Alterations of Bacteroides sp., Neisseria sp., Actinomyces sp., and Streptococcus sp. populations in the oropharyngeal microbiome are associated with liver cirrhosis and pneumonia

Abstract

Background: The microbiomes of humans are associated with liver and lung inflammation. We identified and verified alterations of the oropharyngeal microbiome and assessed their association with cirrhosis and pneumonia.

Methods: Study components were as follows: (1) determination of the temporal stability of the oropharyngeal microbiome; (2) identification of oropharyngeal microbial variation in 90 subjects; (3) quantitative identification of disease-associated bacteria. DNAs enriched in bacterial sequences were produced from low-biomass oropharyngeal swabs using whole genome amplification and were analyzed using denaturing gradient gel electrophoresis analysis.

Results: Whole genome amplification combined with denaturing gradient gel electrophoresis analysis monitored successfully oropharyngeal microbial variations and showed that the composition of each subject's oropharyngeal microbiome remained relatively stable during the follow-up. The microbial composition of cirrhotic patients with pneumonia differed from those of others and clustered together in subgroup analysis. Further, species richness and the value of Shannon's diversity and evenness index increased significantly in patients with cirrhosis and pneumonia versus others (p < 0.001, versus healthy controls; p < 0.01, versus cirrhotic patients without pneumonia). Moreover, we identified variants of Bacteroides, Eubacterium, Lachnospiraceae, Neisseria, Actinomyces, and Streptococcus through phylogenetic analysis. Quantitative polymerase chain reaction assays revealed that the populations of Bacteroides, Neisseria, and Actinomycetes increased, while that of Streptococcus decreased in cirrhotic patients with pneumonia versus others (p < 0.001, versus Healthy controls; p < 0.01, versus cirrhotic patients without pneumonia).

Conclusions: Alterations of Bacteroides, Neisseria, Actinomyces, and Streptococcus populations in the oropharyngeal microbiome were associated with liver cirrhosis and pneumonia.

Fig. 1

The applicability, feasibility, and validity of the WGA method for enriching microbial DNA. (a) DGGE profiles of WGA-amplified oropharyngeal DNAs (lanes A1, B1, C1, and D1) and the original oropharyngeal DNA (lanes A0, B0, C0, and D0); (b) the WGA-amplified fecal DNA (lanes A1–3 and B1–3) represented the most predominant microbial bands that were similar to those detected in the original fecal DNA (lanes A0 and B0), and were reproducible; (c) Seven oropharyngeal microbial DNA samples were amplified in duplicate using WGA for DGGE analysis. Clustering was performed using Dice’s coefficient and UPGMA with BioNumerics software

Fig. 2

Confirmation of the temporal stability of the oropharyngeal microbiome in a 5-day follow-up study. Cluster analysis of DGGE profiles of oropharyngeal bacteria in groups HC (a), CC (a), and CI (b). Cluster analysis was performed using Dice’s coefficient and UPGMA. Lanes were designated by patient number (1, 2, 3, 4), and visit number (1, 2, 3, and 4).CIn-1, before antibiotic treatment; CIn-2, during antibiotic treatment; CIn-3 and -4, after antibiotic treatment. The diversity index of these follow-up samples is shown (c). *p < 0.01; **p < 0.001

Fig. 3

Cluster and diversity analyses of DGGE profiles of the oropharyngeal mucosal microbiomes of 90 PCR-DGGE profiles. (a) Cluster analysis of DGGE profiles of the oropharyngeal mucosal microbiome; (b) Multidimensional scaling analysis (MDS) based on the predominant oropharyngeal bacterial PCR-DGGE profiles. The plot is an optimized three-dimensional representation of the similarity matrix obtained using BioNumerics software, and the x-, y-, and z-axes represent three different dimensions (Dim 1, Dim 2, and Dim 3, respectively). (c) Principal component analysis (PCA) based on bacterial PCR-DGGE profiles. The plots were reoriented to maximize the variation among lanes along the first three principal components (contributions: 18.2, 8.2, and 7.4 %, respectively) obtained using BioNumerics software. Cubes, group CI; spheres, group CC; cylinders, group HC. (d) Comparison of the diversity index, species richness, and evenness index of oropharyngeal mucosal microbiomes. *p < 0.01; **p < 0.001

Fig. 4

Phylogenetic tree analysis based on bacterial PCR-DGGE profiles. (a) Nucleotide sequence identification of the bands from the PCR-DGGE profiles. The number corresponds to the number of band-classes in the phylogenetic tree. (b) The phylogenetic tree generated using a neighbor-joining method of the sequences derived from the DGGE profiles. The fragment sequences were designated according to their positions in gels using the band-matching tool with BioNumerics software version 6.01. The plot was prepared using MEGA5 software

Fig. 5

Identification and analysis of the key band-classes based on bacterial PCR-DGGE profiles of the groups. (a) Comparisons intensities of the key band-classes based on bacterial PCR-DGGE profiles of the different groups. (b) Analysis of the frequencies of key band-classes based on bacterial PCR-DGGE profiles of the different groups. (P-values for the comparison between the groups were shown in supplementary table S3)

Fig. 6

RT-qPCR analysis of the key bacterial species accounting for variations in microbiome compositions. Number of key bacteria: log # per 10 ng original swab DNA. The figure shows the numbers of key bacteria, including Streptococcus, Streptococcus mitis, Actinomycetes, Lachnospiraceae, Eubacterium, Bacteroides, and Neisseria