Stimulus-specific Induction of a Novel Nuclear Factor-κB Regulator, IκB-ζ, via Toll/Interleukin-1 Receptor Is Mediated by mRNA Stabilization*

We have recently identified an inducible nuclear factor-κB (NF-κB) regulator, IκB-ζ, which is induced by microbial ligands for Toll-like receptors such as lipopolysaccharide and the proinflammatory cytokine interleukin (IL)-1β but not by tumor necrosis factor (TNF)-α. In the present study, we examined mechanisms for stimulus-specific induction of IκB-ζ. The analysis of the IκB-ζ promoter revealed an essential role for an NF-κB binding sequence in transcriptional activation. The activation, however, did not account for the Toll-like receptor/IL-1 receptor-specific induction of IκB-ζ, because the promoter analysis and nuclear run-on analysis indicated that its transcription was similarly induced by TNF-α. To examine post-transcriptional regulation, we analyzed the decay of IκB-ζ mRNA, and we found that it was specifically stabilized by lipopolysaccharide or IL-1β but not by TNF-α. Furthermore, we found that costimulation with TNF-α and another proinflammatory cytokine, IL-17, elicited the IκB-ζ induction. Stimulation with IL-17 alone did not induce IκB-ζ but stabilized its mRNA. Therefore, IκB-ζ induction requires both NF-κB activation and stimulus-specific stabilization of its mRNA. Because IκB-ζ is essential for expression of a subset of NF-κB target genes, the stimulus-specific induction of IκB-ζ may be of great significance in regulation of inflammatory reactions.

Cells reprogram gene expression upon environmental changes to maintain homeostasis. For multicellular organisms, serious environmental changes include microbial infection, and a major rearrangement of gene expression occurs during inflammatory responses against infection. The expression of a variety of genes for inflammatory mediators is strongly activated upon stimulation with microbial and viral products that stimulate toll-like receptors (TLRs), 1 and with inflammatory cytokines such as interleukin (IL)-1␤ or tumor necrosis factor (TNF)-␣. Location, degree, and duration of expression of these genes are strictly regulated at the level of transcription, processing and stabilization of mRNA, and/or protein synthesis and degradation. Loss of the regulation could lead to severe disorders exemplified by systemic inflammatory response syndrome, rheumatoid arthritis, or sclerosis.
Nuclear factor-B (NF-B) is an evolutionarily conserved transcription factor that plays pivotal roles in inflammation as well as cell growth, survival, and differentiation (1)(2)(3)(4)(5). It consists of homo-and heterodimers of Rel/NF-B family proteins with the Rel homology domain, which is implicated in both DNA binding and dimer formation. In resting cells, NF-B is sequestered in the cytoplasm via association with cytosolic IB proteins, IB-␣, -␤, and -⑀, through interaction between the ankyrin repeats of the IB proteins and the Rel domains of NF-B. Upon stimulation, the cytosolic IB proteins are phosphorylated followed by ubiquitin/proteasome-mediated degradation, and NF-B translocates into the nucleus where it engages in active transcription of the target genes. One of the NF-B target genes is IB-␣, and the re-synthesized IB-␣ binds to the nuclear NF-B to export it to the cytoplasm and terminate the sequence of the reactions (1,2,4,6). The activity of NF-B is also modulated in the nucleus. Not only phosphorylation and acetylation of NF-B itself (7,8) but interactions with other nuclear proteins play critical roles in the regulation of the transcriptional activity of NF-B. Modulators of the nuclear NF-B include histone deacetylases (9), transcriptional coactivator and corepressor proteins (10), and other transcription factors (11)(12)(13).
Recently, we and others (14 -16) have identified a new nuclear IB protein, IB-(also named as MAIL or INAP), which interacts with NF-B via a C-terminal ankyrin repeat domain in the nucleus (14). IB-is barely able to be detected in resting cells but is robustly induced in response to lipopolysaccharide (LPS). The induced IB-localizes in the nucleus and preferentially interacts with the NF-B p50 subunit rather than the p65 subunit. The initial characterization of IB-has shown that it negatively regulates NF-B activity, because transfection of IB-inhibits NF-B reporter activity stimulated with LPS, IL-1␤, or TNF-␣, and inhibits activity induced by transfection of the NF-B p65 subunit. On the other hand, subsequent functional analyses of the N-terminal region of IBrevealed that it had a latent activity of transcriptional activation. 2 In addition, the most recent studies using IB--deficient mice demonstrated that IB-is essential for the induction of a subset of the inflammatory genes including IL-6, the IL-12 p40 subunit, and granulocyte-macrophage colony-stimulating factor (17). The transcriptional activity of IB-depends on the p50 subunit, which exhibits target DNA binding activity without the transactivation activity (17).
