A plasmacytoid dendritic cell (DC) can produce large amounts of type I IFNs after sensing nucleic acids through TLR7 and TLR9. IκB kinase α (IKKα) is critically involved in this type I IFN production through its interaction with IFN regulatory factor-7. In response to TLR7/9 signaling, conventional DCs can also produce IFN-β but not IFN-α in a type I IFN-independent manner. In this study, we showed that IKKα was required for production of IFN-β, but not of proinflammatory cytokines, by TLR7/9-stimulated conventional DCs. Importantly, IKKα was dispensable for IFN-β gene upregulation by TLR4 signaling. Biochemical analyses indicated that IKKα exerted its effects through its interaction with IFN regulatory factor-1. Furthermore, IKKα was involved in TLR9-induced type I IFN-independent IFN-β production in vivo. Our results show that IKKα is a unique molecule involved in TLR7/9-MyD88–dependent type I IFN production through DC subset-specific mechanisms.
Dendritic cells (DCs) sense nucleic acid immune adjuvants ssRNA and unmethylated CpG DNA via TLR7 and TLR9, respectively, and produce proinflammatory cytokines or type I IFNs (1–4). Accumulating evidence indicates that TLR7/9-induced type I IFN induction is important not only for antiviral defense but also for the pathogenesis of certain autoimmune diseases (5, 6). Therefore, clarifying the underlying mechanisms of this type I IFN production should contribute to the development of immunomanipulation for these diseases.
TLR7 and TLR9 associate with the cytoplasmic adaptor, MyD88, through the homophilic interaction of their respective Toll/IL-1R homologous domains (3). MyD88 is essential for all TLR7/9-mediated effects. Signaling downstream of MyD88 bifurcates into two pathways leading to the activation of NF-κB and IFN regulatory factor (IRF)-7 (7). NF-κB activation leads to production of proinflammatory cytokines and requires phosphorylation and degradation of IκB. IκB phosphorylation depends mainly on a serine threonine kinase, IκB kinase (IKK)β (8).
IRF-7 activation leads to type I IFN production (9). This pathway functions mainly in a specialized DC subset, the plasmacytoid DC (pDC) (10, 11). The pDC expresses TLR7 and TLR9 exclusively among TLRs, has constitutively high levels of IRF-7, and is notable for its potent ability to produce type I IFNs, including IFN-α and IFN-β, following TLR7/9 signaling. IRF-7 activation requires its phosphorylation, and we have found that IKKα, also known as Chuk, is involved in this IRF-7 activation (12).
In response to TLR7/9 signaling, another DC subset, the conventional DC (cDC), can also produce IFN-β, but not IFN-α, through the distinct molecular mechanisms from pDCs (13, 14). In this study, we have investigated the involvement of IKKα in IFN-β production by TLR7/9-stimulated cDCs.
Materials and Methods
Ikka−/−, Myd88−/−, and Tlr4−/− mice have been described previously (15–17). TNFR-associated factor 3 (TRAF3)-deficient (Traf3−/−) mice were generated by T. Yasui and H. Kikutani (manuscript in preparation). C57BL/6, osteopontin (Opn)-deficient (Opn−/−), and IFN-α/β R-deficient (Ifna/br−/−) mice were purchased from CLEA Japan (Shizuoka, Japan), The Jackson Laboratory (Bar Harbor, ME), and B&K Universal (Hull, U.K.), respectively. IL-1R–associated kinase-1 (IRAK-1)–deficient (Irak1−/−) mice were provided by Dr. J. A. Thomas (University of Texas Southwestern Medical Center, Dallas, TX) (18). Bone marrow (BM) chimeric mice were generated by transferring wild-type, Ikka−/−, or Traf3−/− fetal liver cells into C57BL/6 (CD45.1) mice (19, 20). Ikka+/+Ifna/br−/− and Ikka−/−Ifna/br−/− chimeric mice were generated by transferring fetal liver cells from Ikka+/+Ifna/br−/− or Ikka−/−Ifna/br−/− mice into irradiated Ikka+/+Ifna/br−/− mice. All mice were maintained under specific pathogen-free conditions, and animal experiments were conducted according to the institutional guidelines.
GM-CSF–induced BM DCs were generated as previously described (21) and used as cDCs.
