E. coli strains in this study included the pathogenic murine strain EcNC101 wild-type (WT), the EcNC101 depleted for the colibactin P gene (ΔclbP) (both EcNC101 strains were a gift from Dr. Christian Jobin, Cancer Microbiota & Host Response, UF Health Cancer Center, University of Florida), the probiotic E. coli Nissle 1917 (EcN) (Mutaflor®, Pharma-Zentrale GmbH, Germany) and the E. coli K12 (EcK12) (ER2738, New England BioLabs Ltd., Whitby, ON, Canada). EcNC101 were transformed with the plasmid pROP-IP-FBFP (a gift from Srivatsan Raman; Addgene plasmid #122138¹; RRID: Addgene_122,138) that encodes a flavin mononucleotide-based fluorescent protein (FbFP) to obtain EcNC101-FbFP. All strains of E. coli were grown from glycerol stocks in lysogeny broth (LB) with adequate antibiotic supplementation at 37°C at 150 revolution per minute (rpm) overnight and subsequently subculture in appropriate media. The Lactobacillus plantarum 299 V® (Digestive care, Jamieson Wellness Inc., Toronto, ON, Canada) was grown overnight at 37°C in Man-Rogosa-Sharpe (MRS) broth (Sigma–Aldrich Canada Co, Oakville, ON, Canada) in anaerobic conditions.

Following the Canadian Council of Animal Care guidelines, all procedures were performed after approval by the Institutional Animal Care Committee of the Centre de recherche du Centre Hospitalier de l’Université de Montréal (CRCHUM). Breeding colonies of WT and Apc^Min/+ C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME, United States) were established in a specific-pathogen-free (SPF) facility, and offspring were genotyped using allele-specific polymerase chain reaction (PCR) analysis. Mice were 4 weeks old at the beginning of the experiments and were maintained under standard 12:12 light/dark conditions. They were co-housed at two to three mice per cage and were allowed ad libitum access to food and water. Mice received a standard diet (Envigo Teklad Diets, TD2918) until 5 weeks old. At 4 weeks old, Apc^Min/+ females were treated with 2% (w/v) dextran sulfate sodium (DSS) (TDB consultancy AB, Uppsala, Sweden) in drinking water for a week. The following day after DSS treatment, mice were fed a diet containing 10% (wt/wt) inulin (Quadra Chemicals Ltd., Vaudreuil-Dorion, QC, Canada) (Envigo Teklad Diets, TD.190651) or a control diet containing 10% (wt/wt) cellulose (Envigo Teklad Diets, TD.190723), a non-fermentable fiber. Apart from fiber supplementation, both diets were matched in caloric intake in terms of carbohydrate, protein, and fat (Supplementary Table S2). At the same time, mice received an oral gavage of 200 μL of bacterial suspension (10⁸ colony-forming units, CFUs) or sterile Dulbecco’s phosphate-buffered saline (D-PBS; WISENT Inc., St-Bruno, QC, Canada) as control. For the competition experiment, mice received an oral gavage of 200 μL of pasteurized EcN (10⁸ CFU) weekly for 3 weeks. At 9 weeks of age, mice were anesthetized with an intraperitoneal injection of sodium pentobarbital and mechanically killed by cervical dislocation. The small intestine and colon were cut in the longitudinal part, and two persons, blinded to the groups, separately counted the number of tumors. Colonic “Swiss rolls” were fixed by 10% buffered formalin (ChapTec Inc., Montreal, QC, Canada) and then embedded in a paraffin block.

Total DNA was extracted from fecal samples from mice with the PowerSoil® DNA Extraction Kit (Qiagen Inc., Toronto, ON, Canada) and PCR was performed using PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific, Waltham, MA, United States) in the RG 3000A R PCR machine (Qiagen Inc.) (Fillebeen et al., 2015). Simultaneous amplification of colibactin A gene (clbA) and E. coli 16 s rRNA were done with the pair of primers clbA and Ecol16S, respectively (Supplementary Table S1). Amplification of colibactin P gene (clbP) and 16S rRNA were performed using the primers clbP and universal 16S, respectively (Supplementary Table S1). To verify the depletion of the clbP gene in the EcNC101 mutant, we used three primers CheckClbP (Supplementary Table S1).