Detailed analyses on the induction of IB-showed that not only LPS but also other microbial TLR ligands including peptidoglycan, the bacterial lipopeptides, and CpG DNA elicited IB- (17,18). In addition to these microbial products, the proinflammatory cytokine IL-1␤ strongly induced IB- (14). TLRs and IL-1 receptor (IL-1R) share similar cytoplasmic domains known as Toll/IL-1R (TIR) domain and the common adaptor molecule MyD88 (19,20). As expected, the induction of IB-by the TLR ligands is dependent on MyD88 (17). Prominent cellular responses triggered by activation of TLR/IL-1R include activation of NF-B and mitogen-activated protein (MAP) kinases, and our previous studies have revealed that NF-B is essential for IB-induction (18). Another proinflammatory cytokine TNF-␣ also activates both NF-B and MAP kinases, which is similar to TLR ligands or IL-1␤. However, IB-induction by TNF-␣ was marginal, indicating that NF-B activation is not sufficient for the induction of IB- (14). Furthermore, overexpression of the NF-B subunits resulted in robust induction of IB-␣ but not IB-, leading to the conclusion that an additional signal derived from the TIR domain is required for IB-induction (18).
Although IL-1␤ and TNF-␣ do not share sequence homology with each other and they bind to distinct receptors, the intracellular signaling activated by the two cytokines converges to similar adaptor molecules, TNF receptor-associated factor (TRAF) 2 and TRAF6, each of which leads to the activation of NF-B and MAP kinases (1,21). Consequently, many of the target genes for IL-1␤ and TNF-␣ overlap, and the two proinflammatory cytokines give rise to similar biological effects exemplified by inflammation, coagulation, or pyrexia.
The biological effects, however, are not identical between the two cytokines. In addition to IB-, several other genes are preferentially induced by IL-1␤ but not by TNF-␣. These genes are neutrophil gelatinase-associated lipocalin (NGAL) (22), the growth-related oncogene (GRO)␣ homolog/the KC chemokine (23,24), the extracellular metalloprotease MMP-3 (25), and IL-6 (26 -28). Conversely, expression of complement factor H was up-regulated by TNF-␣ but not by IL-1␤ (28). Promoter analysis of NGAL showed that it is preferentially activated by IL-1␤ rather than by TNF-␣, implying that the differential induction is determined at the transcriptional level. Therefore, an IL-1␤-specific transcriptional activator(s) or a TNF-␣-specific repressor(s) exists (22). In fact, IB-has been shown to be one of the essential factors for expression of a subset of inflammatory genes such as IL-6 and NGAL (17). The NF-B p50 subunit is required for the function of IB-, and only IB-is preferentially induced upon stimulation with IL-1␤ and TLR ligands but not with TNF-␣. Therefore, IB-might play a role as a master gene that determines the expression of a subset of the inflammatory genes.
In the present study, we analyzed molecular mechanisms for the differential induction of IB-. We found that stability of IB-mRNA was specifically up-regulated by stimulation with LPS or IL-1␤ but not with TNF-␣. This finding indicates that in addition to NF-B activation, the other signal required for IBinduction, which comes from the TIR domain but not from the TNF-␣ receptor, leads to mRNA stabilization. Given the essential function of IB-in the expression of a series of inflammatory genes, the control of IBexpression by different stimuli may play a key role in determining the nature of inflammation.

EXPERIMENTAL PROCEDURES
Cell Culture, Reagents, and Antibodies-NIH3T3, RAW264.7, and A549 cells were cultured in Dulbecco's modified Eagle's medium supplemented with heat-inactivated fetal calf serum, penicillin, and streptomycin. THP-1 cells were in RPMI 1640 with heat-inactivated fetal calf serum, penicillin, and streptomycin. LPS from Escherichia coli 0111:B4 was purchased from List Biological Laboratories Inc. (Campbell, CA). IL-1␤, IL-17, and TNF-␣ were from Genzyme Techne Corp. (Framingham, MA). Actinomycin D and G418 were from Nakalai Tesque (Kyoto, Japan). Anti-IB-polyclonal antibody against the C-terminal region (anti-IB-(C)) or IB-(L)-specific region (anti-IB-(LN)) was raised as follows. A cDNA fragment for the C-terminal region (amino acids 414-728) or the N-terminal region (amino acids 1-99) of mouse IB-was subcloned into pQE-30 (Qiagen, Chatsworth, CA), pGEX-2T (Amersham Biosciences), or pMALg (14). Recombinant proteins were expressed in E. coli as His 6 -tagged, glutathione S-transferase (GST)-or maltose-binding protein-fusion proteins. The His 6tagged recombinant protein of the C-terminal region was immunized to a rabbit, and the antiserum that was obtained was subjected to affinity purification using Affi-Gel-10 resin (Bio-Rad) coupled with the corresponding recombinant GST-fusion protein. The anti-IB-(L)-specific antiserum was raised against the GST-fusion protein of the N-terminal region and was affinity-purified with the maltose-binding protein-fusion protein. Antibodies against IB-␣, ␤-tubulin, and p38 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and anti-active p38 antibody was from Promega Corp. (Madison, WI).
