Murine Model for Measuring Effects of Humanized-Dosing of Antibiotics on the Gut Microbiome

Abstract

There is a current need for enhancing our insight in the effects of antimicrobial treatment on the composition of human microbiota. Also, the spontaneous restoration of the microbiota after antimicrobial treatment requires better understanding. This is best addressed in well-defined animal models. We here present a model in which immune-competent or neutropenic mice were administered piperacillin-tazobactam (TZP) according to human treatment schedules. Before, during and after the TZP treatment, fecal specimens were longitudinally collected at established intervals over several weeks. Gut microbial taxonomic distribution and abundance were assessed through culture and molecular means during all periods. Non-targeted metabolomics analyses of stool samples using Quadrupole Time of Flight mass spectrometry (QTOF MS) were also applied to determine if a metabolic fingerprint correlated with antibiotic use, immune status, and microbial abundance. TZP treatment led to a 5-10-fold decrease in bacterial fecal viability counts which were not fully restored during post-antibiotic follow up. Two distinct, relatively uniform and reproducible restoration scenarios of microbiota changes were seen in post TZP-treatment mice. Post-antibiotic flora could consist of predominantly Firmicutes or, alternatively, a more diverse mix of taxa. In general, the pre-treatment microbial communities were not fully restored within the screening periods applied. A new species, closely related to Eubacterium siraeum, Mageeibacillus indolicus, and Saccharofermentans acetigenes, became predominant post-treatment in a significant proportion of mice, identified by 16S rRNA gene sequencing. Principal component analysis of QTOF MS of mouse feces successfully distinguished treated from non-treated mice as well as immunocompetent from neutropenic mice. We observe dynamic but distinct and reproducible responses in the mouse gut microbiota during and after TZP treatment and propose the current murine model as a useful tool for defining the more general post-antibiotic effects in the gastro-intestinal ecosystem where humanized antibiotic dosing may ultimately facilitate extrapolation to humans.

Keywords: gastro-intestinal tract; humanized antibiotic treatment; metabolome; microbiome dynamics; non-infected mouse model; piperacillin-tazobactam.

Figure 1

Bacterial quantification of the TZP2 experiment. Box and whiskers plots show the fold change of the 16S rRNA gene count as generated by quantitative PCR (qPCR) from DNA extracted from stool for all mice (immunocompetent and neutropenic shown separately) at each time point as compared to Day 14, the time-point immediately prior to antibiotic treatment.

Figure 2

Principal coordinate analysis (PCoA) of taxonomic profiles. PCoA analysis based on the Bray-Curtis distance matrix of OTUs from TZP1 (A) and TZP2 (B) experiments. Each circle represents a single mouse at a single time point. Closed circles represent mice that had >50% Clostridiales (OTU-1 in A, OTU-3 in B) at any time point and open circles represent all other mice. (C) PCoA analysis on Bray-Curtis distance matrix based on combined genera profiles from both experiments. Note that the number of bullets per panel may vary as a consequence of the number of mice per group and the number of samples taken and analyzed in full per individual mouse.

Figure 3

Effect of piperacillin-tazobactam (TZP) on the Alpha Diversity. The alpha diversity of each mouse in each time point was measured using Pielou’s evenness, absolute number of OTUs (richness), and Shannon diversity index (diversity). Significant differences between two time points are indicated by asterisks where p > 0.05 = ns, p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, and p < 0.0001 = ****. Panel (A) shows results obtained during the TZP1 study, whereas panel (B) displays the results obtained during the TZP2 study.

Figure 4

Bacterial community structure. The OTUs for all mice in each experiment (TZP1 in A, TZP2 in B) are plotted by ranked abundance for each time point.

Figure 5

Taxonomic profiles. The relative abundance of taxa is represented by stacked bars for each mouse and each day. The top 25 OTUs are colored individually and all other OTUs are assigned to the “Other Taxa” group (light gray). (A) Portrays TZP1 in which TZP was administered days 7–13, (B,C) portray immunocompetent mice (mice 1–10) and neutropenic mice (mice 11–20), respectively, from TZP2 in which TZP was administered on days 15–18.

Figure 6

Resistant Clostridiales OTUs. The fold change of relative abundance of Clostridiales OTUs 1 and 3, (A,B, TZP1 and TZP2, respectively) compared to the last pre-antibiotic time point (Days 6 and 14, respectively) are plotted for each day. The vertical line indicates the approximate first TZP exposure. Mice 1–10 were immunocompetent while mice 11–20 were neutropenic.

Figure 7

OTU-1 Abundance. Absolute abundance of OTU-1 in TZP1 estimated by targeted qPCR.

Figure 8

Phylogenetic tree for the newly discovered dominating species in some of the mice post antibiotic treatment.

Figure 9

Metabolomic profiling. Principal component analysis of non-targeted TripleTOF mass spectrometer (MS) data of mouse feces successfully grouped treatment from non-treatment mice. The scores plot for PC1 and PC3 of MS features in positive ion mode shows good separation of the antibiotic vs. non-antibiotic time points. Furthermore, clusters of post-antibiotic immunocompetent and post-antibiotic neutropenic features also show good separation from each other with immunocompetent clustering above and neutropenic below the x-axis.