Fusobacterium species are part of the gut microbiome in humans. Recent studies have identified over-representation of Fusobacterium in colorectal cancer (CRC) tissues but it is not yet clear whether this is pathogenic or simply an epiphenomenon. In this study, we evaluated the relationship between Fusobacterium status and molecular features in CRCs through quantitative real-time PCR in 149 CRC tissues, 89 adjacent normal appearing mucosae and 72 colonic mucosae from cancer-free individuals. Results were correlated with CpG island methylator phenotype (CIMP) status, microsatellite instability (MSI) and mutations in BRAF, KRAS, TP53, CHD7 and CHD8. Whole exome capture sequencing data were also available in 11 cases. Fusobacterium was detectable in 111/149 (74%) CRC tissues and heavily enriched in 9% (14/149) of the cases. As expected, Fusobacterium was also detected in normal appearing mucosae from both cancer and cancer-free individuals but the amount of bacteria was much lower compared to CRC tissues (a mean of 250-fold lower for Pan-fusobacterium). We found the Fusobacterium-high CRC group (FB-high) to be associated with CIMP positivity (p=0.001), TP53 wild type (p=0.015), hMLH1 methylation positivity (p=0.0028), MSI (p=0.018) and CHD7/8 mutation positivity (p=0.002). Among the 11 cases where whole exome sequencing data was available, two that were FB-high cases also had the highest number of somatic mutations (a mean of 736 per case in FB-high vs. 225 per case in all others). Taken together, our findings show that Fusobacterium enrichment is associated with specific molecular subsets of CRCs, offering support for a pathogenic role in CRC for this gut microbiome component

Malignant pleural mesothelioma (MPM) is a fatal thoracic malignancy, the epigenetics of which are poorly defined. We performed high-throughput methylation analysis covering 6,157 CpG islands in 20 MPMs and 20 lung adenocarcinomas. Newly identified genes were further analyzed in 50 MPMs and 56 adenocarcinomas via quantitative methylation-specific PCR. Targets of histone H3 lysine 27 trimethylation (H3K27me3) and genetic alterations were also assessed in MPM cells by chromatin immunoprecipitation arrays and comparative genomic hybridization arrays. An average of 387 genes (6.3%) and 544 genes (8.8%) were hypermethylated in MPM and adenocarcinoma, respectively. Hierarchical cluster analysis showed that the two malignancies have characteristic DNA methylation patterns, likely a result of different pathologic processes. In MPM, a separate subset of genes was silenced by H3K27me3 and could be reactivated by treatment with a histone deacetylase inhibitor alone. Integrated analysis of these epigenetic and genetic alterations revealed that only 11% of heterozygously deleted genes were affected by DNA methylation and/or H3K27me3 in MPMs. Among the DNA hypermethylated genes, three (TMEM30B, KAZALD1, and MAPK13) were specifically methylated only in MPM and could serve as potential diagnostic markers. Interestingly, a subset of MPM cases (4 cases, 20%) had very low levels of DNA methylation and substantially longer survival, suggesting that the epigenetic alterations are one mechanism affecting progression of this disease. Our findings show a characteristic epigenetic profile of MPM and uncover multiple distinct epigenetic abnormalities that lead to the silencing of tumor suppressor genes in MPM and could serve as diagnostic or prognostic targets. [Cancer Res 2009;69(23):9073-82]