Inducible Nitric Oxide Synthase Mediates Prostaglandin H₂ Synthase Nitration and Suppresses Eicosanoid Production

Abstract

Nitric oxide (NO) modulates the biological levels of arachidonate-derived cell signaling molecules by either enhancing or suppressing the activity of prostaglandin H₂ isoforms (PGHS-1 and PGHS-2). Whether NO activates or suppresses PGHS activity is determined by alternative protein modifications mediated by NO and NO-derived species. Here, we show that inducible NO synthase (iNOS) and PGHS-1 co-localize in atherosclerotic lesions of ApoE^−/− mouse aortae. Immunoblotting and immunohistochemistry revealed Tyr nitration in PGHS-1 in aortic lesions but markedly less in adjacent nonlesion tissue. PGHS-2 was also found in lesions, but 3-nitrotyrosine incorporation was not detected. 3-Nitrotyrosine formation in proteins is considered a hallmark reaction of peroxynitrite, which can form via NO-superoxide reactions in an inflammatory setting. That iNOS-derived NO is essential for 3-nitrotyrosine modification of PGHS-1 was confirmed by the absence of 3-nitrotyrosine in lesions from ApoE^−/−iNOS^−/− mice. Mass spectrometric studies specifically identified the active site residue Tyr385 as a 3-nitrotyrosine modification site in purified PGHS-1 exposed to peroxynitrite. PGHS-mediated eicosanoid (PGE₂) synthesis was more than fivefold accelerated in cultured iNOS^−/− versus iNOS-expressing mouse aortic smooth muscle cells, suggesting that iNOS-derived NO markedly suppresses PGHS activity in vascular cells. These results further suggest a regulatory role of iNOS in eicosanoid biosynthesis in human atherosclerotic lesions.

The enzyme prostaglandin H₂ synthase (PGHS, also known as cyclooxygenase) regulates the production of eicosanoids that modulate physiological processes in the vessel wall, contributing to atherosclerosis and thrombosis. These processes include platelet aggregation, control of vascular tone, and the local inflammatory response. PGHS-1, the constitutive form of PGHS, is the predominant isozyme in platelets and mediates production of platelet thromboxane A₂ (TxA₂), a potent platelet agonist and vasoconstrictor. The beneficial effect of aspirin in reducing cardiovascular events is often attributed to inhibition of PGHS-1-mediated platelet TxA₂.^1,2 Prostacyclin (PGI₂), the major product of vascular endothelium,³ has opposite effects on platelet function and vascular tone and may inhibit thrombus formation by preventing platelet adhesion.⁴ PGHS-2, the inducible form of PGHS, is a proinflammatory mediator that is associated with tumorigenesis and atherosclerosis.^5,6 In some settings PGHS-2 may be the major source of PGI₂ under physiological conditions⁷ due to its induction in endothelial cells by hemodynamic shear,⁸ but PGHS-1 is also suggested to play an important role.⁹ The opposing properties of PGI₂ and TxA₂ suggest that a balance between these compounds in the regulation of interactions between platelets and the vessel wall may be disrupted during inflammation and atherosclerosis.^10,11

Eicosanoid production may be modulated by the interaction of the nitric oxide (NO) and arachidonic acid biosynthetic pathways. This interaction can lead to either stimulation or inhibition of PGHS activity, as well as altered transcription of the PGHS-2gene.^12–27 Because several different forms of nitrogen oxides (NO_x) may arise in biological systems, it is conceivable that one form of NO_x may activate PGHS whereas another form may lead to inhibition. Indeed, in our previous studies we demonstrated that peroxynitrite (ONOO⁻) leads to PGHS-1 activation whereas NO leads to inhibition.¹⁷ We also found that ONOO⁻ can affect PGHS activity by initiating a signaling cascade that leads to arachidonic acid release and subsequent eicosanoid production.²⁸ Another interaction of ONOO⁻ and PGHS leads to PGHS nitration, which is correlated with a loss in activity.²⁹ In this regard, ONOO⁻ was found to nitrate-purified PGHS-1 as well as PGHS-1 in smooth muscle cells. Importantly, nitrated PGHS-1 has been detected in human atherosclerotic lesions.²⁹

Further evidence that the NO and eicosanoid biosynthetic pathways are linked is supported by recently published in vivo studies. Mice missing the gene that encodes the inducible form of NO synthase (iNOS) demonstrated a reduced ability to synthesize eicosanoids globally while synthesis of thromboxane was enhanced.³⁰ A reduction in eicosanoid biosynthesis in inflammatory lesions was also observed after administration of NOS inhibitors to rats.^31,32 Furthermore, inflamed brain tissue from iNOS knockout mice also contains decreased levels of PGE₂ when compared to wild-type mice.³³ Taken together, these results indicate that iNOS-derived NO, or a NO-derived species, causes decreased biosynthesis of eicosanoids. In contrast, the loss or inhibition of NOS up-regulates eicosanoid production.^34–38 For example, coronary arteries from wild-type mice vasodilate in response to acetylcholine by a mechanism involving NO from endothelial NOS (eNOS), but eNOS-deficient mice respond principally by a mechanism involving PGHS.³⁹ Such observations suggest that the PGHS pathway may compensate to maintain near-normal coronary arterial function in response to chronic loss of eNOS in blood vessels. Either way, given the strong link between NO and eicosanoid pathways, it follows that eicosanoid production in the vessel wall may be affected during disease states in which NO production is perturbed, such as atherosclerosis.^40,41

The apolipoprotein E-deficient (ApoE^−/−) mouse has been used successfully as an animal model of atherogenesis.⁴² The ApoE amphipathic protein stabilizes and solubilizes lipoprotein particles and thus plays a pivotal role in lipoprotein trafficking. ApoE is a constituent of chylomicrons, very low-density lipoprotein, intermediate-density lipoprotein, and high-density lipoprotein and acts as a ligand for the receptor-mediated clearance of these particles.⁴³ ApoE^−/− mice have plasma cholesterol levels that are four to five times greater than normal and develop atherosclerotic lesions spontaneously, even when fed a low-cholesterol diet.⁴² The development of ApoE^−/− mice that are also deficient in iNOS (ApoE^−/−iNOS^−/− mice) has been described.^44–46 ApoE^−/−iNOS^−/− mice are normotensive and, when fed a normal chow diet, develop atherosclerotic lesions in a manner that is indistinguishable from ApoE^−/− mice.⁴⁴ For this reason, it was initially thought that iNOS-derived NO does not influence lesion progression. However, when lesion progression is accelerated by feeding ApoE^−/−iNOS^−/− mice a Western diet, lesions are reduced in size compared to ApoE^−/− mice.^45,46 These results indicate that iNOS-derived NO is proatherogenic.

