Eubacterium rectale contributes to colorectal cancer initiation via promoting colitis

All cell culture media, trypsin and antibiotics were purchased from Gibco (Grand Island, NY, USA), and FBS was purchased from HyClone (Logan, UT, USA). DAPI, lysosyme, proteinase K, DNase and RNase were purchased from Sigma-Aldrich (St Louis, MO, USA). Immobilon membranes were purchased from Merck Millipore (Bedford, MA, USA). ECL Plus substrate was purchased from CWBio (Beijing, China). Nuclear and cytoplasmic protein extraction kit was purchased from Beyotime (Shanghai, China). ZR Fungal/Bacterial DNA Kit was purchased from Zymo Research Corp. (Irvine, CA, USA). Qualitative fecal occult blood detection kit was purchased from Beijing Leagene Biotechnology Co., Ltd. (Beijing, China). Limulus Amebocyte Lysate (catalogs: T7572) was purchased from Solarbio (Beijing, China).

Rabbit anti-nuclear factor kappa B (anti-NF-κΒ) p65 antibody (catalogs: SAB4502610), rabbit anti-Histone H3 antibody (catalogs: SAB4500354), rabbit anti-IKKα antibody (catalogs: SAB4500257), rabbit anti-IκBα antibody (catalogs: SAB1305978), rabbit anti-β-actin antibody (catalogs: SAB2100037), goat anti-rabbit IgG-peroxidase (catalogs: A0545) and goat anti-rabbit IgG FITC (catalogs: AP132F) were purchased from Sigma-Aldrich (St Louis, MO, USA).

Intestinal microflora samples were collected from 75 patients and 26 health people in Tianjin Union Medical Center. Cotton swab was used to dip on intestinal surface of colorectal cancer tissues collected from patients. Three distinct locations from the same tissue of a given cancer patient were collected, including tumor position (T), para-tumor epithelia position (P) and normal epithelia position (N). Intestinal microflora samples of healthy people were collected when people were diagnosed for colonoscopy and determined as healthy people (H). These samples were rinsed in 1 ml physiological saline. Then, 200 μl of these solutions were used for Bacterial DNA extraction. Bacterial DNA of these samples was extracted using ZR Fungal/Bacterial DNA Kit according to the manufacturer’s instructions.

The 16S ribosomal RNA amplicon libraries were constructed according to the Illumina manufactory manual. Briefly, the following primers were used to amplify the V3-V4 region of 16S rRNA gene: forward primer, 5′TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG and reverse primer 5′GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC.

The amplified DNA library was subsequently purified using AMPure XP beads (Beckman Coulter, USA) and quantified using Quant-iT PicoGreen dsDNA assay kit (Thermo Fisher, USA). The paired-end sequencing reads (2 × 300 bp) were generated on Illumina MiSeq platform, according to the Illumina standard protocol. Quality control and filtering of raw sequences were carried out using FastQC (https://​www.​bioinformatics.​babraham.​ac.​uk/​projects/​fastqc/​). The filtered paired-end reads were assembled using PandaSeq [33]. The assembled sequences were loaded on QIIME pipeline (qiime.org) [34] for de novo OTU picking, taxonomic assignment, and diversity analyses. Usearch was used within QIIME to detect and remove de novo chimera. Rarefaction was performed using alpha_rarefaction.py in QIIME pipeline. De novo OTUs were picked based on sequence similarity (97%) within the assembled sequences. Next, taxonomy was assigned to OTU representative sequences that were picked for each OTU. To identify differential abundance of bacterial phylo associated with specific sites of CRC patients and healthy people, we applied LEfSe [35].

E. rectale ATCC 33656 and F. nucleatum subsp. Nucleatum ATCC 25586 were both purchased from Biobw (Beijing, China). E. rectale was grown anaerobically in PYG medium (ATCC medium 1527) composing of 0.5% peptone, 1% yeast extract, 0.5% tryptone, 0.01% resazurin (Sigma), 0.0008% MgSO₄, 0.0008% anhydrous CaCl₂, 0.004% K₂HPO₄, 0.004% KH₂PO₄, 0.002% NaCl, 0.04% NaHCO₃, 0.0001% (v/v) Vitamin K1 (Sigma), 0.05% L-cysteine-HCL (Sigma). The pH was adjusted to 7.0. F. nucleatum ATCC 25586 was maintained anaerobically in Columbia broth (BD Biosciences). PYG medium and Columbia broth were both boiled to remove oxygen and placed in anaerobic bottles. Argon gas was added to the bottles. Then the bottles were sealed by a cap with a rubber septum before being autoclaved for 20 min at 121 °C.

