Alteration in tongue coating microbiota across different stages of hepatitis B virus-related chronic liver disease

Abstract

Background

Hepatitis B virus-related chronic liver disease (HBV–CLD), chronic hepatitis B (CHB), liver cirrhosis (LC), and hepatocellular carcinoma (HCC) patients' tongue coating microbiota dysbiosis has not yet been clearly defined.

Objectives

We aimed to reveal shifts in the bacterial composition of tongue coating microbiota during the progression of HBV–CLD in three different phases.

Design

We examined tongue coating microbiota of 16 healthy individuals and 81 patients with HBV–CLD, including 25 with CHB, 27 with LC, and 29 with HCC, using 16S rRNA gene sequencing technique.

Results

The bacterial richness in tongue coating was higher in patients with HBV–CLD (all P < 0.05) than in healthy controls. A clear clustering pattern between patients with HBV–CLD and healthy controls is shown using beta diversity analysis (all p < 0.05). Linear discriminant analysis revealed multiple taxa that varied significantly in abundance between healthy controls and patients with HBV–CLD; Firmicutes were higher in patients with LC and HCC, whereas CHB patients had higher levels of Bacteroidetes. PICRUSt2 analysis of the sequencing data revealed changes in microbial activity with disease development.

Conclusions

Our investigation revealed tongue coating microbiota dysbiosis in patients with HBV–CLD, which may offer unique diagnostic possibilities and provide microbial biomarkers for monitoring disease progression.

Introduction

It is estimated that the hepatitis B virus (HBV) chronically infects 296 million people globally, with nearly 1.5 million new cases annually. Patients with chronic hepatitis B (CHB) have an increased risk of liver cirrhosis (LC) and hepatocellular carcinoma (HCC), which together account for around one million fatalities worldwide every year [1]. The oral-gut‒liver axis has revealed a connection between the oral microbiome, gut microbiome, and liver health [2]. The oral microbiota is the community of bacteria that inhabits the oral cavity; it is the second largest microbial community in the human body, following the gut [3,4]. Typically, it appears on hard and soft tissue as a biofilm in the mouth cavity [5,6]. The core human microbiomes are universally shared among all people; others are unique microbial profiles that vary by individual lifestyle and physical attributes [7]. These naturally occurring bacteria do not harm and serve as a barrier to pathogenic species from sticking to the oral mucosa [8,9]. It is only when bacteria cross the commensal border that they become infectious and cause disease [10]. The oral cavity is home to the most varied bacteria, which are mainly from the phyla Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria, and Fusobacteria [11]. Some of these bacteria are present in oral biofilms, such as the tongue coat and dental plaque, or they can be freely found in oral fluid [12]. Furthermore, there are obvious variations in microorganisms throughout different locations in the oral cavity [13]. The tongue serves as a reflection of the body, and the microbes within the biofilm are largely stable there [14]. Monitoring the tongue is a primary diagnostic technique used in traditional Chinese medicine (TCM). The size, appearance, colour, and texture of the tongue's body can provide information about the performance of organs and the course of illnesses [15]. Physicians are increasingly adopting oral microbiome analysis to identify systemic illness and track overall health [16]. Long-term HBV infection is linked to immune system dysfunction, which may contribute to an imbalance of bacteria in the mouth and oral discomfort [17]. Enrichment of certain oral bacteria in Hepatitis B virus-related chronic liver disease (HBV–CLD) patients can act as opportunistic pathogens and produce metabolites that may exacerbate chronic inflammation and prolong HBV persistence [15]. Alterations in the oral microbiota of individuals suffering from liver disease are reflected in the gut microbiome [18]. According to the findings of previous investigations, several studies have examined dysbiosis in the microbial composition of liver disorders. Research on HBV–CLD patients has mostly examined microbiota in faeces and duodenal mucosa [19,20]. We hypothesise that dysbiosis in the tongue coating microbiota is linked to HBV infection, which leads to CHB, LC, and HCC [21]. As demonstrated by earlier research, the tongue coating microbiota is distinct from other microbial communities in the human body [22].

