Abstract:
Background:
Head and neck cancer is the sixth most common malignancy worldwide, with 90% of cases derived from distinct mucosal epithelium regions, collectively, known as head and neck squamous cell carcinoma (HNSCC). Early-stage diagnosis for HNSCC has an 80% five-year survival rate, that significantly drops to 35% when it becomes advanced. Recent studies have demonstrated distinct tumour microbiota within the tumour microenvironment in majority of cancers. These microbial signatures have been correlated with therapeutic responses and patient outcome. In HNSCC, majority of studies consist of small cohorts which identified varying bacteria composition within different tissue sample types (tumour, tumour-adjacent, healthy). However, these studies did not report a consensus microbial signature to differentiate these samples. The purpose of this meta-analysis is to determine if there is a consistent tumour microbiome signature that can be identified in HNSCC. These microbial signatures can be then used to predict microbial-host metagenome functions and further studies on the tumour microenvironment.
Method:
The tissue microbiome 16S rRNA amplicon sequences were obtained from NCBI and processed using QIIME2-DADA2. Data were batch-adjusted using sPLSDA-batch and analysed using multivariate (PERMANOVA, sPLSDA, ANOSIM) and univariate (ANOVA) discriminant analysis to differentiate between sample types. Post-hoc analysis was performed using unpaired welch-t test to compare difference in abundance between sample types. Functional prediction was performed using PICRUST2.
Results:
HNSCC tumour tissue samples displayed distinct microbial signature (PERMANOVA – R2 = 0.013, p < 0.001) compared to tumour-tumour adjacent and healthy tissue. Multivariate sPLSDA discriminant analysis identified 84 representative bacteria to differentiate the 3 sample types. Fusobacterium and Streptococcuswas most abundant in healthy and tumour tissues respectively, while Rothia was most abundant in tumour-adjacent tissues. Functional prediction identified acetyl-CoA fermentation to butanoate II and L-glutamate degradation V pathways were most enriched in cancer tissues.
Conclusion:
Distinct microbial signatures were found in cancer, tumour-adjacent and healthy HNSCC tissues. These microbial signatures displayed distinct functional enriched pathway which may influence the tumour microenvironment.