Plasmids-The pELAM1-Luc reporter plasmid was described previously (29), and pNF-B-Luc that contained four tandem B sequences was purchased from Clontech. Genomic fragments of mouse or human IB-were obtained by screening a genomic library or by PCR with genomic DNA and subcloned into pGL3-basic vector (Promega). cDNAs for mouse MyD88, IB-(L), and IB-(S) were obtained by reverse transcriptase-PCR of RNA from RAW264.7 cells and subcloned into pcDNA3 (Invitrogen). A dominant-negative form of MyD88 (amino acids 155 to the C terminus) was created by PCR. An expression plasmid for the full-length IB-mRNA driven by chicken ␤-actin promoter (pAG-mIB-) was constructed by subcloning the following fragment in pBluescript SKII(ϩ) (Stratagene, La Jolla, CA). The subcloned fragment contained the chicken ␤-actin promoter and the rabbit ␤-globin splicing sequence derived from pCNX2 (30) followed by full-length mouse IB-(L) mRNA (ϩ27 to ϩ4005) including G/U-rich and U-rich sequences beyond the polyadenylation site, which is recognized by the 3Ј end-processing complex. The expression plasmids for chimeric mRNAs (pAG--Luc-, pAG---Luc-, and pAG--Luc--) were constructed by replacing various parts of pAG-mIB-with the open reading frame (ORF) of firefly luciferase derived from pGL3-basic. The parts of pAG-mIB-that were replaced were the entire ORF, the C-terminal region from an internal NcoI site, or the N-terminal region to the NcoI site of the ORF of IB-. All sequences amplified by PCR were confirmed by DNA sequencing.
Stable Cell Line-NIH3T3 cells were transfected with pAG-mIBand pEGFP-N1 (Clontech) at a ratio of 10:1. The cells resistant to G418 were selected in the presence of 0.5 mg/ml G418 and cloned. The expression of IB-mRNA was determined by Northern blot analysis.
Luciferase Reporter Assay-Cells were transfected with a luciferase reporter plasmid and the internal control plasmid pRL-TK (Promega) by using FuGENE 6 (Roche Diagnostics) or Effectene (Qiagen). Two days after transfection, the cells were stimulated at 37°C for 6 h. Luciferase activities were measured by the Dual-Luciferase Reporter assay system (Promega). The transfection efficiency was normalized by the Renilla luciferase activity derived from pRL-TK.
Nuclear Run-on Assay-The nuclear run-on assay was performed as described previously (23). Briefly, isolated nuclei from stimulated cells were incubated in the presence of [␣-32 P]rUTP with rATP, rCTP, and rGTP at 30°C for 30 min, and the synthesized transcript was hybridized with a nitrocellulose membrane (Protran, Schleicher & Schuell) that had been blotted with cDNA for a target gene at 65°C overnight. The membrane was washed with 30 mM NaCl and 3 mM sodium citrate (0.2ϫ SSC) that contained 0.1% sodium lauryl sulfate, was exposed to an imaging plate (Fuji Photo Film, Osaka, Japan), and was analyzed by STORM 860 PhosphorImager (Amersham Biosciences).

IB-Is Specifically Induced by Stimulation of TLR/IL-1R-
Our previous studies have demonstrated the essential requirement of an unidentified signal(s) for IB-induction that comes from the TIR domain, in addition to NF-B activation (18). However, an alternative possibility to explain the inability of TNF-␣ to induce IB-could be that a negative signal was specifically generated by TNF-␣ stimulation. In order to examine this possibility, we treated cells with LPS or IL-1␤ together with TNF-␣. LPS-or IL-1␤-stimulation of NIH3T3 cells resulted in the strong induction of IB-. Although TNF-␣ similarly induced other NF-B target genes such as IB-␣ and A20, TNF-␣-stimulation of the cells induced little IB-, as described previously (Fig. 1A) (14). Costimulation with TNF-␣ did not affect the LPS-or IL-1␤-mediated induction of IB- (Fig. 1A), thus ruling out the possibility of a negative signal by TNF-␣. The expression of IB-was dependent on de novo transcription because the induction of IB-by LPS or IL-1␤ was completely abolished by treating the cells with actinomycin D, an inhibitor of transcription. In the current conditions, the stimulation of NIH3T3 cells with LPS, IL-1␤, and TNF-␣ activated NF-B to a similar extent when measured with an NF-B reporter, pELAM1-Luc (29, 31) (Fig. 1B). NF-B activation by LPS or IL-1␤, but not that by TNF-␣, was specifically inhibited by expression of a dominant-negative mutant of MyD88, a TLR/ IL-1R-specific adaptor molecule (19,20). The stimulus-specific induction of IB-was also observed in A549 cells, which are a well characterized human alveolar type II epithelial cell line (Fig. 1C). In addition, the stimulus-specific induction of IBwas observed in HEK293 or THP-1 cells (data not shown). This indicated that the preferential induction of IB-by activation of TLR/IL-1R is not cell type-specific and is conserved between mouse and human.