There has not been a direct comparison between the tongue coating microbiota of healthy individuals and HBV–CLD patients, the characteristics of the tongue coating microbiota of HBV–CLD patients have not yet been clearly defined. The current study's objectives are to analyse variations in the tongue coating microbiota of patients with HBV–CLD and healthy controls, to enhance patient outcomes. Using 16S rRNA sequencing, we examine the tongue coating microbiota in HBV–CLD patients to elucidate its relationship with the clinical phases of HBV–CLD, aiming to identify microbial biomarkers for high-risk individuals and promote novel noninvasive screening and diagnostic methods.

Methods

Subjects' enrolment

Participants were recruited from March 2023 to April 2024 at the Second Affiliated Hospital of Lanzhou University to investigate the tongue coating microbiota of patients associated with HBV-CLD (CHB, LC, and HCC) compared with healthy individuals who enrolled during their annual medical screenings in hospital. Chronic HBV infection verification in an individual based on the presence of hepatitis B surface antigen (HBsAg) for a minimum of six months. The identification of LC was based on the histopathological findings or identification of nodules on the liver epidermis, a rough liver parenchymal appearance, and constricted vessels with uneven intrahepatic vascular outlines on ultrasound scanning [23]. A standard dynamic imaging examination or pathological findings were used to diagnose HCC, under the guidelines released by the American Association for the Examination of Liver Diseases [24]. Any participant who met any of the following criteria was excluded from the study: (i) individuals aged <35 or >65 years who had antibiotics, probiotics, and other immunosuppressive medications within 6 months before enrollment; (ii) patients who suffered from other liver illnesses. (e.g. intrahepatic cholestasis, hepatitis C, steatohepatitis, and nonalcoholic fatty liver disease.); (iii) patients had a tumor beyond stage II (clinical Staging System on TNM Classification for HCC); (iv) patients who suffered from periodontitis, dental cavities, canker sores, and other oral illnesses; (v) current smoking, or heavy alcohol consumption (>20 g/day for males, >10 g/day for women).

Laboratory measurements

At enrollment, basic metabolic indicators were measured, which include high-density lipoprotein (HDL), low-density lipoprotein (LDL), alanine aminotransferase (ALT), aspartate aminotransferase (AST), triacylglycerol test (TG), blood urea nitrogen test (BUN), serum creatinine test (SCr), fasting blood glucose (FBS), platelet count, total cholesterol (T-CHO). The serum levels of HBeAg, anti-HBe, anti-HBs, anti-HCV, and anti-HDV were measured using standard kits. Automated approaches with a commercial standard were used to perform laboratory experiments at the NHC Key Laboratory of Diagnosis and Therapy of Gastrointestinal Tumour, Lanzhou, Gansu Province, China.

Tongue coating samples collection and DNA extraction

Before sampling, the participants were instructed to use sterilised water to gargle twice and stretch out their tongues as far as possible. According to Winkel, the tongue was split into six functional parts [25]. A qualified stomatologist collected the tongue coat from the distal middle area to the anterior middle area of the tongue by using a tongue scraper. The coating samples were suspended in phosphate-buffered saline (PBS), moved to the lab, agitated, and centrifuged promptly; the supernatant was discarded. The pellets were kept at −80 °C for further processing.

Bacterial DNA was extracted from the tongue coating samples using a DNeasy Power Soil kit (Qiagen, Hilden, Germany) according to the manufacturer's guidelines. The concentration and purity of the DNA were determined using agarose gel electrophoresis and a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

16S rRNA gene amplification and sequencing

PCR amplified the V3-V4 hypervariable sections of the bacterial 16S rRNA gene in a 25  μl reaction using universal primer pairs (343F: 5′-TACGGRAGGCAGCAG−3′ and 798 R: 5′-AGGGTATCTAATCCT−3′), [26]. Illumina sequencing adapters were added to the primers. Gel electrophoresis was employed to assess the amplicon purity. Agencourt AMPure XP beads (Beckman Coulter Co., USA) were used to purify the PCR products followed by quantification with a Qubit dsDNA assay kit. The concentrations of pooled libraries were modified and sequenced on the Illumina NovaSeq6000 using 250 bp paired-end reads.

