IKKi/IKK (cid:1) Plays a Key Role in Integrating Signals Induced by Pro-inflammatory Stimuli*

We report that the product of the inducible gene encoding the kinase known as IKKi/IKK (cid:1) (IKKi) is required for expression of a group of genes up-regulated by pro-inflammatory stimuli such as bacterial endotoxin (li-popolysaccharide (LPS)). Here, using murine embryonic fibroblasts obtained from mice bearing deletions in IKK2, p65, and IKKi genes, we provide evidence to support a link between signaling through the NF- (cid:2) B and CCAAA/enhancer-binding protein (C/EBP) pathways. This link includes an NF- (cid:2) B-dependent regulation of C/EBP (cid:3) and C/EBP (cid:4) gene transcription and IKKi-medi-ated activation of C/EBP. Disruption of the NF- (cid:2) B pathway results in the blockade of the inducible up-regula-tion of C/ (cid:5) BP (cid:3) , C/EBP (cid:4) , and IKKi genes. Cells lacking IKKi are normal in activation of the canonical NF- (cid:2) B pathway but fail to induce C/EBP (cid:4) activity and transcription of C/EBP and C/EBP-NF- (cid:2) B target genes in response to LPS. In addition we show that, in response to LPS or tumor necrosis factor (cid:6)

overlapping subsets of genes encoding immune and pro-inflammatory modulators (3,9). Whether or not NF-B and C/EBP family members influence each other is not well understood at this time. Current paradigms suggest that the rapidly and transiently activated NF-B pathway is central in a primary wave of gene induction followed by a second wave of gene transcription some hours later mediated by other transcription factors including members of the C/EBP family (3,5).
Despite participating in regulation of overlapping sets of genes, the NF-B and the C/EBP pathways are activated by distinct intracellular signaling mechanisms. Activation of the canonical NF-B pathway depends on stability of the inhibitor known as IB␣. It stabilizes NF-B complexes so that after its degradation the remaining subunits translocate from the cytoplasm to the nucleus. NF-B-dependent transcription of IB␣ gene also provides a potent feedback mechanism maintaining the balance between cytoplasmic and nuclear localization of the NF-B subunits (8). Treatment of cells with specific inducers, such as TNF, IL-1, or LPS, results in the phosphorylation of IB␣ at two serines (Ser-32 and Ser-36). This is a signal for its rapid ubiquitin-dependent proteolysis and translocation of free NF-B to the nucleus. IB␣ phosphorylation is catalyzed by the IB kinase (IKK), a complex composed of three subunits, I⌲⌲␣/ IKK1, IKK␤/IKK2, and NEMO/IKK␥/IKKAP1/FIP3. IKK1 and IKK2 are the catalytic subunits, whereas NEMO serves a nonenzymatic, regulatory function. Biochemical and genetic analyses demonstrate that IKK2 is essential for NF-B activation in response to TNF, IL-1, and LPS, whereas IKK1 is not required for such responses (10). Targeted deletion of p65 or IKK2 gene affects TNF-induced transcription of the IB␣ gene (11,12). Recently two additional IKK2-related kinases, IKKi/ IKK⑀ (13,14) and TANK-binding kinase 1/NF-B-activating kinase/T2K (15)(16)(17) have been identified. Whether these latter proteins play a role as enzymatic or non-enzymatic regulatory factors is not fully understood (4).
Regulatory mechanisms for the C/EBP pathways differ markedly from those of NF-B and include transcriptional and/or post-translational mechanisms as well as protein-protein interactions via dimerization through leucine-zipper domains (6,7,18,19). Phosphorylation also regulates the C/EBP family by directing nuclear localization and transcription-activating potential (20 -23). Some members of this family, specifically C/EBP␤ and C/EBP␦, have been linked to gene expression in the acute phase response and during inflammation (24 -26). Furthermore, up-regulation of C/EBP␤ and C/EBP␦ gene expression occurs after exposure to pro-inflammatory stimuli such as TNF, IL-1, IL-6, or LPS (3,5).
