Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Tumor necrosis factor (TNF)- is a mediator of experimental endotoxemia, sepsis, and adult respiratory distress syndrome (1). TNF mediates a systemic and pulmonary endothelial inflammatory response characterized by altered expression of messenger RNA (mRNA) for proteins such as E-selectin (4, 5), intercellular adhesion molecule-1 (4, 5), vascular cell adhesion molecule-1 (4, 5), tissue factor (6), macrophage colony-stimulating factor (7), and nitric oxide (NO) synthase (8). TNF mediates regulation of mRNA for phlogistic substances, at least in part, by altered activation of transcription factors such as activating protein-1 (AP-1) (9). Activated AP-1 binds to regulatory sites on the 5' and 3' untranslated regions of introns, resulting in modulation of mRNA polymerase activity and subsequent effects on transcription of DNA into mRNA (6, 10, 11). Thus, AP-1 can modulate induction of protein synthesis by modulation of specific mRNA sequences.

We have shown that TNF induces an endothelial dysfunction mediated, at least in part, by the increased activity of reactive nitrogen species (12, 13). Reactive nitrogen species activity in endothelial cells can be mediated, at least in part, by the following two pathways (4). First, NO or its reaction product peroxynitrite anion ([ONOO], formed via the reaction NO + O₂  ONOO) induces the activation of guanylate cyclase (8, 14, 15), resulting in synthesis of the catalytic factor cyclic guanosine monophosphate (cGMP) and increased activity of cGMP-dependent protein kinase G (PKG) (8, 16). PKG is a serine-threonine kinase that can potentially modulate phosphorylation of molecules that regulate transcription, such as the AP-1 subcomponents cFOS and cJUN (10, 11, 16, 17). Second, NO or ONOO can react directly with amino acids, resulting in altered function of the target protein (14).

The role of NO-mediated PKG activation during TNF-induced AP-1 activation is not known despite the fact that NO agonists, via a PKG pathway, can increase AP-1 activity (17). Gudi and colleages demonstrated that transfection of PKG into initially PKG-deficient cells induced the activation of AP-1 (16). Tabuchi and associates (19) demonstrated in cerebellar granule cells that NO agonist- induced nitrosylation of AP-1 inhibits its DNA-binding activity, suggesting NO indirectly activates AP-1 via alternate pathways such as PKG (19) and ONOO (20). Thus, the role of PKG in the activity of AP-1 may be dependent on the cell type, culturing conditions, and agents used to alter cGMP activity.

In a model of acute pulmonary endothelial dysfunction using TNF, a systematic investigation of the role of NO-induced, PKG-mediated activation of AP-1 has not been done. Thus, in the present study we investigated the role of NO and activation of PKG in TNF-induced activation of AP-1 in pulmonary endothelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

All reagents were obtained from Sigma Chemical Company (St. Louis, MO) unless otherwise noted.

Pulmonary Microvessel Endothelial Cell Culture

Bovine pulmonary microvessel endothelial cells derived from fresh calf lungs were obtained at the ninth passage (Vec Technologies, Rensselaer, NY). The cells were serially cultured, from nine to 23 passages, in Dulbecco's modified Eagle's medium (DMEM) (GIBCO BRL, Grand Island, NY) supplemented with 10% fetal bovine serum (Hyclone; Hyclone Laboratories, Logan, UT) and 1% nonessential amino acids (GIBCO BRL) (12, 13).

Pulmonary microvessel endothelial monolayers (PEM) were maintained in 5% CO₂ plus humidified air at 37°C and reached confluence within two to three population doublings (bovine PEM: 3 to 5 d). The preparations were identified as endothelial monolayers by: (1) the characteristic "cobblestone" appearance using contrast microscopy showing (2) the presence of Factor VIII-related antigen (indirect immunofluorescence), (3) the uptake of acylated low-density lipoproteins, and (4) the absence of smooth-muscle actin (indirect immunofluorescence).

Pharmacologic Reagents

We determined the lowest effective dose of a pharmacologic agent on the basis of our own preliminary studies or from studies obtained from the literature, verified by using assays and controls as discussed later.

TNF and inactivated TNF. Highly purified recombinant human TNF from Escherichia coli (Genzyme, Cambridge, MA) with a specific activity of 588-900 × 10⁸ U/ml was used at 1,000 U/ml (1, 12, 13). The TNF had less than 14 pg of endotoxin contamination per 10⁶ units of TNF activity by standard limulus assay. To control for endotoxin contamination and the possible effects of the TNF vehicle, TNF was heat-inactivated by boiling for 45 min (12, 13).

NO-related reagents. To test the role of NO generation in the endothelial response to TNF, NO synthase activity was inhibited by coincubation with the NO synthase substrate antagonist aminoguanidine (100 µM; LC Laboratories, Woburn, MA) (12). In addition, PEM were preincubated (i.e., 15 min before aminoguanidine + TNF) with L-arginine (1 mM) to verify the effect of endogenous repletion of NO activity (12, 13). We previously showed that similar doses of L-arginine and spermine-NO reversed the aminoguanidine-induced prevention of the increase in NO and ONOO during a 4-h incubation with TNF using pulmonary arterial endothelial cells (12). NO₂/NO₃ levels, an index of NO, were measured with a Nitrate/Nitrite Colometric Assay Kit (Cayman Chemical, Ann Arbor, MI), a two-step assay using nitrate reductase and Griess reagent (12, 13).

