The intestinal microbial community dissimilarity in hepatitis B virus-related liver cirrhosis patients with and without at alcohol consumption

Background Chronic hepatitis B virus (HBV) infection-reduced liver functions are associated with intestinal microbial community dissimilarity. This study aimed to investigate the microbial community dissimilarity in patients with different grades of HBV-related liver cirrhosis. Results Serum endotoxin was increased with Child–Pugh (CP) class (A, B, and C). Veillonellaceae and Lachnospiraceae families were reduced in patients compared with controls. Megamonas and Veillonella genus was reduced and increased in patients compared with controls, respectively, especially in CPB and CPC groups. Correlation analysis showed that endotoxin content was significantly correlated with alcohol consumption (95% CI 0.100, 0.493), CP class (95% CI 0.289, 0.687) and Lachnospiraceae family level (95% CI − 0.539, − 0.122). Firmicutes/Bacteroidetes ratio was correlated with the level of Lachnospiraceae family (95% CI 0.013, 0.481), Veillonellaceae family (95% CI 0.284, 0.696), Megamonas genus (95% CI 0.101, 0.518) and Veillonella genus (95% CI 0.134, 0.545). All aforementioned bacteria were independent risk or protective factors for hepatitis. Alcohol consumption changed microbial community. Conclusions Our study demonstrated that elevated Firmicutes/Bacteroidetes ratio, reduced Megamonas genus level and increased Veillonella genus level were indicators for HBV-related liver cirrhosis. Alcohol-related pathogenesis was associated with the changed microbial community.

chronic hepatitis B or HBV-related liver cirrhosis and healthy controls have been widely reported [4,6]. With comparison to healthy controls, patients with chronic hepatitis B showed abundant Anaerostipes taxon [7]. Liu et al. reported that patients with HBV-related HCC had reduced levels of pro-inflammatory bacteria like Escherichia-Shigella and increased level of anti-inflammatory bacteria like Faecalibacterium, while patients with non-HBV and non-hepatitis C virus (HCV) HCC showed opposite results [8]. Additionally, the altered abundance of gut microbial community has been identified in nonalcoholic fatty liver disease (NAFLD) and is deemed to be associated with progression into non-alcoholic steatohepatitis (NASH) [5,9]. However, alcohol consumption itself induces injury and inflammation in the intestine and liver, and dissimilarity in microbial community [10,11]. For instance, Stearns et al. showed that alcohol consumption increased the ratio of Firmicutes to Bacteroidetes [11], which is used as a parameter of obese [12,13], lipid metabolism [14] and insulin resistance [15]. These observations showed gut microbiota dysbiosis was relevant to HBV infection and might benefit in the progression to severe liver failure like HCC and liver disease. However, the association of microbial community dissimilarity with patients with HBV-related liver cirrhosis was unclear till now.
We performed this study to investigate the microbial community dissimilarity in healthy and patients with different grades of HBV-related liver cirrhosis. The present study included a study cohort of 80 patients with HBVrelated liver cirrhosis at Child-Pugh class A, B, C and 20 healthy controls without known liver diseases. The microbial community diversity in different groups were analyzed and compared. And alcohol history of each patient was asked for the analysis of alcohol consumption-related factors. The bacteria associated with hepatitis were identified. According to the reported relation of Firmicutes/Bacteroidetes ratio to diseases [15], we also performed logistics analysis to identify the correlation of this ratio with disease progression.

Patient population
Before experiments, an ethical approval and consent to participate (LDYYLL2018-142) was obtained from the ethical committees of First Hospital of Lanzhou University, Lanzhou, China. This study enrolled 80 patients (han Chinese) with HBV-related cirrhosis (including 55 males and 25 females, aged 42.30 ± 13.15 years) admitted in our hospital from July 2018 to Dec 2018. All patients with HBV-related cirrhosis were staged according to the Child-Pugh (CP) class A (CPA, n = 30), B (CPB, n = 31) and C (CPC, n = 19). All patients did not receive probiotics and antibiotics for 8 weeks prior to admission. Patients with solid organ transplantation, HCC, drug-induced liver injury, autoimmune liver disease and alcoholic fatty liver were excluded. In addition, patients with specific food habits, like vegetarians and lactose lovers, were excluded from our study. Twenty healthy volunteers (age-and sex-matched) without known diseases were randomly selected from the physical examination center of our hospital. Use of antibiotics within 8 weeks prior sample collection was forbidden. Alcohol history and consumption (white liquor) of all patients was asked. Fasting blood and stool samples were collected from all participants with informed consent. All samples were set at − 80 °C before examination. This study was performed according to the Declaration of Helsinki (1975).