Induction of endogenous IB-protein was examined by immunoblotting with an antibody raised against a recombinant protein containing the C-terminal region of IB-(anti-IB-(C)). An 85-kDa band that reacted with the antibody appeared 40 min after stimulation in LPS-treated RAW264.7 cells ( Fig.  2A) and in NIH3T3 cells stimulated with LPS or IL-1␤ (Fig.  2B). A faint band of 70-kDa also appeared at ϳ60 min after the stimulation. In contrast to the cytosolic IB proteins including IB-␣, -␤, or -⑀, IB-was barely detected in unstimulated cells, and rapid degradation after the stimulation was not observed. The IB-protein level peaked at 1-2 h after stimulation ( Fig.  2A) and then gradually decreased (data not shown). Consistent with the mRNA induction, IB-protein was strongly induced by LPS or IL-1␤, but we observed minimum induction by TNF-␣ (Fig. 2B). Immunoblotting with the anti-active p38 an-FIG. 1. IB-mRNA is induced in response to LPS or IL-1␤ but not to TNF-␣. A, NIH3T3 cells were stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for 1 h in the presence or absence of actinomycin D (5 g/ml). Total RNA was extracted and subjected to Northern blot analyses with a probe for IB-, IB-␣, A20, or G3PDH. B, NIH3T3 cells were transfected with the pELAM1-Luc reporter plasmid and increasing amounts (0, 0.1, or 1.0 g) of an expression vector for a dominant-negative form of MyD88 (MyD88C) together with pRL-TK. Two days after transfection, cells were stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for 6 h, and the luciferase activities were measured. C, total RNA was extracted from A549 cells stimulated with IL-1␤ (50 pg/ml) or TNF-␣ (20 ng/ml) for 1 h and subjected to Northern blot analyses with the indicated probes.

FIG. 2. IB-protein is induced in response to LPS or IL-1␤ but not to TNF-␣.
A, RAW264.7 cells were stimulated with LPS (100 ng/ml) for the indicated periods. Total cell lysate was prepared and subjected to immunoblot analyses using antibodies against the C-terminal region of mouse IB-, active p38 (p-p38), and p38. B, NIH3T3 cells were stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for the indicated periods, and total cell lysate was subjected to immunoblot analyses as described for A. C, total cell lysates from NIH3T3 cells transfected with an empty vector (vec), an expression vector for mouse IB-(S) or IB-(L), and from RAW264.7 cells stimulated with LPS (100 ng/ml) for the indicated periods were subjected to immunoblot analyses using antibodies against the C-terminal region of IB-(anti-IB-(C)) or the N-terminal region of IB-(L) (anti-IB-(LN)). The 70-kDa band corresponding to IB-(S) is indicated by an asterisk. A dot shows a nonspecific band. tibody showed that p38 MAP kinase was activated by TNF-␣, LPS, or IL-1␤. Therefore, the preferential expression of IBprotein was primarily regulated by the level of the mRNA.
Two types of mRNAs for IB-have been identified, which are most likely generated by an alternative splicing (14 -16, 32). The shorter form, IB-(S), lacks the exon 3 harboring the initiation codon for the longer form, IB-(L), thus coding for a shorter protein beginning at a downstream initiation codon without the N-terminal 99 amino acids of IB-(L). The transfected IB-(L) exhibited a mobility similar to that of the main 85-kDa band (Fig. 2C). Immunoreactivity of the 85-kDa band with an IB-(L)-specific antibody (anti-IB-(LN)), raised against the N-terminal IB-(L)-specific sequence, identified this band as IB-(L) (Fig. 2C). The 70-kDa band reacted with the anti-IB-(C) antibody but not with the IB-(L)-specific antibody, suggesting that it was derived from IB-(S). Transfected IB-(S) also showed a mobility similar to that of the 70-kDa band, further supporting its identity as IB-(S). Thus, the dominant product of the IB-gene upon stimulation with LPS or IL-1␤ was IB-(L). The 85-kDa IB-(L) band shifted slightly upward and appeared as a doublet band after stimulation, suggesting covalent modifications. Resistance to phosphatase treatment indicated that the modification was not phosphorylation (data not shown). However, the identity of the modification is unknown at present.
The Preferential Induction of IB-Determines the Expression of the Target Gene-The analyses of IB--deficient mice have revealed that IB-is essential for the expression of a subset of NF-B target genes (17). Therefore, stimulus-specific induction of IB-should be reflected by the subsequent expression of these genes. NIH3T3 cells were stimulated under the same conditions as shown in Fig. 1. We then examined the induction of IL-6, a representative gene that requires IB-for expression. The expression of IL-6 was more strongly induced by LPS or IL-1␤ than by TNF-␣, similar to that observed for IB- (Fig. 3A). The induction pattern was similar between IB-and IL-6, but the kinetics of the induction were different. IL-6 mRNA peaked at 2 h after stimulation with LPS or IL-1␤ in NIH3T3 cells, but it reached a peak at 4 h after stimulation in LPS-stimulated RAW264.7 cells (Fig. 3B). In either case, IB-was induced more rapidly than IL-6 and peaked at 1 h after the stimulation. These observations are consistent with the concept that IB-controls the expression of IL-6.