The current study was designed to determine whether PGHS nitration is associated with atherosclerosis in the ApoE^−/− mouse model and, if so, whether nitration is dependent on the presence of iNOS. Because iNOS^−/− mice in a nonatherosclerotic genetic background display a reduced ability to synthesize eicosanoids,³⁰ we sought to determine whether the absence of iNOS would promote eicosanoid production in a mouse model of atherosclerosis. For this purpose, we cultured smooth muscle cells from ApoE^−/− and ApoE^−/−iNOS^−/− mice and evaluated eicosanoid production under conditions of iNOS activity and inactivity.

Materials and Methods

Animals

The animal protocol used in these studies was reviewed and approved by the Weill Medical College of Cornell University Care and Use Committee. ApoE^−/−iNOS^−/− mice were generated as described previously.⁴⁴ Mice used in these studies were derived from at least six generations of backcross breeding to C57BL/6J mice.⁴⁴ ApoE^−/−iNOS^−/+ heterozygote mice were bred to generate pups that were then genotyped for the iNOS allele by Southern blot analysis. Primers that are complimentary to portions of exon 11, exon 12, and neomycin (neo) were used: AGAGTCCTTCATGAAGCACATGCA (exon 11), TCAGCTTCTCATTCTGCCAGATGT (exon 12), and CAATCCATCTTGTTCAATGGCCGA (neo). The primers were mixed one part neo, two parts exon 11, one part exon 12. Wild-type mice contain both exons 11 and 12 and give rise to a larger band (∼500 bp) whereas iNOS^−/− mice contain exon 11 and neomycin and give rise to a smaller band (∼340 bp). iNOS^+/− mice give rise to both bands.

Study Design

ApoE^−/−iNOS^−/− males and females were designated as the experimental group and ApoE^−/− littermates (containing iNOS^+/+) were designated the control group. ApoE^−/−iNOS^−/+ pups were used as breeding mice. In total, we have used ∼100 ApoE^−/− mice and ∼100 ApoE^−/−iNOS^−/−mice in these studies. At 21 days, the mice were weaned. The mice were fed a Western diet ad libitum comprising 21.2% fat (g/100 g), 0.2% cholesterol, and 0% cholate (Harlan Teklad, Indianapolis, IN). At 6 months, the mice in the control and experimental groups were sacrificed. The mice were weighed before euthanasia and average weights of the ApoE^−/− and ApoE^−/−iNOS^−/− mice (inclusive of both males and females) were 30.0 ± 0.3 g and 29.2 ± 0.7 g, respectively, and were not significantly different. Aortae were removed surgically from the cardiac origination to the iliac bifurcation and frozen at −80°C. Using a dissecting microscope (SMZ-1B; Nikon, Melville, NY), the aortae were cleared of fat, connective tissues, and adventitia, and dissected into portions representing lesions and surrounding tissue, which were then preserved at −80°C. Lesions and surrounding aortic tissue from these animals were used for Western blot and immunoprecipitation experiments, as described below.

Isolation of Smooth Muscle Cells from ApoE^−/− and ApoE^−/−iNOS^−/− Mice

The isolation of vascular smooth muscle cells from murine aortae was performed according to a published method.⁴⁷ Briefly, mice were sacrificed by CO₂ asphyxiation and secured on a dissecting board. The thorax and abdomen were rinsed (70% ethanol), the skin was removed, and the thorax was opened to expose the heart and lungs. The aorta was dissected from its origin at the left ventricle to the iliac bifurcation. Using a syringe with a 26-gauge needle to puncture the left ventricle, the aorta was flushed with sterile phosphate-buffered saline (3 ml), surgically removed, and placed in a Petri dish covered with Dulbecco’s modified Eagle’s medium (DMEM) (50 to 100 μl, containing 0.25 μg/ml filter-sterilized fungizone). Using a dissecting microscope, the aorta was cleared of fat, connective tissues, and adventitia. The aorta was transferred to a new Petri dish, covered with DMEM (50 to 100 μl, containing 10% fetal bovine serum (FBS), 1% glutamine, 1% antibiotic-antimycotic; Invitrogen, Carlsbad, CA), and cut into 1- to 2-mm-square pieces. The pieces of aorta were transferred to a small tissue culture tube containing type II collagenase solution (136 μg in culture medium). The tube was placed with the cap loosely attached in a standard tissue culture incubator (37°C, 5% CO₂) for 4 to 6 hours. The tube was removed from the incubator, gently agitated to resuspend the cells, and DMEM was added (3 ml containing 10% FBS, 1% glutamine, 1% antibiotic-antimycotic). The suspension was transferred to a conical polypropylene tube (15 ml) and centrifuged at 300 × g for 5 minutes at room temperature. The medium was aspirated and cells were resuspended in fresh medium (5 ml). The suspension was recentrifuged, the cells were resuspended in medium (0.7 to 1 ml) and transferred to a Primaria dish (BD Biosciences, San Diego, CA). The cells were placed in an incubator and left undisturbed for 5 days.

Cells were confirmed as smooth muscle cells based on fluorescence staining with α-smooth muscle cell actin antibody and for their reactivity and lack of reactivity toward calponin and CD31 (platelet endothelial cell adhesion molecule-1, PECAM-1) antibodies, respectively (see below).⁴⁷ Cells were grown to confluence in DMEM supplemented with 10% (v/v) FBS and 1% (v/v) glutamine in either six-well plates or 100-mm dishes (containing ∼8.8 × 10⁶ cells/dish). All cells were incubated at 37°C in 5% CO₂ in air. Cells were tested for the presence of mycoplasma (Mycoplasma PCR ELISA kit; Roche Applied Science, Indianapolis, IN) with negative results. For experiments with suppressed PGHS-2 expression, cells were rendered quiescent using DMEM supplemented with 2% FBS and 1% glutamine for 2 days.⁴⁸ Cells between passages one and five were used in experiments.

