Interrelated Effects of Zinc Deficiency and the Microbiome on Group B Streptococcal Vaginal Colonization

Abstract

Group B Streptococcus (GBS) in the vaginal tract is a risk factor for preterm birth and adverse pregnancy outcomes. GBS colonization is also transient in nature, which likely reflects the contributions of pathogen determinants, interactions with commensal flora, and host factors, making this environment particularly challenging to understand. Additionally, dietary zinc deficiency is a health concern on the global scale that is known to be associated with recurrent bacterial infection and increased rate of preterm birth or stillbirth. However, the impact of zinc deficiency on vaginal health has not yet been studied. Here we use a murine model to assess the role of dietary zinc on GBS burden and the impact of GBS colonization on the vaginal microbiome. We show that GBS vaginal colonization is increased in a zinc-deficient host and that the presence of GBS significantly alters the microbial community structure of the vagina. Using machine learning approaches, we show that vaginal community turnover during GBS colonization is driven by computationally predictable changes in key taxa, including several organisms not previously described in the context of the vaginal microbiota, such as Akkermansia muciniphila. We observed that A. muciniphila increases GBS vaginal persistence and, in a cohort of human vaginal microbiome samples collected throughout pregnancy, we observed an increased prevalence of codetection of GBS and A. muciniphila in patients who delivered preterm compared to those who delivered at full term. These findings reveal the importance and complexity of both host zinc availability and native microbiome to GBS vaginal persistence. IMPORTANCE The presence of group B Streptococcus (GBS) in the vaginal tract, perturbations in the vaginal microbiota, and dietary zinc deficiency are three factors that are independently known to be associated with increased risk of adverse pregnancy outcomes. Here, we developed an experimental mouse model to assess the impact of dietary zinc deficiency on GBS vaginal burden and persistence and to determine how changes in GBS colonization impact vaginal microbial structure. We have employed unique animal, in silica metabolic, and machine learning models, paired with analyses of human cohort data, to identify taxonomic biomarkers that contribute to host susceptibility to GBS vaginal persistence. Collectively, the data reported here identify that both dietary zinc deficiency and the presence of A. muciniphila could perpetuate an increased GBS burden and prolonged exposure in the vaginal tract, which potentiate the risk of invasive infection in utero and in the newborn.

Keywords: Akkermansia muciniphila; GBS; Streptococcus agalactiae; dietary zinc; group B Streptococcus; vaginal colonization; vaginal microbiome.

FIG 1

Establishment and assessment of a murine model of dietary zinc deficiency during vaginal colonization. (A) Murine model of dietary zinc deficiency. (B) Weight gain of mice fed a control or zinc-deficient diet. (C and D) Confirmation of control (C) or dietary zinc deficiency (D) using GBS zinc-transport mutant in vivo. (E and F) GBS vaginal lumen colonization (E) and vaginal tissue GBS burden (F) during early stage colonization. Significance was determined by unpaired student’s two-tailed t test (B to D and F) and two-way ANOVA with Sídák’s multiple-comparison test (E), with *, P < 0.05; **, P < 0.01; ****, P < 0.0001; ns, not significant.

FIG 2

GBS colonization impacts mCST and the vaginal microbiome community. (A and B) Longitudinal assessment of mCSTs of mice prior to and throughout the course of GBS colonization. D, day. (C) Principal coordinate analyses (PCoA) of unweighted UniFrac distances with 95% confidence intervals colored by day and shaped by treatment. (D) Linear mixed-effects model of unweighted UniFrac principal coordinate axis 1 change over time and between treatment.

FIG 3

Random Forest machine learning models predict key taxa of the vaginal microbiota that drive diversity throughout GBS colonization. (A) Random forest regressor accuracy for predicting time using normalized species abundance. (B) The most important species for the random forest regressor to predict time. (C to F) Linear mixed-effects models for measuring the impacts of time and treatment on the CLR-transformed abundances of A. muciniphila (C), Muribaculaceae sp. (D), Peptococcaceae rc4-4 sp. (E), and S. agalactiae (GBS) (F).

FIG 4

Deciphering the impact of GBS-A. muciniphila (AM) interactions in murine and human vaginal health. (A) Detection of GBS, A. muciniphila, GBS and A. muciniphila, or neither in a human cohort of vaginal microbiome data from term or preterm births. (B) Cross-feeding direction of the eight metabolites most confidently predicted in silico to be needed for GBS and A. muciniphila cosurvival. (C) Adapted vaginal colonization model to assess GBS and A. muciniphila interactions in vivo. (D and E) GBS burden (D) and percent (E) of mice colonized with GBS in combination with A. muciniphila. Significance was determined by two-way ANOVA with Sídák’s multiple-comparison test (D) and log-rank test (E), with *, P < 0.05; **, P < 0.01; ***; P < 0.001; ****; P < 0.0001; ns, not significant.