Bacterial diversity in saliva and oral health-related conditions: the Hisayama Study

Abstract

This population-based study determined the salivary microbiota composition of 2,343 adult residents of Hisayama town, Japan, using 16S rRNA gene next-generation high-throughput sequencing. Of 550 identified species-level operational taxonomic units (OTUs), 72 were common, in ≥75% of all individuals, as well as in ≥75% of the individuals in the lowest quintile of phylogenetic diversity (PD). These "core" OTUs constituted 90.9 ± 6.1% of each microbiome. The relative abundance profiles of 22 of the core OTUs with mean relative abundances ≥1% were stratified into community type I and community type II by partitioning around medoids clustering. Multiple regression analysis revealed that a lower PD was associated with better conditions for oral health, including a lower plaque index, absence of decayed teeth, less gingival bleeding, shallower periodontal pockets and not smoking, and was also associated with tooth loss. By contrast, multiple Poisson regression analysis demonstrated that community type II, as characterized by a higher ratio of the nine dominant core OTUs, including Neisseria flavescens, was implicated in younger age, lower body mass index, fewer teeth with caries experience, and not smoking. Our large-scale data analyses reveal variation in the salivary microbiome among Japanese adults and oral health-related conditions associated with the salivary microbiome.

Figure 1. The 72 operational taxonomic units (OTUs) commonly (≥75%) identified in saliva of the individuals in the lowest quintile of phylogenetic diversity.

The OTUs are ordered according to a phylogenetic tree built based on the 16S rRNA gene sequences using FastTree; the tree is described in the center of the diagram. Mean relative abundance of each OTU in the individuals in the lowest PD quintile is shown as a blue bar. Of the 72 OTUs, 22 with a mean relative abundances of ≥1% in the saliva of individuals in the lowest PD quintile were shown in bold. Abbreviations: Mor; Moraxella; Ori, Oribacterium; Meg, Megasphaera; Lep, Leptotorichia; Fus, Fusobacterium; Cam, Campylobacter; Lau, Lautropia; All, Alloprevotella; Ber; Bergeyella; Eub, Eubacterium; Ato, Atopobium; Sol, Solobacterium; Abi, Abiotrophia.

Figure 2. OTUs commonly (≥75%) identified in the saliva of the individuals in the highest quintile (Q5) of phylogenetic diversity, except for the 72 core OTUs shown in Fig. 1

. Of the 77 OTUs, only 65 corresponding to the bacterial species in the Human Oral Microbiome Database with a high (≥98%) identity are shown. OTUs shared among the lower quintile (Q2, Q3, Q4 and Q5) are shown separately. The detection rate (%) of each OTU in each quintile is shown in each grid by the color intensity.

Figure 3. Phylogenetic diversity of salivary bacterial populations of the individuals for various general and clinical conditions.

Of 2,343 individuals whose salivary bacterial composition was determined, 5 are excluded because of antibiotics use. Furthermore, 256 individuals with ≤8 teeth and 2 individuals from whom periodontal data were missing are excluded in %DFT, mean PPD, %teeth with BOP and mean plaque index. Dots indicate the mean; the error bars indicate 95% confidence intervals. ***P < 0.001, **P < 0.01, *P < 0.05 in bivariate analyses using Pearson’s correlation test^a or the Student’s t-test^b. The individuals are categorized into two groups by the Student’s t-test in terms of the number of DT and smoking history (presence vs. absence and current smokers vs. the other individuals, respectively). Abbreviation: BMI, body mass index; DT, decayed teeth; DFT, decayed and filled teeth, PPD, periodontal pocket depth; BOP, bleeding on probing.

Figure 4. Microbial co-occurrence in the salivary bacterial population.

Co-occurrence networks were constructed based on the relative abundances of 22 of 72 core OTUs shown in Fig. 1. The nodes represent the OTU; the nodes with the same color are classified as the same genus. Each edge represents a positive correlation between the two OTUs with an adjusted P value < 10⁻¹² in Pearson’s correlation test, and their thickness corresponds to the coefficient values.

Figure 5. A principal component analysis biplot diagram showing similarity relationships among 2,343 relative abundance profiles for the 22 OTUs.

The relative abundance profiles belonging to each community type are depicted using different colors. The direction and length of the arrows indicate how each OTU contributes to the first two components in the biplot. These two components explained 22.4% and 10.3% of the variance, respectively.