NF-B Sites in the Promoter Are Essential for the Transcriptional Up-regulation of the IB-Gene-We obtained genomic fragments corresponding to mouse and human IB-promoter regions to analyze the transcriptional regulation of the IBgene. Rapid amplification of cDNA ends analyses using poly(A)-tailed RNA from RAW264.7 cells or THP-1 cells stimulated with LPS identified several possible transcription initiation sites. The main initiation site is indicated by a box labeled as ϩ1 in Fig. 4A. The nucleotide sequences around the transcription initiation site were well conserved between mouse and human and were preceded by a potential TATA box. The reporter analysis in RAW264.7 cells with serially deleted fragments of the mouse promoter region showed that reporters with promoter fragments containing a region spanning Ϫ359 bp to the initiation site were activated upon LPS stimulation. The minimal IB-promoter responded more strongly to LPS stimulation than an artificial promoter with four tandem copies of the NF-B consensus sequence (4ϫ B). Thus, the ciselement(s) responsible for LPS responsiveness is present within Ϫ359 bp from the transcription initiation site (Fig. 4B).
We identified three canonical B sites in the proximal promoter region of the mouse IB-promoter, and we defined them as B1 (Ϫ256 to Ϫ247), B2 (Ϫ218 to Ϫ209), and B3 (Ϫ83 to Ϫ74). Although the B1 and B2 sites were completely conserved between the mouse and human IB-promoters, the B3 sequence was poorly matched with the B consensus sequence and was not conserved between the two species. We introduced a mutation into each of the sites to examine the contributions of these B sites to the induction of IB-. The mutation at the B2 site completely abolished the LPS-mediated promoter activation, whereas the mutation at the B1 site resulted in moderate activation (Fig. 4B). In contrast, the mutation at the B3 site did not affect the promoter activity.
In Addition to LPS and IL-1␤, TNF-␣ Also Activates the Transcription of IB--We investigated whether the LPS or IL-1␤-specific induction of IB-was also observed by promoter analysis. NIH3T3 cells transfected with various reporters were stimulated as shown in Figs. 1-3. LPS, TNF-␣, and IL-1␤ similarly activated NF-B, p38 MAP kinase, and the induction of the typical NF-B target genes. Stimulation with LPS or IL-1␤ elicited modest activation of the IB-promoter (Fig. 5A). Unexpectedly, TNF-␣ induced similar or even stronger activation of the IB-promoter when compared with LPS and IL-1␤. We examined the upstream regions up to ϳ11 kb from the transcription initiation site but did not find any elements that conferred further responsiveness or LPS/IL-1␤-specific up-regulation of the promoter activity (data not show). The mutation at the B2 site, but not that at the B1 or B3 sites, abolished the activation by all three stimuli, confirming the critical contribution of NF-B in the induction of IB-.
We also tested the IB-promoter activity in human alveolar A549 cells. By titrating IL-1␤ and TNF-␣, we determined the concentration of each stimulus that resulted in similar levels of activation of the NF-B reporter. Stimulation of A549 cells with the determined concentrations of IL-1␤ and TNF-␣ exhibited an even higher activation of the IB-promoters by TNF-␣ than by IL-1␤, regardless of length and species of the promoters (Fig. 5B). On the other hand, a promoter of human NGAL selectively responded to IL-␤ as reported by Cowland et al. (22).
We performed the nuclear run-on analyses to directly measure the transcriptional activity of the IBgene in the stimulated cells. LPS, IL-␤, and TNF-␣ elicited transcriptional activation of the IBgene as well as the other NF-B target FIG. 3. IL-6 mRNA is strongly induced in response to LPS or IL-1␤ but not to TNF-␣. A, NIH3T3 cells were stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for the indicated periods. Total RNA was extracted and subjected to Northern blot analyses with a probe for IB-, IL-6, IB-␣, or G3PDH. B, total RNA from RAW264.7 cells stimulated with LPS (100 ng/ml) for the indicated periods was subjected to Northern blot analyses with the indicated probes.
genes, A20 and IB-␣. Except for the strongest activity observed 35 min after LPS stimulation, the transcriptional activity of the IB-gene was more strongly induced by TNF-␣ than by IL-1␤ or LPS, as observed in the promoter analyses (Fig. 6,  A and B). The transient but strong transcriptional activation in the LPS-stimulated cells appeared to be commonly found with A20 and IB-␣, which are equally induced by the three stimuli and in other cell types (Fig. 6, C and D). Thus, we concluded that the stimulus specificity of IB-induction is not determined at the transcriptional level.