Colon “Swiss rolls” were placed in optimal cutting temperature (OCT) compound and frozen on dry ice. Sections cut at 5-μm thickness were fixed with 4% paraformaldehyde, permeabilized in 0.1% triton in phosphate-buffered saline (PBS) and blocked in PBS containing 1% bovine serum albumin (BSA) and 4% serum. DNA was stained with DAPI, followed by F-actin staining with Alexa Fluor™ 647 Phalloidin (1/200 in methanol 1X) (a gift from Dr. Larochelle, CRCHUM, University of Montreal). Mucus layer was imaged using acid-Schiff-Alcian blue (Artisan Alcian Blue kit, AR160, Dako North America Inc. Carpinteria, CA; PAS, Sigma–Aldrich). All sections were scanned using a VS-110 microscope with a 40× 0.75 NA objective and a resolution of 0.3225 mm (Olympus), and images were then generated using Fiji software.

Sections were cut at 4-μm thickness and stained with hematoxylin and eosin (H&E), and assessed for low grade adenomas, high grade adenomas, and intramucosal carcinomas.

Formalin-fixed paraffin-embedded (FFPE) sections of colonic tissue were stained using the Benchmark XT autostainer (Ventana Medical Systems, Tucson, AZ, United States). Immunohistochemical staining was carried out on frozen sections using specific antibodies anti-Ki-67 (Biocare CRM325A, Biocare Medical, Pacheco, CA, United States) and the anti-phospho-γ-H2AX (Ser139, Cell Signaling Technology #9718, New England BioLabs Ltd.). Reactions were performed using the iView DAB detection kit, and counterstaining was achieved with hematoxylin and bluing reagents at 1/150 dilution. DSBs and proliferative indexes were represented by the number of positive cells in the descending colon.

Samples of digesta (small intestine) and feces (colon) (Wienk et al., 1997) were plated on MacConkey agar (Thermo scientific Oxoid, Nepean, ON, Canada) and E. coli was quantified based on the number of red colonies (small intestine: CFUs/mL; colon: CFUs/g).

Lactobacillus plantarum 299 V® was used as a positive control for inulin fermentation and identified by PCR using the pair of primers Plantarum (Supplementary Table S1). Bacteria were sub-cultured at 1/100 dilution in standard minimal medium (M9) supplemented with 1% inulin for 24 h in anaerobic conditions using an anaerobic sachet (BD BBL™ GasPak™ anaerobic indicator, BD, Mississauga, ON, Canada) and the optical density at 600 nm (OD600) was determined in a Spark® multimode microplate reader (Tecan Group Ltd., Männedrof, Switzerland) every hour for a day. In parallel, the release of free fructose and glucose in the supernatants during inulin fermentation were determined after 3 h of culture using a D-fructose/D-glucose assay kit (Megazyme International Ltd., Wicklow, Ireland).

To measure EcNC101 growth compared to EcN growth, we transformed strains with the plasmid pUCP20T-E2Crimson and pUCP20T-morange (gifts from Mariette Barbier; Addgene plasmid #78473²; RRID:Addgene_78,473/Addgene plasmid #78468³; RRID:Addgene_78,468) to distinguish them by crimson and orange fluorescence (Barbier and Damron, 2016). Electroporation was used to insert plasmids into the bacteria. Beforehand, we measured the effect of the plasmid on the growth of the bacteria and validated the fluorescence intensity by transfecting each strain with each plasmid and compared their growth to a control strain without plasmid. There were no variations in the bacteria’s growth or fluorescence intensity. Next, the competitor strains (10⁷CFUs/mL) were inoculated into a starting volume of 5 ml LB and grown at 37°C for 24 h without shaking. After 1 day, 50 μL of the competitor-strains culture was inoculated into fresh LB (1/100) and grown for 24 h, and this was repeated for 3 days. Fluorescence was recorded every day from the starting point in a Spark® multimode microplate reader (Tecan Group Ltd). To measure the effect of pasteurized EcN on EcNC101 growth, EcN was inoculated in fresh LB medium overnight at 37°C. Then, the suspension was centrifuged at 5,000 rpm for 3 min. Finally, the cell pellet was resuspended in fresh LB medium and pasteurized at 70°C for 30 min. EcNC101 was inoculated at 10⁷ CFUs/mL in LB medium composed of 50% or 100% of pasteurized EcN. OD600 was measured at 0, 3, 6, and 24 h.

All data were analyzed using GraphPad Prism (Version 5.0, GraphPad Software, San Diego, CA, United States). χ² tests were used to compare categorical variables. When the data did not pass the normality test, log(Y) transformation was applied to the data. An analysis of variance (ANOVA) was followed by post hoc Tukey’s multiple comparisons test. For tumor count, non-parametric Mann–Whitney test was used. p-values <0.05 were considered statistically significant.

