A Dominant Function of IKK/NF-κB Signaling in Global Lipopolysaccharide-induced Gene Expression*

Porphyromonas gingivalis is an etiologic pathogen of periodontitis that is one of the most common inflammatory diseases. Recently, we found that P. gingivalis LPS activated the transcription factor nuclear factor-κB (NF-κB) through the IκB kinase complex (IKK). NF-κB is a transcription factor that controls inflammation and host responses. In this study, we examined the role of IKK/NF-κBin P. gingivalis LPS-induced gene expression on a genome-wide basis using a combination of microarray and biochemical approaches. A total of 88 early response genes were found to be induced by P. gingivalis LPS in a human THP.1 monocytic cell lines. Interestingly, the induction of most of these genes was abolished or attenuated under the inactivation of IKK/NF-κB. Among those IKK/NF-κB-dependent genes, 20 genes were NF-κB-inducible genes reported previously, and 59 genes represented putative novel NF-κB target genes. Using transcription factor binding analysis, we found that most of these putative NF-κB target genes contained one or multiple NF-κB-binding sites. Also, some transcription factor-binding motifs were overrepresented in the promoter of both known and putative NF-κB-dependent genes, indicating that these genes may be regulated in a similar fashion. Furthermore, we found that several transcription factors associated with metabolic and inflammatory responses, including nuclear receptors, activator of protein-1, and early growth responses, were induced by P. gingivalis LPS through IKK/NF-κB, indicating that IKK/NF-κB may utilize these transcription factors to mediate secondary responses. Taken together, our results demonstrate that IKK/NF-κB signaling plays a dominant role in P. gingivalis LPS-induced early response gene expression, suggesting that IKK/NF-κB is a therapeutic target for periodontitis.

romonas gingivalis is a Gram-negative bacterium that is recognized as one of the etiologic agents of periodontitis. Periodontitis impinges on a large proportion of the population and is characterized by alveolar bone resorption, ultimately resulting in tooth loss. LPS, 3 a major component of the outer membrane of these bacteria, has been found in infected periodontal tissues and root surfaces (1,2). LPS isolated from P. gingivalis has been shown to activate monocytes/macrophages and gingival fibroblasts (3,4). P. gingivalis LPS induced monocytes/macrophages to produce bone-resorptive cytokines, including interleukin-1 (IL)-1, tumor necrosis factor-␣ (TNF-␣), and IL-6, chemokines, cell adhesion molecules, and matrix metalloproteinases (5)(6)(7). Experimental periodontitis could be induced by delivering a P. gingivalis LPS solution to the oral cavity in mice. The local injection of P. gingivalis LPS has been found to stimulate the production of inflammatory cytokines and bone resorption (6,8,9).
Toll-like receptors have been found to play a critical role in transducing LPS signaling and host responses (2, 9, 10 -12). Interestingly, whereas LPS isolated from Gram-negative bacteria usually utilizes TLR4 to stimulate intracellular signaling pathways, P. gingivalis LPS has been shown to utilize TLR-4 and/or TLR-2 to facilitate cell activation (2,9). It is well known that the LPS/TLR interaction activates the transcription factor nuclear factor B (NF-B), which turns on transcription of inflammatory mediators (2,(13)(14)(15)(16). NF-B was originally identified as a transcription activator that bound to a specific DNA motif (GGGGACTTCCC) in the intronic enhancer of the immunoglobulin light chain gene in B lymphocytes. Subsequently, it was found that NF-B was a ubiquitous cellular factor that was retained in the cytoplasm by inhibitory proteins IBs. NF-B consists of homo-and heterodimeric complexes of the Rel family proteins, including p50, p52, p65/RelA, c-Rel, and RelB. In mammalian cells, the most widely distributed B-binding activity is a heterodimer of p50 and p65/RelA proteins, in which the p65/RelA subunit has potent transactivation activity (17)(18)(19). The IKK complex is critical for NF-B activation induced by LPS and proinflammatory cytokines. The IKK complex is mainly composed of two catalytic subunits, IKK␣ and IKK␤, and IKK␥ (also known as NF-B essential modulator), a scaffold molecule without catalytic activity (17)(18)(19)(20).
