Benzo[a]pyrene induces the transcription of cyclooxygenase-2 in vascular smooth muscle cells. Evidence for the involvement of extracellular signal-regulated kinase and NF-kappaB.

Polycyclic aromatic hydrocarbons, such as benzo[a]pyrene (B[a]P) present in tobacco smoke and tar, have been implicated in the development of atherosclerosis as well as cancer. Increased expression of cyclooxygenase-2 (COX-2) has been detected both in atherosclerotic lesions and in epithelial cancers. To determine whether polycyclic aromatic hydrocarbons might directly affect COX expression in vascular cells, we investigated the effects of B[a]P on COX-2 expression in human and rat arterial smooth muscle cells (SMC). Treatment with B[a]P increased levels of COX-2 protein and mRNA and enhanced prostaglandin synthesis. Nuclear runoff assays and transient transfections revealed increased COX-2 gene transcription after treatment with B[a]P. Experiments were done to define the signaling mechanism by which B[a]P induced COX-2. B[a]P caused a rapid increase in phosphorylation of extracellular signal-regulated kinase (ERK); pharmacologic inhibition of mitogen-activated protein kinase kinase blocked B[a]P-mediated induction of COX-2. Depletion of the intracellular antioxidant, glutathione, with buthionine sulfoximine significantly increased B[a]P-mediated induction of COX-2 while exposure to N-acetylcysteine, a precursor of glutathione, suppressed the induction of COX-2 by B[a]P. Several lines of evidence suggest that the induction of COX-2 by B[a]P is mediated, at least in part, by NF-kappaB. Treatment with B[a]P increased binding of NF-kappaB to DNA. Moreover, B[a]P-mediated stimulation of COX-2 promoter activity was blocked when a construct containing a mutagenized NF-kappaB site was used. Pharmacological inhibitors of NF-kappaB blocked the induction of COX-2 protein and the stimulation of COX-2 promoter activity by B[a]P. Taken together, these data are likely to be important for understanding the atherogenic effects of tobacco smoke.

celerates atherosclerotic plaque formation, and increases ischemic tissue damage. Benzo[a]pyrene (B[a]P), 1 a polycyclic aromatic hydrocarbon present in tobacco smoke, is metabolized in tissue to mutagenic derivatives which form DNA adducts within target cells (4 -6). Considerable experimental evidence suggests that B[a]P accelerates smooth muscle proliferation and promotes atherosclerosis in animals ranging from chickens to rats (7)(8)(9)(10)13). In addition, we and others have detected B[a]P-related DNA adducts in atherosclerotic arteries (11)(12)(13). In some systems, B[a]P up-regulates the expression of proto-oncogenes, such as c-Ha-ras and c-myc, that favor cell proliferation (14).
Cyclooxygenase (COX) is the rate-limiting enzyme that catalyzes the oxygenation of arachidonic acid to prostaglandin endoperoxides which are converted enzymatically into prostaglandins (PGs) and thromboxane A 2 that play both physiologic and pathologic roles in vascular function. COX is capable of cooxidizing B[a]P; the peroxyl radicals formed during arachidonate metabolism catalyze the epoxidation of B[a]P to its mutagenic products (15)(16)(17). Two distinct isoforms of COX have been identified (18). COX-1 is constitutively expressed at low levels in most tissues. In contrast, COX-2, the product of a related inducible gene, is absent from most normal tissues but is expressed in response to proliferative and inflammatory stimuli such as growth factors and cytokines (19 -25). Overexpression of COX-2 in epithelial cells inhibits apoptosis and increases the invasiveness of malignant cells, favoring tumorigenesis and metastasis (22, 26 -31). COX-2 in activated human monocytes generates the isoprostane 8-epi-PGF 2 , which is mitogenic and vasoactive, leading to cellular proliferation and vasoconstriction (32). COX-2 is expressed in atherosclerotic lesions (33), increases after vascular injury (21,23), and has been detected in myocardium of patients with congestive heart failure (34). Aspirin, a COX inhibitor, can reduce the risks of cardiovascular disease (35)(36). These data all suggest that COX-2 may participate in the genesis of atherosclerosis.
This study is designed to explore the effect of B[a]P on COX-2 in human arterial vascular smooth muscle cells (SMC) from atherosclerotic lesions and in rat aortic SMC. We show that exposure to B[a]P increases COX-2 gene expression and protein production and that these effects, which are augmented by oxidant stress, are mediated by ERK1/2-MAPK signaling and the transcription factor NF-B.

