Signal Transduction Pathways Involved in Rheumatoid Arthritis Synovial Fibroblast Interleukin-18-induced Vascular Cell Adhesion Molecule-1 Expression*

Vascular cell adhesion molecule (VCAM)-1 has been implicated in interactions between leukocytes and connective tissue, including rheumatoid arthritis (RA) synovial tissue fibroblasts. Such interactions within the synovium contribute to RA inflammation. Using phosphoinositide 3-kinase (PI3-kinase) inhibitor LY294002 and Src inhibitor PP2, we show that interleukin (IL)-18-induced ERK1/2 activation is Src kinase-dependent. Antisense (AS) c-Src oligonucleotide (ODN) treatment reduced IL-18-induced ERK1/2 expression by 32% compared with control, suggesting an upstream role of Src in ERK1/2 activation. AS c-Src ODN treatment also inhibited Akt expression by 74% compared with sense control. PI3-kinase inhibitor LY294002 or AS PI3-kinase ODN inhibited Akt expression. AS c-Src ODN inhibited Akt phosphorylation, confirming Src is upstream of PI3-kinase in IL-18-induced RA synovial fibroblast signaling. IL-18 induced a time-dependent activation of c-Src, Ras, and Raf-1, suggesting this signaling cascade plays a role in ERK activation. IL-18 directly activated Src kinase by more than 4-fold over basal levels by enzymatic assay. Electrophoretic mobility shift assay showed that activator protein-1 (AP-1) is activated by IL-18 through ERK and Src but not through PI3-kinase. In an alternate pathway, inhibition of IL-1 receptor-associated kinase-1 (IRAK) with AS ODN to IRAK reduced IL-18-induced expression of nuclear factor κB (NFκB). Finally, IL-18-induced cell surface VCAM-1 expression was inhibited by treatment with AS ODNs to c-Src, IRAK, PI3-kinase, and ERK1/2 by 57, 43, 41, and 32% compared with control sense ODN treatment, respectively. These data support a role for IL-18 activation of three distinct pathways during RA synovial fibroblast stimulation: two Src-dependent pathways and the IRAK/NFκB pathway. Targeting VCAM-1 signaling mechanisms may represent therapeutic approaches to inflammatory and angiogenic diseases characterized by adhesion molecule up-regulation.

Interleukin-18 (IL-18) 1 is a proinflammatory cytokine associated with various pathological conditions including rheumatoid arthritis (RA). IL-18 induces the release of Th1 cytokines by T cells and macrophages and also stimulates the production of inflammatory mediators, such as chemokines by synovial fibroblasts or nitric oxide by macrophages and chondrocytes (1,2). Additionally, we have shown that IL-18 acts upon endothelial cells to induce angiogenesis and cell adhesion (3,4). IL-18 is mainly produced by activated macrophages, whereas the IL-18 receptor (IL-18R) is expressed on T lymphocytes, natural killer cells, macrophages, neutrophils, and chondrocytes (1,5,6). The IL-18R complex is composed of two protein chains ␣ and ␤. The IL-18R␣ is the extracellular binding domain of the IL-18R complex, whereas the IL-18R␤ is the signal transducing chain. When IL-18 binds to the IL-18R, it induces the formation of an IL-1R-associated kinase (IRAK)/TNF receptor-associated factor 6 (TRAF-6) complex that subsequently activates nuclear factor B (NFB) in Th1 cells (7) and in EL4/6.1 thymoma cells (8). IL-18 also operates through IRAK-independent pathways that remain to be elucidated. For instance, IL-18 mediates interferon-␥ (INF-␥) production by the mitogen-activated protein kinases p38 and p42/44, also known as ERK1/2 (9).
Phosphatidylinositol 3-kinase (PI3-kinase) is a lipid kinase that consists of catalytic (p110) and regulatory (p85) subunits. PI3-kinase catalyzes the phosphorylation of the inositol phospholipids at position 3 to generate phosphatidylinositol 3-phosphates, phosphatidylinositol 3,4-biphosphates, and phosphatidylinositol 3-5-triphosphates. These phosphorylated lipid products act as second messengers, activating protein kinases such as Akt (also known as protein kinase B). PI3-kinase is activated by a large spectrum of cytokines, growth factors, and hormones (10). This activation of PI3-kinase is generally regulated by receptor tyrosine kinase and non-receptor tyrosine kinase (NRTK). PI3-kinase can also be activated by G-proteincoupled receptors or by the small GTPase Ras (11,12). PI3-kinase has been implicated as a key signaling molecule for transcription factor activation, protein synthesis, angiogenesis, and cell adhesion (13,14). We recently showed that the PI3kinase inhibitor LY294002 inhibited RA synovial fibroblast IL-18-induced VCAM-1 expression by 50% when used alone and by 85% when used with the NFB inhibitor pyrrolidine dithiocarbamate (2). These results suggested the existence of at least two independent pathways involved in IL-18-induced adhesion molecule expression.
Further investigation of IL-18-induced signaling mechanisms revealed involvement of Src, an NRTK that functions in ligand-induced cellular responses, such as leukocyte survival, adhesion, migration, and proliferation. Src has also been implicated in various cancers and in bone resorption (15). Activation of Src requires phosphorylation at its activation site Tyr-418 concomitantly with decreased phosphorylation at its negative regulatory site Tyr-529; regulation also involves phosphorylation on other residues such as Tyr-215. c-Src activation by tyrosine kinase receptors leads to translocation from the plasma membrane to the cytoskeleton, where Src interacts with a host of proteins that orchestrate cell-matrix adhesion and cell migration (16). Here we found that IL-18 directly activates Src with rapid kinetics, and Src activation appears to be an early event common to the PI3-kinase/Akt and ERK1/2 pathways.
There is precedence for the involvement of transcription factor activation in IL-18-mediated immune and inflammatory functions. For instance, IL-18 activates AP-1 and NFB in the Jurkat T cell leading to IL-2 expression (17). We therefore examined involvement of the PI3-kinase and ERK1/2 signaling pathways in AP-1 activation. Because many cytokine-mediated functions, including IL-18-induced IFN-␥ expression, are regulated via NFB activation (18,19), its significance in IL-18induced VCAM-1 expression was examined. One approach for determining the role of the NFB pathway is by way of IRAK involvement, because IRAK is known to be associated with NFB upon IL-18 stimulation (20,21). Furthermore, IRAK was initially cloned and characterized as a kinase associated with the IL-1 receptor (22), suggesting its importance to IL-18 signaling as IL-18 is a member of the IL-1 family.