Gene depletion studies demonstrate that IKK␤, but not IKK␣, plays an essential role in NF-B activation mediated by LPS and proinflammatory cytokines (18). Although IKK␥ lacks catalytic function, it plays a critical role in assembling the IKK complex. Gene knock-out experiments found that IKK␥ was essential for IKK activation induced by proinflammatory cytokines and LPS (18). IKK␥ contains several predicted functional domains, including two coiled-coil regions that are separated by ␣ helices, a leucine zipper motif, and a putative zinc finger domain at the C terminus. The C terminus of IKK␥ has been found to play a regulatory role in IKK activation (3,15,(21)(22)(23)(24)(25). Recently, we examined the zinc finger domain of IKK␥ and found that C417R mutation in IKK␥ zinc finger domain impaired LPS-or TNF-mediated IKK activation and NF-B transcription. The overexpression of IKK␥C417R inhibited the expression of IKK/ NF-B-dependent genes, such as IL-8, in a dominant negative fashion (15,25). To further understand bacterial infection-mediated host response and inflammation, we examined P. gingivalis LPS-induced genes in human monocytes on a genomewide basis. Using the dominant-negative approach to block IKK/NF-B signaling, we compare gene expression patterns induced by P. gingivalis LPS in monocytes. We found that most of the early response genes induced by P. gingivalis LPS were dependent on IKK/NF-B, suggesting that IKK/NF-B is a therapeutic target for periodontitis.

EXPERIMENTAL PROCEDURES
Cell Culture and Retroviral Infection-Human THP.1 monocytes and murine RAW264.7 macrophages (ATCC) were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 100 g/ml penicillin G, and 100 g/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO 2 . LPS from P. gingivalis 7346 was provided by Dr. Ronald Arnold (University of North Carolina, Chapel Hill, NC). A specific IKK␤ inhibitor, IKKVI, was purchased from Calbiochem. For isolation and culture of primary mouse macrophages from bone marrow, C57BL/6j mice were sacrificed and sterilized with 70% alcohol. Under an aseptic condition, femurs were collected and cut with scissors. The bone marrow was flushed with medium using a syringe. Bone marrow cells were transferred to 100-mm tissue culture plates. Nonadherent cells were removed, and adherent cells were maintained for subsequent experiments.
To generate human THP.1 cells stably expressing the dominant negative mutant of IKK␥ (IKK␥-DN), we utilized retrovirus-mediated transduction, as reported previously (14,15,26). Briefly, HEK293T cells were transfected with the retroviral vectors encoding IKK␥-DN or control empty vector using the calcium phosphate method. 24 h after transfection, cells were treated with 0.01 M sodium butyrate (Sigma) overnight to boost retrovirus production. Retrovirus-containing supernatants were collected 48 h later, filtered, and stored at Ϫ70°C. Human THP.1 monocytes were infected with retroviruses in the presence of 6 g/ml Polybrene (Sigma). 48 h after infection, cells were treated with G418 (600 g/ml) for 2 weeks. The resistant cells were pooled, and cells expressing IKK␥-DN were confirmed by Western blot analysis.
Western Blot Analysis-Human THP.1 monocytes were treated with P. gingivalis LPS (500 ng/ml) for different time periods. Cells were harvested and washed once with cold PBS. Cells were pelleted and lysed with cell lysis buffer containing 1% Nonidet P-40, 5% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 100 mM sodium orthovanadate, and 1:100 protease inhibitor mixtures (Sigma). The protein concentration was measured according to the manufacturer's protocol (Bio-Rad). 50-g aliquots of whole cell lysates were subjected to 10% SDS-PAGE and transferred to polyvinylidene difluoride membrane (Bio-Rad) using a semidry gel transfer cell. The blots were blocked with 5% nonfat milk overnight at 4°C and incubated with the primary antibodies. The immunocomplexes were visualized with horseradish peroxidase-coupled goat anti-rabbit or anti-mouse IgG (Promega) using the SuperSignal reagents (Pierce), as described previously (26,27). The primary antibodies were from the following sources: anti-IB␣, anti-ATF-3, anti-c-Fos, and anti-Fra-1 polyclonal antibodies from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA); anti-phospho-IB␣ (serine 32) antibodies from Cell Signaling; and anti-␣-tubulin from Sigma.