MATERIALS AND METHODS
Reagents-DMEM medium was obtained from BioWhittaker (Walkersville, MA); fetal bovine serum was from Gemini-Bio-Products (Calabasas, CA). NS 398 (N-(2-cyclohexyloxy-4-nitrophenyl)methanesulfonamide), COX-2 polyclonal antibody, COX-2 standard, and enzyme immunoassay reagents were from Cayman Co. (Ann Arbor, MI). Human COX-2 cDNA was a gift from Dr. Stephen M. Prescott (University of Utah). Specific rat COX-2 primers and ␤-actin primers were made by Genosys (The Woodlands, TX). FuGENE  Cell Culture-Fisher-344 rats were from the National Institutes of Health Aging Institute Colony. Rat SMC (rSMC), explanted from the aortas of 18-month-old Fisher-344 rats, were maintained in DMEM supplemented with 5% FBS, 1ϫ minimal essential medium vitamins, 10 mM HEPES, 3 mM glutamine, penicillin (100 units/ml), and streptomycin (100 g/ml). These rSMC were previously shown to have atherogenic potential and enhanced proliferative capacity (37). Passage 10 -14 rSMC were grown to 80% confluence and then growth arrested by incubation in 0.5% FBS/DMEM for 48 h prior to use. Human arterial SMC isolated in our laboratory from atherosclerotic carotid artery lesions under Institutional Review Board-approved protocols (E12 cells, hSMC) were cultured in the same medium detailed above and were incubated in 0.5% FBS/DMEM for 48 h prior to use. In studies of both cell types, medium was removed following growth arrest and replaced by fresh 0.5% FBS/DMEM to which B[a]P (0.5-10 M) was added for 18 -24 h to rSMC or for 6 h to hSMC prior to harvest of cells.
Western Analysis-SMC were washed twice and homogenized in lysis buffer (150 mM NaCl, 100 mM Tris, pH 8.0, 1% Tween 20, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride). Protein concentrations of lysates were determined by the method of Lowry. Proteins (50 g/lane) were separated on a denaturing 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Membranes were blocked in 3% non-fat dry milk in phosphate-buffered saline containing 0.1% Tri- ton X-100. For detection of COX-2, polyclonal anti-COX-2 antibody (1:1,000) was incubated with the membranes overnight at 4°C. Secondary antibody linked to horseradish peroxidase was used at 1:10,000, and signals were visualized by the ECL technique using Kodak x-ray film. For detection of ERK1/2-MAPK total protein, the primary antibody was rabbit polyclonal anti-ERK1/2 antibody (1:1,000), whereas for detection of phosphorylated ERK1/2-MAPK, an antibody that recognizes only the phosphorylated forms of ERK1/2-MAPK was used.
Cyclooxygenase Activity-The effects of different treatments on total COX activity of SMC were determined in intact cells cultured in 24-well plates and stimulated with B[a]P for 6 h, following which medium was removed and prewarmed serum-free medium containing 10 M sodium arachidonate was added to stimulate maximum prostaglandin release. After 15 min, this medium was collected to measure formation of 6-keto-PGF1␣, the stable metabolite of PGI 2 , or prostaglandin E 2 (PGE 2 ), by specific enzyme immunoassay as described previously (23). RNA Extraction-Total cellular RNA was extracted from cells using Tri Reagent according to the manufacturer's instructions. Briefly, Tri Reagent (1 ml/T-25 flask) was added to the tissue culture flasks for 10 min. Cell lysates were harvested and centrifuged at 12,000 ϫ g for 15 min at 4°C after adding 1/10 volume of chloroform. The upper aqueous layer was mixed with an equal volume of isopropanol to precipitate RNA. RNA pellets were washed with 70% ethanol, dried, and dissolved in DEPC-treated water. RNA concentrations were determined spectrophotometrically.

Effect of Oxidant Stress on B[a]P-induced COX-2-To
RT-PCR-cDNA was reverse-transcribed from 1 g total cellular RNA using random hexamer primers and murine leukemia virus reverse transcriptase. One g of cDNA was amplified for 30 cycles using the following rat COX-2 gene-specific primers (38): 5Ј-ACTTGCCT-CACTTTGTTGAGTCATTC-3Ј (sense) and 5Ј-TTTGATTAGTACTG-TAGGGTTAATG-3Ј (antisense). The cycling parameters were the following: 30 s at 94°C for denaturation, 30 s at 60°C for primer annealing, and 1 min at 72°C for polymerization. Meanwhile, the same amount of cDNA was amplified for 20 cycles using specific ␤-actin primers: 5Ј-GAGACCTTCAACACCCC-3Ј (sense) and 5Ј-GTGGTGGT-GAAGCTGTAGCC-3Ј (antisense). The products were visualized after electrophoresis on a 2% agarose gel containing ethidium bromide.
