Type 1 IFNs (IFN-αβ) play pivotal roles in the host antiviral response and in TLR-induced signaling. IFN regulatory factor (IRF) and NF-κB transcription factors are thought to be crucial for virus-induced mRNA expression of IFN-β. Although recent studies have demonstrated essential roles for IRF3 and IRF7, the definitive role of NF-κB factors in IFN-β (or IFN-α) expression remains unknown. Using mice deficient in distinct members of the NF-κB family, we investigated NF-κB function in regulating type 1 IFN expression in response to Sendai virus and Newcastle disease virus infection. Surprisingly, IFN-β and IFN-α expression was strongly induced following virus infection of mouse embryonic fibroblasts (MEFs) from p50−/−, RelA/p65−/−, cRel−/−, p50−/−cRel−/−, and p50−/−RelA−/− mice. Compared with wild-type MEFs, only RelA−/− and p50−/−RelA−/− MEFs showed a modest reduction in IFN-β expression. To overcome functional redundancy between different NF-κB subunits, we expressed a dominant-negative IκBα protein in p50−/−RelA−/− MEFs to inhibit activation of remaining NF-κB subunits. Although viral infection of these cells failed to induce detectable NF-κB activity, both Sendai virus and Newcastle disease virus infection led to robust IFN-β expression. Virus infection of dendritic cells or TLR9-ligand CpG-D19 treatment of plasmacytoid dendritic cells from RelA−/− or p50−/−cRel−/− mice also induced robust type 1 IFN expression. Our findings therefore indicate that NF-κB subunits p50, RelA, and cRel play a relatively minor role in virus-induced type 1 IFN expression.
Type 1 IFNs (IFN-αβ) play a central role in the host response to viral infection (1). The two main subtypes of type 1 IFNs are IFN-α and IFN-β. The IFN-α gene family is comprised of more than a dozen members, whereas IFN-β is encoded by a single gene. Recent studies have identified RIG-I and Mda5 in viral detection and initiating signaling via the IFN-β promoter stimulater/mitochondrial antivirus signaling protein adapter to TANK-binding kinase-1/1κB kinase (IKK)3-ε kinases, resulting in activation of key transcription factors (2, 3, 4, 5). The mechanisms governing virus- and TLR-induced signaling leading to expression of type 1 IFN transcription have been a topic of intense interest for several years (6). In particular, numerous studies have investigated regulation of the IFN-β gene (6). Four cis elements are thought to be crucial for virus-induced IFN-β expression (7). These elements, designated positive regulatory domain (PRD)-I, II, III, and IV, bind key transcription factor that mediates virus-induced expression. The PRD-II element binds to NF-κB transcription factors, and is thought to play an especially important role in virus-induced IFN-β enhanceosome assembly (8, 9). Elements PRD-I and PRD-III associate with members of the IFN regulatory factor (IRF) transcription factor family (6). Using gene-targeted mice, recent studies have defined an essential role for IRF3 and IRF7 in regulation of type 1 IFN expression (10, 11). Absence of IRF7 or combined absence of IRF3 and IRF7 prevents induction of IFN-α and IFN-β following viral infection or TLR stimulation of dendritic cells (DCs) (10, 11). In contrast, no definitive study has addressed the function of members of the NF-κB family in virus-induced type 1 IFN expression.
Regulation of type 1 IFN expression is subject to a critically important positive feedback loop (6, 12). Two type 1 IFN gene products, IFN-α4 and IFN-β, are rapidly induced following viral infection. Newly synthesized IFN-α4 and IFN-β bind to the type 1 IFN receptor, resulting in induction of the IFN-stimulated gene factor-3 transcription factor (6, 12). IFN-stimulated gene factor-3 is a complex of STAT1, STAT2, and IRF9, and is critical for expression of non-α4 subtypes of IFN-α (6, 12). This most likely occurs indirectly through transcriptional induction of IRF7 expression (13). Indeed, high-level induction of IFN-α4/IFN-β and other IFN-α genes is dependent on IRF7 (11).
