The presence of Fusobacterium nucleatum (F. nucleatum) in colorectal carcinoma tissue has been associated with microsatellite instability (MSI), lower-level T-cell infiltrates, and poor clinical outcomes. Considering differences in the tumor–immune microenvironment between MSI-high and non–MSI-high carcinomas, we hypothesized that the association of F. nucleatum with immune response might differ by tumor MSI status. Using samples from 1,041 rectal and colon cancer patients within the Nurses’ Health Study and Health Professionals Follow-up Study, we measured F. nucleatum DNA in tumor tissue by a quantitative polymerase chain reaction assay. Multivariable logistic regression models were used to examine the association between F. nucleatum status and histopathologic lymphocytic reactions or density of CD3+ cells, CD8+ cells, CD45RO (PTPRC)+ cells, or FOXP3+ cells in strata of tumor MSI status. We adjusted for potential confounders, including CpG island methylator phenotype; LINE-1 methylation; and KRAS, BRAF, and PIK3CA mutations. The association of F. nucleatum with tumor-infiltrating lymphocytes (TIL) and intratumoral periglandular reaction differed by tumor MSI status (Pinteraction = 0.002). The presence of F. nucleatum was negatively associated with TIL in MSI-high tumors [multivariable odds ratio (OR), 0.45; 95% confidence interval (CI), 0.22–0.92], but positively associated with TIL in non–MSI-high tumors (multivariable OR 1.91; 95% CI, 1.12–3.25). No significant differential association was observed for peritumoral lymphocytic reaction, Crohn-like lymphoid reaction, or T-cell densities. In conclusion, the association of F. nucleatum with immune response to colorectal carcinoma differs by tumor MSI status, suggesting that F. nucleatum and MSI status interact to affect antitumor immune reactions. Cancer Immunol Res; 6(11); 1327–36. ©2018 AACR.

See related Spotlight on p. 1290.

Accumulating evidence continues to illustrate the importance of the host immune system in regulating the evolution and progression of neoplastic diseases (1, 2). The colorectum is rich in microbes, and the gut microbiome plays a role in regulating local and systemic immune status (3). Fusobacterium nucleatum (F. nucleatum) has emerged as a potentially influential bacterium implicated in colorectal carcinogenesis (4–7). Metagenomic and clinical studies have highlighted the enrichment of F. nucleatum in colorectal cancer tissue relative to adjacent normal epithelium (5–9), and its persistence in metastatic lesions (10). Experimental evidence indicates that F. nucleatum can elicit inflammatory reactions, impair T-cell–mediated antitumor immune response, and promote tumor cell proliferation (11–15). Studies have also demonstrated the association of F. nucleatum in colorectal cancer tissue with proximal tumor location, microsatellite instability (MSI), CpG island methylator phenotype (CIMP), lower levels of tumor-infiltrating CD3+ cells, and poor patient prognosis (8, 16–19).

Colorectal cancer consists of a heterogeneous collection of neoplasms that evolves through stepwise accumulation of genetic and epigenetic aberrations, a process influenced by environmental exposures, the microbiota, and host factors (20–22). MSI-high colorectal carcinomas typically contain numerous frameshift mutations and generate immunogenic peptides (“neoantigens”) due to mismatch-repair deficiency, which encourage the antitumor immune response and lead to more favorable prognosis and better response to immunotherapies (23–28). In fact, the FDA approved the use of anti-PDCD1 (programmed cell death 1, PD-1) antibodies pembrolizumab and nivolumab for colorectal cancer with high-level MSI or mismatch-repair deficiency. Considering the differences in the immune microenvironment between MSI-high and non–MSI-high tumors, we hypothesized that the association of F. nucleatum in colorectal carcinoma tissue with immune response might differ by tumor MSI status.

To test this hypothesis, we utilized a molecular pathological epidemiology (MPE) database of more than 1,000 colorectal carcinoma cases in two U.S. prospective cohort studies, with data on clinical, pathologic, and tumor molecular characteristics. This comprehensive data set enabled us to assess whether the association of F. nucleatum in colorectal cancer tissue with immune response to tumor might differ by MSI status while adjusting for a number of potential confounders.

Study population

We utilized two prospective cohort studies in the United States, the Nurses’ Health Study (NHS, 121,701 women ages 30–55 years followed since 1976) and the Health Professionals Follow-up Study (HPFS, 51,529 men ages 40–75 years followed since 1986). Every two years, participants have been sent follow-up questionnaires to report lifestyle factors and newly diagnosed diseases including colorectal cancer. The follow-up rate was >90% for each follow-up questionnaire in both cohort studies. In both cohorts, the National Death Index was used to ascertain deaths of participants and identify unreported lethal colorectal cancer cases. Study physicians, blinded to exposure data, reviewed medical records of identified colorectal cancer cases to confirm the disease diagnosis and to collect data on clinical characteristics (e.g., tumor location and disease stage).

Among participants diagnosed with colorectal cancer up to 2012, we analyzed 1,041 cases with available data on F. nucleatum and immune response in tumor tissue samples. We included both colon and rectal carcinomas, based on the colorectal continuum model (29, 30). We excluded patients who had been preoperatively treated. We collected formalin-fixed paraffin-embedded (FFPE) tumor tissue blocks from hospitals throughout the United States where colorectal cancer patients had undergone surgical resection. A single pathologist (S. Ogino), blinded to other data, reviewed hematoxylin and eosin–stained tissue sections from all collected blocks and recorded pathologic features. Tumor differentiation was categorized into well/moderate versus poor (>50% vs. ≤50% gland formation, respectively).

Written informed consent was obtained from all participants at enrollment in the NHS and HPFS. This study was conducted in accordance with the Declaration of Helsinki, and after approval by the institutional review boards at Harvard T.H. Chan School of Public Health and Brigham and Women's Hospital (Boston, MA).

Quantitative PCR

DNA was extracted from colorectal cancer tissue in archival FFPE tissue sections using QIAamp DNA FFPE Tissue Kit (Qiagen). As previously described and validated (8), we performed a quantitative PCR assay to measure the amount of F. nucleatum DNA in the tumor using custom TaqMan primer/probe sets (Applied Biosystems) for the nusG gene of F. nucleatum and the reference human gene SLCO2A1. Amplification and detection of DNA were performed using the StepOnePlus Real-Time PCR Systems (Applied Biosystems). Amounts of intratumoral F. nucleatum DNA in each specimen were calculated as a relative value normalized to amount of SLCO2A1 using the 2−ΔCt method [where ΔCt (cycle threshold) = “the mean Ct value of F. nucleatum” – “the mean Ct value of SLCO2A1”] (8). Each specimen was analyzed in duplicate for each target in a single batch, and the mean of the two Ct values was used for each target. In our previous validation study, the Ct values linearly decreased with the amount of input F. nucleatum DNA (r2 > 0.99), and the interassay coefficient of variation of the Ct values in five different batches was <1% for all targets (8). Tumors with any detectable F. nucleatum DNA were classified as F. nucleatum–positive, whereas all other tumors were classified as F. nucleatum–negative.

We performed a quantitative PCR assay to measure the amount of Bifidobacterium genus DNA in the tumor using custom TaqMan primer/probe sets (Applied Biosystems) for the 16S ribosomal RNA gene of Bifidobacterium genus and for the reference gene 16S (31). The primer and probe sequences for each TaqMan Gene-Expression Assay were as follows: Bifidobacterium forward primer, 5′-CGGGTGAGTAATGCGTGACC-3′; Bifidobacterium reverse primer, 5′-TGATAGGACGCGACCCCA-3′; Bifidobacterium FAM probe, 5′-CTCCTGGAAACGGGTG-3′; universal 16S forward primer, 5′-CGGTGAATACGTTCCCGG-3′; universal 16S reverse primer, 5′-TACGGCTACCTTGTTACGACTT-3′; and universal 16S FAM probe, 5′-CTTGTACACACCGCCCGTC-3′. We used 80 ng DNA in each reaction under the same reaction conditions as for F. nucleatum (8). Amounts of intratumoral Bifidobacterium genus DNA in each specimen were calculated as a relative value normalized to amounts of 16S using the 2−ΔCt method (8). Each specimen was analyzed in duplicate for each target in a single batch, and the mean of the two Ct values was used for each target. Tumors with any detectable Bifidobacterium genus DNA were classified as bifidobacteria-positive, whereas all other tumors were classified as bifidobacteria-negative.