Detection of Smooth Muscle Cell α-Actin and Calponin in Smooth Muscle Cells Cultured from ApoE^−/− and ApoE^−/−iNOS^−/− Mice by Fluorescence Staining

Cells were plated onto chamber slides (Lab-Tek; Nunc, Naperville, IL). When the cells were 50% confluent, they were washed [phosphate-buffered saline (PBS), three times] and fixed in methanol at −20°C (5 minutes). The cells were washed in PBS (three times), blocked in 5% rabbit serum (R-7136; 500 μl/well for 30 minutes; Sigma Chemical Co., St. Louis, MO) and incubated with primary antibody (monoclonal anti-α-smooth muscle actin, clone 1A4, A-2547, 1:200 dilution; Sigma) in 5% rabbit serum overnight at 4°C. A control for nonspecific interactions was performed by incubating cells with an equivalent concentration of mouse IgG (Sigma). The cells were washed in PBS (three times, 5 minutes each) and incubated with secondary antibody (Alexa Fluor 488 rabbit anti-mouse IgG, A-11078, 1:200 dilution; Molecular Probes, Eugene, OR) in 5% rabbit serum (1 hour at room temperature in the dark). Before the end of the incubation, a blue fluorescent 4′,6-diamidino-2-phenylindole, dihydrochloride, nucleic acid stain (2 μg/ml) was also added (5 minutes). The cells were washed (PBS, three times) and coverslips applied with Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA). The cells were examined using an upright fluorescence microscope with appropriate filters (BX51; Olympus, Melville, NY). All cells defined by 4′,6-diamidino-2-phenylindole nuclear staining also contained the selective α-smooth muscle actin and were additionally recognized by an antibody directed against the smooth muscle cell-specific protein calponin (mouse monoclonal calponin antibody, C-2687; Sigma), but not by a rat monoclonal antibody against the endothelial marker CD31 (PECAM-1, SC-18916, Santa Cruz) (data not shown). Notably, a previous study using an identical smooth muscle cell isolation method, conservatively estimated the purity of a resulting smooth muscle cell population at 99%.⁴⁷ Cells placed in culture are thought to shift from a contractile to a synthetic phenotype that is also adopted by smooth muscle cells found in the intima during atherosclerosis.^49,50 The cells used in our studies possessed similar levels of α-smooth muscle actin and calponin (which are both markers of contractile function⁵¹), indicating that they are in a similar state of differentiation. Cells were maintained at low passages (up to passage 5), although we did not see a significant variation in response to higher passage number (up to passage 12).

Lipopolysaccharide (LPS)/Interferon (IFN)-γ Treatment of Smooth Muscle Cells, PGE₂, and Nitrate/Nitrite Measurements

Smooth muscle cells were preincubated in the presence and absence of LPS (from Escherichia coli 026:B6, 10 μg/ml; Sigma) and recombinant mouse IFN-γ (100 U/ml; Calbiochem, La Jolla, CA) for 24 hours to induce PGHS-2 and iNOS gene transcription.⁵² Culture medium was replaced with fresh DMEM (with 1% glutamine and 10% or 2% FBS) and the cells incubated for 30 minutes to 1 hour at 37°C. The media was removed, stored at −80°C, and fresh DMEM (with 1% glutamine and 10% or 2% FBS) containing arachidonic acid (20 μmol/L) was added to the cells for 30 minutes to 1 hour at 37°C. The supernatant medium was removed and assayed for PGE₂ formation using an enzyme immunoassay kit (GE Health Care, Piscataway, NJ) and for nitrate/nitrite production using a colorimetric assay kit (Cayman Chemical, Ann Arbor, MI). Total protein in the cell lysate was determined by using either the Lowry or modified Lowry method.⁵³

Western Blots of Tissue Homogenates and Smooth Muscle Cells

Lesions and surrounding aortic tissue from ApoE^−/− and ApoE^−/−iNOS^−/− mice were homogenized with a Tissue Tearor (Dremel, Racine, WI) in a minimum volume (300 to 500 μl) of lysis buffer (50 mmol/L Tris-HCl, pH 8, 10 mmol/L EDTA, 1% Tween 20, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mmol/L phenylmethyl sulfonyl fluoride) and sonicated. The homogenate was centrifuged (13,000 rpm for 10 minutes at 4°C), and the supernatant was used for protein assay and Western blotting. Smooth muscle cells were washed once with PBS and lysed with a lysis buffer (20 mmol/L Tris-HCl, pH 7.5, 1 mmol/L EDTA, 1 mmol/L EGTA, 150 mmol/L NaCl, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerolphosphate, 1 mmol/L sodium orthovanadate, 1 mmol/L phenylmethyl sufonyl fluoride, 1% Triton X-100, and 1 μg/ml leupeptin). The cells were scraped, transferred to Eppendorf tubes, vortexed, and placed on ice (30 minutes). Cell lysates were then sonicated and centrifuged (10,000 × g for 10 minutes) for clarification. Cell lysate or tissue protein (15 to 35 μg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10% acrylamide gel), transferred to a nitrocellulose membrane, blocked (90 minutes at 25°C or overnight at 4°C) with 5% nonfat milk in PBS-1% Tween buffer. The membrane was then probed with primary antibody (for PGHS-1, PGHS-2, or iNOS) in 1% nonfat milk in PBS-1% Tween buffer according to the manufacturer’s recommendations. When using PGHS-1 antibody (mouse monoclonal, 160110; Cayman Chemical), the membrane was probed for 1 hour, followed by a horseradish peroxidase-conjugated purified goat anti-mouse antibody for 1 hour. When using PGHS-2 antibody (goat polyclonal, SC-1745; Santa Cruz Biotechnology, Santa Cruz, CA), the membrane was probed either 1 hour or overnight, followed by a rabbit anti-goat secondary antibody for 1 hour. When using iNOS antibody (rabbit polyclonal, SC-650; Santa Cruz), the membrane was probed overnight, followed by a goat anti-rabbit secondary antibody for 1 hour. Actin (goat polyclonal, SC-1615, Santa Cruz) or α-smooth muscle actin antibody was used as a measurement of loading accuracy (90 minutes followed by rabbit anti-goat secondary antibody). The immunoblot signal was visualized through chemiluminescence (ECL plus; GE Health Care).

Immunoprecipitation of Nitrated PGHS-1 from Aortic Lesions and Surrounding Tissue in ApoE^−/− and ApoE^−/−iNOS^−/− Mice

Lesions and surrounding aortic tissue from ApoE^−/−and ApoE^−/−iNOS^−/− mice were homogenized separately and centrifuged as described above for Western blotting. The resulting supernatant (1 mg/ml in PBS) was preclarified with Protein G-agarose beads (25 μl of a 50% slurry in PBS for 30 minutes) to reduce nonspecific binding. The lysate/bead mixture was centrifuged (5000 × g for 5 minutes, followed by 16,000 × g for 1 minute) to remove the beads, and the clarified lysate was then incubated (overnight at 4°C) with polyclonal nitrotyrosine antibody (Upstate Biotechnology, Lake Placid, NY) followed by Protein G-agarose beads (35 μl of a 50% slurry overnight at 4°C). The bound agarose beads were collected by centrifugation (5000 × g for 10 minutes, followed by 16,000 × g for 1 minute), washed three times with PBS, and resuspended in Laemmli sample buffer (40 μl). The bead-immunocomplex was boiled (5 minutes), and the sample (40 μl) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10%) and immunoblotted with monoclonal PGHS-1 antibody as described above.