LPS or IL-1␤, but Not TNF-␣, Specifically Stabilizes IB-mRNA-We investigated the stability of IB-mRNA in the presence of given stimuli in order to explore the post-transcriptional regulation of IB-expression. Because IB-mRNA can barely be detected in unstimulated cells, we constructed an expression plasmid to express the full-length IB-mRNA without stimulation. In the plasmid pAG-mIB-, a DNA fragment coding for IB-mRNA from ϩ27 to ϩ4005 was located downstream of a chicken ␤-actin promoter, which is a strong promoter unresponsive to the stimuli. A stable cell line established with the plasmid constitutively expressed the full-length mRNA for IB- (Fig. 7A). The kinetics of the decay of the mRNA derived from the IB-transgene and the endogenous mRNA for IB-␣ were analyzed after treating the cells with actinomycin D to terminate de novo transcription (Fig. 7, B and C). In the absence of stimuli, IB-mRNA was degraded with a half-life of about 30 min, in a similar manner to IB-␣. This is in contrast to the stable mRNA for glyceraldehyde-3-phosphate dehydrogenase (G3PDH). Most interestingly, the degradation of the IB-mRNA was specifically delayed by stimulation with LPS or IL-1␤ but not with TNF-␣. LPS or IL-1␤ extended the half-life of the IB-mRNA more than three times than that without stimuli. In contrast to IB-mRNA, the half-life of IB-␣ mRNA was not affected by the stimuli. These results indicated that IB-mRNA is specifically stabilized upon LPS or IL-1␤ stimulation.
A cis-Element in IB-mRNA Is Responsible for the Stimulus-specific Stabilization-To identify a cis-element(s) for the stimulus-specific stabilization of IB-mRNA, we replaced the ORF of IB-in the expression vector with that of the luciferase gene (Fig. 8A, -Luc-). NIH3T3 cells were transfected with the plasmid, and the decay of the chimeric mRNA was chased following treatment with actinomycin D. The stimulus-specific stabilization of mRNA was not observed with the chimeric mRNA (Fig. 8B, upper panel), suggesting that replacement of the ORF resulted in deletion of the cis-element(s) responsible for the stabilization. We attempted to identify the cis-element(s) within the ORF of IB-. The ORF was divided by an internal NcoI restriction site into a 765-bp fragment coding for the N-terminal (IB-(N)) and a 1,421-bp fragment for the C-terminal (IB-(C)) regions. Each of the fragments was inserted upstream or downstream of the luciferase ORF in the -Lucconstructs (Fig. 8A). The resulting chimeric mRNA --Lucharboring the 765-bp N-terminal region was specifically stabilized by LPS and IL-1␤, whereas that with the Cterminal region, -Luc--, did not respond to these stimuli (Fig.  7B, the middle and the bottom panels). We tried to further determine the cis-element(s), but all the constructs with the N-terminal region shorter than 765 bp failed to show the stimulus-specific stabilization by LPS/IL-1␤ (data not shown). Thus, the cis-element responsible for the stimulus-specific stabilization is in the 765-bp N-terminal region of the ORF of IB-.
IL-17 Generates an mRNA Stabilization Signal to Induce IB-in Combination with TNF-␣-The proinflammatory cytokine IL-17 has been shown to augment the induction of a subset of inflammatory genes by TNF-␣. Most intriguingly, several groups (33)(34)(35) have recently reported that costimulation with TNF-␣ and IL-17 up-regulates expression of IL-6 and the GRO␣ homolog. In order to examine the effect of IL-17 on IB-induction, NIH3T3 cells were stimulated with IL-17 alone or in combination with TNF-␣ or IL-1␤. Although marginal expression of IB-mRNA was observed in the cells stimulated with IL-17 alone, costimulation with TNF-␣ and IL-17 elicited robust IB-induction comparable with that by LPS or IL-1␤ (Fig. 9A). On the other hand, the induction of IB-␣ was not affected by stimulation of IL-17 alone or in combination with TNF-␣. IL-17 did not affect the IL-1␤-mediated induction of IB-. The kinetics of IB-induction by TNF-␣ and IL-17 was similar to that by LPS or IL-1␤ (Fig. 9B).
The activation of NF-B and p38 MAP kinase was examined in order to identify the signal(s) generated by IL-17 stimulation. The degradation of IB-␣ in the cytoplasm is preceded by activation of NF-B. Immunoblotting analysis with anti-IB-␣ antibody revealed the transient degradation of IB-␣ followed by re-synthesis on stimulation with LPS, TNF-␣, and IL-1␤ (Fig. 10A, top). However, the degradation was not detected with IL-17 stimulation, indicating that the NF-B-activation signal is not significantly activated by IL-17. The NF-B reporter analyses showed that there was no NF-B activation with increasing amounts of IL-17, further supporting this conclusion (Fig. 10B). Immunoblotting with anti-active p38 antibody showed that p38 MAP kinase was activated by LPS, TNF-␣, or IL-1␤ but not by IL-17 (Fig. 10A, middle).
