|Home | About | Journals | Submit | Contact Us | Français|
The innate immune response provides the first line of defense against foreign pathogens by responding to molecules that are a signature of a pathogenic infection. Certain RNA viruses, such as influenza virus, produce double-stranded RNA as an intermediate during the replication life cycle, which activates pathogen recognition receptors capable of inducing interferon production. By engaging interferon receptors, interferon activates the JAK-STAT pathway and results in the positive feedback of interferon production, amplifying the response to viral infection. To examine how deficiencies in interferon signaling affect the cellular response to infection, we performed influenza virus infections of mouse embryonic fibroblasts lacking the alpha/beta interferon receptor, the gamma interferon receptor, or both. In the absence of the alpha/beta interferon receptor, we observed increased viral replication but decreased activation of PKR, Stat1, and NF-κB; the presence or absence of the gamma interferon receptor did not exhibit discernible differences in these readouts. Analysis of gene expression profiles showed that while cells lacking the alpha/beta interferon receptor exhibited decreased levels of transcription of antiviral genes, genes related to inflammatory and apoptotic responses were transcribed to levels similar to those of cells containing the receptor. These results indicate that while the alpha/beta interferon receptor is needed to curb viral replication, it is dispensable for the induction of certain inflammatory and apoptotic genes. We have identified potential pathways, via interferon regulatory factor 3 (IRF3) activation or Hoxa13, Polr2a, Nr4a1, or Ing1 induction, that contribute to this redundancy. This study illustrates another way in which the host has evolved to establish several overlapping mechanisms to respond to viral infections.
During infection by a foreign pathogen, one of the first signaling mechanisms to be initiated is the innate immune response. This first line of defense is initiated when cellular pathogen recognition receptors (PRRs), such as Toll-like receptor 3 (TLR3), recognize double-stranded RNA (dsRNA), one of many pathogen-associated molecular patterns (PAMPs) (17). Viruses as well as all pathogens contain PAMPs, which are conserved structures crucial for the pathogen's replicative life cycle (51). Besides TLR3, the PRRs RIG-I and MDA5 are also activated by dsRNA (21). During the life cycle of influenza virus, dsRNA is produced as an intermediate, and upon engagement with the above-mentioned PRRs, molecules such as interferon (IFN) regulatory factor 3 (IRF3), IRF7, and NF-κB are activated and translocated into the nucleus to induce beta IFN (IFN-β), which is one of many molecules vital to the innate immune response (18).
As one of the initial cytokines produced during the early stages of infection and the innate immune response, IFN-β induces and activates other proteins that can inhibit different steps of the viral life cycle (6). IFN-β activates these downstream processes by initially engaging the IFN-α/β receptors and activating the JAK-STAT pathway (7). This pathway induces a number of early-response, IFN-stimulated genes (ISGs) including IFN-γ, PKR, and tumor necrosis factor alpha (TNF-α) (9, 41). Furthermore, IFN-α/β also activates NF-κB, thus amplifying the IFN response via a positive-feedback loop since increased NF-κB activation results in increased IFN-β induction (28, 37). This feedback can be especially important for the recruitment of specialized immune cells to the site of injury or viral infection and can potentially cause an inflammatory response; that is, IFN-α/β is initially produced by leukocytes and fibroblasts, leading to the recruitment of T and NK cells, which produce IFN-γ (27). Therefore, the innate immune response can induce IFN as one of its downstream targets, which can in turn activate an inflammatory response, among its other functions (2, 42). But how tightly connected are these pathways, and do other mechanisms exist to activate them independently of one another? Previously, influenza virus infections of mice lacking IFN-α/β or IFN-γ receptors were performed. Those studies showed that the loss of either receptor altered the levels of viral replication in the lungs of the mice. However, mice lacking the IFN-α/β receptor exhibited increased levels of neutralizing antibodies and infiltration of granulocytic inflammatory cells into the lungs (10, 38).
While the recruitment of inflammatory cells assists in alleviating viral infection, it can convolute an analysis of the signaling mechanisms that are taking place in specific cell types, since the tissue is composed of a heterogeneous cell population. To better understand how IFN signaling affects influenza virus infection, we have made use of a homogeneous cell population consisting of mouse embryo fibroblasts (MEFs) devoid of either the IFN-α/β receptor (IFN-αβR−/−), the IFN-γ receptor (IFN-γR−/−), or both (IFN-αβγR−/−). We aimed to determine how the loss of each receptor would affect signaling responses downstream of the IFN receptors during influenza virus infection. Furthermore, we hypothesized that there would be redundant mechanisms within the innate immune response to induce inflammatory response genes even in the absence of certain IFN receptors.