Blood biochemical indicator detection
The serum diamine oxidase (DAO), d-lactate, endotoxin and HBV-DNA was detected using highly sensitive human DAO ELISA kit, lactic acid and bacterial endotoxin assay kits (JY-Po-Color DLT Set; Zhongsheng jinyu diagnostic technology co. LTD., Beijing, China), respectively. Three duplicates were setup for each sample.

DNA extraction and sample preparation
Total DNA samples were extracted from stool samples using MoBioPowersoil DNA extraction kits (MoBio, Carlsbad, CA, USA) following the manufacturer's instructions. DNA quality was determined using 1% agarose gel electrophoresis. Amplification of the V4 region of the 16S rRNA gene was conducted using the 515 (Forward, 5′-GTG CCA GCMGCC GCG GTAA-3′)/806 (Reverse, 5′-GGA CTA CHVGGG TWT CTAAT-3′) primers and TransGen AP221-02: TransStartFastpfu DNA Polymerase (TransGen Biotech, Beijing, China). Amplification of 20 µL reactions was implemented on Applied Biosystems GeneAmp ® 9700 (Applied Biosystems, Foster City, CA, USA) at 94 °C for 4 min, followed by 30 cycles of 94 °C for 40 s, 58 °C for 30 s, and 72 °C for 45 s, and followed by a final extension at 72 °C for 10 min. Three duplicates were set up for each sample. PCR products from one sample were pooled and then gel purified (2% agarose gel) using an AxyPrep DNA Gel Extraction kit (Axygen Biosciences, Hangzhou, China). DNA quantification was performed using QuantiFluor ™ (Promega, Lyon, France). DNA library construction was carried out using the TruSeq ® DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, USA) and 16S rRNA gene sequencing was conducted on the Illumina MiSeq platform (Illumina, San Diego, USA; 2 × 250 bp PE).

Statistical analyses
All clinical data were expressed as mean ± standard deviation or median and range (Q1-Q3) and were analyzed using SPSS 22.0. Comparisons in blood biochemical indicators were performed using non-parametric Mann-Whitney U test, Kruskal-Wallis H test or χ 2 test. Differences between alcohol and non-alcohol consumed patients were analyzed using Mann-Whitney U test. For gut microbiota analysis, non-parametric Kruskal-Wallis H test and Wilcoxon rank-sum test was used for the differences in alpha and beta diversities. Differences in taxonomies were analyzed using non-parametric Kruskal-Wallis sum-rank test and LDA Effect Size (LEfSe) software (http://hutte nhowe r.sph.harva rd.edu/ galax y/root?tool_id=lefse _uploa d). Differences in the relative abundances of OTUs of dominant bacteria were analyzed using Welch's t test or Mann-Whitney U test. Spearman correlation coefficients (r) between dominant bacteria and blood biochemical indicators were identified. Logistic regression analysis was performed to identify the indicators for disease progression. p < 0.05 was considered statistically significant.

Differences in serum indicators
There was no difference in the age, male ratio, and alcohol consumption frequency among the three groups ( Table 1) Table 1. Serum DAO and d-lactate contents were equivalent in patients with CP class A (n = 30), B (n = 31) and C (n = 19) HBVrelated cirrhosis. Serum endotoxin and HBV-DNA contents were increased in patients with higher CP classes (p = 0.000 and p = 0.000, respectively; Table 1). Alcohol consumption increased the serum endotoxin in all patients (p = 0.006, Table 2); endotoxin and d-lactate contents in CPA patients (p = 0.008 and 0.017, respectively); but decreased DAO content in CPC patients (p = 0.035).