We finally examined the effect of IL-17 on the stabilization of IB-mRNA using the stable cell line. As expected, the decay of IB-mRNA was delayed in response to IL-17 alone as well as LPS or IL-1␤ (Fig. 10C). Therefore, IL-17 alone provided a stabilizing signal for IB-mRNA, which was lacking in the cells stimulated with TNF-␣. Thus, costimulation with TNF-␣ and IL-17 induces IB-by compensating the two distinct signals required for the induction, namely activation of NF-B and the specific stabilization of IB-mRNA. DISCUSSION Our previous studies have indicated that activation of NF-B is essential but not sufficient for the induction of IB- (14,18). Another different signal originated from TLR/IL-1R and was missing in the cells stimulated with TNF-␣. This signal is required for induction of IB-, but its identity has remained unknown. In the present study, we attempted to find the signaling required for the differential induction of IB-by IL-1␤ or LPS. The promoter analyses revealed that the NF-B-binding sites were critical for the induction of IB-, supporting our conclusion above from the previous studies (18). However, neither the promoter analyses nor the nuclear run-on assay provided any evidence for preferential transcription of the IBgene. Instead, we found that IB-mRNA was preferentially stabilized by IL-1␤ and LPS but not by TNF-␣.
We also attempted to identify certain stimuli that complement the missing signal in the TNF-␣ signaling. After extensive searching, we identified IL-17, which in combination with TNF-␣ activated the induction of IB-. The stimulation with IL-17 alone did not induce IB-but stabilized IB-mRNA without significant activation of NF-B and p38 MAP kinase. We also found that IL-17 augmented the TNF-␣-mediated NF-B reporter activity but not the IL-1␤-mediated activity (data not shown). However, simple augmentation of the NF-B activity induced by TNF-␣ is unlikely to account for the compensating action of IL-17, because overexpression of the NF-B p65 subunit resulted in remarkable NF-B activity but did not induce IB- (18). Therefore, it is likely that costimulation of TNF-␣ and IL-17 induced IB-by activating the two separable signals required for the induction, which are NF-B activation and mRNA stabilization.
The half-life of IB-in unstimulated cells was almost the same as that of IB-␣. This finding suggests that IB-␣ mRNA is also unstable with a short half-life. Therefore, expression of IB-␣ requires strong transcriptional activation or the stabilization signal for IB-␣ mRNA. The stabilization of mRNA is FIG. 5. The IB-promoter is activated to a similar extent in response to LPS, IL-1␤, and TNF-␣. A, NIH3T3 cells were transfected with the indicated luciferase reporter plasmids and pRL-TK. Two days after transfection, the cells were stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for 6 h, and the luciferase activities were measured. B, A549 cells were transfected with the indicated luciferase reporter plasmids and pRL-TK. Two days after transfection, the cells were stimulated with IL-1␤ (50 pg/ml) or TNF-␣ (20 ng/ml) for 6 h, and the luciferase activities were measured. specific to IB-mRNA but not to IB-␣ mRNA. Transcriptional up-regulation of IB-is not sufficient for induction, whereas IB-␣ was strongly induced without the stabilization of mRNA. As expected from these findings, the transcription of IB-␣ was much more strongly induced than that of IB- (Fig.  6A). Thus, induction of IB-␣ appears to be simply regulated by strong activation of transcription but that of IB-requires at least two different signals, NF-B-mediated transcriptional activation and mRNA stabilization, which could provide more delicate mechanisms for regulation of IB-induction.
A growing body of evidence suggests that proinflammatory stimuli induce stabilization of mRNAs encoded by a series of inflammatory genes to up-regulate their expression, as well as transcriptional activation (21,36). For instance, mRNAs for principal inflammatory mediators such as cyclooxygenase 2, TNF-␣, IL-6, and the chemokine GRO␣ homolog have been shown to be stabilized in response to LPS or IL-1␤ stimulation (21,23,24). The involvement of the p38 MAP kinase cascade in the stabilization of these mRNAs has been suggested previously (21,(37)(38)(39)(40). However, in the present study the stabilization of IB-mRNA did not appear to be dependent on p38 MAP kinase because TNF-␣ induced the activation of p38 MAP kinase to a similar extent as LPS or IL-1␤ (Figs. 2 and 10A). In addition, an inhibitor for the kinase, SB203580, did not affect FIG. 6. The transcription of the IB-gene is activated to a similar extent in response to LPS, IL-1␤, and TNF-␣. The transcription-competent nuclei were prepared from NIH3T3 cells (A) or RAW264.7 cells (C) stimulated with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for the indicated periods and were subjected to nuclear run-on analysis using a probe for vector, G3PDH, A20, IB-, or IB-␣, spotted onto nylon membranes. The activity was normalized with that of G3PDH. B and D, the radioactivity of each spot in A and C was quantitated, and the transcriptional activity of each gene was plotted.