The greatest levels of type 1 IFN are produced by members of the DC lineage (14). Thus, virus infection of conventional DCs (cDCs) and especially plasmacytoid DCs (pDCs) results in very high levels of type 1 IFN expression (3, 14, 15). TLR ligand stimulation of both cDCs and pDCs also results in significant induction of type 1 IFN expression (3, 16). Depending on the specific TLR ligand, IFN expression is controlled by TLR-associated adaptor molecules MyD88 or Toll-IL1 domain-containing adaptor-inducing IFN-β (3, 16). An especially potent stimulus for type 1 expression is the TLR9 ligand CpG DNA D19 (14, 17). D19 is specifically retained in endosomal vesicles of pDCs, resulting in continuous high-level stimulation that allows potent type 1 IFN expression (17). Additional adapter molecules have also been shown to be essential for virus-induced type 1 IFN expression through both TLR-dependent and -independent pathways (18, 19).
NF-κB transcription factors are activated by virus infection and by all known TLR ligands (3, 16). NF-κB activation is mediated by proteasomal degradation of the IκB proteins, an event triggered by IκB phosphorylation by IKK-β (20). Five different members of the NF-κB family have been identified in mammals: p50, p52, RelA (p65), cRel, and RelB (21). A potential role for NF-κB in IFN-β expression was implicated nearly two decades ago, with subsequent studies indicating an important role for NF-κB factors in virus-induced IFN-β expression (8, 9, 22, 23). However, no definitive study has addressed the actual requirement for NF-κB in IFN-β (or IFN-α) regulation. Using NF-κB-deficient mouse embryonic fibroblasts (MEFs), cDCs, and pDCs, we therefore examined the role of NF-κB factors in virus- and TLR-induced expression of type 1 IFNs.
Materials and Methods
Mice and materials
C57BL/6 CD45.1 and C57BL/6 CD45.2 mice were purchased from Jackson Laboratories and bred in the animal facility of H. Lee Moffitt Cancer Center and Research Institute. RelA−/−, p50−/−, and cRel−/− mice have been described previously (24). All mice used were backcrossed into a C57BL/6 background for 10 generations. E14 wild-type (WT) and RelA−/− fetal liver hemopoietic precursors (CD45.2) were adoptively transferred into lethally irradiated CD45.1 recipient mice, as previously described (24). Recipient mice were typically used after 6–8 wk. All experiments with mice were conducted in accordance with institutional guidelines. D19 CpG DNA was synthesized by Sigma-Aldrich.
MEFs were cultured in fibroblast medium (DMEM, 10% of donor bovine serum, antibiotics). p50−/−RelA−/− fibroblasts were immortalized by continued culture. A total of 1 × 106 cells was plated in 6-cm tissue culture dishes in 5 ml of medium and cultured overnight before being used in experiments.
Bone marrow-derived DCs (BMDCs) were cultured from mouse bone marrow precursors following a modified procedure (24). Briefly, bone marrow cells were cultured in 24-well plates in DC medium (RPMI 1640, 5% FBS, 100 μM 2-ME and antibiotics) in the presence of 3% supernatant from J558L cells transduced with mouse GM-CSF at 1 × 106 cells/ml. Medium was changed every 2 days. On day 6, DC clusters were collected and resuspended in fresh DC medium containing GM-CSF. Mouse spleen pDCs were purified using Miltenyi Biotec anti-mouse plasmacytoid DC Ag-1 magnetic beads and cultured in DC medium. For all DC experiments, cells were plated at a density of 1 × 106 cells/ml.
Sendai virus (SV) and Newcastle disease virus (NDV) infections
MEFs or BMDCs were washed twice in culture medium without serum. The 200HA U/ml SV (Cantell strain; Charles River Laboratories) or NDV (strain B1 grown in embryonated chicken eggs) at a multiplicity of infection of 5 (for MEFs) or 2 (for DCs) was then added to cells without serum for ∼1 h. MEFs were then washed twice in medium without serum and incubated in culture medium for time periods indicated in the text. After a 1-h infection, serum and GM-CSF supernatant were added to DCs, which were then cultured further for time periods indicated in the text.