To this end, we infected these cells with either a mouse-adapted strain of influenza virus, A/WSN/33 (WSN); the reconstructed 1918 (r1918) human pandemic influenza virus; or the highly pathogenic avian influenza virus A/Vietnam/1203/2004 (VN1203). We found that there were increased levels of virus replication in cells lacking the IFN-α/β receptor, which correlated with a decreased activation of antiviral genes and proteins. However, there was a similar induction of inflammation- and apoptosis-related genes in all cell types, as observed on a global level, as well as similar levels of IRF3 activation. Furthermore, certain genes were induced only in the absence of the IFN-α/β receptor, and these genes may be capable of activating the inflammation- and apoptosis-related genes induced in all cell types. Our findings suggest that while the IFN-α/β receptor is necessary to curb viral replication, it is dispensable for the induction of inflammatory response and apoptosis genes. Thus, redundancies exist in the innate immune response in order to achieve similar terminal responses to combat pathogenic infection. These results can be used to further study differences in rates of mortality for animals infected with influenza virus that are lacking IFN receptors.
Mouse embryonic fibroblasts (MEFs) derived from wild-type (WT) 129/SvEv mice or mice lacking the IFN-γ receptor (IFN-γR−/−), the IFN-α/β receptor (IFN-αβR−/−), or both the IFN-γ and IFN-α/β receptors (IFN-αβγR−/−) generated on the 129/SvEv background (19, 34, 54) were grown as monolayers in high-glucose Dulbecco's modified Eagle's medium (hgDMEM) supplemented to contain 10% heat-inactivated fetal calf serum (FCS) (HyClone Laboratories), 2 mM l-glutamine, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 10 μM 2-mercaptoethanol, penicillin G (50 units/ml), and streptomycin sulfate (50 μg/ml). Madin-Darby canine kidney (MDCK) cells were grown as monolayers in hgDMEM supplemented to contain 10% heat-inactivated fetal calf serum, 2 mM l-glutamine, penicillin G (50 units/ml), and streptomycin sulfate (50 μg/ml). The A/WSN/33 (WSN) strain of influenza virus was grown in Madin-Darby canine kidney cells as previously described (11). The reconstructed 1918 (r1918) and A/Vietnam/1203/2004 (VN1203) viruses were grown as previously described (36). Near-confluent monolayers of cells were mock infected with phosphate-buffered saline (PBS) alone or infected with influenza virus diluted in infection medium (hgDMEM supplemented to contain 2% heat-inactivated calf serum, 2 mM l-glutamine, penicillin G [50 units/ml], streptomycin sulfate [50 μg/ml], and 50 mM HEPES) to the indicated multiplicity of infection (MOI). After 45 min of adsorption at 4°C, virus and medium were removed. Fresh infection medium was added to the cells, and infections were allowed to proceed at 37°C until the indicated time postinfection (p.i.).
At the indicated times p.i., mock- or influenza virus-infected cells were labeled with 30 μCi of Express 35S protein labeling mix (Perkin-Elmer) in methionine- and cysteine-free hgDMEM for 30 min at 37°C. Cells were then washed twice with ice-cold Hanks' balanced salt solution and lysed in disruption buffer (0.5% Triton X-100, 50 mM KCl, 50 mM NaCl, 20 mM Tris-HCl [pH 7.5], 1 mM EDTA, 10% glycerol, 1× Complete protease inhibitor [Roche], 25 mM β-glycerophosphate, 1 mM Na3VO4). Levels of radioactivity for each sample were determined by trichloroacetic acid precipitation and scintillation counting. Lysates were boiled in an equal volume of 2× electrophoresis buffer (3.3% sodium dodecyl sulfate [SDS], 2.4 M β-mercaptoethanol, 16.7% glycerol, 13 mM Tris-HCl [pH 6.8], 8.3% water-saturated bromophenol blue) and separated by SDS-polyacrylamide gel electrophoresis (PAGE). As a loading control, identical counts per minute were loaded into each well.
