Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Nitric Oxide Inhibits Macrophage-Colony Stimulating Factor Gene Transcription in Vascular Endothelial Cells

1995; Elsevier BV; Volume: 270; Issue: 28 Linguagem: Inglês

10.1074/jbc.270.28.17050

ISSN

1083-351X

Autores

Haibing Peng, Tripathi B. Rajavashisth, Peter Libby, James K. Liao,

Tópico(s)

Peroxisome Proliferator-Activated Receptors

Resumo

Macrophage-colony stimulating factor (M-CSF) contributes to atherogenesis by regulating macrophage-derived foam cells in atherosclerotic lesions. Here we report that nitric oxide (NO) inhibits the expression of M-CSF in human vascular endothelial cells independent of guanylyl cyclase activation. The induction of M-CSF mRNA expression by either oxidized low density lipoprotein (ox-LDL) or tumor necrosis factor-α (TNFα) was attenuated by NO donors, S-nitrosoglutathione (GSNO), sodium nitroprusside (SNP), and 3-morpholinosydnonimine, but not by cGMP analogues, glutathione, or nitrite. Inhibition of endogenous NO production by N-monomethyl-L-arginine (L-NMA) also increased M-CSF expression in control and TNFα-stimulated cells. Nuclear run-on assays and transfection studies using M-CSF promoter constructs linked to chloramphenicol acetyltransferase reporter gene indicated that NO repressed M-CSF gene transcription through nuclear factor-κB (NF-κB). Electrophoretic mobility shift assays demonstrated that activation of NF-κB by L-NMA, ox-LDL, and TNFα was attenuated by GSNO and SNP, but not by glutathione or cGMP analogues. Since the induction of M-CSF expression depends upon NF-κB activation, the ability of NO to inhibit NF-κB activation and M-CSF expression may contribute to some of NO's antiatherogenic properties. Macrophage-colony stimulating factor (M-CSF) contributes to atherogenesis by regulating macrophage-derived foam cells in atherosclerotic lesions. Here we report that nitric oxide (NO) inhibits the expression of M-CSF in human vascular endothelial cells independent of guanylyl cyclase activation. The induction of M-CSF mRNA expression by either oxidized low density lipoprotein (ox-LDL) or tumor necrosis factor-α (TNFα) was attenuated by NO donors, S-nitrosoglutathione (GSNO), sodium nitroprusside (SNP), and 3-morpholinosydnonimine, but not by cGMP analogues, glutathione, or nitrite. Inhibition of endogenous NO production by N-monomethyl-L-arginine (L-NMA) also increased M-CSF expression in control and TNFα-stimulated cells. Nuclear run-on assays and transfection studies using M-CSF promoter constructs linked to chloramphenicol acetyltransferase reporter gene indicated that NO repressed M-CSF gene transcription through nuclear factor-κB (NF-κB). Electrophoretic mobility shift assays demonstrated that activation of NF-κB by L-NMA, ox-LDL, and TNFα was attenuated by GSNO and SNP, but not by glutathione or cGMP analogues. Since the induction of M-CSF expression depends upon NF-κB activation, the ability of NO to inhibit NF-κB activation and M-CSF expression may contribute to some of NO's antiatherogenic properties. The 1The abbreviations used are: M-CSFmacrophage colony stimulating factorNOnitric oxideGSNOS-nitrosoglutathioneTNFαtumor necrosis factor αox-LDLoxidized low density lipoproteinRSVRous sarcoma virusTBARSthiobarbituric acid reactive substancesCATchloramphenicol acetyltransferaseSNPsodium nitroprussideL-NMAN-monomethyl-L-arginine. 1The abbreviations used are: M-CSFmacrophage colony stimulating factorNOnitric oxideGSNOS-nitrosoglutathioneTNFαtumor necrosis factor αox-LDLoxidized low density lipoproteinRSVRous sarcoma virusTBARSthiobarbituric acid reactive substancesCATchloramphenicol acetyltransferaseSNPsodium nitroprussideL-NMAN-monomethyl-L-arginine. activation of mononuclear phagocytes in the vessel wall is an important event in atherogenesis(1Libby P. Clinton S.K. Curr. Opin. Lipidol. 1993; 4: 355-363Crossref Scopus (124) Google Scholar). Macrophage-colony stimulating factor (M-CSF)1 regulates macrophage growth (2Stanley E.R. Chen D.M. Lin H.S. Nature. 