Abstract 4560: Fusobacterium nucleatum and tumor immunity status according to microsatellite instability in colorectal cancer

• Published in: Cancer research (Chicago, Ill.) Ill.), 2017-07, Vol.77 (13_Supplement), p.4560-4560
• Main Authors: Mima, Kosuke, Hamada, Tsuyoshi, Nishihara, Reiko, Nowak, Jonathan, Qian, Zhi Rong, Kosumi, Keisuke, Giannakis, Marios, Bullman, Susan, Baba, Hideo, Huttenhower, Curtis, Garrett, Wendy, Chan, Andrew, Meyerson, Matthew, Fuchs, Charles, Ogino, Shuji
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2017-4560

Objective: Fusobacterium nucleatum (F. nucleatum) is a microbial pathogen that contributes to the initiation and progression of colorectal cancer. Although evidence points to its immunosuppressive effects in colorectal cancer, F. nucleatum has been associated with high-level microsatellite instability (MSI) which is characterized by intense immune response. We hypothesized that the association of F. nucleatum with local immunosuppression might be stronger in tumors with high-level MSI or a high neoantigen load, which is a direct consequence of MSI that influences tumor-immune interactions. Design: We utilized molecular pathological epidemiology database of 1,027 colon and rectal cancer cases in the Nurses’ Health Study and the Health Professionals Follow-up Study. We assessed the association of F. nucleatum (negative, low, or high) with tumor immunity status in strata of MSl status. We measured F. nucleatum DNA in tumor tissue using quantitative PCR. As immunity status, we evaluated hystopathological lymphocytic reaction (tumor-infiltrating lymphocytes [TIL], intratumoral periglandular reaction, peritumoral lymphocytic reaction, or Crohn's-like lymphoid reaction) and measured T cell densities (CD3+, CD8+, CD45RO [PTPRC]+ or FOXP3+ cells) using immunohistochemistry and computer-assisted image analysis. Logistic regression model was used to adjust for potential confounders, including CpG island methylator phenotype, LINE-1 methylation level, and KRAS, BRAF, and PIK3CA mutations. Using whole-exome sequencing data (N = 586), we conducted secondary analyses to assess the interaction of F. nucleatum with a neoantigen load. Results: The association between F. nucleatum and TIL differed by MSI status (P for interaction = 0.0017; with the adjusted α level of 0.006). Compared with F. nucleaum-negative cases, F. nucleatum-positive cases were associated with low-level TIL in MSI-high tumors (odds ratio, 0.40 [95% confidence interval, 0.14-1.08] and 0.40 [95% confidence interval, 0.17-0.97] for F. nucleaum low and high cases, respectively), but not in microsatellite stable/MSI-low tumors. We did not find any interaction between F. nucleatum and MSI status in T cell subsets. The association between F. nucleatum and TIL did not differ significantly by levels of a neoantigen load (P for interaction = 0.015). Conclusions: The association of F. nucleatum with the immunosuppressive microenvironment may be stronger in colorectal cancer with high-level MSI. Our findings sugges