Assessing the Effect of Smokeless Tobacco Consumption on Oral Microbiome in Healthy and Oral Cancer Patients

Abstract

Oral cancer is a globally widespread cancer that features among the three most prevalent cancers in India. The risk of oral cancer is elevated by factors such as tobacco consumption, betel-quid chewing, excessive alcohol consumption, unhygienic oral condition, sustained viral infections, and also due to dysbiosis in microbiome composition of the oral cavity. Here, we performed an oral microbiome study of healthy and oral cancer patients to decipher the microbial dysbiosis due to the consumption of smokeless-tobacco-based products and also revealed the tobacco-associated microbiome. The analysis of 196 oral microbiome samples from three different oral sites of 32 healthy and 34 oral squamous cell carcinoma (OSCC) patients indicated health status, site of sampling, and smokeless tobacco consumption as significant covariates associated with oral microbiome composition. Significant similarity in oral microbiome composition of smokeless-tobacco-consuming healthy samples and OSCC samples inferred the possible role of smokeless tobacco consumption in increasing inflammation-associated species in oral microbiome. Significantly higher abundance of Streptococcus was found to adequately discriminate smokeless-tobacco-non-consuming healthy samples from smokeless-tobacco-consuming healthy samples and contralateral healthy site of OSCC samples from the tumor site of OSCC samples. Comparative analysis of oral microbiome from another OSCC cohort also confirmed Streptococcus as a potential marker for healthy oral microbiome. Gram-negative microbial genera such as Prevotella, Capnocytophaga, and Fusobacterium were found to be differentially abundant in OSCC-associated microbiomes and can be considered as potential microbiome marker genera for oral cancer. Association with lipopolysaccharide (LPS) biosynthesis pathway further confirms the differential abundance of Gram-negative marker genera in OSCC microbiomes.

Keywords: dignostic biomarker; microbiome & dysbiosis; oral microbiome shift; oral squamos cell carcinoma; tobacco consumption.

Figure 1

Microbiome diversity in Healthy and OSCC samples. (A) Sampling sites of oral cavity considered in this study. For the healthy individuals, swab samples were collected from the right (abbreviated as “R”) and left (abbreviated as “L”) buccal site and the dental plaque or biofilm (abbreviated as “D”). For the OSCC patients, the swab samples were collected from the cancer lesion site or tumor site (abbreviated as “T”), its anatomically matched contralateral normal buccal site (abbreviated as “B”), and dental plaque or biofilm (abbreviated as “D”). (B) Number of samples used in this study. The healthy group consisted of 48 smokeless tobacco-consuming (TC-H) and 46 non-consuming (NTC-H) samples. The OSCC group consists of 72 smokeless-tobacco-consuming (TC-OSCC) and 30 non-consuming (NTC-OSCC) samples. (C) Principal coordinate analysis considering intersample Bray–Curtis distance between all 196 samples. Samples were tagged based on the health status. (D) Principal coordinate analysis considering intersample Bray–Curtis distance between 94 healthy samples. Samples were tagged based on the smokeless tobacco consumption status. (E) Principal coordinate analysis considering intersample Bray–Curtis distance between 94 healthy samples. Samples were tagged as left buccal site, right buccal site, and dental site. (F) Principal coordinate analysis considering intersample Bray–Curtis distance between 102 OSCC samples. Samples were tagged as tumor site, contralateral healthy site, and dental site. **p < 0.01; ns, not significant.

Figure 2

Analysis based on average intersample Bray–Curtis distance of samples from buccal and dental sites. (A) Comparison between intersample Bray–Curtis distance of oral microbiome (buccal site) of healthy smokeless tobacco non-consumers from healthy smokeless tobacco consumers, tumor site (T-site) of OSCC patients, and contralateral healthy site (B-site) of OSCC patients. (B) Intersample Bray–Curtis distance of oral microbiome from buccal site of healthy smokeless tobacco non-consumers, healthy smokeless tobacco consumers, tumor site (T-site) of OSCC patients, and contralateral healthy site (B-site) of OSCC patients. (C) Principal coordinate analysis considering intersample Bray–Curtis distance between buccal samples of healthy smokeless tobacco non-consumers, healthy smokeless tobacco-consumers, tumor site (T-site) of OSCC patients, and contralateral healthy site (B-site) of OSCC patients. (D) Comparison between intersample Bray–Curtis distance of dental microbiome of healthy smokeless tobacco non-consumers from healthy smokeless tobacco consumers and OSCC patients. **p < 0.01; ***p < 0.001; ns, not significant.

Figure 3

Differentially abundant bacterial phyla and core microbiome (genus level) of healthy and OSCC samples. (A) Differentially abundant bacterial phyla in healthy and OSCC samples identified using Boruta. All 196 samples were considered for this analysis. (B) Differentially abundant bacterial phyla in smokeless-tobacco-consuming and non-consuming healthy samples identified using Boruta. All 94 healthy samples were considered for this analysis. (C) Differentially abundant bacterial phyla in the tumor site and contralateral healthy site of OSCC samples identified using Boruta. All 102 OSCC samples were considered for this analysis. (D) Relative abundance of core genera with >1% abundance in healthy and OSCC samples (n=196). The significance levels were indicated based on Wilcoxon rank-sum test. (E) Relative abundance of core genera with >1% abundance in healthy smokeless-tobacco-consuming and non-consuming samples (n=94). The significance levels were indicated based on Wilcoxon rank-sum test. (F) Relative abundance of core genera with >1% abundance in the tumor site and contralateral healthy site of OSCC samples (n=102). The significance levels were indicated based on Wilcoxon rank-sum test.

Figure 4

Differentially abundant genera in healthy and OSCC oral microbiome. (A) Differentially abundant genera in healthy and OSCC samples and corresponding linear discriminant analysis (LDA) score using LEfSe. The discriminating genera reported by the analysis using Boruta were highlighted in blue color. (B) Differentially abundant genera in the tumor site and contralateral healthy buccal site of OSCC samples and corresponding LDA score using LEfSe. The discriminating genera reported by the analysis using Boruta were highlighted in blue color. (C) Differentially abundant genera in smokeless-tobacco-consuming and non-consuming healthy samples and corresponding LDA score using LEfSe. The discriminating genera reported by the analysis using Boruta were highlighted in blue color.

Figure 5

Differentially abundant species. (A) Relative abundance of core microbial species in smokeless-tobacco-consuming and non-consuming healthy samples, and tumor site and contralateral healthy buccal site of OSCC samples. The significance levels were indicated based on Wilcoxon rank-sum test. Species indicated by black filled dots are discriminating core microbial species identified by Boruta. (B) Differentially abundant core microbial species in smokeless-tobacco-consuming and non-consuming healthy samples, and tumor site and contralateral healthy buccal site of OSCC samples. LDA score using LEfSe was also indicated in this figure. **p < 0.01; ***p < 0.001.

Figure 6

Comparative analysis of Indian and Chinese oral microbiome of OSCC samples. (A) Principal coordinate analysis of Indian and Chinese OSCC samples based on intersample unweighted-UniFrac distance. (B) Relative abundance of core-oral microbiome (genus level) in Chinese samples. The significance levels were indicated based on Wilcoxon rank-sum test. (C) Differentially abundant microbial genera in the tumor site and healthy site of OSCC samples. LDA score using LEfSe were also indicated in this figure.

Figure 7

Schematic representation of the design and outcome of this study.