Retroviral infection of immortalized MEFs
The IκBα Ser32/36 S→A mutant gene was cloned into murine stem cell virus internal ribosome entry site GFP (MIG) retroviral expression vector. Retroviruses were produced by transfecting 293T cells with retroviral plasmid DNA and the pCL-Eco packaging plasmid. p50−/−RelA−/− fibroblasts were spin infected at 2600 rpm for 1 h at room temperature. GFP-positive cells were isolated by cell sorting and then cultured in fibroblast medium.
RNase protection assay (RPA) and real-time PCR
Total RNA was prepared by TRIzol reagent for RPA (Ambion) for detection of IFN-β and IL-12 expression. RPA was performed according to manufacturer’s recommendations (Ambion). For real-time PCR, RNA was further purified using Qiagen RNeasy kit. Purified RNA samples were then subjected to real-time PCR analysis by a standard curve method for relative quantitation from Applied Biosystems. Briefly, standards were prepared by serial dilution of a pool of all samples. RNA samples and standards were reverse transcribed into cDNA using TaqMan reverse-transcription reagents from Applied Biosystems. cDNAs were then subjected to real-time PCR analysis by standard curve method in an Applied Biosystems 7900HT Sequence Detection System using SYBR Green I Dye. All samples were run in triplicate, with SDs indicated in figures. Results for target genes are presented after normalizing to β-actin. PCR primers used were: 5′-AGCTCCAAGAAAGGACGAACAT-3′ (IFN-β-F), 5′-GCCCTGTAGGTGAGGTTGATCT-3′ (FIN-β-R); 5′-GTTCCTGCACTTGGCAATCATCCA-3′ (IκBα-F), 5′-TCAGGATCACAGCCAGCTTTCAGA-3′ (IκBα-R); 5′-ATGGAATCCTGTGGCATCCAT-3′ (β-actin-F), 5′-CCACCAGACAACACTGTGTTGG-3′ (β-actin-R); 5′-ARSYTGTSTGATGCARCAGGT-3′ (IFN-α (non-α4)-F), 5′-GGWACACAGTGATCCTGTGG-3′ (IFN-α (non-α4)-R) (sequence provided by D. Levy, New York, NY); 5′-CCTGTGTGATGCAGGAACC-3′ (IFN-α4-F), 5′-TCACCTCCCAGGCACTGA-3′ (IFN-α4-R); 5′-CTGGAGCCATGGGTATGCA-3′ (IRF-7-F), 5′-AAGCACAAGCCGAGACTGCT-3′ (IRF-7-R).
EMSA and ELISA
For detection of NF-κB activation, nuclear extracts were prepared and EMSA was performed, as described (24). After virus infection or D19 treatment, cells were cultured for indicated time periods and supernatants were collected. Production of mouse IFN-α and IFN-β was determined by ELISA kits from PBL Biomedical Laboratories following the manufacturer’s recommendations. Readings below the first standard curve sample were considered undetected.
Regulation of type 1 IFN expression in NF-κB-deficient embryonic fibroblasts following SV infection
Using MEFs, we investigated the individual and combined roles of NF-κB p50, RelA, and cRel in regulating IFN-β expression. To this end, MEFs were cultured for different time periods after infection with SV, following which IFN-β mRNA expression was determined by RPA. In WT MEFs, IFN-β expression was detected at 4, 6, or 12 h after SV infection, with maximal expression at 6 h (Fig. 1⇓A). IFN-β mRNA expression was indistinguishable among WT, p50−/−, and cRel−/− MEFs at all three time points (Fig. 1⇓A). Compared with other genotypes, IFN-β expression was slightly decreased at 4- and 6-h time points in RelA−/− MEFs (Fig. 1⇓A; also see below). IFN-β expression may be controlled by redundant function of p50, RelA, and cRel. To address this possibility, we first used p50−/−cRel−/− MEFs. Compared with WT MEFs, p50−/−cRel−/− MEFs showed similar induction of IFN-β at the different time points tested (Fig. 1⇓A). To quantify changes in IFN-β expression, we performed real-time PCR on mRNA from SV-infected MEFs. Consistent with RPA results, only RelA−/− MEFs showed a modest reduction in IFN-β expression at 4- and 6-h, but not 12-h time points (Fig. 1⇓A).