Peroxynitrite-related reagents. In separate studies, SIN-1 (1 mM; Calbiochem, San Diego, CA) was used to verify that the TNF effect was due to the simultaneous generation of superoxide anion (O₂) and NO on the endothelium. SIN-1 generates NO + O₂, resulting in formation of ONOO (NO + O₂  ONOO) (14, 21). In separate studies, superoxide dismutase (SOD) (10 U/ml, 4,166 U/mg), the enzyme that dismutates O₂ into hydrogen peroxide and oxygen, was coincubated with TNF or SIN-1 to verify the role of O₂ because SOD removes the reactant O₂. We demonstrated that this dose of SOD is not a direct ONOO scavenger (13, 22). To verify antioxidant specificity, heat-inactivated SOD (autoclaving SOD for 45 min) was coincubated with the TNF (13, 22).

PKG-related reagents. To test the role of guanylate cyclase activity in the endothelial response to TNF, guanylate cyclase was inhibited by coincubation with ODQ (100 µM; Alexis Corp., San Diego, CA) (18). To test the role of PKG activity in the endothelial response to TNF, PKG activity was specifically inhibited by coincubation with the adenosine triphosphate (ATP)-binding site antagonist KT5823 (1 µM; Calbiochem) (18, 23) or the cGMP-binding site antagonist 8-bromo-cGMP-thioate (100 µM, Alexis Corp.). 8-Bromo-cGMP (100 µM), a phosphodiesterase-resistant cGMP analog, was used both to verify that the effect of ODQ was due to inhibition of guanylate cyclase and to assess the effect of cGMP on the endothelium (16).

Treatments

PEM in 35-mm wells were treated with TNF, SIN-1, spermine-NO, 8-bromo-cGMP, or appropriate controls for 0.5, 2, or 4 h before preparation of cell lysate. The TNF, SIN-1, spermine-NO, 8-bromo-cGMP, or controls were coincubated with the pharmacologic probes during the 0.5-, 2-, or 4-h period of incubation. The appropriate lysis buffer was added to the PEM after aspiration of the treated culture media.

Electrophoretic Mobility Shift Assay

Nuclear extracts. Nuclear extracts were prepared as previously described (24). PEM were seeded (1 × 10⁶) in six-well 35-mm plastic dishes and allowed to reach confluence (2 × 10⁶ cells) in 3 to 5 d. PEM were washed once with Tris-buffered saline, followed by treatment with buffer A (10 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], 10 mM KCl, 100 µM ethylenediaminetetraacetic acid [EDTA], 100 µM ethyleneglycol-bis-(-aminoethyl ether)-N-N'-tetraacetic acid [EGTA], 1 mM dithiothreitol [DTT], 1 µg/ml leupeptin, 5 µg/ml pepstatin, 5 µg/ml antipain, 700 µM phenylmethylsufonyl fluoride [PMSF], 5 µg/ ml antitrypsin, 5 µg/ml aprotinin, and 100 µM benzamidine. The PEM were scraped using a rubber policeman. The lysate was placed in a test tube and incubated on ice for 15 min. After incubation, 25 µl of Nonidet P-40 (10% solution) was added to 400 µl of lysate and the test tube was vortexed for 10 s and centrifuged at 15,000 × g for 30 s at 4°C. The pellet was resuspended in 40 µl of buffer C (5 mM Hepes, 100 mM NaCl, 250 µM EDTA, 250 µM EGTA, 1 mM DTT, 1 µg/ml leupeptin, 5 µg/ml pepstatin, 5 µg/ml antipain, 1.5 mM PMSF, 5 µg/ml antitrypsin, 5 µg/ ml aprotinin, and 100 µM benzamidine) and shaken on ice for 15 min. The suspension was centrifuged at 15,000 × g for 5 min at 4°C. The supernatant was aliquotted and stored at 70°C. A necessary protein concentration of > 1 µg/µl was determined using the Bradford assay.

Labeling of oligonucleotides. Oligonucleotide cognates for AP-1 (5'-CGC TTG ATG AGT CAG C-3'; Promega, Madison, WI) and nuclear factor (NF)-B (5'-AGT TGA GGG GAC TTT CCC AGG C-3'; Promega) were prepared using a modification of Promega Technical Bulletin #110. After the labeling of the oligonucleotides with ³²P-deoxyATP using T4 polynucleotide kinase, the unincorporated label was removed using Nick Spin Columns (Pharmacia, Piscataway, NJ). The oligonucleotide's final specific activity was 50,000 to 200,000 counts per minute (cpm)/µl.

Electrophoretic mobility shift assay. The reaction mixture was composed of nuclear extract (5 µg), ³²P-oligonucleotide cognate (1 µl), and 0.4 mg/ml of calf thymus DNA in a total volume of 20 µl in binding buffer (50 mM Tris-HCl, 250 mM NaCl, 2.5 mM EDTA, 2.5 mM DTT, and 20% glycerol, pH 7.5). The 5 µg of extract was within the linear range of a curve obtained in preliminary studies that indicated that the increase in AP-1 DNA binding versus the amount of sample protein was similar between the Control and TNF Groups. The reaction mixture was incubated for 5 to 10 min at room temperature, loaded on a 6% polyacrylamide gel, and electrophoresed at 500 mA and 125 V for 3.5 h. The gel was dried at 37°C for 24 h and autoradiographed at 70°C (overnight or until an adequate signal developed).