Here we have performed experiments to identify possible regulatory links between the NF-B and C/EBP pathways, with an emphasis on events related to innate immune re-* 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.
§ To whom correspondence should be addressed: The Scripps Research Institute, IMM-12, 10550 North Torrey Pines Rd. La Jolla, CA 92037. Tel.: 858-784-8219; Fax: 858-784-8239; E-mail: ulevitch@ scripps.edu. 1 The abbreviations used are: C/EBP, CCAAA/enhancer-binding protein; RANTES, regulated on activation normal T cell expressed and secreted; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; LPS, lipopolysaccharide; TNF, tumor necrosis factor; IL-1, interleukin-1; IKK, IB kinase; MEF, murine embryonic fibroblast; imMEF, immortalized MEF; siRNA, small interfering RNA; EMSA, electrophoretic mobility shift assay; ChIP, chromatin immunoprecipitation; IRF-3, interferon regulatory factor-3; NEMO, NF-B essential modulator. sponses. We have used murine embryonic fibroblasts (MEFs) obtained from various strains of mice bearing targeted gene deletions in components of the NF-B pathway. We show that fibroblasts isolated from p65Ϫ/Ϫ or IKK2Ϫ/Ϫ, but not from control, mice fail to induce C/EBP␤ and C/EBP␦ transcripts in response to cell stimulation with LPS. Furthermore, we observed that both p65 and IKK2 are required for induction of IKKi mRNA. In turn, in primary fibroblasts lacking the IKKi gene LPS induces C/EBP␤ and C/EBP␦ mRNAs and activates NF-B, as observed in control cells, but importantly, fail to induce C/EBP␦-specific DNA binding activity. Furthermore, we report that LPS treatment of IKKiϪ/Ϫ MEFs reveals deficits in expressions of genes associated with immune and pro-inflammatory responses. In totality, these data support the contention that there is a link between the NF-B and C/EBP pathways and that IKKi may be a key element in integration of signals from both pathways during inflammatory and immune responses.
Metabolic Labeling-10 6 cells were plated per 60-mm dish. On the second day, the cells were washed 3 times with phosphate-free Dulbecco's modified Eagle's medium containing 5% of dialyzed fetal bovine serum and then incubated in the same medium containing 400 mCi/ml [ 32 P]H 3 PO 4 for 4 h. The last 2 h some cells were incubated with 100 ng/ml LPS. The cells were then washed three times with cold phosphate-buffered saline and used for preparation of nuclear extract according to EMSA protocol (see above). The nuclear extracts were diluted by the addition of 10 volumes of standard radioimmune precipitation assay buffer, and C/EBP␦ or p65 proteins were recovered by immunoprecipitation with specific antibodies as indicated in Fig. 5C. The immunoprecipitates were analyzed by SDS-PAGE and autoradiography.

RESULTS
We first sought an experimental system to investigate relationships between the C/EBP and NF-B pathways. The promoter region of C3 gene contains C/EBP sites (32); it is known that C3 expression is regulated by C/EBP␦ (26). Despite the fact that there are no identifiable NF-B sites in the C3 promoter, transcription of the C3 gene is induced by NF-B activators including LPS (9, 33). Thus, we reasoned that measure-ments of LPS induction of C3 mRNA would be an appropriate marker for initial studies to investigate possible relationships between the NF-B and C/EBP pathways. Here we have used MEFs from mice bearing targeted deletions of genes encoding IKK2, p65, and IKKi. We first examined induction of C3 mRNA in LPS-treated MEFs derived from IKK2Ϫ/Ϫ and control (IKK2ϩ/ϩ) embryos; the induction of C3 mRNA was observed in control cells but not in the IKK2-deficient cells (Fig. 1A). In contrast, LPS-mediated induction of c-jun mRNA was nearly identical in both types of MEFs. Moreover, as shown here and as previously noted, IKK2Ϫ/Ϫ MEFs showed LPS-induced IB␣ mRNA and NF-B DNA binding activity that was partially reduced when IKK2Ϫ/Ϫ and IKK2ϩ/ϩ cells were compared (Ref. 12; Fig. 1, A and B). Similar findings were noted in experiments with spontaneously imMEFs derived from IKK2Ϫ/Ϫ and control cells (Fig. 1C). Thus, the absence of IKK2 revealed a deficiency in LPS-induced C3 induction that is not likely to result from loss of LPS responsiveness. To further probe the role of the NF-B pathway we also examined the effects of p65 deficiency on LPS induction of C3 and IB␣ mRNA. LPS-mediated induction of both IB␣ and C3 mRNAs was completely abolished in p65Ϫ/Ϫ imMEFs (Fig. 1D). These data suggest that regulation of C3 gene expression requires an intact NF-B pathway. However, we hypothesize that the role of NF-B involves indirect mechanisms and requires additional gene expression under the control of NF-B. In fact it has been shown that protein synthesis is required for LPS-induced expression of C3 mRNA (34).