MSI status and other molecular characteristics

MSI status was analyzed using 10 microsatellite markers (D2S123, D5S346, D17S250, BAT25, BAT26, BAT40, D18S55, D18S56, D18S67, and D18S487), as previously described (29). MSI-high was defined as presence of instability in ≥30% of the markers and non–MSI-high as instability in <30% of the markers. Using bisulfite-treated DNA, we determined methylation status of eight CIMP-specific promoters (CACNA1G, CDKN2A, CRABP1, IGF2, MLH1, NEUROG1, RUNX3, and SOCS1) and long interspersed nucleotide element-1 (LINE-1) as previously described (29). CIMP-high was defined as ≥6 methylated promoters of eight promoters, and CIMP-low/negative as 0–5 methylated promoters. PCR and pyrosequencing were performed for KRAS (codons 12, 13, 61, and 146), BRAF (codon 600), and PIK3CA (exons 9 and 20), as previously described (29). We constructed tissue microarrays of colorectal cancer cases with sufficient tissue materials, including up to four tumor cores from each case in one tissue microarray block (32). As previously described (33, 34), IHC for CD274 (PD-L1) and PDCD1LG2 (PDCD1 ligand 2, PD-L2) was performed using an anti-CD274 (clone MIH1; dilution, 1:50; eBioscience) and anti-PDCD1LG2 (clone 366C.9E5; dilution, 1:6,000; provided by the laboratory of G.J. Freeman), respectively. We assessed overall tumor CD274 expression as an ordinal scale of 0 to 4 by summing cytoplasmic expression score [absent (0), weak (1), moderate (2), or strong (3)] and membrane expression score [absent (0) or present (1)], and categorized CD274 as low (overall score of 0 to 1) or high (2 to 4; ref. 33). We calculated the percentage of tumor cells expressing PDCD1LG2 in the cytoplasm or membrane and grouped the data into bins of 0%–20%, 21%–50%, 51%–80%, or 81%–100% (34).

Lymphocytic reaction to colorectal cancer

Lymphocytic reaction to tumor was histopathologically evaluated on whole tissue sections, as previously described (23). Four components of lymphocytic reaction were examined, including tumor-infiltrating lymphocytes (TIL), intratumoral periglandular reaction, peritumoral lymphocytic reaction, and Crohn-like lymphoid reaction (Fig. 1). TIL was defined as lymphocytes on top of cancer cells in a tissue section. Intratumoral periglandular reaction was defined as lymphocytic reaction in tumor stroma within a tumor mass. Peritumoral lymphocytic reaction was defined as discrete lymphoid reaction surrounding a tumor mass. Crohn-like lymphoid reaction was defined as transmural lymphoid reaction. Each of the four components was graded as negative/low, intermediate, or high. A subset of cases (n = 398) were independently reviewed by a second pathologist (J.N. Glickman) with a good interobserver correlation as previously described (23). We constructed tissue microarrays and measured densities (cells/mm2) of CD3+ cells, CD8+ cells, CD45RO (an isoform of the PTPRC protein)+ cells, and FOXP3+ cells in colorectal cancer tissue based on IHC and image analysis using an automated scanning microscope and the Ariol image analysis system (Genetix) as previously described (24).

Figure 1.

Lymphocytic reaction to colorectal cancer. A, Crohn-like lymphoid reaction (arrows) and peritumoral lymphocytic reaction (asterisks; original magnification, 20×). B, Peritumoral lymphocytic reaction (asterisks; original magnification, 40×). C, Intratumoral periglandular reaction (asterisks; original magnification, 200×). D, Tumor-infiltrating lymphocytes (arrows; original magnification, 400×).

Figure 1.

Lymphocytic reaction to colorectal cancer. A, Crohn-like lymphoid reaction (arrows) and peritumoral lymphocytic reaction (asterisks; original magnification, 20×). B, Peritumoral lymphocytic reaction (asterisks; original magnification, 40×). C, Intratumoral periglandular reaction (asterisks; original magnification, 200×). D, Tumor-infiltrating lymphocytes (arrows; original magnification, 400×).

Close modal

Statistical analysis

Our primary hypothesis test assessed a statistical interaction between F. nucleatum status in colorectal cancer tissue (positive vs. negative) and MSI status (high vs. non-high) in relation to immune response status. We assessed statistical interactions using the Wald test for the cross-product in multivariable-adjusted logistic regression models. In our primary hypothesis testing on new discoveries, we used the α level of 0.005. We conducted all other assessments including evaluations of individual odds ratios (ORs) as secondary analyses, and interpreted the results cautiously in addition to use of the α level of 0.005. Multivariable logistic regression models initially included age (continuous), sex, year of diagnosis (continuous), family history of colorectal cancer (absent vs. present), tumor location (proximal colon vs. distal colon vs. rectum), CIMP status (high vs. low/negative), LINE-1 methylation level (continuous), KRAS (mutant vs. wild-type), BRAF (mutant vs. wild-type), and PIK3CA (mutant vs. wild-type). A backward elimination with a threshold P of 0.05 was used to select variables for the final models. Cases with missing data [family history of colorectal cancer (0.8%), tumor location (0.3%), CIMP status (5.4%), KRAS (7.7%), BRAF (1.1%), and PIK3CA (6.6%)] were included in the majority category of a given categorical covariate to limit the degrees of freedom of the models. For cases with missing data on LINE-1 methylation level (2.6%), we assigned a separate indicator variable. We confirmed that the exclusion of cases with missing data in any of the variables did not substantially alter our results. The proportional odds assumption was assessed using the ordinal logistic regression model. We observed evidence of violation of this assumption in three of the four lymphocytic reaction patterns (TIL, intratumoral periglandular reaction, and peritumoral lymphocytic reaction) and T-cell quartile variables. We therefore used binary histologic lymphocytic reaction variables and T-cell variables dichotomized at the median value as outcome variables in logistic regression analyses.

In secondary analyses, we assessed a statistical interaction between F. nucleatum status in colorectal cancer tissue (positive vs. negative) and MSI status (high vs. non-high) in relation to colorectal cancer–specific or overall mortality. We assessed statistical interactions using the Wald test for the cross-product in multivariable-adjusted Cox proportional hazards regression models. Multivariable Cox regression models initially included tumor differentiation (well-moderate vs. poor) and disease stage (I/II vs. III/IV) in addition to the same set of covariates as in the multivariable logistic regression models. A backward elimination with a threshold P of 0.05 was used to select variables for the final models. Cases with missing data on tumor differentiation (0.1%) were included in the majority category, and the other covariates were dealt as in the multivariable logistic regression models. We confirmed that the exclusion of cases with missing data in any of the variables did not substantially alter our results. The assumption of proportionality of hazards was generally satisfied using the assessment of a time-varying covariate, that is, the cross-product of F. nucleatum status and survival time in strata of MSI status (P > 0.2). Cumulative survival probabilities were estimated using the Kaplan–Meier method and compared using the log-rank test. For colorectal cancer–specific mortality, deaths from other causes were treated as censored events. Using a single regression model with reparameterization of an interaction term (32), we calculated ORs for immune response status and hazard ratios for colorectal cancer mortality comparing F. nucleatum–positive to F. nucleatum–negative tumors in the two strata of MSI status.

All statistical analyses were performed using SAS software (version 9.4; SAS Institute), and all P values were two-sided.

Among the 1,041 colorectal carcinoma cases from the two prospective cohort studies, F. nucleatum DNA was detected using the quantitative PCR assay in 135 (13%) cases. Table 1 shows clinical, pathologic, and molecular features of colorectal cancer cases according to combined tumor tissue F. nucleatum and MSI status. As previously reported (8), MSI-high tumors were associated with detectable F. nucleatum DNA. There was a trend toward a higher rate of detectable intratumor F. nucleatum in more recent cases, particularly for MSI-high tumors. High persistence of F. nucleatum has been observed in metastatic liver lesions from F. nucleatum–positive colorectal cancer (10). We examined the correlations between F. nucleatum and disease stage in strata of tumor MSI status. We observed a significant positive correlation between F. nucleatum and disease stage in non–MSI-high tumors (P = 0.004, with the α level of 0.005), but not in MSI-high tumors (P = 0.34). However, statistical power was limited in the analysis of MSI-high tumors.

Table 1.