Quantitation of autoradiographs. Autoradiographs were digitized using a scanner (Scan Jet 4C/T; Hewlett-Packard, Downers Grove, IL). The bands were analyzed with Sigma Scan/Image (Jandel Scientific Software, San Rafael, CA) and quantified by: fill area × intensity = total intensity units. The digitized data were within the linear range of a curve obtained in preliminary studies that indicated that the increase in density versus the amount of sample protein was similar between the Control and TNF Groups.

Assay of PKG

PKG activity in cell lysate was measured by the ³²P-ATP-mediated phosphorylation of the peptide substrate BPDEtide (16, 17).

Preparation of sample. PEM were seeded (400,000 cells/ well) and grown until confluent in 35-mm well dishes. After the respective drug procedures, PEM were washed three times with ice-cold phosphate-buffered saline (PBS) (2 ml/well) and scraped with 500 µl ice-cold homogenization buffer: potassium phosphate (10 mM, pH 6.8), EDTA (10 mM), 2-mercaptoethanol (10 mM), and 3-isobutylmethylxanthine (IBMX) (1 mM; Alexis). The cell suspension was vortexed (5 s) and immediately placed on ice. An aliquot (40 µl) was removed for determination of cell counts. The remainder of the cell suspension was lysed by three sonications (a 10-s pulse) and clarified by centrifugation at 11,000 × g (30 s, 4°C). The supernatant (sample) was snap-frozen with dry ice/2-propanol and stored at 70°C until used.

Activity of sample. The sample (8 µl) was added to 20 µl of reaction mixture (BPDEtide [150 µM; Alexis], Tris [40 µM, pH 7.4], magnesium acetate [20 mM], -³²P-ATP [200 µM/0.6 Ci/; Amersham, Arlington Heights, IL], IBMX [100 µM], and PKI [0.9 µm; Alexis]) and incubated for 10 min on ice. The reaction was terminated by spotting aliquots (25 µl) onto 2-cm² phosphocellulose filter paper (Whatman P81; Fisher Scientific, Pittsburgh, PA). The filter paper was washed three times in phosphoric acid (75 mM) and put into Ecoscint (10 ml; National Diagnostic, Atlanta, GA), and radioactivity was measured using a beta counter (Packard 2500 TR; Canberra Company, Downers Grove, IL). The cpm were corrected for cell counts and specific activity of the ³²P-ATP.

Assay of PKA

PKA activity in cell lysate was measured by the ³²P-ATP- mediated phosphorylation of the peptide substrate KEMPtide (23).

Preparation of sample. The technique for the preparation of the samples was identical to the preparation technique described previously in the assay of PKG.

Activity of sample. The technique for the determination of the activity of the sample was identical for the PKG activity technique described earlier except for the following differences. First, the reaction mixture contained the PKA-specific substrate KEMPtide (150 µm; Alexis) instead of BPDEtide. Second, the PKG inhibitor KT5823 (1.0 µm) was included in all assays instead of PKI. Third, the positive control was 8-bromo-cyclic adenosine monophosphate (cAMP) (100 µM) rather than 8-bromo-cGMP. Fourth, the activity was corrected for total protein (Bradford Assay; Bio-Rad, Hercules, CA) because it did not require separate studies as did the determination of cell counts.

Transfection Studies: Biologic Relevance of Pulmonary Endothelial AP-1 Activation by TNF

Plasmids. PEM were transfected with a plasmid with a response element interacting exclusively with AP-1 (kindly supplied by Dr. D. Leonardo, New York Medical Center, Tuxedo, NY) to demonstrate the functional role of AP-1 in TNF-induced gene expression in the pulmonary endothelium. The plasmid was constructed by linking a pBL-CAT2 plasmid (Accession #X64410) that contains a chloramphenicol acetyltransferase (CAT) reporter and a minimal thymidine kinase promoter (25) to an AP-1 binding site (5'-TGAGTCA). The efficiency of DNA expression was monitored by cotransfection with the reporter plasmid pCH-110 containing the -galactosidase (GAL) gene driven by the active simian virus 40 promoter (Promega). CAT activity was assessed with enzyme-linked immunosorbent assay (Boehringer-Mannheim, Indianapolis, IN) and GAL was measured using the -Galactosidase Enzyme Assay System (Promega). CAT activity was reported as the ratio of CAT/GAL activity (24).

Transfection. Plasmids (1 µg/µl) were mixed with Tranfectam (Promega Technical Bulletin #116) in serum-free DMEM and were immediately added to PEM. The PEM were initially incubated with the plasmids for 2.5 h, overlaid with complete media, and grown for 48 h before treatment with TNF.

Assay of Cell Viability

Trypan blue exclusion. PEM were seeded (2 × 10⁶ cells/ well) and grown until confluent in 100-mm well dishes. After the respective drug procedures, PEM were washed with PBS (2 ml) followed by treatment with 0.05% trypsin (1 ml) for 1 min at 37°C. The cells were then washed and suspended in PBS (5 ml). An aliquot of cell suspension (50 µl) was combined with 0.4% trypan blue (50 µl) for 3 min, then 20 µl of the mixture was counted for cells using a hemocytometer. Cell viability was defined by the following formula: Cell viability = (cells excluding trypan blue/ total cells) × 100.