A phosphorothioate oligodeoxynucleotide containing an unmethylated CpG motif, ODN1668, was used as CpG DNA (22). LPS derived from Salmonella minnesota Re595, polyinosinic:polycytidylic acid [poly(I:C)], and R848 were purchased from Sigma-Aldrich (St. Louis, MO), Amersham Biosciences (Piscataway, NJ), and InvivoGen (San Diego, CA), respectively.
Measurement of cytokine production
Cells were treated for 20–24 h with the indicated stimuli, and cytokine production was measured by ELISA as described previously (12).
Northern blot analysis
Northern blot analysis was performed as described previously (21).
IFN-stimulated response element-binding activities were analyzed as described previously (12).
Nuclear translocation of IRF-1 and NF-κB
Nuclear extracts were subjected to immunoblot analyses with anti–IRF-1 or anti-p65 Abs (Santa Cruz Biotechnology, Santa Cruz, CA) as described previously (12).
The 293T human embryonic kidney cell line was transiently transfected with expression vectors for FLAG-tagged mouse IRF-1 (FLAG-IRF-1) and/or Myc-tagged mouse IKKα (Myc-IKKα) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Subsequently, immunoprecipitation and immunoblotting experiments were performed as described previously (12).
In vitro kinase assay
One hundred nanograms of purified recombinant IKKα (Upstate Biotechnology, Lake Placid, NY) was incubated with 1 μg substrates, GST, GST fused to residues 2–329 of mouse IRF-1 (GST-IRF-1), or GST fused to residues 5–55 of mouse IκBα (GST-IκBα) and subjected to an in vitro kinase assay as described previously (12).
Serum IFN-β levels after CpG DNA injection
Each mouse was injected i.p. with 20 nmol CpG DNA and 20 mg d-(+)-galactosamine. At the indicated times, the mice were bled, and serum IFN-β levels were measured by ELISA (PBL InterferonSource, Piscataway, NJ).
A two-tailed, unpaired Student t test was used for assessment of the differences between groups. Prism (Graphpad Software, La Jolla, CA) was used for statistical analysis. Differences were considered to be significant when the value of p < 0.05.
Results and Discussion
Type I IFN production from TLR9-stimulated cDC
An Ifna/br−/− pDC has severe defects in TLR9-induced type I IFN production (9), suggesting the involvement of positive feedback loop through IFN-α/β R. We have analyzed whether this feedback loop is also working in cDCs. Wild-type cDCs produced IFN-β, but not IFN-α, in response to the TLR9 agonist CpG DNA, and the IFN-β production increased in a dose-dependent manner. Ifna/br−/− cDCs also produced comparable amounts of IFN-β (Supplemental Fig. 1). This IFN-β production was dependent on TLR9, because Tlr9−/− cDCs failed to produce any IFN-β upon stimulation with CpG DNA (data not shown). Thus, TLR9-stimulated cDCs can produce IFN-β in an IFN-α/β R-independent manner.
Critical roles of IKKα in IFN-β production by TLR7/9-stimulated cDCs
We have analyzed involvement of IKKα. TLR9-induced IFN-β induction was severely impaired in Ikka−/− cDCs (Fig. 1A). This was a specific effect, because wild-type and Ikka−/− cDCs produced comparable amounts of proinflammatory cytokines, such as IL-12p40 and TNF-α, in response to TLR9 stimuli (Fig. 1A).
We next investigated expression of these cytokine genes by Northern blot analysis (Fig. 1B, 1C). CpG DNA could upregulate expression of IFN-β as well as proinflammatory cytokine genes in wild-type but not in Myd88−/− cDCs, indicating their dependence on MyD88. Ikka−/− cDCs showed a severe defect in TLR9-induced IFN-β gene expression (Fig. 2B) but normal induction of IL-12p40 and TNF-α gene expression. Similar defects were observed also in TLR7-stimulated Ikka−/− cDCs (Fig. 2C). Thus, as with pDCs, IKKα is critical for TLR7/9-provoked induction of type I IFN.
