Exploring the oral microbiota of children at various developmental stages of their dentition in the relation to their oral health

Abstract

Background: An understanding of the relation of commensal microbiota to health is essential in preventing disease. Here we studied the oral microbial composition of children (N = 74, aged 3 - 18 years) in natural transition from their deciduous to a permanent dentition and related the microbial profiles to their oral health status. The microbial composition of saliva was assessed by barcoded pyrosequencing of the V5-V6 hypervariable regions of the 16 S rRNA, as well as by using phylogenetic microarrays.

Results: Pyrosequencing reads (126174 reads, 1045 unique sequences) represented 8 phyla and 113 higher taxa in saliva samples. Four phyla--Firmicutes, Bacteriodetes, Proteobacteria and Actinobacteria--predominated in all groups. The deciduous dentition harboured a higher proportion of Proteobacteria (Gammaproteobacteria, Moraxellaceae) than Bacteroidetes, while in all other groups Bacteroidetes were at least as abundant as Proteobacteria. Bacteroidetes (mainly genus Prevotella), Veillonellaceae family, Spirochaetes and candidate division TM7 increased with increasing age, reflecting maturation of the microbiome driven by biological changes with age. Microarray analysis enabled further analysis of the individual salivary microbiota. Of 350 microarray probes, 156 gave a positive signal with, on average, 77 (range 48-93) probes per individual sample. A caries-free oral status significantly associated with the higher signal of the probes targeting Porphyromonas catoniae and Neisseria flavescens.

Conclusions: The potential role of P. catoniae and N. flavescens as oral health markers should be assessed in large-scale clinical studies. The combination of both open-ended and targeted molecular approaches provides us with information that will increase our understanding of the interplay between the human host and its microbiome.

Figure 1

Relative abundance of the main bacterial phyla (99.9% of the reads) identified in saliva of children at various developmental stages of their dentition. Average and standard deviations from three pools of samples - healthy, treated and caries - per each developmental stage. * - Adult saliva data are added for comparison from Keijser et al. 2008 [1].

Figure 2

Relative abundance of the predominant (92-95% of the reads) bacterial groups (family or genus if only one genus comprised the family, or unclassifiable bacteria within one phylum) in saliva samples of children at various developmental stages of the dentition. Average and standard deviations from three pools of samples - healthy, treated and caries - per each developmental stage. NA - not assigned.

Figure 3

Results of the Significance Analysis of Microarrays (SAM) on individual sample profiles obtained by phylogenetic microarray. Individual saliva samples are shown in the columns, the probes - in the rows. Sample groups by health status and by dentition are shown in different colours. High probe signal abundance is depicted red, low - dark blue. Only probes which had significant association with the oral health status are shown. No association with the stage of dentition and individual probes was found.

Figure 4

Results of the Principal Component Analysis (PCA) on individual sample profiles obtained by phylogenetic microarray. A) the plot of the PCA axis 1 (accounting for 21% of intersample variation) and the axis 2 (14% of intersample variation) B) the plot of the PCA axis 2 and the axis 3 (11.4% of intersample variation). Green dots - samples from the caries-free group, blue dots - samples from the treated group, red dots - samples from children with dental caries. Samples identified with the same number belonged to the siblings from one family. Samples without numbers belonged to unrelated individuals.

Figure 5

Visualization of similarities between salivary profiles by hierarchical clustering analysis (average linkage method, Pearson correlation). Samples are depicted in the columns. Siblings, who participated in the study, are marked by the same colour, each colour representing different family. Unrelated individuals have no colour code.

Figure 6

Comparison of cluster profiling results of oral microbiota data obtained through 454 pyrosequencing (genus level), left part, and taxonomic microarray analysis (right). Hierarchical clustering for both datasets was performed in MeV v4.5.