Inulin is a common functional ingredient and supplement with prebiotic benefits (Christoforou et al., 2021). However, our previous findings showed that in vitro, inulin may enhance the genotoxicity of pks + E. coli that produce colibactin (Oliero et al., 2021). To determine the potential risk of inulin supplementation for CRC development, we investigated whether early colonization with pks + E. coli affected the response to inulin supplementation on tumor development in Apc^Min/+ mice. As shown in Figure 1A, DSS treatment resulted in temporary damage to the mucus layer, which would promote the direct contact between pks + E. coli and intestinal epithelial cells necessary for colibactin action (Nougayrede et al., 2006). Mice received an oral gavage of PBS control or the murine strain of colibactin-producing EcNC101. While all mice had E. coli in their gut, colibactin-producing bacteria was only present in mice inoculated with EcNC101 (Figure 1B).

Body weight progress was similar in all four experimental groups during the first 4 weeks, after which, mice colonized with EcNC101 on the inulin diet had a significantly lower body weight compared to the uninfected mice fed the same diet (Figure 1C). As shown in Figure 1D, the number of tumors in the small intestine was significantly higher in Apc^Min/+ mice colonized with EcNC101 and fed inulin than in uninfected mice on the same diet (1.7-fold higher). However, EcNC101 colonization did not affect the number of intestinal tumors when mice received the cellulose diet. Finally, when comparing the effect of diet on mice colonized with EcNC101, mice fed the inulin diet had a higher tumor count compared with those fed cellulose (1.7-fold change). These results indicate that the tumorigenic effect of EcNC101 in Apc^Min/+ mice in the small intestine was dependent on the diet.

In the colon, the number of tumors was similar in the four groups (Figure 1E). However, the size of the tumors found in EcNC101 positive Apc^Min/+ mice fed inulin was significantly larger compared to EcNC101 negative mice on the same diet (1.92 ± 0.15 mm vs. 1.37 ± 0.13 mm, ±SEM; Figures 1F,G). In addition, EcNC101 colonization was associated with a significantly higher tumor grade based on the appearance of intramucosal carcinomas that were never observed in EcNC101 negative mice. In addition, 71% of mice fed the inulin diet had intramucosal carcinomas compared to 11% of mice fed cellulose (Figure 1H; Supplementary Figure 1), indicating that tumor progression in the colon of mice colonized with pks + bacteria depends on the diet.

In line with these findings, EcNC101 colonization led to increased cell proliferation as assessed by Ki-67 expression (Luo et al., 2019), regardless of diet, and to increased levels of γ-H2AX staining, a marker of DSBs (Mah et al., 2010), in mice fed the inulin diet (Figures 1I,J; Supplementary Figures 2A,B). Finally, higher levels of the clbP gene were present in fecal samples from Apc^Min/+ mice fed the inulin diet (Figure 1K).

Overall, these results show that in the presence of colibactin-producing E. coli, inulin accelerates tumor progression in the gastrointestinal tract of Apc^Min/+ mice.

To investigate whether colibactin is necessary for the enhanced tumorigenesis in mice colonized with pks + E. coli, Apc^Min/+ mice fed an inulin diet were inoculated with EcNC101 WT or with the colibactin-deficient mutant strain EcNC101 ΔclbP (Dubois et al., 2011). The quantification of CFUs revealed similar abundance levels between EcNC101 WT and EcNC101 ΔclbP in the feces 4 weeks post-inoculation (Figures 2A,B) indicating that further effects were not caused by colonization levels of the various mutants.

Apc^Min/+ mice colonized with WT or ΔclbP mutant had a significantly lower body weight compared to the uninfected mice (Figure 2C). However, tumor counts in the small intestine in mice colonized with the ΔclbP mutant was comparable to uninfected mice (25 ± 4.0 vs. 25 ± 5.8 tumors; ±SEM) and was significantly reduced compared to mice colonized with the WT strain (41 ± 4.9 tumors; ±SEM Figure 2D). In addition, tumor progression was also reduced in mice colonized with the ΔclbP mutant compared to WT as evidenced by tumors of smaller size (1.45 ± 0.11 mm vs. 2.04 ± 0.13 mm; ±SEM) and of lower grade in the colon (Figures 2E,F). Body weight changes in both WT and ΔclbP EcNC101 positive mice may be due to altered metabolic processes since chronic colonization with pathogenic E. coli strains can induce changes in inflammatory-associated genes, such as adiponectin, leptin and thrombopoietin (Stromberg et al., 2018). Alternatively, or concomitantly, E. coli colonization may lead to significant changes in gut microbiota composition, known to affect body weight (Million et al., 2013), which are independent of colibactin expression (Tronnet et al., 2020).