At the indicated times p.i., 100 μl of influenza virus-infected cell supernatant was removed and then assayed in triplicate for viral yield by standard plaque assay on MDCK cells. Viral yields were calculated according to the following formula: yieldt=x = [log10(PFU/ml)t=x]/[log10(PFU/ml)t=0], where t is time and x is the time postinfection.
Following influenza virus infection, cells were lysed at the indicated times p.i., as described above. Total protein content was determined for clarified cell lysates by using the BCA protein assay kit (Pierce). Lysates were separated by SDS-PAGE with the same amount of total protein being loaded into each lane and then transferred onto polyvinylidene difluoride (PVDF) paper. Immunoblots were blocked for 1 h in PBS containing 0.5% Tween 20 and 5% nonfat dry milk, washed in PBS containing 0.05% Tween 20, and incubated at 4°C overnight with a mouse monoclonal actin antibody (MP Biochemicals) in PBS containing 0.5% Tween 20 and 1% nonfat dry milk. Blocking and primary antibody incubation for total PKR (Santa Cruz Biotechnology), pT451 PKR (Biosource International), total Stat1, or pS727 Stat1 (Upstate) were performed according to instructions provided by the manufacturer. Subsequently, membranes were washed and incubated for 2 h with horseradish peroxidase-conjugated donkey anti-mouse or anti-rabbit immunoglobulin G (Jackson Immunoresearch), and bound antibodies were detected with ECL Western blotting detection reagent (Amersham Biosciences/GE Healthcare).
Following influenza virus infection of cells cultured on glass coverslips, cells were fixed in 2% paraformaldehyde in PBS, permeabilized in 0.1% Triton X in PBS, washed with 2.5% fetal bovine serum (FBS) and 10 mM glycine in PBS, and then blocked with 10% FBS in PBS. Cells were then incubated for 2 h with primary antibodies recognizing NF-κB, IRF3 (Santa Cruz Biotechnology), or influenza virus nucleoprotein (NP) (a kind gift from Adolfo García-Sastre) diluted in 10% FBS in PBS. Subsequently, cells were washed and incubated for 1 h with fluorescein isothiocyanate (FITC)- or tetramethyl rhodamine isothiocyanate (TRITC)-conjugated donkey anti-mouse or anti-rabbit immunoglobulin G (Jackson Immunoresearch). Coverslips were mounted onto glass slides by using Vectashield Hardset medium with DAPI (4′,6′-diamidino-2-phenylindole) (Vector Laboratories) supplemented with p-phenylenediamine (1 mg/ml) to prevent photobleaching. Cells were imaged with a Nikon Eclipse E600 microscope and an RT Slider charge-coupled-device (CCD) camera by using SPOT software (Diagnostic Instruments).
At the indicated times p.i., cells were lysed in solution D (4 M guanidinium thiocyanate, 25 mM sodium citrate, 0.5% sarcosyl, 0.1 M β-mercaptoethanol), and total RNA was isolated by using RNeasy (Qiagen). The quantity of total RNA was determined by spectrophotometry using a NanoDrop ND-1000 fluorospectrometer. Contaminating DNA was removed by treating samples with RNase-free DNase and removal reagents (Ambion). Reverse transcription (RT) was performed by using TaqMan reverse transcription reagents (Applied Biosystems). RT-PCR was performed as previously described (31). Each target was run in quadruplicate, with 100 ng of sample in reaction mixture volumes of TaqMan 2× PCR universal master mix (Applied Biosystems). Genome copy numbers were normalized to microtubule-associated protein fraction (Mapf) values determined in parallel by using TaqMan gene expression assay endogenous control primer-probe sets (Applied Biosystems). The quantification of each gene, relative to the calibrator, was calculated by the instrument with the equation 2ΔCT(infected) − ΔCT(mock) within Applied Biosystems Sequence Detection software version 1.3. When calculating the relative levels of influenza virus M1 RNA in infected cells compared to mock-infected ones, we used the maximum cycle number, 40, in the ΔCT(mock) term of the equation. The minor-groove binding probe and primer sets for each gene were part of an Applied Biosystems assay set as follows: Mm01207402_m1 for mouse TLR3, Mm00440966_m1 for mouse PKR, Mm01257286_m1 for mouse Stat1, and Mm00439546_s1 for mouse IFN-β. The probe for the consensus influenza virus M1 transcript was 6-carboxyfluorescein (FAM)-CGATTCAAGTGATCCTC-minor groove binding nonfluorescent quencher (MGBNFQ).
