Systems-level comparison of host-responses elicited by avian H5N1 and seasonal H1N1 influenza viruses in primary human macrophages

Abstract

Human disease caused by highly pathogenic avian influenza (HPAI) H5N1 can lead to a rapidly progressive viral pneumonia leading to acute respiratory distress syndrome. There is increasing evidence from clinical, animal models and in vitro data, which suggests a role for virus-induced cytokine dysregulation in contributing to the pathogenesis of human H5N1 disease. The key target cells for the virus in the lung are the alveolar epithelium and alveolar macrophages, and we have shown that, compared to seasonal human influenza viruses, equivalent infecting doses of H5N1 viruses markedly up-regulate pro-inflammatory cytokines in both primary cell types in vitro. Whether this H5N1-induced dysregulation of host responses is driven by qualitative (i.e activation of unique host pathways in response to H5N1) or quantitative differences between seasonal influenza viruses is unclear. Here we used microarrays to analyze and compare the gene expression profiles in primary human macrophages at 1, 3, and 6 h after infection with H5N1 virus or low-pathogenic seasonal influenza A (H1N1) virus. We found that host responses to both viruses are qualitatively similar with the activation of nearly identical biological processes and pathways. However, in comparison to seasonal H1N1 virus, H5N1 infection elicits a quantitatively stronger host inflammatory response including type I interferon (IFN) and tumor necrosis factor (TNF)-alpha genes. A network-based analysis suggests that the synergy between IFN-beta and TNF-alpha results in an enhanced and sustained IFN and pro-inflammatory cytokine response at the early stage of viral infection that may contribute to the viral pathogenesis and this is of relevance to the design of novel therapeutic strategies for H5N1 induced respiratory disease.

Figure 1. Heatmap showing microarray gene expression profiles of influenza A infected primary human macrophages at different post-infection time points.

Expression of genes with p-values <0.05 and fold change ≥±1.5 in at least 1 of the 6 conditions in primary human macrophages infected with H5N1 or H1N1 viruses at 1, 3 and 6 h post-infection are shown. Note that the majority of gene expression changes were found at 6 h post-infection in both H5N1 and H1N1 infected cells. Data presented are averaged gene expression changes for three different individuals.

Figure 2. Validation of microarray data by real time PCR.

Expression of six genes was assessed at 1, 3 and 6 h after infection by influenza A compared to mock infection. Data presented was from one representative donor and showed similar expression patterns compared with the microarray experiment.

Figure 3. Expression of selected genes annotated are related to cytokine and chemokine activity.

Increased gene expression levels were seen in response to H5N1 compared to H1N1 infection.

Figure 4. Components of the RIG-I, TNF and type I IFN pathways are up-regulated in response to H5N1 to greater extent than H1N1.

Viral ligands are shown as diamonds, genes/proteins as circles, and complexes or pathways as rectangles. Green arrows indicate activation, red T-bars represent inhibition, black lines indicate binding, and blue lines indicate stimulation of gene expression. For clarity, blue lines feeding back to upstream pathway components were removed; many of these upstream components, however, are regulated by type I IFN and/or NFκB. The colour of a node reflected the H5N1:H1N1 expression ratio: blue nodes were not significantly differentially expressed in response to either virus, pink/red nodes were up-regulated more in response to H5N1 than H1N1 (red  = >1.5-fold more in response to H5N1, pink  = 1.0-1.49-fold higher in response to H5N1), and the green node was down-regulated to a greater extent in response to H5N1. Larger nodes had a fold-change value of >10 in H5N1 infection.