Northern Analysis-Ten g of total cellular RNA per lane were electrophoresed in a formaldehyde-containing 1.2% agarose gel and transferred to a nitrocellulose membrane. After baking, membranes were prehybridized in a solution containing 50% formamide, 5ϫ sodium chloride/sodium phosphate/EDTA buffer (SSPE), 5ϫ Denhardt's solution, 0.1% SDS, and 100 g/ml salmon sperm DNA for 3 h and hybridized in the above solution containing 32 P-labeled COX-2 cDNA probes for 16 h at 42°C. Following hybridization, the membranes were washed twice for 20 min in 2ϫ SSPE and 0.1% SDS at room temperature, twice for 20 min in the same solution at 55°C, and twice for 20 min in 0.1ϫ SSPE and 0.1% SDS at 55°C. Washed membranes were then exposed to x-ray film. To verify the equivalency of RNA loading in the different lanes, the membranes were rehybridized with a probe for 18 S rRNA. The signal level of the bands was quantified densitometrically.
Nuclear Run-off-Serum-deprived hSMC plated in T150 cell culture dishes were grown to 80% confluence, serum-starved for 48 h, and then stimulated with 1 M B[a]P for 3 h. Nuclei were isolated and stored in liquid nitrogen until use for transcription assays as described previously (39). Briefly, nuclei were incubated in reaction buffer containing 100 Ci of uridine 5Ј-[␣-32 P]triphosphate and unlabeled nucleotides. After 30 min, labeled nascent RNA transcripts were isolated. COX-2 and 18S rRNA cDNAs were immobilized onto nitrocellulose membranes and prehybridized overnight. Hybridization was carried out at 42°C for 24 h using 5 ϫ 10 5 cpm of labeled nascent RNA transcripts. The filters were washed twice with 2ϫ SSC buffer at 55°C for 1 h and were then treated with 10 mg/ml RNase A in 2ϫ SSC at 37°C for 30 min, dried, and autoradiographed.
Transient Transfection Assay-Human SMC were seeded at a density of 1 ϫ 10 5 cells/well in 6-well culture dishes and grown for 24 h in medium containing 5% FBS. COX-2 promoter-luciferase plasmid DNA (40) and pSV-␤-galactosidase were co-transfected into cells using Fu-GENE 6 transfection reagent according to the manufacturer's protocol. After 6 h of incubation, the cells were maintained in 5% FBS/ DMEM for another 24 h and then stimulated with B[a]P for 3 h. Cells were lysed and luciferase activity was measured in the cellular extracts using an enhanced luciferase assay kit. Luciferase activity was normalized to ␤-galactosidase activity (41).
Electrophoretic Mobility Shift Assay-Human vascular smooth muscle cells were treated with B[a]P in DMEM containing 0.5% FBS for 2 h. Cells were harvested, and nuclear extracts were prepared as described previously (42). For binding studies, an NF-B consensus oligonucleotide was used with the following sequence: 5Ј-AGTTGAGGGGACTTTC-CCAGGC-3Ј (sense) and 3Ј-TCAACTCCCCTGAAAGGGTCCG-5Ј (antisense). A mutant NF-B oligonucleotide was obtained from Santa Cruz Biotechnology Inc. for additional binding studies.
The complementary oligonucleotides were annealed in 20 mM Tris (pH 7.6), 50 mM NaCl, 10 mM MgCl 2 , and 1 mM dithiothreitol. The annealed oligonucleotides were phosphorylated at the 5Ј-end with [␥-32 P]ATP and T 4 polynucleotide kinase. The binding reaction was performed by incubating 2 g of nuclear protein in 20 mM HEPES (pH 7.9), 10% glycerol, 300 g of bovine serum albumin, and 1 g of poly(dI-dC) in a final volume of 10 l for 10 min at 25°C. The labeled NF-B consensus oligonucleotide was added to the reaction mixture and allowed to incubate for an additional 20 min at 25°C. To detect an antibody supershift, 2 l of antibody to NF-B p65 or p50 were added to the reaction mixture for 30 min at 25°C. The samples were electrophoresed on a 4% nondenaturing polyacrylamide gel. The gel was then dried and subjected to autoradiography at Ϫ80°C.
Statistical Analysis-Data are presented as mean Ϯ S.D., and statistical comparisons were made using the Student's t test for paired observations or analysis of variance for multiple comparisons. Significance was defined at the p Ͻ 0.05 level.