Bioinformatics and statistical analysis

The FASTQ format bioinformatics analysis was utilised for raw sequencing data. After that, the cutadapt software was used to preprocess the paired end reads to identify and clip off the adaptors. Shortly after trimming, paired-end reads were filtered for low-quality sequences, denoised, merged, and detected cut-off chimera reads using DADA2 [27] with the default settings of QIIME2 [28] (2020.11). Finally, the QIIME2 programme was used to pick the corresponding reads of each ASV. Using a q2-feature classifier with the default settings, all sample reads were noted and blasted against the Silva database (Version 138). The Alpha and beta diversity were analysed with QIIME2 software. Alpha diversity parameters, the Chao1 index [29], and the Shannon diversity [30] were assessed to estimate species richness in the samples. The R tool uses principal coordinate analysis (PCoA) on a UniFrac distance matrix to determine beta diversity. The permutational multivariate analysis of variance (PERMANOVA) with 999 permutations was performed using the Adonis2 function, and the pairwise Adonis package in R was used to analyse the statistical significance of beta diversity. The R programme was employed to examine statistically significant variation between the groups utilising ANOVA/Kruskal–Wallis/T test/Wilcoxon statistical test. Every test was duplicated, and a p-value less than 0.05 denoted empirical significance. The taxonomy abundance range was compared using the linear discriminant analysis (LDA) effect size (LEfSe) approach. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States 2 (PICRUSt2) was used to predict microbial community functions based on the Kyoto Encyclopaedia of Genes and Genomes (KEGG) database.

Results

Clinical characteristics of subjects

In this investigation, we examined the tongue coating microbiota of 16 healthy individuals and 81 patients with HBV–CLD identified as having CHB, LC, and HCC (25, 27, and 29, respectively). Table 1 provides a summary of the participants' baseline information. For healthy controls, the average age was 46 years, 49 years for CHB patients, 57 years for LC patients, and 60 years for HCC patients; Accordingly, the percentage of males was 68.75, 60, 59.25, and 68.96%, respectively. Overall, patients with CHB, LC, and HCC disease had higher levels of ALT, AST, BUN, and creatinine test (CRE) compared to healthy controls (all p < 0.05). Platelet counts were decreased in patient groups compared to healthy individuals (p < 0.05).

Table 1.

Demographic information and clinical characteristics of the research population.

Parameters	Healthy	CHB	LC	HCC	Statistic	p-value
Number of subjects	16	25	27	29	NA	NA
Gender (M/F)	11/5	15/10	16/11	20/9	Pearson chi-square	0.634
Age (years)	46 ± 6	49 ± 8	57 ± 5	60 ± 5	Kruskal–Wallis test	0.079
BMI (kg/m2)	21.15 ± 3.22	21.06 ± 4.19	22.54 ± 4.06	22.70 ± 4.78	One-way ANOVA	0.934
LDL (mg/dL)	103.17 ± 25.07	118.24 ± 21.19	107.04 ± 23.09	94.86 ± 19.28	One-way ANOVA	0.628
HDL (mg/dL)	1.477 ± 0.428	0.999 ± 0.869	0.819 ± 0.821	1.076 ± 0.77	Kruskal–Wallis test	0.695
TG (mg/dL)	130.13 ± 15.32	141.27 ± 9.08	105.11 ± 20.17	102.29 ± 14.02	Kruskal–Wallis test	<0.05
AST (U/L)	16.27 ± 3.05	28.15 ± 6.23	35.19 ± 8.05	37.03 ± 11.10	Kruskal–Wallis test	<0.05
ALT (U/L)	9.03 ± 3.12	47.27 ± 28.09	31.56 ± 14.12	30.11 ± 12.16	Kruskal–Wallis test	<0.05
Prothrombin time	11.25 ± 2.16	12.48 ± 1.21	14.49 ± 1.51	16.87 ± 2.99	Kruskal–Wallis test	<0.05
Platelet count (109/L)	231.10 ± 15.07	169.13 ± 34.11	121.21 ± 20.13	136.11 ± 17.03	Kruskal–Wallis test	<0.05
CRE (mg/dL)	0.5 ± 0.11	0.9 ± 0.20	1.1 ± 0.31	1.3 ± 0.42	Kruskal–Wallis test	<0.05
Bun (mg/dL)	10.23 ± 0.90	13.43 ± 0.61	14.43 ± 0.31	16.06 ± 0.62	Kruskal–Wallis test	<0.0001
GLU-AC (mg/dL)	74.85 ± 8.42	92.43 ± 7.21	112.68 ± 19.43	117.38 ± 26.32	Kruskal–Wallis test	<0.05
T-CHO (mg/dL)	151.21 ± 30.12	196.41 ± 31.21	136.6 ± 29.31	106.56 ± 25.72	Kruskal–Wallis test	<0.05

