Chemokine-Cytokine Cross-talk

It is well established that cytokines can induce the production of chemokines, but the role of chemokines in the regulation of cytokine expression has not been fully investigated. Exposure of rat cardiac-derived endothelial cells (CDEC) to lipopolysaccharide-induced CXC chemokine (LIX), and to a lesser extent to KC and MIP-2, activated NF-κB and induced κB-driven promoter activity. LIX did not activate Oct-1. LIX-induced interleukin-1β and tumor necrosis factor-α promoter activity, and up-regulated mRNA expression. Increased transcription and mRNA stability both contributed to cytokine expression. LIX-mediated cytokine gene transcription was inhibited by interleukin-10. Transient overexpression of kinase-deficient NF-κB-inducing kinase (NIK) and IκB kinase (IKK), and dominant negative IκB significantly inhibited LIX-mediated NF-κB activation in rat CDEC. Inhibition of Gi protein-coupled signal transduction, poly(ADP-ribose) polymerase, phosphatidylinositol 3-kinase, and the 26 S proteasome significantly inhibited LIX-mediated NF-κB activation and cytokine gene transcription. Blocking CXCR2 attenuated LIX-mediated κB activation and κB-driven promoter activity in rat CDEC that express both CXCR1 and -2, and abrogated its activation in mouse CDEC that express only CXCR2. These results indicate that LIX activates NF-κB and induces κB-responsive proinflammatory cytokines via either CXCR1 or CXCR2, and involved phosphatidylinositol 3-kinase, NIK, IKK, and IκB. Thus, in addition to attracting and activating neutrophils, the ELR+ CXC chemokines amplify the inflammatory cascade, stimulating local production of cytokines that have negative inotropic and proapoptotic effects.

Chemokines are small molecular weight cytokines involved in activation of specific subsets of immune cells and their recruitment to the site of injury and inflammation (1)(2)(3)(4)(5)(6)(7). They are classified into C, CC, CXC, and CX3C families (2)(3)(4). In the CXC family, the first two conserved cysteines are separated by one nonconserved amino acid (X in CXC). CXC chemokines that have a glutamic acid-leucine-arginine (ELR) sequence immediately preceding the CXC motif are potent neutrophil chemoattractants (ELR ϩ CXC chemokines) (2,8,9). Neutrophil migration is stimulated by a gradient of these chemokines from blood toward the site of inflammation.
We have recently shown up-regulation of the ELR ϩ CXC chemokines, LIX 1 (lipopolysaccharide-induced CXC chemokine; CXCL5), KC (cytokine-induced neutrophil chemoattractant; CXCL1), and MIP-2 (macrophage inflammatory protein-2; CXCL2) in a rat model of myocardial ischemia/reperfusion injury (10). High levels of myocardial neutrophil infiltration coincided with peak levels of LIX and MIP-2 expression. Neutralization of LIX, KC, and MIP-2 inhibited myeloperoxidase activity, a measure of neutrophil infiltration, by 79, 28, and 37%, respectively, indicating that LIX may be the predominant neutrophil chemoattractant in this model of reperfusion injury (10). Furthermore, the proinflammatory cytokine expression preceded the chemokine expression in this model, suggesting that chemokine expression was a downstream effect of cytokine production (10). This was confirmed in in vitro studies where exposure of cardiomyocytes to TNF-␣ induced LIX expression via NF-B activation (10).
NF-B is a ubiquitous, multisubunit, inducible transcription factor that regulates the expression of various genes involved in the immune and inflammatory processes (11,12). The p50/ p65 heterodimer, which has been most studied, resides in the cytoplasm in an inactive state because of binding of p65 to an inhibitory subunit IB. The IB family, including IB-␣, IB-␤, IB-␥, IB-⑀, all prevent activation and subsequent nuclear translocation of the heterodimer. Various stimuli including cytokines, growth factors, and oxidative stress induce IB hyperphosphorylation leading to its selective degradation in the cytoplasm by the ubiquitin-26 S proteasome system, resulting in NF-B activation (11,12).
A multiprotein complex comprised of IKK (IB kinase)-␣, IKK-␤, and a regulatory subunit IKK-␥/NEMO was shown to mediate phosphorylation of IB by various cytokines (13)(14)(15)(16)(17). The cytokine-initiated signal transduction cascade leading to IB phosphorylation has been shown to converge at activation of the IKK by NF-B-inducing kinase (NIK). NIK associates with IKK-␥ and activates the IKK signalsome. PI 3-kinase, PI-phospholipase C, protein kinase C, and p38 mitogen-activated protein kinase were implicated as upstream regulators of NIK and IKK. Furthermore, poly(ADP-ribose) polymerase 1 (PARP-1), a nuclear protein involved in DNA repair, has been shown to physically and functionally associate with NF-B in the nucleus and modulate NF-B-dependent cytokine gene transcription (18,19). The role of these various regulatory subunits in chemokine-mediated NF-B activation and cytokine gene transcription has not been investigated.
Whereas the agonistic effects of cytokines on chemokine expression are well described, very little is known about chemokine-mediated cytokine expression. In the present study we investigated the role of the ELR ϩ CXC chemokines, LIX, KC, and MIP-2, in NF-B activation and induction of IL-1␤ and TNF-␣ expression. Furthermore, we explored the chemokine receptor usage and signal transduction pathway involved in chemokine-mediated NF-B activation. Our results indicate that the ELR ϩ CXC chemokines activate NF-B, induce proinflammatory cytokine expression, and signal through CXCR2, and presumably also through CXCR1.

Materials and Reagents
Recombinant mouse LIX, KC, and MIP-2 were obtained from Pepro-Tech, Inc. (Rocky Hill, NJ). Recombinant carrier-free rat IL-1␤ and rat TNF-␣ were from R&D Systems (Minneapolis, MN). The recombinant proteins contained Ͻ1 ng of endotoxin per g of protein.

