Microbiome composition and effects in tumors, with a focus on gastric carcinomas

Abstract

An estimated 15-20% of cancers worldwide are linked to viral, parasitic, or bacterial infections, and in 2008 these cases were responsible for 1.5 million cancer deaths worldwide. The best-studied examples of microbes contributing to cancer are hepatitis B virus, human papillomavirus, Epstein-Barr virus, human immunodeficiency virus, Schistosoma haematobium and Helicobacter pylori. Of these, H. pylori is the only bacteria considered to be a definite carcinogen. These microbes are primarily known to contribute to cancerous mutations through chronic inflammation. We hypothesized that other bacteria may also influence carcinogenesis through inflammation and other mechanisms. We sought to use the latest sequencing techniques and publicly available sequence data to identify novel bacteria in tumors and to further our understanding of the relationship between H. pylori and the gastric epithelium with a dual-species transcriptomics approach. During this process, sequence alignment techniques were optimized and are described here as best practices. Our analysis of publicly available cancer data led to the conclusion that Pseudomonas spp. and Acinetobacter spp. may be biologically relevant in gastric cancer samples and acute myeloid leukemia samples, respectively. Some bacterial species were found to be contaminant bacteria in each cancer type studied, while Mycobacterium spp. was identified as a contaminant in only the ovarian cancer samples sequenced at The Broad Institute. In our examination of H. pylori, the only known definite carcinogen, we identified differences in inflammatory pathways between tumor and adjacent samples both colonized with H. pylori. A co- transcriptomics approach was then used to determine the gastric epithelial response to H. pylori 26695 with and without cagE, a necessary gene for the type IV secretion system and secretion of the oncoprotein CagA. Multiple inflammatory pathways were activated in response to the H. pylori 26695 cagE+ compared to the cagE- strain. The bacterial response to the switch from broth to co-culture was also studied and both strains were determined to activate oipA and urease genes in response to gastric epithelial cells. Collectively, these findings advance our understanding of bacteria and their relationship to cancer.