Saliva and Lung Microbiome Associations with Electronic Cigarette Use and Smoking

Abstract

The microbiome has increasingly been linked to cancer. Little is known about the lung and oral cavity microbiomes in smokers, and even less for electronic cigarette (EC) users, compared with never-smokers. In a cross-sectional study (n = 28) of smokers, EC users, and never-smokers, bronchoalveolar lavage and saliva samples underwent metatranscriptome profiling to examine associations with lung and oral microbiomes. Pairwise comparisons assessed differentially abundant bacteria species. Total bacterial load was similar between groups, with no differences in bacterial diversity across lung microbiomes. In lungs, 44 bacteria species differed significantly (FDR < 0.1) between smokers/never-smokers, with most decreased in smokers. Twelve species differed between smokers/EC users, all decreased in smokers of which Neisseria sp. KEM232 and Curvibacter sp. AEP1-3 were observed. Among the top five decreased species in both comparisons, Neisseria elongata, Neisseria sicca, and Haemophilus parainfluenzae were observed. In the oral microbiome, 152 species were differentially abundant for smokers/never-smokers, and 17 between smokers/electronic cigarette users, but only 21 species were differentially abundant in both the lung and oral cavity. EC use is not associated with changes in the lung microbiome compared with never-smokers, indicating EC toxicity does not affect microbiota. Statistically different bacteria in smokers compared with EC users and never-smokers were almost all decreased, potentially due to toxic effects of cigarette smoke. The low numbers of overlapping oral and lung microbes suggest that the oral microbiome is not a surrogate for analyzing smoking-related effects in the lung.

Prevention relevance: The microbiome affects cancer and other disease risk. The effects of e-cig usage on the lung microbiome are essentially unknown. Given the importance of lung microbiome dysbiosis populated by oral species which have been observed to drive lung cancer progression, it is important to study effects of e-cig use on microbiome.

Figure 1.

Lung Microbiome associated with Smoking Status. A) Distribution of alpha diversity, measured by the Shannon Diversity H-Index, in the samples stratified by smoking group; smoking groups were similar (P>0.05, Kruskal-Wallis) B) MDS plot clustered samples using Bray-Curtis dissimilarity index as a distance metric to compare beta diversity by smoking status (blue – never-smoker, green – e-cig users, red – smokers); never-smokers and e-cig users clustered together, compared to smokers. C) The relative abundances of bacteria species with at least 1% abundance in any smoking group identified in the lung. Bacteria species colored in red were species with > 1% abundance in both the lung and oral microbiome. D) Number of significant bacteria species in common from the multiple pairwise smoking status comparisons. There were 10 bacterial species in common between SM vs. NS and SM vs. EC. E) Volcano plot, of log2 fold change vs –log2(P-value), for SM vs. NS, SM vs. EC and EC vs. NS. Significant bacteria species were colored red. (FDR-adjusted p-value <0.1). Most of the bacteria species were decreased in SM vs. NS, while all bacteria species were decreased in SM vs. EC.

Figure 2.

Relative abundances of significant bacteria from SM/NS and SM/Ecig pairwise comparisons. A) Aerobic bacteria species’ relative abundances were decreased in SM compared to EC users and NS. B) Aerobic bacteria species’ relative abundances were increased in SM compared to NS. C) Anaerobic bacteria species’ relative abundance in SM compared to EC users and NS. The first 3 bacteria species have decreased relative abundance in SM while the last 2 bacteria species have increased relative abundance in SM.

Figure 3.

Oral microbes associated with smoking status. A) Distribution of alpha diversity, measured by the Shannon Diversity H-Index, in the samples stratified by smoking group; smoking groups were similar (P>0.05, Kruskal-Wallis) B) MDS plot clustered samples using Bray-Curtis dissimilarity index as a distance metric to compare beta diversity by smoking status (blue – never-smoker, green – e-cig users, red – smokers) never-smokers and e-cig users clustered together, compared to smokers. C) The relative abundances of bacteria species with at least 1% abundance in any smoking group in the oral microbiome are depicted. Bacteria species colored in red are those with > 1% abundance in both the lung and oral samples. D) Significant bacteria species in common from the multiple pairwise smoking status comparisons. E) Volcano plot, of log2 fold change vs –log2(P-value), for SM vs. NS, SM vs. EC and EC vs. NS. Significant bacteria species are colored red (FDR-adjusted p-value <0.1). F) Relative abundances of commensal oral microbiome species (Neisseria elongata, Haemophilus parainfluenzae and Fusobacterium periodonticum) by smoking groups. These common oral microbiome commensal species were all decreased in smokers.

Figure 4.

Differences in Significant Bacteria Species of the Oral and Lung Microbiome. A-B) Upset plots show the number of bacteria species observed in oral microbiome, lung microbiome and both the oral and lung microbiome by pairwise smoking comparison. A) In the SM/NS pairwise comparison, 131 species were observed in the oral microbiome, 23 species observed in the lung microbiome, and 21 species were observed both the lung and oral microbiome. B) In the SM/EC pairwise comparison, 13 species were observed in the oral microbiome, 8 species were observed in the lung microbiome, and 4 species were observed in both the lung and oral microbiome. C-D) Heatmap generated from all significant microbes from pairwise comparison. C) Significant lung microbes were used to produce the heatmap, demonstrating clustering of all SM together in one group, with 2 EC users and 1 NS. D) Significant oral microbes do not cluster the smokers together in one group.