The observation that LPS induces IKKi mRNA (13) prompted us to evaluate the effects of IKKi deletion on C3 gene expression. Thus, we next measured LPS-induced C3, IB␣, and IKKi mRNA in IKKiϪ/Ϫ and control MEFs as well as in IKK2Ϫ/Ϫ cells ( Fig. 2A). As shown above, LPS-mediated induction of IB␣ mRNA was reduced in cells lacking IKK2, whereas IKKi deficiency had no effect on LPS-induced expression of IB␣ mRNA. In contrast, the induction of C3 mRNA was abolished in the IKKiϪ/Ϫ cells as well as in the IKK2Ϫ/Ϫ cells. Furthermore, we failed to observe induction of IKKi mRNA in the IKK2Ϫ/Ϫ cells. Nearly identical results were also obtained in similar experiments with LPS-stimulated IKKiϪ/Ϫ and IKKiϩ/ϩ imMEFs derived from the corresponding primary isolates of MEFs (Fig. 2B). As an additional specificity control, we also examined the LPS responses in MEFs derived from Egr-1-deficient embryos (28). The LPS-mediated induction of C3 or IKKi mRNA was practically identical in Egr1Ϫ/Ϫ and Egr1ϩ/ϩ cells (Fig. 2C), emphasizing the specificity of the effects observed here with IKK2Ϫ/Ϫ, p65Ϫ/Ϫ, and IKKiϪ/Ϫ MEFs. Thus, IKKi appears to be a key molecule for transcriptional induction of C3 gene in response to LPS, and its expression requires an intact NF-B pathway.
To further define how IKKi participates in gene regulation we performed the following experiments. Extracts from IKK2Ϫ/Ϫ, IKK2ϩ/ϩ, IKKiϪ/Ϫ (as negative control), or IKKiϩ/ϩ (as positive control) were immunoprecipitated to enrich the samples for IKKi, and the resultant immunoprecipitates were electrophoresed and then subjected to Western blot analysis (Fig. 3A). As was expected, expression of IKKi protein was not detected in IKKiϪ/Ϫ cells. In contrast, IKKiϩ/ϩ or IKK2ϩ/ϩ cells showed detectable IKKi protein expression that was increased after LPS addition. Compared with IKK2ϩ/ϩ or IKKiϩ/ϩ cells, the basal level of IKKi protein was significantly reduced in IKK2-deficient cells. Importantly, LPS addition failed to up-regulate IKKi expression in this cell type. These data together with Northern blot studies support the contention that IKK2 is required for the inducible expression of IKKi mRNA and protein. These data prompted us to address the question of how IKKi regulates LPS-induced C3 gene expression. Specifically. we asked whether the failure to induce C3 results from the absence of IKKi protein or whether essential signaling including expression and/or activation of IKK2 is also affected by IKKi deficiency. To address these issues, we used Western blot analysis to compare the levels of IKK2 protein in extracts from IKKiϪ/Ϫ, IKKiϩ/ϩ, IKK2Ϫ/Ϫ (as negative control), or IKK2ϩ/ϩ (as positive control) cells. The results showed that IKKiϪ/Ϫ, IKKiϩ/ϩ, and IKK2ϩ/ϩ cells express nearly identical levels of IKK2 protein (Fig. 3B). Expression of IKK1 and NEMO subunits of the IKK complex was also unchanged in IKKiϪ/Ϫ cells (Fig. 3B). Thus, the absence of IKKi does not reduce the protein expression of other key members of the IKK complex. Furthermore, it is unlikely that IKKi is required for activation of IKK complex because LPS treatment up-regulated the kinase activity of the IKK similarly in IKKiϪ/Ϫ and IKKiϩ/ϩ cells (Fig. 3C). Thus, the absence of IKKi does not alter signaling that is directly related to NF-B activation. The same extracts were subjected in parallel to an immunoprecipitation/kinase assay for IKKi. In addition LPS treatment did not alter the kinase activity of IKKi (Fig. 3C), in keeping with observations of others that pro-inflammatory mediators do not alter IKKi kinase activity but, rather, up-regulate its expression (13).
