Microbiome Signatures and Dysbiotic Patterns in Oral Cancer and Precancerous Lesions

Abstract

Background: The oral microbiome has been shown to be associated with the development of oral squamous cell carcinoma (OSCC). Research has primarily focused on elucidating the oncogenic mechanisms of specific pathogens by comparing the microbiomes of OSCC and normal tissues. However, the characteristics of the microbiome in the precancerous state remain less understood, as does the influence of metabolic and environmental factors on OSCC-associated microbiomes.

Methods: In this study, we analyzed mucosa-associated microbiomes in normal, precancerous, and OSCC lesions from a cohort of 51 patients using 16S rRNA amplicon sequencing. We investigated compositional changes in the microbiome, including the specific abundances and co-occurrences of OSCC-associated bacteria.

Results: Our findings indicate that the microbiome associated with precancerous lesions is indistinguishable from that of the normal mucosa, whereas the OSCC microbiome significantly differs from both normal and precancerous conditions. Specifically, the OSCC microbiome harbors less Streptococcus, coupled with an increase in amino-acid-degrading anaerobes such as Fusobacterium and Prevotella. The metabolic properties of individual microbes reported suggest that the overrepresentation of OSCC-specific bacteria is a result of metabolic adaptation to tumor microenvironments, although this possibility needs to be experimentally confirmed.

Conclusions: Our results demonstrate oral microbiome patterns across OSCC progression, offering insights into microbial ecological perspectives.

Keywords: 16S rRNA; Illumina sequencing; microbial ecology; microbial metabolism; oral microbiome; oral squamous cell carcinoma.

FIGURE 1

Microbiome composition and diversity on normal buccal mucosa, precancerous lesions, and OSCC. (a) Venn diagram of total observed species in three groups. (b) Alpha diversity (observed features and Shannon index) with p‐values determined by Mann–Whitney U test. (c) Relative abundance of the 20 most prevalent genera in each sample. Percentages were calculated by raw counts. All other minor genera were pooled in the “Others” category. (d) Heatmap showing unsupervised hierarchical clustering of samples and bacterial genera. Shades of color indicate values of log10 pseudocounts.

FIGURE 2

Dissimilarity of microbiomes between OSCC and normal/precancerous mucosa. (a, b) PCA (euclidean distance) and Bray–Curtis dissimilarity visualized with PCoA. p values were calculated using ANOSIM with 99,999 random permutations. (c) Confusion matrix of betel nut chewing versus disease status. (d) PCA plot of normal and precancerous samples of all subjects without betel nut usage.

FIGURE 3

Differentially abundant taxa between OSCC and normal/precancerous mucosa. (a) Linear discriminant analysis (LEfSe). Genus‐level taxa are marked by arrowheads. (b, c) Genera with the highest LDA scores identified from (a) that were enriched in OSCC (b) or normal/precancerous sites (c). p values were determined by the Mann–Whitney U test.

FIGURE 4

Co‐occurrence of differentially abundant bacteria indicative of OSCC. (a) Genus co‐occurrence analysis measured by Spearman's correlation. Only significant correlation values were shown (Benjamini–Hochberg adjusted p ≤ 0.05) in colors. Nonsignificant correlation values were set to 0, that is, white color. Black triangles indicate “Pro‐OSCC” genera that are inversely related to Streptococcus. (b, c) Score (b) and ROC curve (c) for the prediction of OSCC. The score was defined by the sum of read counts of the 10 “Pro‐OSCC” genera (Solobacterium, Catonella, Oribacterium, Atopobium, Prevotella, Capnocytophaga, Fusobacterium, Alloprevotella, Peptostreptococcus, and Dialister) divided by the count of Streptococcus. AUC, area under curve.