The mediating roles of the oral microbiome in saliva and subgingival sites between e-cigarette smoking and gingival inflammation

Here, we assess the disparity in each microbial taxon at each taxonomic rank (i.e., phylum, class, order, family, genus, species) between EC users and non-users while adjusting for age, gender, and the frequency of brushing teeth [Fig. 5, Fig. 6]. We found significant disparities in relative abundance between EC users and non-users for: 36 microbial taxa (i.e., 2 phyla, 3 classes, 6 orders, 7 families, 8 genera, 10 species) in the saliva site [Fig. 5]; 71 microbial taxa (i.e., 4 phyla, 7 classes, 11 orders, 15 families, 21 genera and 13 species) in the subgingival site [Fig. 6]. Of these, we identified 21 microbial taxa in common (i.e., from both the saliva and subgingival sites) (i.e., 2 phyla, 2 classes, 4 orders, 5 families, 4 genera, 4 species), 15 microbial taxa (i.e., 1 class, 2 orders, 2 families, 4 genera, 6 species) only in the saliva site, and 50 microbial taxa (i.e., 2 phyla, 5 classes, 7 orders, 10 families, 17 genera, 9 species) only in the subgingival site.

Fig. 5figure 5

The results from taxonomic differential abundance analysis using saliva samples. The fitted random effects model to assess the disparity in each microbial taxon at each taxonomic rank (i.e., phylum, class, order, family, genus, species) between EC users and non-users while adjusting for age, gender and the frequency of brushing teeth. *Est. represents the estimated coefficient on the effect of each taxon in the fitted random effects model. *Q-value represents the P-value after the FDR control. *Only the statistically significant taxa after the FDR control (i.e., Q-value < 0.05) are included. For the species, the one before semi-colon (;) is the genus name and the one after semi-colon (;) is the species name

Fig. 6figure 6

The results from taxonomic differential abundance analysis using subgingival samples. The fitted random effects model to assess the disparity in each microbial taxon at each taxonomic rank (i.e., phylum, class, order, family, genus, species) between EC users and non-users while adjusting for age, gender and the frequency of brushing teeth. *Est. represents the estimated coefficient on the effect of each taxon in the fitted random effects model. *Q-value represents the P-value after the FDR control. *Only the statistically significant taxa after the FDR control (i.e., Q-value < 0.05) are included. For the species, the one before semi-colon (;) is the genus name and the one after semi-colon (;) is the species name

Shared by saliva and subgingival sites [Fig. 5, Fig. 6]: A total of 21 microbial taxa were discovered in both the saliva and subgingival sites to be differentially abundant. Of these, 4 taxa showed an increase in relative abundance for EC users, while 17 microbial taxa showed a decrease in relative abundance for EC users. Interestingly, both the saliva and subgingival sites had the same effect direction (increase or decrease) for all these shared taxa in relative abundance between EC users and non-users. At the genus level, Bergeyella, Neisseria, Enterococcus and Haemophilus showed a decrease in relative abundance for EC users.

Unique to the saliva site [Fig. 5]: A total of 15 microbial taxa were discovered only in the saliva site to be differentially abundant. Of these, 10 taxa showed an increase in relative abundance for EC users, while 5 taxa showed a decrease in relative abundance for EC users. At the genus level, Catonella showed a decrease in relative abundance for EC users while Alloscardovia, Cryptobacterium and Dialister showed an increase in relative abundance for EC users.

Unique to the subgingival site [Fig. 6]: A total of 50 microbial taxa were discovered only in the subgingival site to be differentially abundant. Of these, 16 taxa showed an increase in relative abundance for EC users, while 34 taxa showed a decrease in relative abundance for EC users. At the genus level, Bacteroidetes_[G-3], Olsenella, Lachnospiraceae_[G-7], Filifactor, Peptostreptococcaceae_[XI][G-1] and Treponema showed an increase in relative abundance for EC users, while Porphyromonas, Capnocytophaga, Leptotrichia, Actinomyces, Corynebacterium, Rothia, Peptidiphaga, Kingella, Cardiobacterium, Streptococcus and Abiotrophia showed a decrease in relative abundance for EC users.

