Importance
Evidence indicates a complex link between gut microbiome, immunity, and intestinal tumorigenesis. To target the microbiota and immunity for colorectal cancer prevention and therapy, a better understanding of the relationship between microorganisms and immune cells in the tumor microenvironment is needed. Experimental evidence suggests thatFusobacterium nucleatummay promote colonic neoplasia development by downregulating antitumor T cell–mediated adaptive immunity. Objective
To test the hypothesis that a greater amount ofF nucleatumin colorectal carcinoma tissue is associated with a lower density of T cells in tumor tissue. Design, Setting, and Participants
A cross-sectional analysis was conducted on 598 rectal and colon carcinoma cases in 2 US nationwide prospective cohort studies with follow-up through 2006, the Nurses’ Health Study (participants enrolled in 1976) and the Health Professionals Follow-up Study (participants enrolled in 1986). Tissue collection and processing were performed from 2002 through 2008, and immunity assessment, 2008 through 2009. From 2013 through 2014, the amount ofF nucleatumin colorectal carcinoma tissue was measured by quantitative polymerase chain reaction assay; we equally dichotomized positive cases (high vs low). Multivariable ordinal logistic regression analysis was conducted in 2014 to assess associations of the amount ofF nucleatumwith densities (quartiles) of T cells in tumor tissue, controlling for clinical and tumor molecular features, including microsatellite instability, CpG island methylator phenotype, long interspersed nucleotide element-1 (LINE-1) methylation, andKRAS,BRAF, andPIK3CAmutation status. We adjusted the 2-sided α level to .013 for multiple hypothesis testing. Main Outcomes and Measures
Densities of CD3+, CD8+, CD45RO (protein tyrosine phosphatase receptor type C [PTPRC])+, and FOXP3+T cells in tumor tissue, determined by means of tissue microarray immunohistochemical analysis and computer-assisted image analysis. Results
F nucleatumwas detected in colorectal carcinoma tissue in 76 (13%) of 598 cases. Compared withF nucleatum–negative cases,F nucleatum–high cases were inversely associated with the density of CD3+T cells (for a unit increase in quartile categories of CD3+T cells as an outcome: multivariable odds ratio, 0.47 [95% CI, 0.26-0.87];Pfor trend = .006). The amount ofF nucleatumwas not significantly associated with the density of CD8+, CD45RO+, or FOXP3+T cells (P fortrend = .24, .88, and .014, respectively). Conclusions and Relevance
The amount of tissueF nucleatumis inversely associated with CD3+T-cell density in colorectal carcinoma tissue. On validation, our human population data may provide an impetus for further investigations on potential interactive roles ofFusobacteriumand host immunity in colon carcinogenesis.