The mean ± standard deviation was used to describe continuous data variables. Pearson's chi-square test the difference in Gender. BMI and LDL were tested using a one-way analysis of variance (ANOVA).

Abbreviations: BMI, body mass index; LDL, low‐density lipoprotein; HDL, high‐density lipoprotein; TG, triglycerides; AST, aspartate aminotransferase; ALT, alanine aminotransferase; CRE, creatinine; BUN, blood urea nitrogen; GLU-AC, fasting blood glucose; T-CHO, total cholesterol.

Overview of the 16S rRNA sequencing data output

The amplicon sequencing of the 16S ribosomal RNA (rRNA) gene V3–V4 generated 7,761,550 reads; the average number of sequence reads per specimen was 80,016 (minimum: 78,020; maximum: 81,950). To consider sample size-related variability, rarefaction was used to standardise the reads to 75,925 per sample. After rarefaction, an overall 4,100 amplicon sequence variants (ASVs) were obtained.

Differences in tongue coating microbiome structure

The shared and distinct ASVs seen in the different phases of HBV–CLD, as well as the total number of identified ASVs, are highlighted in Figure 1a. Alpha diversity indices were computed according to rarefied ASVs. In comparison to healthy controls, patients with CHB had higher levels of bacterial richness and evenness according to the calculated values of the Shannon diversity index Figure 1c. The Observed species, Chao1, and ACE indices showed that patients with LC and HCC had a higher degree of species richness than healthy controls Figure 1b, d, and e. The predicted values of PD_whole_tree showed higher levels of bacterial richness in patients with LC when compared to healthy controls, Figure 1f (all p < 0.05).

Figure 1.

Comparative analysis of bacterial diversity and richness between healthy controls and the three patient groups: CHB, LC, and HCC. (a) A Venn diagram illustrates shared and unique ASVs among the four groups. (b) Total observed species, with the lowest observed species in healthy controls. (c–f) Alpha diversity indices: (c) Shannon index, (d) Chao1 index, (e) ACE index, and (f) PD_whole_tree index, depicting variations in richness and evenness between the groups; statistical significance is indicated by the Kruskal‒Wallis test *p < 0.05 and **p < 0.01.

The PCoA weighted UniFrac, unweighted UniFrac, Bray-Curtis distance, and Binary Jaccard showed separate clusters for healthy controls and three patient groups: CHB, LC, and HCC Figure 2a−d. These clusters were distinguished by the significant influence of disease status (all p < 0.05). Pairwise comparisons were performed using PERMANOVA, which revealed that the tongue coating microbiota of three patient groups significantly differed from that of healthy controls.

Figure 2.

Beta diversity patterns in tongue coating microbiota. PCoA of bacterial beta diversity in healthy controls and patients with CHB, LC, and HCC groups based on (a) weighted UniFrac, (b) unweighted UniFrac, (c) Bray–Curtis distances, and (d) Binary–Jaccard, revealing distinct clustering patterns between healthy controls and three patient groups.