These results indicate that the deleterious effect of inulin in Apc^Min/+ mice colonized with EcNC101 is dependent on the ability of the E. coli strain to produce colibactin.

To further assess the effect of inulin on EcNC101 growth, we cultured the bacteria in the presence of inulin. The addition of inulin increased the growth of EcNC101 in a concentration-dependant manner (Figure 3A). We then investigated whether EcNC101 was able to ferment inulin for their growth and measured the concentration of monosaccharides released in the minimal culture medium inoculated with EcNC101. The inulin-fermenting strain Lactobacillus plantarum was used as a control and as expected (Nordstrom et al., 2021), increased the concentration of monosaccharides in the medium (1.4-fold change) with the addition of 1% inulin (Figure 3B). EcK12, used as a control strain unable to ferment inulin (Rossi et al., 2005; Jung et al., 2015), did not change the monosaccharide concentration. In contrast, EcNC101 cultures released monosaccharides into medium when supplemented with 1% inulin (1.4-fold change; Figure 3B). These data suggest that EcNC101 ferments inulin and thus, may explain the effect of inulin in promoting EcNC101 growth.

Next, we quantified EcNC101 colonization in Apc^Min/+ mice overtime by culturing fecal homogenates on MacConkey agar plates and counting CFUs. The EcNC101 colonization level were similar until week 2, after which EcNC101 levels in fecal samples were higher in the Apc^Min/+ mice fed an inulin supplemented diet compared to those fed the cellulose diet (Figure 3C). CFUs quantification at the endpoint confirmed that EcNC101 was present at higher concentrations in the small intestine and in fecal samples of Apc^Min/+ mice fed the inulin diet (ranging from 10⁵ to 10⁷ CFUs/mg feces) compared to mice fed the cellulose diet (ranging from10³ to 10⁶ CFUs/mg feces) (Figure 3C).

Together, these results suggest that EcNC101 may ferment inulin to fuel its growth, resulting in increased expansion of EcNC101 in Apc^Min/+ mice fed an inulin supplemented diet.

Competition for space and nutrients in the gut is prevalent among bacterial species belonging to the same family as resources and niche requirements are similar (Hibbing et al., 2010). Since complete eradication of pks + E. coli from the gut microbiome is not possible (Raimondi et al., 2019), we tested whether limiting the growth of EcNC101 via competition decreased tumorigenesis by using a well-characterized probiotic, E. coli Nissle 1917 (EcN) (Scaldaferri et al., 2016) in co-cultures with EcNC101. As shown in Figure 4A, EcN was able to outcompete EcNC101 after 3 days of successive cycles of inoculations. Unfortunately, previous studies have shown that the probiotic properties of EcN cannot be dissociated from the production of colibactin (Olier et al., 2012), which may complicate the use of live EcN in the context of CRC. To circumvent the ability of EcN to produce colibactin, we tested pasteurized EcN and showed that in vitro, it significantly inhibited the growth of pathogenic EcNC101 as shown in Figure 4B.

We next tested pasteurized EcN in Apc^Min/+ mice. EcN101-colonized mice that received oral gavage with pasteurized EcN showed lower concentrations of colibactin in fecal samples as assessed by quantitative PCR (Figure 4C). Accordingly, CFUs of EcNC101 were lowered in the small intestine (Figure 4D) and 7-fold reduced in Apc^Min/+ mice gavaged with pasteurized EcN compared to mice receiving PBS (Figure 4E). As expected, EcNC101 colonization led to a decrease of body weight gain interdependent of further treatments (Figure 4F). In mice colonized with EcNC101, pasteurized EcN lowered tumor counts in the small intestine compared to mice treated with PBS (27 ± 3.0 vs. 38 ± 3.4 tumors; ±SEM Figure 4G) and was associated with smaller colonic tumors (1.53 ± 0.09 mm vs. 1.89 ± 0.13 mm; ±SEM Figures 4G–I). In addition, intramucosal carcinomas in the colon were only found in one out of 10 mice receiving pasteurized EcN whereas 6 out of 10 mice treated with PBS developed intramucosal carcinomas (Figure 4J).

Taken together, these results show that the deleterious effect of inulin in mice colonized by EcNC101 can be mitigated by using pasteurized EcN.