Transient Cell Transfections and Reporter Assays
The NF-B driven luciferase reporter plasmid (pNF-B-Luc) was obtained from Stratagene (La Jolla, CA) and contains five copies of NF-B consensus sequence linked to the minimal E1B promoter-luciferase reporter gene. pEGFP-Luc was used as a control. The phosphorylation-deficient S32A/S36A mutant of IB-␣ (pCMX-IB-␣(S32A/ S36A)) was a gift from Inder Verma (The Salk Institute, La Jolla, CA), and the Myc-tagged phosphorylation-deficient S19A/S32A mutant of IB-␤ in pCMV-Tag3B (Stratagene) has been described earlier (22). Kinase-deficient NIK (pRK7-NIK(KK429 -430AA)-Flag), IKK-␤ (pRK5-IKK-␤-Flag), and dominant negative IKK-␥ (pcDNA3-IKK-␥-HA) were obtained from David V. Goeddel (Tularik Inc., South San Francisco, CA), Tom Maniatis (Harvard University, Cambridge, MA), and Gabriel Nunez (University of Michigan Medical School, Ann Arbor, MI), respectively. Rat CDEC were plated on six-well tissue culture dishes and transfected the following day at ϳ70 -80% confluency using Lipo-fectAMINE 2000 TM (Invitrogen, Carlsbad, CA) as described by the manufacturer. pRL Renilla-luciferase reporter gene (100 ng; pRL-TK vector; Promega, Madison, WI) was used as an internal control. The empty vectors pCMX, pCMV-Tag3B, pRK5, pRK7, and pcDNA3 were used as controls. Data were normalized for transfection efficiency by dividing firefly luciferase activity with that of corresponding Renilla luciferase, and expressed as mean relative stimulation Ϯ S.E. for a representative experiment from three separate experiments, each performed in triplicate. The amount of DNA transfected was kept constant (2 g) in all transfection experiments. After transfection, the cells were found to be viable (trypan blue dye exclusion). 24 h after transfection, the media was changed, and the cells were exposed to LIX, KC, or MIP-2 at the indicated concentrations and for the specified time periods. Cell extracts were prepared, and luciferase activity was determined with a TD 20/20 luminometer (Turner Designs, Sunnyvale, CA) using the Promega Biotech TM dual-luciferase reporter assay system. Transfection efficiency was determined by transfecting rat CDEC with pEGFP-N1 vector (Clontech, Palo Alto, CA) that constitutively expresses the enhanced green fluorescent protein (EGFP) under the regulation of CMV promoter and enhancer. Once the cells reached ϳ70% confluency, the media was replaced with M199 ϩ 0.5% BSA. After overnight culture, cells were transfected with pEGFP-N1 and LipofectAMINE 2000. 24 h later, the cells were trypsinized, seeded onto Lab-Tek II chamber slide (Nunc TM ), and cultured for an additional 48 h. Cells were then washed in PBS (pH 7.4), fixed in 4% paraformaldehyde in PBS for 30 min at room temperature. After washing in PBS, coverslips were mounted using ProLong TM Antifade kit (Molecular Probes, Eugene, OR). After the mounting media was dried, the coverslips were sealed with black nail polish, and stored at 4°C in the dark. The cells were visualized by a fluorescent microscope (Nikon Eclipse TE200, Nikon Inc., Melville, MA), and 1,000 cells were counted under ϫ20 objective, and bright to very-bright green fluorescent cells were considered positive for the expression of EGFP, and the others as nontransfected (controls). The transfection efficiency varied between 37 and 46% with an average of 38.1 Ϯ 2.9%. To determine the role of CXC receptors in LIX-mediated NF-B activation, rat and mouse CDEC were treated with SB447232 (GlaxoSmithKline Beecham), a specific CXCR2 antagonist, for 10 min before the addition of LIX.

Interleukin-1␤ and TNF-␣ Promoter Analyses
Murine IL-1␤ Promoter-The murine IL-1␤ promoter (Ϫ4093 to ϩ45) construct in pBluescript vector was a kind gift from Clifford J. Bellone This promoter construct has been demonstrated to confer a strong responsiveness to lipopolysaccharide (27). Rat CDEC were transfected with 3 g of either the IL-1␤ Ϫ4093 to ϩ45-CAT (chloramphenical acetyltransferase) or a mock plasmid that contains CAT reporter gene alone (pFR-CAT; Stratagene). To compensate for variations in transfection, cells were cotransfected with a ␤-galactosidase reporter construct (pSV-␤-galactosidase control vector, Promega) in which the SV40 early promoter and enhancer drives transcription of the lacZ gene, which encodes the ␤-galactosidase enzyme. 24 h later, the media was changed, and the cells were treated with LIX (100 ng/ml), neutralized LIX, or vehicle. Seven hours later the cells were processed for CAT levels using a CAT enzyme-linked immunosorbent assay kit (Roche Molecular Biochemicals) and ␤-galactosidase levels by using the ␤-galactosidase assay kit (Invitrogen) essentially as described by the manufacturers.
Murine TNF-␣ Promoter-Rat CDEC were transfected with pTNFϪ1080/ϩ138, pTNFϪ85/ϩ135, or empty vector (pGL3 basic). These murine TNF-␣ promoter reporter constructs were described earlier (28). The 1.1-kb TNF-␣ promoter construct (Ϫ1080/ϩ138 nucleotides; relative to the transcription start site) contains 4 NF-B response elements, and has been demonstrated to confer responsiveness to a variety of stimuli including lipopolysaccharide. Its deletion mutant construct (Ϫ85/ϩ135) lacks all the four NF-B response elements but contains only TATA box and Sp1 site, and responds poorly to lipopolysaccharide (28). The cells were co-transfected with pRK-Renilla to compensate for transfection efficiency. 24 h after transfection, the media was changed, and the cells were treated with LIX, neutralized LIX, or vehicle. Seven hours later the cells were processed for luciferase activity by the dual luciferase assay kit.
mRNA Stability-Actinomycin D Pulse-Rat CDEC were cultured in M199 medium containing 10% fetal calf serum. At 70 -80% confluency, the complete medium was replaced with M199 ϩ 0.5% BSA, and cul-tured for an additional 16 h. The cells were then treated with either LIX (100 ng/ml) or vehicle (control) for 4 h. Actinomycin D (5 g/ml; Sigma), a potent inhibitor of RNA polymerase II-dependent transcription, was then added. At the indicated time periods 1.5, 3, 4.5, and 6 h), cells were harvested for total RNA isolation. RNA was isolated using TRIzol reagent, and analyzed by Northern blot hybridization to quantitate IL-1␤ and TNF-␣ mRNA levels as described above.