Clinical, pathologic, and molecular characteristics of patients with colorectal cancer and their samples, according to F. nucleatum status in tumor tissue, stratified by MSI status

Non–MSI-highMSI-high
F. nucleatum in tumorF. nucleatum in tumor
All casesaNegativeaPositiveaNegativeaPositivea
Characteristic(n = 1,041)(n = 787)(n = 77)(n = 119)(n = 58)
Sex 
 Female (NHS) 609 (59%) 437 (56%) 42 (55%) 85 (71%) 45 (78%) 
 Male (HPFS) 432 (41%) 350 (44%) 35 (45%) 34 (29%) 13 (22%) 
Mean age ± SD (years) 69.5 ± 8.9 68.9 ± 9.0 69.5 ± 9.6 72.1 ± 7.6 72.5 ± 8.4 
Race/ethnicity 
 White 995 (97%) 751 (97%) 74 (99%) 114 (97%) 56 (98%) 
 Black 16 (1.6%) 13 (1.7%) 2 (1.7%) 1 (1.8%) 
 Other 11 (1.1%) 8 (1.0%) 1 (1.3%) 2 (1.7%) 
Year of diagnosis 
 1995 or before 343 (33%) 285 (36%) 21 (27%) 29 (24%) 8 (14%) 
 1996–2000 286 (27%) 216 (27%) 25 (32%) 33 (28%) 12 (21%) 
 2001–2012 412 (40%) 286 (36%) 31 (40%) 57 (48%) 38 (66%) 
Family history of colorectal cancer 
In first-degree relative(s) 
 Absent 829 (80%) 632 (81%) 66 (88%) 84 (71%) 47 (81%) 
 Present 204 (20%) 150 (19%) 9 (12%) 34 (29%) 11 (19%) 
Prediagnosis body mass indexb 
 <25 kg/m2 417 (40%) 311 (40%) 30 (39%) 49 (42%) 27 (47%) 
 25–29.9 kg/m2 424 (41%) 330 (42%) 35 (46%) 43 (36%) 16 (28%) 
 ≥30 kg/m2 194 (19%) 142 (18%) 11 (14%) 26 (22%) 15 (26%) 
Prudent dietary patternc 
 Quartile 1 (lowest) 258 (25%) 187 (24%) 25 (33%) 30 (25%) 16 (28%) 
 Quartile 2 258 (25%) 206 (26%) 19 (25%) 21 (18%) 12 (21%) 
 Quartile 3 257 (25%) 187 (24%) 21 (28%) 32 (27%) 17 (29%) 
 Quartile 4 (highest) 258 (25%) 199 (26%) 11 (14%) 35 (30%) 13 (22%) 
Western dietary patternc 
 Quartile 1 (lowest) 258 (25%) 198 (25%) 10 (13%) 31 (26%) 19 (33%) 
 Quartile 2 258 (25%) 192 (25%) 21 (28%) 32 (27%) 13 (22%) 
 Quartile 3 258 (25%) 200 (26%) 20 (26%) 21 (18%) 17 (29%) 
 Quartile 4 (highest) 258 (25%) 190 (24%) 25 (33%) 34 (29%) 9 (16%) 
Tumor location 
 Proximal colon 522 (50%) 332 (42%) 33 (43%) 104 (87%) 53 (91%) 
 Distal colon 297 (29%) 264 (34%) 20 (26%) 10 (8.4%) 3 (5.2%) 
 Rectum 219 (21%) 189 (24%) 23 (30%) 5 (4.2%) 2 (3.5%) 
Tumor differentiation 
 Well to moderate 936 (90%) 745 (95%) 67 (88%) 86 (72%) 38 (66%) 
 Poor 104 (10%) 42 (5.3%) 9 (12%) 33 (28%) 20 (34%) 
AJCC disease stage 
 I 239 (25%) 197 (27%) 8 (11%) 25 (22%) 9 (17%) 
 II 314 (33%) 199 (28%) 21 (30%) 64 (56%) 30 (56%) 
 III 280 (29%) 222 (31%) 28 (39%) 18 (16%) 12 (22%) 
 IV 127 (13%) 103 (14%) 14 (20%) 7 (6.1%) 3 (5.6%) 
CIMP status 
 CIMP-low/negative 794 (81%) 688 (92%) 68 (94%) 24 (21%) 14 (25%) 
 CIMP-high 191 (19%) 56 (7.5%) 4 (5.6%) 89 (79%) 42 (75%) 
Mean LINE-1 methylation level ± SD (%) 63.6 ± 10.4 62.6 ± 10.2 62.2 ± 9.4 68.4 ± 10.3 70.5 ± 9.5 
KRAS mutation 
 Wild-type 541 (56%) 390 (52%) 30 (41%) 88 (85%) 33 (82%) 
 Mutant 420 (44%) 353 (48%) 44 (59%) 16 (15%) 7 (18%) 
BRAF mutation 
 Wild-type 858 (83%) 709 (91%) 73 (95%) 55 (47%) 21 (37%) 
 Mutant 172 (17%) 70 (9.0%) 4 (5.2%) 62 (53%) 36 (63%) 
PIK3CA mutation 
 Wild-type 812 (84%) 617 (84%) 59 (81%) 91 (84%) 45 (85%) 
 Mutant 160 (16%) 121 (16%) 14 (19%) 17 (16%) 8 (15%) 
Tumor CD274 (PD-L1) expression 
 Low 263 (41%) 176 (37%) 29 (52%) 37 (46%) 21 (64%) 
 High 377 (59%) 294 (63%) 27 (48%) 44 (54%) 12 (36%) 
Tumor PDCD1LG2 (PD-L2) expression 
 0%–20% 160 (25%) 111 (24%) 19 (37%) 22 (28%) 8 (25%) 
 21%–50% 148 (23%) 110 (23%) 11 (21%) 17 (21%) 10 (31%) 
 51%–80% 197 (31%) 150 (32%) 12 (23%) 25 (31%) 10 (31%) 
 81%–100% 129 (20%) 99 (21%) 10 (19%) 16 (20%) 4 (13%) 
Bifidobacterium genus DNA 
 Negative 748 (73%) 576 (75%) 54 (71%) 82 (69%) 36 (63%) 
 Positive 276 (27%) 197 (25%) 22 (29%) 36 (31%) 21 (37%) 
Non–MSI-highMSI-high
F. nucleatum in tumorF. nucleatum in tumor
All casesaNegativeaPositiveaNegativeaPositivea
Characteristic(n = 1,041)(n = 787)(n = 77)(n = 119)(n = 58)
Sex 
 Female (NHS) 609 (59%) 437 (56%) 42 (55%) 85 (71%) 45 (78%) 
 Male (HPFS) 432 (41%) 350 (44%) 35 (45%) 34 (29%) 13 (22%) 
Mean age ± SD (years) 69.5 ± 8.9 68.9 ± 9.0 69.5 ± 9.6 72.1 ± 7.6 72.5 ± 8.4 
Race/ethnicity 
 White 995 (97%) 751 (97%) 74 (99%) 114 (97%) 56 (98%) 
 Black 16 (1.6%) 13 (1.7%) 2 (1.7%) 1 (1.8%) 
 Other 11 (1.1%) 8 (1.0%) 1 (1.3%) 2 (1.7%) 
Year of diagnosis 
 1995 or before 343 (33%) 285 (36%) 21 (27%) 29 (24%) 8 (14%) 
 1996–2000 286 (27%) 216 (27%) 25 (32%) 33 (28%) 12 (21%) 
 2001–2012 412 (40%) 286 (36%) 31 (40%) 57 (48%) 38 (66%) 
Family history of colorectal cancer 
In first-degree relative(s) 
 Absent 829 (80%) 632 (81%) 66 (88%) 84 (71%) 47 (81%) 
 Present 204 (20%) 150 (19%) 9 (12%) 34 (29%) 11 (19%) 
Prediagnosis body mass indexb 
 <25 kg/m2 417 (40%) 311 (40%) 30 (39%) 49 (42%) 27 (47%) 
 25–29.9 kg/m2 424 (41%) 330 (42%) 35 (46%) 43 (36%) 16 (28%) 
 ≥30 kg/m2 194 (19%) 142 (18%) 11 (14%) 26 (22%) 15 (26%) 
Prudent dietary patternc 
 Quartile 1 (lowest) 258 (25%) 187 (24%) 25 (33%) 30 (25%) 16 (28%) 
 Quartile 2 258 (25%) 206 (26%) 19 (25%) 21 (18%) 12 (21%) 
 Quartile 3 257 (25%) 187 (24%) 21 (28%) 32 (27%) 17 (29%) 
 Quartile 4 (highest) 258 (25%) 199 (26%) 11 (14%) 35 (30%) 13 (22%) 
Western dietary patternc 
 Quartile 1 (lowest) 258 (25%) 198 (25%) 10 (13%) 31 (26%) 19 (33%) 
 Quartile 2 258 (25%) 192 (25%) 21 (28%) 32 (27%) 13 (22%) 
 Quartile 3 258 (25%) 200 (26%) 20 (26%) 21 (18%) 17 (29%) 
 Quartile 4 (highest) 258 (25%) 190 (24%) 25 (33%) 34 (29%) 9 (16%) 
Tumor location 
 Proximal colon 522 (50%) 332 (42%) 33 (43%) 104 (87%) 53 (91%) 
 Distal colon 297 (29%) 264 (34%) 20 (26%) 10 (8.4%) 3 (5.2%) 
 Rectum 219 (21%) 189 (24%) 23 (30%) 5 (4.2%) 2 (3.5%) 
Tumor differentiation 
 Well to moderate 936 (90%) 745 (95%) 67 (88%) 86 (72%) 38 (66%) 
 Poor 104 (10%) 42 (5.3%) 9 (12%) 33 (28%) 20 (34%) 
AJCC disease stage 
 I 239 (25%) 197 (27%) 8 (11%) 25 (22%) 9 (17%) 
 II 314 (33%) 199 (28%) 21 (30%) 64 (56%) 30 (56%) 
 III 280 (29%) 222 (31%) 28 (39%) 18 (16%) 12 (22%) 
 IV 127 (13%) 103 (14%) 14 (20%) 7 (6.1%) 3 (5.6%) 
CIMP status 
 CIMP-low/negative 794 (81%) 688 (92%) 68 (94%) 24 (21%) 14 (25%) 
 CIMP-high 191 (19%) 56 (7.5%) 4 (5.6%) 89 (79%) 42 (75%) 
Mean LINE-1 methylation level ± SD (%) 63.6 ± 10.4 62.6 ± 10.2 62.2 ± 9.4 68.4 ± 10.3 70.5 ± 9.5 
KRAS mutation 
 Wild-type 541 (56%) 390 (52%) 30 (41%) 88 (85%) 33 (82%) 
 Mutant 420 (44%) 353 (48%) 44 (59%) 16 (15%) 7 (18%) 
BRAF mutation 
 Wild-type 858 (83%) 709 (91%) 73 (95%) 55 (47%) 21 (37%) 
 Mutant 172 (17%) 70 (9.0%) 4 (5.2%) 62 (53%) 36 (63%) 
PIK3CA mutation 
 Wild-type 812 (84%) 617 (84%) 59 (81%) 91 (84%) 45 (85%) 
 Mutant 160 (16%) 121 (16%) 14 (19%) 17 (16%) 8 (15%) 
Tumor CD274 (PD-L1) expression 
 Low 263 (41%) 176 (37%) 29 (52%) 37 (46%) 21 (64%) 
 High 377 (59%) 294 (63%) 27 (48%) 44 (54%) 12 (36%) 
Tumor PDCD1LG2 (PD-L2) expression 
 0%–20% 160 (25%) 111 (24%) 19 (37%) 22 (28%) 8 (25%) 
 21%–50% 148 (23%) 110 (23%) 11 (21%) 17 (21%) 10 (31%) 
 51%–80% 197 (31%) 150 (32%) 12 (23%) 25 (31%) 10 (31%) 
 81%–100% 129 (20%) 99 (21%) 10 (19%) 16 (20%) 4 (13%) 
Bifidobacterium genus DNA 
 Negative 748 (73%) 576 (75%) 54 (71%) 82 (69%) 36 (63%) 
 Positive 276 (27%) 197 (25%) 22 (29%) 36 (31%) 21 (37%) 