Immunohistochemical Localization of PGHS-1 and iNOS

Immunohistochemistry was performed on frozen aortic root sections removed from ApoE^−/− and ApoE^−/−iNOS^−/−mice that were maintained on a Western diet for 4 months. Murine hearts were removed and cryo-preserved in a 30% sucrose:OCT solution (1:1; Sakura Finetek USA, Inc., Torrance, CA). Serial frozen sections were obtained by cryostat sectioning in the region of the aortic leaflet. The frozen sections were air-dried onto microscopic slides and were treated with H₂O₂ (0.1%) in methanol at −20°C for 30 minutes to quench endogenous peroxidase activity. Adjacent sections were incubated overnight at 4°C with either PGHS-2 (goat polyclonal, SC-1745; Santa Cruz) or iNOS (rabbit polyclonal, SC-650; Santa Cruz) antibodies. Sections were exposed at room temperature for 1 hour to biotinylated horse anti-goat (BA9500; Vector Laboratories) or biotinylated goat anti-rabbit (BA1000, Vector Laboratories) secondary antibodies, respectively. Immunoreactive proteins were detected using an avidin-biotin-based peroxidase system, using VIP chromogenic substrate (Vector Laboratories). Counterstaining was achieved using hematoxylin. Control sections were treated with PGHS-2 or iNOS antibodies that were previously incubated with blocking peptides (SC-1745P, PGHS-2 blocking peptide; SC-650P iNOS blocking peptide; Santa Cruz) and then developed as described above. Sections were visualized by light microscopy, and digital images were captured using an Olympus microscope.

Double Immunofluorescence of Co-Localized PGHS-2 and iNOS

Sections, prepared as above, were incubated overnight at 4°C with both PGHS-1 (goat polyclonal, SC-1754; Santa Cruz) and nitrotyrosine (rabbit polyclonal, 06284; Upstate Biotechnology) antibodies. The sections were exposed at room temperature for 1 hour to a biotinylated horse anti-goat secondary antibody (BA9500, Vector Laboratories), followed by exposure to rhodamine/avidin (A-2012, Vector Laboratories) for visualization of PGHS-1 and to a fluorescein-labeled goat anti-rabbit secondary antibody (FI-100, Vector Laboratories) for visualization of nitrotyrosine. Control sections were treated with PGHS-1 and nitrotyrosine antibodies that were previously incubated with a PGHS-1 blocking peptide (SC-1754P, Santa Cruz) and 3-nitrotyrosine (10 mmol/L). The fluorescence from immunoreactive proteins was observed by fluorescence microscopy and digital images were captured using an Olympus microscope.

LC-MS/MS Analysis and Database Search

Nanoflow liquid chromatography-tandem mass spectrometry (nLC-MS/MS) was used to identify sites of Tyr nitration in ovine PGHS-1. Ovine PGHS-1 was purified from sheep seminal vesicles as described previously.⁵⁴ MS/MS analyses were performed using an 1100 series LC/MSD XCT plus ion trap mass spectrometer (Agilent, Palo Alto, CA). The mobile phase contained 0.1% formic acid in 3% acetonitrile (solvent A) and 0.1% formic acid in 100% acetonitrile (solvent B). PGHS-1 (8 μmol/L in 13 μl, containing heme) was incubated with ONOO⁻ (1000 μmol/L) for 1 hour at ambient temperature. ONOO⁻-treated and nontreated control proteins were precipitated in 2 vol of acetone, centrifuged, and then resolubilized in ammonium bicarbonate (20 mmol/L, pH 8). After reduction by dithiothreitol (10 mmol/L, 30 minutes at 50°C) and alkylation by iodoacetamide (50 mmol/L, 30 minutes at ambient temperature), the samples were digested with trypsin (sequencing grade, 1:100, trypsin:PGHS-1; Promega, Madison, WI) for 8 hours at 37°C. Samples were diluted 100-fold with solvent A, and 8 μl of the digested peptide mixture was injected onto a 0.3 × 5 mm Zorbax 300SB-C18 enrichment column at a flow rate of 10 μl/minute. Peptides were subsequently resolved on a 0.075 × 150 mm Zorbax 300SB-C18 analytical column (3.5 μm particle size) at a flow rate of 0.3 μl/min with a gradient of 10 to 40% solvent B for 40 minutes and 40 to 80% solvent B for 30 minutes. Mass spectra were acquired in the automated MS/MS mode, in which MS/MS scans were performed on the three most intense ions after each MS scan. The MS/MS spectra were used to identify the sites of Tyr nitration in PGHS-1 by a database search using SpectrumMill software (Millenium Pharmaceuticals, Cambridge, MA). The parameters for searching were minimum matched peak intensity of 50%, precursor mass tolerance of 2.5 Da, and product mass tolerance of 0.7 Da. The program was instructed to account for Tyr nitration when matching peptide fragments to PGHS-1.

Isolation of Total RNA and Northern Blotting

Aortic smooth muscle cells cultured from ApoE^−/− and ApoE^−/−iNOS^−/− mice were lysed in RNA-BEE (Tel-Test, Friendswood, TX). RNA was extracted with chloroform, and then precipitated with isopropanol. Total RNA from each sample was loaded on 1% formaldehyde-agarose gels (20 μg/lane). After electrophoresis, RNA was transferred to a Zeta-probe blotting membrane (Bio-Rad, Hercules, CA). The blot was UV cross-linked and prehybridized with Hybrisol I (Chemicon, Temecula, CA). The blot was then probed using ³²P-labeled PGHS-1, followed by rehybridization with ³²P-labeled GAPDH (glyceraldehyde-3-phosphate dehydrogenase) to verify that loading in the wells was equal. The PGHS-1 and GAPDH probes were kindly provided by Dr. David L. DeWitt (Michigan State University, East Lansing, MI). The PGHS-2 probe was kindly provided by Dr. Raymond Dubois (Vanderbilt University Medical Center, Nashville, TN).

Statistical Analyses

Data are presented as means ± SE with significant differences determined by t-test, with P < 0.05 defined as statistically significant.

Results

PGHS-2 is Present in Aortic Lesions in ApoE^−/− and ApoE^−/−iNOS^−/− Mice but Absent from the Surrounding Tissue

Figure 1 shows the levels of PGHS-2 and iNOS protein found by Western blot analysis in the plaque (P) and surrounding media (M) in aortae removed surgically from ApoE^−/− and ApoE^−/−iNOS^−/− mice after feeding a Western diet for 6 months. Because the tissue samples were small, it was necessary to pool samples from several mice before analysis. Although PGHS-2 and iNOS proteins were both present in the aortic plaque of ApoE^−/− mice, they were missing or barely detectable in the media surrounding these lesions. In addition, PGHS-2 protein was present in the aortic plaque of ApoE^−/−iNOS^−/− mice, but absent from the media. Western blot results confirmed that iNOS protein was missing in both the plaque and the surrounding media of theApoE^−/−iNOS^−/− mice.