Amplification of mRNA was performed as described previously by using equal masses of total RNA isolated from cells lysed in solution D (16). An equal-mass pool of mRNA isolated from time-matched, mock-infected cells was prepared as a reference sample. Microarray slide hybridization was performed by using mouse oligonucleotide genome CGH arrays (G4426B; Agilent Technologies). The data presented are the error-weighted average changes in expression calculated from four technical replicate arrays. Any analysis of the microarray data used a ≥2-fold cutoff (P < 0.01) within each independent set of viral infections, and sets from each analysis of a viral infection were then merged. All data were entered into a custom-designed relational database and subsequently uploaded into Rosetta Resolver System 7.1 (Rosetta Biosoftware), Spotfire Decision Site 9.1 (Spotfire/Tibco), or Ingenuity Pathways Analysis (Ingenuity Systems, Inc.). Primary microarray data are available at http://viromics.washington.edu.
To begin characterizing how the presence or absence of the IFN-α/β and IFN-γ receptors affects influenza virus infection in a controlled, homogeneous system, we infected wild-type, IFN-γR−/−, IFN-αβR−/−, or IFN-αβγR−/− MEFs (19, 34, 54) with the A/WSN/33 (WSN) strain of influenza virus. Previously, García-Sastre et al. showed that WSN infection of MEFs derived from mice lacking IFN-γ did not generate increased numbers of viral progeny but that those derived from mice lacking the IFN-α receptor did (12). In the present study, we performed a different characterization of these cells to determine the levels of viral replication. MEFs were infected with the WSN strain of influenza virus at an MOI of 2 PFU/cell, and levels of viral protein synthesis were assessed at 24 h p.i. by labeling infected cells with [35S]methionine and analyzing total protein synthesis by SDS-PAGE. By 24 h p.i., there was no noticeable viral protein synthesis in wild-type or IFN-γR−/− MEFs, but IFN-αβR−/− or IFN-αβγR−/− MEFs showed considerably higher levels of viral protein synthesis (Fig. (Fig.1A).1A). We further analyzed levels of infection by staining cells for the NP of influenza virus at 24 h p.i. At 24 h p.i., there were increased levels of NP staining in IFN-αβR−/− and IFN-αβγR−/− MEFs compared to wild-type and IFN-γR−/− MEFs (Fig. (Fig.1B).1B). Finally, we determined the levels of infectious virions present in the cell culture supernatant at 24 h p.i. by plaque assay with MDCK cells. IFN-αβR−/− and IFN-αβγR−/− MEFs produced 100-fold-more infectious virus than wild-type and IFN-γR−/− MEFs (Fig. (Fig.1C1C).
Since we observed increased levels of viral replication in cells lacking the IFN-α/β receptor, we next sought to determine the activation status of certain antiviral and IFN-inducible proteins. PKR is induced by IFN-α treatment and activated by dsRNA (32). Also, influenza virus infection induces IFN-β (16), which then induces and activates Stat1 downstream of the IFN-α/β receptor (33). To determine if the increased viral replication in cells lacking the IFN-α/β receptor is correlated with decreased levels of PKR or Stat1 activation, we determined the phosphorylation levels of these proteins via Western blotting. During influenza virus infection, there were decreased PKR and Stat1 phosphorylation levels in IFN-αβR−/− and IFN-αβγR−/− MEFs compared to wild-type and IFN-γR−/− MEFs (Fig. (Fig.2).2). Furthermore, the treatment of these cells with IFN-α resulted in increased PKR and Stat1 phosphorylation levels, albeit modest, only in the presence of the IFN-α/β receptor. These results indicate that decreased PKR or Stat1 activation may be contributing to increased viral replication in the absence of the IFN-α/β receptor.