Abbreviations: AJCC, American Joint Committee on Cancer; CIMP, CpG island methylator phenotype; LINE-1, long interspersed nucleotide element-1; SD, standard deviation.

aPercentage indicates the proportion of cases with a specific clinical, pathologic, or molecular characteristic in all cases or in strata of F. nucleatum and MSI status. Total percentages may not equal 100% due to rounding.

bCalculated by dividing weight (kilogram) by squared height (meter).

cThe prudent dietary pattern is characterized by high intake of vegetables, fruits, fish, poultry, and whole grains, and the Western dietary pattern is characterized by high intake of red and processed meats, added sugar, and refined grains (42).

Table 2 shows the distribution of colorectal carcinoma cases according to F. nucleatum status and histologic lymphocytic reaction patterns in strata of MSI status. The presence of F. nucleatum in tumor tissue was negatively correlated with intratumoral periglandular reaction in MSI-high tumors (P = 0.002, with the α level of 0.005) and was positively correlated with Crohn-like lymphoid reaction in non–MSI-high tumors (P = 0.001).

Table 2.

Distribution of colorectal cancer samples according to F. nucleatum status in tumor tissue and lymphocytic reaction status, stratified by MSI status

F. nucleatum in colorectal cancer tissue
All casesaNegativeaPositiveaPb
Non–MSI-high TILs (n = 862)    0.013 
  Negative/low 699 (81%) 644 (82%) 55 (71%)  
  Intermediate 119 (14%) 107 (14%) 12 (16%)  
  High 44 (5.1%) 34 (4.3%) 10 (13%)  
 Intratumoral periglandular reaction (n = 863)    0.57 
  Negative/low 124 (14%) 106 (13%) 18 (23%)  
  Intermediate 660 (76%) 613 (78%) 47 (61%)  
  High 79 (9.2%) 67 (8.5%) 12 (16%)  
 Peritumoral lymphocytic reaction (n = 859)    0.89 
  Negative/low 138 (16%) 123 (16%) 15 (20%)  
  Intermediate 615 (72%) 567 (72%) 48 (63%)  
  High 106 (12%) 93 (12%) 13 (17%)  
 Crohn-like lymphoid reaction (n = 723)    0.001 
  Negative/low 586 (81%) 544 (83%) 42 (66%)  
  Intermediate 111 (15%) 92 (14%) 19 (30%)  
  High 26 (3.6%) 23 (3.5%) 3 (4.7%)  
MSI-high TILs (n = 177)    0.055 
  Negative/low 46 (26%) 25 (21%) 21 (36%)  
  Intermediate 64 (36%) 45 (38%) 19 (33%)  
  High 67 (38%) 49 (41%) 18 (31%)  
 Intratumoral periglandular reaction (n = 177)    0.002 
  Negative/low 12 (6.8%) 1 (0.8%) 11 (19%)  
  Intermediate 100 (56%) 68 (57%) 32 (55%)  
  High 65 (37%) 50 (42%) 15 (26%)  
 Peritumoral lymphocytic reaction (n = 177)    0.062 
  Negative/low 13 (7.3%) 4 (3.4%) 9 (16%)  
  Intermediate 91 (51%) 62 (52%) 29 (50%)  
  High 73 (41%) 53 (45%) 20 (34%)  
 Crohn-like lymphoid reaction (n = 151)    0.13 
  Negative/low 64 (42%) 41 (41%) 23 (44%)  
  Intermediate 48 (32%) 26 (26%) 22 (42%)  
  High 39 (26%) 32 (32%) 7 (13%)  
F. nucleatum in colorectal cancer tissue
All casesaNegativeaPositiveaPb
Non–MSI-high TILs (n = 862)    0.013 
  Negative/low 699 (81%) 644 (82%) 55 (71%)  
  Intermediate 119 (14%) 107 (14%) 12 (16%)  
  High 44 (5.1%) 34 (4.3%) 10 (13%)  
 Intratumoral periglandular reaction (n = 863)    0.57 
  Negative/low 124 (14%) 106 (13%) 18 (23%)  
  Intermediate 660 (76%) 613 (78%) 47 (61%)  
  High 79 (9.2%) 67 (8.5%) 12 (16%)  
 Peritumoral lymphocytic reaction (n = 859)    0.89 
  Negative/low 138 (16%) 123 (16%) 15 (20%)  
  Intermediate 615 (72%) 567 (72%) 48 (63%)  
  High 106 (12%) 93 (12%) 13 (17%)  
 Crohn-like lymphoid reaction (n = 723)    0.001 
  Negative/low 586 (81%) 544 (83%) 42 (66%)  
  Intermediate 111 (15%) 92 (14%) 19 (30%)  
  High 26 (3.6%) 23 (3.5%) 3 (4.7%)  
MSI-high TILs (n = 177)    0.055 
  Negative/low 46 (26%) 25 (21%) 21 (36%)  
  Intermediate 64 (36%) 45 (38%) 19 (33%)  
  High 67 (38%) 49 (41%) 18 (31%)  
 Intratumoral periglandular reaction (n = 177)    0.002 
  Negative/low 12 (6.8%) 1 (0.8%) 11 (19%)  
  Intermediate 100 (56%) 68 (57%) 32 (55%)  
  High 65 (37%) 50 (42%) 15 (26%)  
 Peritumoral lymphocytic reaction (n = 177)    0.062 
  Negative/low 13 (7.3%) 4 (3.4%) 9 (16%)  
  Intermediate 91 (51%) 62 (52%) 29 (50%)  
  High 73 (41%) 53 (45%) 20 (34%)  
 Crohn-like lymphoid reaction (n = 151)    0.13 
  Negative/low 64 (42%) 41 (41%) 23 (44%)  
  Intermediate 48 (32%) 26 (26%) 22 (42%)  
  High 39 (26%) 32 (32%) 7 (13%)  

aPercentage indicates the proportion of cases with a specific level of lymphocytic reaction in strata of MSI status or in strata of F. nucleatum and MSI status. Total percentages may not equal 100% due to rounding.

bP value was calculated using the Spearman correlation test between F. nucleatum status (positive vs. negative) and each lymphocytic reaction variable [ordinal; negative/low (0), intermediate (1), and high (2)].