Cell number. Total cell numbers of PEM were determined as previously described in conjunction with the trypan exclusion studies.

Nuclear protein. Nuclear protein concentrations were determined in diluted samples with a Bradford Assay (Bio-Rad) (12).

Statistics

A one-way analysis of variance was used to compare values among the treatments. If significance among treatments was noted, a Bonferroni multiple comparison test (post hoc) was used to determine significant differences among the groups (26). A t test was used when appropriate. Each well and flask represents a single experiment. All data are reported as means ± standard error of the mean. Significance was at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

TNF Did Not Affect Cell Viability

The levels of cell counts, trypan blue exclusion, and nuclear protein concentrations were similar among the groups, indicating that the various treatments did not induce toxicity in the endothelial monolayer (e.g., Control: 3.7 ± 0.1 versus TNF: 3.5 ± 0.1 cells × 10⁶/100 mm; Control: 45 ± 3.1 versus TNF: 39.8 ± 3.4 µg protein/35 mm; Control: 97.4 ± 0.4 versus TNF: 96.2 ± 0.4% trypan blue negative).

TNF Increased AP-1 Activity in PEM

TNF induced a time-dependent increase in DNA binding of AP-1. Figure 1 shows the time course for the TNF-induced increase in AP-1 activity. TNF from 0.5 to 4 h induced an augmentation in density of the AP-1 band (i.e., decrease in migration of the oligonucleotide probe) that became statistically significant at 4 h compared with the Control Group.

View larger version (44K): [in this window] [in a new window]   Figure 1.   TNF induces a time-dependent increase in DNA binding of AP-1. Graph shows the mean ratio of relative density of the AP-1 band for each group from PEM incubated with control (Con) media and TNF for 0.5, 2, and 4 h. *Significantly different from Control Group value.

Role of NO

TNF activated DNA binding of AP-1 via a NO-dependent pathway. Figure 2 shows the effect of inhibition of NO synthase by aminoguanidine on TNF-induced enhancement of AP-1 DNA-binding activity. Figure 2A shows representative assays of nuclear lysates isolated from PEM incubated for 4 h with TNF, inactive TNF, aminoguanidine, L-arginine, and control media. Figure 2B shows the mean values (n = 6 per group) of the groups represented in Figure 2A compared to the Control Group. TNF induced an increase in density of the AP-1 band at 4 h compared with the Control Group, which was not observed in the Inactive TNF Group. In the Aminoguanidine + TNF Group, the AP-1 DNA-binding activity was similar to that of the Control Group but higher compared with the Aminoguanidine Group. In the L-Arginine + Aminoguanidine + TNF Group, TNF did induce an increase in AP-1 DNA-binding activity compared with the Control and L-Arginine + Aminoguanidine Groups, indicating the specificity of aminoguanidine as an antagonist of L-arginine activity. There were no significant densities detected in the lanes loaded with free oligonucleotide and TNF + 50× cold oligonucleotide (data not shown). Thus, the data indicate that NO synthase activity and additional pathways mediate TNF-induced AP-1 activation.

View larger version (53K): [in this window] [in a new window]   View larger version (42K): [in this window] [in a new window]   Figure 2.   TNF activates DNA binding of AP-1 via a NO-dependent pathway. (A) Autoradiograph of representative EMSAs of samples obtained from PEM incubated with control (CON) media, aminoguanidine (AMINO), L-arginine (L-ARG), TNF, and inactive TNF for 4 h. Lanes are: 1 = CON, 2 = TNF, 3 = AMINO, 4 = CON, 5 = TNF, 6 = AMINO + TNF, 7 = L-ARG + AMINO + TNF, 8 = CON, 9 = TNF, and 10 = Inactive TNF. (B) Graph shows the mean ratio of relative density of the AP-1 band for each group. In addition, shows the Aminoguanidine, Spermine-NO, and Spermine Groups. The autoradiographs were scanned and each AP-1 band was analyzed for total pixels (area × pixels/area unit). *Significantly different from both Control Group and Aminoguanidine Group values; #Significantly different from Aminoguanidine Group value.

TNF recruited a concurrent pathway: DNA binding of AP-1 was inhibited by SOD. Our previous data indicate that a potential concurrent mediator of TNF-induced AP-1 activation is O₂ (13, 14, 22). Thus, PEM were coincubated with SOD to remove the reactant O₂. Figure 3 shows the mean relative density of each group (n = 3 per group) for nuclear lysates isolated 4 h after incubation with TNF, SOD, inactive SOD, and control media. The density of the AP-1 band increased in the TNF Group compared with the density in the Control Group. The density of the AP-1 band did not change in the SOD + TNF Group compared with the density in the Control Group; however, the density of the AP-1 band did increase in the TNF + Inactive SOD Group, indicating the specific O₂ dismutase effect of the SOD treatment. There were no significant densities detected in the lanes loaded with free oligonucleotide (no sample) and TNF + 50× cold oligonucleotide (data not shown). Therefore, the pharmacologic profile derived from the results of Figures 2 and 3 supports the hypothesis that the TNF-induced increase in AP-1 DNA-binding activity is dependent on both NO and O₂.