In addition to IFN-β, the α4 subtype of IFN-α can be directly induced by virus infection, i.e., in the absence of IFN autocrine signaling. Notably, the role of NF-κB p50, RelA, or cRel in regulating virus-induced IFN-α4 or IFN-α (non-α4) has not been determined. Using real-time PCR, we first determined the roles of the above NF-κB subunits in IFN-α4 mRNA expression. As shown in Fig. 1⇑B, no substantial difference in IFN-α4 expression was noticed between WT and the different knockout MEFs. IFN-α (non-α4) requires IFN-β- and IFN-α4-induced autocrine signaling for induction of mRNA expression. RelA−/− MEFs showed substantially reduced expression of IFN-α (non-α4) at 4 and 6 h, but elevated expression at 12 h (Fig. 1⇑C). This is most likely due to reduced expression of IFN-β and IFN-α4 at early time points, resulting in decreased IFN-α (non-α4) induction by autocrine signaling. Other knockout MEFs failed to show appreciable reduction in IFN-α (non-α4) expression (Fig. 1⇑C). Another crucial target of autocrine signaling is the gene encoding for the IRF7 transcription factor. As with IFN-α (non-α4), RelA−/− MEFs showed slightly reduced IRF7 expression, whereas, interestingly, p50−/−cRel−/− MEFs showed slightly elevated expression (Fig. 1⇑D). Because only RelA−/− and p50−/−cRel−/− MEFs showed any appreciable change in virus-induced gene expression, we specifically tested these cells in two additional experiments. First, we tested how SV-induced IκBα mRNA expression, a bona fide NF-κB target gene, was affected in RelA−/− and p50−/−cRel−/− MEFs. Unlike modest reductions in IFN gene expression, IκBα expression was dramatically reduced in RelA−/− MEFs, and to a lesser extent in p50−/−cRel−/− MEFs (Fig. 1⇑E). Thus, whereas RelA absence does not have a major impact on IFN-β, it is critical for induction of a bona fide NF-κB target gene. Second, we determined SV-induced IFN-β protein expression in RelA−/− and p50−/−cRel−/− MEF supernatants. Consistent with mRNA results, RelA−/− MEFs, but not p50−/−cRel−/− MEFs, showed a 30% decrease in IFN-β protein levels compared with WT MEFs.
Type 1 IFN expression in p50−/−RelA−/− MEFs expressing a dominant-negative variant of IκBα
Hetreodimers of p50 + RelA comprise the major NF-κB activity in MEFs. We next used immortalized p50−/−RelA−/− MEFs to determine virus-induced NF-κB activity. SV infection of WT immortalized MEFs led to highly significant increase in NF-κB levels that bound to the MHC class I (H-2Kb) and IFN-β κB sites (Fig. 2⇓A). Consistent with predominance of p50 and RelA in MEFs, overall SV-induced NF-κB activity was dramatically reduced in p50−/−RelA−/− MEFs (Fig. 2⇓A). However, a low-level activity was detected, which was partly inhibited by Abs against the three remaining NF-κB subunits (p52, cRel, and RelB) (Fig. 2⇓A). To inhibit virus-induced activation of these subunits, we transduced p50−/−RelA−/− MEFs with a retrovirus expressing a dominant-negative variant of IκBα, which is refractory to signal-induced phosphorylation and degradation (IκBαSA). Control retrovirus (MIG)- and IκBαSA retrovirus (MIG-IκBα)-infected p50−/−RelA−/− MEFs were FACS sorted based on GFP expression and then subjected to SV infection. Importantly, MIG-IκBα-transduced cells showed virtually no inducible NF-κB activity that bound to the MHC class I (H-2Kb) and IFN-β κB sites after SV infection (Fig. 2⇓A). Thus, IκBαSA expression in p50−/−RelA−/− MEFs can inhibit the three remaining NF-κB subunits (p52, cRel, and RelB), thereby providing a cell type virtually devoid of detectable virus-induced NF-κB activity.
