In this study, the pathogenic properties of NS1 proteins encoded by different subtypes of influenza A viruses were compared by measuring cytokine/chemokine expression and apoptosis induced in transfected cells.
While the key target of influenza virus is lung epithelial cells [48, 49], most studies on influenza pathogenesis have been based on macrophages and monocytes infected in-vitro or in-vivo [50–52]. It is important to note that at the time of early infection these immune cells would not be present in great numbers until they have been recruited into the area. The mechanism concerning bronchial infiltration of inflammatory cells, particularly lymphocytes and eosinophils, and the subsequent hyperresponsiveness of the bronchial wall induced by viruses remain unclear . Therefore, in this study we have used a cell line derived from human lung epithelial cells as an in-vitro model to study the pathogenicity of influenza NS1 proteins.
Previous in-vitro studies have shown that influenza infection induces the production of cytokines IFN-α, tumor necrosis factor (TNF)-α, IL-1, IL-6, IL-8 and the mononuclear cell attractant chemokines CCL-3/MIP-1α, CCL-4/MIP-1β, CCL-2/MCP-1, CCL-7/MCP-3, CXCL-10/IP-10 and CCL-5/RANTES in human monocytes, epithelial cells, rat alveolar or murine macrophages [48, 50, 53–62]. Based on the findings of these studies, we identified the six key cytokines/chemokines for the current study.
Recently, it has been shown that the inflammatory response is played out over time in a reproducible and organized way with different induction kinetics after an initiating stimulus . Cytokines released following infection can be classified broadly into "early" and "late" cytokines. Our results showed that CCL-2/MCP-1, CCL-3/MIP-1α and CCL-5/RANTES were produced early post-transfection; while IL-6, CXCL-10/IP-10 and CXCL-9/MIG were produced later. This time course of cytokine/chemokine production was consistently observed across different subtypes of influenza viruses with different pathogenicity. It would be worthwhile to further investigate whether this temporal sequence is unique to influenza or generally true for other acute respiratory viruses.
The most remarkable observation in this study was the distinct cytokine/chemokine profiles induced by the NS1 protein of H5N1. Our in-vitro observation is in line with previous reports that the peripheral blood of patients infected with H5N1 have much higher serum levels of CXCL-10/IP-10 and CCL-2/MCP-1 than patients infected with seasonal influenza [13, 15]. Furthermore, the in-vitro model used in our study by measuring the levels of cytokines in lung tissue may be more relevant to pathogenesis than levels in blood . Another in-vitro study in macrophages also showed a stronger cytokine induction by H5N1/1997 viruses compared to H3N2 .
We found that NS1 protein encoded by 2009 pdmH1N1 virus induced similar levels of cytokine/chemokine compared to seasonal H1N1 and H3N2 strains. This observation is in line with a recent report which showed that pro-inflammatory cytokine expression in the 2009 pandemic H1N1 virus-infected macrophages was similar to that of seasonal H1N1/1999, and was much lower than in H5N1/2004-infected cells [64–67].
In this study, we observed a similar temporal profile of apoptosis as induced by different subtypes of influenza. This is in contrast to previous studies based on whole virions. For instance, Geiler et al. (2011)  reported a delay in the induction of apoptosis for 2009 pdmH1N1 compared to H5N1; whereas Mok et al. (2007)  reported a delayed apoptosis of H5N1 compared to seasonal H1N1. The reason for these different observations remains to be verified. Both of these two studies [67, 68] used whole virions, and therefore the observation may be partly related to the time required for sufficient virus replication and hence the synthesis of a certain amount of NS1 protein; whereas the current study used transfection where the same amount of NS1 was expected to be synthesized at any one time for different subtypes. If NS1 protein per se was the question of interest, using transfection methods with the same transfection efficiencies among different subtypes, might avoid biases result from differences in replication efficiency of the viruses being studied. Another major difference is that in contrast to the lung epithelial cells used in our study, the two previous studies used macrophages. These two cell types may display differences in apoptotic response to different subtypes of influenza.
Another remarkable observation of the current study is the high apoptosis inducing ability conferred by NS1 protein encoded by H5N1 compared to all other subtypes. This is reminiscent of the rapid development of severe primary pneumonitis in patients infected with H5N1 [4, 13–16]. Our data showed that the apoptosis inducing ability of NS1 protein encoded by 2009 pdmH1N1 virus was similar to H7, H9 and seasonal subtypes; but much lower than H5N1. This is in line with a previous observation based on macrophages infected with whole virions, where the level of apoptosis induced by 2009 pdmH1N1 was much lower than H5N1 .
NS1 have both pro- and anti-apoptotic functions, and the level of apoptosis observed reflects a balance between the two [30, 34, 69–73]. It would be worthwhile to further investigate whether the NS1 encoded by H5N1 is more pro-apoptotic or less anti-apoptotic as compared to other subtypes.