Our results confirms a previous study (Lopès et al., 2020) that demonstrates the carcinogenic potential of colibactin in mice fed a regular diet. We observed an increase in the number of tumors in the small intestine, and in the size and grade of colonic tumors in Apc^Min/+ mice colonized with the pathogenic strain EcNC101. Discrepancies with previous studies regarding enhanced tumor counts in the colon of EcN101 positive mice may be related to differences in E. coli strains employed and/or the duration of the studies (Bonnet et al., 2014; Lucas et al., 2020). Another colibactin-producing strain, E. coli MP13, was shown to promote carcinogenesis in the azoxymethane (AOM)/DSS CRC model (Zhu et al., 2019).

Inulin enhanced carcinogenesis in Apc^Min/+ mice colonized with pks + E. coli. Similarly, a diet supplemented with 10% inulin was found to increase the tumor promotion explained by higher level of cytosolic β-catenin (Pajari et al., 2003) and higher expression of cyclin D1 (Misikangas et al., 2008). In contrast, others have reported that dietary supplementation with 15% inulin decreased tumor sizes in syngeneic mice transplanted with melanoma cells subcutaneously (Li et al., 2020) and both 5 and 15% inulin supplemented diet reduced colonic tumoral load in AOM-treated Apc^Min/+ mice (Moen et al., 2016). In humans, inulin consumption was suggested to prevent CRC (Pool-Zobel and Sauer, 2007). However, more recent studies failed to find any association between inulin-type fructans and CRC prevention (Turati et al., 2022). Here we show that pks + E. coli negates the effect of inulin supplementation. When evaluating the effect of inulin, none of the previous studies assessed the presence of colibactin-producing bacteria, which could, at least partially, explain the variations between studies.

Increased carcinogenesis was associated with a greater abundance of EcNC101 in mice fed an inulin diet, further supporting previous studies showing that inulin promotes the growth of pks + E. coli in vitro (Oliero et al., 2021). Inulin fermentation supports the growth of many strains of Bifidobacterium and Lactobacillus (Zhu et al., 2020), including the probiotic L. plantarum 299 V, whereas E. coli strains are often compared as negative controls (Rossi et al., 2005; Jung et al., 2015). However, we found that the pathogenic strain EcNC101 may ferment inulin as a source for its growth, releasing monosaccharides into the culture medium.

We showed that the deleterious effect of EcNC101 increased-growth was mediated by the production of colibactin and was mitigated by addition of the probiotic EcN strain outcompeting pathogenic bacteria (Scaldaferri et al., 2016). EcN has been shown to be able to suppress the growth of Salmonella typhimurium by competing for iron acquisition through siderophore production (Deriu et al., 2013) and to limit the growth of related competitors, including pathobionts from the Enterobacteriaceae family, through the release of small antimicrobial molecules called microcins (Sassone-Corsi et al., 2016). This growth inhibition could be mediated by outer membrane vesicles (OMVs) produced by EcN that are able to stimulate the host immune responses (Cañas et al., 2016; Fábrega et al., 2016), to reinforce the epithelial barrier (Alvarez et al., 2016), and to reduce colitis in the DSS mouse model (Fábrega et al., 2017). OMVs from gram negative bacteria play a key role in nutrients intake, and in the release of bacteriocins (Schwechheimer and Kuehn, 2015) and proteases (Jan, 2017). Since EcN has been identified as potentially harmful for the host due to the presence of colibactin genes (Nougayrède et al., 2021), its pasteurization would have the advantage of resulting into colibactin degradation due to its high instability (Dougherty and Jobin, 2021). Here, we show that pasteurized EcN was able to suppress inulin-enhanced EcNC101 colonization in the gut of Apc^Min/+ mice resulting in lower tumorigenesis. Importantly, heat treatment has been shown to preserve the ability of EcN to inhibit bacterial competitive growth (Lagos, 2013). In addition, pasteurized EcN has been reported to have a direct anti-cancer effect through the regulation of signaling pathways in colonic cells (Alizadeh et al., 2020) and may lower inflammation in the intestine (Fábrega et al., 2017). The use of pasteurized EcN probiotic (paraprobiotic) instead of live bacteria has many advantages, the most important being the abolition of colibactin secretion, followed by greater safety, and improved logistic parameters of supplementation such as storage and shelf life (Piqué et al., 2019).

In conclusion, we show that inulin promotes EcNC101 growth in Apc^Min/+ mice, resulting in enhanced colonization, increased DSBs, cell proliferation, and tumorigenesis (Figure 5). These results suggest that inulin supplementation may not be appropriate for all individuals, depending on the composition of their gut microbiota. Our findings highlight the necessity of screening patients for pks + bacteria and providing them with appropriate preventive dietary counseling. Further studies are needed to investigate the interaction between dietary supplements and cancer-promoting bacteria, such as colibactin-producing bacteria.