We then used WT, p50−/−RelA−/−, MIG p50−/−RelA−/−, and MIG-IκBα p50−/−RelA−/− MEFs to investigate IFN-β expression after SV infection. The lowest expression of IFN-β was seen in MIG-IκBα p50−/−RelA−/− MEFs (Fig. 2⇑B). However, this was <2-fold reduced compared with WT MEFs at 4 and 6 h, whereas MIG-IκBα p50−/−RelA−/− MEFs showed higher expression than WT MEFs at the 9-h time point (Fig. 2⇑B). Furthermore, IFN-α4, IFN-α (non-α4), or IRF7 mRNA levels were also strongly induced in the different cell types (Fig. 2⇑, C–E). These results therefore indicate that SV induces reduced, yet robust expression of type 1 IFNs in the absence of detectable NF-κB activity.
Induction of IFN-β expression in NDV-infected NF-κB-deficient fibroblasts
It is possible that lack of a requirement for NF-κB in type 1 IFN expression reflects a unique property of SV. We therefore used select MEF subsets to determine IFN-β expression following infection with NDV. NDV-induced IFN-β expression in WT MEFs was strongest at 12 h after infection (Fig. 3⇓A). Thus, SV and NDV induce IFN-β expression with different kinetics, whereas SV shows maximal expression at 6 h postinfection and NDV shows maximal expression at 12 h postinfection. Notably, high levels of IFN-β expression were induced in WT, RelA−/−, and p50−/−cRel−/−MEFs (Fig. 3⇓A), although, as with SV infection, NDV-induced IFN-β was also reduced in RelA−/− MEFs. In addition, IFN-β expression was strongly induced in WT and p50−/−RelA−/−, MIG p50−/−RelA−/−, or MIG-IκBα p50−/−RelA−/− MEFs (Fig. 3⇓B). Taken together, our results thus indicate that both SV and NDV can induce IFN-β expression in the virtual absence of detectable NF-κB activity.
Regulation of virus- and TLR-induced IFN-β expression in p50-, cRel-, and RelA-deficient DCs
DCs are critically important for type 1 IFN expression following virus infection or TLR stimulation (14, 25). NF-κB proteins play a crucial role in DC development (24). Thus, combined absence of p50 and RelA severely impairs DC generation (24). In contrast, DC development can still take place in the combined absence of p50 and cRel (24). To determine how individual absence of RelA or combined absence of p50 and cRel impacts virus-induced IFN-β expression, we generated DCs from bone marrow of RelA−/− and p50−/−cRel−/− mice. NDV infection of DCs resulted in robust induction of IFN-β at 6 h postinfection (Fig. 4⇓A), whereas MEFs showed high expression at 12 h (Fig. 3⇑). These distinct kinetics are most likely due to differences in NDV infection kinetics of these two cell types. NDV or SV infection of RelA−/− BMDCs showed a modest reduction in IFN-β mRNA expression (Fig. 4⇓A), a result similar to virus infection of RelA−/− MEFs. In contrast, p50−/−cRel−/− BMDC infection with SV or NDV showed higher induction of IFN-β compared with WT DCs (Fig. 4⇓B). Similar to MEFs, we therefore conclude that neither RelA nor p50 + cRel function is essential for virus induction of IFN-β expression. Virus infection of DCs also leads to production of inflammatory and T cell-stimulatory cytokines such as TNF-α, IL-6, and IL-12 (26). Importantly, IFN responsiveness is crucial for this because DCs lacking the type 1 IFN receptor exhibit reduced cytokine expression (26). Previously, NF-κB has been shown to regulate LPS-induced expression of the p40 subunit of IL-12 in DCs and macrophages (24, 27). We therefore examined whether, in addition to the type 1 IFN pathway, virus-induced IL-12 p40 expression also requires NF-κB proteins. NDV, and to a significantly lesser extent SV, induced expression of IL-12 p40 in WT DCs (Fig. 4⇓). NDV-induced IL-12 p40 expression was reduced in RelA−/− DCs, but completely abolished in p50−/−cRel−/− DCs (Fig. 4⇓). Thus, despite robust IFN-β expression, p50−/−cRel−/− DCs are unable to induce IL-12 p40 expression. These results therefore indicate that p50 + cRel are independently involved in virus-induced IL-12 p40 expression.