Colitis was experimentally induced by administration of 2.5% Dextran sodium sulfate (DSS) in the drinking water of 7-week-old Balb/c mice for 7 days. Afterwards, the mice were allowed to drink water without DSS until the end of experiment. After 2 days of DSS drinking, mice were inoculated with E. rectale or F. nucleatum by coloclysis with 1 × 10⁷ colony forming units (CFU) in 200 μl PBS. Briefly, intraperitoneal injection of 4% chloral hydrate (10 μl/g) was used to deeply anesthetize mice. The mice were kept to prone position. Colon was inserted with paraffin oil applied plastic hose (2 mm internal diameter) through anus. When the top of hose was inserted 4 cm from anus, the mice were kept to handstand. The hose was pulled out after bacterium was injected. Anus was pressed with cotton swab within 1 min, and the mice were put back to cage. Mice were separated into 6 groups: (Water) control group, water drinking and no bacterial inoculation; (DSS) DSS drinking and no bacterial inoculation; (E. rectale + water) water drinking and E. rectale inoculation; (F. nucleatum + water) water drinking and F. nucleatum inoculation; (E. rectale + DSS) DSS drinking and E. rectale inoculation; (F. nucleatum + DSS) DSS drinking and F. nucleatum inoculation. Each group consisted of five mice. The mice were monitored and weighed daily until they had lost > 20% of their initial body weight, And upon being sacrificed, the colon and spleen were both harvested for determination of colon length, histological examination and spleen weight. Stool was collected and measured daily for the presence of occult and gross blood. Occult blood was measured by a qualitative fecal occult blood detection kit.

Harvested colons were cleaned in physiological saline solution to remove fecal residue. After 24 h of fixation in 10% buffered formalin, the colons were embedded in paraffin and sectioned. Then the tissues were stained with hematoxylin and eosin (H&E). Images of histology slides were taken by an Olympus IX51 microscope. Histology of colon was scored on a scale of 0–5 where 0 = normal and 5 = severe inflammation and complete loss of surface and crypt epithelium.

Endotoxic LPSs were isolated from bacteria using the hot phenol-water method as previously described [36]. In brief, bacterial pellets were digested with lysozyme. Then, cell lysate was incubated with DNase and RNase to digest nucleic acids. 90% phenol extraction procedure was used. The water layer was dialyzed against deionized water and digested with proteinase K. Then the solution was dialyzed again. The purity of the phenol extract was tested by the detection of nucleic acid (260 nm), coomassie brilliant blue stained protein (595 nm). Isolated LPSs were also detected by SDS-PAGE. Finally, LPSs were lyophilized using Freeze Dry System (FD-1A-80, BILON Ltd., Shanghai, China). The biological effectiveness of isolated LPSs was tested in Limulus Amebocyte Lysate assay according to the manufacturer’s instructions.

The human normal colon epithelial cell lines NCM460 and HCoEpiC were both purchased from Tong Pai Technology, Inc. (Shanghai, China). These cells were cultured in RPMI 1640 medium supplemented with 10% FBS, 100 μg/mL streptomycin, and 100 U/mL penicillin. LPS was dissolved in RPMI 1640 medium and vortexed for 10 min by ultrasonic bath (Shenglan Ltd., Jiangsu, China) before use. A concentration series of LPS dilutions was used to treat cells for 2.5 h.

For detection of NFκΒ p65 expression in nuclei, nuclear proteins were isolated from harvested cells using a nuclear and cytoplasmic protein extraction kit and following the manufacturer’s protocol. Western blotting was carried out as previously described [37]. Briefly, protein samples were suspended in SDS loading buffer and boiling for 5 min. Then 10 μg protein was separated by SDS-PAGE and transferred to immobilon membranes by semi-dry blotting method. The membranes were probed with antibodies using standard techniques. Finally, the protein bands were visualized using ECL Plus and exposed film. Each assay was carried out in triplicate.

