Abstract 5138: Comparative profiling for bacterial inhabitance in pancreatic ductal adenocarcinoma and matched adjacent normal tissues

Objective: Composition of resident microbes of the human body is associated with different disease manifestations. The microbiome in bodily fluids and different organs is altered with diverse health outcomes including cancer. Microbiome composition is a potential source for identifying biomarkers in cancers. Studies have shown the association of specific bacterial communities with the occurrence of pancreatic ductal adenocarcinoma (PDAC). However, investigations for the altered microbiota in an individual with the emergence of the disease are not studied. So, this study presents a comparative profiling for bacterial inhabitance in cancer tissues with that of matched adjacent normal tissues. This association will enable discovery of potential microbial biomarker(s) for PDAC. Methods: Fresh frozen cancer and matched adjacent normal (N= 20 each) tissues were collected from PDAC patients. Genomic DNA was extracted by MO BIO PowerSoil DNA Isolation kit and enriched using NEBNext Microbiome DNA Enrichment kit. Next-generation sequencing (NGS) method, pyrosequencing of 16S rRNA was used to assess the diversity of the tissue-associated microbiota. Operational Taxonomic Units (OTUs) were defined at >97% similarity followed by taxonomic classification using BLASTn against a curated GreenGenes/NCBI/RDP derived database (bTEFAP®). The Bray-Curtis coefficient was used to quantify the compositional dissimilarity between microbiome profiles. Statistical analysis was done using “R” and NCSS 2010. Alpha and beta diversity analyses were conducted to cluster samples by their microbiome profiles. Overall associations between PDAC and microbiome compositions were assessed based on weighted or unweighted UniFrac distances using PERMANOVA. Results: 16S taxonomic profiling reveal 10 and 11 dominant bacterial genera associated with normal and cancer lesion, respectively. Prominent genera that are only found in cancer lesions include Bacteroides, Enhydrobacter, Lautropia, Mycobacterium, and Phascolarctobacterium, whereas Fusobacterium, Acinetobacter, Propionibacterium, and Pseudomonas are the genera with high relative abundance in both cancer lesion and adjacent normal tissues. The most prevalent bacterial species in all observed samples are Propionibacterium acnes and Pseudomonas pseudoalcaligenes. We found Fusobacterium nucleatum in 4 samples (2 each in tumor and adjacent normal) in our preliminary work (10% prevalence). The PCoA multidimensional plot showed clustering of “tumor”