Benzo[a]pyrene Treatment Increases COX-2 Gene Expression and Protein Synthesis in E12 Human Arterial Smooth Muscle Cells and Rat Aortic Smooth Muscle Cells and Increases Prostaglandin Production-B[a]
P treatment of hSMC produced a dose-dependent induction of COX-2 protein, maximal at 1 M B[a]P (Fig. 1A). Treatment of rSMC with B[a]P also produced a dose-and time-dependent induction of COX-2 protein (Fig. 1, B and C, respectively.) Maximal effects were observed at 5 M B[a]P and 18 h, respectively. To determine if B[a]P-induced COX-2 enzyme was functional, prostaglandin production was measured. Synthesis of PGE 2 , the main prostaglandin produced by hSMC, increased by 2-3-fold; this increase was blocked by 2 M NS 398, a known specific COX-2 inhibitor (data not shown). Serum-deprived rSMC similarly treated with B[a]P increased prostaglandin production (mainly PGI 2 ) up to 8-fold in a dose-dependent manner, with maximal production in response to 1-5 M B[a]P (data not shown).
Northern blot analysis was done to determine whether B[a]P induced COX-2 mRNA. Steady state levels of COX-2 mRNA increased after 1-3 h of exposure to B[a]P in both types of vascular SMC (Fig. 2). To determine if B[a]P-mediated induction of COX-2 mRNA involved increased transcription, de novo COX-2 mRNA synthesis was determined by nuclear runoff assay in hSMC treated for 3 h with 1 M B[a]P. Fig. 3A shows that B[a]P stimulated the transcription of COX-2. This increase was similar in extent to that induced by PMA, a known inducer of COX-2 gene expression. In complementary studies, transient transfections were performed with a human COX-2 promoter-luciferase reporter construct (Ϫ1432/ϩ59) demonstrating that exposure to B[a]P resulted in more than a 100% increase in COX-2 promoter activity (Fig. 3B).  (Fig. 4A). Markedly increased levels of COX-2 protein were detected in SMC pretreated with BSO and then exposed to B[a]P (Fig. 4A). Conversely, NAC treatment blocked B[a]P-mediated induction of COX-2 protein (Fig. 4B.)  (Fig. 5B). Treatment of rSMC with 20 M PD98059 also decreased B[a]P-mediated induction of COX-2 mRNA (Fig. 6A), COX-2 protein (Fig. 6B), and prostaglandin synthesis (Fig. 6C). The decline in both basal and B[a]P-stimulated PG synthesis in the presence of PD98059 can be attributed to the known inhibitory effect of PD98059 on enzymatic activity of COX-1 and COX-2 (43), in addition to suppression of any increase in COX-2 protein. In contrast, SB203580, an inhibitor of p38 MAPK, had little effect on B[a]P-mediated induction of COX-2 protein although it also decreased COX-2 enzymatic activity (data not shown).

Induction of COX-2 by Benzo[a]pyrene Is Mediated by ERK1/2 MAP Kinase-To define the mechanism of B[a]P-me-
Role of NF-B in Benzo[a]pyrene-induced COX-2 Expression-Activation of NF-B signaling can stimulate COX-2 transcription. Therefore, to study whether NF-B binding mediates B[a]P-induced COX-2 gene transcription, electrophoretic mobility shift assays were performed. NF-B binding activity increased in SMC treated with B[a]P (Fig. 7, A and B). The specificity of the binding was confirmed by using excess unlabeled NF-B oligonucleotides (Fig. 7A) and an oligonucleotide containing a mutant NF-B (Fig. 7B). Supershift analyses identified both NF-B p65 and p50 proteins in the binding complex (Fig. 7B). The functional importance of NF-B in mediating the induction of COX-2 by B[a]P was then tested by transient transfections. B[a]P treatment led to a 2-fold increase in COX-2 promoter activity. Stimulation by B[a]P was blocked by mutating the NF-B (Ϫ223/Ϫ214) site in the 5Ј-region of the COX-2 promoter (Fig. 7C). Furthermore, inhibitors of NF-B activation, Bay 11-7085 and pyrrolidinedithiocarbamate (PDTC), blocked B[a]P-mediated induction of COX-2 promoter activity (Fig. 8A) and COX-2 protein (Fig. 8B). DISCUSSION We have previously shown that mechanical injury induced COX-2 in vascular tissue and that the increased expression of COX-2 persisted within vascular SMC during development of proliferative lesions in the injured vessel (23). Since B[a]P may cause vascular injury (44), and COX-2 participates in conversion of B[a]P to more cytotoxic derivatives, we have investigated whether B[a]P stimulates the expression of COX-2 in vascular SMC.
