Sequential Infection with Common Pathogens Promotes Human-like Immune Gene Expression and Altered Vaccine Response

Abstract

Immune responses differ between laboratory mice and humans. Chronic infection with viruses and parasites are common in humans, but are absent in laboratory mice, and thus represent potential contributors to inter-species differences in immunity. To test this, we sequentially infected laboratory mice with herpesviruses, influenza, and an intestinal helminth and compared their blood immune signatures to mock-infected mice before and after vaccination against yellow fever virus (YFV-17D). Sequential infection altered pre- and post-vaccination gene expression, cytokines, and antibodies in blood. Sequential pathogen exposure induced gene signatures that recapitulated those seen in blood from pet store-raised versus laboratory mice, and adult versus cord blood in humans. Therefore, basal and vaccine-induced murine immune responses are altered by infection with agents common outside of barrier facilities. This raises the possibility that we can improve mouse models of vaccination and immunity by selective microbial exposure of laboratory animals to mimic that of humans.

Figure 1. Co-infection of mice with chronic viruses and a helminth leads to reduced antibody production after vaccination with YFV-17D

(A) Experimental outline. See experimental procedures and Figure S1 for more detail on number of mice and groups. (B) Measurement of total anti-YFV-17D IgG from serum. (C) Total serum neutralizing anti-YFV-17D antibody was quantified 34-40 days after vaccination. (D) Genome copies (normalized to actin) of YFV-17D were measured in draining lymph node (LN) of mice on days 3 and 7 following vaccination. Numbers above bar are the number of mice positive for YFV genome over the number of mice analyzed. One-way ANOVA was used to calculate statistical differences *** p< 0.01. Dots represent individual animals, and bars represent averages +/− SEM.

Figure 2. Microarray analysis of blood samples of vaccinated animals reveals kinetically distinct gene expression in co-infected mice relative to mock

(A) Heatmap of significant kinetically different genes between mock and co-infected. See Table S1. (B) Kinetics of each cluster, represented in individual points for each of three experiments as the median expression value of all genes in a given cluster across a sample. Line represents the mean of three replicates median expression values. (C) Heatmap of top significantly enriched GO terms in each cluster. Lack of ontological agreement between mock and co-infected high clusters suggests highly differential response to vaccine. (D) Analysis of cytokines in serum of vaccinated animals over time in pictograms (pg)/ml. One-way ANOVA was used to calculate statistical differences * p< 0.5, ** p<0.1, *** p<0.01. Symbols are individual animals, and bars are averages of all animals +/− SEM.

Figure 3. Comparison of metagenes from co-infected and mock mice with pet store-raised and laboratory-raised mice indicates overlap of co-infected mice with pet store mice and mock mice with laboratory-raised mice

(A) Enrichment analysis of a co-infected d0 vs. mock d0 metagene (type-I interferon response) and mock vs. co-infected d0 metagene (naive lymphocyte) in pet store vs. laboratory mice (B) Enrichment analysis of a co-infected d0 vs. mock d0 metagene (type-I interferon response) and mock vs. co-infected d0 metagene (naive lymphocyte) in cohoused vs. laboratory array data. See Figure S4 and Table S2.

Figure 4. Co-infected mice have parallel gene expression with human maternal blood, whereas mock mice have parallel gene expression with human cord blood

Enrichment analysis of co-infected d0 vs. mock d0 metagene (type-I interferon response) and mock vs. co-infected d0 metagene (naive lymphocyte) in human maternal and cord blood array data.