Although PKR and Stat1 were activated only in the presence of the IFN-α/β receptor, we sought to determine if the receptor was necessary for the activation of proteins downstream of PKR and Stat1 signaling. Previously, it was shown that PKR activation results in the activation of NF-κB (14, 59). Additionally, there is evidence that alternative mechanisms exist for the activation of NF-κB via IFN signaling via phosphatidylinositol 3-kinase (PI-3K) or Tyk2 (56, 57). It was also shown previously that influenza virus infection activates interferon regulatory factor 3 (IRF3) (48). We therefore used nuclear localization assays to test for the activation of these proteins in MEFs infected with the WSN virus. While mock infection did not cause a nuclear localization of NF-κB or IRF3 in any cell type (Fig. 3A and B), we observed decreased NF-κB nuclear localization in IFN-αβR−/− and IFN-αβγR−/− MEFs compared to wild-type and IFN-γR−/− MEFs (Fig. (Fig.3C).3C). However, we observed a nuclear localization of IRF3 in all cell types during WSN infection (Fig. (Fig.3D).3D). In some cases, we observed NF-κB or IRF3 nuclear localization in cells that did not exhibit NP staining. This may be because the levels of NP staining were below the limits of detection or because infected cells secreted cytokines that activated NF-κB or IRF3 in neighboring cells that had not yet been infected. Collectively, these results indicate that the loss of NF-κB activation during influenza virus infection is attributable to the loss of IFN-α/β signaling but that IRF3 activation is not altered by the presence or absence of the IFN-α/β or IFN-γ receptor.
Since all of our previous experiments used WSN, a mouse-adapted strain of influenza virus, we also evaluated how human and avian influenza virus infections progressed in these cell types. Previous studies have shown that the reconstructed 1918 (r1918) human pandemic influenza virus and the A/Vietnam/1203/2004 (VN1203) avian influenza virus are highly pathogenic in mice, with the latter causing greater mortality (36). Cells were infected with WSN, r1918, or VN1203 at an MOI of 2 PFU/cell, and RNA was collected at 24 h p.i. for quantitative RT-PCR (qRT-PCR) analysis. The results showed that the level of M1 expression was highest during VN1203 infection and lowest during WSN infection (Fig. (Fig.4).4). Furthermore, during WSN infection, there was increased M1 expression levels in IFN-γR−/−, IFN-αβR−/−, and IFN-αβγR−/− MEFs compared to wild-type MEFs. During r1918 infection, the levels of M1 expression were the same among all cell types. However, VN1203 infection resulted in increased M1 expression levels in IFN-αβR−/− and IFN-αβγR−/− MEFs compared to wild-type MEFs. Furthermore, levels of viral replication were at least 10-fold higher in IFN-αβR−/− and IFN-αβγR−/− MEFs than in wild-type MEFs during VN1203 infection but not r1918 infection (Table (Table11).
In addition to comparing levels of viral replication among different viruses, we also determined how antiviral genes, namely, TLR3, PKR, and Stat1, were induced during infection with the r1918 and VN1203 viruses. We determined levels of TLR3 induction since it was previously shown that TLR3 is induced in the presence of dsRNA and IFN-α treatment (20). Using qRT-PCR, we observed that TLR3, PKR, and Stat1 were all induced to a lesser extent in IFN-αβR−/− or IFN-αβγR−/− MEFs than in wild-type or IFN-γR−/− MEFs (Fig. 5A to C). This was also dependent on the reported pathogenicity of the virus in mice; that is, VN1203 induced these genes to the greatest extent, r1918 induced them to an intermediate extent, and WSN induced them to the least extent, which is correlated to the levels of viral replication for each type of viral infection. However, the induction of IFN-β did not follow the same pattern, as its level of induction was decreased in IFN-αβR−/− or IFN-αβγR−/− MEFs compared to wild-type MEFs only during WSN infection, although IFN-αβγR−/− MEFs also exhibited decreased levels of IFN-β induction during VN1203 infection (Fig. (Fig.5D).5D). Furthermore, we observed no IFN-γ or IFN-λ induction in any cell type (data not shown). This indicates that IFN-β gene expression may be induced independently of the presence of its receptor, perhaps via IRF3 or other mechanisms. It may also be that WSN, but not r1918, depends on the positive amplification loop through the IFN-α/β receptor to produce as much IFN-β as wild-type cells. Furthermore, IFN-γ induction is not being induced in fibroblasts to cause downstream signaling through the IFN-γ receptor; rather, IFN-γ is produced by infiltrating immune cells at the site of infection in a whole-animal model (27).