LIX-mediated IL-1␤ and TNF-␣ Transcription (Nuclear Run-on)
After treating rat CDEC with LIX (100 ng/ml) for 4 h in M199 medium containing 0.5% BSA, nuclei were isolated, counted in a hemocytometer, and resuspended (2 ϫ 10 8 /ml) in a storage buffer (50 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 0.1 mM EDTA, 2 mM dithiothreitol, 40% glycerol) as described in detail previously (29). The nuclei were aliquoted and snap frozen in methanol/dry ice bath and stored in liquid N 2 until further use. For labeling RNA, nuclei were thawed (100 l), mixed with equal volumes of labeling mixture (200 mM KCl, 8 mM MgCl 2 , 1 mM each of ATP, UTP, and CTP, 100 M GTP) and 100 Ci of [␣-32 P]UTP (800 Ci/mmol). The mixture was incubated at 30°C for 30 min and 20 l (4 g) of RQ1 RNase-free DNase I (Promega Corp.) was added and incubated for an additional 10 min. After digestion with proteinase K (60 g in 20 l) in a buffer containing 1% SDS, 50 mM Tris-HCl (pH 7.0), 50 mM EDTA for 30 min at 42°C, it was subjected to phenol/chloroform/ isoamyl alcohol and chloroform extractions. The aqueous phase was ethanol-precipitated in the presence of 20 g of carrier RNA (Escherichia coli transfer RNA, RNase-free, Roche Molecular Biochemicals). The pellet was dissolved in 80 l of STE buffer (100 mM NaCl, 20 mM Tris-HCl, pH 7.5, 10 mM EDTA), and the unincorporated label was removed using NucTrap TM probe purification columns (Stratagene), and the incorporated radioactivity was determined in a scintillation counter.
Equal amounts of plasmid vectors containing IL-1␤, TNF-␣, glyceraldehyde-3-phosphate dehydrogenase (GenBank TM accession number M17701; 339-bp product, sense: 5Ј-TCCGCCCCTTCCGCTGATG-3Ј (bases 388 -406), antisense: 5Ј-CACGGAAGGCCATGCCAGTGA-3Ј (bases 707-727)), or empty plasmid (pCR TM 2.1-TOPO TM vector) were alkaline denatured, applied to nitrocellulose membranes using a slotblot apparatus (HYBRI-SLOT TM MANIFOLD, Invitrogen). After fixing the DNA to the membranes by UV cross-linking, prehybridization was performed at 52°C overnight, followed by hybridization for 3 days with ϳ10 6 cpm/ml of labeled RNA. The filters were then washed three times for 10 min each in 2ϫ SSC plus 0.1% SDS, two times in 0.1ϫ SSC plus 0.1% SDS at 65°C for 15 min each, and then treated with RNase A (10 mg/ml) for 30 min at 37°C in 2ϫ SSC. Finally the membranes were washed with 2ϫ SSC at 37°C for 30 min, and subjected to autoradiography, and the visualized bands were semiquantitated by densitometry.

Protein Extraction and Western Blot Analysis
30 g of cell extract in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml each of aprotinin, leupeptin, and pepstatin, 1% Nonidet P-40, 1 mM sodium orthovanadate, and 1 mM NaF) from untreated (control) and treated CDEC were subjected to SDS-PAGE under reducing conditions, and electrotransferred onto polyvinylidene difluoride membranes (Millipore, MA). Nonspecific sites were blocked with 10% normal goat serum (preimmune; DAKO) for 1 h at room temperature, drained, and incubated overnight at 4°C with the primary antibody in TBST (Trisbuffered saline containing 0.5% Tween 20) containing normal goat serum, washed in TBST, and incubated further for 1 h with the secondary antibody conjugated to horseradish peroxidase. After extensive washings with TBST, the membranes were incubated with an enhanced chemiluminescence reagent (Amersham Biosciences). The membranes were then washed, exposed to Kodak X-Omat AR film, and the autoradiographic bands were semiquantified and normalized to ␤-actin levels (10,26).

Enzyme-linked Immunosorbent Assay
TNF-␣ (sensitivity 0.7 pg/ml) and IL-1␤ (sensitivity Ͻ3.0 pg/ml) levels in culture supernatants were measured by enzyme-linked immunosorbent assay using commercially available kits (BIOSOURCE International; Ref. 30). Studies were performed as per the manufacturer's instructions.

Measurement of PI3K Activity
PI3K lipid kinase assays were performed essentially as described by Foukas et al. (31). After overnight incubation in 0.5% BSA, M199 media, rat CDEC were treated with rLIX (100 ng/ml) for 5 min with and without LY 294002. Cleared cell lysates were prepared by centrifugation at 10,000 ϫ g for 30 min at 4°C, and protein concentration was determined. Equal amounts of protein was immunoprecipitated with affinity purified antibodies against the p85 regulatory subunit of PI3K (Santa Cruz Biotechnology, Inc., number sc-423) for 2 h followed by protein A-Sepharose (Amersham Biosciences) for 1 h at 4°C. After washing the immunoprecipitates (IP) in Tris-HCl (100 mM, pH 7.4) containing 0.5 M LiCl and kinase assay buffer (2ϫ, 100 mM HEPES-NaOH, pH 7.4, 200 mM NaCl, 2 mM dithiothreitol), the immunoprecipitates were resuspended in 50 l of 1ϫ kinase assay buffer containing 5 mM MgCl 2 , 100 M ATP (plus 0.1 Ci of [␥-32 P]ATP/assay), and 200 g/ml phosphatidylinositol as a substrate. The reaction was incubated at 25°C for 20 min. The reaction was stopped by the addition of 100 l of 0.1 M HCl and 200 l of chloroform/methanol (1:1). The lower organic phase containing phospholipids was recovered and spotted on silica gel thin-layer chromatography plates (Gel-60, Merck), impregnated with 1% (w/v) potassium oxalate, 1 mM EDTA in water/methanol (6:4), and developed in a mixture of chloroform, methanol, 4 M NH 3 (9:7:4). The radioactivity on the dried plate was visualized and quantified by autoradiography and densitometry.