Immunofluorescence staining was used to analyze the location of NFκΒ p65 expression. The procedure was carried out as previously described [37]. In brief, cells were washed with PBS for three times, fixed in 10% formalin at room temperature for 20 min, treated with 0.5% Triton X-100 for 5 min at 4 °C and blocked with 5% normal goat serum overnight at 4 °C. Then the slides were incubated with primary antibody for 1 h at 37 °C. After washing with PBS, the slides were incubated with FITC-conjugated secondary antibody and DAPI simultaneously for 1 h, at 37 °C. Subsequently, the slides were washed with PBS and then sealed. The slides were photographed immediately using a fluorescence microscope (Olympus, Tokyo, Japan).

All data represent mean ± standard deviation. Statistical analysis was carried out by Student’s t test using SPSS software. P value < 0.05 was considered statistically significant.

To reveal microbiome profile differences among gut mucosal phenotypes, we performed 16S ribosomal RNA gene sequencing on biopsy samples collected from on-tumor sites (n = 75), tumor-adjacent sites (n = 75), and off-tumor sites (n = 75) of CRC patients, and healthy people (n = 26) at Tianjin Union Medical Center (Table 1). In the discovery cohort, severity of pathological changes in intestinal epithelium was strongly associated with a decrease of intraindividual diversity of on-tumor sties compared to off-tumor sites, as measured by Shannon diversity index (p < 0.05) (Fig. 1a). Alpha-diversity in microbiome profiles may be affected by individual difference. It has been reported that Shannon diversity index is similar in fecal samples of normal, adenoma and cancer group [12], but it is also found that bacterial diversity of gut mucosal samples is lower in CRC group than normal group [13]. To assess the overall diversity of microbiome profiles, we performed principal component analysis (PCA) based on genera abundances in different mucosal sites from CRC patients and healthy people. Figure 1b showed the microbial composition differences in different mucosal sites of patients and healthy controls.

Using the reference Greengenes database, clustered operational taxonomic units (OTU) were assigned for bacterial phylotypes. Consistent with previous findings [13], we observed the same major bacterial phyla across the different mucosal sites of patients and healthy controls, including Firmicutes, Bacteroides, and Proteobacteria. To identify differentially abundant taxa in mucosal phenotypes as site-specific bacterial taxon markers, we performed linear discriminant analysis (LDA) on the microbial compositions of the different mucosal sites from CRC patients and healthy people (LDA score > 2.5, p value < 0.05). No significant changes were observed at phylum level among different mucosal sites from CRC patients and healthy people. In total, 15, 5, 8, and 15 families were identified to be associated with on-tumor site, adjacent-tumor site, off-tumor site, and healthy control, respectively (Fig. 2). At family level, Fusobacteriaceae showed the strongest correlation with on-tumor site (LDA score = 4.86, p value = 3.28E-07) (Fig. 2a), consistent with previous findings [5]. Bradyrhizobium (LDA score = 3.71, p value = 1.32E-11) and Corynebacteriaceae (LDA score = 3.49, p value = 1E-04) showed the strongest correlation with tumor-adjacent sites of CRC patients, distinct different from the taxon markers for on-tumor sites (Fig. 2b). Erysipelotrichaceae (LDA score = 4.02, p value = 0.03) and Ruminococcaceae (LDA score = 4.84, p value = 0.03) showed the strongest correlation with off-tumor sites and healthy controls, respectively (Fig. 2c and d). Our data suggest that microbial compositions in different mucosal sites and healthy controls show distinct microbial profiles. In addition, bacteria associated with off-tumor site, including the family Eubacteriaceae, may serve as driver bacteria that initiate CRC.

Among the eight bacterial families associated with off-tumor site of CRC patients, the family Eubacteriaceae contains a butyrate-producing genus Eubacterium. Here we chose E. rectale as an example to examine the role of Eubacterium in CRC initiation. As shown in Fig. 3a, abundances of E. rectale and F. nucleatum were calculated by QIIME. As a typical bacterium associated with CRC, according to previous reports [8], F. nucleatum exhibits high abundance in intestinal tract of CRC patients and is practically nonexistent in samples from healthy people. Furthermore, tumor location has the highest abundance of F. nucleatum among all three locations of patients’ intestine, suggesting that F. nucleatum is suitable to grow in the tumor microenvironment. It has been reported that E. rectale is commonly distributed in healthy intestinal tract, and is less abundant in CRC patients [14]. Our results showed that there was lower abundance of E. rectale in on-tumor site of CRC patients than off-tumor site, as well as that E. rectale exhibits higher abundance in on-tumor site of CRC patients than healthy people samples. Local intestinal microenvironment of off-tumor sites of CRC patients is expected to be more like that of healthy people samples than that of on-tumor sites of CRC patients. Higher abundance of E. rectale in off-tumor site compared with healthy people may be due to other reasons, such as eating habits or individual difference. This difference in E. rectale abundance may also be due to its presence in these people’s intestinal tract before CRC onset, which is also one reason for CRC initiation. Therefore, we speculate that E. rectale may be a ‘driver’ bacterium, and next investigated the role E. rectale in CRC initiation (Fig. 3b).

