and CCl₄+ Bacteroides group. Animals in all groups were provided a standard commercial rodent diet and were maintained on a 12-h light/dark cycle at constant temperature and humidity. All experimental procedures were approved by the Ethical Committee of Sichuan Provincial People’s Hospital. In addition, the National Institutes of Health guide for the care and use of laboratory animals was strictly followed.

Rats in the CCl₄ group or CCl₄+ Bacteroides group were gavaged with CCl₄ dissolved (MACKLIN, C805329) in olive oil (40% v/v, 3 mL/kg) twice per week for 12 weeks. During this 12-week treatment course, rats in the CCl₄+ Bacteroides group were gavaged with 1*10⁶ pieces/mL of bacterial (BeNa Culture, BNCC336948) solution (5 mL/day). At weeks 8, 9, 10, 11, and 12, portal blood flow and thrombosis were detected in rats using small animal ultrasound. The animals were euthanized at 12 weeks under deep anesthesia by intraperitoneal injection with an overdose of pentobarbital sodium (200 mg/kg) to obtain their samples for biochemical and molecular analyses.

For bioinformatics statistical analysis, an independent sample t-test was used to analyze normally distributed variables. Non-normally distributed variables were analyzed using the Mann–Whitney U test. A value of p < 0.05 was considered statistically significant. The α-diversity index analysis was performed using Qiime 2 software. Each index of Alpha diversity was analyzed by rank sum test. If two sets of samples are compared, wilcox.test function in R is used, and if more than two sets of samples are compared, the kruskal.test function in R is used.

For clinical data analysis, the statistical analysis was conducted with SPSS software. Continuous variables that exhibited normal distribution were presented as mean ± SD, and the t-test was utilized to compare the differences. We utilized nonparametric tests to compare continuous variables that were not normally distributed. Categorical variables were represented using percentages and integers, and groups were compared using Fisher’s exact test. Statistical significance was determined at a p value of <0.05.

Twelve patients with liver cirrhosis with PVT and 21 patients without PVT as controls were included in this cross-sectional study. The demographic characteristics and clinical parameters of the participants are presented in Table 1. The LC with PVT group had higher levels of platelet and D-Dimer compared to the control group. However, there were no statistically significant differences in body mass index, sex, other laboratory indicators, primary etiology of cirrhosis, or Child-Pugh grade between the two groups (Table 1).

Initially, we examined the species composition and abundance of gut microorganisms in both groups at both the phylum and genus levels. Firmicutes, Proteobacteria, Actinobacteria, and Bacteroidetes collectively accounted for >95% of the relative abundance in both groups. Compared to the control group, the abundances of Firmicutes (66.8 vs. 66.0%) and Proteobacteria (16.9 vs. 14.8%) in the LC with PVT group were relatively low, whereas the abundances of Actinobacteria (9.0 vs. 10.0%) and Bacteroidetes (6.1 vs. 2.9%) were higher (Figure 1A). At the genus level, the five most abundant bacterial taxa in the LC group were Escherichia-Shigella (14.9%), Blautia (13.4%), Streptococcus (10.7%), Lactobacillus (7.9%), and Enterococcus (7.2%). However, in the LC with PVT group, Blautia (20.2%), Lactobacillus (11.4%), Streptococcus (11.2%), Escherichia-Shigella (8.9%), and Bifidobacterium (8.0%) was detected (Figure 1B). When comparing the alpha diversity of the gut microbiota between the two groups, we found that the Chao1, observed species, and PD diversity indices were significantly higher in the group with LC and PVT than in the control group. The PCoA of beta diversity based on unweighted UniFrac revealed significant differences in community composition between the two groups (Figures 2A–D).

To identify differences in gut microbiota between the two groups, LDA and LEfSe analyses were conducted. Using a logarithmic LDA score with a cutoff value of 2.0, we identified important taxonomic differences and presented the top 30 differential gut microbiota. Our results suggest that there was a remarkable difference in gut microbiota between the two groups. The highest relative abundances in LC with PVT were for Akkermansia, Eubacterium hallii group, Fusicatenibacter, Anaerostipes, and Pseudomonas. In contrast, Enterococcus, Weissella, Bacteroides, Subdoligranulum, Lactococcus, Collinsella, RF39, and Hungatella were enriched in LC controls (Figures 3A,B).

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were analyzed using linear discriminant analysis (LDA) effect size (LEfSe) to predict the potential association between PVT and gut microbial metabolism. There were significant differences in gut microbial metabolism between the two groups; the top five candidate metabolic pathways in the LC group were other glycan degradation, pentose phosphate pathway, pentose and glucuronate interconversions, glycosaminoglycan degradation, and glycolysis/gluconeogenesis. The top five metabolic pathways in the LC with PVT group were bisphenol degradation, limonene and pinene degradation, linoleic acid metabolism, toluene degradation, and atrazine degradation (Figure 4A). We further analyzed the associations between the microbes and the derived metabolites (Figure 4B).

We analyzed the potential correlations between differentially abundant gut microbiota and coagulation function parameters (PLT, PT, INR, FIB, and D-Dimer). PLT and D-Dimer were correlated with a partial abundant of gut microbiota. Vibrio, Fusicatenibacter, Halomonas, Ralstonia, Thiomicrorhabdus, Peptostreptococcus, Guyparkeria, and Subgroup-22 significantly and actively correlated with D-Dimer levels, whereas Bacteroides, Adlercreutzia, Incertae-Sedis, Weissella, CAG-352, Lactococcus, Kurthia, and Pediococcus negatively correlated. Monoglobus, Bacteroides, and Adlercreutzia were significantly and negatively correlated with PLT (Figure 5).

The Bacteroides, a type of intestinal probiotic, is widely considered as source of novel beneficial candidates for attenuating inflammation (Tan et al., 2019; Guo et al., 2021). In our results, Bacteroides was significantly decreased in the PVT group, so we explored its effect on PVT by supplementing with Bacteroides. Initially, we successfully induced a cirrhosis model using CCL₄ and confirmed PVT formation using ultrasound. Supplementation with Bacteroides significantly reduced PVT volume as well as increased blood flow (Figures 6A,B). We examined fibrinogen, D-Dimer, and P-Selectin in plasma by ELISA and showed that they were significantly elevated in the CCL₄ group and decreased after Bacteroides treatment (Figures 6C-E). These data indicate that Bacteroides supplementation ameliorates CCl₄-induced PVT in Wistar rats.