Measurement of Intracellular Calcium
Intracellular calcium measurements were made in rat CDEC using the calcium-sensitive probe Fura-2/AM (Molecular Probes). The cells were loaded with Fura-2/AM pentapotassium salt (5 M) in M199 medium supplemented with 10% fetal calf serum. After incubation for 45 min at 37°C, the cells were washed and resuspended at 2 ϫ 10 6 cells/ml in 137 mM NaCl, 4.5 mM KCl, 1.2 mM MgCl 2 ⅐7H 2 O, 4.9 mM KCl, 1.2 mM NaH 2 PO 4 , 20 mM HEPES, 15 mM D-glucose, 1.8 mM CaCl 2 (pH 7.4). The cell suspension was placed in a fluorimetry cuvette and stirred continuously at 37°C. After equilibrating at 37°C for 10 min, rLIX (100 ng/ml), neutralized LIX, or PBS were added, and fluorescence was monitored at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm on a Hitachi F-2000 fluorescence spectrophotometer; the results were calculated as the ratio of emission following excitation at 340 nm with that produced by excitation at 380 nm.
The Fura-2/AM-loaded cells were treated with either Triton X-100 (1%) or 100 mM EGTA to obtain maximal and minimal fluorescence, and the data were normalized as a percentage of the maximal fluorescence.

Inhibition of Radioligand Binding and Calcium Mobilization
Procedures utilized for 125 I-IL-8 binding to membranes of Chinese hamster ovary cells stably expressing CXCR1 and CXCR2 were done as previously described (32,33). Inhibition of 10 nM IL-8-induced calcium mobilization in RBL 2H3 cells stably expressing CXCR1 or CXCR2 was done as previously described (33). Inhibition of calcium mobilization with human polymorphonuclear leukocytes was done using 1 nM IL-8 or 10 nM GRO␣ as described (32). In addition, the same procedure was used for inhibition of rat GRO␤-induced calcium mobilization in rat polymorphonuclear leukocytes isolated from peripheral blood (32,33).

Statistical Analysis
Comparisons between controls and various treatments were performed for measures of NF-B DNA-binding activity, B-driven luciferase activity, and cytokine mRNA and protein levels by analysis of variance with post-hoc Dunnett's t-tests. Error bars in figures indicate the S.E.

RESULTS
The Proinflammatory Cytokines IL-1␤ and TNF-␣ Induce LIX Expression-Both proinflammatory cytokines and chemokines are induced during inflammation and endotoxemia. We have previously demonstrated induction of the neutrophil chemoattractants LIX, KC, and MIP-2, members of the ELR ϩ CXC chemokines, during ischemia/reperfusion injury (26). In isolated adult rat cardiomyocytes, TNF-␣ induced LIX expression in a NF-B-dependent manner (26). In the present study, we investigated whether IL-1␤ and TNF-␣ induce LIX expression in rat cardiac-derived endothelial cells. EMSA showed rapid and sustained induction of NF-B DNA-binding activity by After rat CDEC reached 70 -80% confluency in M199 ϩ 10% fetal calf serum, the media was replaced with M199 ϩ 0.5% BSA. After overnight culture, the cells were treated with rLIX (100 ng/ml). At the indicated time periods, the media was removed and the cells were rinsed with ice-cold PBS. The cells were then stored at Ϫ80°C until further analysis. Total RNA was isolated from frozen cells, and quantitated at 260 nm. 20 g of total RNA was electrophoresed in 0.8% agarose-formaldehyde gels, electroblotted onto nitrocellulose membrane, fixed by UV irradiation, and analyzed for cytokine mRNA expression by Northern blot analysis. The same membrane was used after stripping off its previous label. 28 S rRNA was used as an internal control. Our results indicate that both IL-1␤ and TNF-␣ are potent inducers of NF-B (A and B). Both cytokines rapidly increased NF-B DNA-binding activity that sustained up to 48 h. However, when IL-1␤ and TNF-␣ were added together, no further increase in NF-B levels (C) was detected. Similarly, IL-1␤-induced LIX mRNA persisted up to 24 h, and that induced by TNF-␣ remained high up to 48 h (D and E). Lanes 1-3 in panels A-C: lane 1, competition with mutant NF-B oligonucleotide. Protein extract from CDEC treated with cytokine for 30 min was preincubated with 100-fold molar excess of unlabeled double-stranded mutant NF-B oligonucleotide followed by the addition of 32 P-labeled consensus B probe. Lane 2, competition with consensus NF-B oligonucleotide. Protein extract from CDEC treated with cytokine for 30 min was preincubated with 100-fold molar excess of unlabeled double-stranded consensus NF-B oligonucleotide followed by the addition of 32 P-labeled consensus B probe. Lane 3, no protein extract but contains 32 P-labeled consensus B probe. either cytokine in rat CDEC (Fig. 1, A and B). The induction was observed at 5 min after addition of IL-1␤ (100 ng/ml) or TNF-␣ (10 ng/ml) and persisted up to 48 h. However, no synergy was observed when the cells were treated with IL-1␤ and TNF-␣ together (Fig. 1C). Furthermore, IL-1␤ and TNF-␣ induced LIX mRNA expression in a sustained manner (Fig. 1, D  and E). These results indicate that cytokines induce chemokine expression, probably through activation of NF-B.
ELR ϩ CXC Chemokines Activate NF-B DNA-binding Activity-To determine whether cytokine-chemokine cross-talk amplifies the proinflammatory cytokine cascade, we investigated whether ELR ϩ CXC chemokines induce cytokine expression, and the role of NF-B in chemokine-mediated cytokine expression. EMSA showed that unstimulated rat CDEC contained low levels of NF-B in the nuclear protein extracts. Treatment with all three ELR ϩ CXC chemokines increased NF-B activity in a dose-dependent manner ( Fig. 2A). LIX-induced NF-B activity increased to near peak levels at 100 ng/ml, with a slight further increase when the concentration was increased to 1000 ng/ml. Even at 1000 ng/ml, KC-and MIP-2-induced NF-B activity in rat CDEC did not reach levels comparable with that induced by LIX. To confirm the EMSA results we performed transient transfections with a pNF-B luciferase reporter vector. LIX, KC, and MIP-2 induced B-driven luciferase activity (Fig. 2B), and LIX was the most potent inducer of B activation. Neutralization of LIX with anti-LIX antibodies completely blocked LIX-induced B-driven luciferase activity. Semiquantitative analysis by densitometry of autoradiographic bands indicated that LIX induced a rapid increase in NF-B activation (15 min; 2.6-fold, versus untreated controls) that persisted up to 48 h (Fig. 2C). Whereas increased NF-B activity was readily detected upon treatment of rat CDEC with LIX,