pDCs are a critically important type 1 IFN-producing cell type. One of the most potent signals for pDC-induced type 1 IFN expression is D19-CpG, which signals though TLR9. To determine the roles of RelA, p50, and cRel in TLR9-induced type 1 IFN expression, we purified pDCs from RelA−/− and p50−/−cRel−/− mouse spleens. After a 20-h stimulation with D19, pDC culture supernatants were collected to determine IFN-α and IFN-β expression by ELISA. Remarkably similar to MEFs and BMDCs, a modest reduction in IFN-β (and IFN-α) expression was noticed in RelA−/− pDCs (Figs. 5⇓A and 1⇑F). In contrast, p50−/−cRel−/− pDCs showed moderately elevated expression of both IFN-α and IFN-β (Fig. 5⇓B), as well as elevated constitutive expression of IFN-β (Fig. 5⇓B). We therefore conclude that RelA, p50, or cRel is not essential for TLR9-induced type 1 IFN expression in pDCs.
Many studies in the past have proposed critically important roles for NF-κB in virus-induced IFN-β expression (8, 9, 22, 23). In the first definitive study of its kind, we have found that NF-κB/Rel transcription factors play only a modest role in virus-induced IFN-β expression. Only the individual absence of RelA, but not p50 or cRel, resulted in reduction of IFN-β mRNA expression. Our most definitive studies were performed on p50−/−RelA−/− fibroblasts. By expressing IκBαSA in p50−/−RelA−/− MEFs, we generated a fibroblast cell line that exhibited virtually no detectable virus-induced NF-κB activity. Even in these cells, albeit reduced compared with WT cells, SV or NDV induced robust IFN-β expression. Nonetheless, we cannot exclude the possibility that low-level complexes of p52 and RelB, which appeared to be inhibited following IκBαSA expression, can still play a role in modulating IFN-β expression. Based on studies of knockout cells, we can conclude that whereas NF-κB subunits p50, RelA, and cRel play a role, they are apparently not essential for virus-induced IFN-β expression.
Three families of transcription factors have been implicated in IFN-β expression, as follows: IRF, NF-κB, and AP-1. It has been shown that IRF3/7 factors play an essential role in transcriptional induction of IFN-β (6). In contrast, our results indicate that NF-κB factors are largely dispensable for this induction. However, NF-κB factors may still play an important regulatory role in fine-tuning the kinetics and robustness of IFN-β induction in response to viral infection or other stimuli. It is also possible that NF-κB factors are more important for IFN-β expression in subset of cells or cell types with limiting amounts of other transcription factors, such as IRF3/7 or AP-1. In any case, our results indicate that NF-κB plays a more subtle role in IFN-β expression than previously thought. Virus infection leads to formation of the IFN-β enhanceosome (8), a structure that has been used as a paradigm for regulation of inducible transcription. Furthermore, NF-κB has been proposed to play a crucial role in assembly of the IFN-β enhanceosome (8, 9). Our results, however, indicate that either NF-κB factors are not required for enhanceosome assembly or virus-induced IFN-β expression can take place in the absence of the enhanceosome. An additional means for regulating mRNA levels is through control of stability. Indeed, mRNA levels of many cytokines, including IFN-β, are subject to regulation through control of stability. Although we have not examined this possibility in the present study, it is plausible that NF-κB proteins can positively or negatively modulate IFN-β (or IFN-α; see below) mRNA levels posttranscriptionally. Additional studies will be required to test this interesting possibility.