To investigate whether IKKi deficiency alters the rate and extent of LPS-induced expression of a variety of genes associated with innate immune and inflammatory responses we measured LPS-induced expression of TNF, IL-1, IL-6, IP-10, RANTES, and COX-2 mRNA. LPS treatment increased mRNA expression for each of these genes in IKKiϩ/ϩ cells (Fig. 4A); we also showed that TNF, IL-1, and IL-6 protein levels also increased (Fig. 4, B and C). In contrast IKKi deficiency resulted in a marked reduction of LPS-induced mRNA expression for each of this group of genes (Fig. 4A). Parallel decreases in protein expression for TNF, IL-1, and IL-6 were also noted (Figs. 4, B and C). The fact that both cell lines (IKKiϪ/Ϫ and IKKiϩ/ϩ MEFs) remain LPS-responsive is supported by the observation that identical levels of LPS-induced mRNA for c-jun, C/EBP␤, and C/EBP␦ genes were observed.
Comparative analysis of the promoters of the group of IKKimodulated genes depicted in Fig. 4A identified the presence of binding sites for multiple transcription factors including NF-B, interferon regulatory factor-3 (IRF-3), and C/EBP. Although NF-B is known to be involved in the regulation all of these genes (8,9), a number of observations suggest that the absence of IKKi has no effect on activation of the canonical NF-B pathway. First, we have shown that LPS-mediated activation of IKK complex activity (see Fig. 3C) and translocation of NF-B were indistinguishable when IKKiϩ/ϩ and IKKiϪ/Ϫ MEFs were compared (Fig. 4D). Second, analysis of the subunit composition of the B binding activity with antibodies specific for NF-B-related proteins revealed that treatment of IKKiϪ/Ϫ and IKKiϩ/ϩ cells with LPS induces DNA binding complexes containing essentially the same proteins including p65 (Fig.  4E). Phosphorylation of p65 subunit increases its ability to activate transcription of NF-B target genes (8,35). Although our data do not exclude a role for IKKi in this process, we believe that to be unlikely because LPS-induced expression of IB␣ gene, a classical NF-B target gene (8,11), is not affected by IKKi deficiency (see Fig. 2, A and B). Finally we examined LPS-induced activation of IRF-3 in IKKiϪ/Ϫ MEFs since the promoter region of IP-10 gene contains interferon stimulusresponsive element, a binding site for transcription factor IRF-3 (36,37). The results showed that LPS treatment induces essentially the same levels of interferon stimulus-responsive element binding activity in IKKiϪ/Ϫ and IKKiϩ/ϩ MEFs (Fig.  4F), suggesting that IKKi is not involved in regulation of IRF-3.