FIG. 2. The ELR ؉ CXC chemokines LIX, KC, and MIP-2 activate NF-B and induce B-driven luciferase activity in rat cardiacderived endothelial cells.
Rat CDEC were treated with LIX, KC, and MIP-2 (0, 1, 10, 100, or 1000 ng/ml) for 1 h, and nuclear protein extracts were analyzed by EMSA. Our results indicate that while all three chemokines activated NF-B, LIX was the most potent (A). Similar to the EMSA results, LIX (100 ng/ml) was also found to be the most potent inducer of B-driven luciferase activity in rat CDEC, and neutralizing LIX with anti-LIX antibodies abrogated LIX-induced B-activation (B). C1, untransfected and untreated cells; C2, cells transfected with the control vector pEGFP-Luc; C3, cells transfected with pEGFP-Luc and treated with LIX. ␣LIX, anti-LIX neutralizing antibodies. **, p Ͻ 0.01; *, p Ͻ 0.05 (versus C1); †, p Ͻ 0.01 (versus LIX). A rapid and persistent increase in NF-B activation was detected upon LIX treatment (C). However, LIX treatment had no effect on the basal expression of the constitutively expressed transcription factor Oct1 (D). LIX-induced IB-␣ degradation was associated with transient increase in the phosphorylated form of IB-␣ (P-IB-␣; E). However, ␤-actin levels showed no variations between samples. Furthermore, LIX-induced NF-B contained both p50 and p65 complexes as assessed by gel supershift assay (F).

FIG. 3. LIX treatment up-regulated the B-responsive proinflammatory cytokines IL-1␤ and TNF-␣, increased transcription rate and mRNA half-life, and induced cytokine promoter activity in rat CDEC.
After rat CDEC reached 70 -80% confluency, the complete media was replaced with M199 containing 0.5% BSA. After overnight culture, LIX (100 ng/ml) was added and the incubation continued up to 48 h. At the end of experimental period, media was separated and the cells were rinsed with ice-cold PBS. Cells were then processed for mRNA expression by Northern blotting. Culture supernatants were assayed for cytokine protein levels by enzyme-linked immunosorbent assay. Our results indicate that LIX up-regulated both IL-1␤ and TNF-␣ mRNA expression in a time-dependent manner with significant increases detected around 4 h post-treatment (A). Similarly, treatment with LIX, but not LIX after neutralization, significantly increased cytokine protein levels in culture supernatants at 4 h (B; *, p Ͻ 0.001 (versus control); †, p Ͻ 0.001 (versus LIX)). Because increased transcription and/or mRNA stability contribute to mRNA expression, we then studied the effects of LIX on cytokine transcription by nuclear run-on assay and mRNA stability by actinomycin D pulse. Our results indicate that LIX treatment significantly increased transcripts for IL-1␤ (C) and TNF-␣ (D) in nuclei isolated from LIX-treated rat CDEC. Actinomycin D pulse following LIX treatment showed increased stability of IL-1␤ mRNA as compared with control (E). However, mRNA half-life for TNF-␣ and controls was similar (F). To demonstrate the effects of LIX on cytokine promoter activity, rat CDEC were transiently transfected with 3 g of IL-1␤ promoter construct (IL-1␤-4093 to ϩ45-CAT) or a mock plasmid (pFR-CAT). Cells were cotransfected with pSV-␤-galactosidase vector to compensate for variability in transfection efficiency. CAT and ␤-galactosidase levels were measured, and CAT expression was normalized to that of ␤-galactosidase, and represented as normalized CAT expression. The results indicate that LIX significantly induced IL-1␤ promoter activity (G), and neutralizing LIX with anti-LIX antibodies prevented its stimulatory effects on the promoter activity. To demonstrate LIX effects on TNF-␣ promoter activity, rat CDEC were transfected with a 1.1-kb TNF-␣ promoter that contained 4 NF-B response elements (TNF-␣ Ϫ1080/ϩ138), its deletion mutant that lacks all NF-B response elements (TNF-␣ Ϫ85/ϩ138), or the empty vector (pGL3 basic).
LIX treatment had no effect on the expression of Oct1, a constitutively expressed transcription factor (Fig. 2D).
Activation of NF-B results from phosphorylation and dissociation of IB from the NF-B complex. The phosphorylated IB is then polyubiquitinated and degraded in the cytoplasm by the 26S proteasome. Therefore, we examined the levels of IB-␣ and the phosphorylated form of IB-␣ in the nucleus-free cellular extracts from rat CDEC treated with LIX. Treatment with LIX rapidly but transiently induced phosphorylation of IB-␣ as seen by increased P-IB-␣ levels at 10 min with a corresponding decrease in IB-␣ levels (Fig. 2E).
Cells were cotransfected with pRK-Renilla to compensate for variability in transfection efficiency. The results are represented as a ratio of firefly luciferase to Renilla luciferase. The results indicate that LIX significantly increased TNF-␣ promoter activity (H), and lack of B-response elements abrogated LIX-mediated luciferase activity. *, p Ͻ 0.01 (versus LIX-treated empty vector transfected cells).
CDEC treated with LIX (Fig. 2F). Because LIX was the most potent inducer of NF-B activation, in all subsequent experiments we used LIX at a concentration of 100 ng/ml. LIX Induces Proinflammatory Cytokine Expression-The