We have also determined the roles of NF-κB RelA, p50, and cRel subunits in virus- and TLR-induced IFN-β expression in DCs. Although our findings with DCs also demonstrate that RelA, p50, and cRel are not essential for IFN-β expression, interesting differences were noticed in IFN-β expression between WT and knockout DCs. As in fibroblasts, a modest reduction in IFN-β mRNA and protein expression was noticed in RelA−/− BMDCs and pDCs. Interestingly, virus infection or D19 treatment of p50−/−cRel−/− BMDCs and pDCs, respectively, showed elevated IFN-β mRNA and protein expression compared with WT DCs. Thus, although not essential for virus-induced IFN-β expression, NF-κB proteins may still influence aspects of IFN-β expression. Previously, the NF-κB-activating kinase IKK-β was reported to control type 1 IFN expression, whereas more recently, IKK-α was shown to participate in IRF7 activation (28, 29). In light of our findings, it is therefore plausible that both IKK-α and IKK-β control type 1 IFN expression by regulating the IRF3/IRF7 activation pathway. Importantly, we have found that the combined functions of p50 + cRel are crucial for virus-induced IL-12 p40 expression. Thus, unlike their role in virus-induced IFN-β expression, NF-κB proteins play critically important roles in mediating specific host responses to viral infection.
We have also performed the first detailed analysis of the role of NF-κB proteins in IFN-α expression. Similar to IFN-β, we found that virus infection leads to robust IFN-α expression in NF-κB-deficient MEFs. In addition, RelA−/−- or p50−/−cRel−/− CpG D19-treated pDCs also strongly induced IFN-α expression, consistent with a recent report (30). Virus-induced transcriptional induction of IRF7 requires type 1 IFN autocrine signaling. We also examined virus-induced expression of IRF7 to determine a potential role for NF-κB proteins in regulating IFN autocrine signaling. However, no highly significant differences were seen in IRF7 expression between WT and the different knockout cells, indicating that neither type 1 IFN expression nor downstream signaling events are absolutely dependent on NF-κB proteins. Interesting differences, especially between WT and p50−/−RelA−/− cells, were nonetheless noted. Thus, although IRF7 regulates IFNα (non-α4), IRF-7 levels were decreased, whereas IFNα (non-α4) levels were increased in p50−/−RelA−/− cells in comparison with WT cells. Additionally, the main difference in IFN-α (non-α4) mRNA levels was observed at 9 and 12 h postinfection, when levels in WT cells were rapidly decreasing. We believe that a possible explanation for these results may be that decay of IFN-α (non-α4) mRNA levels is slower in p50−/−RelA−/− cells, suggesting possible control of mRNA stability by NF-κB proteins, as mentioned above.
Members of the NF-κB and IRF families of transcription factors are among the most crucial for mediating host responses to microbial agents. Recent studies using IRF3−/− and IRF7−/− fibroblasts and DCs have demonstrated an essential role for these transcription factors in virus-induced type 1 IFN expression (6). In contrast to these findings, our results indicate that robust virus-induced type 1 IFN expression can take place in the absence of key NF-κB subunits and following dominant-negative IκBα expression. Together with previous studies, the results shown in this study indicate discrete functions for these two transcription factor families in mediating host responses to virus infection: IRF3/7 being crucial for induction of type 1 IFN responses and NF-κB factors playing essential roles in inducing key cytokine genes.
We thank Dr. Christian Schindler (Columbia University, New York, NY) for helpful comments on this manuscript. We also acknowledge help provided by Flow Cytometry and Molecular Biology core facilities at the Moffitt Cancer Center.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by National Institutes of Health Grant R01 AI059715 and institutional funds from Moffitt Cancer Center (to A.A.B.) and National Institutes of Health U19AI62623 funded Center to Investigate Viral Immune Antagonism (to A.G.-S.).
↵2 Address correspondence and reprint requests to Dr. Amer A. Beg, Department of Interdisciplinary Oncology, H. Lee Moffitt Cancer Center & Research Institute, University of South Florida, 12902 Magnolia Drive, Mail Stop: SRB-2, Tampa, FL 33612. E-mail address:
↵3 Abbreviations used in this paper: IKK, IκB kinase; BMDC, bone marrow dendritic cell; cDC, conventional dendritic cell; DC, dendritic cell; IRF, IFN regulatory factor; MEF, mouse embryonic fibroblast; MIG, murine stem cell virus internal ribosome entry site; NDV, Newcastle disease virus; pDC, plasmacytoid DC; PRD, positive regulatory domain; RPA, RNase protection assay; SV, Sendai virus; WT, wild type.
- Received December 27, 2006.
- Accepted March 16, 2007.
- Copyright © 2007 by The American Association of Immunologists