Affymetrix Human Microarray-Cells were treated with P. gingivalis LPS for 1 h. After treatment, cells were harvested, and total RNAs were extracted with Trizol reagents in accordance with the manufacturer's instruction (Invitrogen). To eliminate contaminated genomic DNA, total RNAs were cleaned with an RNeasy kit (Qiagen). 10-g aliquots of total RNA from each sample were utilized for microarray analysis as described previously (28). Briefly, RNAs were transcribed to double-stranded cDNA using SuperScript II reverse transcriptase (Invitrogen) with an oligo(dT) primer that has a T7 RNA polymerase site on the 5Ј-end. Then, the cDNAs were used in an in vitro transcription reaction in the presence of biotin-modified ribonucleotides to generate single-stranded RNAs. The biotin-labeled RNAs were fragmented and hybridized with an Affymetrix (Santa Clara, CA) human U133A gene chip at 45°C for 16 h in a mix that contained 10 g of fragmented RNA, 6ϫ SSPE, 0.005% Triton X-100, and 100 g/ml herring sperm DNA in a total volume of 200 l. Labeled bacterial RNAs were spiked into the hybridization mix for an internal standard and normalization. Chips were washed and stained with Streptavidin R-phycoerythrin (Molecular Probes, Inc., Eugene, OR). The arrays were scanned with the GeneArray scanner (Affymetrix). Signal intensity was calculated using the one-step Tukey's biweight estimate (28). Affymetrix Microarray Suite 5.0 was used for data analysis. Scaling was performed to facilitate the comparison of multiple arrays. We chose a target intensity of 500 in scaling. The Microarray Suite 5.0 detection algorithm generates a detection p value from probe pair intensities and assigns a present ( p), marginal (M), or absent (A) call to each probe set. We applied the default cut-off value of 0.04 (i.e. a p value under 0.04 indicates a P call). The Microarray Suite change algorithm utilized Wilcoxon's signed rank test to compare each probe set on the experiment array with its counterpart on the base-line array, and a change p value was calculated. We assigned increase (I) (p Ͻ 0.0025), decrease (D) (p Ͼ 0.9975), or no change call (NC) (p between 0.0025 and 0.9975) to each probe set. Significant increase was assigned to a probe set if it had a detection call of P on experiment array, a change call of I, and a -fold change of at least 2. In the same way, significant decrease was assigned to a probe set if it had a detection of P on base-line array, a change call of D, and a -fold change of 0.5 or lower.
Gene Function Annotation-Gene Ontology (GO) terms for genes were obtained from the hgu74av2 and mgu74av2 package of bioconductor (29) (available on the World Wide Web at www.bioconductor.org/packages/data/annotation/stable/src/ contrib/html/index.html). We used those packages to build an association file as an input for TermFinder (available on the World Wide Web at search.cpan.org/dist/GO TermFinder/) (30), which calculates a p value for the overexpression of specific GO terms in gene sets using a hypergeometric distribution, and output overrepresented GO terms.
Gene Set Enrichment Analysis (38) was performed using the stand alone javaGSEA package (available on the World Wide Web at www.broad.mit.edu/gsea/software/software_index.html) and the "s2.hgul133agmt" functional gene sets downloaded from the same site. Gene sets achieving a nominal p value of less than 0.05 are defined as "enriched." Transcription factor binding site analysis was performed by obtaining sequences from the ENSEMBL data base (available on the World Wide Web at www.ensembl.org/index. html) (32) and scanning the genomic sequence regions spanning 3 kb around each transcription start site using the TRANSFAC MATCH program (23,34,35). Default weight matrix score cutoffs were used.
