Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma

Abstract

An estimated 15% or more of the cancer burden worldwide is attributable to known infectious agents. We screened colorectal carcinoma and matched normal tissue specimens using RNA-seq followed by host sequence subtraction and found marked over-representation of Fusobacterium nucleatum sequences in tumors relative to control specimens. F. nucleatum is an invasive anaerobe that has been linked previously to periodontitis and appendicitis, but not to cancer. Fusobacteria are rare constituents of the fecal microbiota, but have been cultured previously from biopsies of inflamed gut mucosa. We obtained a Fusobacterium isolate from a frozen tumor specimen; this showed highest sequence similarity to a known gut mucosa isolate and was confirmed to be invasive. We verified overabundance of Fusobacterium sequences in tumor versus matched normal control tissue by quantitative PCR analysis from a total of 99 subjects (p = 2.5 × 10(-6)), and we observed a positive association with lymph node metastasis.

Figure 1.

Relative abundance of microbial genomes in tumor and control specimens. Numbers of read-pairs that matched known microbial sequences were normalized according to sequencing depth for both tumor and matched normal samples. The abundance of normalized bacterial read-pairs ranged from zero to a maximum of 66,896 represented by a transition from green to red on a log₁₀ scale. F. nucleatum sequences were present in the tumor samples at levels twofold or greater than in normal samples in nine out of the 11 subjects. The mean over abundance across all subjects was 79-fold.

Figure 2.

Relative abundance of Fusobacterium in tumor versus normal colorectal carcinoma biopsies. Relative amounts of Fusobacterium DNA were determined between tumor and matched normal biopsies in 99 subjects, using quantitative real-time PCR (qPCR). The cycle threshold (Ct) values for the normal samples had a Ct range of 25.5 to 40, and the Ct range for the tumor samples was between 21.4 and 40. The data shown are mean values from two independent experiments. Fusobacterium load, as determined by qPCR, was found to be significantly higher in the tumor samples versus the matched control samples (two-tailed ratio t-test, p = 2.52 × 10⁻⁶).

Figure 3.

Hive plots showing alignment of three Fusobacterium genomes. Approximately 32 million high-quality WGS Illumina HiSeq reads (≥99 consecutive Q30 bases) from Fusobacterium tumor isolate CC53 were assembled with SSAKE (v3.7, default options) into 379 contigs. The contigs were aligned using cross_match (-minmatch 29 -minscore 59 -masklevel 101) to the complete F. nucleatum susb. nucleatum ATCC 25586 genome and, independently, to the 12-contig HMP Fusobacterium sp. 3_1_36A2 assembly, respectively; and ordered/oriented based on the highest identity to the latter sequence. Three-way cross_match (http://www.phrap.org) alignments between each Fusobacterium genome were performed and represented visually using hive plots (http://www.hiveplot.com). For each, the top, left, and right axes are proportional to genome size and represent the Fusobacterium tumor isolate CC53 (2.07 Mb), the HMP sp. 3_1_36A2 (2.25 Mb), and the ATCC 25586 type strain (2.17 Mb), in that order. Synteny between the isolates is depicted by green and red links that show direct and inverted alignments, respectively. Sequence similarity and synteny is highest between CC53 and sp. 3_1_36A2, as evidenced by a greater density of high similarity sequence matches between them, relative to ATCC 25586, and shared patterns of inversions compared to this reference strain. Three regions of sequences present in sp. 3_1_36A2 but absent from CC53 are apparent as conspicuous gaps on the sp. 3_1_36A2 axis. Sequence segments unique to CC53 are not visible at this scale.

Figure 4.

Representative differentially stained immunofluorescence image showing strain CC53 invading Caco-2 cells. (A) The differential staining method allows for delineation between bacteria that have penetrated the host cells (labeled for actin in green) to reside within them (orange, also indicated with orange arrows), and bacteria present on the outside of the cell (purple). CC53 shows a very long, fine, thread-like cell morphology. (B) Detail of CC53 invasion. (Top right) A representative CC53 cell in the process of invading the Caco-2 host cell (image is differentially stained as for 4A). The long, thread-like cells appear to penetrate host cells pole first. (Orange arrow) The CC53 cell that is internalized; (white arrow) the external portion of the same bacterial cell that has looped around on itself, demonstrating apparent flexibility. (Purple arrow) A single CC53 cell that has not invaded the host cells, for contrast. Bar, 15 μm. In the immunofluorescence micrograph, CC53 shows a very long, flexible cell morphology. (Green) Actin (Caco-2 cells); (orange) invasive and internalized bacteria; (purple) bacteria external to the cell.