See the article "An innate immune response has developed as a rapid and regulated defense mechanism in which the recognition of an invading pathogenic organism can occur on binding to a specific viral receptor or Toll-like receptor (TLR) that can recognize the conserved patterns of the proteins, lipoproteins, double-stranded RNA (dsRNA), or unmethylated CpG DNA (1).
As an early response to infection, cells induce a profile of the early inflammatory genes, including antiviral cytokines and chemokines. Two families of the transcriptional factors play a major role in the transcriptional activation of these genes: a well characterized family of NF-κB factors and a newly emerging family of IFN regulatory factors (IRFs). The IRFs play a critical role in the induction of type I IFN and chemokine genes and genes mediating antiviral, antibacterial, and inflammatory responses. Three of these IRFs, IRF-3, IRF-5, and IRF-7, function as direct transducers of virus-mediated signaling (2). In uninfected cells, these IRFs reside in the cytoplasm, whereas in infected cells they are phosphorylated on serine residues in the carboxyl-terminal region of the polypeptide and are transported to the nucleus. In infected cells, ubiquitously expressed IRF-3 mediates induction of IFN-B genes and some of the IFN-induced genes, whereas IRF-5 and IRF-7, expression of which is limited to lymphoid cells, are required for induction of most of the IFN-A genes. Recent data indicate that the binding of a ligand, dsRNA to TLR-3 and lipopolysaccharide (LPS) to TLR-4, also activates IRF-3 and IRF-7 (3). Binding of LPS to TLR-4 (Fig. 1) activates signaling events not only through the MyD88 adapter but also by the MyD88-independent pathway. This second pathway is mediated by two adapters named TRIF and TRAM that activate IRF-3 and induce IFN-β (3, 4). In contrast to TLR-4, TLR-3 does not associate with MyD88, MAL, and TRAM, and thus binding of dsRNA to this receptor, which is believed to mimic viral infection, activates IRF-3 only through the TRIF adapter (Fig. 1).
 | Fig. 1. Signaling pathways for TLR-3 and TLR-4. The interaction of LPS with TLR-4 that requires cooperation of two other proteins, CD14 and MD-2, initiates two major signaling pathways. The MyD88-dependent signaling pathway leads to the phosphorylation of IRAK, (more ...) |
These observations show that after initial recognition of the pathogen, the signaling pathways merge and indicate the existence of cross talk between virus- and bacteria-induced signaling.
In a recent issue of PNAS, McWhirter et al. (5) addressed the identity of the kinase that activates IRF-3 in infected cells and on binding of the ligands to TLR-3 or TLR-4. Thus, this study is a follow-up of previous reports (6, 7) in which two noncanonic IκB kinases, IKKε and TBK-1, were implicated in the phosphorylation and activation of IRF-3 in cultured human cell lines in vitro. There is high homology between these two kinases (8, 9). Both kinases are synergistic with TANK (TRAF family-associated NF-κB activator), which links them to the IKK complex by TANK association with IKKγ/nemo (10). Silencing studies with a specific small interfering RNA indicated that both of these kinases are involved in the Sendai virus-mediated activation of IRF-3 in cultured cells (6, 7). McWhirter et al. show that recombinant IKKε and TABK-1 phosphorylate IRF-3 directly and that the primary target of the phosphorylation is a cluster of four serines (Ser-396, -398, -402, and -405) and Thr-404 that is localized in the carboxyl-terminal region of IRF-3. Both IKKε or TABK-1 kinase also phosphorylate IKBα, but only on Ser-36, and because the degradation of IKBα by the ubiquitination pathway requires phosphorylation of Ser-32 and Ser-36, neither IKKε nor TABK-1 is able to target IKBα for degradation and activate NF-κB (8, 9).
McWhirter et al. (5) then examined the role of TABK-1 in activation of IRF-3 in cultured cells by using TABK-1-deficient mouse embryo fibroblasts. The TABK-1 deficiency is embryonic-lethal, and mice die from massive liver degeneration and apoptosis that is strikingly similar to that observed in relA-, IKKβ-, and IKKγ-deficient mice (11). Consequently, it was suggested that TABK-1 has a unique role in the NF-κB response induced by proinflammatory cytokines. McWhirter et al. show that the transcriptional activation of the promoters of IRF-3-targeted genes, such as IFN-B and Rantes, by Sendai virus, LPS, or dsRNA was completely abrogated in TBK-1-deficient cells. In contrast, activation of NF-κB-regulated promoters was not affected by TBK-1 deficiency. To further prove the essential role of TBK-1 in the activation of IRF-3 in response to viral infection and TLR-3 and TLR-4 ligands, McWhirter et al. demonstrated that IRF-3 is not activated and translocated to the nucleus in TBK-1-deficient cells and that, in contrast to WT MEM, expression of IRF-3-targeted genes, such as type I IFN and Rantes, is profoundly inhibited in virus-infected, poly(IC)-treated, or LPS-stimulated TBK-1-deficient cells. Tumor necrosis factor α-mediated induction of the Rantes gene was the same in WT and TBK-1-deficient cells, however, indicating that TBK-1 was not required for activation of the IKKα,β,γ complex and IKBα degradation.
