A Longitudinal Study of the Human Oropharynx Microbiota Over Time Reveals a Common Core and Significant Variations With Self-Reported Disease

Abstract

Our understanding of human microbial communities, in particular in regard to diseases is advancing, yet the basic understanding of the microbiome in healthy subjects over time remains limited. The oropharynx is a key target for colonization by several important human pathogens. To understand how the oropharyngeal microbiome might limit infections, and how intercurrent infections might be associated with its composition, we characterized the oropharyngeal microbiome of 18 healthy adults, sampled weekly over a 40-weeks using culture-independent molecular techniques. We detected nine phyla, 202 genera and 1438 assignments on OTU level, dominated by Firmicutes, Bacteroidetes, and Proteobacteria on phylum level. Individual microbiomes of participants were characterized by levels of high alpha diversity (mean = 204.55 OTUs, sd = 35.64), evenness (19.83, sd = 9.74) and high temporal stability (mean Pearson's correlation between samples of 0.52, sd = 0.060), with greater differences in microbiome community composition between than within individuals. Significant changes in community composition were associated with disease states, suggesting that it is possible to detect specific changes in OTU abundance and community composition during illness. We defined the common core microbiota by varying occurrence and abundance thresholds showing that individual core microbiomes share a substantial number of OTUs across participants, chiefly Streptococci and Veillonella. Our results provide insights into the microbial communities that characterize the healthy human oropharynx, community structure and variability, and provide new approaches to define individual and shared cores. The wider implications of this result include the potential for modeling the general dynamics of oropharynx microbiota both in health and in response to antimicrobial treatments or probiotics.

Keywords: community; core; microbial; microbiome; oral; oropharynx.

FIGURE 1

(A) Mean alpha diversity (i.e., count of unique OTUs in each sample) for all participants (±sd) and (B) mean evenness (Inverse Simpson). Horizontal bars show significant pairwise differences (p < 0.05) across participants. Note the difference in scale.

FIGURE 2

Hierarchical clustering of log10 + 1 transformed microbial abundance data of all participants across time points, with samples representing one sampling week. Highlighted are samples collected from two participants (HA in red, HT in blue), numbers indicate sampling weeks.

FIGURE 3

Example of temporal changes in community composition over the course of the data collection for two participants of this study: (A) HA (self-reporting a cold during weeks 14 annotated C1 and 26 annotated C2, C1, and C2, respectively) and (B) HT (self-reportedly healthy throughout the sampling period): (A) Relative abundance (%) of dominant phyla, (B) heatmap of the 20 most abundant OTUs, (C) non-metric multidimensional scaling ordination of bacterial community structure [dashed ellipse represents the 95% confidence interval for the centroid of each stratification group as calculated by ordiellipse (Oksanen et al., 2018)].

FIGURE 4

Phylogenetic core of the oropharyngeal microbiome of participants HA (A) and HT (B) based on abundance (% relative abundance) and prevalence (% occurrence in samples) in accordance with previous studies (Jalanka-Tuovinen et al., 2011), data was bootstrapped to standardize sampling effort. Relative abundance values ranged from 0.001 to 0.2 and prevalence from 5 to 100%. Overlapping OTUs between participants are highlighted bold.

FIGURE 5

Bacterial core of the oropharyngeal microbiome from the 18 participants of the study based on selected abundance (% relative abundance) and prevalence (% occurrence in samples), data was bootstrapped to standardize sampling effort. Relative abundance values ranged from 0.001 to 0.2 and prevalence from 5 to 100%. (A) Core size as a function of prevalence and (B) bacterial taxa as a function of detection threshold and prevalence.

FIGURE 6

Heat tree matrix comparing OTUs of the bacterial core of the oropharyngeal microbiome from the participants of this study. Lower left-hand side diagram shows the phylogeny of the pooled data set and the sizes of the circles associated with different taxa indicate their relative abundances. Brown and cyan colors indicate significant differences across pairwise abundances, while gray represent no significant difference in relative abundance.