Our data clearly show that B[a]P induces COX-2 mRNA, protein, and prostaglandin synthesis in SMC derived from aortas of normal rats or from atherosclerotic human arteries. These effects are time-and dose-dependent. Nuclear run-off assays and transient transfections demonstrated that B[a]P treatment of SMC increases the rate of COX-2 transcription. The increased prostaglandin synthesis in SMC following treatment with B[a]P reflects an increase in functional COX-2 protein, since NS398, a specific inhibitor of COX-2 enzyme activity, effectively blocked prostaglandin synthesis in the B[a]Ptreated cells.
The extent of COX-2 induction by B[a]P in SMC also depends upon the antioxidant potential of the SMC, being strongly enhanced when glutathione was depleted by pretreatment with BSO, an inhibitor of glutathione synthesis, and strongly diminished by treatment with NAC, a precursor of glutathione. These data suggest that the oxidant effects of B[a]P and its metabolites mediate the induction of COX-2 in vascular SMC, a finding consistent with previous reports that oxidative stress can induce COX-2 (45,46). Oxidative stress can activate MAPK signaling, which, in turn has been implicated in the response to injury as well as the regulation of cell growth and differentiation (45,(47)(48)(49).
COX-2 expression has been linked with activation of MAPK pathways (50 -53). In the present study, we demonstrated that B[a]P induced COX-2 by activating ERK1/2 MAPK. In contrast, p38 MAPK did not appear to contribute to B[a]P-mediated COX-2 induction. COX-2 induction via preferential activation of ERK1/2 has been associated with resistance to oxidative stress in cardiomyocytes (45), and increased ERK1/2 activation protects against apoptosis in neurons (54). ERK1/2 MAPK activation can regulate the expression of numerous genes by activating NF-B (55). Our data suggest that NF-B is impor- This study demonstrates that the tobacco protoxin, B[a]P, stimulates expression of COX-2 and enhances prostaglandin synthesis in vascular SMC from two different species. Increased COX-2 levels, characteristic of tissues undergoing inflammatory reactions, have been observed in arteries after experimental injury and in clinical settings of atherosclerosis and heart transplantation (10,23,33,34). COX catalyzes the conversion of B[a]P 7,8-diol to highly reactive B[a]P diolepoxides (BPDE) that bind DNA, and BPDE selectively bind to CG-rich sequences in specific genes that may be involved in cell proliferation (6,17,56,57). For example, in lung epithelial or HeLa cells, BPDE has been shown to bind to mutagenic "hotspots" in the p53 gene (6). Whether BPDE-DNA adducts detected in atherosclerotic lesions (10) produce mutations that contribute to plaque development or represent tissue markers for oxidative vascular injury, or both, has not yet been determined. Mutations in atherosclerotic lesions resulting from frameshifts in repetitive DNA sequences within the TGF-␤ II receptor gene have been reported that render the atherosclerotic cells resistant to the growth-limiting effects of TGF-␤ (58,59). Since COX contributes to the formation of toxic derivatives of B[a]P, B[a]P-mediated induction of COX-2 in vascular tissues suggests a feedback mechanism capable of amplifying the toxic effects of this xenobiotic, resulting in vascular injury and potentiating atherogenesis.
The capacity for COX-2 to promote production of tissue damaging derivatives of B[a]P would be expected to increase in settings where COX-2 levels are high, particularly if antioxidant defenses are low. In addition, B[a]P itself is known to depress intracellular antioxidant mechanisms, for example, to decrease glutathione levels, another mechanism by which B[a]P could depress tissue capacity to detoxify xenobiotics and to resist oxidant stress (60,61). In epithelial cells, high levels of COX-2 are also associated with inhibition of apoptosis, and inhibition of COX-2 by aspirin or more selective COX-2 inhibitors appears to decrease the risk of carcinogenesis (26,27,29). In a developing atherosclerotic plaque, it is possible that a similar anti-apoptotic effect of high levels of COX-2 could augment plaque growth by decreasing cell death rates and depressing normal vascular remodeling. Decreased apoptosis in the presence of B[a]P in the setting of elevated COX-2 levels might also favor accumulation of BPDE-DNA adducts and potential mutations. Because interactions between B[a]P and COX-2 within vascular smooth muscle cells present several mechanisms for worsening progression of atherosclerosis, compounds that inhibit COX-2-gene activation and COX-2 activity may have a particularly useful role in prevention and treatment of atherosclerosis.