Others show that C/EBP␦ is involved in IL-1-induced regulation of the C3 gene (26). Moreover, C/EBP␦ appears to synergize with C/EBP␤, another member of the C/EBP family, in transcriptional regulation of the IL-6 gene (24,25,38). Consistent with previous reports (24,25), we also observed that LPS induces C/EBP␦ and C/EBP␤ mRNA. Here we show that this occurs to an identical extent in both IKKiϪ/Ϫ and IKKiϩ/ϩ MEFs (see Fig. 4A). As both nuclear localization and transcriptional activating potential of C/EBP family members may be regulated by phosphorylation (20,21,23,39), we also investigated whether IKKi-deficiency affects the function of C/EBP␤ and C/EBP␦ proteins in response to LPS. Treatment of MEFs with LPS induces a DNA binding complex containing C/EBP␦ and C/EBP␤ proteins in IKKiϩ/ϩ cells. Surprisingly, IKKi-deficiency resulted in the reduction of C/EBP DNA binding activity induced by LPS and affected the induction of DNA complexes containing specifically the C/EBP␦ but not the C/EBP␤ proteins (Fig. 5A). Western blot analysis of the C/EBP␦ proteins expression demonstrated that extracts from IKKiϪ/Ϫ and IKKiϩ/ϩ cells contains essentially the same levels of C/EBP␦ (Fig. 5B). Metabolic labeling experiments using 32 P

FIG. 3. LPS induces expression of IKKi protein in MEFs.
A, Western blot analysis for IKKi protein expression in IKK2Ϫ/Ϫ, IKK2ϩ/ϩ, IKKiϪ/Ϫ, and IKKiϩ/ϩ MEFs before and after 5 h of treatment with LPS (100 ng/ml). An anti-IKKi immunoprecipitation was performed followed by Western blot analysis with an anti-IKKi. Arrows indicate the position of IKKi and IgG (heavy chain). B, Western blot analysis for IKK2, IKK1, or NEMO protein expression in IKK2Ϫ/Ϫ, IKK2ϩ/ϩ, IKKiϪ/Ϫ, and IKKiϩ/ϩ MEFs before and after 5 h of treatment with LPS (100 ng/ml). C, IKK or IKKi kinase assay (KA) from extracts of IKKiϪ/Ϫ or IKKiϩ/ϩ MEFs treated with LPS for the indicated times. An anti-NEMO or anti-IKKi immunoprecipitation (IP) was performed followed by kinase assay with glutathione S-transferase-IB␣ (1-44) as a substrate. The samples were also subjected to Western blot (WB) analysis for the relevant proteins as indicated.
revealed that IKKi-deficiency results in a significant reduction of a phosphoprotein precipitated by antibodies specific for C/EBP␦. In contrast the absence of IKKi did not effect p65/RelA phosphorylation (Fig. 5C). These results support the contention that IKKi is involved in the regulation of C/EBP␦ activity.
To confirm and extend the findings obtained using MEFs we

FIG. 4. Absence of IKKi affects LPS-mediated induction of genes for immune and inflammatory modulators in MEFs.