Because antisense oligodeoxynucleotides (ODNs) offer a potential gene therapy strategy to block transcription or translation of specific genes, antisense ODNs to relevant signaling molecules were employed to examine RA synovial fibroblast signaling events following IL-18 stimulation. Ultimately, the treatment effect of antisense ODN on IL-18-induced VCAM-1 expression implicated specific signaling molecules involved in RA synovial fibroblast VCAM-1 expression. The ability of these antisense ODNs to inhibit IL-18-induced activation of the transcription factor NF-B and downstream VCAM-1 expression was also assessed.
We examined the signal transduction mechanisms by which IL-18 induces VCAM-1 expression in RA synovial fibroblasts. Our findings demonstrate that IL-18 induced VCAM-1 expression through Src kinase, PI3-kinase/Akt, and ERK1/2 pathways. The important role of the IRAK/NFB pathway in VCAM-1 expression was also elucidated. Finally, we describe a new signaling cascade involving Src/Ras/Raf/ERK/AP-1 in IL-18-stimulated RA synovial fibroblasts.
Cell Culture-Fibroblasts were isolated from synovium obtained from RA patients meeting the American College of Rheumatology criteria for RA who had undergone total joint replacement surgery or synovectomy (23). Fresh synovial tissues were minced and digested in solution of dispase, collagenase, and DNase. Cells were used at passage 5 or older, at which time they were a homogeneous population of fibroblasts. Synovial fibroblasts were grown in 175-mm tissue culture flasks (Falcon, Franklin Lakes, NJ) at 37°C, in a humidified atmosphere with 5% CO 2. Upon confluence, cells were passaged by brief trypsinization as described previously (24).
Cell Lysis and Immunoblotting-RA synovial fibroblasts were plated onto 6-or 10-cm Petri dishes (Falcon) at 1 ϫ 10 5 cells/ml and allowed to adhere for 24 h at 37°C in 5% CO 2 atmosphere. Alternatively, for ODN experiments, fibroblasts were plated at 4 ϫ 10 6 cells/well on 6-well plates. Fibroblasts were serum-starved for at least 14 h before stimulation with IL-18 (10 nM) for 0, 5, 10, and 20 min. At the end of each period, supernatants were gently aspirated, and fibroblasts were lysed in extraction buffer containing 100 mM Tris, pH 7.4, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM NaP 2 O 4 , 2 mM Na 3 VO 4 , 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% deoxycholate, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitors (1 tablet/10 ml). For experiments with signaling inhibitors, RA synovial fibroblasts were preincubated with the respective inhibitor for 60 -120 min before activation with IL-18. For the ODN experiments the cells were treated as detailed and lysed similarly. Nuclei were pelleted (1250 ϫ g at 4°C for 5 min), and supernatants of different samples were collected for determination of protein content using a BCA protein assay kit. Cell lysates were mixed 1:1 with Laemmli's sample buffer, boiled for 5 min, and then centrifuged at 10,000 ϫ g for 10 min. Equal amounts (or 15 g) of each sample was subjected to 10% SDS-PAGE. Separated proteins were electrophoretically transferred from the gel onto nitrocellulose membranes using a semi-dry transblotting apparatus (Bio-Rad). To block nonspecific binding, membranes were incubated with 5% milk in Tris-buffered saline containing 0.1% Tween 20 (TBST) for 1 h at room temperature. Blots were incubated in respective primary antibody in TBST ϩ 5% milk at 4°C overnight. After washing with TBST, blots were incubated with horseradish peroxidase-conjugated sheep anti-mouse IgG (1:10,000) or with goat anti-rabbit IgG (1:10,000) for 1 h at room temperature. An ECL detection system (Amersham Biosciences) was used to detect specific protein bands. Different bands were then scanned and quantitated using the software UN-SCAN-IT version 5.1 (Silk Scientific, Orem, UT). Blots were subsequently stripped and restained with antibody to tubulin to determine relative band densities.
PI3-Kinase Assay-RA synovial fibroblasts (8 ϫ 10 5 cells) were plated in 10 cm-dishes in RPMI 1640 containing 10% FBS, 1% penicillin/streptomycin. Once cells were 80 -90% confluent, they were further incubated in serum-free RPMI 1640 for at least 14 h. RA synovial fibroblasts were then stimulated with IL-18 (10 nM) for 0, 10, and 20 min at 37°C. At the end of the incubation, cell lysate was prepared. Protein content of each sample was quantitated using a BCA protein assay kit and normalized according to the protein concentration. 500 g of each sample in 500 l of lysis buffer was incubated overnight at 4°C with 5 l of rabbit anti-PI3-kinase antibody directed against the 85-kDa regulatory subunit. 60 l of protein A-agarose conjugate (50% slurry in PBS) was added to each sample and further incubated for 1 h at 4°C. The immunoprecipitates were collected by centrifugation at 14,000 ϫ g for 10 s. The immunoprecipitates were then washed 3 times with buffer A (137 mM NaCl, 20 mM Tris-HCl, pH 7.4, 1 mM CaCl 2 , 1 mM MgCl 2 , and 0.1 mM sodium orthovanadate) containing 1% nonionic detergent IGE-PAL CA-630, followed by 3 washes with buffer B (0.1 M Tris-HCl, pH 7.4, 5 mM lithium chloride, and 0.1 mM sodium orthovanadate) and 3 washes with TNE (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 5 mM EDTA) containing 0.1 mM sodium orthovanadate. To each sample, the following reagents were added sequentially: 50 l of TNE, 10 l (20 g) of phosphatidylinositol (PI, in 10 mM Tris-HCl, pH 7.4 containing 1 mM EGTA), and 10 l of 100 mM MgCl 2 . The PI reaction was initiated with addition of 5 l of [␥-32 P]ATP. The reaction mixture was incubated at 37°C for 15 min with continuous agitation. The reaction was stopped by addition of 20 l of 6 N HCl. Radiolabeled lipid was extracted from the reaction sample by adding 160 l of CH 3 Cl/MeOH (1:1), vortexing, and separating the organic and aqueous phases by centrifugation for 10 min at 14,000 ϫ g. 50 l of radiolabeled lipid containing the lower organic phase were spotted onto oxalate-treated TLC plates (Fisher) and developed in CHCl 3 /MeOH/H 2 O/NH 4 OH (60:47:11.3:2). TLC plates were dried and autoradiographed.