Sequencing depth estimation
Illumina sequencing generated 4,661,253 sequences (ranging 251-300 bp), with an average length of 255.76 bp. These sequences represented 4414 OTUs with 97% identity after data processing. Rarefaction curves and Shannon-Wiener curves reached plateaus, indicating sufficient sequencing depths for microbial diversity analysis ( Fig. 1a, b). Venn diagram showed there were 1724 overlapped OTUs (Fig. 1c). PCoA based on the Bray-Curtis distance showed there was no distinct clustering of species groups (Fig. 1d).

Microbial community dissimilarity in patients
Microbial diversity analysis showed that the relative abundances of Firmicutes and Bacteroidetes phylum in patients were equivalent with that in control (Fig. 2a) Fig. 3a, b).

Microbial communities in patients with different CP classes
We further performed the pairwise comparison between patients with different CP classes and controls in order to investigate the microbial community dissimilarity in patients with different disease severities. Microbial  community analysis showed there were no differences in the relative abundances of bacteria at the phylum (Fig. 4a) and family level (Fig. 4b, c).  Fig. 4d, e). Species taxonomy dissimilarity analysis showed uncultured bacteria of Faecalibacterium, Megamonas and Bacteroides were dominant bacteria in all groups (Fig. 5a, b). The relative abundance of an uncultured species of Megamonas in patients with CPB (1.61 ± 0.74%) and CPC (5.13 ± 1.82%) disease was dramatically lower than that in controls (14.77 ± 3.39%, p < 0.0001) and patients with CPA disease (11.93 ± 2.92%, p < 0.0001, Fig. 5b). LEfSe analysis showed that Mollicutes family (Tenericutes phylum), Lachnospiraceae family, Micrococcales family (Actinobacteria phylum) and Pasteurellales (Proteobacteria phylum) were important bacteria in CPA, CPB, CPC and control groups (Fig. 5c, d).

Microbial community dissimilarity by alcohol consumption
In order to investigate the influence of alcohol consumption on microbial community dissimilarity, patients were divided into two groups: with and without alcohol consumption. There was no microbial community dissimilarity in the relative abundance of dominant phyla (data not shown), families (Fig. 6a) and genera (Fig. 6b) between the two groups. In patients within the same CP class, alcohol consumption induced significant microbial community dissimilarity at the family and genus levels (Fig. 6c, Fig. 6c). At the genus level, we found alcohol consumption increased Bacteroides (CPA and CPC) and Megamonas (CPA) and Veillonella (CPB) level, respectively (Fig. 6d). These differences demonstrated that alcohol consumption caused microbial community dissimilarity in patients with HBVrelated cirrhosis.