Our virology and biochemical assays indicated that influenza virus infection of cells lacking the IFN-α/β receptor resulted in greater viral protein synthesis, virion production, and viral gene expression, which were inversely correlated to the induction and activation of antiviral proteins. In order to uncover additional differences in the host response that may impact viral replication, we used oligonucleotide microarrays to profile the cellular transcriptional response to infection. For our microarray analyses, wild-type, IFN-γR−/−, IFN-αβR−/−, or IFN-αβγR−/− MEFs were mock infected or infected with the WSN, r1918, or VN1203 strain of influenza virus at an MOI of 2 PFU/cell. Analyses were performed by comparing RNA isolated from each individual cell type against a pool of RNA from genotype-matched mock-infected MEFs.
An initial evaluation of the data showed the greatest differential gene expression at later time points and in response to infection with the VN1203 virus. Thus, we began by examining the data obtained from infections with the VN1203 strain at 24 h p.i. Since our initial observations revealed that viral replication was greatly affected by the presence or absence of the IFN-α/β receptor but not the IFN-γ receptor, we analyzed the transcriptional profile data by creating a gene set consisting of genes that were at least 2-fold (P < 0.01) upregulated in wild-type and IFN-γR−/− MEFs but either downregulated or unchanged in IFN-αβR−/− and IFN-αβγR−/− MEFs. A functional analysis of this gene set allowed us to determine which genes were activated only in the presence of the IFN-α/β receptor. We then created another gene set that contained genes that were at least 2-fold (P < 0.01) upregulated among all four cell types. Analysis of this gene set allowed us to determine which genes were activated independently of the IFN-α/β receptor. This type of analysis was then repeated for infections with the r1918 and WSN strains of influenza virus, and the gene sets from the three separate analyses for each virus were merged.
We then performed a functional analysis of these gene sets by using Ingenuity Pathways Analysis. The top-ranked functional categories, determined by the P value of the enrichment analysis, of the set containing genes upregulated only in wild-type and IFN-γR−/− MEFs were related to the interferon response, containing genes such as Irf5, Irf7, Mx1, Mx2, and Oas3. The top functional categories of the set containing genes upregulated among all four cell types were related to inflammatory and apoptotic pathways, such as Ccl5, Il6, Irf1, Il1b, and Tnf. Genes from these categories were selected and are presented in Fig. Fig.66 and in Fig. S1 and S2 in the supplemental material. The gene expression data show that certain IFN response genes do not need to be induced for the induction of genes related to inflammatory and apoptotic responses; other potential mechanisms for the induction of these genes in the absence of the IFN-α/β receptor are described below. However, the increased level of induction of these IFN response genes in wild-type and IFN-γR−/− MEFs is correlated with decreased viral replication; without the induction of these genes, there is increased replication, as observed for IFN-αβR−/− and IFN-αβγR−/− MEFs. Second, Fig. Fig.66 demonstrates that the virus that is most pathogenic in animals, VN1203 (C. Cillóniz and M. G. Katze, unpublished data), elicits the greatest induction of these inflammatory response genes, perhaps due to levels of viral replication. Different methods of analyses, namely, analysis of variance (ANOVA) and gene set enrichment analysis (GSEA) (47), were also employed to examine these transcriptional profiles, and the results of each analysis were similar to the results presented here.
To evaluate the functional relationships of these genes more closely, we used Ingenuity Pathways Analysis to create a network of the genes shown in Fig. Fig.6.6. Only those genes that exhibited direct interactions among gene sets were included in the network (Fig. (Fig.7).7). Dotted lines represent interactions between the gene sets shaded in light blue and orange, indicating potential mechanisms by which the presence of the IFN-α/β receptor induces genes related to inflammatory and apoptotic responses. For example, Stat1 was previously shown to induce Irf1 (22), and Rela (NF-κB) induces Ifnb1 expression (49). In general, the signaling pathways that occur to initiate an inflammatory or apoptotic response in the presence of the IFN-α/β receptor are well known. To determine potentially novel mechanisms for the induction the same inflammatory and apoptotic response genes in the absence of the IFN-α/β receptor, we included genes that were at least 2-fold (P < 0.01) upregulated only in IFN-αβR−/− and IFN-αβγR−/− MEFs but not in WT and IFN-γR−/− MEFs during any of the three types of viral infection (Fig. (Fig.7,7, yellow, and see Fig. S3 in the supplemental material). Solid lines represent interactions between gene sets in yellow and orange or within the set in orange. Of particular interest are Ing1 and Nr4a1, which induce apoptosis via Mdm2, and Polr2a, which induces apoptosis via Myc (15, 24, 29, 30, 58). The genes in yellow are on the periphery of the network diagram because the direct mechanisms for how they may initiate inflammatory or apoptotic responses and interact with many of the genes in orange are not yet known. Nevertheless, we highlight a potential mechanism for the induction of inflammatory and apoptotic response genes in the absence of the IFN-α/β receptor (Fig. (Fig.7,7, blue edges). Via the Hoxa13-mediated activation of Bmp2, signaling cascades that ultimately induce IFN-β or interleukin-1β (IL-1β), a key component of the inflammasome, can be initiated (8, 23, 39, 46). It is important that the genes shown in yellow in Fig. Fig.77 may initiate pathways among the genes shown in orange without signaling through the genes shown in blue, establishing potentially novel mechanisms for the activation of genes related to inflammatory and apoptotic responses in the absence of signaling through the IFN-α/β receptor. A hyperactivation of these pathways may be responsible for the increased mortality for animals lacking the IFN-α/β receptor.
