Application of a neutral community model to assess structuring of the human lung microbiome

Abstract

DNA from phylogenetically diverse microbes is routinely recovered from healthy human lungs and used to define the lung microbiome. The proportion of this DNA originating from microbes adapted to the lungs, as opposed to microbes dispersing to the lungs from other body sites and the atmosphere, is not known. We use a neutral model of community ecology to distinguish members of the lung microbiome whose presence is consistent with dispersal from other body sites and those that deviate from the model, suggesting a competitive advantage to these microbes in the lungs. We find that the composition of the healthy lung microbiome is consistent with predictions of the neutral model, reflecting the overriding role of dispersal of microbes from the oral cavity in shaping the microbial community in healthy lungs. In contrast, the microbiome of diseased lungs was readily distinguished as being under active selection. We also assessed the viability of microbes from lung samples by cultivation with a variety of media and incubation conditions. Bacteria recovered by cultivation from healthy lungs represented species that comprised 61% of the 16S rRNA-encoding gene sequences derived from bronchoalveolar lavage samples.

Importance: Neutral distribution of microbes is a distinguishing feature of the microbiome in healthy lungs, wherein constant dispersal of bacteria from the oral cavity overrides differential growth of bacteria. No bacterial species consistently deviated from the model predictions in healthy lungs, although representatives of many of the dispersed species were readily cultivated. In contrast, bacterial populations in diseased lungs were identified as being under active selection. Quantification of the relative importance of selection and neutral processes such as dispersal in shaping the healthy lung microbiome is a first step toward understanding its impacts on host health.

FIG 1

(A) Anatomical connections between the lungs, stomach, and oral and nasal cavities permit movement of microbes between these sites. (B) Implementation of a neutral model for analysis of microbial community structure. The solid black line represents the best-fit neutral model generated using a probability distribution. Dashed lines represent 95% confidence intervals around this best-fit neutral model. Species within the confidence intervals (gray points) are likely present in the lungs because of neutral processes (i.e., dispersal from the source community and ecological drift within the lungs). Species deviating from the neutral model (green and red points) either are candidates for selection in the lungs or differ in their ability to disperse compared to other microbes in the source community.

FIG 2

Results of neutral model testing with (A) tonsils, (B) gingiva, and (C) female urogenital tract as potential source communities for microbes in the lungs. The coefficient of determination (R²) is the goodness of fit of the neutral model. It ranges from ≤0 (no fit) to 1 (perfect fit). Note the correspondence between decreasing values of R² from panels A to C and the proximity of points to the line denoting the best-fitting neutral model.

FIG 3

The neutral model applied to each body site as a potential source of microbes for the lung microbiome. Shown are the abundance of neutrally distributed, overrepresented, and underrepresented OTUs in lungs from each body site and the associated goodness of fit of the neutral model (R²).

FIG 4

List of OTUs that most commonly deviated from the neutral model and had a relative abundance of >0.5% in the lungs. Green circles indicate the sites from which the OTU appeared as being overrepresented in the lungs, and gray circles are the sites from which the OTU was neutrally distributed.

FIG 5

Cultivar taxonomy mapped onto the rank abundance curve of 16S rRNA-encoding gene surveys of BAL samples. Colored OTUs represent those scored as having a “cultured representative.” Open circles represent OTUs that were not cultivated (a list of uncultivated OTUs is provided in Table S1 in the supplemental material).