View larger version (44K): [in this window] [in a new window]   Figure 3.   TNF-induced DNA binding of pulmonary endothelial AP-1 is inhibited by SOD. Graph shows the mean ratio of the relative density of each group at 4 h in the Control, SOD, Inactive SOD, and TNF Groups. *Significantly different from Control Group value.

SIN-1 induced an increase in DNA binding activity of AP-1. We used SIN-1 (SIN  NO + O₂) as a positive control because the pharmacologic profile of Figures 2 and 3 suggests that the activity of both NO and O₂ mediate TNF-induced effects on the endothelium. Figure 4 shows a representative electrophoretic mobility shift assay (EMSA) (n = 3 per group) of nuclear lysate isolated 4 h after incubation with SIN-1, SOD, and control media. The density of the AP-1 band increased in the SIN-1 Group compared with the density in the Control Group. The density of the AP-1 band did not change significantly in the SOD + SIN-1 Group compared with the SOD Group. Thus, the data indicate that SIN-1-induced AP-1 activation is dependent on O₂.

View larger version (82K): [in this window] [in a new window]   Figure 4.   SIN-1 activates pulmonary endothelial DNA-binding activity of AP-1. Figure shows a representative EMSA indicating densities of the AP-1 band at 4 h in the Control, SIN-1, and SOD Groups. *Significantly different from Control Group value.

Role of PKG

TNF activated DNA binding of AP-1 via cGMP-PKG. We studied the role of the guanylate cyclase and PKG in TNF-induced AP-1 DNA-binding activity because cGMP-mediated PKG activity is the classic effector pathway downstream from NO and ONOO (8, 16). Figure 5 (n = 5 per group) shows the effect of inhibition of guanylate cyclase by ODQ and inhibition of PKG by KT5823 and 8-bromo-cGMP-thioate on TNF-induced increases in AP-1 DNA-binding activity. There was an increase in the density of the AP-1 band in the TNF Group. The density of the AP-1 band did not change in the 8-bromo-cGMP, ODQ, and ODQ + TNF Groups; however, the density of the AP-1 band increased in the ODQ + 8-Bromo-cGMP + TNF Group. These data indicate that 8-Bromo-cGMP alone had no effect on AP-1 activation but restored TNF-induced AP-1 activation in PEM also treated with ODQ; thus, the effect of ODQ was specific to inhibition of guanylate cyclase-mediated cGMP generation.

View larger version (41K): [in this window] [in a new window]   Figure 5.   TNF activates pulmonary endothelial DNA-binding activity of AP-1 via cGMP-PKG. Graph shows the mean ratio of relative density of each group at 4 h in the Control, ODQ, KT5823, 8-Bromo-cGMP-thioate, 8-Bromo-cGMP, and TNF Groups. *Significantly different from Control Group value.

SIN-1 activated DNA binding of AP-1 via cGMP-PKG. We wanted to verify that SIN-1 induces a PKG-dependent increase in AP-1 activity. There was a greater density in the SIN Group (95.8 ± 4.2 AP-1 relative density units) compared with the densities in the Control (37.5 ± 3.5, AP-1 relative density units), KT5823 (38.2 ± 17.0 AP-1 relative density units), and KT5823 + SIN-1 (33.9 ± 9.7 AP-1 relative density units) Groups.

TNF Increased PKG Activity in PEM

Role of NO. We wanted to verify that the pharmacologic profile of TNF-induced, NO-mediated, AP-1 DNA-binding activity is supported by an appropriate pharmacologic profile of PKG activity. Table 1 shows the PKG values in the Control and treated groups (n = 5-10 per group). TNF induced an increase in PKG activity at 0.5 h compared with the Control Group; however, there was no increase in PKG activity at 4 h. TNF did not induce a change in the PKG activity in the Aminoguanidine + TNF Group but did induce a PKG activity increase in the L-Arginine + Aminoguanidine + TNF Group, indicating that NO synthase activity mediates TNF-induced PKG activation. TNF increased PKG activity in the SOD + TNF Group, indicating that the TNF-induced PKG activation is not mediated by O₂. Thus, the data in Table 1 support the concept that TNF-induced activation of PKG is mediated by NO.

Role of guanylate cyclase. TNF did not increase PKG activity in the ODQ + TNF Group, indicating that guanylate cyclase activity may mediate TNF-induced PKG activation. However, TNF did induce an increase in PKG activity in the ODQ + 8-Bromo-cGMP + TNF Group, indicating that cGMP mediates TNF-induced activation of PKG. In the KT5823 + TNF and 8-Bromo-cGMP-thioate + TNF Groups, there was no increase in PKG activity, indicating that TNF-induced increases in AP-1 activity were mediated by PKG.