FIG. 7. IB-mRNA was stabilized in response to LPS or IL-1␤ but not to TNF-␣. A, schematic illustration of the construct for expression of the full-length IB-mRNA. A fragment for the fulllength IB-mRNA is located downstream of the chicken ␤-actin promoter. A hatched box and closed ovals in the 3Јuntranslated region (UTR) indicate the ORF of IBand the AREs, respectively. The polyadenylation signal followed by the G/U-rich and U-rich sequences is shown in a shaded box. B, NIH3T3 cells stably expressing the full-length IB-mRNA were treated with actinomycin D (5 g/ml) together with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for the indicated periods. Total RNA was extracted and subjected to Northern blot analyses with a probe for the IBtransgene (Tg), IB-␣, or G3PDH. C, the radioactivities in B were quantitated and shown after they had been normalized with those of G3PDH. the induction of IB-(data not shown). Thus, molecular mechanisms underlying the stabilization of IB-mRNA and other gene products could be different.
We first expected that the cis-element(s) responsible for the stabilization of IB-mRNA would be present in either the 5Јor 3Ј-untranslated regions. The 3Ј-untranslated region of IB-mRNA harbors four AU-rich elements (AREs), which are known to destabilize mRNA. We examined the role for the AREs in the stability of IB-mRNA, but mutations in the elements did not severely affect the decay kinetics or the preferential stabilization (data not shown). We further attempted to analyze the effect of the trans-acting factors that stabilize mRNA by binding to the AREs, such as HuR and Apobec-1 (39,(41)(42)(43)(44). However, overexpression of HuR or Apobec-1 failed to induce IB-mRNA, even after TNF-␣ stimulation (data not shown). The analyses of the chimeric mRNAs identified the cis-element in the region corresponding to the ORF. There have been several reports that showed the determinants for mRNA stability in the coding region (45)(46)(47).
IL-17 is a proinflammatory cytokine produced exclusively from activated T cells (48 -50). In contrast to its restricted expression, the receptor for IL-17 is expressed in a wide variety of cell types. Recent progress on genome-wide sequencing has identified several IL-17-related genes (IL-17 B to F) with redundant biological activities. IL-17 receptor is a type I transmembrane cell surface protein without significant homology to other known proteins (48,50), and its intracellular signaling is poorly understood. Thus, mechanisms for IL-17-mediated mRNA stabilization remain to be investigated. Because mRNA stabilization was measured in the presence of actinomycin D, de novo synthesis of the mediators is not required for IL-17-mediated stabilization. Al- FIG. 8. The cis-element(s) responsible for the stimulus-specific induction of IBis in the ORF of IB-mRNA. A, schematic illustrations of the constructs for the chimeric mRNA composed of IBand luciferase. B, NIH3T3 cells were transiently transfected with the indicated expression plasmids and treated with actinomycin D (5 g/ml) together with LPS (100 ng/ml), TNF-␣ (10 ng/ml), or IL-1␤ (10 ng/ml) for the indicated periods. Total RNA was extracted and subjected to Northern blot analysis with a probe for luciferase, IB-␣, or G3PDH. The radioactivities were quantitated and shown after they had been normalized with those of G3PDH. though IL-17 has been reported to activate NF-B and MAP kinases (35,51), these activities were not detected under our current experimental conditions. It has been suggested that IL-17 signaling requires TRAF6 but not TRAF2 (52,53). Therefore, IL-17 signaling may be qualitatively similar to that of TLR/IL-1R. Our observation that IL-17 augmented the TNF-␣-mediated NF-B activation, but not the IL-1␤mediated activation, appears consistent with this possibility.
Although IL-17 has been reported to elicit the production of inflammatory mediators by itself, recent studies (33)(34)(35) have focused on the synergistic effects of IL-17 and TNF-␣ on the expression of chemokines and cytokines, including IL-6 and the chemokine GRO␣ homolog. The stabilization of IL-6 mRNA (35) and the induction of the CCAAT/enhancer-binding protein (C/EBP)-␦ that engages in the transcription of IL-6 (34) have been shown to be synergistic effects. However, the induction of IB-by costimulation with TNF-␣ and IL-17 is likely to be involved in the induction of IL-6, because IB-is essential for the TLR/IL-1R-mediated induction of IL-6 (17).
The physiological significance of IB-induction by collaboration between TNF-␣ and IL-17 is currently unknown. However, a critical role for IB-in expression of a diverse array of genes raises various possibilities. Recent reports including the present study have suggested that T cell-derived IL-17 and proinflammatory cytokines produced by monocytes/macrophages such as TNF-␣ can cooperatively modulate inflammation. Collaboration between TNF-␣ and IL-17 has been shown to regulate the production of inflammatory cytokines or tissue remodeling factors from synoviocytes (49). IB-might play a crucial role in the induction of inflammatory genes in synovial tissues, which are closely associated with pathophysiology of rheumatoid arthritis or osteoarthritis.
In addition to the efforts to elucidate molecular mechanisms of IB--mediated transcription, investigation of the molecular machinery that leads to IB-induction could be crucial to understanding the precise regulation of inflammation and pathological conditions derived from disorders of this regulation.