1978; 274: 168-170Crossref PubMed Scopus (151) Google Scholar) and differentiation (3Munn D.H. Armstrong E. Cancer Res. 1993; 53: 2603-2613PubMed Google Scholar) and may contribute to the development of macrophage-derived foam cells in atherosclerotic lesions (4). Expression of M-CSF in vascular endothelial cells is induced by minimally modified low density lipoprotein (LDL) (5Rajavashisth T.B. Andalibi A. Territo M.C. Berliner J.A. Navab M. Fogelman A.M. Lusis A.J. Nature. 1990; 344: 254-257Crossref PubMed Scopus (608) Google Scholar) and various cytokines such as interleukin-1 and TNFα(6Clinton S.K. Underwood R. Hayes L. Sherman M.L. Kufe D.W. Libby P. Am. J. Pathol. 1992; 140: 301-316PubMed Google Scholar). Atherosclerotic lesions contain both oxidized lipids (7Ylä-Herttuala S. Palinski W. Rosenfeld M.E. Parthasarathy S. Carew T.E. Butler S. Witztum J.L. Steinberg D. J. Clin. Invest. 1989; 84: 1086-1095Crossref PubMed Google Scholar) and inflammatory cytokines (8Tipping P.G. Hancock W.W. Am. J. Pathol. 1993; 142: 1721-1728PubMed Google Scholar) which may induce the local expression of M-CSF. Indeed, human and rabbit atherosclerotic lesions contain increased levels of M-CSF compared to normal arterial tissues(5Rajavashisth T.B. Andalibi A. Territo M.C. Berliner J.A. Navab M. Fogelman A.M. Lusis A.J. Nature. 1990; 344: 254-257Crossref PubMed Scopus (608) Google Scholar, 6Clinton S.K. Underwood R. Hayes L. Sherman M.L. Kufe D.W. Libby P. Am. J. Pathol. 1992; 140: 301-316PubMed Google Scholar, 7Ylä-Herttuala S. Palinski W. Rosenfeld M.E. Parthasarathy S. Carew T.E. Butler S. Witztum J.L. Steinberg D. J. Clin. Invest. 1989; 84: 1086-1095Crossref PubMed Google Scholar). Consequently, factors which regulate the expression of M-CSF may modulate atherogenesis. macrophage colony stimulating factor nitric oxide S-nitrosoglutathione tumor necrosis factor α oxidized low density lipoprotein Rous sarcoma virus thiobarbituric acid reactive substances chloramphenicol acetyltransferase sodium nitroprusside N-monomethyl-L-arginine. macrophage colony stimulating factor nitric oxide S-nitrosoglutathione tumor necrosis factor α oxidized low density lipoprotein Rous sarcoma virus thiobarbituric acid reactive substances chloramphenicol acetyltransferase sodium nitroprusside N-monomethyl-L-arginine. Nitric oxide exerts many antiatherogenic actions via stimulation of guanylyl cyclase activity(9Garg U.C. Hassid A. J. Clin. Invest. 1989; 83: 1774-1777Crossref PubMed Scopus (1991) Google Scholar, 10Radomski M.W. Palmer R.M. Moncada S. Br. J. Pharmacol. 1987; 92: 639-646Crossref PubMed Scopus (1006) Google Scholar, 11Bath P.M. Hassall D.G. Gladwin A.M. Palmer R.M. Martin J.F. Arterioscler. Thromb. 1991; 11: 254-260Crossref PubMed Scopus (300) Google Scholar). Abnormal endothelial-derived nitric oxide activity contributes to impaired vascular responses in atherosclerotic vessels of humans and animals(12Bossaler C. Habib G.B. Yamamoto H. Williams C. Wells S. Henry P.D. J. Clin. Invest. 1987; 79: 174-179Crossref Scopus (430) Google Scholar, 13Tanner F.C. Noll G. Boulanger C.M. Lüscher T.F. Circulation. 1991; 83: 2012-2020Crossref PubMed Scopus (303) Google Scholar). Inhibition of endogenous NO production by Nω-nitro-L-arginine methyl ester promotes vasoconstriction and endothelial-leukocyte adhesion, processes which are mitigated, to some extent, by addition of cGMP analogues(14Kubes P. Suzuki M. Granger D.N. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 4651-4655Crossref PubMed Scopus (2808) Google Scholar, 15Kurose I. Kubes P. Wolf R. Anderson D.C. Paulson J. Miyasaka M. Granger D.N. Circ. Res. 1993; 73: 164-171Crossref PubMed Scopus (337) Google Scholar). Furthermore, enriching the diets of cholesterol-fed rabbits with L-arginine, the precursor of NO, improves endothelial-dependent relaxation, reduces leukocyte attachment to the endothelial surface, and limits the extent of atherosclerotic lesions(16Cooke J.P. Singer A.H. Tsao P. Zera P. Rowan R.A. Billingham M.E. J. Clin. Invest. 1992; 90: 1168-1172Crossref PubMed Scopus (635) Google Scholar). Although many effects of nitric oxide are attributed to its stimulation of guanylyl cyclase, little is known regarding other cellular pathway(s) mediated by nitric oxide. The findings of our recent study indicate that the regulation of endothelial vascular cell adhesion molecule-1 expression by NO is not mediated by cGMP, but rather is associated with the inhibition of nuclear binding protein, NF-κB(17De Caterina R. Libby P. Peng H.