Interpersonal Gut Microbiome Variation Drives Susceptibility and Resistance to Cholera Infection

Abstract

The gut microbiome is the resident microbial community of the gastrointestinal tract. This community is highly diverse, but how microbial diversity confers resistance or susceptibility to intestinal pathogens is poorly understood. Using transplantation of human microbiomes into several animal models of infection, we show that key microbiome species shape the chemical environment of the gut through the activity of the enzyme bile salt hydrolase. The activity of this enzyme reduced colonization by the major human diarrheal pathogen Vibrio cholerae by degrading the bile salt taurocholate that activates the expression of virulence genes. The absence of these functions and species permits increased infection loads on a personal microbiome-specific basis. These findings suggest new targets for individualized preventative strategies of V. cholerae infection through modulating the structure and function of the gut microbiome.

Keywords: bile; cholera; colonization resistance; infection; interpersonal variation; microbiome; pathogenesis.

Figure 1.. Model human gut microbiomes replicate structure of communities affected by diarrhea-induced dysbiosis.

(A) Defined human gut communities. (B) Composition of healthy US human donor fecal microbiomes. (C) Principal coordinates analysis of defined and complete human gut microbiomes based on weighted UniFrac distance, % variance explained shown in parentheses. Ellipses show 95% confidence intervals. (D) Weighted UniFrac distance to indicated defined human model microbiomes of fecal samples from cholera patients at the end of diarrhea (left) and healthy human donors (right) *P<0.05, ****P<0.0001, Mann-Whitney U-test. Boxplots show inter-quartile range, whiskers minimum to maximum.

Figure 2.. Gut microbiome composition contributes to V. cholerae infection resistance.

V. cholerae colonization in germfree adult mice harboring defined model communities co-gavaged with V. cholerae in (A) feces and (B) small intestines 4 days post infection. (C) Fecal V. cholerae colonization in GF adult mice harboring defined communities for 2 weeks and then gavaged with V. cholerae. Mix: 10 days DS colonization, followed by a SR microbes 4 days prior to V. cholerae infection. (D) Intestinal V. cholerae colonization of GF suckling mice co-gavaged with model communities and V. cholerae. (E) Intestinal V. cholerae colonization of antibiotic-treated CD-1 suckling mice co-gavaged with indicated communities. (F) 16S gene abundance in small intestine. (G) Expression of tcpA in infected mice with model human microbiomes. (H) T6SS effects on small intestine colonization in antibiotic-treated CD-1 pups. * P<0.05, ** P<0.01, *** P<0.001 (Mann-Whitney U-test); n.s. not significant. Error bars represent mean ± SEM. n=6–12 animals for all experiments.

Figure 3.. Addition of the CR model human microbiome to mice hosting DS microbes yields a community structure closer to complete fecal communities of healthy human volunteers.

(A, D) Principal coordinates analysis (PCoA) of microbial community diversity based on weighted UniFrac distance, % variance explained shown in parentheses for each axis. Ellipses show 95% confidence intervals. (A) PCoA of fecal samples and distal third of small intestine of GF mice with model communities during V. cholerae infection compared to healthy US donor fecal samples, with (B) PC1 positions and (C) all pairwise weighted UniFrac distances to healthy US donor fecal samples. (D) PCoA of model communities and healthy human donor communities in suckling mice, with (E) PC1 positions and (F) all pairwise weighted UniFrac distances to healthy US donor fecal samples. (G) Microbiome structure during infection with V. cholerae and host reads filtered (left) and in antibiotic-treated suckling mice without V. cholerae (right). (H) B. obeum abundance in adult GF mice containing indicated microbiomes during V. cholerae infection (4d post infection). * P<0.05, **** P<0.0001; n.s. not significant (Mann-Whitney U-test). Error bars represent mean ± SEM.

Figure 4.. Gut microbiome composition contributes to inter-personal differences in V. cholerae infection resistance.

Intestinal V. cholerae colonization of antibiotic-treated CD-1 pups colonized with complete fecal microbiomes from healthy US human volunteers. n=5–7 animals for all experiments.

Figure 5.. An unbiased combinatorial strategy for identifying commensal correlates of protection against and susceptibility to V. cholerae colonization.

(A) Combinatorial strategy. (B) 5-member microbiome embodiments randomly generated using CR/DS members (left). Resulting V. cholerae infection of antibiotic-treated suckling mice containing defined microbiome embodiments are shown at right. (C) Mean V. cholerae colonization in suckling mice bearing communities containing B. obeum or Streptococcus species alone and in combination. Data normalized across experiments as fold CFU gavaged V. cholerae recovered after infection. (D) Abundance of DS member species in randomized microbiomes correlated to resulting V. cholerae abundance after infection. Points represent mice receiving different microbiome embodiments. * P<0.05, ** P<0.01 (Mann-Whitney U-test); n.s. not significant. Error bars represent mean ± SEM.

Figure 6.. B. obeum exerts effects on V. cholerae colonization through degradation of the in vivo virulence gene activating signal taurocholate (TC).

P_tcpA activity normalized to tcpA induction by 125μM TC unless noted. (A) Modulation of tcpA-activating signals in suckling CD-1 mouse intestinal homogenates by pure cultures of B. obeum and S. salivarius, with heat treatment. (B) Bile effects on tcp gene expression in vitro. (C) Effects of CR and DS pure cultures on TC activation of virulence in vitro. (D) Effects of B. obeum bsh enzyme expression on TC-mediated tcp activation in vitro. (E) Mass spectrometry measurement of TC in suckling CD-1 mouse intestines after incubation with pure cultures of indicated strains. (F) V. cholerae infection of suckling CD-1 mice after 1-day of colonization with indicated E. coli strains. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001 (unpaired Student’s t-test). Error bars represent mean ± SEM. n=3–10 for all experiments.

Figure 7.. Levels of B. obeum bsh enzymes in human samples correlate to infection outcome, and can independently modulate V. cholerae colonization.

(A) Levels of phylotype 1 bsh enzymes in metagenomic sequencing of fecal microbiomes of cholera patients pre- (“Diarrhea (d0)”) and post-antibiotic (“Diarrhea + abx”) treatment, compared to healthy adults in Bangladesh. (B) Mass spectrometry measurement of bile levels in 125μM solutions of TC incubated with indicated cultured human fecal communities in vitro, †: TC not detected. (C) B. obeum bsh levels in intestines of antibiotic-cleared suckling CD-1 mice colonized with complex human fecal samples without V. cholerae, compared to V. cholerae colonization of antibiotic-cleared suckling animals bearing human donor microbiomes. * P<0.05, ** P<0.01, *** P<0.001 (Mann-Whitney U-test). Error bars represent mean ± SEM.