Ras Activation Assay-Ras activation was studied using a Ras activation kit (Upstate Biotechnology, Inc.). RA synovial fibroblasts were stimulated with IL-18 (10 nM) for different times. At each time point, synovial fibroblast extracts were prepared with cell lysis buffer, and protein content in each sample was quantitated. The Ras activation assay involved two steps. In the first step, the activated form of Ras was immunoprecipitated from 800 g of each synovial fibroblast lysate sample with an immobilized Raf-1-Ras binding domain (Raf-1-RBD) and subsequently run on 10% SDS-PAGE as described above. The presence of activated Ras in samples was then detected by probing with a specific mouse monoclonal anti-Ras antibody (1 g/ml). Different bands were then scanned and quantitated using an imaging densitometer.
Src Kinase Activation Assay-Src kinase was first immunoprecipitated with agarose beads conjugated with anti-Src antibody (mouse monoclonal IgG1). RA synovial fibroblasts were plated at 1 ϫ 10 6 cells on 10-cm culture dishes, adhered overnight in complete media, serumstarved overnight, and stimulated with IL-18 (10 nM) for 10 min. Cells were washed in cold PBS and lysed with 0.5 ml of RIPA buffer. The total cell lysate was diluted to 1 g/l with PBS, and 1 mg was mixed with 4 g (8 l) of anti-Src antibody-conjugated agarose beads and gently rocked at 4°C for 2 h. Beads were collected by microcentrifuging (5 s at 14,000 ϫ g), the supernatant was drained, and beads were then washed (3 times) with ice-cold PBS. Samples containing agarose beads, Src antibody, and Src protein in complex were then used to determine Src kinase activation.
Direct activation of Src was examined with n Src kinase assay kit (Upstate Biotechnology, Inc.) to measure the ability of activated Src to act on known substrate. Stock solutions necessary for the assay were prepared as follows: Src substrate peptide (600 M, diluted in SrcRB), purified Src kinase (p60 c-Src, 20units/10 l/test), [␥-32 P]ATP (PerkinElmer Life Sciences, 1 mCi/100 l, 3000 Ci/mmol, further diluted to 1 Ci/l with Src manganese/ATP mixture), and 0.075% phosphoric acid (diluted from 85% with PBS). Substrate peptide (10 l, 150 M final concentration) was added to 10 l of SrcRB, to which was added 10 l of Src (p60, c-Src, 20 units) or immunoprecipitated sample (200 g minimum) and 10 l of diluted [␥-32 P]ATP (10 Ci) in a microcentrifuge tube. The reagent mixture was incubated for 10 min at 30°C with agitation. To precipitate peptides, 20 l of 40% trichloroacetic acid was added to each mixture and incubated at room temperature for 5 min. Onto the center of P81 phosphocellulose paper squares, 25 l of each sample was spotted. Squares were washed with 0.75% phos-phoric acid (5 times for 5 min) and acetone (once for 3 min). Samples were read in a scintillation counter, and counts/min of immunoprecipitated enzyme samples were compared with counts/min of the background control samples (no enzyme). This assay was performed with fibroblasts from 4 different RA donors.
Preparation of Oligonucleotides and Lipofection of RA Synovial Fibroblasts-Sequences of the ODNs employed in this study are listed in Table I. Antisense ODNs were selected for sequence target to c-Src (25), IRAK-1 (26), PI3-kinase (27,28), and ERK1/2 (29). The corresponding sense ODN was used as control for each antisense ODN. The ODNs were synthesized and purified by the Northwestern University Biotechnology Laboratory and modified with phosphorothioate. Lipofectionencapsulated ODNs were prepared using LipofectAMINE and Lipo-fectAMINE Plus Reagent from Invitrogen. Each ODN was reconstituted to 0.4 g/l concentration in double distilled water and then diluted in Opti-MEM media as detailed. For transient transfection of cells, ODN/Plus reagent-LipofectAMINE complex was prepared with 40 g of ODN/100 l of serum-free Opti-MEM with 2.5 l of Plus reagent and 4 g of LipofectAMINE Reagent/100 l of serum-free Opti-MEM and incubation at room temperature for 30 min, followed by dilution with another 800 l of Opti-MEM serum-free medium for an additional 15-min incubation. RA synovial fibroblasts were plated at 90% confluency at 4 ϫ 10 5 cells per well on 6-well plates or per 6-cm dish and allowed to adhere in RPMI, 10% FBS, 1% penicillin and streptomycin. After attachment, cells were then treated with 5 M of antisense or sense ODN by incubation with 1 ml of ODN/Plus reagent/ LipofectAMINE complex for 5 h, followed by media change to complete RPMI, 10% FBS, 1% penicillin and streptomycin overnight. The media were also changed to serum-free RPMI, 1% penicillin and streptomycin for 8 -10 h prior to stimulation with IL-18 (10 nM, 10 min, for the Western blot and Src kinase experiments; 5 nM, 8 h, for VCAM-1 expression by flow cytometry experiments).

FIG. 1. IL-18 activates PI3-kinase in RA synovial fibroblasts.
Serum-starved cells were stimulated for the indicated times with IL-18 (10 nM). Extracts were subjected to immunoprecipitation with a rabbit antibody against the p85 subunit of PI3-kinase. PI3-kinase activity was determined using an in vitro kinase assay as described under "Experimental Procedures." Products were separated by thin layer chromatography. Activity of PI3-kinase is presented as production of phosphatidylinositol phosphate (PIP), indicated by arrows. Experiments were repeated two times with essentially identical results.

GCA GCC ATG GCC GGG GGG
Antisense c-Src

ATG GGG AGC AGC AAG AGC AAG CCC
ATG GCG GCG GCG GCG GC cells/ml and allowed to adhere overnight, and cells were pretreated with pertussis toxin (100 ng/ml) for 12 h before stimulation with IL-18 (10 nM) for 8 h. For the ODN experiments, RA synovial fibroblasts were plated and treated as detailed above prior to IL-18 (5 nM) stimulation for 8 h. Cells were harvested with a cell scraper and transferred to fluorescence-activated cell sorting (FACS) tubes (BD PharMingen). Cells were then treated with mouse anti-VCAM-1 or isotype-matched control (5 g/ml) as primary antibody followed by incubation for 30 min with PE-conjugated goat anti-mouse antibody (1.5 g/ml). Samples were washed twice with PBS, 1% FBS, and then fixed with 1% paraformaldehyde. Samples were assayed using an Epics XL-MCL flow cytometer (Beckman Coulter). Prior to data acquisition, the PE channel was standardized using fluorescent beads (Rainbow Beads, Spherotech, Libertyville, IL). Isotype-matched control values were subtracted from the test results. Cellular ELISA-RA ST fibroblasts were plated on 96-well tissue culture plates in RPMI 1640, 10% fetal calf serum and adhered for 14 h. Cells were preincubated with specific inhibitors or Me 2 SO vehicle control in RPMI 1640, 2% fetal calf serum before stimulation with IL-18 (5 nM) for 8 h. Inhibitors (LY294002, PP2, SB203580, and PD98059) or vehicle Me 2 SO were applied to cells for 60 min, and then cells were stimulated with IL-18 (5 nM) for 12 h. Cell viability was judged by trypan blue exclusion and was Ͼ90%. RA synovial fibroblasts were successively fixed in 3.7% formalin in PBS and blocked in PBS, 1% bovine serum albumin, 5% goat serum for 15 min. After successive incubations in mouse anti-human VCAM-1 or isotype-matched control for 2 h, goat anti-mouse IgG peroxidase-conjugated antibody was added for 1 h, and the ELISA was developed with tetramethylbenzidine sub-strate. The reaction was stopped with 1 N H 2 SO 4 before reading at 450 nm with a Bio-Rad model 550 microplate reader.