Although a pDC fails to respond to a TLR4 agonist, LPS, a cDC can upregulate expression of proinflammatory cytokines and IFN-β upon stimulation with LPS (Fig. 1D). In Myd88−/− cDCs, induction of proinflammatory cytokine genes was abolished, but IFN-β gene induction was preserved as reported previously (21, 23). LPS-mediated IFN-β induction does not depend on MyD88 but on another cytoplasmic adaptor, Toll/IL-1R domain-containing adaptor-inducing IFN-β (TRIF) (23, 24). When an Ikka−/− cDC was stimulated with LPS, expression of IFN-β as well as proinflammatory cytokines was upregulated (Fig. 1D). These results indicate that IKKα is critical for TLR7/9/MyD88- but not TLR4/TRIF-mediated IFN-β gene induction in cDCs.
cDCs can also produce type I IFNs in response to the dsRNA analog poly(I:C). This response depends on a cytosolic RNA sensor, MDA5, which belongs to the RIG-I–like receptor (RLR) family (25, 26). When stimulated with poly(I:C), wild-type and Ikka−/− cDCs produced comparable amounts of IFN-α and IFN-β (Supplemental Fig. 2), indicating that IKKα is dispensable for RLR-induced type I IFN production from cDCs.
Roles of MyD88-associated molecules in IFN-β gene induction in TLR9-stimulated cDCs
Several molecules have been reported to be involved in TLR7/9-mediated type I IFN production by pDCs. TRAF3 associates with the TLR adaptors MyD88 and TRIF as well as IKK family members, such as TANK-binding kinase 1, and is critically involved in TLR- and RLR-induced type I IFN production. TRAF3 deficiency leads to defective induction of type I IFNs by RLRs and TLRs, including TLR4, TLR7, and TLR9 (27, 28). Opn functions as an intracellular signaling molecule and, by associating with MyD88, plays critical roles in IRF-7 activation and type I IFN production in pDCs (29). Furthermore, a serine threonine kinase, IRAK-1, also associates with MyD88 and IRF-7 and is critical for TLR7/9-induced type I IFN production (30).
As with wild-type cDCs, Traf3−/−, Opn−/−, and Irak1−/− cDCs upregulated IFN-β and IL-12p40 gene expression after TLR9 stimuli (Fig. 2). Thus, IKKα is distinguished from these three molecules in terms of the involvement in TLR9-induced type I IFN production by both pDCs and cDCs.
Impaired activation of IRF-1 and NF-κB in TLR9-stimulated Ikka−/− cDCs
IFN-β gene expression is regulated by several IRF family members or NF-κB (31). Which signaling molecules are activated depends on the stimuli or stimulated cell types. IRF-3 dimer formation is required for LPS-induced IFN-β gene expression (32). However, CpG DNA fails to induce IRF-3 dimer formation (Supplemental Fig. 3). IRF-7 activation was severely impaired in Ikka−/− pDCs (12). However, amounts of IFN-stimulated response element-binding complexes containing IRF-7 were increased in TLR9-stimulated Ikka−/− cDCs (Supplemental Fig. 4). Thus, it is unlikely that IKKα is critically involved in IFN-β gene upregulation in cDCs through IRF-3 or IRF-7.
A cDC does not require IRF-7 or IRF-3 but instead requires IRF-1 for TLR9-induced IFN-β production (13, 14). We have investigated the nuclear translocation of IRF-1 and the NF-κB subunit, p65, in LPS or CpG DNA-stimulated cDCs (Fig. 3A). Upon stimulation with LPS, wild-type cDCs increased nuclear IRF-1 levels, and this increase was not impaired in Ikka−/− cDCs. When stimulated with CpG DNA, wild-type cDCs also had elevated levels of IRF-1 in the nucleus, but this response was defective in Ikka−/− cDCs. Nuclear p65 protein levels were also increased in wild-type cDCs after stimulation with LPS or CpG DNA. This increase, however, was significantly impaired in CpG DNA- but not LPS-stimulated Ikka−/− cDCs. These results indicate that IKKα is required for IRF-1 and p65 activation in TLR9-stimulated cDCs.
We then analyzed whether IKKα can interact with IRF-1 (Fig. 3B). When FLAG-tagged IRF-1 was expressed with Myc-tagged IKKα in 293T cells, IKKα was coimmunoprecipitated with IRF-1. Moreover, an in vitro kinase assay revealed that IKKα could phosphorylate IRF-1 as well as IκBα (Fig. 3C).