DSS mouse model of colitis was used to examine whether E. rectale promotes colitis which contributes to initiation of CRC. Treatment of mice in each group was shown in Fig. 4a. Histopathological analysis of H&E stained colon sections (Fig. 4b and c) revealed that, compared to the normal colonic tissue of water drinking group, DSS exposure led to segmental regions of intestinal mucosa rupture, marked acute inflammatory cell infiltration, crypt abscesses, and early architectural distortion. E. rectale but not F. nucleatum inoculation increased DSS induced colitis. Neither E. rectale nor F. nucleatum inoculation can induce obvious colitis without DSS exposure. Group ‘E. rectale + DSS’ showed the highest histology score, suggesting that E. rectale increased DSS induced colitis. Group ‘DSS’ and ‘F. nucleatum + DSS’ showed similar histology score, -suggesting that F. nucleatum did not increase DSS-induced colitis.

Figure 4d–g showed that colon length from the distal end to the colon-cecum junction and spleen weight changes. E. rectale or F. nucleatum inoculation in water drinking mice did not induce a decrease in colon length or increase spleen weight. DSS exposure caused significant decrement in colon length or increment in spleen weight, and additional E. rectale inoculation further increased this effect. By contrast, F. nucleatum inoculation did not increase the effect of DSS on colon length shorten or spleen weight increment.

As shown in Fig. 4h and j, Body weight in inoculated groups showed no significant difference with Water group. Body weight in DSS group showed significant difference starting at day 10, additional F. nucleatum inoculation shift this time point to day 9, and additional E. rectale inoculation shift this time point to day 7. Body weight in group ‘E. rectale + DSS’ continued to decrease until the end of experiment, but body weight in groups ‘DSS’ and group ‘F. nucleatum + DSS’ declined slowly from day 13 to the end of experiment, the reason of which may be due to E. rectale but not F. nucleatum inoculation enhance colitis induced by DSS. In addition, all mice survived until the end of experiment.

As shown in Fig. 4i and j, there are no occult or gross blood in stool in Water group and inoculated groups. DSS drinking groups had occult blood in stool at day 4 and gross blood in stool at day 5. These groups began to exhibit gross blood in all stool at day 8. DSS group and F. nucleatum + DSS group both showed a decrease in gross blood and an increase in adult blood in stool from day 10. Occurrence of occult blood in stool of these two groups also continued to decrease from day 12 to the end of experiment. Compared to these two groups, E. rectale + DSS group has more time of gross blood occurrence, and incidence rate of occult blood decrease slower.

NF-κΒ is a critical transcription regulator that regulates the expression of many genes of inflammatory functions and the immune system. NF-κΒ activation facilitates both tumor development and metastatic progression cancer [15, 16]. NF-κΒ has two subunits, p50 and p65, to form a homodimer or heterodimer. The dimer maintains normal physiological function in vivo. In normal conditions, NF-κΒ is sequestered by members of the IκB family to form an inactivated compound in the cytoplasm. Under external stimuli, such as enteroinvasive bacterial infection or cytokine stimulation to intestinal epithelial cells [17, 18], NF-κΒ is activated. During this process, IκB kinase (IKK) is activated and caused IκB degradation, which induced NF-κΒ dissociation from IκΒ and transferred to the nucleus. Therefore, we measured expression levels of IKKα and IκBα in cytoplasm, and p65 in nucleus to determine NF-κΒ activation.

Two human normal colon epithelial cell lines, HCoEpiC and NCM460, were used to investigate whether E. rectale LPS induces inflammatory in vitro. Our results (Fig. 5) showed that NF-κΒ is activated in both HCoEpiC and NCM460 when they were treated by 20 ng/ml of E. rectale LPS for 2.5 h. NF-κΒ was significantly expressed in nuclei of HCoEpiC and NCM460 cells after treating with E. rectale LPS. By contrast, F. nucleatum LPS does not show this effect at the same concentration. Purity measurement of LPS is shown in Additional file 1: Figure S4.