Figure 1.

PGHS-2 is present in aortic plaque in ApoE^−/− and ApoE^−/−iNOS^−/−mice but is absent in the surrounding medial tissue. Western blot demonstrating PGHS-2 and iNOS protein levels in atherosclerotic plaque (P) and surrounding media (M) removed fromApoE^−/−(1) and ApoE^−/−iNOS^−/− (2) mice after 6 months on a Western diet. Actin protein levels are also included to demonstrate equal loading of wells. The results shown were obtained by pooling tissue from 11ApoE^−/−and 19 ApoE^−/−iNOS^−/−mice. The results from this experiment have been reproduced three times from three different pools of tissue.

To investigate the potential co-localization of PGHS-2 and iNOS in aortic tissue, immunohistochemical analysis was performed on aortic root sections from ApoE^−/− and ApoE^−/−iNOS^−/− mice. Immunohistochemical analysis of aortic sinus lesions demonstrated increased PGHS-2 immunoreactivity in the plaque (P) (Figure 2) with little or no expression in the media (M) in both ApoE^−/− and ApoE^−/−iNOS^−/− mice. iNOS was detected in lesions from ApoE^−/− mice, but was missing, as expected, in aortic sections from ApoE^−/−iNOS^−/− mice. Some iNOS staining was also detected in the medial layer of sections from ApoE^−/− mice, which is consistent with the observation that by Western blotting some iNOS was present in the tissue that surrounds the lesions in ApoE^−/− mice (Figure 1). Increased levels of iNOS have previously been localized to vascular smooth muscle cells and mononuclear leukocytes in lesions.^55,56 The expression pattern for both iNOS and PGHS-2 in ApoE^−/− and ApoE^−/−iNOS^−/− lesions was similar to that described by other investigators.^45,46,57

Figure 2.

A comparison of PGHS-2 and iNOS localization in aortic sinus lesions from ApoE^−/−and ApoE^−/−iNOS^−/−mice. Immunohistochemical detection of iNOS and PGHS-2 in aortic sinus lesions from ApoE^−/−(top) and ApoE^−/−iNOS^−/− (bottom) mice fed a Western diet for 4 months. L, lumen; P, plaque; M, media; con, control sections stained with antibody preincubated with antigen. Sections were counterstained with hematoxylin to visualize nuclei. Similar results were observed in aortic sinus lesions from three ApoE^−/− and three ApoE^−/−iNOS^−/− mice.

Lesions from ApoE^−/− Mice Contain Nitrated PGHS-1 that Is Absent in ApoE^−/−iNOS^−/− Mice

Immunoprecipitation experiments were performed to determine whether nitrated PGHS accumulates in aortic tissue obtained from ApoE^−/− and ApoE^−/−iNOS^−/− mice. Animals were maintained on a Western diet for 6 months, and aortic tissue was dissected under a microscope into portions comprising the plaque (P) and surrounding media (M). Because the tissue samples were small, it was necessary to pool samples from several mice for ample immunoprecipitation of PGHS-1. Anti-oxidants were added to limit potential ex vivo oxidation. Lesions were readily visualized and were commonly ∼1 mm in length but sometimes reached 3 to 4 mm. For every immunoprecipitation experiment, tissues were pooled from ∼10 ApoE^−/− and ∼10 ApoE^−/−iNOS^−/− mice, providing ∼300 μg of total protein. Nitrated proteins from the atherosclerotic plaque and surrounding media were immunoprecipitated using a polyclonal 3-nitrotyrosine antibody. Immunoprecipitated proteins were resolved by gel electrophoresis and immunoblotted with a PGHS-1 monoclonal antibody (Figure 3). Although nitrated PGHS-1 was observed in the plaque from ApoE^−/− mice, little or no PGHS-1 nitration was detected in plaque obtained from the ApoE^−/−iNOS^−/− animals. In addition, the surrounding media showed little or no evidence of PGHS-1 nitration in either the ApoE^−/− or ApoE^−/−iNOS^−/−mice. These data indicate that lesional PGHS-1 undergoes significant Tyr nitration in vivo and iNOS is the principal source of NO required for this modification. Surprisingly, on stripping of blots and reprobing for PGHS-2, we did not observe PGHS-2 nitration (data not shown).

Figure 3.

Lesions from ApoE^−/− mice contain nitrated PGHS-1 that is reduced in ApoE^−/−iNOS^−/− mice. Aortic tissue from ApoE^−/− and ApoE^−/−iNOS^−/− mice fed a Western diet for 6 months was dissected into portions representing the plaques (P) and noninvolved surrounding medial tissue (M). The tissue was homogenized as described in the Materials and Methods. Proteins containing 3-nitrotyrosine were immunoprecipitated using a polyclonal 3-nitrotyrosine antibody and then immunoblotted with a PGHS-1 monoclonal antibody. This experiment has been reproduced three times using three different pools of tissue. Each tissue pool is from ∼10 ApoE^−/− and ∼10 ApoE^−/−iNOS^−/− mice.

To confirm that PGHS-1 is nitrated in situ, double-label immunofluorescence imaging was performed on sections of murine hearts from ApoE^−/− and ApoE^−/−iNOS^−/− mice fed a Western diet for 4 months. Double immunofluorescence detected PGHS-1 (Figure 4a) and nitrotyrosine (Figure 4b) in lesions of the ApoE^−/− mouse. The overlay demonstrated co-localization of PGHS-1 and nitrotyrosine in lesions of the ApoE^−/− mouse (yellow; indicated by arrows in Figure 4c). In contrast, in lesions of ApoE^−/−iNOS^−/− mice, PGHS-1 was present in the plaque (Figure 4d), but nitrotyrosine levels were low (Figure 4e). Moreover, co-localization of PGHS-1 and nitrotyrosine in ApoE^−/−iNOS^−/− mice was reduced (Figure 4f). These data indicate that nitrated PGHS-1 predominates in the plaque of lesions obtained from ApoE^−/− mice relative to ApoE^−/−iNOS^−/− mice, although the presence of some nitrated PGHS-1 in aortic lesions of ApoE^−/−iNOS^−/− mice cannot be discounted.

Figure 4.