In cDCs, IRF-1 interacts and colocalizes with MyD88. After TLR9 stimuli, IRF-1 is licensed and migrates into the nucleus. IRF-1 is phosphorylated in a MyD88-dependent manner (13), although it remains unclear whether IRF-1 phosphorylation is critical for its activation or nuclear translocation. A critical involvement of IRF-1 was shown by the finding that Irf1−/− cDCs fail to produce IFN-β in response to TLR9 (13, 14). Similar to Ikka−/− cDCs, Irf1−/− cDCs retained the ability to produce TNF-α or IL-12p40 in response to CpG DNA. Importantly, we have found that nuclear translocation of IRF-1 was decreased in TLR9-stimulated Ikka−/− cDCs, indicating that IKKα is required for IRF-1 activation by TLR9 signaling. Our present results also demonstrated that IKKα could associate with IRF-1 and phosphorylate IRF-1. Thus, it can be assumed that IKKα is involved in TLR7/9-mediated IFN-β induction through its interaction with IRF-1. Nuclear translocation of the NF-κB p65 subunit was also decreased in TLR9-stimulated Ikka−/− cDCs. However, this decrease did not lead to defective production of proinflammatory cytokines.
In vivo roles of IKKα in TLR9-induced type I IFN-independent IFN-β gene induction
We have further investigated whether and how the mechanisms defined in the in vitro cDCs contribute to in vivo responses. For this purpose, we have analyzed the role of IKKα in TLR9-induced IFN-β production in the absence of IFN-α/β R signaling. As with wild-type mice, Ifna/br−/− mice injected with CpG DNA showed significant or even exaggerated elevation of serum IFN-β (Fig. 4A). The results indicate that CpG DNA can induce IFN-β in vivo in a type I IFN-independent manner.
We have further examined the importance of IKKα in Ifna/br−/− genetic backgrounds. Compared with Ikka+/+Ifna/br−/− chimeric mice, the Ikka−/−Ifna/br−/− mice had a severe defect in the elevation of serum IFN-β levels after the injection of CpG DNA (Fig. 4B). Thus, IKKα is required for type I IFN-independent IFN-β induction in vivo by TLR9.
IFN-β production from cDCs is less than that from pDCs. However, the cDC has a potent ability to activate T cells because of a higher expression of costimulatory molecules than the pDC and, moreover, is numerically much more plentiful than the pDC. Therefore, type I IFN produced by cDCs should play critical roles in shaping immune responses. Notably, unlike pDCs, cDCs can produce type I IFN in an IFN-α/β R-independent manner, and Ifna/br−/− mice had a significant increase in serum IFN-β levels after injection of a TLR9 agonist. Although it is currently unclear why the increase is more evident in Ifna/br−/− mice than in wild-type mice, the results clearly indicate that TLR9-mediated IFN-α/β R-independent IFN-β induction also functions in an IKKα-dependent manner in vivo.
We have also examined other types of cDCs including splenic or Flt3 ligand-induced BM cDCs (11). IKKα was required for these cDCs to produce IFN-β in response to TLR9 signaling (Supplemental Fig. 5). In these cDCs, IKKα was, although partially, involved also in other cytokine gene induction. IKKα should be considered as a unique molecular target for manipulating type I IFN induction by TLR7/9 independently of the DC subsets.
We thank N. Iwami, Y. Fukuda, and E. Haga for technical assistance and S. Haraguchi for secretarial assistance. We also thank P. Burrows for critical reading of the manuscript.
Disclosures The authors have no financial conflicts of interest.
This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology and the Japan Science and Technology Corporation, the Uehara Memorial Foundation, the Mochida Memorial Foundation for Medical and Pharmaceutical Research, and the Japan Intractable Diseases Research Foundation.
The online version of this article contains supplemental material.
Abbreviations used in this paper:
- bone marrow
- Coomassie brilliant blue
- conventional dendritic cell
- dendritic cell
- IκB kinase
- IL-1R–associated kinase-1
- IFN regulatory factor
- plasmacytoid dendritic cell
- polyinosinic:polycytidylic acid
- RIG-I–like receptor
- TNFR-associated factor 3
- Toll/IL-1R domain-containing adaptor-inducing IFN-β.
- Received May 27, 2009.
- Accepted January 27, 2010.
- Copyright © 2010 by The American Association of Immunologists, Inc.