Overall, the greater abundance and variety of taxa cataloged at the subgingival site [Fig. 6] than at the saliva site [Fig. 5] may indicate that the oral microbiome at the subgingival site is affected more strongly or preserves microbiome changes longer than the oral microbiome at the saliva site. The difference in the taxonomic discoveries may also indicate that the oral microbiomes in the saliva and subgingival sites are affected by EC smoking in different ways.

We also include the results of univariate analysis with no covariate adjustments in the Supplementary Information [Additional file 4: Figure S4, Additional file 5: Figure S5]. We found fewer taxonomic discoveries in the analysis with covariate adjustments (i.e., a total of 36 microbial taxa) [Fig. 5] than in the analysis with no covariate adjustments (i.e., a total of 61 microbial taxa) for the saliva site [Additional file 4: Figure S4]; as such, the proportion of taxonomic discoveries in the analysis with covariate adjustments compared with the one with no covariate adjustments in the saliva site was 36/61 (i.e., about 59%). In contrast, for the subgingival site, we found relatively less fewer taxonomic discoveries in the analysis with covariate adjustments (i.e., a total of 71 taxa) [Fig. 6] than in the analysis with no covariate adjustments (i.e., a total of 82 taxa) [Additional file 5: Figure S5]; as such, the proportion of taxonomic discoveries in the analysis with covariate adjustments compared with the one with no covariate adjustments in the subgingival site was 71/82 (i.e., about 87%). As in the α-diversity and β-diversity analyses, this may indicate that age, gender, and the frequency of brushing teeth have stronger confounding effects on the disparity in microbiome composition between EC users and non-users in the saliva site.

Mediation analysis

To assess whether the microbial taxa that are significantly affected by ECs [Fig. 5, Fig. 6] in turn cause gingival inflammation, we fitted mediation models [26] adjusting for EC use as well as the covariates age, gender, and the frequency of brushing teeth. We identified 1 microbial taxon (i.e., 1 species) in the saliva site [Fig. 7] and 18 microbial taxa (i.e., 1 phylum, 2 classes, 4 orders, 5 families, 4 genera and 2 species) in the subgingival site [Fig. 8] as microbial taxa that mediate the effects of ECs on gingival inflammation.

Fig. 7figure 7

The results from mediation analysis using saliva samples. The fitted generalized linear mixed model to assess the mediating roles of the saliva microbiome between EC smoking and gingival inflammation. *Est. represents the estimated coefficient on the effect of each taxon in the fitted generalized linear mixed model. *Q-value represents the P-value after the FDR control. *Only the statistically significant taxa after the FDR control (i.e., Q-value < 0.05) are included. For the species, the one before semi-colon (;) is the genus name and the one after semi-colon (;) is the species name

Fig. 8figure 8

The results from mediation analysis using subgingival samples. The fitted generalized linear mixed model to assess the mediating roles of the subgingival microbiome between EC smoking and gingival inflammation. *Est. represents the estimated coefficient on the effect of each taxon in the fitted generalized linear mixed model. *Q-value represents the P-value after the FDR control. *Only the statistically significant taxa after the FDR control (i.e., Q-value < 0.05) are included. For the species, the one before semi-colon (;) is the genus name and the one after semi-colon (;) is the species name

In the saliva site, EC smoking significantly decreased the relative abundance of Absconditabacteria_(SR1)_[G-1];bacterium_HMT_875 (species) [Fig. 5], and this decrease consequently resulted in gingival inflammation [Fig. 7].