TNF Did Not Alter PKA Activity in PEM

We wanted to verify that the pharmacologic manipulation of PKG activity indicated in Table 1 was not confounded by changes in PKA activity. Table 2 shows the PKA values in the Control and treated groups (n = 5-10 per group). In the 8-Bromo-cAMP Group, there was an increase in PKA activity at 0.5 h but not at 4 h, which was prevented by KT5720 in the KT5720 + 8-Bromo-cAMP Group. In the TNF and KT5823 + TNF Groups, there was no change in PKA activity compared with the Control Group. Interestingly, there were trends for PKA activity to increase in the Aminoguanidine + TNF and ODQ + TNF Groups that were prevented by L-arginine (L-Arginine + Aminoguanidine + TNF Group) and 8-bromo-cGMP (8-Bromo-cGMP + ODQ + TNF Group), respectively. Thus, the data indicate that a 4-h treatment with TNF does not change the activity of PKA and that direct inhibition of PKG has no effect on the activity of PKA in response to TNF.

cGMP without TNF Did Not Activate DNA Binding of AP-1

We assessed the effect of 8-bromo-cGMP for 0.5 h on AP-1 DNA-binding activity because Table 1 indicates that TNF and 8-bromo-cGMP increased PKG activity only at 0.5 h. Figure 6 shows a representative EMSA (n = 3 per group) of nuclear lysate isolated 0.5 and 4 h after incubation with no sample, TNF, 8-bromo-cGMP, and control media. In the 8-Bromo-cGMP Group, the density of the AP-1 band at 0.5 and 4 h did not increase compared with the Control Group (as opposed to the increases in densities after incubation with TNF at 4 h). Thus, the results in Figures 5 and 6 indicate that the cGMP-PKG-mediated increase in AP-1 DNA-binding activity is dependent on concurrent pathways induced by TNF.

View larger version (46K): [in this window] [in a new window]   Figure 6.   PKG without TNF does not activate pulmonary endothelial DNA-binding activity of AP-1. Figure shows a representative EMSA indicating densities of the AP-1 band at 0.5 and 4 h in the Control (Con), TNF, and 8-Bromo-cGMP (cGMP) Groups. *Significantly different from Control Group value.

TNF Induced AP-1-Dependent Pulmonary Endothelial Gene Transcription via NO, Guanylate Cyclase, and PKG

We wanted to verify that TNF-activated, AP-1-dependent gene regulation has a similar pharmacologic profile as AP-1 DNA-binding activity and PKG activation as indicated earlier. Figure 7 shows the CAT/GAL activity ratio of PEM cotransfected with the CAT (AP-1 construct) and the GAL (control construct) reporter plasmids. After transfection, the PEM were incubated for 4 h with TNF, aminoguanidine, L-arginine, ODQ, 8-bromo-cGMP, 8-bromo-cGMP-thioate, 8-bromo-cAMP, and control media (n = 4-6 for each group). The CAT/GAL activity ratio was greater in the TNF Group than in the Control Group, which verifies the ability of TNF-activated AP-1 to alter gene expression. TNF did not induce an increase in the CAT/GAL activity ratio in the Aminoguanidine + TNF Group but did increase the CAT/GAL activity ratio in the L-Arginine + Aminoguanidine + TNF Group, verifying that TNF-induced AP-1 activation is mediated by NO synthase. TNF did not induce an increase in the CAT/ GAL activity ratio in the ODQ + TNF and 8-Bromo-cAMP + ODQ + TNF Groups; however, there was an increase in the CAT/GAL activity ratio in the 8-Bromo-cGMP + ODQ + TNF Group. Moreover, 8-bromo-cGMP or 8-bromo-cAMP given alone had no effect on the CAT/ GAL activity ratio. These results indicate that cGMP- induced, AP-1-dependent gene expression is TNF-dependent and not related to cAMP. The CAT/GAL activity ratio was lower in the 8-Bromo-cGMP-thioate Group than in the Control Group and did not increase in response to TNF. Thus, the results shown in Figure 7 corroborate the conclusion that a  [NO]   guanylate cyclase activity   [cGMP]   PKG activity pathway mediates TNF-induced increases in AP-1-mediated gene transcription. In PEM transfected with only the GAL reporter there was no consistent effect of any of the treatments on GAL activity (data not shown), indicating no nonspecific effect on transcription.

View larger version (46K): [in this window] [in a new window]   Figure 7.   TNF induces AP-1-dependent pulmonary endothelial gene transcription via NO and PKG. Figure shows the CAT/ GAL activity ratio of PEM cotransfected with the CAT and GAL reporter plasmids. In the transfected PEM, the assays of CAT and GAL were performed after 4 h of treatment with TNF, aminoguanidine, L-arginine, ODQ, 8-bromo-cGMP, 8-bromo-cGMP-thioate, 8-bromo-cAMP, and control media. The ratios of CAT and GAL activity (CAT/GAL) compared with the Control Group are shown. *Significantly different from the Control Group.

TNF Activated DNA Binding of NF-B Independent of NO

We studied the effect of TNF treatment on NF-B DNA binding activity to verify that the result of NO synthase inhibition was not due to a nonspecific effect on TNF- induced DNA binding of transcription factors. Figure 8 shows a representative EMSA (n = 2) of nuclear lysate isolated from PEM incubated for 4 h with TNF, aminoguanidine, L-arginine, and control media. The density of the NF-B band increased in all the groups treated with TNF, indicating that the effect of aminoguanidine on DNA binding is not due to nonspecific effects on transcription factor activity.