-B. Thannickal V.J. Rajavashisth T.B. Gimbrone M.A. Shin W.S. Liao J.K. J. Clin. Invest. 1995; 96: 60-68Crossref PubMed Scopus (1573) Google Scholar). The induction of various inflammatory cytokines important in atherogenesis requires activation of NF-κB(18Libermann T.A. Baltimore D. Mol. Cell. Biol. 1990; 10: 2327-2334Crossref PubMed Google Scholar, 19Kunsch R. Lang R.K. Rosen C.A. Shannon M.F. J. Immunol. 1994; 153: 153-164PubMed Google Scholar, 20Neish A.S. Williams A.J. Palmer H.J. Whitley M.Z. Collins T. J. Exp. Med. 1992; 176: 1583-1593Crossref PubMed Scopus (384) Google Scholar). NF-κB was originally described as a heterodimeric cytosolic protein in B-cells which, upon activation, translocated into the nucleus where it binds to specific decameric sequences in the IgG κ light chain enhancer(21Sen R. Baltimore D. Cell. 1986; 46: 705-716Abstract Full Text PDF PubMed Scopus (1924) Google Scholar). Subsequent studies have shown that this pleiotropic binding protein can also activate viral enhancer elements as well as transcriptionally induce the expression of many proinflammatory cytokines and cellular adhesion molecules(22Nabel G. Baltimore D. Nature. 1987; 326: 711-713Crossref PubMed Scopus (1448) Google Scholar, 23Lenardo M.J. Baltimore D. Cell. 1989; 58: 227-229Abstract Full Text PDF PubMed Scopus (1255) Google Scholar, 24Baeuerle P.A. Biochim. Biophys. Acta. 1991; 1072: 63-80PubMed Google Scholar). The NF-κB family includes p65, p105/p50, p100/p52, c-rel, and relB which bind as homo- or heterodimers to promoter regions of target genes(23Lenardo M.J. Baltimore D. Cell. 1989; 58: 227-229Abstract Full Text PDF PubMed Scopus (1255) Google Scholar, 24Baeuerle P.A. Biochim. Biophys. Acta. 1991; 1072: 63-80PubMed Google Scholar). In endothelial cells, NF-κB consists predominantly of the p65 and p50 heterodimer(25Collins T. Lab. Invest. 1993; 68: 499-508PubMed Google Scholar). Since cellular adhesion molecules and proinflammatory cytokines participate in atherogenesis and share common κB binding motifs in their transcriptional promoters, we hypothesized that NO may regulate their gene expression through NF-κB. This study, therefore, tested whether NO could regulate the expression of an important proatherogenic molecule, M-CSF, through NF-κB. All standard culture reagents were obtained from JRH Bioscience (Lenexa, KS). Glutathione, nitrite, sodium nitroprusside, dimethyl sulfoxide, dithiothreitol, L-arginine, heparin sulfate, cupric sulfate (CuSO4), polymyxin B, butylated hydroxytoluene, thiobarbituric acid, and 1,1,3,3-tetramethoxypropane, phenylmethylsulfonyl fluoride, and cGMP analogues, 8-bromo-cGMP and dibutyryl cGMP, were purchased from Sigma. GSNO was synthesized from glutathione and nitrite as described previously(26Kowaluk E.A. Fung H.-L. J. Pharmacol. Exp. Ther. 1990; 255: 1256-1264PubMed Google Scholar). Purified human low density lipoprotein (LDL, Lot No. 730793) and N-monomethyl-L-arginine (L-NMA) were obtained from Calbiochem. The Limulus amebocyte lysate kinetic assay was performed by BioWhittaker (Walkersville, MD). Recombinant human TNFα was purchased from Endogen, Inc. (Boston, MA). [α-32P]CTP (3000 Ci/mmol), [γ-32P]ATP (3000 Ci/mmol), 32Pi (1000 Ci/mmol), and [3H]chloramphenicol (37 Ci/mmol) were supplied by DuPont NEN. The oligonucleotide corresponding to the two tandem κB sequences in the M-CSF promoter was synthesized by Genosys Biotechnologies, Inc. (The Woodlands, TX). Rabbit polyclonal antisera to NF-κB subunits, p65 and p50, were obtained from Santa Cruz Biotechnologies (Santa Cruz, CA). Nylon transfer membranes were purchased from Schleicher and Schuell. The expression vectors containing the RSV promoter linked to NF-κB subunits, p65 and p50, were kindly provided by G. Nabel (University of Michigan). The human M-CSF promoter constructs linked to the chloramphenicol acetyltransferase (CAT) reporter gene were generously provided by D. Kufe (Dana Farber Cancer Institute, Boston, MA). Human saphenous vein and bovine aortic