FIG. 6. The LIX-mediated increase in intracellular Ca 2؉
i and NF-B activation was sensitive to pertussis toxin. Rat CDEC loaded with Fura-2/AM were treated with pertussis toxin (PTx; 100 ng/ml) for 1 h prior to the addition of LIX (100 ng/ml). Intracellular calcium levels were determined in a spectrophotofluorometer. Our results indicate that whereas PBS had no effect (A), treatment with LIX increased a rapid but transient increase in Ca 2ϩ i (B), pretreatment with PTx significantly attenuated LIX-mediated increase in intracellular Ca 2ϩ i (C) indicating that the inhibitory G proteins are involved in LIX signaling. In the next series of experiments we determined the effects of PTx on LIX-mediated NF-B activation. Rat CDEC were treated with pertussis toxin (100 ng/ml) for 1 h prior to the addition of LIX (100 ng/ml) either for 1 (D) or 4 h (E). Our results indicate that pretreatment with PTx inhibits LIX-mediated NF-B activation (D) and B-responsive cytokine mRNA expression (E). C1, PBS-treated; C2, untreated.
promoter/enhancer regions of proinflammatory cytokines contain binding elements for various stress-responsive transcription factors that are regulated by oxidative stress and proinflammatory stimuli. Therefore, we assessed the effects of LIX on IL-1␤ and TNF-␣ expression in rat CDEC. The results are shown in Fig. 3. IL-1␤ and TNF-␣ mRNA were detected at low levels under basal conditions, and were up-regulated by LIX, with peak levels detected around 4 h (Fig. 3A). Whereas IL-1␤ expression returned to near basal level by 48 h, LIX-induced TNF-␣ expression remained high. In addition to IL-1␤ and TNF-␣ gene transcription, LIX treatment also increased cytokine protein levels in the culture supernatants (Fig. 3B). Both cytokines were induced at high levels by LIX at 4 h, and their expression was blocked when rat CDEC were treated with LIX after antibody neutralization. Thus, LIX induces both transcription and translation of IL-1␤ and TNF-␣.
Increased mRNA levels reflect enhanced transcription and/or increased mRNA half-life. To determine whether increase in cytokine mRNA expression is because of increased gene transcription, we performed nuclear run-on analyses. As illustrated in Fig. 3, C and D, low levels of IL-1␤ and TNF-␣ transcripts were detected in control rat CDEC, but were significantly increased at 4 h after LIX treatment (Il-1␤, 4.8-fold; TNF-␣, 11.3-fold; p Ͻ 0.001). These results indicate that LIXinduced cytokine expression is regulated at the transcriptional level.
To determine mRNA half-life, cells were treated with LIX for 4 h, followed by actinomycin D pulse for up to 6 h. Fig. 3, E and F, shows that the half-life of LIX-induced IL-1␤ mRNA was twice that of untreated controls (t1 ⁄2 ; control, 2.5 h, IL-1␤, ϳ4.25 h). However, the half-life of TNF-␣ mRNA was similar in untreated and LIX-treated cells (t1 ⁄2 , control ϳ2.5 h, TNF-␣, ϳ2.75 h). These results indicate that while increased transcription and mRNA stability contributed to LIX-induced IL-1␤ induction, increased transcription contributed to LIX-induced TNF-␣ expression.
Because LIX up-regulated cytokine mRNA expression, we next determined whether LIX regulates cytokine promoter activity. Rat CDEC were transiently transfected with IL-1␤ or TNF-␣ promoter-reporter constructs. Fig. 3G shows that LIX, but not neutralized LIX, significantly increased IL-1␤ promoter activity, as seen by increased CAT expression (p Ͻ 0.01). To determine TNF-␣ promoter activity, we used a 1.1-kb TNF-␣ promoter (TNF-␣ (Ϫ1080/ϩ135)) that contains 4 NF-B sequence and a deletion construct (TNF-␣ (Ϫ85/ϩ135)) that lacks all four NF-B response elements (Fig. 3H). Treatment with LIX, but not neutralized LIX, induced a 3.5-fold increase in luciferase activity as compared with control (untreated and untransfected cells) and LIX-treated empty vector-transfected rat CDEC (p Ͻ 0.01). In contrast, cells transfected with the TNF-␣ promoter construct that lacks all four NF-B sites failed to respond to LIX (Fig. 3H).
LIX-induced Cytokine Expression Is Dependent on Activation of NF-B, and Involves NIK, IKK, and IB-The signaling cascade initiated by free radicals and proinflammatory cytokines resulting in NF-B activation has previously been shown to converge at IKK. The NIK associates with IKK-␥, and activates IKK-␣ and IKK-␤ of the IKK signalsome resulting in phosphorylation and degradation of IB. Because LIX activated NF-B (Fig. 1, A and B) and induced proinflammatory cytokines (Fig. 3, A and B) in rat CDEC, we next determined if LIX-mediated cytokine expression was dependent on activation of NF-B, and whether LIX-mediated NF-B activation proceeds via NIK, IKK, and IB. We used a series of vectors that expressed dominant negative IB-␣, IB-␤, or IKK-␥ or kinasedeficient NIK or IKK-␤. We confirmed that the dominant negative or kinase-deficient vectors expressed the appropriate protein (Fig. 4, A and B). We then studied the effects of LIX on rat CDEC that had been transiently transfected with the above expression vectors. The results are shown in Fig. 4. LIX-mediated NF-B activation was significantly inhibited by overexpression of dnIB-␣, dnIB-␤, kdNIK, kdIKK-␤, and dnIKK-␥ but not by the corresponding empty vectors (Fig. 4C). LIXmediated cytokine mRNA (Fig. 4D) and protein (Fig. 4E) was also similarly inhibited by these dominant negative and kinase-deficient expression vectors as compared with empty vector-transfected cells.
