Fusobacterium nucleatum Load Correlates with KRAS Mutation and Sessile Serrated Pathogenesis in Colorectal Adenocarcinoma

Abstract

Fusobacterium nucleatum (Fn) has been frequently detected in colorectal cancer. A high load of Fn has been associated with subtypes of colorectal cancers, located in the proximal colon, exhibiting microsatellite instability-high (MSI-H), MLH1 promoter hypermethylation, the CpG island hypermethylation phenotype-high, or BRAF mutation in some studies. Although these features characterize the sessile serrated pathway (SSP) of colon cancers, other studies have shown that Fn infection is associated with KRAS mutations mainly characteristic of non-serrated neoplasia. It is also not clear at what point the association of Fn infection with these genomic alterations is established during colorectal carcinogenesis. Here we show that MSI-H, MLH1 hypermethylation, BRAF mutation or KRAS mutations were independently associated with Fn infection in colorectal cancer. On the other hand, increasing Fn copy number in tissues was associated with increased probability to exhibit MSI-H, MLH1 hypermethylation or BRAF mutations but not KRAS mutations in colorectal cancer. We also show that Fn load was significantly less than that of colorectal cancer and no association was detected between BRAF/KRAS mutations or MLH1 hypermethylation and Fn infection in adenomas. Our combined data suggest that increasing loads of Fn during and/or after adenomacarcinoma transition might promote SSP but not KRAS-driven colorectal carcinogenesis. Alternatively, Fn preferentially colonizes colorectal cancers with SSP and KRAS mutations but can expand more in colorectal cancers with SSP.

Significance: The authors demonstrated that Fn is enriched in colorectal cancers exhibiting the SSP phenotype, and in colorectal cancers carrying KRAS mutations. Fn infection should be considered as a candidate risk factor specific to colorectal cancers with the SSP phenotype and with KRAS mutations.

FIGURE 1

A, Comparison of Fn loads among subgroups of colorectal cancers. Copy number of Fn per nanogram of tumor DNA (log) among L/E-CRC (n = 129), MSS-CRC (n = 137), MSI-H-CRC (n = 38) and non-MSI-H-CRC (n = 266), MLH1 hypermethylated-CRC (n = 25) and non-MLH1 hypermethylated-CRC (n = 279), BRAF V600E-CRC (n = 25) and non-BRAFV600E-CRC (n = 279) and KRAS mutated-CRC (n = 101) and non-KRAS mutated-CRC (n = 203) were compared. Data are depicted in each boxplot. The thick horizontal line within each box represents the median copy number of Fn. Dots in each column represent maximum (top), mean (middle), and minimum (bottom) copy number of Fn. Dots in MSS, non-MSI-H, non-MLH1 hypermethylated, non-BRAFV600E, and non-KRAS mutated-CRC column represent outliers. B, Comparison of Fn loads among subgroups of colorectal cancer that were infected with Fn. Fn DNA content between Fn-infected L/E (n = 52), MSS-CRC (n = 34), MSI-H-CRC (n = 23), non-MSI-H-CRC (n = 86), MLH1 hypermethylated-CRC (n = 15), non-MLH1-hypermethylated-CRC (n = 94), BRAFV600E (n = 14) and non-BRAFV600E (n = 95) and KRAS mutated-CRC (n = 45) and non-KRAS mutated-CRC (n = 64) were compared using Kruskal–Wallis test. Each number represents the P value for each comparison. A P value that is less than 0.05 is considered significant.