Quantitative Real Time PCR-Total RNAs were isolated using Trizol reagent. 2-g aliquots of total RNAs from each sample were subjected to reverse transcription using a Superscript first strand cDNA synthesis kit (Invitrogen) according to the manufacturer's protocol. Quantitative real time PCR was performed on an iCycler (Bio-Rad) using SYBR Green PCR master mixture (Applied Biosystems) according to the manufacturer's instructions. The human primers used for THP-1

Inhibition of P. gingivalis LPS-induced IKK Activation by Overexpression of IKK␥ Dominant Negative Mutant (IKK␥-DN)-
Previously, we have found that the overexpression of IKK␥C417R could act as a dominant negative mutant to inhibit TNF-and LPS-mediated IKK activation and NF-B transcription in human THP.1 monocytic cells (15). However, human THP.1 cells expressing IKK␥-DN were not very stable. After prolonged passages, we found that the level of IKK␥-DN was significantly reduced and was not sufficient to inhibit the IKK/ NF-B signaling pathway. Therefore, to globally examine IKK/  Fig. 1, B and C, although P. gingivalis LPS rapidly induced the phosphorylation and degradation of IB␣ by acti-vating IKK activities in THP.1/V cells, the phosphorylation and degradation of IB␣ was inhibited in THP.1/IKK␥-DN cells. Also, P. gingivalis LPS-induced NF-B transcription was inhibited in THP.1/IKK␥-DN cells using a B-dependent luciferase reporter assay (data not shown).

P. gingivalis LPS-induced Genes in Monocytes by Microarray
Analysis-Next, we utilized these cells to globally compare IKK/NF-B-dependent genes induced by P. gingivalis LPS in monocytes. Both THP.1/IKK␥-DN and THP.1/V cells were treated with P. gingivalis LPS for 1 h only so that we could examine the primary IKK/NF-B-dependent genes induced by P. gingivalis LPS. After treatment, total RNAs were extracted and assayed using an Affymetrix human gene chips U133A. Affymetrix Microarray Suite 5.0 was used for data analysis. Genes that were differentially expressed between THP.1/V cells and THP.1/IKK␥C417R cells were compared following P. gingivalis LPS stimulation. An overview of -fold changes of these genes (Ն2-fold) was shown in Fig. 2A. Compared with untreated cells, we found that P. gingivalis LPS significantly induced 88 genes by Ն2-fold in THP.1/V cells. Although only 10% of these genes (nine genes) were not affected, the -fold induction of 90% of these genes (79 genes) was inhibited or reduced in THP.1/IKK␥-DN cells compared with THP.1/V cells following P. gingivalis LPS stimulation (Fig. 2B). Among those genes that were up-regulated upon P. gingivalis LPS stimulation in THP/V cells, almost 72% of them remained unchanged upon P. gingivalis LPS stimulation in THP.1/ IKK␥-DN cells. Although another 18% were up-regulated in THP/IKK␥-DN, most of them were induced with a smaller -fold change compared with THP.1/V cells. Since we previously showed that overexpression of IKK␥-DN severely inhibited LPS-induced IKK activation and NF-B transcription in a dominant-negative fashion (15), both of these sets of genes were considered as NF-B-dependent genes. However, we observed that only five genes (two SEC23A, HMGA1, ARHGDIA, and LOC40257) were repressed by P. gingivalis LPS, suggesting that it mainly induces gene transcription during early gene responses. Taken together, our results suggest that P. gingivalis LPS predominantly utilizes IKK/NF-B signaling to induce primary gene responses.
Previously, we demonstrated that the overexpression of IKK␥-DN did not affect the expression of IKK␣, IKK␤, and p65, which are the signaling components of NF-B (15). Consistently, our microarray found that the expression of these genes did not show statistically significant differences between THP.1/V and THP.1/IKK␥-DN cells. Importantly, we found that the inhibition of IKK/NF-B by IKK␥-DN suppressed the expression of IL-8, a well known NF-B target gene that is induced by P. gingivalis LPS. Also, we observed that the -fold induction of IL-8 was significantly reduced in THP.1/IKK␥-DN cells compared with THP.1/V cells upon P. gingivalis LPS stimulation. Moreover, a number of common inflammatory mediators that are well known NF-B-dependent genes, including IL-1, cyclooxygenase-2, and chemokine ligands 2, 3, and 4 (18,19,36), were identified in our microarray. Previously, excellent studies by Mangan et al. (26) found that LPS could prolong monocyte survival by inhibiting apoptosis. We and others have demonstrated that NF-B played a critical role in regulation of apoptosis (37)(38)(39)(40).Consistently,NF-B-dependentantiapoptotic gene a20 and bcl-2-related genes a1 and mcl-1 (8,27,39) were found to be induced by P. gingivalis LPS in an IKK/NF-B-dependent fashion (Tables 1 and 2). These results demonstrate that our microarray data are highly reliable.