In our primary hypothesis test using logistic regression models (Table 3; Supplementary Table S1), we observed a statistically significant interaction between F. nucleatum and MSI status in relation to TIL or intratumoral periglandular reaction (Pinteraction = 0.002, with the α level of 0.005). The presence of F. nucleatum in tumor tissue was negatively associated with degrees of TIL in MSI-high colorectal cancer [multivariable OR, 0.45; 95% confidence interval (CI), 0.22–0.92], but positively associated with TIL in non–MSI-high cancer (multivariable OR, 1.91; 95% CI, 1.12–3.25). Similarly, the presence of F. nucleatum in tumor tissue was negatively associated with intratumoral periglandular reaction in MSI-high tumors (multivariable OR, 0.43; 95% CI, 0.21–0.87), but positively associated with intratumoral periglandular reaction (multivariable OR, 1.97, 95% CI, 1.00–3.86) in non–MSI-high tumors. The statistical interaction of F. nucleatum and MSI status in relation to peritumoral lymphocytic reaction or Crohn-like lymphoid reaction was not significant with the α level of 0.005 (Pinteraction > 0.01).

Table 3.

Logistic regression analyses to assess the association of F. nucleatum status in colorectal cancer tissue with lymphocytic reaction status, in strata of MSI status

UnivariableMultivariable
OR (95% CI)OR (95% CI)a
Model for TILsb (as an outcome variable; n = 1,039)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.83 (1.08–3.09) 1.91 (1.12–3.25) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.47 (0.23–0.94) 0.45 (0.22–0.92) 
Pinteractionc  0.002 0.002 
Model for intratumoral periglandular reactionb (as an outcome variable; n = 1,040)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.98 (1.02–3.85) 1.97 (1.00–3.86) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.48 (0.24–0.96) 0.43 (0.21–0.87) 
Pinteractionc  0.004 0.002 
Model for peritumoral lymphocytic reactionb (as an outcome variable; n = 1,036)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.53 (0.81–2.89) 1.44 (0.76–2.76) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.66 (0.34–1.26) 0.54 (0.27–1.05) 
Pinteractionc  0.068 0.038 
Model for Crohn-like lymphoid reactionb (as an outcome variable; n = 874)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 2.48 (1.42–4.31) 2.86 (1.62–5.06) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.89 (0.45–1.76) 0.99 (0.49–1.98) 
Pinteractionc  0.022 0.020 
UnivariableMultivariable
OR (95% CI)OR (95% CI)a
Model for TILsb (as an outcome variable; n = 1,039)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.83 (1.08–3.09) 1.91 (1.12–3.25) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.47 (0.23–0.94) 0.45 (0.22–0.92) 
Pinteractionc  0.002 0.002 
Model for intratumoral periglandular reactionb (as an outcome variable; n = 1,040)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.98 (1.02–3.85) 1.97 (1.00–3.86) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.48 (0.24–0.96) 0.43 (0.21–0.87) 
Pinteractionc  0.004 0.002 
Model for peritumoral lymphocytic reactionb (as an outcome variable; n = 1,036)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 1.53 (0.81–2.89) 1.44 (0.76–2.76) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.66 (0.34–1.26) 0.54 (0.27–1.05) 
Pinteractionc  0.068 0.038 
Model for Crohn-like lymphoid reactionb (as an outcome variable; n = 874)   
 Non–MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 2.48 (1.42–4.31) 2.86 (1.62–5.06) 
 MSI-high F. nucleatum-negative 1 (referent) 1 (referent) 
 F. nucleatum-positive 0.89 (0.45–1.76) 0.99 (0.49–1.98) 
Pinteractionc  0.022 0.020 

aThe multivariable logistic regression model initially included sex, age at diagnosis, year of diagnosis, family history of colorectal cancer, tumor location, CpG island methylator phenotype-specific promoter status, long interspersed nucleotide element-1 methylation level, and KRAS, BRAF, and PIK3CA mutations. A backward elimination with a threshold P of 0.05 was used to select variables for the final models. The variables that remained in the final models for TILs and intratumoral periglandular reaction were shown in Supplementary Table S1.

bTo avoid violation of the proportional odds assumption, the most common score or less (i.e., 0 for TILs and Crohn-like lymphoid reaction; and 0–1 for intratumoral periglandular reaction and peritumoral lymphocytic reaction) was categorized as low, and higher score was categorized as high.

cPinteraction (two-sided) was calculated using the Wald test for the cross-product of F. nucleatum status (positive vs. negative) and MSI status (high vs. non-high) in the logistic regression model.

As secondary analyses using a subset of cases with available tissue microarray data (up to 580 cases), we examined the association of F. nucleatum with densities of T-cell subsets (CD3+ cells, CD8+ cells, CD45RO+ cells, or FOXP3+ cells) in strata of MSI status. Although we did not observe a significant interaction between F. nucleatum and MSI status in relation to the density of CD3+ cells, CD8+ cells, CD45RO+ cells, or FOXP3+ cells (Pinteraction > 0.1, with the α level of 0.005; Supplementary Table S2), statistical power was limited in these analyses.

We did not observe a statistically significant interaction between Bifidobacterium genus and MSI status in relation to lymphocytic reaction patterns or densities of T-cell subsets (Pinteraction > 0.1, with the α level of 0.005).

As secondary analyses to assess the prognostic association of F. nucleatum in strata of MSI status, we conducted Kaplan–Meier analyses and Cox regression analyses, and did not observe a significant interaction between F. nucleatum and MSI status in relation to colorectal cancer–specific or overall mortality (Supplementary Fig. S1; Supplementary Table S3).

Using two large prospective cohort studies in the United States, we tested the hypothesis that the association of intratumoral F. nucleatum status with local immune response to colorectal cancer might differ by tumor MSI status. We found negative associations of F. nucleatum status with degrees of TIL and intratumoral periglandular reaction in MSI-high colorectal carcinomas and positive associations with those reaction patterns in non–MSI-high carcinomas. Our findings suggest an interplay between F. nucleatum, MSI status, and immune cells in colorectal tumor.

High-level infiltration of lymphocytes in colorectal cancer has been associated with better clinical outcomes (23, 24, 35, 36), and therapeutic activation of antitumor immune response in the tumor microenvironment has become an attractive strategy for cancer immunotherapy (37–39). In particular, immune-checkpoint inhibitors that target the PDCD1 (PD-1) or CD274 (PDCD1 ligand 1, PD-L1) protein have shown promise in treating various cancer types (37–39), including colorectal cancer with high MSI (26–28). However, despite the accompanying intense immune response, some MSI-high colorectal cancers can evade the immune eradication, and not all MSI-high tumors respond to the immune-checkpoint blockade. Therefore, immunosuppressive factors in MSI-high colorectal cancer need to be elucidated to improve the efficacy of immunotherapies in colorectal cancer. In the tumor microenvironment, host cells, neoplastic cells, and microbiota form complex interactive networks. A better understanding of these interactions would have relevance to existing avenues of treatment and may also reveal additional strategies for combating cancer (40, 41).

