The presence of bacteria varies between colorectal adenocarcinomas, precursor lesions and non-malignant tissue

Abstract

Background: A causal association has been suggested between certain bacteria and colorectal cancer (CRC). Only a few studies have, however, investigated the presence of these bacteria directly in colon tissue with conflicting results. It is thus uncertain which role they may have in prognosis and carcinogenesis of CRC.

Methods: Formalin-fixed and paraffin-embedded (FFPE) colorectal tissue samples from patients diagnosed with colorectal cancer (CRC)(tumor and paired normal tissue, n = 99), adenomas (n = 96), or diverticular disease (n = 104) were tested for the presence and bacterial load of Streptococcus gallolyticus (S. gallolyticus), Fusobacterium nucleatum (F. nucleatum), and Bacteroides fragilis (B. fragilis) using quantitative PCR. A subsequent broader search was conducted on a subset of samples using 16S ribosomal RNA gene sequencing. Finally, to evaluate the prognostic value, the bacterial status was compared to patient outcome.

Results: S. gallolyticus was not detected by qPCR in any of the investigated tissue samples and F. nucleatum and B. fragilis were found to be equally distributed in tumors, paired normal tissue, and diverticula, but significantly less present in adenomas compared to both tumors and diverticula. Neither, F. nucleatum nor B. fragilis status affected the five-year prognosis of the patients. The 16S rRNA gene sequencing data revealed that tumors were associated with the Prevotella genus while conversely adenomas and diverticula were associated with Acinetobacter genus.

Conclusion: These findings do not support a role of F. nucleatum or B. fragilis during colorectal beginning, while S. gallolyticus was not implicated in the colorectal tissue of a Danish population. A potential role of the bacterial genera Prevotella and Acinetobacter was indicated, and requires further investigations.

Keywords: Acinetobacter; Bacteroides; Cancer microbiota; Colorectal adenomas; Colorectal cancer; Fusobacterium; Prevotella; Streptococcus.

Fig. 1

Presence and quantity of F. nucleatum and B. fragilis in colorectal tissue. qPCR determination of presence and quantity of bacterial DNA in 99 colorectal tumor tissue, 76 paired normal tissue, 96 adenomas and 104 diverticula. a Prevalence of F. nucleatum and B. fragilis in colorectal tissue. Positivity was determined as bacterial species with a DNA quantity above the LOD of the primers. b Prevalence of F. nucleatum and B. fragilis in different stages of CRC. No statistical significant differences were observed. c Difference in quantity of F. nucleatum and B. fragilis DNA in colorectal tumor tissue compared to paired normal tissue, adenomas and diverticula as well as in adenomas compared to diverticula. Brackets denote standard deviation. * P < 0.05, ** P < 0.001

Fig. 2

Five-year follow-up based on presence of F. nucleatum or B. fragilis. Survival (a, c, e) and disease-free survival (b, d, f) of patients presenting with CRC (a and b, n = 99), adenomas (c and d, n = 96) or diverticula (e and f, n = 104) depending on presence or absence of F. nucleatum or B. fragilis. Five-year follow-up data was not available for two study participants belonging to the diverticula group. These patients were excluded from the follow-up analysis

Fig. 3

Gut microbiome richness and diversity between tissue types in a subsection of samples. a OTU richness and b Shannon diversity index was compared between a subsection of the tumors, paired normal tissue, adenomas and diverticula included in this study. A total of 35 tissue samples were investigated, with 10 tumors samples, 6 paired normal tissue, 9 adenomas and 10 diverticula

Fig. 4

Variation in bacterial composition between individual samples and tissue types. β-diversity was investigated in 35 tissue samples using a) PCA and subsequent b) RDA plots with Hellinger Distance of OTU abundances. Colored boxes represent different tissue types

Fig. 5

Bacterial composition in a subsection of tissue types. The bacterial composition of the 35 samples analyzed using 16S rRNA gene sequencing, was visualized. a Heatmap of the investigated samples. Colors represent bacterial composition, with stronger red indicating higher percentage of total read abundance, while light blue indicate absence of the bacteria. The 20 most common bacteria are depicted on the y axis, while the x-axis contains the 35 samples included in this analysis. b qPCR results for the investigated subsection of tissue samples. Samples starting with “U” indicated diverticula, “P” indicate adenomas while samples starting with “K” indicate samples originating from patients diagnosed with CRC (tumors or paired normal tissue)