View larger version (92K): [in this window] [in a new window]   Figure 8.   TNF activates pulmonary endothelial DNA-binding activity of NF-B independent of NO. PEM were coincubated with control (CON) media, aminoguanidine (AMINO), L-arginine (L-ARG), and TNF for 4 h before the EMSA for NF-B. A representative EMSA of two experiments is shown. Lanes are: 1 = No sample, 2 = CON, 3 = AMINO, 4 = TNF, 5 = AMINO + TNF, and 6 = L-ARG + AMINO + TNF. *Significantly different from Control Group value.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

NO Mediated PKG Activation in Response to TNF

In the present study, TNF induced an increase in DNA binding of AP-1 that was associated with an increase in PKG activity. The TNF-induced increase in PKG activity was prevented by the NO synthase inhibitor aminoguanidine, the guanylate cyclase inhibitor ODQ, and the PKG inhibitors KT5823 and 8-bromo-cGMP-thioate. L-Arginine reversed the effect of aminoguanidine on TNF-induced PKG activation. However, SOD did not prevent TNF- induced activation of PKG. Thus, the data indicate that NO is necessary for TNF-induced activation of guanylate cyclase, which results in an increase in PKG activity in response to TNF. NO can directly activate guanylate cyclase (14) and increase cGMP, possibly via formation of RSNO (14, 15).

PKG with Concurrent Pathways Mediated AP-1 Activation in Response to TNF

The TNF-induced increase in DNA binding of AP-1 was prevented by the guanylate cyclase inhibitor ODQ and the PKG inhibitors KT5823 and 8-bromo-cGMP-thioate. Thus, the data indicate that guanylate cyclase-mediated activation of PKG is necessary for TNF-induced activation of AP-1. Mechanisms for a change in activation of AP-1 include an alteration in the phosphorylation, redox state, and quantity of the AP-1 subcomponents (i.e., cFOS and cJUN ) (10, 11, 16, 17). PKG may mediate the direct phosphorylation of the AP-1 subcomponents; however, a more likely mechanism is modulation of downstream transduction pathways (e.g., kinases, phosphatases, oxidant generation; 6, 10, 11, 27) by PKG.

The present data indicate that TNF-induced DNA-binding activity of AP-1 requires the  [NO]   guanylate cyclase activity   [cGMP]   PKG activity pathway; however, the integration of PKG with other signal transduction pathways is noted. First, TNF induced an increase in PKG activity at 0.5 but not at 4 h; whereas DNA-binding activity of AP-1 increased significantly only between 2 and 4 h. Second, 8-bromo-cGMP given alone did not induce significant activation of AP-1 despite the fact that 8-bromo-cGMP increased activity of PKG. However, the same dose of 8-bromo-cGMP prevented the suppression of TNF-induced activation of PKG, AP-1, and gene transcription in the ODQ-treated groups. Finally, spermine-NO alone did not induce the activation of AP-1; however, the same dose of spermine-NO prevented the suppression of TNF-induced activation of AP-1 in the aminoguanidine-treated groups. These data indicate that the effect of cGMP is dependent on coincubation with TNF; thus, recruitment of pathways in addition to PKG is required for TNF-induced AP-1 activation.

Reactive Oxygen Species Are a Concurrent Pathway Mediating AP-1 Activation in Response to TNF

The results of the pharmacologic profile support the hypothesis that O₂, possibly via ONOO, is a concurrent pathway that mediates the TNF-induced increase in DNA-binding activity of AP-1 (8, 14) for the following reasons. First, TNF-induced DNA-binding activity of AP-1 was prevented by the NO synthase inhibitor aminoguanidine and the O₂ scavenger SOD, reagents that prevent formation of the reactants for generation of ONOO. Second, the inhibition of NO synthase and SOD was specific to their respective effects on NO and O₂ because L-arginine reversed the effect of aminoguanidine and there was no effect of heat-inactivated SOD. SOD had no effect on the TNF-induced increase in PKG activity, indicating that the protective effect of SOD is via removal of the reactant O₂, rather than the scavenging of NO. Finally, KT5823 prevented SIN-1-induced AP-1 activation, which supports the hypothesis that PKG mediates the response to TNF-induced generation of NO and O₂. We demonstrated previously that a similar dose of SOD prevents TNF- induced alterations in glutathione, NO, and ONOO (13, 22), indicating that the activity of O₂ is involved in the response to TNF.

TNF has also been shown to induce the release of O₂ (28) and induce lipid peroxidation (22) in endothelial cells. In addition, our data indicate that TNF incubation for 4 h induces the accumulation of extracellular NO to a final concentration of 3.7 µM (12), indicating a rate of NO accumulation of ~ 2.54 nM/min. ONOO is generated by the reaction of NO with O₂, which has a rate constant of ~ 6.7 × 10⁹ M¹s¹ (14). In the present study, the extracellular SOD concentration is ~ 77 nM because SOD is not cell-permeable (approximate molecular weight = 31,200 kD) (29). SOD reacts with O₂ with a rate constant of ~ 2.0 × 10⁹ M¹s¹ (29). Thus, SOD functioning as an online "O₂-sink" can theoretically compete with NO for the reaction with O₂, preventing generation of ONOO. The notion that O₂, possibly via ONOO, is an additional pathway that mediates TNF-induced AP-1 activation is supported by the observation that SIN-1 induced an increase in AP-1 activation that was prevented by SOD. Also, we have new evidence using pulmonary microvessel endothelial cell lysates indicating that TNF caused an 11% increase in Western blot nitrotyrosine immunoreactivity, a probable indicator of ONOO (29). Finally, an exclusive role for NO during TNF-induced AP-1 activation is unlikely because spermine-NO alone had no effect on AP-1 activation but still reversed the effect of aminoguanidine on TNF-induced AP-1 activation. We showed previously that the identical treatment of PEM with spermine-NO induced the accumulation of extracellular NO to a final concentration of 2.9 µM (similar to TNF-induced NO accumulation) (12). This dose of spermine-NO will generate NO at a rate of 25 to 0.8 nM/min based on a half-life of 39 min at 37°C (30). The rate of NO accumulation and the extracellular site of action of SOD suggest that both NO and O₂ are generated by membrane-bound enzymes such as endothelial NO synthase (8) and nicotinamide adenine dinucleotide phosphate oxidase (28), respectively. Manna and coworkers showed that Mn-SOD prevents TNF-induced AP-1 activation in the human breast-cancer cell line MCF-7 (20). It is well known that O₂ activates AP-1 (10), but its role with NO in response to TNF is not known. The NO + O₂   ONOO pathway can alter other components (6, 10, 11, 27, 31, 32), which may contribute to TNF-induced AP-1 activation. Specifically, the role of ONOO-induced oxidation, nitrosylation and nitration of AP-1 during TNF-induced AP-1 activation is not clear (10, 14) and is a subject of our future investigations. H₂O₂ is not a critical component of TNF-induced AP-1 activation under our conditions because SOD mediates the dismutation of O₂ to H₂O₂ (14).