Alterations in the gut microbial composition are considered to play a role in colorectal cancer development (3, 42, 43). Fiber-rich diets are associated with a lower risk of colorectal cancer containing abundant F. nucleatum (42). Our findings of a possible temporal increase in detectable intratumor F. nucleatum might be due to growing popularity of colonoscopy screening, changes in dietary and lifestyle patterns, and/or aging of the cohort participants. Such interactions might become evident with a longitudinal epidemiologic investigation. Mechanistic studies indicate that F. nucleatum can exert carcinogenic effects through binding of its adhesin FadA to CDH1 (E-cadherin) and resultant activation of the CTNNB1 (beta-catenin)/WNT signaling pathway in colorectal carcinoma cells (44) as well as through suppression of the adaptive immune response (12). WNT signaling activation has been linked to immunosuppression in colorectal carcinoma (45). Experimental evidence suggests that binding of the virulence factor FAP2 to a host factor Gal-GalNAc may lead to interaction with the inhibitory immune cell receptor TIGIT, thereby inhibiting T-cell activity and NK cell cytotoxicity (11, 13, 46). In line with these studies, our previous population-based study has shown inverse associations of the enrichment of F. nucleatum in colorectal cancer tissue with CD3+ pan-T-cell density and patient survival (8, 18). F. nucleatum also increases numbers of myeloid-derived suppressor cells and tumor-associated macrophages in the colorectal tumor microenvironment (12, 47), contributing to proinflammatory reactions and impairment of tumor-specific immune responses.

Previous studies have linked F. nucleatum to MSI-high colorectal cancer (8, 16, 17), which is characterized by a vigorous antitumor immune response in the tumor microenvironment resulting from frameshift mutations and high expression of immunogenic peptides (23–27). However, there is a paradoxical finding that the presence of detectable F. nucleatum in tumor tissue is inversely associated with CD3+ T-cell density (8). Our current findings suggest that F. nucleatum may exert suppressive effects on adaptive antitumor immune response in MSI-high colorectal cancer with abundant immunogenic neoantigens (Fig. 2). During the development of colorectal carcinomas, MSI-high tumors need to acquire the means to resist antitumor immune activity. Our study provides evidence suggesting that infection with F. nucleatum may promote immune evasion in the immune cell–rich microenvironment of MSI-high colorectal carcinomas. We did not observe differential associations of F. nucleatum with T-cell subsets (CD3+ cells, CD8+ cells, CD45RO+ cells, and FOXP3+ cells) by tumor MSI status. These seemingly conflicting findings might be due to the roles of other cell types in immune reactions to tumor, measurement errors in T-cell densities, and/or a smaller overall sample size in analyses of T cells.

Figure 2.

F. nucleatum and lymphocytic reaction to colorectal carcinoma according to tumor MSI status. The current study suggests that F. nucleatum may attenuate adaptive immune response to MSI-high colorectal carcinomas, which exhibit abundant immunogenic neoantigens. In contrast, the proinflammatory properties of F. nucleatum may lead to lymphocytic reaction to non–MSI-high carcinomas that contain lower amounts of neoantigens and tumor-infiltrating lymphocytes than MSI-high carcinomas.

Figure 2.

F. nucleatum and lymphocytic reaction to colorectal carcinoma according to tumor MSI status. The current study suggests that F. nucleatum may attenuate adaptive immune response to MSI-high colorectal carcinomas, which exhibit abundant immunogenic neoantigens. In contrast, the proinflammatory properties of F. nucleatum may lead to lymphocytic reaction to non–MSI-high carcinomas that contain lower amounts of neoantigens and tumor-infiltrating lymphocytes than MSI-high carcinomas.

Close modal

The trend toward more lymphocytic reaction associated with F. nucleatum in non–MSI-high tumors warrants further discussion. In addition to suppression of the antitumor immune response, F. nucleatum can exert proinflammatory effects in the colorectal mucosa (12, 14, 15) and has been implicated in pathogenesis of inflammatory bowel diseases (14). Studies have shown that F. nucleatum may increase expression of proinflammatory cytokines and upregulate the NF-κB pathway (12, 48) in colorectal tumors. Therefore, we speculate that the proinflammatory effects of F. nucleatum may outweigh its immunosuppressive effects in non–MSI-high tumors, which contain fewer neoantigens and fewer tumor-infiltrating immune cells on average relative to MSI-high tumors (Fig. 2). These divergent associations of tumor tissue F. nucleatum with lymphocytic reactions between MSI-high and non–MSI-high tumors underscore the importance of interactive effects of environmental and tumor molecular features on host immune response in the tumor microenvironment (40).

We acknowledge limitations of our study. Our study focused on limited numbers of microbial species and immune reaction components. Accumulating evidence points to the roles of a variety of microbes and immune cells in antitumor immune responses and carcinogenesis (2, 3, 49). Therefore, more comprehensive analyses are warranted to investigate the interactions between microbes other than F. nucleatum (e.g., enterotoxigenic Bacteroides fragilis and pks-positive Escherichia coli; ref. 4) and other immune parameters (e.g., myeloid-derived suppressor cells, tumor-associated macrophages, and TIGIT-expressing cells). High-throughput sequencing-based analyses including 16S rRNA gene sequencing, metagenomic analyses, gene-expression profiling, and T-cell receptor sequencing would also provide insights into the interrelationship of tumor molecular markers and microbiota in relation to antitumor immune response. Detailed data on antibiotics and cancer treatments including immune-checkpoint inhibitors were not available in our study. There might be confounding factors that were unaccounted for; nonetheless, we adjusted for a variety of demographic, clinical, and tumor characteristics, as we utilized the MPE database. Another limitation is the cross-sectional study design, which dictates that we cannot rule out reverse causation. It is possible that immune responses may alter the amount of F. nucleatum in the tumor microenvironment. However, our hypothesis was based on accumulating evidence for the suppressive effects of F. nucleatum on T-cell–mediated antitumor immunity (8, 11–13). We used the quantitative PCR assay for F. nucleatum in FFPE tissue specimens; therefore, tissue processing procedures and storage conditions might have affected the detection rate. Nonetheless, our previous validation study using quantitative PCR showed concordance in detecting F. nucleatum in paired FFPE and frozen tissue specimens and linearity and reproducibility of F. nucleatum measurements in FFPE tissue specimens (8). Finally, our findings need to be validated in independent cohorts.

The current study has strengths, including the use of the MPE (40, 41) database derived from two U.S. prospective cohort studies, and this study represents the integration of immunology into the MPE paradigm (i.e., immunology-MPE; ref. 50). The data on the amount of F. nucleatum DNA in the tumor, tumor molecular characteristics, and pathologic findings in over 1,000 cases allowed us to comprehensively examine possible differential roles of F. nucleatum in MSI-high and non–MSI-high colorectal cancers. Our study population was derived from a large number of incident colorectal cancer cases that occurred in the well-defined populations of the prospective cohorts, and cases were also derived from hospitals in 48 states in the United States; both features increase the generalizability of our findings. Nevertheless, our findings need to be validated in independent studies.

In conclusion, the current study provides evidence for divergent effects of F. nucleatum in the tumor microenvironment according to tumor MSI status. There may be dominant suppressive effects of F. nucleatum on adaptive antitumor immune responses in MSI-high colorectal cancer and apparent proinflammatory effects in non–MSI-high cancer. Our findings, if validated, would inform future mechanistic studies examining the interplay of F. nucleatum and tumor characteristics in colorectal tumor evolution.

No potential conflicts of interest were disclosed.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.

Conception and design: T. Hamada, X. Zhang, K. Mima, K. Kosumi, M. Song, L. Liu, A.D. Kostic, W.S. Garrett, M. Meyerson, E.L. Giovannucci, M. Giannakis, S. Ogino

Development of methodology: T. Hamada, K. Mima, S. Bullman, Y. Sukawa, K. Kosumi, L. Liu, C. Huttenhower, W.S. Garrett, A.T. Chan, S. Ogino

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Hamada, K. Mima, Y. Sukawa, K. Kosumi, Y. Masugi, Y. Shi, M. Gu, W. Li, H. Koh, K. Inamura, E.L. Giovannucci, A.T. Chan, C.S. Fuchs, M. Giannakis, S. Ogino

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Zhang, K. Mima, S. Bullman, J.A. Nowak, K. Kosumi, Y. Cao, M. Song, L. Liu, Y. Shi, H. Koh, K. Nosho, N. Keum, K. Wu, J.A. Meyerhardt, C.S. Fuchs, R. Nishihara, M. Giannakis, S. Ogino

Writing, review, and/or revision of the manuscript: T. Hamada, X. Zhang, S. Bullman, J.A. Nowak, K. Kosumi, Y. Masugi, T.S. Twombly, Y. Cao, M. Song, L. Liu, A. da Silva, Y. Shi, M. Gu, H. Koh, N. Keum, K. Wu, J.A. Meyerhardt, W.S. Garrett, M. Meyerson, E.L. Giovannucci, A.T. Chan, M. Giannakis, S. Ogino

