Early-life antibiotics attenuate regulatory T cell generation and increase the severity of murine house dust mite-induced asthma

Abstract

Introduction: Early-life exposure to antibiotics (ABX) has been linked to increases in asthma severity and prevalence in both children and laboratory animals. We explored the immunologic mechanisms behind this association using a mouse model of house dust mite (HDM)-induced asthma and early-life ABX exposure.

Methods: Mice were exposed to three short courses of ABX following weaning and experimental asthma was thereafter induced. Airway cell counts and differentials; serum immunoglobulin E (IgE); pulmonary function; lung histopathology; pulmonary regulatory T cells (Tregs); and the fecal microbiome were characterized following ABX exposure and induction of experimental asthma.

Results: Asthma severity was increased in mice exposed to ABX, including: airway eosinophilia, airway hyper-reactivity, serum HDM-specific IgE, and lung histopathology. ABX treatment led to sharp reduction in fecal microbiome diversity, including the loss of pro-regulatory organisms such as Lachnospira. Pulmonary Tregs were reduced with ABX treatment, and this reduction was directly proportional to diminished microbiome diversity.

Conclusion: Intermittent exposure to ABX early in life worsened the severity of experimental asthma and reduced pulmonary Tregs; the latter change correlated with decreased microbiome diversity. These data may suggest targets for immunologic or probiotic therapy to counteract the harmful effects of childhood ABX.

Figure 1. Early-Life Antibiotic Exposure Increases the Severity of Experimental Asthma

Experimental schematic is shown in (A), with colored symbols corresponding to groups in Figure 4. At sacrifice, broncho-alveolar lavage (BAL) was performed, total leukocytes counted (B), cellular differentials determined (C), and total eosinophils determined (D). Typical naïve BAL cell count is represented by a dotted line at 130,000 in (B). Serum taken at sacrifice was assessed for HDM-specific IgE using ELISA (E). Values in (B), (D), and (E) represent mean ± the SEM while data in (C) represent the mean. n=8 per group. ** p < 0.01- a: p < 0.01, HDM vs HDM + Antibiotic Mix eosinophil percent- b: p < 0.05, HDM vs HDM + Antibiotic Mix macrophage percent- c: p < 0.05, HDM vs HDM + Antibiotic Mix AUC.

Figure 2. Antibiotic Exposure Is Associated with Increased Disease Pathology

Formalin-fixed tissue sections were stained with hematoxylin and eosin (top row, A & B, 4X) or periodic acid- Schiff with hematoxylin counterstain (bottom row, C & D, 10X). Inflammation scores (E) and mucus scores (F) were determined in a blinded manner on a severity scale of 0-3 by five independent graders. Data represent mean ± interquartile ranges. n=3 slides per group. ** p < 0.01.

Figure 3. Antibiotic Exposure Significantly Increases Airway Hyper-Reactivity

Total respiratory system resistance (Rrs) was measured using the flexiVent system following increasing doses of aerosolized methacholine. Rrs was compared using Area Under the Curve (AUC). Data represent mean ± SEM. n=4 for HDM, n=9 for HDM + Antibiotic Mix. a: p < 0.05, HDM vs HDM + Antibiotic Mix AUC.

Figure 4. Antibiotic Exposure Reduces the Diversity of the Fecal Microbiome

QIIME was used to analyze 16S sequence data and pick operational taxonomic units (OTUs). Symbols in (A-C) refer to those in Figure 1A. OTUs are represented as bar graphs of the average percent of total organisms represented by each OTU in each group (A). In (A), p__ refers to phylum, c__ to class, o__ to order, f__ to family, and g__ to genus. OTUs lacking a lower classification refer to a single OTU classifiable only to the lowest taxonomic level noted. Principal coordinate analysis (PCoA) of the unweighted UniFrac distances was performed on samples before and after HDM exposure (B). In (B), the closer two points are to each other, the more similar they are to each other. Variance among samples is represented by the X and Y axes, where the percentage noted represents the proportion of all variance represented on that axis. The higher the number, the greater the dissimilarity a given distance along that axis represents. Simpson’s Diversity Index for the fecal microbiome following HDM exposure was calculated (C). n=8 for No Treatment and HDM + Antibiotic Mix, n=5 for HDM, n=6 for Antibiotic Mix. ** p < 0.01.

Figure 5. Antibiotic Exposure Results in Proportional Changes in Multiple Taxa

Linear Discriminant Analysis with Effect Size (LEfSe) was performed on the taxonomy table of abundances for all taxa based on (A) antibiotic treatment as the class and HDM exposure as the subclass or (B) HDM exposure as the class and antibiotic exposure as the subclass. In (A) bars represent taxa associated with antibiotic treatment only (left) or no antibiotic treatment only (right). In (B) bars represent taxa associated with HDM exposure only (left) or not associated with HDM exposure only (right). LDA scores represent the strength of the association between an identified taxa and the variable (HDM, antibiotics) in question. n=13 for No Antibiotics and n=16 for Antibiotic Mix. n=14 for HDM and n=15 for Pre-HDM.

Figure 6. Fecal Microbiome Diversity Directly Correlates with Regulatory T Cell Proportions

At sacrifice, Lung-Draining (Hilar, Mediastinal) Lymph Nodes (HLNs) were removed and processed to identify CD3+CD4+Foxp3+ regulatory T cells (Tregs) (A). Simpson’s Diversity Index was plotted against the proportion of all CD4+ T Cells in the HLN that are Foxp3+ Tregs (B). The Spearman r coefficient of correlation was calculated. The solid line is the linear regression line with the equation shown below the graph. The curved, dotted lines represent the 95% confidence band for the linear regression line. Error bars represent mean ± SEM. For (A), n=7 for HDM and n=8 HDM + Antibiotic Mix. For (B), n=5 for HDM and n=8 for HDM + Antibiotic Mix. Not all animals from (A) were able to be assessed for (B) due to poor DNA extraction and/or 16S gene amplification. * p < 0.05.