endothelial cells were cultured and characterized as described previously(27Liao J.K. Shin W.S. Lee W.Y. Clark S.L. J. Biol. Chem. 1995; 270: 319-324Abstract Full Text Full Text PDF PubMed Scopus (533) Google Scholar). Only endothelial cells of less than three passages were used. Cells were pretreated with NO donors for 30 min prior to addition of LDL or TNFα. Cellular viability was determined by morphology and trypan blue exclusion. Native LDL (density 1.02-1.06 g/ml) from a single donor was isolated using a sequential ultracentrifugation method in the presence of butylated hydroxytoluene and polymyxin B as described previously(28Liao J.K. J. Biol. Chem. 1994; 269: 12987-12992Abstract Full Text PDF PubMed Google Scholar). Its identity was confirmed by SDS-polyacrylamide gel electrophoresis. Cholesterol, triglyceride, and protein content were determined as described previously(27Liao J.K. Shin W.S. Lee W.Y. Clark S.L. J. Biol. Chem. 1995; 270: 319-324Abstract Full Text Full Text PDF PubMed Scopus (533) Google Scholar). Oxidized LDL (80% lipid, 20% protein) was prepared by exposing samples of native LDL to CuSO4 (5 μM) at 37°C for 2 to 24 h. Both native and oxidized LDL were dialyzed with three changes of sterile buffer (150 mM NaCl, 0.01% EDTA, and 100 μg/ml polymyxin B, pH 7.4) before filtering through a 0.2-μm membrane. The degree of LDL oxidation was estimated by measuring the amounts of thiobarbituric acid reactive substances (TBARS) produced using a colorimetric assay standardized with malondialdehyde(29Yagi K. Biochem. Med. 1976; 15: 212-216Crossref PubMed Scopus (2044) Google Scholar). The TBARS value is expressed as nanomoles of malondialdehyde per mg of LDL protein. RNA was extracted using guanidinium isothiocyanate and purified by cesium chloride ultracentrifugation (30). Equal amounts of total RNA (20 μg/lane) were separated by 1.2% formaldehyde-agarose gel electrophoresis, transferred overnight onto nitrocellulose membrane by capillary action, and baked (72°C) for 2 h prior to prehybridization. Radiolabeling of a 1.8-kilobase human M-CSF cDNA probe was performed using random hexamer priming with [α-32P]CTP and a Klenow fragment of DNA polymerase I (Pharmacia Biotech). The membranes and probe were hybridized overnight at 52°C in a buffer containing 50% formamide, 5 × SSC, 2.5 × Denhardt's solution, 25 mM sodium phosphate buffer (pH 6.5), 0.1% SDS, and 250 μg/ml salmon sperm DNA and washed in 0.2 × SSC, 0.1% SDS at 65°C before autoradiography at −80°C for 24-72 h. All blots were subsequently rehybridized with β-actin cDNA probe as an internal control (ATCC 37997, Rockville, MD). Confluent endothelial cells (5 × 106) were incubated with 32Pi (500 μCi) for 1 h prior to the addition of 8-bromo-cGMP at the indicated concentrations and incubated for an additional 1 h. The study was terminated by the addition of sodium phosphate (50 mM), trichloroacetic acid (20%), and sodium vanadate (1 mM). Cells were scraped and lysed by a Dounce homogenizer. Protein concentrations from cellular extracts were determined by the method of Lowry et al.(31Lowry O.H. Rosebrough N.J. Farr A.L. Randall R.J. J. Biol. Chem. 1951; 193: 265-275Abstract Full Text PDF PubMed Google Scholar). Proteins (50 μg) were suspended in denaturing buffer containing Tris-HCl (125 mM, pH 6.8), SDS (4%), glycerol (20%), and 2-mercaptoethanol (10%) and centrifuged at 12,000 × g for 10 min. The supernatants and known molecular weight markers (Bethesda Research Laboratory) were separated by SDS-polyacrylamide gel electrophoresis (10% running, 4% stacking gel). The gels were then fixed with Coomassie Blue (0.4%), methanol (20%), and glacial acetic acid (10%) and dried by a gel dryer before autoradiography at −70°C for 12-24 h. Nuclei from 108 endothelial cells were prepared, and in vitro transcription with [32P]UTP was performed as described(27Liao J.K. Shin W.S. Lee W.Y. Clark S.L. J. Biol. Chem. 1995; 270: 319-324Abstract Full Text Full Text PDF PubMed Scopus (533) Google Scholar). Linearized plasmids (1 μg) containing M-CSF, pGEM (Promega), and rat β-tubulin cDNAs were immobilized on nylon membranes using a vacuum-transfer slot blot apparatus (Schleicher & Schuell), and the membranes were hybridized to radiolabeled transcripts (∼5-8 × 107 cpm/ml) at 45°C for 48 h in a buffer containing 50% formamide, 5 × SSC, 2.5 × Denhardt's solution, 25 mM sodium phosphate buffer (pH 6.5), 0.1% SDS, and 250 μg/ml salmon sperm DNA. The membranes were then washed with 1 × SSC, 0.1% SDS for 1 h at 65°C before autoradiography for 72 h at −80°C. For transient transfections, bovine rather than human endothelial cells were used because of their higher transfectional efficiency by the calcium-phosphate precipitation method (32). Two different M-CSF promoter constructs, [-565]M1 and [-248]M4, linked to the chloramphenicol acetyltransferase (CAT) gene were used(33Yamada H. Iwase S. Mohri M. Kufe D. Blood. 1991; 78: 1988-1995Crossref PubMed Google Scholar). Cells were transfected with the indicated promoter constructs (30 μg): p.CAT (no promoter), pSV2.CAT (SV40 early promoter), M1, or M4. Approximately 60 h after transfection, cells were treated with ox-LDL (50 μg/ml) or TNFα (10 ng/ml). For co-transfection studies with RSVp65 and RSVp50, GSNO (0.2 mM) was added 12 h after transfection, and media were changed and GSNO was renewed every 12 h. As an internal control for transfection efficiency, pRSV.βGAL plasmid (10 μg) was co-transfected in all experiments. Seventy-two hours after transfection, cellular extracts were prepared using lysis buffer (100 μg/ml leupeptin, 50 μg/ml aprotinin, 0.1 mM phenylmethylsulfonyl fluoride, 5 mM EDTA, 5 mM EGTA, 100 mM NaCl, 5 mM Tris-HCl, pH 7.4) and one freeze-thaw cycle. CAT activity was determined by incubating the cellular extracts (100 μl) with [3H]chloramphenicol (50 μCi/ml) and n-butyryl coenzyme A (250 μg/ml) for 20 h at 37°C as described previously(34Seed B. Sheen J.-Y. Gene (Amst.). 1988; 67: 271Crossref PubMed Scopus (830) Google Scholar). The relative CAT activity was calculated as the ratio of CAT to β-galactosidase activity. M-CSF promoter activity (-fold induction) was expressed as the ratio of relative CAT activity to the relative basal CAT activity of [-565]M1.CAT. Each experiment was performed three times in duplicate. Nuclear extracts were prepared as described(35Dignam J.D. Lebovitz R.M. Roeder R.G. Nucleic Acids Res. 1983; 11: 1475-1489Crossref PubMed Scopus (9150) Google Scholar). The NF-κB oligonucleotide corresponding to the two tandem κB sites in the M-CSF promoter (GGGGATTTTCAGGGCC TGGAGGGAAAGTCCCTT) was end-labeled with [γ-32P]ATP and T4 polynucleotide kinase (New England Biolabs) and purified by Sephadex G-50 columns (Pharmacia Biotech). Nuclear extracts (10 μg) were added to 32P-labeled NF-κB oligonucleotide (∼20,000 cpm, 0.2 ng) in buffer containing 2 μg of poly[dI·dC], 10 μg of bovine serum albumin, 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, and 5% glycerol. DNA·protein complexes were resolved on 4% nondenaturing polyacrylamide gel electrophoresed at 12 V/cm for 3 h in low ionic strength buffer (0.5 × TBE) at 4°C. For supershift assays, the indicated antibody (15 μg/ml) was added to the nuclear extracts for 10 min before addition of radiolabeled probe. In some studies, GSNO or unlabeled NF-κB oligonucleotide (20 ng) was added directly to the nuclear extracts 10 min prior to addition of radiolabeled probe. Band intensities from Northern and in vitro transcription assay blots were analyzed densitometrically by the National Institutes of Health IMAGE program(36Rasband W. NIH Image Program. Vol. 1. National Institutes of Health, Bethesda1993: 49Google Scholar). All values are expressed as mean ± S.E. compared to controls and among separate experiments. Paired and unpaired Student's t tests were employed to determine the significance of changes in CAT activity and densitometric measurements. A significant difference was considered for p values of less than 0.05. There were no observable adverse effects of oxidized LDL, GSNO, or SNP on cellular morphology, and cellular confluency (∼7 × 106 cells/T-150 cm2 flask) and viability were maintained for all treatment conditions described. LDL samples, L-NMA, and 8-bromo-cGMP had no detectable levels of endotoxin ( 0.05). In vitro transcription studies showed a moderate basal transcriptional