Colorectal cancer specific conditions promote Streptococcus gallolyticus gut colonization

Abstract

Colonization by Streptococcus gallolyticus subsp. gallolyticus (SGG) is strongly associated with the occurrence of colorectal cancer (CRC). However, the factors leading to its successful colonization are unknown, and whether SGG influences the oncogenic process or benefits from the tumor-prone environment to prevail remains an open question. Here, we elucidate crucial steps that explain how CRC favors SGG colonization. By using mice genetically prone to CRC, we show that SGG colonization is 1,000-fold higher in tumor-bearing mice than in normal mice. This selective advantage occurs at the expense of resident intestinal enterococci. An SGG-specific locus encoding a bacteriocin ("gallocin") is shown to kill enterococci in vitro. Importantly, bile acids strongly enhance this bacteriocin activity in vivo, leading to greater SGG colonization. Constitutive activation of the Wnt pathway, one of the earliest signaling alterations in CRC, and the decreased expression of the bile acid apical transporter gene Slc10A2, as an effect of the Apc founding mutation, may thereby sustain intestinal colonization by SGG. We conclude that CRC-specific conditions promote SGG colonization of the gut by replacing commensal enterococci in their niche.

Keywords: APC/Notch; S. bovis; S. gallolyticus; bacteriocin; colorectal cancer.

Fig. 1.

Presence of colonic polyps enhances gut colonization by SGG UCN34. (A) cfu counts of SGG UCN34 in the stool of tamoxifen (Tmx)-injected Notch and Notch/APC mice. SGG was inoculated by oral gavage 2 mo following tamoxifen injection, and cfu counts were enumerated 7 d later (Prt 1). (B) cfu counts of SGG UCN34 in the stool of Notch and Notch/APC mice at different time points following its oral inoculation. SGG was administered 1 mo before tamoxifen injection following a 1-wk antibiotic (Ab) treatment. Macroscopic adenoma count in the colon of SGG colonized vs. PBS solution-treated Notch/APC mice at the time of euthanasia, 2 mo following tamoxifen injection (mean ± SEM; **P ≤ 0.01 and *P ≤ 0.05, t test). (C) Mean ± SEM cfu counts of SGG UCN34 in the ileum and proximal (Pxi) and distal (Ds) colon of Notch, Notch/APC, WT, and APC mice 3 mo after oral inoculation (Prt 2; *P ≤ 0.05, t test). (D) Immunofluorescence staining of SGG UCN34 in Notch/APC colonic tissues 3 mo after its inoculation (Prt 2) shows tumoral (marked as “T”) and adjacent nontumoral (NT) areas. Bacteria (in red) were stained with a specific rabbit anti-UCN34 polyclonal antibody and anti-rabbit IgG coupled to Alexa Fluor 568, respectively. Mucus was visualized with wheat germ agglutinin (WGA) lectin coupled to Alexa Fluor 488 (green), and nuclei were labeled with DAPI (cyan blue). ns, not significant. (Scale bar, 100 µm.)

Fig. 2.

Involvement of a specific bacteriocin in the persistence of SGG in tumor-bearing mice. (A) Mean ± SEM cfu counts of SGG UCN34 in the stool of SPF and GF Notch and Notch/APC mice 3 mo after SGG oral inoculation (**P ≤ 0.01, t test). (B) cfu counts of SGG UCN34 in the stool of GF mice colonized with microbiota from Notch/APC or Notch mice (n = 5). (C) cfu counts of enterococci in the stool of Notch and Notch/APC mice at different time points following SGG UCN34 oral inoculation (D0). Enterococci cfu counts in the stool of Notch/APC mice receiving PBS solution instead of SGG are also depicted (mean ± SEM; **P ≤ 0.01, paired t test). (D) Agar diffusion assay showing the capacity of SGM and different strains of SGG [UCN34, Δblp, bWT, Δblp(pTCVblp)] to inhibit the growth of commensal E. faecalis. Δblp, bacteriocin-deficient mutant; Δblp(pTCVblp), complemented Δblp mutant; bWT, back to WT strain. (E) Mean ± SEM cfu counts of the Δblp mutant in the ileum, proximal colon, and stools of Notch, Notch/APC, WT, and APC mice 3 mo after its oral inoculation. ns, not significant.

Fig. 3.

Bile acid deoxycholate enhances SGG bacteriocin activity in vitro and in vivo. (A) Inhibitory potential of cell-free (CF) sterile-filtered supernatant (SN) of different SGG strains (UCN34, bWT, Δblp mutant) against E. faecalis growth. Supernatants from overnight cultures were used alone or in combination with Tween 20 (Tw20) or DCA. (B) Growth curves of E. faecalis in TH broth containing 50% of SGG UCN34 sterile filtered supernatant (SN) from the bWT or Δblp mutant strain and supplemented with different concentrations of DCA or GDCA. (C) Mean ± SEM cfu counts of SGG UCN34 or the Δblp mutant in the stools 7 d after oral inoculation in C57BL/6 mice treated with different doses of DCA for 14 d (0.1% or 0.3%, wt/wt ratio; *P ≤ 0.05, t test). ns, not significant.

Fig. 4.

Increased excretion of secondary bile acids in Notch/APC mice. (A) Mean ± SEM cfu counts of SGG UCN34 in the stool of Notch/APC mice 1 d before the beginning and 1 d after the end of a 10-d CS oral treatment (*P ≤ 0.05, paired t test). (B) Mean ± SEM fecal concentration of primary (Iry), secondary (IIry), and conjugated (Conj) bile acids (BA), DCA, and LCA in Notch and Notch/APC mice 2 mo after Notch receptor activation (**P ≤ 0.01, t test). (C) Slc10A2 gene relative expression level in the ileum and proximal colon of WT, Notch, and Notch/APC mice 2 mo after Notch receptor activation (mean ± SEM; **P ≤ 0.01, t test). (D) Slc10A2 gene expression level in the ileum and proximal colon of 3-mo-old WT and APC mice. (E) Slc10A2 and Mrp3 gene expression level in small intestinal organoids produced from C57BL/6 mice and cultivated for 20 h in the presence of different concentrations of Wnt3A (0, 100, 200, and 500 ng/mL). ns, not significant.