A Case for Antibiotic Perturbation of the Microbiota Leading to Allergy Development

Lisa A Reynolds; B Brett Finlay


Expert Rev Clin Immunol. 2013;9(11):1019-1030. 

In This Article

Absence of Microbial Signals & Allergy Development

Although GF mice have many developmental immune defects,[14] these mice have been extremely valuable to identify mechanisms by which the microbiota can potentially influence immune development and modulate disease susceptibility. GF mice develop more severe allergic disease in the ovalbumin/alum model of allergic airway inflammation than their conventionally housed counterparts.[27,28] The exacerbated airway inflammation is measured through increased airway resistance, total bronchoalveolar lavage fluid (BALF) cell numbers, BALF eosinophilia, serum immunoglobulin (Ig)E levels and lung tissue eosinophil infiltration.[27,28] Several mechanisms for this exacerbated inflammation have been proposed (Figure 1). One group has described how GF Swiss-Webster and GF C57BL/6 mice have higher levels of invariant natural killer T (iNKT) cells in their colonic LP and lung tissue than their conventionally housed counterparts.[28] iNKT cells express the chemokine receptor CXCR6, which recognizes the chemokine ligand CXCL16. In GF conditions, the increased recruitment of iNKT cells can be explained by hypermethylation of the Cxcl16 locus, leading to 5-hydroxymethylcytosine incorporation into DNA, which is associated with higher transcription of Cxcl16 mRNA in the epithelium of these sites.[28] DNA hypermethylation leading to the incorporation of 5-hydroxymethylcytosine, in contrast to the incorporation of 5-methylcytosine, may be associated with higher levels of gene transcription, rather than repression of transcription.[29] iNKT cells recognize antigen presented by CD1d, and the inflammatory phenotype appears CD1d-dependent, as elimination of the asthmatic response was observed after administration of a CD1d-blocking antibody (19G11). It is not clear whether CD1d was responsible for the exacerbation of inflammation in GF mice, or the development of any inflammation in this model, since 19G11 treatment also blocks disease development in conventionally housed mice.[28]

Figure 1.

Mechanisms leading to exacerbated inflammation in germ-free mice. (A) In conventionally raised mice, the presence of intestinal microbiota limits circulating IgE levels,31 by as yet undescribed mechanisms. At steady state, the locus for the iNKT chemokine ligand CXCL16 is not hypermethylated, leading to a basal level of transcription and translation of CXCL16 in the epithelium of the lung and colon, causing minimal iNKT cell recruitment to the lamina propria at these sites.28 (B) In the absence of microbial colonization, in GF mice, no inhibition of IgE production results in high levels of circulating IgE, and basophil surface-bound IgE.27,31 Additionally, hypermethylation of the Cxcl16 locus occurs, leading to incorporation of 5-hydroxymethylcytosine into DNA, which is associated with increased transcription and translation of CXCL16 in the colonic and lung epithelium.28 The chemokine receptor for CXCL16, CXCR6, is expressed on iNKT cells, causing increased recruitment of iNKT cells to the colonic and lung lamina propria.28 In a manner dependent on CD1d, iNKT cells mediated exacerbated allergic inflammation, likely due to their ability to produce large amounts of Type 2 cytokines following their activation.28

GF mice also have reduced numbers of alveolar macrophages in their lungs compared with conventionally housed mice,[27] a cell type implicated in inducing Tregs and promoting tolerance in the lungs,[30] although whether it is a lack of alveolar macrophages that is directly responsible for allowing exaggerated inflammation in the airways of GF mice is not yet known. Circulating cells and antibodies may contribute to the increased propensity for allergy of microbiota-deficient mice as in both GF and in mice deficient for the TLR-adaptor protein MyD88, serum IgE levels, circulating basophils numbers and basophil-surface bound IgE levels are elevated compared with conventionally housed and MyD88-sufficient controls.[31]

It will be critical to examine whether similar mechanisms lead to an increased propensity toward atopy in antibiotic-treated mice, which will clarify whether the phenotype of GF mice is a result of developmental immune defects, or due to an absence of microbial stimulation at the time of antigen sensitization.