Dysbiosis and alterations in predicted functions of the subgingival microbiome in chronic periodontitis

Abstract

Chronic periodontitis is an inflammatory disease of the periodontium affecting nearly 65 million adults in the United States. Changes in subgingival microbiota have long been associated with chronic periodontitis. Recent culture-independent molecular studies have revealed the immense richness and complexity of oral microbial communities. However, data sets across studies have not been directly compared, and whether the observed microbial variations are consistent across different studies is not known. Here, we used 16S rRNA sequencing to survey the subgingival microbiota in 25 subjects with chronic periodontal disease and 25 healthy controls and compared our data sets with those of three previously reported microbiome studies. Consistent with data from previous studies, our results demonstrate a significantly altered microbial community structure with decreased heterogeneity in periodontal disease. Comparison with data from three previously reported studies revealed that subgingival microbiota clustered by study. However, differences between periodontal health and disease were larger than the technical variations across studies. Using a prediction score and applying five different distance metrics, we observed two predominant clusters. One cluster was driven by Fusobacterium and Porphyromonas and was associated with clinically apparent periodontitis, and the second cluster was dominated by Rothia and Streptococcus in the majority of healthy sites. The predicted functional capabilities of the periodontitis microbiome were significantly altered. Genes involved in bacterial motility, energy metabolism, and lipopolysaccharide biosynthesis were overrepresented in periodontal disease, whereas genes associated with transporters, the phosphotransferase system, transcription factors, amino acid biosynthesis, and glycolysis/gluconeogenesis were enriched in healthy controls. These results demonstrate significant alterations in microbial composition and function in periodontitis and suggest genes and metabolic pathways associated with periodontal disease.

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

Figures

FIG 1

Microbial diversity, evenness, and richness in subgingival microbiota in subjects with chronic periodontitis and healthy controls, shown in box plots. (A) The Shannon diversity index was used to estimate microbial diversity for each group. The species evenness index was calculated by using the formula J′ = H′/H′max, where H′ is the Shannon diversity index and H′max is the maximal value of H′ (i.e., ln S, where S is the total number of species in the community). Species richness was defined as the number of OTUs identified in each sample. Each point represents an individual subgingival sample. (B) CP sites were stratified by pocket depths (6 mm or 7 mm/8 mm). P values (Student's t test) are shown above the bars for each comparison.

FIG 2

Comparison of subgingival microbial community composition. Weighted UniFrac analysis was used to generate distances among different samples. Scattered plots were then generated by using principal coordinate analysis. The percentage of variation explained by each principal coordinate (PC) is indicated on the axes. Each point represents a microbial community. (A) Microbial communities in subjects with chronic periodontitis versus healthy controls. (B) Microbial communities in subjects with chronic periodontitis at pocket depths of 6 mm, 7 mm, and 8 mm versus healthy controls. (C) Subgingival microbial communities in this study versus the subgingival data set from the Human Microbiome Project. (D) Bleeding sites and nonbleeding sites from the study by Abusleme et al. (31) versus sites from subjects with chronic periodontitis in this study. (E) Healthy sites from the study by Abusleme et al. versus healthy sites from this study. (F) Chronic periodontitis sites from the study by Griffen et al. (30) versus CP sites from this study. (G) Healthy sites from subjects with chronic periodontitis and healthy sites from healthy controls from the study by Griffen et al. versus healthy sites from this study. (H) Average UniFrac distance between pairs of samples within each group, indicating lower heterogeneity in subgingival microbial communities in the chronic periodontitis group. Error bars indicate standard errors of the means.

FIG 3

Relative abundance of bacterial taxa at the phylum level in subjects with chronic periodontitis (CP) and healthy controls (HC). (A) Relative proportion of sequence reads for each phylum. (B) Relative proportions of bacterial phylotypes for each phylum. (C) Top 10 most abundant families in HCs. (D) Top 10 most abundant families in subjects with CP.

FIG 4

Differentially abundant bacterial phylotypes identified by linear discriminant analysis (LDA) coupled with effect size measurements (LEfSe). Bacterial taxa enriched in healthy sites are indicated with positive linear discriminant analysis scores, and taxa enriched in periodontitis sites are indicated with negative linear discriminant analysis scores. Only taxa that met the significant linear discriminant analysis threshold of 3.5 are shown. Phylotypes that were also significantly different between the two groups by multivariate analysis are indicated by an asterisk. The oral taxon numbers are derived from the Human Oral Microbiome Database.

FIG 5

Clustering for subgingival microbial communities. (A) Prediction scores based on weighted UniFrac distances showing strong support for two independent clusters. One cluster is dominated by periodontitis sites (red), and the second cluster consists of mostly healthy microbial communities (blue). (B) Measures of microbial diversity and species richness in the two clusters (P > 0.5). (C) Relative proportions of sequence reads according to the taxa indicated.

FIG 6

Differentially abundant gene functions in subjects with chronic periodontitis and healthy controls. Functional categories of genes of the subgingival metagenome were predicted by using PICRUSt, and differentially abundant functions were then identified by using linear discriminant analysis (LDA) coupled with effect size measurements (LEfSe). Gene functions enriched in healthy sites are indicated with positive linear discriminant analysis scores, and functions differentially enriched in periodontitis sites are indicated with negative linear discriminant analysis scores. Only gene functions that have a linear discriminant analysis score threshold of 2.75 are shown.