Tongue Coating and the Salivary Microbial Communities Vary in Children with Halitosis

Abstract

Halitosis is a common symptom mainly caused by microbial activities in the oral cavity. Here, we used 16S rRNA gene pyrosequencing and metagenomic sequencing to examine oral microbial compositions and their functional variations in children with halitosis. We found that the tongue coating of subjects with halitosis had greater bacterial richness than those of healthy subjects. The relative abundance and prevalence of Leptotrichia wadei and Peptostreptococcus stomatis were higher in tongue coating samples from children with halitosis than those from children without halitosis; Prevotella shahii had higher relative abundance and prevalence in saliva samples from children with halitosis. We present the first comprehensive evaluation of the co-occurrence networks of saliva and tongue coating communities under healthy and halitosis conditions, and investigated patterns of significant differences between these communities. Moreover, we observed that bacterial genes associated with responses to infectious diseases and terpenoid and polyketide metabolism were enriched in subjects with halitosis, but not in healthy subjects. Hydrogen sulphide (H2S)-related metabolic pathways suggested that there was higher microbial production and less usage of H2S in subjects with halitosis. Thus, the mechanism of halitosis was implied for the first time via metagenomic sequencing.

Figure 1. Comparisons of different α diversity indices in healthy and halitosis samples.

A number of sequences (7,180) were randomly subsampled to obtain equal numbers of sequences from each dataset. (A) Observed OTUs, CatchAll (estimated number of OTUs) and Chao1, representing community richness, were calculated for the healthy tongue coating samples (TH) and the halitosis tongue coating samples (TD). (B) Observed OTUs, CatchAll and Chao1 comparisons between the healthy saliva (SH) and halitosis saliva samples (SD). *P < 0.05.

Figure 2. OTUs with different relative abundances and prevalences between the healthy and halitosis samples.

(A) The graph depicts OTUs with different relative abundances and is based on Wilcoxon signed-rank test results from comparisons between healthy (TH) and halitosis tongue coating samples (TD). The bars represent the mean relative abundance (±SD). (B) OTUs with different prevalences were calculated based on Fisher’s exact test. Similar comparisons between the healthy (SH) and halitosis saliva samples (SD) are shown in (C) and (D). The OTUs framed by red boxes had different relative abundances and prevalences between the healthy and halitosis samples.

Figure 3. Co-occurrence networks between tongue coating and saliva samples under healthy and halitosis conditions.

A co-occurrence network was constructed using the relative abundances of all OTUs with Pearson’s correlation coefficients |r| > 0.4 and Permutation test P < 0.01. OTUs with the most linkages (>30) were selected as the hub species. Each node represents an OTU. Lines between nodes show positive correlations (red) or negative correlations (blue). (A,B) Inter-group correlations (namely, correlations of OTUs between tongue coating and saliva) in healthy and halitosis groups. (C,D) Intra-group correlations (namely, correlations of OTUs within tongue coating samples and of OTUs within saliva samples) in healthy and halitosis groups. TC: tongue coating nodes, SA: saliva nodes.

Figure 4. Comparisons of microbial gene sets in different samples.

(A) Venn diagram of gene sets from a healthy tongue coating sample (TH) and a halitosis tongue coating sample (TD). The number of genes in each group and the overlapping area are shown. (B) Venn diagram of gene sets from a healthy saliva sample (SH) and a halitosis saliva sample (SD).

Figure 5. Comparisons of differential genes in tongue coating and saliva samples and enrichment of KO groups in tongue coating and saliva samples.

(A) Genes with relative abundance differences of >5-fold and P-values < 0.001 in the DEGexp results were defined as differential genes in the tongue coating samples and are shown in the overlapping areas. (B) Differential genes in the saliva samples are shown. (C) KO enrichment groups in tongue coating and saliva samples are shown as percentages. Groups framed by a red box were enriched in both halitosis samples (tongue coating and saliva) but not in the healthy samples. *P < 0.05.

Figure 6. Functional variations in the H2S metabolic process between halitosis and healthy microbiomes.

Genes with relative abundance ratios of >2 were defined as differentially abundant genes and are indicated by arrows. The upward arrows represent genes in the halitosis samples that were more abundant than in the healthy samples, whereas the downward arrows represent genes that were less abundant in the halitosis samples. Genes with relative abundance ratios of <2 are shown as “−”. Variations in tongue coating samples are labelled outside the “[]”, whereas variations in saliva samples are labelled inside the “[]”.

Figure 7. The core tongue coating microbiome and the health-associated and halitosis-associated tongue coating microbiomes.

Core microbiomes were defined as OTUs with prevalences above 75% in both the healthy and halitosis groups (green). Health-associated microbiomes were defined as OTUs with prevalences above 75% in the healthy tongue coating samples, but less than 75% in the halitosis tongue coating samples (blue). Conversely, halitosis-associated microbiomes were defined as OTUs with prevalences above 75% in the halitosis tongue coating samples, but less than 75% in the healthy tongue coating samples (orange). Prevalence and relative abundance were used for further filtering. Circles labelled with “1” contain OTUs that were present in all 19 subjects with a mean relative abundance of ≥1% of the total sequences. Circles labelled with “2” contain OTUs that were present in all healthy (dark blue) or halitosis samples (dark orange), but with low abundances (<1%). The outer circles labelled with “3” contain OTUs with greater than 75% but less than 100% prevalence in the healthy (light blue) and halitosis samples (light orange) with a mean relative abundance of <1%. TH: healthy tongue coating sample, TD: tongue coating sample with halitosis.