In addition, we have also tested the effects of IL-10, an FIG. 7. CXCR2-specific blockade inhibits LIX-mediated NF-B activation and B-driven luciferase activity. Reverse transcriptase-PCR was performed to demonstrate expression of CXCR1 and CXCR2 in rat CDEC using primers designed based on published sequences for CXCR1 and CXCR2 in rats. Our results show that rat CDEC expresses both CXCR1 and CXCR2 at basal conditions (A). B, NF-B DNA binding activity by EMSA; C, B-driven luciferase activity. Treatment with a specific CXCR2 antagonist (SB447232, 10 nM, 10 min) followed by LIX treatment for 1 h attenuated NF-B activation in rat CDEC that express both CXCR1 and -R2, and abrogated LIX-mediated NF-B activation in mouse CDEC that express only R2, indicating that LIX signals via both CXCR1 and -R2.
anti-inflammatory cytokine, on LIX-mediated NF-B activation. Our results indicate that pretreatment with IL-10 for 4 h significantly inhibited LIX-induced NF-B activation and cytokine expression (Fig. 4, B-D).
LIX-mediated NF-B Activation Involves PI3K, PKC, PARP-1, and Proteasome-The signal transduction pathway(s) initiated by ELR ϩ CXC chemokines in the activation of NF-B are not fully known. To determine the role of PI3K, PKC, PARP-1, and proteasome in LIX-mediated activation of NF-B, rat CDEC were pretreated with selective inhibitors. Inhibition of PI3K with the selective inhibitor LY 294002 significantly inhibited LIX-mediated B activation and cytokine expression (Fig. 5A). Similar results were obtained with wortmannin. Furthermore, 3-aminobenzamide, a specific PARP-1 inhibitor, and MG-132, a proteasome inhibitor, also inhibited LIX-mediated B activation and cytokine expression (Fig. 5, B and C). On the other hand, inhibition of PKC by chelerythrine chloride partially, but significantly, attenuated B activation and cytokine expression.
To confirm activation of PI3K by LIX, we performed PI3K lipid kinase assays. Fig. 5D shows that LIX, but not neutralized LIX, activates PI3K (2.4-fold increase, p Ͻ 0.01). Whereas Me 2 SO and LY 294002 had no effect on basal levels of PI3K, pretreatment with LY 294002 significantly inhibited LIX-induced activation of PI3K (p Ͻ 0.01; Fig. 5D).
LIX-mediated NF-B Activation Is G Protein-dependent-The ELR ϩ CXC chemokines signal through the 7 transmembrane domain G protein-coupled receptors CXCR1 (R1) and CXCR2 (R2). Damaj et al. (34) have shown that addition of IL-8 to cell lines that specifically express either R1 or R2, or to neutrophils that express both R1 and R2 resulted in the formation of immunoprecipitable complexes containing the receptors and the ␣ subunits of G i2 proteins. In addition, IL-8mediated increase in cytosolic-free calcium was inhibited by pertussis toxin indicating a G protein-coupled signal transduction (35). Therefore, we determined the role of G proteins in LIX-mediated cellular and NF-B activation. Rat CDEC were pretreated with pertussis toxin for 1 h followed by LIX stimulation. G-protein function was determined by measuring intracellular calcium levels in Fura-2/AM-loaded rat CDEC. Fig. 6 illustrates that treatment with LIX (100 ng/ml), but not PBS, increased intracellular calcium levels rapidly but transiently (compare Fig. 6, A and B). Pretreatment with pertussis toxin significantly attenuated LIX-induced calcium transient (Fig.  6C). Furthermore, LIX-mediated NF-B activation and cytokine expression were also blocked by pertussis toxin (Fig. 6, D and E) indicating the role of inhibitory G proteins in LIXmediated NF-B activation.
LIX-mediated NF-B Activation Involves Both CXCR1 and CXCR2-CXCR1 and -R2 have a 78% sequence homology within the transmembrane domains, but differ in the extracel-lular, intracellular, and NH 2 -and COOH-terminals, leading to distinct ligand specificity and signaling (36 -39). Therefore, we determined the role of CXCR1 and -R2 in LIX-mediated NF-B activation. Rat CDEC that express both R1 and R2 (Fig. 7A), and mouse CDEC that express only R2 were pretreated with SB447232, a CXCR2-specific antagonist (32), and LIX-induced NF-B activation was determined. SB 447232 is a potent CXCR2 antagonist with similar selectivity to SB 225002 (32) and SB 265610 (33), i.e. Ͼ100-fold higher affinity for human CXCR2 versus CXCR1 (binding IC 50 values of 26.8 Ϯ 3.5 (n ϭ 6) and 5,610 Ϯ 1,433 nM, respectively (Table I). In addition, SB 447232 was a potent inhibitor of CXCR2 (0.1 nM rat GRO␤induced calcium mobilization) on rat neutrophils with an IC 50 of 16 nM (Table I). The advantage of SB 447232 over SB 225002 and SB 265610 is that the former compound has much better bioavailability in rodents and will be useful for in vivo studies in the future. In mouse CDEC, CXCR2 blockade completely abrogated NF-B activation and B-driven luciferase activity. Blockade of R2 in rat CDEC attenuated LIX-induced NF-B activation and B-driven luciferase activity by 50% (Fig. 7, B and C) indicating that LIX signals through CXCR2, and presumably CXCR1. DISCUSSION Our results indicate for the first time that the ELR ϩ CXC chemokines LIX, KC, and MIP-2 up-regulate the proinflammatory cytokines IL-1␤ and TNF-␣ in cardiac-derived endothelial cells via activation of NF-B. LIX-mediated NF-B activation and B-responsive gene transcription involves IB hyperphosphorylation leading to its selective degradation in the cytoplasm by the proteasome system. The LIX signaling was inhibited by IL-10 and NF-B pathway-specific mutant expression vectors. LIX signals via both CXCR1 and -R2 in inducing NF-B activation. Specific blocking of CXCR2 attenuated NF-B activation in rat cardiac-derived endothelial cells that express both R1 and R2, and completely abrogated LIX-induced NF-B activation and B-driven luciferase activity in mouse cardiac-derived endothelial cells that express only R2.