Discussion
Our present study demonstrated that intestinal microbial community diversity in patients with HBV-related liver cirrhosis was different from that in healthy controls. The intestinal microbial community dissimilarity, like reduced Megamonas genus (Veillonellaceae family) level, Lachnospiraceae family level as well as increased Veillonella genus (Veillonellaceae family) level, was associated with CP classes in patients. In addition, we confirmed that the Firmicutes/Bacteroidetes ratio was positively correlated with the levels of above-mentioned bacteria. Additionally, alcohol consumption did not obviously altered the levels of these bacteria and serum parameters including d-lactate, endotoxin and DAO contents in patients with HBV-related liver cirrhosis. The Firmicutes/Bacteroidetes ratio was reported to be associated with the morbidity and pathogenesis of several chronic diseases like obesity or diabetes [13][14][15]. Obese population, both children and adults, had elevated Firmicutes/Bacteroidetes ratio compared with controls with normal weight [13,16]. On contrast, decreased Firmicutes/Bacteroidetes ratio in obese individuals was identified to be related with weight loss [17]. Cui et al. showed that the elevated the Firmicutes/ Bacteroidetes ratio in pigs was correlated with reduced fatty acid synthase (FASN) and acetyl-CoA carboxylase α (ACCα) mRNA levels in liver, but opposed trends in subcutaneous fat. The inhibition of ACCα suppresses FASN and thus inhibits the synthesis of fatty acid from free fatty acids (FFAs) like saturated straight-chain FAs [18]. Decreased FASN level was detected in adipose tissue from hypertensive individuals and subcutaneous adipose tissue from obese subjects [18][19][20]. What's more, FASN has a negative correlation with insulin resistance [21]. Consistently, our present study suggested that the increased Firmicutes/Bacteroidetes ratio was an independent risk facor for HBV-related liver cirrhosis, indicating the alterations in lipid metabolism and FAs biosynthesis in patients.
Megamonas and Veillonella genus were identified to be significantly contributed to the pathogenesis of liver disease. Both Megamonas and Veillonella genera are poorly characterized members of Veillonellaceae family, Firmicutes phylum. Veillonella genus is anaerobic genus [22]. However, both Megamonas and Veillonella are frequently elevated in cystic fibrosis cases [23], individuals with type I diabetes [22] and patients with primary sclerosing cholangitis with or without inflammatory bowel disease (IBD) compared with healthy controls [24,25]. Veillonella genus is a producer of butyrate and shortchain fatty acids (SCFAs) [26] which contributes to antiinflammatory response in the host [27]. The elevated and decreased level of Megamonas was identified in the host with IBD [24] and Behcet's disease [28], respectively. Chakravarthy et al. detected the reduced Veillonella dispar level in patients with uveitis, an inflammatory disease of the eye [29]. By contrast, Matera et al. reported the increased level of Veillonella in the mouth flora in individuals with periodontal disease [30]. Additional in vitro evidence suggested that Veillonella LPS stimulated the release of cytokines like interleukin-6 (IL-6), IL-1β, IL-10 and tumor necrosis factor alpha (TNF-α) in human peripheral blood mononuclear cells [30]. These results suggested that intestinal Veillonella metabolites like LPS and Veillonella-modulated metabolites like SCFAs affect the liver pathology and inflammation in the host [9]. We found the level of Veillonella and Megamonas was positively and negatively correlated with Child-Pugh classes in patient with HBV-related liver cirrhosis, respectively. The elevated Veillonella genus and decreased Megamonas genus might suggest the unbalanced inflammatory status in patients with HBV-related liver cirrhosis compared with control.
Several observations suggested that alcohol consumption itself not only induces injury and inflammation in the intestine and liver but also causes dissimilarity in microbial community [10,11]. Our present study confirmed that there was no dissimilarity in microbial community between patients with and without alcohol consumption. But differences were observed in microbial community between healthy controls and patients with or without alcohol consumption. This suggested the influence of alcohol consumption on the pathogenesis of disease.  Alcohol consumption increases intestinal permeability and therefore induces endotoxaemia [31,32]. Some researchers showed the insignificantly influence of alcohol consumption on intestinal permeability and endotoxin [33]. Our correlation analysis showed that alcohol consumption was positively correlated with increased endotoxin level. Alcohol consumption upregulated endotoxin and d-lactate levels in patients with Child-Pugh class A disease, and DAO in patients with Child-Pugh class C disease. However, alcohol consumption showed non-benefit to endotoxin in patients with Child-Pugh class B and C diseases. Further regression analysis showed alcohol consumption was not an risk factor for endotoxin and aforementioned intestinal bacteria in patients with HBV-related liver cirrhosis. The increased endotoxin, d-lactate and DAO in patients with different Child-Pugh classes by alcohol consumption might due to the increased intestinal permeability. These data might reveal that alcohol consumption played an important inducer role in the pathogenesis of but a faint function the progression of HBV-related liver cirrhosis.

Conclusions
Our present study demonstrated that the intestinal microbial community dissimilarity in patients with HBVrelated liver cirrhosis compared with healthy controls. Elevated Firmicutes/Bacteroidetes ratio, reduced Megamonas genus (Veillonellaceae family) level, increased Veillonella genus (Veillonellaceae family) level was associated with Child-Pugh classes in patients with HBVrelated liver cirrhosis. We additionally identified that alcohol consumption play important roles in the pathogenesis of HBV-related liver cirrhosis.