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Sukawa, Y. Masugi, L. Liu, A. da Silva, A.T. Chan, S. Ogino

Study supervision: J.A. Nowak, A.T. Chan, C.S. Fuchs, M. Giannakis, S. Ogino

This work was supported by NIH grants (P01 CA87969 to M.J. Stampfer; UM1 CA186107 to M.J. Stampfer; P01 CA55075 to W.C. Willett; UM1 CA167552 to W.C. Willett; U01 CA167552 to W.C. Willett and L.A. Mucci; P50 CA127003 to C.S. Fuchs; R01 CA118553 to C.S. Fuchs; R01 CA169141 to C.S. Fuchs; R01 CA137178 to A.T. Chan; K24 DK098311 to A.T. Chan; R35 CA197735 to S. Ogino; R01 CA151993 to S. Ogino; K07 CA190673 to R. Nishihara; and K07 CA188126 to X. Zhang); by Nodal Award (2016-02) from the Dana-Farber Harvard Cancer Center (to S. Ogino); and by grants from the Project P Fund, The Friends of the Dana-Farber Cancer Institute, Bennett Family Fund, and the Entertainment Industry Foundation through National Colorectal Cancer Research Alliance. Research supported by a Stand Up To Cancer Colorectal Cancer Dream Team Translational Research Grant (grant number: SU2C-AACR-DT22-17) was awarded to C.S. Fuchs and M. Giannakis. Stand Up To Cancer is a division of the Entertainment Industry Foundation. Research grants are administered by the American Association for Cancer Research, the Scientific Partner of SU2C. T. Hamada was supported by a fellowship grant from the Mitsukoshi Health and Welfare Foundation. K. Mima was supported by a grant from Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers from Japan Society for the Promotion of Science. K. Kosumi was supported by grants from Overseas Research Fellowship (JP2017-775) and Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers, from Japan Society for the Promotion of Science. M. Song was supported by the 2017 AACR–AstraZeneca Fellowship in Immuno-oncology Research (grant number 17-40-12-SONG). L. Liu was supported by a scholarship grant from Chinese Scholarship Council and a fellowship grant from Huazhong University of Science and Technology. N. Keum was supported by grants from the National Research Foundation of Korea (NRF-2018R1C1B6008822; NRF-2018R1A4A1022589). M. Giannakis was supported by a KL2/Catalyst Medical Research Investigator Training award (NIH award, KL2 TR001100).