activity of the M-CSF gene under standard tissue culture conditions (Fig. 5B). Treatment with oxidized LDL (50 μg/ml, TBARS 13.4 nmol/mg) or TNFα (10 ng/ml) increased M-CSF gene transcription 7.8- and 18-fold relative to β-tubulin gene transcription, respectively. NO essentially abolished all transcriptional activity of the M-CSF gene induced by oxidized LDL or TNFα, but did not substantially affect β-tubulin gene transcription. Preliminary studies using different amounts of radiolabeled RNA transcripts demonstrate that under our experimental conditions, hybridization was linear and nonsaturable. The density of each M-CSF band was standardized to the density of its corresponding β-tubulin band. The specificity of each band was determined by the lack of hybridization to the nonspecific pGEM cDNA vector. To characterize further the effects of NO on M-CSF gene transcription, we transfected bovine aortic endothelial cells using two M-CSF promoter constructs, M1 and M4, linked to the chloramphenicol acetyltransferase (CAT) reporter gene(33Yamada H. Iwase S. Mohri M. Kufe D. Blood. 1991; 78: 1988-1995Crossref PubMed Google Scholar). Analyses of the M-CSF promoter revealed putative DNA binding sequences for NF-κB, SP1, SSRE (shear-stress responsive element), "CAT" and "TTAAA" boxes, and initiation start site (Fig. 6A). [-565]M1 promoter contains two tandem κB sites, while the deletional [-248]M4 promoter lacks these κB sites. Stimulation with TNFα (10 ng/ml) or oxidized LDL (50 μg/ml, TBARS 13.4 nmol/mg) increased M1 promoter activity by 9.2- and 7.5-fold, respectively (Fig. 6B). Co-transfection of M1.CAT with the expression vector RSVp65 alone resulted in a 13-fold induction in M1 promoter activity compared to a 5.7-fold induction with a combination of RSVp65 and RSVp50 and 2.2-fold induction with RSVp50 alone. Treatment with GSNO (0.2 mM) was effective in decreasing M1 promoter activity induced by TNFα (65% reduction), oxidized LDL (72% reduction), and co-transfections with p65 alone (61% reduction) or in combination with p50 (52% reduction). Basal M4 promoter activity was 1.8-fold lower than that of basal M1. Co-transfection with RSVp65 with M4.CAT produced essentially no promoter activity. Stimulation with TNFα or oxidized LDL produced only a 2.5-fold induction of the non-κB containing M4 promoter activity. GSNO did not significantly affect M4 promoter activity induced by TNFα, ox-LDL, or co-transfection with RSVp65. Transcriptional repression was not due to general toxicity since GSNO did not affect basal M1 and M4 or the SV40 promoter activity. Electrophoretic mobility shift assays demonstrated rapid and near-maximal activation of NF-κB by TNFα after 30 min (Fig. 7). In contrast, the degree of NF-κB activation by native LDL (50 μg/ml, TBARS 0.2 nmol/mg) was smaller compared to TNFα and occurred only after 6 h when the measured TBARS value was 1.6 nmol/mg, presumably secondary to endothelial cell modification of native LDL. Activation of NF-κB by oxidized LDL (50 μg/ml, TBARS 13.4 nmol/mg) occurred in a time-dependent manner, and, after 6 h, resemble that of TNFα after 30 min. GSNO (0.2 mM) and sodium nitroprusside (SNP) attenuated the activation of NF-κB by both oxidized LDL and TNFα (Figure 7:, Figure 8:). Addition of 8-bromo-cGMP (1 mM) did not affect TNFα-induced activation of NF-κB suggesting that NO's inhibitory effect was not due to guanylyl cyclase activation. GSNO was effective only when added to whole cells rather than directly to nuclear extracts suggesting that NO does not interfere with NF-κB binding to DNA (Fig. 8). Activation of NF-κB was also observed when endogenous NO production was inhibited by L-NMA (1 mM). However, L-NMA produced NF-κB activation to a lesser extent than did TNFα. Preliminary studies indicate that treatment with L-NMA (1 mM) reduced basal NO synthase activity by 80% (data not shown). Although the level of NF-κB activation by L-NMA treatment was much less compared to that caused by TNFα, L-NMA-induced NF-κB activation was more completely abolished by treatment with GSNO (0.2 mM). The indicated band was specific for NF-κB since, in the presence of antibodies to p50 and p65, this band was "supershifted" and attenuated. We have shown that both endogenous and exogenous nitric oxide (NO) can limit the expression of a proatherogenic cytokine, macrophage-colony stimulating factor (M-CSF) induced by two pro-atherogenic mediators, TNFα and oxidized LDL. Since M-CSF may contribute to the development of macrophage-derived foam cells(4Rosenfeld M.E. Ylä-Herttuala S. Lipton B.A. Ord V.A. Witztum J.L. Steinberg D. Am. J. Pathol. 