Bioinformatics Analysis of P. gingivalis LPS-stimulated Early Gene Responses-We performed GO analysis to categorize those LPS-induced NF-B-dependent genes listed in Table 1. As listed in Table 3, consistent with NF-B biological functions, these genes were associated with response to stress, immune and inflammatory response, chemotaxis, antimicrobial humoral responses, and apoptosis. Additionally, gene set enrichment analysis also showed a similar pattern (Table 4).

IKK in LPS-induced Gene Expression
suggest that NF-B may utilize these transcription factors to further modulate or amplify inflammatory responses.
To confirm these results, we first performed Western blot analysis to examine whether AP-1 family proteins were induced by LPS through IKK/NF-B. To block the IKK/NF-B signaling pathway, we utilized a specific IKK␤ inhibitor, IKKVI. As shown in Fig. 3A, IKKVI potently inhibited the phosphorylation and degradation of IB␣ induced by P. gingivalis LPS in THP.1 cells. Consistent with our microarray data, we found that LPS induced the expression of c-Fos, Fra-1, and ATF-3 expression in THP.1 cells. The pretreatment of the IKKVI inhibitor abolished their induction (Fig. 3B). In RAW264.7 cells, we found that the induction of ATF-3 and Fra-1 by P. gingivalis LPS was also dependent on the IKK/NF-B signaling pathway. Interestingly, c-Fos was constitutively expressed in RAW264.7 cells. Although P. gingivalis LPS could not induce its expression, the IKKVI inhibitor abolished its basal expression.
We utilized real time PCR to confirm the expression of egr-1, -2, and -3, nr4a1, nr4a2, and pnrc1 induced by P. gingivalis LPS. As shown in Fig. 4, P. gingivalis LPS significantly induced their expression in THP.1 cells, and pretreatment of the IKKVI inhibitors suppressed their induction, validating our microarray results. Except for nr4a2, we also confirmed that P. gingivalis LPS induced their expression in RAW264.7 cells through IKK/NF-B signaling (Fig. 5).
We also further examined whether P. gingivalis LPS induced these genes in primary macrophages through IKK/NF-B signaling. Primary macrophages were prepared from mouse bone marrows. As shown in Fig. 6, real time PCR revealed that the expression of egr-1, -2, and -3 and nr4a1 was also induced by P. gingivalis LPS and that the addition of IKKVI inhibitor suppressed their expression. However, probably due to cell type, we could not detect that P. gingivalis LPS induced nr4a2 and pnrc1 in primary bone marrow macrophages.
NF-B Directly Binds to the egr-2 Promoter-It is known that NF-B usually stimulates gene transcription by binding to a specific DNA motif. 22 known NF-B target genes that we identified as induced by P. gingivalis LPS contain NF-B-binding sites. Since 57 genes were putative novel NF-B-dependent genes, we further analyzed the 3-kb promoter region upstream of the transcription start site of these genes using transcription factor binding site analysis. The sequences of  Cells were pretreated with the IKKVI inhibitor and then treated with P. gingivalis LPS for 1 h. Total RNAs were extracted with Trizol reagents. The expression of egr-1, -2, and -3, nr4a1, nr4a2, and pnrc1 was quantitatively measured by real time PCR. A, the induction of egr-1 by P. gingivalis LPS was dependent on IKK. B, the induction of egr-2 by P. gingivalis LPS was dependent on IKK. C, the induction of egr-3 by P. gingivalis LPS was dependent on IKK. D, the induction of pnrc1 by P. gingivalis LPS was dependent on IKK. E, the induction of nr4a1 by P. gingivalis LPS was dependent on IKK. F, the induction of nr4a2 by P. gingivalis LPS was dependent on IKK. FIGURE 5. Examination of P. gingivalis LPS-induced genes by real time PCR in RAW264.7 cells. Cells were pretreated with the IKKVI inhibitor and then treated with P. gingivalis LPS for 1 h. Total RNAs were extracted with Trizol reagents. The expression of egr-1, -2, and -3, nr4a1, and pnrc1 was quantitatively measured by real time PCR. A, the induction of egr-1 by P. gingivalis LPS was dependent on IKK. B, the induction of egr-2 by P. gingivalis LPS was dependent on IKK. C, the induction of egr-3 by P. gingivalis LPS was dependent on IKK. D, the induction of pnrc1 by P. gingivalis LPS was dependent on IKK. E, the induction of nr4a1 by P. gingivalis LPS was dependent on IKK.
