A pathway of interest to study influenza interactions with the host cell
In this present study, we report a microarray analysis revealing that infection by four human or avian influenza viruses (H3N2, H5N1, H5N2 and H7N1) significantly alters the gene expression of several host factors belonging to the p53 pathway. This pathway is composed of numerous genes, the protein products of which have a direct impact on host cell homeostasis, via the regulation of cell death and cell cycle progression as well as several other processes such as metabolism or the cellular immune response . The transcriptional activity of p53 is tightly regulated by Hdm2 in response to several intracellular signalling cascades triggered by different stimuli (DNA damage, UV irradiation and hypoxia). The central core of the pathway includes the p53 protein itself, which can directly or indirectly shut down the growth of stressed cells mainly by inducing cell-cycle arrest and apoptosis. Considering the central role of the p53 pathway, the impact of influenza infection on this cellular pathway is of particular interest.
Several studies have underlined the interplay between influenza viruses and different cellular signalling pathways, such as the PI3K/Akt, innate immune or apoptotic signalling pathways [4, 36–39]. Intriguingly however, only a few have been dedicated to changes in p53 upon influenza infection, highlighting the need to further characterize the interplay between influenza viruses and the p53 pathway [25, 26, 34, 40].
Influenza infection is associated with a global down-regulation of signalling cascades both upstream and downstream of p53 protein
We first observed that upstream signals of the p53 pathway that directly connect stress (UV, hypoxia, DNA damage) -induced signalling cascades to regulators of p53, were generally down-regulated at the transcriptional level by infection. Among such regulating kinases, DNA-PK, Chk2, JNK1 and Gsk3β genes are all under-expressed following infection.
Similarly, our results also revealed a significant decrease in gene expression for several endogenous p53 transcriptional targets, including p21, 14-3-3, PERP, FAS, DR4/5, PIG3, BAX, Bcl-XL, PAI-1 and PCNA during infection by our set of viruses. Moreover, these results were confirmed at the protein level for p21, Bax and Bcl-XL.
Altogether, our results indicate a global down-regulation in the mRNA expression of principal factors belonging to both upstream and downstream parts of the p53 pathway, suggesting a global negative effect on p53 pathway activity.
Discrepancy between p53 mRNA, protein levels and activity in influenza infected cells
The down-regulation of p53-target genes suggests that p53 expression and consequently its transcriptional activity is decreased during infection. We verified whether this was the case using luciferase assay and the monitoring of p53 phosphorylation status by western blot in H3N2-infected A549 cells (data not shown) and revealed some discrepancies. On the one hand, no significant change in p53 mRNA expression was detected in infected cells, except for a significant decrease with H5N1. On the other hand, levels of p53 protein expression were significantly increased by all the viruses, seemingly incoherent with the reduced activity of p53 in infected cells.
Altogether, these results suggest a likely complex regulation of TP53 expression during infection. Regulating factors include transcriptional activators (ISGF3 PRKCδ, HOXA5, CREBPB), inhibitors (BCL6, RBPJ) and stabilizing factors (ELAVL1 and ZMAT3) . Influenza infection might therefore induce the over-expression of both inhibitors (BLC6 and RBPJ) and activators (ISGF3) (data not shown). In accordance with this notion, a recent study reported the stress-induced regulation of p53 activity through the control of p53 mRNA stabilization and translation in addition to the well-described alteration of protein stability . Other stabilizing/destabilizing factors, such as host microRNAs may also be involved and should be investigated in the context of infection , in particular with H5N1, which induces a down-regulation of p53 mRNA expression unlike the other viruses.
Our results also clearly reveal that the impact that influenza viruses have on p53 is not exclusively on gene expression. Further investigations are necessary, in particular on the stabilization and activation of p53 protein, considering the putative interaction of p53 with viral proteins during the time-course of infection.
In this perspective, a recent study reported that the viral protein NS1 interacts with p53 and inhibits its activity . Furthermore, transient expression of H5N1 NS1 (wild-type or with mutations in regions 80-84 and 92) reduced the transcriptional activity of p53 . We can hypothesize that p53 protein is stabilized upon infection in an inactive form through interaction with the viral NS1 protein . Such a scenario would explain our seemingly contradicting results concerning p53 expression and activity. This may even reflect what happens with a number of DNA viruses for which viral proteins like BZLF1 (Epstein Barr Virus) or HBx (Hepatitis B virus) inhibit the transcriptional activity of a stabilized form of p53 [42, 43]. Interestingly, the disruption of p53 signalling to p21 was described in human lung A549 cells in the case of BZLF1 viral protein .
Viral down-regulation of the p53 pathway and modulation of the cell cycle and apoptosis by influenza A infection
We report the down-regulation of the p53-target genes p21 (G1/S-arrest) and 14-3-3 (G2/M-arrest by sequestering Cdc2-complex Cyclin B), both of which are involved in cell cycle control. As expected, the down-regulation of p21 and 14-3-3 in association with an over-expression of the cell-cycle regulators TP53INP1 and GADD45G [35, 45], could modulate cell cycle progression and resulted in G1-arrest. Our results are in line with one recent study describing the effect of influenza A infection on the cell cycle and show that viruses can induce an arrest in G0/G1 . They also suggest a putative involvement of p21, TP53INP1 and Gadd45gamma in the influenza-induced cell cycle arrest. A thorough functional evaluation of these results is now needed which should contribute to a better understanding of the viral mechanisms involved which permit an optimal viral protein expression and progeny production.
The other group of down-regulated p53-target genes that we identified during infection encodes the pro-apoptotic proteins PERP , Fas, DR4/5, PIG3 and BAX. In addition, the Bax-inhibiting anti-apoptotic factor Bcl-XL, was also under-expressed during infection, as was PAI-1 whose role in regulating apoptosis remains unclear . At the protein level, Bcl-XL expression was decreased by most viral infections. It is worth noting that the down-regulation of the three p53-target genes, p21, BAX and Bcl-XL, was considerably greater after H5N1 infection than with the other influenza A viruses tested.
Concordantly, the knock-down of BAX was recently shown to reduce the replication of influenza virus . Thus the down-regulation of a negative regulator of Bax, such as the anti-apoptotic Bcl-XL, at both mRNA and protein levels, such as we have reported here, would be expected to increase viral replication. Further silencing experiments are required to confirm such a result.
Altogether, our observations suggest that influenza viruses may at least partially control intrinsic and extrinsic pathways of apoptosis by decreasing the expression of both pro- and anti-apoptotic factors which are under the control of the p53 transcription factor, as suggested by several studies .
Is inhibition of the p53 pathway essential for the replication cycle of influenza viruses?
In agreement with the above hypothesis, we showed that the level of H3N2 viral production is significantly higher in HCT116 p53 -/- than in HCT116 p53 +/+ cells, at both 24 and 48 hpi (figure 6). These results are in accordance with initial observations reported by Turpin and colleagues in a p53 dominant-negative A549 model and more recent data reported on siRNA high-throughput screening for host functional interactants [25, 6].
These and our results suggest a marked antiviral facet for p53 and its pathway in the context of influenza infection, as has already been observed for several other viruses. The antiviral effect of p53 depends largely on its pro-apoptotic activity, which limits virus replication . However, other works have also demonstrated that p53 contributes to the innate cellular antiviral response by enhancing type I interferon (IFN)-dependent antiviral activity, independently of its function as a pro-apoptotic gene . In light of our results on p53 targets implicated in cell-cycle regulation and a recent study on human circulating leukocytes from infected patients , the p53 antiviral role might involve not only its activity in mediating the IFN response and apoptosis, but also in controlling cell cycle progression.