Taxonomic distribution and bacterial abundance

The relative abundance of each taxon was used to describe the distribution of bacteria in HBV–CLD. In total, 84 genera, 49 families, 29 orders, 13 classes, and 10 phyla were found in the tongue coating samples. Figure 3 displays the taxonomic distributions of significantly abundant bacteria at various levels.

Figure 3.

Comparison of the tongue coating microbiome at the family (a) and genus (b) levels. Box plots represent the relative abundance of tongue coating bacteria in healthy controls and patients with CHB, LC, and HCC. The upper and lower ranges of the box represent the 75th and 25th percentiles. Statistical comparisons were performed using the Wilcoxon rank-sum test; significant differences are indicated by *p < 0.05, **p < 0.01, and ***p < 0.001.

The phylum-level screening of tongue coating microbiota is shown in Figure S1. The six most abundant phyla were Bacteroidetes, Proteobacteria, Firmicutes, Fusobacteria, Actinobacteria, and Patescibacteria; these six major taxa collectively made up most of the sequences. Compared to healthy controls, the LC and HCC patient groups had considerably higher relative abundances of Actinobacteria and Firmicutes, while Bacteroidetes were more abundant in CHB patients, and Fusobacteria had a lower relative abundance in all three patient groups.

At the family level, Fusobacteriaceae and Pasteurellaceae were more prevalent in the tongue coating microbiome of healthy controls. Streptococcaceae and Micrococcaceae were more abundant in the tongue coating microbiome of LC and HCC patients. The Prevotellaceae were more abundant in CHB patients Figure 3a.

At the genus level, out of 10 discriminatory genera, the relative abundance of Rothia and Streptococcus was higher in the tongue coating microbiome of the LC and HCC patient groups. Fusobacterium and Haemophilus were higher in the tongue coating microbiome of healthy controls Figure 3b. The genus level alteration of tongue coating bacteria in HBV–CLD patients is shown in a heatmap Figure S2.

Differences in tongue coating microbiome composition in HBV–CLD

To identify biomarkers and variations in tongue coating microbiota composition for each HBV–CLD group, the LEfSe method was employed, using the logarithmic LDA score to ascertain the primary taxonomic alterations between the clinical stages of HBV–CLD. Notable differences were found between healthy controls and patients with HBV–CLD Figure 4a–f. Table S1 summarises the LDA score and p value of the most prevalent genera in the LEfSe analysis.

Figure 4.

LEfSe analysis of 16S rRNA sequencing data to identify differentially abundant taxa in tongue coating microbiota of healthy controls and patients with HBV–CLD. Comparisons include: (a) CHB vs. Healthy, (b) LC vs. Healthy, (c) HCC vs. Healthy, (d) CHB vs. LC, (e) HCC vs. CHB, and (f) HCC vs. LC, highlighting distinct taxa with disease progression.

The comparison between patients with CHB and healthy controls, Figure 4a revealed that the most abundant genera in patients with CHB were Prevotella, Alloprevotella, and Actinomyces; whereas, the most prevalent genera in healthy controls were Haemophilus, Fusobacterium, and Porphyromonas.

Investigation between patients with LC and healthy controls, Figure 4b revealed that the most enriched genera in patients with LC were Streptococcus and Rothia, whereas the most abundant genera in healthy controls were Fusobacterium, Haemophilus, and Porphyromonas.

Assessment between patients with HCC and healthy controls, Figure 4c showed that the most prevalent genera in patients with HCC were Streptococcus, Actinomyces, Rothia, and Lactobacillus. Conversely, Haemophilus, Fusobacterium, and Porphyromonas were more abundant in healthy controls.

Comparability between individuals with CHB and LC, Figure 4d demonstrated that patients with CHB had an increased population of Prevotella and Alloprevotella, whereas the most enriched genera in patients with LC were Streptococcus and Rothia.

Evaluation between patients with CHB and HCC, Figure 4e revealed that individuals with CHB were enriched in Prevotella, Alloprevotella, and Haemophilus. Conversely, Streptococcus, Rothia, and Granulicatella were more prevalent in patients with HCC.