We would like to thank the participants and staff of the Nurses’ Health Study and the Health Professionals Follow-up Study for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, and WY. The authors assume full responsibility for analyses and interpretation of these data.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Zitvogel
L
,
Pietrocola
F
,
Kroemer
G
. 
Nutrition, inflammation and cancer
.
Nat Immunol
2017
;
18
:
843
50
.
2.
Fridman
WH
,
Zitvogel
L
,
Sautes-Fridman
C
,
Kroemer
G
. 
The immune contexture in cancer prognosis and treatment
.
Nat Rev Clin Oncol
2017
;
14
:
717
34
.
3.
Chen
J
,
Domingue
JC
,
Sears
CL
. 
Microbiota dysbiosis in select human cancers: evidence of association and causality
.
Semin Immunol
2017
;
32
:
25
34
.
4.
Tilg
H
,
Adolph
TE
,
Gerner
RR
,
Moschen
AR
. 
The intestinal microbiota in colorectal cancer
.
Cancer Cell
2018
;
33
:
954
64
.
5.
Kostic
AD
,
Gevers
D
,
Pedamallu
CS
,
Michaud
M
,
Duke
F
,
Earl
AM
, et al
Genomic analysis identifies association of Fusobacterium with colorectal carcinoma
.
Genome Res
2012
;
22
:
292
8
.
6.
Castellarin
M
,
Warren
RL
,
Freeman
JD
,
Dreolini
L
,
Krzywinski
M
,
Strauss
J
, et al
Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma
.
Genome Res
2012
;
22
:
299
306
.
7.
Li
YY
,
Ge
QX
,
Cao
J
,
Zhou
YJ
,
Du
YL
,
Shen
B
, et al
Association of Fusobacterium nucleatum infection with colorectal cancer in Chinese patients
.
World J Gastroenterol
2016
;
22
:
3227
33
.
8.
Mima
K
,
Sukawa
Y
,
Nishihara
R
,
Qian
ZR
,
Yamauchi
M
,
Inamura
K
, et al
Fusobacterium nucleatum and T cells in colorectal carcinoma
.
JAMA Oncol
2015
;
1
:
653
61
.
9.
Repass
J
,
Iorns
E
,
Denis
A
,
Williams
SR
,
Perfito
N
,
Errington
TM
. 
Replication study: Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma
.
Elife
2018
;
7
:
e25801
.
10.
Bullman
S
,
Pedamallu
CS
,
Sicinska
E
,
Clancy
TE
,
Zhang
X
,
Cai
D
, et al
Analysis of Fusobacterium persistence and antibiotic response in colorectal cancer
.
Science
2017
;
358
:
1443
8
.
11.
Kaplan
CW
,
Ma
X
,
Paranjpe
A
,
Jewett
A
,
Lux
R
,
Kinder-Haake
S
, et al
Fusobacterium nucleatum outer membrane proteins Fap2 and RadD induce cell death in human lymphocytes
.
Infect Immun
2010
;
78
:
4773
8
.
12.
Kostic
AD
,
Chun
E
,
Robertson
L
,
Glickman
JN
,
Gallini
CA
,
Michaud
M
, et al
Fusobacterium nucleatum potentiates intestinal tumorigenesis and modulates the tumor-immune microenvironment
.
Cell Host Microbe
2013
;
14
:
207
15
.
13.
Gur
C
,
Ibrahim
Y
,
Isaacson
B
,
Yamin
R
,
Abed
J
,
Gamliel
M
, et al
Binding of the Fap2 protein of Fusobacterium nucleatum to human inhibitory receptor TIGIT protects tumors from immune cell attack
.
Immunity
2015
;
42
:
344
55
.
14.
Bashir
A
,
Miskeen
AY
,
Hazari
YM
,
Asrafuzzaman
S
,
Fazili
KM
. 
Fusobacterium nucleatum, inflammation, and immunity: the fire within human gut
.
Tumour Biol
2016
;
37
:
2805
10
.
15.
Ye
X
,
Wang
R
,
Bhattacharya
R
,
Boulbes
DR
,
Fan
F
,
Xia
L
, et al
Fusobacterium nucleatum subspecies animalis influences proinflammatory cytokine expression and monocyte activation in human colorectal tumors
.
Cancer Prev Res (Phila)
2017
;
10
:
398
409
.
16.
Flanagan
L
,
Schmid
J
,
Ebert
M
,
Soucek
P
,
Kunicka
T
,
Liska
V
, et al
Fusobacterium nucleatum associates with stages of colorectal neoplasia development, colorectal cancer and disease outcome
.
Eur J Clin Microbiol Infect Dis
2014
;
33
:
1381
90
.
17.
Nosho
K
,
Sukawa
Y
,
Adachi
Y
,
Ito
M
,
Mitsuhashi
K
,
Kurihara
H
, et al
Association of Fusobacterium nucleatum with immunity and molecular alterations in colorectal cancer
.
World J Gastroenterol
2016
;
22
:
557
66
.
18.
Mima
K
,
Nishihara
R
,
Qian
ZR
,
Cao
Y
,
Sukawa
Y
,
Nowak
JA
, et al
Fusobacterium nucleatum in colorectal carcinoma tissue and patient prognosis
.
Gut
2016
;
65
:
1973
80
.
19.
Yu
T
,
Guo
F
,
Yu
Y
,
Sun
T
,
Ma
D
,
Han
J
, et al
Fusobacterium nucleatum promotes chemoresistance to colorectal cancer by modulating autophagy
.
Cell
2017
;
170
:
548
63
e16
.
20.
Ogino
S
,
Chan
AT
,
Fuchs
CS
,
Giovannucci
E
. 
Molecular pathological epidemiology of colorectal neoplasia: an emerging transdisciplinary and interdisciplinary field
.
Gut
2011
;
60
:
397
411
.
21.
Rajpoot
M
,
Sharma
AK
,
Sharma
A
,
Gupta
GK
. 
Understanding the microbiome: Emerging biomarkers for exploiting the microbiota for personalized medicine against cancer
.
Semin Cancer Biol
2018
in press
.
22.
Morgillo
F
,
Dallio
M
,
Della Corte
CM
,
Gravina
AG
,
Viscardi
G
,
Loguercio
C
, et al
Carcinogenesis as a result of multiple inflammatory and oxidative hits: a comprehensive review from tumor microenvironment to gut microbiota
.
Neoplasia
2018
;
20
:
721
33
.
23.
Ogino
S
,
Nosho
K
,
Irahara
N
,
Meyerhardt
JA
,
Baba
Y
,
Shima
K
, et al
Lymphocytic reaction to colorectal cancer is associated with longer survival, independent of lymph node count, microsatellite instability, and CpG island methylator phenotype
.
Clin Cancer Res
2009
;
15
:
6412
20
.
24.
Nosho
K
,
Baba
Y
,
Tanaka
N
,
Shima
K
,
Hayashi
M
,
Meyerhardt
JA
, et al
Tumour-infiltrating T-cell subsets, molecular changes in colorectal cancer, and prognosis: cohort study and literature review
.
J Pathol
2010
;
222
:
350
66
.
25.
Giannakis
M
,
Mu
XJ
,
Shukla
SA
,
Qian
ZR
,
Cohen
O
,
Nishihara
R
, et al
Genomic correlates of immune-cell infiltrates in colorectal carcinoma
.
Cell Rep
2016
;
15
:
857
65
.
26.
Li
SK
,
Martin
A
. 
Mismatch repair and colon cancer: mechanisms and therapies explored
.
Trends Mol Med
2016
;
22
:
274
89
.
27.
Llosa
NJ
,
Cruise
M
,
Tam
A
,
Wicks
EC
,
Hechenbleikner
EM
,
Taube
JM
, et al
The vigorous immune microenvironment of microsatellite instable colon cancer is balanced by multiple counter-inhibitory checkpoints
.
Cancer Discov
2015
;
5
:
43
51
.
28.
Le
DT
,
Durham
JN
,
Smith
KN
,
Wang
H
,
Bartlett
BR
,
Aulakh
LK
, et al
Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade
.
Science
2017
;
357
:
409
13
.
29.
Yamauchi
M
,
Morikawa
T
,
Kuchiba
A
,
Imamura
Y
,
Qian
ZR
,
Nishihara
R
, et al
Assessment of colorectal cancer molecular features along bowel subsites challenges the conception of distinct dichotomy of proximal versus distal colorectum
.
Gut
2012
;
61
:
847
54
.
30.
Mima
K
,
Cao
Y
,
Chan
AT
,
Qian
ZR
,
Nowak
JA
,
Masugi
Y
, et al
Fusobacterium nucleatum in colorectal carcinoma tissue according to tumor location
.
Clin Transl Gastroenterol
2016
;
7
:
e200
.
31.
Sivan
A
,
Corrales
L
,
Hubert
N
,
Williams
JB
,
Aquino-Michaels
K
,
Earley
ZM
, et al
Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy
.
Science
2015
;
350
:
1084
9
.
32.
Nosho
K
,
Irahara
N
,
Shima
K
,
Kure
S
,
Kirkner
GJ
,
Schernhammer
ES
, et al
Comprehensive biostatistical analysis of CpG island methylator phenotype in colorectal cancer using a large population-based sample
.
PLoS One
2008
;
3
:
e3698
.
33.
Hamada
T
,
Cao
Y
,
Qian
ZR
,
Masugi
Y
,
Nowak
JA
,
Yang
J
, et al
Aspirin use and colorectal cancer survival according to tumor CD274 (Programmed Cell Death 1 Ligand 1) expression status
.
J Clin Oncol
2017
;
35
:
1836
44
.
34.
Masugi
Y
,
Nishihara
R
,
Hamada
T
,
Song
M
,
da Silva
A
,
Kosumi
K
, et al
Tumor PDCD1LG2 (PD-L2) expression and the lymphocytic reaction to colorectal cancer
.
Cancer Immunol Res
2017
;
5
:
1046
55
.
35.
Pages
F
,
Mlecnik
B
,
Marliot
F
,
Bindea
G
,
Ou
FS
,
Bifulco
C
, et al
International validation of the consensus Immunoscore for the classification of colon cancer: a prognostic and accuracy study
.
Lancet
2018
;
391
:
2128
39
.
36.
Rozek
LS
,
Schmit
SL
,
Greenson
JK
,
Tomsho
LP
,
Rennert
HS
,
Rennert
G
, et al
Tumor-infiltrating lymphocytes, Crohn's-like lymphoid reaction, and survival from colorectal cancer
.
J Natl Cancer Inst
2016
;
108
:
djw027
.
37.
Basile
D
,
Garattini
SK
,
Bonotto
M
,
Ongaro
E
,
Casagrande
M
,
Cattaneo
M
, et al
Immunotherapy for colorectal cancer: where are we heading?
Expert Opin Biol Ther
2017
;
17
:
709
21
.
38.
Grizzi
F
,
Basso
G
,
Borroni
EM
,
Cavalleri
T
,
Bianchi
P
,
Stifter
S
, et al
Evolving notions on immune response in colorectal cancer and their implications for biomarker development
.
Inflamm Res
2018
;
67
:
375
89
.
39.
Gang
W
,
Wang
JJ
,
Guan
R
,
Yan
S
,
Shi
F
,
Zhang
JY
, et al
Strategy to targeting the immune resistance and novel therapy in colorectal cancer
.
Cancer Med
2018
;
7
:
1578
603
.
40.
Ogino
S
,
Nowak
JA
,
Hamada
T
,
Phipps
AI
,
Peters
U
,
Milner
DA
 Jr
, et al
Integrative analysis of exogenous, endogenous, tumour and immune factors for precision medicine
.
Gut
2018
;
67
:
1168
80
.
41.
Hamada
T
,
Keum
N
,
Nishihara
R
,
Ogino
S
. 
Molecular pathological epidemiology: new developing frontiers of big data science to study etiologies and pathogenesis
.
J Gastroenterol
2017
;
52
:
265
75
.
42.
Mehta
RS
,
Nishihara
R
,
Cao
Y
,
Song
M
,
Mima
K
,
Qian
ZR
, et al
Association of dietary patterns with risk of colorectal cancer subtypes classified by Fusobacterium nucleatum in tumor tissue
.
JAMA Oncol
2017
;
3
:
921
7
.
43.
Park
CH
,
Han
DS
,
Oh
YH
,
Lee
AR
,
Lee
YR
,
Eun
CS
. 
Role of Fusobacteria in the serrated pathway of colorectal carcinogenesis
.
Sci Rep
2016
;
6
:
25271
.
44.
Rubinstein
MR
,
Wang
X
,
Liu
W
,
Hao
Y
,
Cai
G
,
Han
YW
. 
Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating E-cadherin/beta-catenin signaling via its FadA adhesin
.
Cell Host Microbe
2013
;
14
:
195
206
.
45.
Grasso
CS
,
Giannakis
M
,
Wells
DK
,
Hamada
T
,
Mu
XJ
,
Quist
M
, et al
Genetic mechanisms of immune evasion in colorectal cancer
.
Cancer Discov
2018
;
8
:
730
49
.
46.
Abed
J
,
Emgard
JE
,
Zamir
G
,
Faroja
M
,
Almogy
G
,
Grenov
A
, et al
Fap2 mediates Fusobacterium nucleatum colorectal adenocarcinoma enrichment by binding to tumor-expressed Gal-GalNAc
.
Cell Host Microbe
2016
;
20
:
215
25
.
47.
Park
HE
,
Kim
JH
,
Cho
NY
,
Lee
HS
,
Kang
GH
. 
Intratumoral Fusobacterium nucleatum abundance correlates with macrophage infiltration and CDKN2A methylation in microsatellite-unstable colorectal carcinoma
.
Virchows Arch
2017
;
471
:
329
36
.
48.
Yang
Y
,
Weng
W
,
Peng
J
,
Hong
L
,
Yang
L
,
Toiyama
Y
, et al
Fusobacterium nucleatum increases proliferation of colorectal cancer cells and tumor development in mice by activating toll-like receptor 4 signaling to nuclear factor-kappaB, and up-regulating expression of MicroRNA-21
.
Gastroenterology
2017
;
152
:
851
66
e24
.
49.
Pushalkar
S
,
Hundeyin
M
,
Daley
D
,
Zambirinis
CP
,
Kurz
E
,
Mishra
A
, et al
The pancreatic cancer microbiome promotes oncogenesis by induction of innate and adaptive immune suppression
.
Cancer Discov
2018
;
8
:
403
16
.
50.
Ogino
S
,
Giannakis
M
. 
Immunoscore for (colorectal) cancer precision medicine
.
Lancet
2018
;
391
:
2084
6
.