1992; 140: 291-300PubMed Google Scholar, 6Clinton S.K. Underwood R. Hayes L. Sherman M.L. Kufe D.W. Libby P. Am. J. Pathol. 1992; 140: 301-316PubMed Google Scholar), inhibition of M-CSF expression by NO may be one mechanism by which NO can attenuate atherogenesis. NO's inhibitory effect on M-CSF mRNA expression was not mediated by stimulation of guanylyl cyclase since cGMP analogues did not inhibit M-CSF expression. This is in contrast to other antiatherogenic effects of NO which are mediated by cGMP such as vascular smooth muscle relaxation (12Bossaler C. Habib G.B. Yamamoto H. Williams C. Wells S. Henry P.D. J. Clin. Invest. 1987; 79: 174-179Crossref Scopus (430) Google Scholar) and inhibition of platelet aggregation(10Radomski M.W. Palmer R.M. Moncada S. Br. J. Pharmacol. 1987; 92: 639-646Crossref PubMed Scopus (1006) Google Scholar). Thus, our findings provide a novel mechanism by which NO can modulate the expression of an important atherogenic cytokine, M-CSF. Actinomycin D studies and nuclear run-on assays indicated that the regulation of M-CSF expression occurred at the level of M-CSF gene transcription. Analyses of the M-CSF promoter revealed that two tandem κB binding sites located approximately 400 bp upstream from the initiation start site were necessary for full transcriptional induction by TNFα and oxidized LDL. These results agree with previous studies showing the obligatory role of nuclear binding protein NF-κB in transcriptionally activating the M-CSF promoter(33Yamada H. Iwase S. Mohri M. Kufe D. Blood. 1991; 78: 1988-1995Crossref PubMed Google Scholar). However, the deletional construct lacking the κB sites still exhibits substantial promoter activity in response to TNFα and oxidized LDL suggesting that other non-NF-κB binding proteins could also participate and perhaps act synergistically in transactivating the M-CSF promoter. Electrophoretic mobility shift assays demonstrated that the transcriptional repression of the M-CSF gene by NO was due principally to the inhibition of NF-κB activation. NO did not physically inhibit the binding of NF-κB to its cognate DNA since the addition of NO directly to nuclear extracts of TNFα-stimulated cells did not affect the activation of NF-κB. Thus, cellular factor(s) must be present in the intact cell which mediate NO's inhibitory effect on NF-κB activation. Interestingly, co-transfection with p65 alone resulted in a greater increase in M-CSF promoter activity compared to that achieved with the combination of p65 and p50. Since the p50 homodimer can bind κB sites, but is a relatively weak transactivator(37Shu H.-B. Agranoff A.B. Nabel E.G. Leung K. Duckett C.S. Neish A.S. Collins T. Nabel G.J. Mol. Cell. Biol. 1993; 13: 6283-6289Crossref PubMed Google Scholar), co-transfection with p65 and p50 presumably leads to a competition between the p65 homodimer with the p65/p50 heterodimer and the p50 homodimer for κB sites, resulting overall in less promoter activity compared to that of the p65 homodimer. Furthermore, the ability of NO to attenuate M-CSF promoter activity in cells co-transfected with either p65 alone or in combination with p50 suggests that NO's inhibitory effects are likely mediated through p65. NF-κB mediates transcriptional activation of the M-CSF gene in response to TNFα(33Yamada H. Iwase S. Mohri M. Kufe D. Blood. 1991; 78: 1988-1995Crossref PubMed Google Scholar). Recent studies indicate that the activation of NF-κB by TNFα and bacterial lipopolysaccharide involves the generation of reactive oxygen species such as superoxide anion(38Marui N. Offermann M.K. Swerlick R. Kunsck C. Rosen C.S. Ahmad M. Alexander R.W. Medford R.M. J. Clin. Invest. 