Structural and Functional Alterations in the Microbial Community and Immunological Consequences in a Mouse Model of Antibiotic-Induced Dysbiosis

Abstract

The aim of this study was to establish continuous therapeutic-dose ampicillin (CTDA)-induced dysbiosis in a mouse model, mimicking typical adult exposure, with a view to using this to assess its impact on gut microbiota, intestinal metabolites and host immune responses. Mice were exposed to ampicillin for 14 days and antibiotic-induced dysbiosis was evaluated by alteration of microbiota and gut permeability. The cecal index was increased in the CTDA group, and the gut permeability indicated by fluorescent dextran, endotoxin and D-Lactate in the serum was significantly increased after antibiotic use. The tight-junction proteins ZO-1 and occludin in the colon were reduced to half the control level in CTDA. We found that alpha-diversity was significantly decreased in mice receiving CTDA, and microbial community structure was altered compared with the control. Key taxa were identified as CTDA-specific, and the relative abundance of Enterococcus and Klebsiella was particularly enriched while Lachnospiraceae, Coprobacillus and Dorea were depleted after antibiotic treatment. In particular, a significant increase in succinate and a reduction in butyrate was detected in CTDA mice, and the triggering of NF-κB enhancement reflected that the host immune response was influenced by ampicillin use. The observed perturbation of the microbiota was accompanied by modulation of inflammatory state; this included increase in interferon-γ and RegIIIγ, and a decrease in secretory IgA in the colon mucosa. This study allowed us to identify the key taxa associated with an ampicillin-induced state of dysbiosis in mice and to characterize the microbial communities via molecular profiling. Thus, this work describes the bacterial ecology of antibiotic exposure model in combination with host physiological characteristics at a detailed level of microbial taxa.

Keywords: ampicillin; gut dysbiosis; inflammation; metabolite; mice model; microbiota.

FIGURE 1

α- and β-diversity in Control, CTDA1, and CTDA2 fecal samples. (A) α-diversity as measured by the Shannon diversity index of microbiota in fecal samples from different groups. The CTDA was administrated for 14 days and fecal samples were collected at day 7 (CTDA1 group) and day 14 (CTDA2 group), and the 0.9% saline solution was administrated for 14 days and fecal samples were collected at day 14 (Control group), n = 8 per treatment group. The Shannon diversity index for each microbial community was calculated at a depth of 5000 sequences per sample. Two-tailed student’s t-test was used to determine statistical significance, ^∗∗P < 0.01(B) Visualization of principal coordinates analysis (PCoA) of unweighted UniFrac distances to show differences in bacterial community structure. The first principal component (PC1) and second principal component (PC2) explained 19.93 and 8.21% of the variance in the unweighted UniFrac metrics, respectively. Each point represents the bacterial fecal microbiota in a single sample. (C) Pairwise unweighted Unifrac distance measurements of microbiota in fecal samples from Control, CTDA1, and CTDA2 mice. Means ± SEM of the distances are shown. A two-tailed student’s t-test was used to determine statistical significance, ^∗∗P < 0.01. C = Control group, C1 = CTDA1 group, C2 = CTDA2 group.

FIGURE 2

Characteristics of microbiota from Control and CTDA fecal samples. (A) Relative abundance of bacterial phyla in fecal samples from Control, CTDA1 and CTDA2 groups of mice. (B) Heatmap of differential taxa. Taxonomic Heatmap using Bray-Curtis Dissimilarity index distance, combined with Average (unweighted pair-group method with arithmetic means) clustering for 41 of the most statistically significant OTUs between Control and CTDA samples. Yellow and blue represent high and low abundance, respectively. (C) Identification of significant differences in bacterial taxa between CTDA2 and Control group. Cladogram depicting differences and phylogenic location are shown. In each section, the diameter of the circle is proportional to the abundance of the taxon. Fecal microbial communities from CTDA2 and Control mice were compared using LEfSe (green = taxon significantly enriched in Control; red = taxon significantly enriched in CTDA2; yellow = non-significant). (D) Histogram of the LDA scores computed for features differentially abundant between the CTDA2 and Control groups. LEfSe scores represent the degree of consistent difference in relative abundance between features in two groups of analyzed microbial communities. The clades of the histogram (red indicating CTDA2 group, green indicating Control group) identifies statistical and biological differences between the communities. (E) Taxa co-occurrence network. Cytoscape was used to visualize co-occurrence patterns for CTDA2 and Control mice at the genera level. In the network, samples were colored by treatment group (Control samples = gray; CTDA2 samples = orange). Individual genera are represented by small black nodes. Lines connecting a sample to a genus, colored by treatment group, indicate that the genus was observed in that specific sample. (F) Relative abundance of Enterococcus, Klebsiella, Enterobacter, Akkermansia, and Lachnospiraceae in fecal samples from the CTDA and Control groups of mice. Boxes represent the interquartile range (IQR), median (horizontal line within the box), range (whiskers) and outliers (crosses) (>1.5^∗IQR). A two-tailed student’s t-test was used to determine statistical significance, ^∗∗P < 0.01.

FIGURE 3

Metabolic and immunological outcomes of the CTDA mouse treatment. (A) Distribution of intensities for selected metabolites in the feces from Control and CTDA groups of mice. Horizontal lines show mean and SEM. A two-tailed student’s t-test was used to determine statistical significance, ^∗P < 0.05, ^∗∗P < 0.01. (B) Levels of inflammatory cytokines measured by ELISA in colonic tissues from Control and CTDA2 groups of mice. A two-tailed student’s t-test was used to determine statistical significance, ^∗∗P < 0.01. Error bars represent mean ± SEM (n = 8).

FIGURE 4

Establishing a CTDA mouse model based on key host biomarkers (A) = weight and (B) = food intake of mice. Day 4 was the first day of the adaptive phase, ampicillin administration started on day 0, and the end point of ampicillin use was day 14. A two-tailed student’s t-test was used to determine statistical significance, ^∗P < 0.05. Error bars represent mean ± SEM (n = 8). (C) Main factors including cecal indices in cecum, biomarkers associated with the gut barrier isolated from the serum and expression of tight-junction proteins in the colon of Control and CTDA2 mice. A two-tailed student’s t-test was used to determine statistical significance, ^∗∗P < 0.01. Error bars represent mean ± SEM (n = 8). (D) Images representing tight junction disruption (occludin) of the ileum in the Control and CTDA2 groups of mice as determined by immunofluorescence (Representative images, n = 4/group). (E) Non-parametric monotonic relationship between the dissimilarities in the samples of CTDA2 group (blue, right) and Control group (red, left) matrix. The spots site based on the abundance of microbial taxa, measurement of cecal, colon, and serum samples which listed in C. (F) Non-metric multidimensional scaling plot of family and genus compositions for the Control (red) and CTDA2 samples (blue) with the three featured variables (purple arrows) (DAO = diamine oxidase, ET = endotoxin, and ZO1 = tight-junction protein ZO-1) plotted using the EnvFit function of package in R. The gray arrows give the top eight families or genera that most differ in expected proportion. f_Enterobacteriaceae, g_K = f_Enterobacteriaceae, g_Klebsiella, f_Erysipelotrichaceae, g_Coproba = f__Erysipelotrichaceae, g__Coprobacillus, f_Enterococcaceae, g_Enteroco = f_Enterococcaceae, and g_Enterococcus.