A, Northern blot analysis of steady-state mRNA levels in IKKiϪ/Ϫ and IKKiϩ/ϩ MEFs treated with LPS (100 ng/ml) for the indicated times; the specificity of each probe is shown on the left. JE-MCP-1, macrophage chemotactic protein-1. B, Western blot analysis of TNF in culture medium from LPS-treated IKKiϪ/Ϫ and IKKϩ/ϩ MEFs. C, cytokine (IL-1 or IL-6) production by IKKϪ/Ϫ and IKKϩ/ϩ MEFs in response to LPS. Cytokine production was measured by enzyme-linked immunosorbent assay after 6 h of treatment with LPS (1 g/ml). D, EMSA of nuclear extracts from IKKiϩ/ϩ and IKKiϪ/Ϫ MEFs treated with LPS (100 ng/ml) for the indicated times. The DNA binding activity of NF-B or Oct-1 (as a loading control) transcription factor is shown. E, EMSA of a polypeptide composition of the B-specific complexes in IKKiϪ/Ϫ and IKK2ϩ/ϩ MEFs exposed to 100 ng/ml LPS for 1 h. The nuclear extracts were incubated in the presence of specific antibody (Ab) to members of the NF-B/Rel family (c-Rel, p50, and p65) of proteins indicated on top of each lane or with normal rabbit IgG (NR). F, EMSA of nuclear extract from IKKiϪ/Ϫ and IKKiϩ/ϩ MEFs untreated or treated with LPS for 2 h (the nuclear extracts were the same as indicated in panel D). Analysis of the interferon stimulus-responsive element binding activity is shown. also utilized an approach relying on siRNAs targeted against IKKi. In this experiment we evaluated LPS induction of IL-6 gene since this gene is known to be regulated by C/EBP␦ and C/EBP␤ (25,38). Results obtained from siRNA-transfected human umbilical vein endothelial cells demonstrated that specific reduction of IKKi transcripts is accompanied by significant inhibition of IL-6 mRNA induced by LPS (Fig. 5D). As an additional specificity control, we also examined the effects of siRNAs on LPS induction of NF-B regulated gene for IB␣. As expected, LPS induction of IB␣ mRNA was unchanged by the reduction of IKKi expression (Fig. 5D).
Our findings support the contention that IKKi acting through C/EBP␦ links the NF-B and C/EBP pathways. It appears that control of IKKi gene expression plays an important role in this process. Studies with IKK2Ϫ/Ϫ MEFs indicate that activation of NF-B is required for induction of IKKi and C3 mRNA. We further addressed this possibility using p65Ϫ/Ϫ and IKK2Ϫ/Ϫ cells. Northern blot analysis revealed that when compared with p65ϩ/ϩ imMEFs, the p65Ϫ/Ϫ cells fail to induce IKKi, C/EBP␤ and C/EBP␦ mRNAs in response to LPS. In contrast basal levels of these transcripts were normal in p65Ϫ/Ϫ cells (Fig. 6A). Similar experiments using IKK2Ϫ/Ϫ cells also revealed a blockade in induction of C/EBP␤ and C/EBP␦ mRNA (Fig. 6B). Thus, these data support the contention that a functional NF-B pathway is required for regulation of C/EBP␤ and C/EBP␦ genes. The data also raise the question of how IKKi is regulated -primarily through NF-B or through C/EBP or through combined effects of both pathways.
To address questions of IKKi regulation we analyzed the 5Ј-untranslated regions of human and mouse DNA in the vicinity of the IKKi gene. Interestingly our analysis did not reveal the presence of NF-B sites but rather indicated that both the murine and human genes have identical sequences containing a C/EBP-like DNA binding site (Fig. 6C). By using electrophoretic gel shift assay carried out on a nuclear extract prepared from LPS-treated cells we found that this sequence has the C/EBP-specific binding activity in vitro (Fig. 6D). The interaction of C/EBP with IKKi promoter was also demonstrated via chromatin immunoprecipitation (ChIP) assay. The results of ChIP assay performed on chromatin samples from untreated or LPS-treated imMEFs confirmed that both C/EBP␤ and C/EBP␦ are able to bind with the promoter region of IKKi gene in vivo (Fig. 6E). Similar results were obtained on chromatin samples from TNF-treated cells (Fig. 6E). The binding of p65 subunit of NF-B with the IB␣ promoter was used as positive control, whereas the chromatin samples from IKK2Ϫ/Ϫ imMEFs were tested as negative control for C/EBPspecific binding with IKKi promoter. Therefore our data provide support for the contention that up-regulation of the IKKi promoter by pro-inflammatory mediators such as LPS and TNF is likely to involve members of the C/EBP family.