In the subgingival site, EC smoking significantly decreased the relative abundance of Actinobacteria (phylum), Gammaproteobacteria and Betaproteobacteria (classes), Actinomycetales, Pasteurellales, Burkholderiales and Neisseriales (orders), Corynebacteriaceae, Micrococcaceae, Burkholderiaceae, Neisseriaceae and Steptococcaceae (families), Actinomyces, Rothia, Neisseria and Enterococcus (genera), and Bergeyella;sp._HMT_322 (species) [Fig. 6], and these decreases consequently resulted in gingival inflammation [Fig. 8]. In contrast, in the subgingival site, EC smoking significantly increases the relative abundance of Olsenella;uli, (species) [Fig. 6], and this increase consequently resulted in gingival inflammation [Fig. 8]. Interestingly, Olsenella;uli, (species) is a gram-positive bacterium that is known to cause endodontic infections [40].

Functional differential abundance analysis

We additionally assessed the disparity in each functional annotation (i.e., KEGG pathway [25]) between EC users and non-users while adjusting for age, gender, and the frequency of brushing teeth [Fig. 9]. We found significant regulations in relative abundance between EC users and non-users for 71 KEGG pathways in the subgingival site [Fig. 9], while no metabolic pathways were significantly regulated in the saliva site.

Fig. 9figure 9

The results from functional differential abundance analysis using subgingival samples. The fitted random effects model to assess the disparity in each functional annotation (i.e., KEGG pathway) between EC users and non-users while adjusting for age, gender and the frequency of brushing teeth. *Est. represents the estimated coefficient on the effect of each pathway in the fitted effects model. *Q-value represents the P-value after the FDR control. *Only the statistically significant pathways after the FDR control (i.e., Q-value < 0.05) are included

Here, we can exhibit an upregulation in metabolic pathways for obtaining energy observed in anoxic niches (e.g., fermentation) which has been confirmed in other studies of periodontal diseases [41, 42]. For example, we found an upregulation in the reductive acetyl coenzyme A pathway (CODH-PWY) [Fig. 9], which is used in anaerobic environments to fix carbon dioxide by forming acetyl-CoA that is then fermented into acetate [43]. We also detected other upregulated anaerobic metabolisms in EC users, specifically the breakdown of amino acids found in anoxic environments. The fermentation of lysine and glutamate is especially relevant to oral disease. These pathways that were observed to be upregulated in EC users are, for example, Fermentation of L-Lysine Produces Acetate and Butanoate (P163-PWY) and L-Glutamate Degradation V via hydroxylate (P162-PWY) [Fig. 9] that produce butanoate, which is an active signaling molecule in the oral cavity of the host that has been associated with inflammation, an important determinant of periodontitis [44,45,46]. Interestingly, our results also showed that EC users have upregulated functions observed exclusively in Archaeans. For example, there is increased formation of methane and carbon dioxide from acetate (acetyclastic) by upregulated anaerobic processes only observed in Archaeans (see methanogenesis from acetate (METH-ACETATE_PWY) [Fig. 9]) [47]. There are also synthetic pathways that are associated with increased population of these methanogens (e.g., see pyrimidine deoxyribonucleotides de novo biosynthesis IV (PWY-7198) [Fig. 9]). These results suggest that for EC users Archaeans are seen in combination with anaerobic fermentative bacteria in a type of symbiosis known as syntrophic association, which is represented in the functional pathways by an upregulation in both Archaeans-specific functions and fermentative metabolism [48,49,50]. In addition, we identified upregulated functional pathways that lead to the production of A-LPS, which is a specific O-antigen in the LPS from Porphyromonas gingivalis (e.g., see UDP-2,3-diacetamido-2,3-dideoxy-&alpha;-D-mannuronate biosynthesis (PWY-7090) [Fig. 9]), signaling the presence of this periodontal pathogen in EC users [51, 52]. Most downregulated functions were evidence of shutdown of biosynthesis in EC users (e.g., downregulated biosynthesis of amino acids, carbohydrates and aerobic processes), as there is evidence of a shift towards reductive chemistry to produce energy as described above. This is also corroborated by downregulated functions related to oxidative pathways, for example reduced TCA, glycolysis, fatty acid beta-oxidation, and aerobic respiration [Fig. 9].

留言 (0)

沒有登入
gif