1993; 92: 1866-1874Crossref PubMed Scopus (974) Google Scholar). Indeed, antioxidants such as N-acetylcysteine or pyrrolidine dithiocarbamate attenuate the activation of NF-κB(38Marui N. Offermann M.K. Swerlick R. Kunsck C. Rosen C.S. Ahmad M. Alexander R.W. Medford R.M. J. Clin. Invest. 1993; 92: 1866-1874Crossref PubMed Scopus (974) Google Scholar, 39Schreck R. Albermann K. Baeuerle P.A. Free Radical Res. Commun. 1992; 17: 221-237Crossref PubMed Scopus (1297) Google Scholar). NF-κB, therefore, is an attractive candidate for inhibition by nitric oxide (NO) since, under certain conditions, NO can function as an antioxidant through its scavenging effects on superoxide anion(40Schreck R. Rieber P. Baeuerle P.A. EMBO J. 1991; 10: 2247-2258Crossref PubMed Scopus (3413) Google Scholar, 41Huie R.E. Padmaja S. Free Radical Res. Commun. 1993; 18: 195-200Crossref PubMed Scopus (2009) Google Scholar). NO interacts with superoxide anion to form peroxynitrite, thereby diverting superoxide anion away from its dismutation product, hydrogen peroxide(41Huie R.E. Padmaja S. Free Radical Res. Commun. 1993; 18: 195-200Crossref PubMed Scopus (2009) Google Scholar). Thus, less hydrogen peroxide would be available to activate NF-κB(38Marui N. Offermann M.K. Swerlick R. Kunsck C. Rosen C.S. Ahmad M. Alexander R.W. Medford R.M. J. Clin. Invest. 1993; 92: 1866-1874Crossref PubMed Scopus (974) Google Scholar, 39Schreck R. Albermann K. Baeuerle P.A. Free Radical Res. Commun. 1992; 17: 221-237Crossref PubMed Scopus (1297) Google Scholar, 40Schreck R. Rieber P. Baeuerle P.A. EMBO J. 1991; 10: 2247-2258Crossref PubMed Scopus (3413) Google Scholar). However, in the presence of NO, the consequence of peroxynitrite formation remains to be determined, but can lead to tyrosine phosphorylation(42Beckman J.S. Beckman T.W. Chen J. Marshall P.A. Freeman B.A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 1620-1624Crossref PubMed Scopus (6675) Google Scholar). The activation of NF-κB and increase in M-CSF expression also occurred in the presence of the NO synthase inhibitor, L-NMA. Endogenous NO production by the constitutive NO synthase may therefore tonically inhibit the expression of proinflammatory genes through suppression of NF-κB. Interestingly, both TNFα and oxidized LDL decrease the expression of endothelial NO synthase (27Liao J.K. Shin W.S. Lee W.Y. Clark S.L. J. Biol. Chem. 1995; 270: 319-324Abstract Full Text Full Text PDF PubMed Scopus (533) Google Scholar, 43Yoshizumi M. Perrella M.A. Burnett Jr., J.C. Lee M.E. Circ. Res. 1993; 73: 205-209Crossref PubMed Scopus (704) Google Scholar) which, in turn, may serve to augment the activation of NF-κB. Treatment with NO donors produced further inhibition of basal and TNFα-stimulated expression of M-CSF. Such higher levels of NO may be encountered by endothelial cells at sites of inflammation where induction of NO synthase activities in macrophages and vascular smooth muscle cells could generate concentrations of NO comparable to that given exogenously in this study(44Radomski M.W. Rees D.D. Dutra A. Moncada S. Br. J. Pharmacol. 1992; 107: 745-749Crossref PubMed Scopus (302) Google Scholar). In solution, NO has a very short half-life(42Beckman J.S. Beckman T.W. Chen J. Marshall P.A. Freeman B.A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 1620-1624Crossref PubMed Scopus (6675) Google Scholar). However, in the local environment of an inflammatory atherosclerotic lesion, NO can act at short distances, given its intracellular origin and the close proximity of endothelial cells to macrophages and vascular smooth muscle cells and, thus, be less subject to inactivation. Our findings provide a novel antiatherogenic effect of NO which is independent of its classically recognized effect on soluble guanylyl cyclase. Although we report here the inhibitory effects of NO on NF-κB activation and M-CSF expression, these effects may extend to other inflammatory cytokines and adhesion molecules which contain functional κB sites in their transcriptional promoters. We propose that NO is an important physiological mediator of both homeostasis and inflammation. We thank Drs. G. Nabel for RSVp65 and RSVp50 expression vectors and H. Yamada and D. Kufe for M-CSF promoter constructs.

Referência(s)