DISCUSSION
The members of the C/EBP and NF-B families are known to regulate overlapping cellular processes, including the control of inflammation and immunity (2,3,5,8). C/EBP and NF-B DNA binding motifs have been identified in the promoter regions of genes encoding immune and pro-inflammatory modulators, including IL-6 (24, 40), IL-1␤ (41), IP-10 (36), TNF (42), RANTES (43) and COX-2 (44). Previous studies have also shown that the p50 subunit of NF-B associates with the C/EBP␤/NF-IL6, suggesting cross-talk between NF-B and C/EBP in regulation of immune and acute-phase responses (45). The nature of the mechanisms responsible for coordination of C/EBP and NF-B pathways has been unclear.
Here we provide experimental data to support the contention that the IKK2-related kinase IKKi is a key molecule for functional coordination between these two pathways. The IKKidirected functional link between NF-B and C/EBP seems to be a "nonlinear" mechanism that includes a NF-B-dependent regulation of C/EBP␤ and C/EBP␦ gene transcription and IKKi-mediated activation of C/EBP-specific DNA binding activity. Consistent with this mechanism, our experiments carried out on MEFs derived from mice with targeted deletion of IKKi gene showed that IKKi is required for LPS induction of genes encoding IL-1␤, IL-6, IP-10, TNF, RANTES and COX-2. Further support for the role of IKKi was obtained in the experiments showing that siRNA-directed reduction of IKKi expression in human umbilical vein endothelial cells results in significant inhibition of LPS-mediated up-regulation of the C/EBP-NF-B-regulated target gene for IL-6.
Previous work by others demonstrated that the promoter of the C3 gene contains C/EBP binding sites (32) and is transactivated by C/EBP␦ (26). Our findings extend these observations by showing that IKKi is required for LPS-induced activation of a C/EBP-specific DNA binding complex containing C/EBP␦ and driving expression of C3. Moreover, we found that IKKi-deficiency results in a significant reduction of LPS-induced phosphorylation of a protein precipitated by antibodies specific for C/EBP␦. Although further work is needed to characterize the nature of this phosphoprotein, these results are consistent with the assumption that IKKi is involved in the regulation of C/EBP␦ activity induced by LPS. In contrast to the effect on C/EBP we did not find evidence for linkage of IKKi deficiency to other pathways initiated by LPS.
Previous studies have demonstrated that activators of the NF-B pathway, such as LPS, TNF and IL-1 (9), induce upregulation of C/EBP␤, C/EBP␦ and IKKi mRNA (24,25,13). However the exact sequence of events in this process has not been defined. Here we show that LPS-mediated activation of the genes for both C/EBP␤ and C/EBP␦ requires intact NF-B pathway, since neither gene is induced by LPS in IKK2-or p65-deficient MEFs. Further examination of the C/EBP␤ and C/EBP␦ promoters revealed the presence of B-like sites, suggesting that NF-B directly involves in the LPS-mediated regulation of C/EBP␤ and C/EBP␦ genes. However additional studies are needed to more completely understand the role of these sites. In contrast, the analysis of the promoter sequences in both human and mouse IKKi genes failed to identify a B-like site. Surprisingly we also noted that IKKi promoter contains a C/EBP-like site with identical sequences in human and mouse genes. Our analysis revealed that both ␤ and ␦ FIG. 6. IKKi is a key molecule coupling NF-B and C/EBP pathways. A, Northern blot analysis of IKKi, C/EBP␤, C/EBP␦, and GAPDH steady-state mRNA levels in p65Ϫ/Ϫ and p65ϩ/ϩ imMEFs treated with LPS (100 ng/ml) for the indicated times. B, Northern blot analysis of C/EBP␤, C/EBP␦, and GAPDH steady-state mRNA levels in IKK2Ϫ/Ϫ and IKK2ϩ/ϩ imMEFs treated with LPS (100 ng/ml) for the indicated times. C, the 36-nucleotide sequence of mouse or human (shown in bold) chromosome 1 identified 480 bp upstream of the translation initiation site of IKKi gene. A sequence of C/EBP-like binding site is underlined. D, EMSA of nuclear extract from normal imMEFs treated with LPS for 2 h. The nuclear extracts were incubated with a 32 P-labeled 36-bp DNA fragment (see a sequence of the upper strand on the panel C) in the absence or in the presence of competitive (Comp) unlabeled oligonucleotide (2 pmol) containing the wild type 36-bp fragment (W), the C/EBP consensus sequence (C), or the wild type 36-bp fragment containing the mutation of a C/EBP binding motif (M). Ab, antibody. In addition, some samples were incubated in the presence of normal rabbit IgG (N) or in the presence of specific antibodies (2 g per reaction) against C/EBP␣, -␤, or -␦ as indicated. E, ChIP assays were carried out on chromatin samples from IKK2Ϫ/Ϫ (negative control), IKKiϩ/ϩ, and IKKiϪ/Ϫ (positive control) imMEFs untreated or treated with LPS or TNF for 2 h. The chromatin was immunoprecipitated with antibodies to C/EBP (␤ or ␦) or p65 as an additional control. Shown is an IKKi or an IB␣ (positive control) promoter fragment amplified by PCR from the ChIP samples. subunits of C/EBP interact with an IKKi promoter fragment containing this sequence. This observation thus supports the conclusion that IKKi gene expression is regulated by C/EBP.
