Salivary Microbiome and Cigarette Smoking: A First of Its Kind Investigation in Jordan

Abstract

There is accumulating evidence in the biomedical literature suggesting the role of smoking in increasing the risk of oral diseases including some oral cancers. Smoking alters microbial attributes of the oral cavity by decreasing the commensal microbial population and increasing the pathogenic microbes. This study aims to investigate the shift in the salivary microbiota between smokers and non-smokers in Jordan. Our methods relied on high-throughput next-generation sequencing (NGS) experiments for V3-V4 hypervariable regions of the 16S rRNA gene, followed by comprehensive bioinformatics analysis including advanced multidimensional data visualization methods and statistical analysis approaches. Six genera-Streptococcus, Prevotella, Vellionella, Rothia, Neisseria, and Haemophilus-predominated the salivary microbiota of all samples with different percentages suggesting the possibility for the salivary microbiome to restored after quitting smoking. Three genera-Streptococcus, Prevotella, and Veillonella-showed significantly elevated levels among smokers at the expense of Neisseria in non-smokers. In conclusion, smoking has a definite impact on shifting the salivary microbiota in smokers. We can suggest that there is microbial signature at the genera level that can be used to classify smokers and non-smokers by Linear Discriminant Analysis Effect Size (LEfSe) based on the salivary abundance of genera. Proteomics and metabolomics studies are highly recommended to fully understand the effect of bacterial endotoxin release and xenobiotic metabolism on the bacterial interrelationships in the salivary microbiome and how they affect the growth of each other in the saliva of smokers.

Keywords: 16S rRNA; Jordan; bioinformatics; microbiome; microbiota; next-generation sequencing; operational taxonomic unit (OTU); saliva; smoking.

Figure A1

Sequence counts per sample for raw, filtered, and classified sequences, ordered by increasing classified sequence counts. The dashed gold line indicates the 25,000 sequence threshold used for rarefaction.

Figure A2

Rarefication curve using filtered dataset.

Figure 1

Workflow for the salivary microbiome data generation, pre-processing, and analysis.

Figure 2

Boxplots of the three studied alpha diversity metrics: (A) Chao1 (community richness), (B) observed OTUs (community uniqueness), and (C) Shannon (community evenness or entropy). Red boxplots represent females and green boxplots represent males. Statistically significant differences were determined using analysis of variance with alpha diversity as the response variable, and smoking status and sex as crossed predictor variables, with Tukey’s HSD for group-specific differences. * Statistically significant at adjusted (p-value < 0.05), ** Statistically significant at adjusted (p-value < 0.01) and *** Statistically significant at adjusted (p-value < 0.001).

Figure 3

Principal Coordinates Analysis (PCoA) of the distance matrix generated using three distance metrics: (A) Bray-Curtis dissimilarity data, (B) Weighted UniFrac, and (C) Unweighted UniFrac. The x and y axes correspond to the first and second major principal coordinates (PC1 and PC2) identified from the PCoA analysis. Each principal coordinate explains a certain fraction of the variability (indicated by the percentage between brackets on each axis) observed in the data set. The principal coordinates PC1 and PC2 are plotted to create a visual two-dimensional (2D) representation of the multidimensional microbial community compositional differences between tested samples. Each sample is represented by a point and colored based on the smoking status and the sex of tested human subjects: Female non-smokers (red), male non-smokers (green), female smokers (teal green), and male smokers (magenta). The distance between the points represents the similarity of those samples (closer together = more similar).

Figure 4

Taxonomic composition represented by the abundance (percentage) of phyla per smoking status per gender for each sample in each comparison group.

Figure 5

The relative abundance of genera per sample group per smoking status and gender.

Figure 6

Boxplot for the three statistically impacted features identified by univariate analysis at the phylum level for non-smokers (red box) versus smokers (blue box) regardless of gender, statistically significant p-values with a significance level set at 0.05.

Figure 7

A plot of the LDA scores of the top 15 genera showing statistically significant differences between smokers and non-smokers. LDA scores on the x-axis and genera on the y-axis. The color-coding in the squares of the right side of the plot refers to the cumulative abundance of each genus in each binned group, where red means high cumulative abundance and blue means low cumulative abundance.

Figure 8

A plot of the LDA scores of the top 15 genera showing statistically significant differences between binned years of smoking. LDA scores on the x-axis, and genera on the y-axis. The color-coding in the squares of the right side of the plot refers to the cumulative abundance of each genus in each binned group, where red means high cumulative abundance and blue means low cumulative abundance.

Figure 9

A plot of the LDA scores of the top 15 genera showing statistically significant differences between binned numbers of cigarettes smoked by human subjects. LDA scores are on the x-axis and genera on the y-axis. The color-coding in the squares of the right side of the plot refers to the cumulative abundance of each genus in each binned group, where red means high cumulative abundance and blue means low cumulative abundance.