There have been a number of studies that made use of influenza virus-infected mice devoid of IFNs or their receptors (10, 13, 25, 38). In general, those studies have shown that the lack of IFN results in increased mortality rates and levels of viral replication, especially in the presence of the Mx1 gene (25). In the work presented here, we observed increased levels of viral replication in the absence of the IFN-α/β receptor, and this correlated with decreased levels of TLR3, PKR, Stat1, and NF-κB induction or activation. However, we observed that IRF3 was activated even in the absence of the IFN-α/β receptor and that the absence of the receptor did not preclude cells from inducing genes related to inflammatory and apoptotic pathways. Finally, we used highly pathogenic viruses, r1918 and VN1203, along with a mouse-adapted laboratory strain, WSN, to show that while each virus exhibited similar patterns of antiviral, inflammatory, and apoptotic response gene expression among the four cell types, more-pathogenic viruses caused a greater induction of these genes. For these experiments, we used MEFs, a homogeneous cell population, since they allowed us to study the signaling pathways without immune cell infiltration, which can confound results observed for an animal system. However, it should be stated that one may be able to better understand immunity during influenza virus infection by infecting macrophages, dendritic cells, or lung epithelial cells isolated from mice lacking interferon receptors. Nevertheless, fibroblasts were shown to play a role in lung pathogenesis during influenza virus infection; lung fibroblasts can produce IFN during infection, and the interaction of them with T cells prevents the activation of CD4+ cells (27, 53).
In the presence of the IFN-α/β receptor, we observed that the induction of genes related to inflammatory and apoptotic responses was achieved in part via NF-κB (Rela), Stat1, or PKR (Eif2ak2) signaling (17, 59); these “classical” pathways are represented in Fig. Fig.77 by dotted lines. Furthermore, it was previously shown that the activation of these proteins is dependent on the presence of the IFN-α/β receptor (12, 32, 57). However, in the absence of the IFN-α/β receptor, the inflammatory and apoptotic responses may be initiated through alternative mechanisms, such as Ing1, Nr4a1, Polr2a, or Hoxa13, as shown in Fig. Fig.77 (solid lines). Furthermore, other PAMPs that are part of the innate immune response, such as IRF3, which we observed to be activated in both the presence and the absence of the IFN-α/β receptor, may be responsible for the induction of inflammatory genes even when IFN-α/β receptor signaling is absent (5).