The ELR ϩ CXC chemokines primarily attract and activate neutrophils to the site of injury/inflammation (1-7). We have previously shown that LIX, KC, and MIP-2 are expressed in the post-ischemic myocardium, and most notably LIX is expressed by all myocardial constituent cells (10). Although activated neutrophils at the site of myocardial ischemic injury play a role in scavenging damaged tissue and subsequent remodeling, at least initially they exacerbate tissue injury through generation of free radicals, and secretion of various proteolytic enzymes and proinflammatory cytokines (40 -48). The results presented here demonstrate that chemokines (LIX) in addition to the recruitment of neutrophils to the site of myocardial ischemic injury may contribute to myocardial inflammation by the direct induction of cytokine expression. Interleukin-1␤ and TNF-␣ are B-responsive proinflammatory cytokines with known negative myocardial inotropic effects. In isolated cardiomyocytes, papillary muscles, myocardial segments, Langendorff preparations, and in whole animals addition/infusion of TNF-␣ has been shown to depresses contractile function via induction of the inducible form of nitricoxide synthase and sustained generation of high levels of nitric oxide (49 -54). High levels of TNF-␣ expression have also been shown to induce cell death by apoptosis (49). Both TNF-␣ and IL-1␤ are known free radical generators and NF-B activators (11). We have previously shown that TNF-␣ induces LIX via activation of NF-B in isolated cardiomyocytes (10). Furthermore, we demonstrated activation of NF-B and induction of LIX by IL-1␤ and TNF-␣ in rat CDEC (Fig. 1). In the present study, we describe the converse, that is, the induction of proinflammatory cytokines by chemokines via NF-B activation. Together, these observations indicate that NF-B activation plays a central role in regulating cross-talk between chemokines and cytokines in myocardial cells.
The ELR ϩ CXC chemokines bind and exert their biological effects via the seven-transmembrane heterotrimeric G proteincoupled receptors CXCR1 and CXCR2. The sequences of CXCR1 and -R2 within the seven-transmembrane domains and the connecting loops are homologous, but differ in the N and C-terminal domains, leading to overlapping as well as distinct ligand-binding and selective signal transduction pathways (36 -39). Whereas all ELR ϩ CXC chemokines bind with high affinity to CXCR2, IL-8, because of the presence of Tyr 13 and Lys 15 in the N terminus has been shown to also bind R1 with high affinity (55). Recently, granulocyte chemotactic protein-2 has also been shown to bind R1 with high affinity because of the presence of Arg 20 (56) indicating that other ELR ϩ CXC chemokines may bind to both R1 and R2 with high affinity. In the present study, we demonstrated that LIX-induced NF-B activation is mediated in part through R2 in rat CDEC that express both R1 and R2, and fully through R2 in mouse CDEC that express only this receptor. Presumably, in the rat CDEC that express both receptors, the R2-independent signaling occurs through R1.
Pretreatment of endothelial cells with pertussis toxin, which specifically blocks the coupling of CXC receptor to G i proteins, attenuated the LIX-induced increase in intracellular calcium levels and completely inhibited LIX-induced NF-B activation. Similarly, treatment with LY 294002, a specific PI3K inhibitor, inhibited LIX-induced activation of PI3K activity and completely blocked NF-B activation and B-responsive cytokine gene transcription. In contrast, chelerythrine chloride, a PKC inhibitor, partially inhibited LIX-mediated NF-B activation and cytokine expression. Collectively, these data indicate that LIX signals via inhibitory G proteins and PI3K, and partially via PKC. Although other G proteins may be involved in chemokine-mediated cell signaling (57,58), our results exclude this probability because LIX-mediated NF-B activation was completely blocked by pertussis toxin.
Interleukin-10 is an anti-inflammatory cytokine, and has been shown to block expression of various proinflammatory cytokines via inhibition of NF-B activation (59). In the present study we demonstrate that IL-10 blocked LIX-induced NF-B activation and B-responsive gene transcription. It has been previously shown that IL-10 could block NF-B activation by inhibiting IKK-mediated IB phosphorylation and degradation (55). Because the inhibitory effects of IL-10 are not cell-specific, and can inhibit activation of NF-B in response to various proinflammatory stimuli, IL-10 may have a therapeutic potential in ischemia/reperfusion injury by blocking induction of proinflammatory cytokines and chemokines (60).
In the present study, we demonstrate for the first time that inhibition of PARP-1 activation prevents LIX-mediated NF-B activation and IL-1␤ and TNF-␣ expression. It has been demonstrated recently that PARP-1, a nuclear protein involved in repairing DNA strand breaks, has also been shown to activate NF-B (18). PARP-1 activation has been described during endotoxemia and inflammation (62)(63)(64)(65)(66). Administration of lipopolysaccharide to mice activated PARP-1 and resulted in PARP-1-dependent B-responsive IL-1, IL-6, TNF-␣, iNOS gene expression, and iNOS-mediated NO generation (63). Furthermore, in the murine system PARP-1 gene disruption or pharmacological inhibition of PARP-1 activation has been shown to reduce free radical generation, attenuate B-responsive gene transcription, and reduce neutrophil infiltration in the lungs (64). Whether PARP-1 may be a logical target for inhibition to attenuate post-ischemic myocardial injury will require further study.
Taken together, our results indicate that the ELR ϩ CXC chemokines, besides being potent neutrophil chemoattractants, also induce proinflammatory cytokine expression via activation of NF-B. Blunting the activation of NF-B or other components of the signaling cascade, rather than targeting inhibition of individual cytokines, chemokines, or adhesion molecules, may be a valid strategy to attenuate myocardial tissue injury during various inflammatory conditions.