Lastly, an enquiry was made between patients with LC and HCC, Figure 4f showed that patients with LC had a higher prevalence of Neisseria and Ruminococcus, whereas the most abundant genera in patients with HCC were Pantoea, Faecalibaculum, and Parasutterella.

Functional analysis of tongue coating microbiota based on 16S rRNA sequencing data

We performed a comparison of Level 2 KEGG pathways to estimate the functional content of the tongue coating microbiota using PICRUSt 2 to illustrate changes in bacterial functions as the illness progresses. A total of 20 differentially abundant pathways were found among the HBV–CLD patient groups and healthy controls (FDR < 0.05), Figure 5a–f. The enriched pathways in HBV–CLD were glycan biosynthesis and metabolism, transport and catabolism, cell growth and death, signal transduction, cellular community-prokaryotes, cancers overview, bacterial infectious disease, endocrine system, xenobiotic biodegradation and metabolism. Ten Clusters of Orthologous Groups (COG) categories, including glycosyltransferase in cell wall biosynthesis, site-specific recombinase, outer membrane receptor proteins, transferase activity in cell wall synthesis, Na+ -driven multidrug efflux pump, DNA-binding transcriptional regulator, DNA-directed RNA polymerase, lipoprotein export system, pseudouridylate synthase, and DNA-binding response regulator, showed differences among the three patient groups and healthy controls Figure S3.

Figure 5.

PICRUSt2 analysis of 16S rRNA sequencing data predicts Level 2 KEGG functional categories in tongue coating microbiota of healthy controls and patients with HBV–CLD, illustrating the distribution of differentially expressed functional categories for: (a) CHB vs. healthy, (b) LC vs. healthy, (c) HCC vs. healthy, (d) CHB vs. LC, (e) HCC vs. CHB, and (f) HCC vs. LC, highlighting microbial functional shifts.

Discussion

The aetiology and pathogenesis of HBV–CLD are complex. Despite the availability of antiviral therapies, curing HBV infection remains challenging; therefore, encouraging the investigation of innovative approaches, like focusing on the oral microbiota in HBV–CLD, may provide new insights into the field. This study utilised 16S rRNA sequencing to characterise the diversity and composition of tongue coating microbiota in patients with CHB, LC, and HCC, comparing them to healthy controls to identify microbial dysbiosis patterns and potential microbial biomarkers.

In our investigation, the operational taxonomic unit (OTU) numbers and observed ASVs were higher in the three patient categories than in the healthy controls. The Chao1, observed species, PD_whole_tree, and ACE were elevated in the LC and HCC patient groups. Shannon diversity was higher in the CHB patient group compared to healthy controls. A large Chinese cohort study found no correlation between age and microbial alpha diversity, suggesting that, rather than the patient's baseline characteristics, the distinct illness states of the three patient groups may have contributed to the variations in tongue coating microbiota [20]. To minimise confounding, patients with substantial alcohol intake were specifically excluded from this critical investigation on tongue coating microbiota in patients with HBV–CLD, as alcohol can independently alter microbial composition [31,32].

PCoA analysis (weighted/unweighted UniFrac, Bray-Curtis, Binary Jaccard) demonstrated distinct clustering of HBV–CLD groups from healthy controls. Patients with HBV–CLD and healthy individuals have different microbiome composition qualitatively and quantitatively, as measured in terms of beta diversity. Our results agree with those of Zeng et al. [32]. In another enquiry beta diversity analyses performed using Bray-Curtis-based nonmetric multidimensional scaling also exhibited dysbiosis in microbiota composition in patients with CHB and LC [33].