Regarding the highly pathogenic viruses used in this study, r1918 and VN1203, we observed increased levels of induction of genes capable of activating inflammatory and apoptotic responses compared to the WSN strain of influenza virus. This may be due in part to increased levels of viral replication during infection with the more-pathogenic viruses. We further characterized these observations by determining the levels of transcripts that encode antiviral proteins, and we observed the highest levels of Stat1, TLR3, and PKR during VN1203 infection.Infection with r1918 produced an intermediate phenotype with regard to these transcripts compared to WSN infection. It was previously shown that VN1203 causes more rapid mortality in mice than does r1918 infection (52). Current studies in our laboratory not only have confirmed this but also have shown that wild-type mice exhibited decreased rates of mortality and viral replication in the brain and spleen compared with IFN-αβR−/− mice; levels of viral replication in the lungs were similar between animal genotypes (Cillóniz and Katze, unpublished). Furthermore, there was increased viral replication in VN1203-infected animals compared to r1918-infected ones. The results from these animal experiments can be explained in part by the experiments with a homogeneous fibroblast population devoid of signaling from immune cells that infiltrate the lung during infection; that is, cells and mice lacking the IFN-α/β receptor exhibited increased viral replication, and in cells, this was anticorrelated with a decreased activation of the antiviral proteins PKR, Stat1, and NF-κB. We are currently evaluating the activation status of these proteins using mice lacking the IFN-α/β receptor. Additionally, there were no discernible differences in lung or spleen pathogenesis between wild-type and IFN-αβR−/− mice at late times p.i., characterized by moderate to severe bronchiolitis at 4 days p.i. However, pathogenesis was greater for VN1203-infected animals than for r1918-infected ones. Similarly, in MEFs, the presence or absence of the IFN-α/β receptor did not impact the induction of genes related to inflammatory and apoptotic responses, but VN1203-infected MEFs exhibited a greater induction of these genes than did r1918-infected MEFs. Therefore, we have shown how experiments with a homogeneous cell culture population can help interpretations of whole-animal studies; that is, even though the level of viral replication was lower in wild-type animals than in IFN-αβR−/− mice, presumably due to the IFN response, the pathogenesis remained the same for both, presumably due to the inflammatory response. Further analysis of the gene expression profiles from these infected animals will lead to more mechanistic detail regarding viral replication and pathogenesis pathways.
In showing that potential pathways exist to reach similar expressions of genes related to the inflammatory and apoptotic responses in both the presence and absence of the IFN-α/β receptor, we have identified another redundancy in intracellular signaling that exists to combat viral infections. Du and colleagues have shown that NF-κB, a transcription factor vital to the cellular response of external stimuli, can be activated by both IFN-dependent and -independent pathways (9). Furthermore, NF-κB can initiate signaling through a number of different molecules such as TRAF2, PI-3K, or Tyk2 (55-57). Previously, a novel type of IFN was discovered, IFN-λ, which functions through its own receptor (26, 43). While the receptor for IFN-λ is different than that of IFN-α/β and IFN-γ, IFN-λ still functions through a Jak/Stat signaling pathway, and many of the downstream biological activities are similar between IFN-α/β and IFN-λ (4). Additionally, IFN-λ induction can be stimulated by TLR3 signaling and viral infection and has antiviral activity, similar to IFN-α/β and IFN-γ (3, 35, 40, 44). While we did not observe any production of IFN-λ in our experiments, since it is produced in a tissue-specific fashion, it performs functions similar to those of IFN-α/β albeit on different cell types (45). The same is true for IFN-γ: it was not produced in the cells used in our experiments and thus does not provide a level of redundancy in fibroblasts. However, in a whole-animal system, IFN-α/β signaling recruits T and NK cells, which produce IFN-γ to elicit antiviral effects (27, 38). Therefore, to use MEFs to study the role of IFN-γ or IFN-λ in the absence of IFN receptors, specialized immune cells would have to be isolated from the mutant mice for in vitro experimentation.
Our results indicate that while the IFN-α/β receptor is needed to curb viral replication, it is dispensable for the induction of certain inflammatory and apoptotic genes. We identify potential pathways, via IRF3 or IL-1β activation or Hoxa13, Polr2a, Nr4a1, or Ing1 induction, that may contribute to this redundancy. Further experimentation is needed to interrogate these potential mechanisms and how the proteins encoded by each gene may elicit inflammatory or apoptotic responses in the absence of the IFN-α/β receptor. Of particular interest is the mechanism of IL-1β activation in the absence of the IFN-α/β receptor, since recent studies have shown that this molecule is central to inflammasome signaling (1, 50). Together, our study and those described above illustrate ways in which the host has established overlapping mechanisms to respond to viral infections and that redundancies occur within host signaling mechanisms, which likely developed from the coevolution of pathogen and host.
We thank Herley Beyene for growing the A/WSN/33 strain of influenza virus and Jenny Tisoncik for critical reading of the manuscript. We thank Joy Loh, Megan Goodwin, David Strong, and Skip Virgin for providing us with the cohort of IFN receptor knockout MEFs.
This work was supported by NIH grant R01AI022646 (M.G.K.).
Published ahead of print on 25 November 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.