Electrophoretic Mobility Shift Assay-RA synovial fibroblasts (1 ϫ 10 6 ) were grown on 10-cm dishes (Falcon) and starved for more than 12 h in serum-free RPMI 1640, 1% penicillin/streptomycin, before stimulation with 10 nM of IL-18 for 0, 1, 2, and 4 h. For experiments using specific inhibitors, cells were pretreated for 2 h before stimulation with IL-18 (10 nM). Cytoplasmic and nuclear extracts were prepared as described previously (4), and 10 g of nuclear extract from each sample were incubated with 1.5 g of poly(dI-dC) and 47 fmol of 32 P-end-labeled ODN probe containing the DNA-binding site for AP-1 (Promega, Madison, WI). Cold ODN competition control was performed using labeled and unlabeled ODNs at a ratio of 1:2 and 1:4. Protein-DNA complexes were resolved on 4% polyacrylamide gels and visualized by autoradiography.
Statistical Analysis-Data were analyzed using the Wilcoxon rank order test, and p values less than 0.05 were considered statistically significant.

FIG. 2. IL-18-induced RA synovial fibroblast Akt phosphorylation and inhibition by the PI3-kinase inhibitor LY294002 or PI3kinase antisense ODN.
A, serum-starved cells were stimulated for the indicated times with IL-18 (10 nM). Cell lysates were blotted for phospho-Akt (*p-Akt). The same blot was stripped and reprobed with a mouse anti-human tubulin antibody. B, blots were scanned and analyzed for quantification with the UN-SCAN-IT software. Band intensities for phospho-Akt were normalized to the corresponding band intensities for tubulin. Data from three RA donors were averaged and are represented as mean Ϯ S.E., expressed as fold increase with respect to nonstimulated (NS) cells; *, p Ͻ 0.05 versus time 0; n ϭ number of donors. C, RA synovial fibroblasts were serum-starved for 16 h before treatment with LY294002 (10 M) or Me 2 SO control. After 2 h of pretreatment with inhibitors, IL-18 was added to the culture media for the indicated times. Cell lysates were prepared and subjected to Western blot analysis for activation of Akt. D, blots were scanned and quantitated. Values for phospho-Akt were normalized to the corresponding values for tubulin. Data from three RA donors were averaged and are represented as the mean Ϯ S.E., expressed as the percentage of inhibition of Akt phosphorylation for cells pretreated with Me 2 SO for 10 and 20 min; *, p Ͻ 0.05 versus time 10 or 20 min; n ϭ number of donors. E, antisense PI3-kinase (PI3K) ODN inhibited RA synovial fibroblast Akt expression. RA synovial fibroblasts were treated with antisense (AS) PI3-kinase ODN or the corresponding control sense (S) ODN and incubated with or without IL-18 (10 nM) for 20 min. Total cell lysates were prepared for equal protein loading and separation by Western blot. Phospho-Akt was detected with mouse monoclonal anti-human phospho-Akt antibody. F, data from four RA donors were scanned and analyzed for quantification with the UN-SCAN-IT software, and band intensities for Akt were normalized to tubulin to quantitate relative Akt protein expression and presented as the mean Ϯ S.E.; *, p Ͻ 0.05; n ϭ number of donors; S, sense ODN; AS, antisense ODN. resulted in a 5 Ϯ 1.3-fold increase of PI3-kinase activity, which peaked at 10 min. This IL-18-induced increase in PI3-kinase activity declined rapidly at 20 min.
IL-18 Activates Phosphorylation of Akt through PI3-Kinase-The lipid products of PI3-kinase are known to activate serinethreonine protein kinase Akt through phosphorylation of serine 473 and threonine 308 (30). We therefore examined the ability of IL-18 to induce Akt phosphorylation by probing Western blots of proteins from IL-18-stimulated RA synovial fibroblasts using antibodies to phosphorylated Akt (serine 473). IL-18 induced Akt phosphorylation in a time-dependent manner (Fig. 2, A and B), and after 5 min there was a 4-fold increase in Akt activation, with maximum activation (11-fold) observed at 20 min (p Ͻ 0.05, n ϭ 3). This increase in Akt phosphorylation could be blocked when RA synovial fibroblasts were preincubated with the PI3-kinase inhibitor LY294002 (10 M) (Fig.  2, C and D). Pretreatment resulted in ϳ80 Ϯ 17% (mean Ϯ S.E.) decrease in Akt phosphorylation (p Ͻ 0.05, n ϭ 3). The role of PI3-kinase in Akt activation was more specifically demonstrated through inhibition with PI3-kinase AS ODN.
Antisense PI3-Kinase Inhibits RA Synovial Fibroblast Akt Expression-To examine involvement of the Akt pathway in IL-18 signaling through a more specific approach, antisense ODN targeting PI3-kinase and corresponding control sense ODN were used to treat RA fibroblasts with subsequent IL-18 (10 nM) stimulation. Levels of Akt expression were determined by Western blot of total cell lysates. IL-18 (10 nM) stimulated Akt protein expression from nearly undetectable basal levels. To confirm the above data and, more importantly, to demonstrate specificity of inhibition by antisense ODN, treatment with antisense PI3-kinase ODN reduced the level of Akt protein expression almost completely versus treatment with corresponding sense ODN, 100 Ϯ 17.92 to 1.03 Ϯ 0.11 (p Ͻ 0.05, n ϭ 4), which is consistent with the expectations for Akt as a target of PI3-kinase-activated signals (Fig. 2, E and F).