PGHS-1 and nitrotyrosine co-localize in aortic sinus lesions from ApoE^−/− mice. Double-immunofluorescence detection of PGHS-1 and nitrotyrosine in aortic sinus lesions removed from ApoE^−/− (a–c) and ApoE^−/−iNOS^−/− (d–f) mice maintained on a Western diet for 4 months. For visualization of PGHS-1, sections were incubated with a goat polyclonal PGHS-1 antibody followed by exposure to a biotinylated horse anti-goat secondary antibody and a rhodamine/avidin mixture. Sections were co-incubated with a fluorescein-labeled goat anti-rabbit secondary antibody for visualization of nitrotyrosine. The left sides of these panels depict autofluorescence from the internal elastic lamina in the media. Asterisks denote areas of nonspecific binding or autofluorescence. Insets (bottom parts of a, b, d, and e) represent control sections in which the PGHS-1 and nitrotyrosine antibodies were pretreated with a PGHS-1 blocking peptide and 3-nitrotyrosine (10 mmol/L), respectively. PGHS-1 and nitrotyrosine co-localization was assessed in aortic sinus sections from three ApoE^−/− and three ApoE^−/−iNOS^−/− mice.

Reaction of PGHS-1 with Peroxynitrite Results in 3-Nitrotyrosine Accumulation at Tyr385

We previously demonstrated that purified PGHS-1 can be nitrated by ONOO⁻.²⁹ Spectroscopic data indicated that ONOO⁻ nitrates ∼2 Tyr residues/PGHS-1 monomer.^29,58 To identify which Tyr residues were nitrated, ONOO⁻-treated purified PGHS-1 was subjected to trypsinolysis and analyzed by nLC-MS/MS. PGHS-1 was incubated with ONOO⁻ (1 mmol/L) for 1 hour and isolated by acetone precipitation. PGHS-1 was resuspended in ammonium bicarbonate, reduced, and alkylated (see Materials and Methods). Solubilized PGHS-1 was proteolyzed with 0.3% trypsin, followed by peptide analysis using nLC-MS/MS (Figure 5). Figure 5a shows the extracted ion chromatograms for the ONOO⁻-treated PGHS-1 and control PGHS-1 samples of a fragment corresponding to m/z (645.6) of the quadruply charged peptide ion IAMEFNQLYHWHPLMPDSFR (Ile377 to Arg396), which is nitrated on Tyr385. Notably, this nitrated peptide from ONOO⁻-treated PGHS-1 (∼38 minutes, indicated by an arrow) was undetectable in nontreated, control PGHS-1. Tandem mass spectrometric analysis of this peptide (Figure 5b) confirms that this peak indeed represents the tryptic peptide containing nitrated Tyr385. Several y- and b-series ions are observed, including a long, continuous series of y ions that stretches from y12 to y19 of this peptide fragment. It is notable that the fragmentation pattern of peptides is sequence-dependent, and many peptides do not display an ideal fragmentation pattern with a more complete b- and y-ion series. This is particularly relevant for a peptide such as IAMEFNQLYHWHPLMPDSFR, which contains two Pro and two His residues. It is well known that both of these residues induce atypical fragmentation.⁵⁹ With two His and Pro each in the sequence, it is, therefore, not surprising that the fragmentation pattern observed for this peptide is somewhat limited. Figure 5b shows that the MS/MS spectrum is dominated by a triply charged y18 ion at m/z = 798.9, in accord with the presence of two positively charged His residues in the sequence. The identity of this ion is further supported by another strong peak at m/z = 599.4 representing the quadruply charged ion of the same fragment. The strong dominancy of this particular fragment may account for the relatively less complete ion series detected. In a previously published tandem mass spectrometric study of a variant form of the human hemoglobin α chain, a very similar fragmentation pattern of a quadruply charged tryptic peptide ion bearing two His residues was observed: the MS/MS was dominated by a few triply charged y ions near the N terminus, with incomplete b- and y-ion series.⁶⁰ Site-directed mutagenesis studies have demonstrated that Tyr385 is essential to cyclooxygenase activity.^61,62 Technical limitations have precluded the rigorous MS/MS-based structural characterization of PGHS-1 from lesional tissues, owing mainly to the very limited amount of available tissue. These data establish that ONOO⁻ selectively nitrates Tyr385 in PGHS-1, but it is not yet known whether this specific Tyr residue is nitrated in PGHS-1 in atherosclerotic tissue, and if so, whether this is the preferred Tyr residue that undergoes nitration under biological conditions.

Figure 5.

Peroxynitrite nitrates purified PGHS-1 at Tyr385. Purified ovine PGHS-1 (8 μmol/L) was incubated with peroxynitrite (ONOO⁻; 1 mmol/L) for 1 hour and subjected to acetone precipitation, resuspension in ammonium bicarbonate, and reduction and alkylation (as described in Materials and Methods). A control PGHS-1 sample was prepared in a similar manner. The samples were digested with 0.3% trypsin overnight and analyzed by nLC-MS/MS. a: Extracted ion chromatograms for the peptide fragment (m/z = 645.6) representing the quadruply charged peptide IAMEFNQLYHWHPLMPDSFR that is nitrated at the Tyr residue. The top and bottom extracted ion chromatograms monitor this peptide fragment with time for the ONOO⁻-treated and control PGHS-1 samples, respectively. The arrow at ∼38 minutes in the top profile indicates the presence of a peak that is missing in the control PGHS-1 sample. b: MS/MS analysis of the peptide fragment at ∼38 minutes confirms the identity of this peptide sequence as IAMEFNQLYHWHPLMPDSFR containing nitrated Tyr385. The black diamond at m/z = 645.6 indicates the position of the precursor fragment from which the ions observed are derived.

Smooth Muscle Cells Cultured from ApoE^−/−iNOS^−/− Mice Express Significantly Higher Levels of PGHS-1 and PGHS-2 Protein and mRNA than Smooth Muscle Cells from ApoE^−/− Mice

Smooth muscle cells were cultured from ApoE^−/− and ApoE^−/−iNOS^−/− mice to examine the levels of PGHS-1 and PGHS-2 expressed. Cells were maintained in FBS, which has previously been shown to induce PGHS-2 in rat smooth muscle cells.^48,63 To determine the effect of serum on PGHS-2 expression, smooth muscle cells were grown in media containing 10% FBS, which was then replaced with 2% FBS. Figure 6 demonstrates that PGHS-2 was expressed when smooth muscle cells from ApoE^−/− mice were incubated in media containing 10% FBS (day 0). However, PGHS-2 levels progressively diminished with time when replaced by 2% FBS (days 1 to 4). Thus, smooth muscle cells cultured in FBS express PGHS-2.

Figure 6.

Serum induces PGHS-2 in smooth muscle cells. Aortic smooth muscle cells were cultured from ApoE^−/− mice as described in Materials and Methods. The cells were grown to ∼80% confluency in DMEM containing 10% FBS (lane 0) and in DMEM containing 2% FBS for 1, 2, 3, or 4 days (lanes 1–4, respectively). The cells were lysed, analyzed by Western blotting, and probed for PGHS-2, followed by actin. Std represents standard purified ovine PGHS-2.

