Background/Purpose: Behcet’s syndrome (BS) exhibits features of both innate and acquired immunity.   Oral ulceration is the cardinal lesion which usually precedes other manifestations.  Despite the presence of clues from clinical and basic research suggesting the role of an infectious trigger in pathogenesis, a specific microorganism has not been identified.  16S rRNA sequencing provides an advantage over conventional techniques as it is quantitative and not reliant on tissue culture.  This study aims to assess the microbiome in oral ulcer tissue in Turkish BS patients compared to non-lesional and healthy control mucosal tissue in parallel with pathogen recognition and pro-inflammatory cytokine pathways.  

Methods: 5mm punch biopsies were obtained from ulcerated oral tissue and the contralateral unaffected side in 16 BS patients and 5 ethnically matched controls.  Each sample was divided in half with16S rRNA sequencing of bacterial genomes performed in one half and mRNA-Seq performed in the other half.  RNA-Seq data was probed by gene set enrichment analysis (GSEA) for the following pathways important in pathogen recognition:  NOD-like receptor (NLR), Toll-like receptor (TLR), IL-1 receptor and IL-10 pathways. 

Results: Bacterial sequence taxonomic representation of 16S rRNA sequencing data by Jaccard indices did not segregate 6 ulcerated, 6 unaffected and 4 normal control tissue samples on the basis of taxonomic membership.  In contrast, a principal component analysis of RNA-Seq data showed the 16 ulcerated samples to cluster together with the 16 unaffected and 4 control samples clustering separately (See Figure.  Green – ulcer, blue – unaffected, red – control).  Of the 34,799 reads, 4595 transcripts were 2 fold or greater up-regulated in ulcerated compared to unaffected BS samples with 971 down-regulated greater than 2 fold.  GSEA of NLR, TLR, IL-1 receptor and IL-10 pathways all showed enrichment in ulcerated compared to unaffected samples with normalized enrichment scores of 1.45, 1.48, 1.39 and 1.36 and nominal p-values of 0.02, 0.02, 0.11 and 0.05, respectively. These pathways did not differ between unaffected and control samples.

Conclusion: Taken together, these results show that while bacterial patterns did not differ in BS oral ulcers compared to unaffected mucosa or controls, pathways of pathogen recognition were enriched in ulcerated BS tissue.  These pathways were not enriched in unaffected mucosa compared to controls.  This may be due to non-bacterial triggers or may indicate an abnormal tissue response to normal flora supporting the previous contention that several microorganisms may be using a common pathway to trigger a series of events.  Understanding these mechanisms is an area of active investigation.