IL-18 Activates the Nonreceptor Protein Tyrosine Kinase c-Src-We next investigated signaling pathways involved in PI3kinase/Akt activation. Because PI3-kinase is generally activated by protein-tyrosine kinase (PTK), we examined PTKs activated by IL-18 using a mouse monoclonal antibody (4G10), which specifically reacts with tyrosine-phosphorylated proteins. IL-18 (10 nM)-induced protein tyrosine phosphorylation in RA synovial fibroblasts in a time-dependent manner, showing maximum activation at 10 and 20 min by Western blot. Two major bands were detected at 50 -75-and 35-50-kDa ranges (Fig. 3A). Proteins phosphorylated at tyrosine residues corresponding to molecular mass ranges of 60 -65 and 42-44 kDa are Src kinases and MAPK, respectively. Like PI3-kinase, Src kinases are located near the cell membrane, and their activation occurs relatively early in the signaling sequence (31). Moreover, it has been demonstrated that c-Src recruits and activates PI3-kinase in NIH3T3 fibroblasts (32). Thus, we examined the effect of IL-18 stimulation on Src activation by Western blot and found that IL-18 induces the phosphorylation of Src in a time-dependent manner (p Ͻ 0.05, n ϭ 3). As compared with nonstimulated RA synovial fibroblasts, a 4-fold increase in phosphorylated Src was observed at 10 and 20 min (Fig. 3, B and C).
IL-18 Activates Src Kinase in RA Synovial Fibroblasts-A direct role for IL-18 activation of Src was examined by an enzymatic assay. Confluent RA synovial fibroblasts were stimulated with IL-18 (10 nM) or control (media alone) for 20 min; their total lysates were collected and immunoprecipitated to isolate Src; and the Src kinase assay was performed with Src substrate peptide and [␥-32 P]ATP. Results represent fold increase in counts/min versus non-stimulated control cells. IL-18 (10 nM) significantly induced Src activation greater than a 4-fold over basal activation (4.35 Ϯ 1.92) (mean Ϯ S.E.) in the four different RA patient samples examined (p Ͻ 0.05) (Fig.  4A).
Antisense c-Src ODN Inhibits RA Synovial Fibroblast Phospho-Akt Expression-To examine the involvement of the Akt pathway in IL-18 signaling, antisense ODN targeting of the gene for c-Src and corresponding control sense ODN were used to treat cells with subsequent IL-18 (10 nM) stimulation. The level of phospho-Akt expression was determined by Western blot of total cell lysates. Again, IL-18 (10 nM) stimulated phospho-Akt protein expression from nearly undetectable basal (or non-stimulated) levels, similarly to the data in Fig. 2E. Antisense c-Src ODN caused a reduction in the level of phospho-Akt expression, apparently through specific reduction in c-Src (Fig.  4, B and C). Antisense c-Src ODN inhibited phospho-Akt expression by 75%, from 100 Ϯ 15.13 to 25.7 Ϯ 10.96 (mean Ϯ S.E.) (p Ͻ 0.05, n ϭ 4). This finding suggests that not only does IL-18-induced phospho-Akt expression occur through the activation of c-Src but that this phospho-Akt expression can be inhibited by c-Src blockade. With the knowledge that Akt is a downstream substrate of PI3-kinase activity, which was confirmed above to validate the use of antisense PI3-kinase ODN, these results showing Akt down-regulation via c-Src ODN treatment support the critical position of Src upstream of PI3-kinase/Akt. IL-18-induced ERK1/2 Activation Is PI3-Kinase-independent but Src-dependent-Other studies have shown a role for PI3-kinase in signaling ERK activation in a number of cell systems (33,34). Hence, we looked at the kinetics of ERK activation in IL-18-stimulated RA synovial fibroblasts by Western blot and found a time-dependent response, which was maximal at 20 min (p Ͻ 0.05, n ϭ 3) (Fig. 5, A and B). To determine whether ERK1/2 activation was dependent on Src kinases, we studied the effect of the Src kinase inhibitor PP2 on IL-18induced ERK1/2 activation. PP2 (25 M) strongly inhibited ERK1/2 activation (Fig. 5C). To assess the role of PI3-kinase in ERK activation by IL-18, cells were preincubated with LY294002, a potent inhibitor of PI3-kinase. LY294002 did not change IL-18-mediated ERK stimulation suggesting PI3-ki-

FIG. 4. IL-18 directly activates Src kinase in RA synovial fibroblasts, and antisense c-Src ODN treatment inhibits IL-18-induced Akt phosphorylation.
A, serum-starved RA synovial fibroblasts were stimulated for 10 min with IL-18 (5 nM). Cell lysates were prepared with RIPA buffer, immunoprecipitated with anti-Src antibody-conjugated agarose beads, and then assayed for Src kinase activity in the presence of peptide substrate, [␥-32 P]ATP and manganese/ATP mixture. Data from five RA donors were averaged and are represented as mean Ϯ S.E., expressed as fold increase relative to nonstimulated control; *, p Ͻ 0.05; n ϭ number of donors. B, antisense c-Src ODN inhibits Akt expression. RA synovial fibroblasts were treated with antisense (AS) c-Src ODN or the corresponding control sense (S) ODN and incubated with or without IL-18 (10 nM) for 20 min. Total cell lysates were prepared for equal protein loading and separation by Western blot. Akt was detected with mouse monoclonal anti-human phospho-Akt antibody (*p-Akt). C, data from four RA donors were scanned and analyzed for quantification with the UN-SCAN-IT software, and band intensities for Akt were normalized to tubulin to quantitate relative Akt protein expression and presented as the mean Ϯ S.E.; *, p Ͻ 0.05; n ϭ number of donors; S, sense ODN; AS, antisense ODN. nase is not involved in the signaling route leading to ERK activation (Fig. 5D).

IL-18 Activates the Ras/Raf Pathway-A number of studies
have shown Raf kinase is a predominant MEK kinase which, in turn, is activated by binding to Ras-GTP (35). Src family kinases have also been shown to be linked to MAPK through Ras/Raf kinases (36,37). To study the activation of Ras/Raf kinases, we first immunoprecipitated activated Ras with a Raf-1-Ras binding domain (RBD) conjugated to agarose beads and subsequently probed with mouse monoclonal anti-Ras antibody. IL-18 stimulation of RA synovial fibroblasts showed a time-dependent activation of the Ras/Raf pathway with maximal activation at 10 and 20 min (p Ͻ 0.05, n ϭ 3) (Fig. 6A). We confirmed activation of Raf by IL-18 using an antibody specific for phosphorylated Raf-1 (p Ͻ 0.05, n ϭ 3) (Fig. 6B).