these genes from the ENSEMBL data base (available on the World Wide Web at www.ensembl.org/index.html) (32) were obtained and scanned using TRANSFAC MATCH. We discovered that, of the 57 genes that were not previously identified as NF-B target genes, 41 genes have at least one NF-B binding site around the transcription start site. More than 88% of those 41 genes have more than one NF-B binding site, half have more than two binding sites, and a few have as many as eight or nine binding sites. Since we found that P. gingivalis LPS induced egr-2 in three different cells through IKK/NF-B and because egr-2 is a novel NF-B target gene, we were interested in determining whether NF-B could directly bind to the egr-2 promoter by ChIP assay, thereby stimulating its transcription. As shown in Fig. 7A, a putative NF-B-binding site in the egr-2 promoter was revealed by TRANS-FAC MATCH. Quantitative real time PCR analysis of ChIP samples revealed that the active subunit of NF-B p65 was recruited to the egr-2 promoter following P. gingivalis LPS stimulation. The pretreatment of the IKKVI inhibitor significantly suppressed p65 bound to the egr-2 promoter (Fig. 7B). These results suggest that egr-2 is a direct target of IKK/ NF-B signaling.

DISCUSSION
Our microarray analysis demonstrated here that P. gingivalis LPS induced a broad range of immune and inflammatory response genes in monocytes. Importantly, we found that IKK/NF-B signaling played a dominant role in P. gingivalis LPS-induced early gene responses in monocytes by activating IKK, since the inhibition of IKK activities abolished or attenuated expression of most of the genes induced by P. gingivalis LPS. Our results suggest that IKK may be a critical target for inhibiting periodontal inflammation induced by oral pathogens, such as P. gingivalis LPS.
During the early stage of periodontal infection, it is known that bacteria or bacterial products induce monocytes/macrophages to secrete proinflammatory mediators and chemokines. Subsequently, these cytokines and chemokines recruit neutrophils and lymphocytes to infected sites (1,4,9,41,42). Consistently, we found that these proinflammatory cytokines and chemokines were induced by P. gingival LPS in monocytes on a genome-wide basis. More importantly, the inhibition of the IKK/NF-B signaling pathway suppressed or attenuated expression of these genes in monocytes. Our results highlight the critical role of the IKK/NF-B signaling pathway in the initial phase of periodontitis, which leads to recruitment of neutrophils and lymphocytes. Additionally, we have found that several NF-B-dependent antiapoptotic genes, including a20, a1, and mcl-1, were induced by P. gingivalis LPS. Since monocytes/ macrophages spontaneously undergo apoptosis in inflamed sites, the life span of monocytes/macrophages in periodontitis may be associated with the resolution of inflammation (43). The induction of antiapoptotic genes by P. gingivalis LPS may promote monocyte/macrophage survival, which contributes to the prolonged inflammation in chronic periodontitis (43).
Importantly, we have identified a group of transcription factors, including AP-1 family proteins, nuclear receptor subfamily proteins, and EGR-1 family proteins, which were induced by P. gingivalis LPS through the IKK/NF-B signaling pathway. Although functional roles of some transcription factors are unclear in the context of LPS-induced inflammation and immune responses, activation of AP-1 has been found to play FIGURE 6. Examination of P. gingivalis LPS-induced genes by real-time PCR in primary macrophages. Primary macrophages were isolated from mouse bone marrow. Cells were pretreated with the IKKVI inhibitor and then treated with P. gingivalis LPS for 1 h. Total RNAs were extracted with Trizol reagents. The expression of egr-1, -2, and -3 and nr4a1 was quantitatively measured by real time PCR. A, the induction of egr-1 by P. gingivalis LPS was dependent on IKK. B, the induction of egr-2 P. gingivalis LPS was dependent on IKK. C, the induction of egr-3 by P. gingivalis LPS was dependent on IKK. D, the induction of nr4a1 by P. gingivalis LPS was dependent on IKK.