Another recruited-concurrent component required for TNF-induced AP-1 activation may be PKC. Our previous studies indicate that TNF activates PKC and our studies in progress indicate that the PKC inhibitor calphostin C (200-1,000 nM) significantly reduces TNF-induced DNA-binding activity of AP-1 (13, 22). Phorbol ester-induced activation of PKC and reactive oxygen species classically modulates the activation of AP-1 (10, 27) via JUN aminotermini kinase activation (31).

The data indicate that the effect of reactive nitrogen species inhibition was not due to a nonspecific effect on the TNF response because aminoguanidine and L-arginine did not affect the increase in NF-B activity in response to TNF and there was similar cell viability (i.e., percent of trypan blue, cell counts, and nuclear protein levels) among all the groups. It has been reported that NO inhibits TNF-induced activation of NF-B both in systemic endothelial cells and in astrocytes (33), possibly by stabilization of IB (5, 7, 21, 34). Thus, the role of reactive nitrogen species on the activity of NF-B is dependent on the model used for investigation (4, 34).

Potential for "Cross Talk" between PKG and PKA during TNF-Induced AP-1 Activation

We needed to characterize the effect of the pharmacologic reagents used in the present study on the cAMP-PKA system because of the potential "cross talk" between the cGMP-PKG and cAMP-PKA pathways (23). Our data indicate that the cAMP-PKA system does not confound interpretation of the role for PKG activation because neither KT5823 nor TNF altered PKA activity; whereas KT5823 prevented TNF-induced AP-1 activation. Moreover, 8-bromo-cAMP did not change the inhibitory effect of ODQ on TNF-induced gene transcription despite the fact that the same dose of 8-bromo-cAMP increased PKA activity in our system. Interestingly, there was a trend for aminoguanidine and ODQ to increase PKA activity in response to TNF that was prevented by L-arginine and 8-bromo-cGMP, respectively. Thus, the notion that NO-induced, cGMP-mediated suppression of PKA activity may have a permissive role in AP-1 activation in response to TNF cannot be ruled out. A TNF-induced increase in the level of cGMP can decrease the activity of PKA, possibly via the activation of the cGMP-stimulated phosphodiesterase Type II (23, 31). These data underscore the complexities of signal transduction and indicate that the "cross talk" among specific pathways must be considered in a biologic response.

Reactive Nitrogen Species and PKG Mediated AP-1- Dependent Gene Transcription in Response to TNF

We verified the ability of TNF-activated AP-1 to alter gene regulation because TNF-treated PEM transfected with the AP-1-CAT reporter plasmid exhibited greater CAT/GAL activity than did the Control Group. The TNF-induced increase in CAT activity was dependent on the increase in AP-1 activity because the CAT activity was corrected by the expressed activity of an AP-1-independent reporter plasmid for GAL. NO synthase inhibition did not affect constitutive CAT activity; however, PKG inhibition decreased constitutive CAT activity. It is possible that subtle differences in the time course (0-4 h) for the effects of the antagonists caused the different constitutive expression of CAT. However, the inhibition of NO synthase, guanylate cyclase, and PKG prevented the increase in CAT activity in response to TNF. The effects of aminoguanidine and ODQ on the response to TNF were ablated by cotreatment with L-arginine and 8-bromo-cGMP, respectively. The specificity of the role of cGMP was further verified because 8-bromo-cAMP did not affect CAT expression, nor did it ablate the effect of ODQ. Thus, the data indicate that the  [NO]   guanylate cyclase activity   PKG activity pathway modulates AP-1-dependent gene transcription in response to TNF.

Summary

The data indicate the novel observation that TNF-induced DNA-binding and transcription activity of AP-1 is mediated by a  [NO]   guanylate cyclase activity   [cGMP]   PKG activity pathway in pulmonary microvessel endothelial cells. Moreover, the TNF-induced activation of AP-1 is dependent on concurrent pathways via the activities of both NO and O₂, perhaps as a result of the generation of ONOO. Thus, NO-mediated increases in PKG activity are a mechanism for the modulation of AP-1-mediated gene transcription by TNF.