After the submission of the present work, Sharma et al. (46) and Fitzgerald et al. (47) reported on experiments with human cell lines showing a link between TANK-binding kinase 1 and IKKi/IKK⑀ and phosphorylation of endogenous IRF-3 (46) and reporter gene expression regulated by interferon regulatory factor-3 (IFR-3) and IRF-7 in response to virus infection. Here we have not examined effect of IKKi-deficiency on virus induced responses in MEFs so it is difficult to compare our studies with those recently published (46,47). However we believe each report provides new evidence linking IKKi to innate immune responses to both viral and bacterial infection.
Here we showed that LPS-mediated activation of IRF-3 DNAbinding activity was normal in IKKiϪ/Ϫ cells, whereas the activation of IRF family members target genes for IP-10 and RANTES was blocked in IKKi-deficient cells. This result may reflect differences in signaling via TLR4-and TLR3-dependent pathways. IKKi-deficiency also resulted in the reduction of LPS-mediated expression of TNF, IL-6, IL-1 and COX-2 genes; in none of these cases does the promoter appear to contain IRF sites. In contrast a common characteristic of the promoters for this group of IKKi-affected genes, including IP-10 and RANTES, is the presence of binding sites for C/EBP and NF-B. Here we provide clear evidence that IKKi is not involved in the activation of NF-B but significantly affects the activation of C/EBP. It has been shown that similar to IRF-3 and IRF-7, C/EBP is also involved in the regulation of viral oncogenesis and infection (48). Further it was shown that VSV replication was significantly inhibited by a kinase-dead variant of IKKi in 293 cells over-expressing IRF-3 (46). On the other hand, interferon-mediated antiviral response against VSV is normal in IRF-3Ϫ/Ϫ MEFs (49). Thus all of the most recent data suggest that IKKi is involved in the innate immune responses to viral and bacterial infection. This may well occur via its effects on multiple transcription factors including members of NF-B, IRF and C/EBP families.
In summary, our findings identify the function of IKKi as a signaling molecule required for LPS-inducible transcriptional regulation of genes for inflammatory and immune mediators in fibroblasts and endothelial cells, suggesting an important role for IKKi in innate immune responses of cells of non-hematopoietic origin. Studies under way with the IKKiϪ/Ϫ mice will provide more direct information about these pathways in vivo. It follows from our data that modulation of IKKi function may have implications for the development of the anti-inflammatory drugs. Interestingly, IKKi exists as a high molecular weight complex (14). We are currently performing studies to identify and characterize IKKi-associated proteins that may provide, in addition to IKKi, molecular targets for development of anti-inflammatory therapies. The recent observation that diminished local synthesis of C3 was beneficial for the transplanted kidney (50) suggests that inhibition of C3 expression by modulation of IKKi function could have therapeutic value in tissue transplantation.