Transcriptional Profiling of the Immune Response to Marburg Virus Infection

Abstract

Marburg virus is a genetically simple RNA virus that causes a severe hemorrhagic fever in humans and nonhuman primates. The mechanism of pathogenesis of the infection is not well understood, but it is well accepted that pathogenesis is appreciably driven by a hyperactive immune response. To better understand the overall response to Marburg virus challenge, we undertook a transcriptomic analysis of immune cells circulating in the blood following aerosol exposure of rhesus macaques to a lethal dose of Marburg virus. Using two-color microarrays, we analyzed the transcriptomes of peripheral blood mononuclear cells that were collected throughout the course of infection from 1 to 9 days postexposure, representing the full course of the infection. The response followed a 3-stage induction (early infection, 1 to 3 days postexposure; midinfection, 5 days postexposure; late infection, 7 to 9 days postexposure) that was led by a robust innate immune response. The host response to aerosolized Marburg virus was evident at 1 day postexposure. Analysis of cytokine transcripts that were overexpressed during infection indicated that previously unanalyzed cytokines are likely induced in response to exposure to Marburg virus and further suggested that the early immune response is skewed toward a Th2 response that would hamper the development of an effective antiviral immune response early in disease. Late infection events included the upregulation of coagulation-associated factors. These findings demonstrate very early host responses to Marburg virus infection and provide a rich data set for identification of factors expressed throughout the course of infection that can be investigated as markers of infection and targets for therapy.

Importance: Marburg virus causes a severe infection that is associated with high mortality and hemorrhage. The disease is associated with an immune response that contributes to the lethality of the disease. In this study, we investigated how the immune cells circulating in the blood of infected primates respond following exposure to Marburg virus. Our results show that there are three discernible stages of response to infection that correlate with presymptomatic, early, and late symptomatic stages of infection, a response format similar to that seen following challenge with other hemorrhagic fever viruses. In contrast to the ability of the virus to block innate immune signaling in vitro, the earliest and most sustained response is an interferon-like response. Our analysis also identifies a number of cytokines that are transcriptionally upregulated during late stages of infection and suggest that there is a Th2-skewed response to infection. When correlated with companion data describing the animal model from which our samples were collected, our results suggest that the innate immune response may contribute to overall pathogenesis.

FIG 1

Comparison of changes in PBMC cytokine mRNAs to serum cytokine protein levels during MARV infection. Changes (fold) in IL-6 (A), MCP-1 (B), and MIP-1α (C) mRNA induction from PBMCs and IL-6 (A), MCP-1 (B), and MIP-1α (C) protein induction from serum are shown. Error bars represent standard errors over three separate microarrays for each time point.

FIG 2

Strongly upregulated genes in MARV-exposed NHPs. (A) Heat map illustration showing 295 hierarchically clustered genes that showed at least a 3-log₂-fold differential expression following challenge. Each row in the heat maps represents data from an individual gene, and each column represents the individual PBMC sample taken at a specific infection stage. Samples from the data set were grouped as in Table 1. Red and blue colors denote expression levels greater or less than baseline (white), respectively, and colored outlines (rust, green, and yellow) identify significant clusters of genes induced during early, middle, and late disease, respectively, and are labeled accordingly. (B) Expanded view of these gene clusters. The most significant functional groups (assigned by DAVID; P < 0.001) found in the respective clusters are listed to the right of the heat maps, along with the names of some representative genes.

FIG 3

Signaling network of genes that are strongly upregulated following MARV challenge of NHPs. By using Ingenuity Pathway Analysis (IPA) to map the genes which were found to be strongly upregulated following MARV challenge, 34 gene products (nodes with official gene symbols) were found to be directly or indirectly connected according to Ingenuity's database of published interactions. Edges without arrows indicate a direct interaction between two gene products, e.g., a protein-protein interaction. Edges with arrows represent a regulation of transcription, directed from the gene regulator to a regulated gene. Loops indicate self-regulation.

FIG 4

Innate immune response as indicated by mRNA expression following MARV challenge. Heat maps show the mRNA expression levels of a selection of day 1-responsive genes (A), genes upregulated beginning at day 3 postexposure (B), and complement-associated genes showing little change in gene expression (C). Red and blue colors denote expression levels greater or less than baseline (white), respectively.

FIG 5

Expression of inflammatory mediators in MARV-exposed NHPs. Heat maps illustrate the expression of mRNAs from inflammation-associated genes. (A) Classical mediators, such as IL-6, IL-1β, and IL-8, and receptors, such IL-1R1 and -2, are upregulated. (B) Other inflammatory mediators, such as IL-4, IL-5, and CXCL10, are upregulated. (C) Downregulated mediators, such as CXCR4 and IL-23, are shown.

FIG 6

Alteration of Th2-associated cytokines in animals exposed to MARV. (A) Average mRNA fold changes over time for IL-4, IL-5, IL-12, CCL5, and TNF-α. (B) Levels of IL-10, IL-5, and IL-4 protein in serum from animals infected with Marburg virus. Values were measured using a Bioplex 3D analyzer.