The TIR containing adapter (TRIF) is a very potent activator of the IFN-B promoter (4). TRIF was shown to associate with TLR-3 and directly interact with TRAF-6, thus linking TLR-3 and TLR-4 for NF-κB activation that is independent on MyD88 and Irak-1 (12) (Fig. 1). Mice that are TRIF-deficient (13) or have a mutation in the TRIF gene (14) have a profound defect in NF-κB activation and fail to produce IFN-β. Addressing the molecular mechanism of IRF-3 activation and stimulation of IFN-B and Rantes genes in infected, dsRNA, or LPS-treated cells, Fitzgerald et al. (6) have shown that IKKε and TBK-1 associate with TRIF. In contrast, Sato et al. (12) only detected an association of TRIF with TBK-1 and formation of a TRIF, TBK-1, and IRF-3 complex that depended on kinase activity of TBK-1 and phosphorylation of TRIF. Those authors also found that the association of TBK-1 and TRAF-6 with TRIF is mutually exclusive. The availability of TBK-1-deficient mouse embryo fibroblasts enabled McWhirter et al. (5) to examine directly whether TBK-1 activity is required for TRIF to function. They found that TRIF-mediated activation of IRF-3 and stimulation of IRF-3-activated promoters, such as IFN-B, IP-10, and Rantes, were completely abrogated in TABK-1-deficient cells, whereas TRIF-mediated activation of NF-κB-dependent promoters, such as ELAM, did not require TBK-1 activity.
It is noteworthy that whereas TRIF was shown to be solely responsible for the TLR-3-mediated activation of IRF-3 and IFN-β, a recently identified adapter protein, TRAM (3), that associates with TLR-4 also is required for LPS-mediated activation of NF-κB and IRF-3. Taken together, these results indicate that the initial signaling events leading to the activation of TBK-1 by TLR-3 and TLR-4 may not be identical.
Although McWhirter et al. (5) have clearly documented the critical role of TBK-1 in the activation of IRF-3 and IFN-B genes by dsRNA, LPS, and viral infection in mouse embryo fibroblasts, certain questions remain. What is the role of IKKε? Does it function in a cell-specific manner? TBK-1 is ubiquitously expressed and could therefore activate IRF-3 during the primary viral response, whereas IKKε that is expressed in lymphoid cells could target IRF-7 and IRF-5. Do IKKε and TBK-1 phosphorylate identical or distinct serines in the IRF-3 polypeptide? At present, identification of the phosphorylation site in the IRF-3 peptide is based on analysis of the mutated IRF-3 peptides, but the IRF-3 sites phosphorylated in infected or dsRNA-treated cells have not yet been identified. Recent observations that the induction of IFN genes and activation of IRF-3 mediated by a ligand binding to TLR-7 and TLR-9 seem to be MyD88-dependent are unexpected findings (15). The induction of IFN has been associated with the utilization of TRIF by TLR-3 and TLR-4; whether TRIF associates with TLR-7 and TLR-9 and whether all these different sets of adapters activate the same kinase, TBK-1, is worthwhile addressing.
There are profound similarities in cellular response to bacterial and viral infections.
Another issue that needs to be resolved is the mechanism of TLR-3-independent dsRNA signaling. It has been common knowledge that dsRNA can induce IFN-β in fibroblasts and cell lines that do not express TLR-3. In addition, two recently published studies (16, 17) show that dsRNA can mediate signaling in a TLR-3-independent manner. Diebold et al. (16) found that nonplasmoid dendritic cells induced high levels of IFN in response to dsRNA introduced into the cytoplasm in the absence of TLR-3 and MyD88. Hoebe et al. (17) identified an allele on chromosome 7, designated dsRNA1, that controls induction of IFN-β in the absence of TLR-3 and TRIF. The role of TABK-1 in the TLR-independent dsRNA induction of IFN is not known.
Finally, it is noteworthy that several virus-specific autosomal loci, named If loci, were identified by a genetic analysis (18). These If loci, which have high and low alleles expressed on macrophages and hematopoetic cells, determine the levels of IFN production in infected and poly(IC)-treated inbred strains of mice. Although ≈10 different If loci were found, only 2 of them were mapped. Thus the If-1 locus influencing IFN induction by Newcastle disease virus was assigned to chromosome 3, and herpes simplex virus-induced IFN production was shown to be X-linked and assigned If-x locus. Clearly it will be of great interest to determine whether the virus-specific If loci represent distinct TLRs or a critical component of the TLR signaling pathway.
Perhaps the most important message from the work of McWhitler et al. (5) is that the more closely we look, the more profound similarities we find in the cellular response to bacterial and viral infection. Although different TLRs use distinct sets of adapters, the fact that a single kinase (TBK-1) and a single transcription factor (IRF-3) play critical roles in the induction of IFN-β and some chemokines indicates that the activity of this TBK-1 could be targeted by an inhibitor in situations when induction of type I IFNs is detrimental. Alternatively, a selective stimulation of TBK-1 activity may maximize the immune response to vaccine antigens, as IRF-3 seems to function as an effective DNA vaccine adjuvant (19). Some of these clinical interventions will be harder to implement than others, but more detailed insight into the signaling mechanisms generated during the innate response to the infection will outweigh the possible obstacles associated with the genetic variability and alternative pathways.