At the phylum level, Patescibacteria, Proteobacteria, and Fusobacteriota were depleted in three patient groups; by contrast, the relative abundance of Bacteroidetes (Prevotella and Alloprevotella) was higher in CHB patients with the highest LDA score, followed by the phyla Firmicutes (Streptococcus, Granulicatella, and Lactobacillus) and Actinobacteria (Actinomyces and Rothia), which were higher in LC and HCC patients. Patescibacteria were identified by next-generation sequencing as commensals in different human habitats, including the oral cavity, gastrointestinal, and genital tracts [34]. Proteobacteria were reduced in the gut microbiota of viral hepatitis [35]. Fusobacteria are considered a nonpathogenic commensal phylum [36]. Patients with advanced liver disease had higher levels of Actinobacteria [20]. The predominant phyla Firmicutes and Bacteroidetes exhibit dysbiosis, which shows an association with HBV–CLD [20]. Most Bacteroidetes are gram-negative bacteria, which may be primarily responsible for carrying out metabolic processes, such as the breakdown of complex sugar polymers and proteins [37]. On the other hand, Firmicutes are gram-positive bacteria, probiotic organisms, and opportunistic pathogens that are representatives of this phylum [38,39]. Probiotics are examples of microbiota-based therapy that may help to understand the role of oral bacteria, which can assist in creating targeted treatments and restore a healthy oral microbiome by reducing the course of illness [40,41]. Oral hygiene and dental treatment procedures, including oral prophylaxis and scaling, can effectively reduce the bacterial load in the oral cavities of patients with liver disease [42–44]. This reduction may mitigate pathogenic translocation and systemic inflammation, potentially lowering the risk of disease progression.

It is reported that the oral-gut‒liver axis mediates the bidirectional microbial interactions, which are an essential part of the pathophysiology of CLD [2]. It facilitates the translocation of oral bacteria to the gut via saliva or systemic circulation, where they elevate LPS levels, cytokine production (TNF-α, IL-6), and systemic inflammation that may reshape the tongue coating microenvironment [2,17,18,42,43,45].

In our study, the genera Porphyromonas, Haemophilus, and Fusobacterium were depleted in patients with HBV–CLD. Whereas, Prevotella and Alloprevotella were enriched in CHB patients. Streptococcus, Lactobacillus, Actinomyces, and Rothia were enriched in LC and HCC patients. Haemophilus and Porphyromonas were found at lower levels in the tongue coating microbiota of digestive system tumours and pancreatic head cancer patients [46,47]. Fusobacterium spp. are described as commensal organisms in the human oral microbiota [48]. Granulicatella was more prevalent in oral cancer patients [49]. Phylogenetic analysis revealed that the Actinomyces and Streptococcus are predominant genera in the saliva of CHB-related CLD [21]. Lactobacillus spp. are widely utilised probiotics and are efficient in ameliorating gastrointestinal tract infections, diarrhoea, inflammatory bowel disease, nonalcoholic fatty liver disease, steatohepatitis, and digestive diseases [50,51]. Streptococcus and Rothia may play a role as opportunistic bacteria, increasing oral microbiome diversity in inflammatory and systemic diseases [39,52,53]. This process is probably relevant to HBV–CLD. Notably, a proportion of Streptococcus continuously accumulated from CHB to the HCC condition [20,54]. Alloprevotella spp. enrichment was found in the CHB patient with a low viral load of HBV DNA [55]. Prevotella was more abundant in the gut microbiota of CHB patients [35,56]. LEfSe modelling and LDA scores revealed that the CHB, LC, and HCC groups were comprised of tongue coating microbial taxa that could differentiate healthy individuals from HBV–CLD patients. The distinct patterns of tongue coating microbiota dysbiosis suggest that tongue coating bacteria could serve as a noninvasive early detection tool in patients with HBV–CLD [57].

PICRUSt2 analysis, applied to 16S rRNA sequencing data, infers the functional potential of tongue coating microbiota in HBV–CLD, identifying metabolic pathways that help to clarify the mechanism of illness by connecting microbial taxa to the KEGG reference database, indicating the role of microbes in persistent viral infection and chronic inflammation. We identified 20 metabolic pathways differentially expressed among the three patient groups and healthy controls. Our investigation revealed that the glycan biosynthesis and metabolism, transport and catabolism, cell growth and death increased in CHB patients, which is possibly due to the higher prevalence of Bacteroidetes in these subjects [58]. Consequently, interactions between microbiota composition and functional categories may help in the development of HBV–CLD.