Western blot analyses were performed on smooth muscle cells from ApoE^−/− and ApoE^−/−iNOS^−/− mice cultured under identical conditions to determine the relative expression of PGHS-1 and PGHS-2 protein levels. Figure 7 shows that significantly higher levels of both PGHS-1 (Figure 7a) and PGHS-2 (Figure 7b) protein were found in aortic smooth muscle cells obtained from adult ApoE^−/−iNOS^−/− mice compared to ApoE^−/− mice. Under these conditions, iNOS protein was missing in the ApoE^−/−iNOS^−/− smooth muscle cells, but traces were occasionally detected by Western blotting in the ApoE^−/− cells (data not shown). Under quiescent conditions (ie, in DMEM containing 2% FBS), PGHS-1 protein levels were also significantly higher in smooth muscle cells from ApoE^−/−iNOS^−/− mice compared to ApoE^−/−, although no differences were observed in the low levels of PGHS-2 (data not shown). Figure 7 also shows that significantly higher levels of PGHS-1 (Figure 7c) and PGHS-2 (Figure 7d) mRNA levels were expressed in cells obtained from ApoE^−/−iNOS^−/− mice compared to ApoE^−/− control mice (when normalized to GAPDH mRNA levels).

Figure 7.

PGHS-1 and PGHS-2 protein and mRNA levels are significantly higher in smooth muscle cells cultured from ApoE^−/−iNOS^−/− than ApoE^−/− mice. Aortic smooth muscle cells were cultured from ApoE^−/− and ApoE^−/−iNOS^−/− mice as described in Materials and Methods. The cells were grown to confluency in DMEM containing 10% FBS, then lysed and analyzed in triplicate for PGHS-1 (a) and PGHS-2 (b) protein by Western blotting. In each case, the blot was also probed for actin. The bar graphs represent the protein/actin ratio for a and b. The results are representative of one experiment (with three replicates) that was repeated three times. RNA was also isolated from ApoE^−/− and ApoE^−/−iNOS^−/− smooth muscle cells that were cultured in DMEM containing 10% FBS. c: After isolation, RNA (20 μg per lane) was loaded onto a gel in triplicate and separated by electrophoresis, transferred to a nylon membrane, and hybridized with a ³²P-labeled probe for PGHS-1 to yield a signal at 2.8 kb. The blot was then rehybridized with a ³²P-labeled GAPDH cDNA probe (signal at 1.3 kb). d: The same blot was later rehybridized with a ³²P-labeled cDNA probe for PGHS-2 to yield a signal at 4.3 kb. The bar graphs represent the PGHS-1/GAPDH (c) and PGHS-2/GAPDH (d) ratios. The results are representative of one experiment repeated twice.

Smooth Muscle Cells Cultured from ApoE^−/− and ApoE^−/−iNOS^−/− Mice Produce Significantly Different Levels of Eicosanoids

Smooth muscle cells cultured from ApoE^−/− and ApoE^−/−iNOS^−/− mice were exposed to arachidonic acid (20 μmol/L) and the supernatant medium was analyzed for PGE₂ formation. Figure 8 shows that in the absence of added arachidonic acid, basal levels of PGE₂ produced by cultured ApoE^−/− and ApoE^−/−iNOS^−/− smooth muscle cells were low. On addition of arachidonic acid (20 μmol/L), increased levels of PGE₂ were produced in both cases compared to basal levels. However, ApoE^−/−iNOS^−/− smooth muscle cells produced significantly higher levels of PGE₂ compared to ApoE^−/− smooth muscle cells. After incubation with IFN-γ and LPS, which are known to synergistically induce both PGHS-2 and iNOS expression,⁵² significantly higher levels of PGE₂ were once again observed in the supernatant medium from ApoE^−/−iNOS^−/− compared to ApoE^−/− smooth muscle cells after arachidonic acid addition (20 μmol/L). When challenged with LPS/IFN-γ, ApoE^−/− and ApoE^−/−iNOS^−/− smooth muscle cells produced similar levels of PGHS-2 with no significant differences (data not shown). Nitrite/nitrate production and iNOS expression under these conditions, however, was observed in ApoE^−/− mice but not in ApoE^−/−iNOS^−/− mice (data not shown). Interestingly, there was no significant difference between PGE₂ produced by arachidonic acid-stimulated ApoE^−/−iNOS^−/− cells in the presence or absence of LPS/IFN-γ, suggesting that the ApoE^−/−iNOS^−/− cells produce maximal levels of PGHS-2.

Figure 8.

PGE₂ production is significantly higher in ApoE^−/−iNOS^−/− than in ApoE^−/− smooth muscle cells. Aortic smooth muscle cells were cultured from ApoE^−/− and ApoE^−/−iNOS^−/− mice as described in Materials and Methods. The cells were grown in DMEM containing 10% FBS to 90% confluency, and then incubated in DMEM with and without LPS (10 μg/ml) and IFN-γ (100 U/ml) at 37°C. After 24 hours, the supernatant was replaced with fresh DMEM containing 10% FBS and the cells incubated at 37°C. After 1 hour, the medium was removed and stored at −80°C. The cells were then exposed to DMEM (10% FBS) containing arachidonic acid (20 μmol/L) and incubated at 37°C. After 1 hour, the supernatant was removed and stored at −80°C. The supernatants were analyzed for PGE₂ formation. The data represent the average from three experiments in which three measurements were recorded per experiment.

These data indicate that significantly higher PGE₂ levels were produced by smooth muscle cells cultured from ApoE^−/−iNOS^−/− than ApoE^−/− mice. This result is consistent with the observation that ApoE^−/−iNOS^−/− cells express more PGHS-1 and PGHS-2 mRNA and protein than ApoE^−/− cells. In addition, smooth muscle cells cultured from ApoE^−/−iNOS^−/− mice were maximally stimulated with respect to PGHS-2 expression and were not influenced by the addition of LPS/IFN-γ. Although LPS/IFN-γ challenge in ApoE^−/− smooth muscle cells produced a modest increase in PGE₂ over basal levels, the levels produced did not match those achieved by the ApoE^−/−iNOS^−/− cells. These results further suggest a regulatory role of iNOS on PGHS activity.

Discussion

This study demonstrates for the first time that iNOS influences PGHS expression and activity and contributes to PGHS-1 nitration during the disease state of atherosclerosis. Owing to the numbers of mice required for this study, both male and female ApoE^−/− and ApoE^−/−iNOS^−/− mice were used and no attempt was made to determine the effects of gender.