IL-18 Activates AP-1 through ERK but Not PI3-Kinase-Because in our previous study inhibitors of the transcription factor NFB were only able to reduce IL-18-induced VCAM-1 expression by 50%, we explored whether other signaling pathways and transcription factors might be involved in induction of VCAM-1 transcription. AP-1 was an appropriate candidate because AP-1 is activated by PI3-kinase and MAPK and has also been involved in the induction of adhesion molecule gene transcription (38,39). By using gel shift assays, we show that DNA binding of AP-1 was potently induced by IL-18 at 1 h (Fig.  7A). The specificity of the AP-1-DNA binding complexes was

FIG. 5. IL-18-induced ERK1/2 phosphorylation in RA synovial fibroblasts, and inhibition by the Src inhibitor PP2 or the antisense c-Src ODN.
A, serum-starved RA synovial fibroblasts were stimulated with IL-18 for various time points as indicated. Cell lysates were blotted for phospho-ERK1/2 with mouse monoclonal anti-human phospho-ERK1/2 antibody (*p-ERK1/2). The same blot was stripped and reprobed with a mouse anti-human tubulin antibody. B, blots were scanned and analyzed for quantification with the UN-SCAN-IT software. Band intensities for phospho-ERK1/2 were normalized to the corresponding band intensities for tubulin. Data from three RA donors were averaged and are represented as the mean Ϯ S.E., expressed as fold increase with respect to nonstimulated (NS) cells. *, p Ͻ 0.05 versus time 0; n ϭ number of donors. C and D, synovial fibroblasts from four RA donors (indicated RA number above the blot) were serum-starved for 16 h before treatment with specific kinase inhibitors (25 M) or vehicle control Me 2 SO. After 2 h of pretreatment with inhibitors, IL-18 was added to the culture media for 20 min. Cells lysates were prepared and subjected to Western blot analysis for the activation of ERK1/2 (*p-ERK1/2). The same blots were stripped and reprobed with a mouse anti-human tubulin antibody. Blots were scanned and analyzed for quantification. Shown in C are the blots for the cells pretreated with PP2, and shown in D is the blot for cells pretreated with LY294002. E, RA synovial fibroblasts were treated with antisense (AS) ERK1/2 ODN or antisense c-Src ODN or the corresponding control sense (S) ODNs and incubated with or without IL-18 (10 nM) for 20 min. Total cell lysates were prepared for equal protein loading and separation by Western blot. ERK1/2 was detected with mouse monoclonal anti-human phospho-ERK1/2 antibody (*p-ERK1/2). F, data from four RA donors were scanned and analyzed for quantification with the UN-SCAN-IT software, and band intensities for ERK1/2 were normalized to tubulin to quantitate relative ERK1/2 protein expression and presented as the mean Ϯ S.E.; *, p Ͻ 0.05; n ϭ number of donors.
confirmed by the fact that their formation could be eliminated by the addition of a 4 M excess of unlabeled ODN (Fig. 7B). This AP-1 binding to DNA was completely inhibited in RA synovial fibroblasts pretreated with PD98059, but had no effect in cells pretreated with LY294002 or PDTC, and had a partial effect with PP2 pretreatment (Fig. 7C). PDTC served as a control due to its known absence of effect on AP-1 (40).
Antisense IRAK-1 ODN Inhibits IRAK-1 Gene Expression-Initial studies involved specificity determination to validate that the antisense ODN could inhibit expression of its corresponding protein. To verify that antisense IRAK-1 ODN inhibits target gene expression, IRAK-1 protein in ODN-treated cells stimulated with IL-18 (10 nM) was detected by Western blot analysis. The expression of tubulin was examined as the housekeeping protein, and optical density measurement of the band representing the protein of interest was normalized to that of tubulin to quantitate relative protein expression (Fig. 8). Antisense IRAK-1 ODN inhibited IL-18-induced IRAK-1 protein expression. Western blot analysis showed 76% inhibition of IRAK-1 protein expression by antisense IRAK-1 ODN, whereas sense IRAK-1 ODN was unable to block IRAK-1 protein expression in IL-18-stimulated cells, 24.11 Ϯ 1.23 versus 100 Ϯ 18.1 (p Ͻ 0.05, n ϭ 4). To demonstrate further the relative specificity of the antisense IRAK-1 ODN, as a control, cells were also treated with antisense or sense c-Src ODN. c-Src ODNs were also unable to block IRAK-1 protein expression, which was comparable with control (no ODN). As shown, IL-18 clearly up-regulated IRAK-1 expression, whereas basal levels by non-stimulated control cells were not readily detectable in these Western blots.
Antisense IRAK-1 ODN Inhibits RA Synovial Fibroblast IL-18-induced NF-B Activation-IL-18 (10 nM) significantly stimulated NF-B activation from basal levels of 25.7 Ϯ 3.4 to 100 Ϯ 7.5, a 4-fold increase. This IL-18-induced NF-B activation was inhibited by antisense IRAK-1 ODN treatment for 5 h. The expression of NF-B was decreased by 32.9% from control expression of 100 Ϯ 7.5 to 67.12 Ϯ 6.6 (p Ͻ 0.05, n ϭ 4) (Fig. 9). There was no difference between antisense and sense ODN nonstimulated control cells. Again, as a control, cells were also treated with antisense or sense c-Src ODN, neither of which was able to block IL Cells were extracted with cell lysis buffer, and the protein content in each sample was estimated. Each sample (100 g) was immunoprecipitated with a Raf-1-RBD-agarose conjugate and subsequently subjected to 10% SDS-PAGE. A, the presence of activated Ras in the samples was then detected by probing with a specific mouse monoclonal anti-Ras antibody (1 g/ml). C, phospho-c-Raf (*p-raf-1) was detected by direct Western blot analysis. Blots were stripped and reprobed with mouse anti-human tubulin. B and D, in each case, blots were scanned and analyzed for quantification, and data from three RA donors were averaged and are represented as the mean Ϯ S.E., expressed as fold increase with respect to nonstimulated (NS) cells. *, p Ͻ 0.05 versus time 0; n ϭ number of donors.
(PP2) at 25 M significantly decreased IL-18-induced VCAM-1 expression by 32 and 35% (0.329 Ϯ 0.06 OD for Me 2 SO versus 0.223 Ϯ 0.04 OD for LY294002, and 0.212 Ϯ 0.03 OD for PP2, each value representing the mean Ϯ S.E., p Ͻ 0.05). The inhibition was greater with 50 M of these inhibitors, 43 and 46%, respectively (data not shown). Interestingly, an inhibitor of ERK (PD98059) at these concentrations did not have an effect on IL-18-induced VCAM-1 expression measured by cell ELISA, which appeared to suggest the lack of ERK involvement, but results of inhibition with antisense ODN described below showed otherwise. Because studies have suggested that G proteins can activate PI3-kinase and Src (41,42), we also examined the potential role of G proteins in IL-18-induced VCAM-1 expression. RA synovial fibroblasts were pretreated with the G o/i protein inhibitor, pertussis toxin, before stimulation with IL-18. The expression of VCAM-1 on RA synovial fibroblasts was measured by FACS analysis and was not different for cells in the presence or absence of pertussis toxin (data not shown).