This study has a few limitations. First, its cross-sectional design limits causal inferences on the microbiota's role in HBV–CLD progression, necessitating longitudinal studies. Secondly, we were limited to making primary findings due to the small sample size of individuals associated with HBV–CLD. Third, the interaction between tongue coating microbiota and intestinal microflora is necessary to understand the development of HBV–CLD. Finally, the rarefaction of reads may bound the estimation of the microbial community. In the future, shotgun metagenomic sequencing should be utilised to investigate the processes and pathophysiology of chronic liver diseases associated with HBV, providing detailed species-level information and elucidating metabolic activities. Despite these limitations, our study offers a fresh perspective on the alteration of tongue coating microbiota and its impact on the prevalence of HBV–CLD. It will be a valuable resource for the identification and therapeutic management of HBV–CLD.

Conclusions

In conclusion, our investigation revealed that the diversity and composition of tongue coating microbiota were significantly altered in HBV–CLD, including CHB, LC, and HCC. Interestingly, the opportunistic pathogen Streptococcus and Rothia were consistently enriched in HBV–CLD, which simplified the network of the oral‒liver axis in HBV–CLD. This dysbiosis may serve as a microbial biomarker for monitoring disease progression and offers unique diagnostic possibilities for patients with HBV–CLD. However, further research should focus on metagenomic insights and explore microbiota-targeted interventions, such as probiotics and prebiotics, to mitigate dysbiosis and improve therapeutic outcomes in HBV–CLD.

Supplementary Material

Supplementary material

Supplementary fig

Supplementary material

Supplementary TAB

Funding Statement

The work was supported by the National Natural Science Foundation of China [Nos. 82071695, 81902364, 82203630, and 82471640], the Young Scientific and Technological Talent Innovation Project of Lanzhou (2023-QN−79), and the Open Fund Project of the NHC Key Laboratory of Diagnosis and Therapy of Gastrointestinal Tumour (NLDTG2020006). The authors thank all the participants in this study and the staff of the department (School of Basic Medical Sciences & Second Affiliated Hospital of Lanzhou University).

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/20002297.2025.2596454.

Acknowledgements

The work was supported by the National Natural Science Foundation of China [Nos. 82071695, 81902364, 82203630, and 82471640], the Young Scientific and Technological Talent Innovation Project of Lanzhou (2023-QN−79), and the Open Fund Project of the NHC Key Laboratory of Diagnosis and Therapy of Gastrointestinal Tumour (NLDTG2020006). The authors thank all the participants in this study and the staff of the department (School of Basic Medical Sciences & Second Affiliated Hospital of Lanzhou University).

Author contributions

CRediT: Umar Pervaiz: Writing – original draft, Methodology, Formal analysis; Fuxia Wu: Visualisation, Data curation; Pervaiz Nabeel: Conceptualisation, Formal analysis; Rui Zhao: Project administration; Zhengbin Zhao: Investigation, Data curation; Yibao Zhang: Methodology, Formal analysis; Peng Xia: Investigation, Data curation; Pengfei Ji: Visualisation, Data curation; Xinyi Yuan: Investigation, Data curation; Xiaohui Hu: Investigation, Data curation; Zhao Guo: Investigation, Data curation; Kun Xie: Investigation, Data curation; Fang Wang: Visualisation, Conceptualisation; Degui Wang: Writing – review & editing, Funding acquisition, Supervision, Conceptualisation.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. All authors have read and approved the final manuscript.

Data availability statement

The 16S rRNA sequencing datasets generated during this study are deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1145884 and will be publicly available upon publication of this manuscript [https://www.ncbi.nlm.nih.gov/sra/PRJNA1145884].

Ethical approval and consent to participate

The study was conducted in compliance with the Helsinki Declaration (updated version, 2013) and was approved by the Ethics Committee of the Second Affiliated Hospital of Lanzhou University. Research was carried out under the approved protocols. All participants provided written informed consent.