DISCUSSION
Because the PI3-kinase inhibitor partially blocked IL-18induced VCAM-1 expression, we hypothesized IL-18 could stimulate PI3-kinase activity. Indeed, IL-18 potently stimulated PI3-kinase activity with rapid kinetics. The serine-threonine protein kinase Akt is one of the major targets of PI3kinase-generated signals (43). In our experiments, we showed that Akt becomes phosphorylated after IL-18 stimulation and that the PI3-kinase-specific inhibitor LY294002 could inhibit such stimulation. This was confirmed by nearly complete abrogation of IL-18-induced Akt activation through antisense PI3-kinase ODN treatment. These results demonstrate IL-18 is a novel activator of the PI3-kinase/Akt signaling cascade. PI3kinase activation was first described in signaling of hormones and growth factors (11,44). Most recently, Reddy et al. (45,46) reported that PI3-kinase was also involved in signaling of inflammatory cytokines such as IL-1 and TNF-␣. In comparison to these cytokines, the magnitude and kinetics of induction of PI3-kinase by IL-18 are more comparable with those observed for TNF-␣ than IL-1.
Receptor and non-receptor tyrosine kinases and G proteincoupled receptors are known to be main activators of PI3kinase (11). There are a number of reports demonstrating the role of NRTK Src kinase in PI3-kinase activation (47,48). We showed IL-18 induced a time-dependent increase in Src kinase phosphorylation. In addition, IL-18 directly activated Src kinase as shown by enzymatic assay. Members of the Src family NRPKs are critically involved in triggering a number of intracellular signaling pathways. Affected pathways are involved in cell migration, adhesion, cell cycle, apoptosis, and phagocytosis (49,50). Src has been implicated in a variety of diseases, such as breast, liver and colon cancer, especially those with liver metastasis (51)(52)(53). Some Src PTKs have tissue-restricted expression, which would provide some advantage as therapeutic targets in inflammatory, autoimmune, allergic, and malignant disease. Of note, our data showed a trend toward a greater degree of inhibition of surface VCAM-1 expression by blocking IL-18-stimulated RA synovial fibroblasts with c-Src antisense ODN relative to blockade of other signaling molecules. This difference may suggest a more prominent role for Src in IL-18 signaling and VCAM-1 expression, although the ability of the various ODNs to inhibit the target gene of interest may be having more influence than would allow for comparisons of relative roles. Inhibition of Src and Src-dependent pathways may present potentially useful approaches for treatment of RA and diseases characterized by adhesion molecule up-regulation.
Interestingly, IL-18-induced Src phosphorylation was temporally correlated with PI3-kinase and Akt activation. Experiments using antisense c-Src ODN further confirmed the upstream position of Src in this pathway. Our data indicate that the initial event in the IL-18 response is Src activation, which leads to activation of PI3-kinase and ERK1/2. Our observations appear to support this hypothesis because the Src kinase inhibitor PP2 strongly inhibits IL-18-induced ERK activation, and inhibition with c-Src antisense ODN significantly reduced ERK1/2 expression. Moreover, the Src inhibitor PP2 and the PI3-kinase inhibitor LY294002 have very similar effects on IL-18-induced VCAM-1 expression, suggesting that Src and PI3-kinases belong to the same signaling cascade. However, ERK1/2 and PI3-kinase/Akt are probably independent of each other, based on the failure of the PI3-kinase inhibitor to alter IL-18-mediated ERK1/2 activation.
Our results are consistent with several lines of evidence suggesting that the ERK1/2 and PI3-kinase signaling pathways do not cross-talk (54, 55). Kalina et al. (9) reported previously that IL-18 activates ERK1/2 in the natural killer cell Lysates were prepared and analyzed as described in previous figures. NF-B was detected with rabbit polyclonal anti-human NFB p65 antibody. Shown is a representative Western blot, which was stripped and reprobed with mouse anti-human tubulin antibody. B, data from four RA donors were scanned and analyzed for quantification with the UN-SCAN-IT software, and band intensities for NF-B were normalized to tubulin and presented as the mean Ϯ S.E.; *, p Ͻ 0.05; n ϭ number of donors. line 92, but the underlying mechanism was unclear. Here we demonstrate IL-18 stimulates ERK1/2 through an Src/Ras/Raf pathway, as evidenced by decreased ERK1/2 expression by c-Src inhibition with antisense ODN. Src and PI3-kinases may not be activated by G proteins, based on the fact that the G o/i protein inhibitor pertussis toxin has no effect on IL-18-induced VCAM-1 expression. These data suggest Src and G protein are independent of each other. It is more likely that Src may recruit and activate PI3-kinase via tyrosine phosphorylation of IL-18R. Indeed, Src has already been shown to phosphorylate receptors such as epidermal growth factor receptor (56), and it is known that phosphorylated receptors like IL-1 are able to recruit PI3-kinase (57). Further investigations are needed to clarify possible interactions between activation of Src, PI3kinase, and the IL-18R following IL-18 stimulation. Furthermore, the IL-18R␤ chain has been shown to be essential for IL-12-independent proinflammatory activity of IL-18-induced IL-8 in COS-1 fibroblasts (58), suggesting the IL-18 receptor itself is potentially an important regulatory site in signaling. The IL-18R components, IL-18R␤ (previously IL-1R-related protein, IL-1Rrpl) and IL-18R␣ (or IL-1R accessory proteinlike, IL-1RacPL), confer a highly specific and unique IL-18 responsiveness (59). Both have been cloned and shown to be critical for IL-18 signaling (60,61), and their potential interactions with relevant signaling molecules are of interest.
We next investigated downstream events of PI3-kinase/Akt and ERK1/2 activation. Our previous work (4) suggested that transcription factors other than NFB existed that may be involved in IL-18-induced VCAM-1 expression. AP-1 was of particular interest for three reasons. First, AP-1 is activated by PI3-kinase and ERK (45,62,63). Second, AP-1 regulates the transcription of adhesion molecules (38,39). Third, AP-1 is an important transcription factor in RA synovial fibroblasts (64). The data showing that IL-18 induces AP-1 binding activity made this hypothesis plausible and is in accord with previous studies reporting AP-1 activation by IL-18 in different cells types (17,65,66). Although the EMSAs only demonstrate AP-1 binding and cannot provide information on its phosphorylation status, our data using specific inhibitors suggest that, unlike PI3-kinase, ERK1/2 regulates AP-1 activation in IL-18-stimulated synovial fibroblasts. For PI3-kinase, the signaling event downstream of Akt remains unanswered. STAT3 might be an interesting candidate because IL-18, c-Src, and PI3-kinase have all been shown to stimulate STAT3 (9,67,68). Interestingly, the nuclear translocation of STAT4 is strictly dependent on the presence of IL-18 for IL-12/IL-18-induced IFN-␥-mediated nitric oxide production by mouse peritoneal macrophages (69).
Finally, we showed that PI3-kinase is involved in IL-18induced VCAM-1 expression. In the literature, there are conflicting reports regarding the effect of PI3-kinase on adhesion molecule expression. Radisavljevic et al. (70) showed that vascular endothelial growth factor up-regulates ICAM-1 expression via a PI3-kinase/Akt pathway in endothelial cells. In contrast, using the same stimulus and a similar cell type, Kim et al. (71) reported an up-regulation of ICAM-1, VCAM-1, and E-selectin expression by PI3-kinase inhibition. The experimental conditions including the different concentrations and types of PI3-kinase inhibitors used by these and other authors may explain this discrepancy. In our hands, blocking PI3-kinase with LY294002 had a repressive effect on IL-18-induced RA synovial fibroblast VCAM-1 expression using two different techniques (FACS and cell ELISA). More specific inhibition with PI3-kinase antisense ODN treatment was confirmatory. These results are consistent with the known properties of PI3kinase and VCAM-1. Indeed, we and others (13,14,72,73) have implicated these two molecules in cell adhesion and angiogenesis. Moreover, we showed in recent reports (4) that IL-18 mediates leukocyte-fibroblast adhesiveness and stimulates endothelial cell tube formation in vitro and in vivo (3). Hence, it is possible that PI3-kinase may be a key signaling molecule responsible for IL-18-induced VCAM-1 synthesis, and subsequently cell adhesion and angiogenesis.
As with PI3-kinase, selective ERK1/2 inhibition with antisense ODN treatment decreased IL-18-induced VCAM-1 expression. However, ERK inhibitor PD98059 did not show significant inhibition, which perhaps was due to the lack of ERK inhibition at the concentrations used without affecting viabil- ity. Thus, the use of antisense ODN to block signaling molecules specifically may be more appropriate for investigating these pathways and for developing potential therapeutic strategies, at least in the case of ERK1/2. In addition, antisense ODNs may be an alternative to avoid the toxicity that prevents these chemical inhibitors from becoming optimal therapeutics. Taken together, our observations support the perspective that specifically targeting PI3-kinase (71), or perhaps ERK1/2, may represent a valid approach to blocking adhesion molecule expression, cell adhesion, and angiogenesis. Furthermore, that c-Src has a signaling role in IL-18-induced VCAM-1 expression is consistent with the findings that c-Src activation by IL-18 lies upstream to the induction of PI3-kinase and ERK1/2, both of which we showed to affect VCAM-1 expression.
Additionally, the IRAK/NFB pathway, distinct from the Src-dependent pathways, appears to play a significant role in RA synovial fibroblast VCAM-1 expression. Indeed, antisense IRAK ODN treatment inhibited IL-18-induced VCAM-1 expression. This finding is in accordance with the fact that the IRAK/NFB signaling pathway has been implicated in various IL-18 functions. IL-18 binding to its receptor induces IRAK recruitment, and IRAK activation leads to nuclear translocation of NFB and subsequent expression of immune response genes, such as IFN-␥ in Th1 cells (20). This NFB activation is diminished in fibroblasts from IRAK-deficient mice, although additional IL-1 and IL-18 cytokine responses are independent of the IRAK pathway (21), which could explain Src-mediated pathways. IRAK is reportedly necessary for NFB and JNK activation in response to IL-18, as in the mutant cell line I1A generated to lack IRAK protein; specifically, the N-proximal region of IRAK is essential for NFB activation but not for JNK activation, suggesting IRAK may serve as a branch point (74). Guo and Wu (75) reported inhibition of IL-18-induced NFB activation by blocking PI3-kinase with antisense ODN. Whether IRAK has a regulatory role in uniting these pathways is one question to be clarified. Impaired NFB and JNK activation in IRAK-deficient mice is further evidence of an important role for IRAK in IL-18 signaling (76). Upon IL-18 stimulation, in addition to IRAK recruitment and phosphorylation, TRAF-6 associates with IRAK in mouse EL-4 cells (77). Thus, the interplay of other key molecules with IRAK may also be critical for regulation of IL-18 functions.
There is additional precedence for the involvement of transcription factor activation in IL-18-mediated functions. For instance, IL-18 activates AP-1 and NFB in Jurkat T cells leading to IL-2 expression (17). The significance of the NFB pathway in the regulation of IL-18-induced IFN-␥ production was shown by Kojima et al. (18) in the human myelomonocytic cell line KG-1 cells. The essential role of NFB activation in IL-18-induced IFN-␥ production was also shown in murine Th1 cells (19). Again, the possibility that ligand-binding receptor components may be critical for IL-18-mediated signaling is also of interest, given that IL-18R␤ is necessary for NFB and JNK activation, based on studies with Th1-developing splenic CD4 ϩ T cells derived from IL-18R-deficient mice (60).
In summary, we describe three distinct pathways activated by IL-18: the Ras/Raf-1/ERK/AP-1, PI3-kinase/Akt, and the IRAK/NFB pathways. Furthermore, Src kinase appears to be the initial event common to the first of these two novel pathways (Fig. 11). We also demonstrate that all three described pathways control IL-18-induced VCAM-1 expression. The potential benefit of antisense ODN as selective inhibitors in down-regulating cellular signaling events is also supported by the significant decrease in IL-18-induced VCAM-1 expression seen after antisense ODN treatment. Although modulation of cellular signaling events by chemical inhibi-tors or antisense ODNs to provide insight to pathway is possible, the resultant downstream cellular effect is most relevant. Given the critical role of VCAM-1 in the initiation of leukocyte migration and angiogenesis, it is possible that targeting signaling events important for VCAM-1 expression can have application in inflammatory and angiogenesisdriven diseases. Through involvement in IL-18 signaling and induction of inflammatory mediators like VCAM-1, Src PTKs, PI3-kinase, ERK1/2, and transcription factors such as NFB and AP-1 may serve as excellent therapeutic targets in the treatment of RA and conditions characterized by adhesion molecule up-regulation.