PGHS-2 and iNOS have previously been observed in human atherosclerotic tissue,^5,6,64–66 and levels of PGHS-2 were also found to be increased in murine atherosclerotic aortae.^57,67 In this current study, we demonstrate that PGHS-2 is present in the murine atherosclerotic lesions of ApoE^−/− and ApoE^−/−iNOS^−/− mice fed a Western diet for 6 months but is absent in the tissue surrounding the lesion. In addition, we show that iNOS is induced in lesions in ApoE^−/− mice and is virtually absent in the surrounding tissue. As expected, iNOS is absent in both lesions and surrounding noninvolved tissues obtained from the ApoE^−/−iNOS^−/− mice.

We have previously observed PGHS-1 nitration in human atherosclerotic tissue obtained from endarterectomies.²⁹ We now report that PGHS-1 is nitrated in lesions obtained from ApoE^−/− mice but markedly less so in lesions from ApoE^−/−iNOS^−/− mice. Notably, immunofluorescence data does not exclude the possibility that PGHS-1 nitration may occur at low levels in the absence of iNOS. Accordingly, eNOS may contribute to PGHS-1 nitration, although PGHS-1 nitration is primarily dependent on iNOS protein expression. The coincidental lack of PGHS-1 nitration and iNOS in the surrounding noninvolved aortic tissue in the ApoE^−/− mouse seems to suggest, however, that PGHS-1 nitration is iNOS-dependent.

Our data demonstrate that iNOS is involved in PGHS-1 nitration at the site of the atherosclerotic lesion and that iNOS provides the first step in a mechanism of PGHS-1 nitration. NO release from iNOS could combine with superoxide to form peroxynitrite (ONOO⁻), a reactive species with enhanced oxidizing capability.^68,69 Of the deleterious effects, ONOO⁻ has been implicated in myocardial reoxygenation injury,⁷⁰ initiating lipid peroxidation of membranes,⁷¹ reacting with sulfhydryl groups of proteins,⁷² and causing oxidative damage by introducing Tyr nitration.⁷³ We have previously demonstrated that ONOO⁻ causes nitration of purified PGHS-1 and PGHS-1 in smooth muscle cells.²⁹ In this study, we now identify Tyr385 as a specific site of nitration by ONOO⁻ in purified PGHS-1 by mass spectrometry. Tyr385 is located at the apex of the cyclooxygenase binding channel and is thought to be involved in the catalytic mechanism of PGHS enzymes.^61,62 Tyr385 is involved in radical formation that initiates arachidonic acid oxygenation, leading to eicosanoid formation,^74,75 although the radical may also be localized on a different Tyr residue.^62,76–79 Nitration of Tyr385 would predictably lead to a complete loss of arachidonic acid metabolism.⁶¹ Although NO has previously been demonstrated to couple with the Tyr385 radical of PGHS-1,⁸⁰ leading to the eventual nitration of Tyr385,⁸¹ this represents the first identification of a specific nitrated Tyr residue in PGHS-1 by ONOO⁻.

Under pathophysiological conditions, ONOO⁻ formation may occur, because both iNOS expression and superoxide production are concomitantly elevated,^64,65,82 and the rate of ONOO⁻ production (6.7 × 10⁹ mol/L⁻¹second⁻¹) outcompetes the rate at which superoxide dismutase scavenges superoxide (2 × 10⁹ mol/L⁻¹second⁻¹).^68,83 Thus, the presence of ONOO⁻ during atherosclerosis seems likely, although doubts over its formation under physiological or pathophysiological conditions persist.^84–86 Other mechanisms can also lead to nitrotyrosine formation. For instance, peroxidase enzymes (eg, myeloperoxidase, MPO) can use nitrite (NO₂⁻) and hydrogen peroxide (H₂O₂) as substrates to catalyze tyrosine nitration in proteins.^87–89 The MPO/NO₂⁻/H₂O₂ system has been demonstrated to nitrate and to cause LDL oxidation, which may contribute to the development of atherosclerosis.⁹⁰ On the other hand, others have shown that MPO^−/− mice show greater levels of nitrotyrosine formation (and larger infarct volumes) than wild-type mice, leading to the conclusion that the presence of MPO counters nitration.⁹¹ In the absence of a peroxidase, NO₂⁻ and H₂O₂ can nitrate at low pH (<2).⁹² Nitrotyrosine formation in proteins may also occur in the presence of hypochlorous acid (HOCl) and NO₂⁻, but this reaction also yields 3-chlorotyrosine,^93,94 which may also cause oxidative damage in tissues.⁹⁵ Despite these alternative mechanisms, nitration of Tyr residues in proteins produces a stable end-product that can be detected immunologically. Identification of nitrated PGHS-1 in lesions from ApoE^−/− mice by mass spectrometry has not yet proven feasible (due to the large amounts of tissue required coupled with the necessity of having reasonably pure samples), but it is the subject of ongoing efforts. Further studies will be required to elucidate the mechanism by which Tyr nitration in PGHS-1 occurs in atherosclerotic tissue and whether Tyr385 in PGHS-1 represents a physiological target for nitration.

We have previously demonstrated that PGHS nitration coincides with a loss in activity,²⁹ but it is unclear whether the extent of nitration and associated loss in function contributes to the progression of atherosclerosis. Surprisingly, PGHS-2 nitration was not detected, suggesting that this modification does not occur to a significant extent or possibly that antibodies used are not sufficiently sensitive for detection. Although it is generally assumed that PGHS-2 is the most important isozyme in atherosclerosis and inflammation, PGHS-1 may also play a key role in maintaining the balance of eicosanoids during this disease state.⁶⁷

Previous studies have found that endothelial NO synthase (eNOS) is an important regulator of vascular function. Thus,ApoE^−/−eNOS^−/− mice were found to have higher blood pressure and increased atherosclerotic lesion size compared to normotensive, atherosclerotic ApoE^−/− mice.⁴⁴ The lack of eNOS in ApoE^−/− mice caused them to develop kidney damage and have decreased kidney weight with glomerular lipid deposition and calcification.⁴⁴ The inducible form of NOS is usually absent in nondiseased vessels, and thus may not be expected to contribute to vascular function under normal conditions. However, under conditions of inflammation, iNOS is induced and is capable of generating NO, possibly in the μm range.⁹⁶ A genetic lack of iNOSin ApoE^−/− mice does not affect blood pressure nor alter the development of atherosclerotic lesions as compared to ApoE^−/− mice when fed a normal chow diet.⁴⁴ However, when ApoE^−/−iNOS^−/− mice are fed a high-fat diet, the atherosclerotic lesions are diminished in size compared to ApoE^−/− mice.^45,46 Our experience from performing immunoprecipitation experiments indicated that lesions from ∼5 ApoE^−/− and ∼9 ApoE^−/−iNOS^−/− mice (at 6 months on a high cholesterol diet) typically provided enough protein for one such experiment. The difference in the numbers of mice required likely reflects the fact that less lesion formation occurs in ApoE^−/−iNOS^−/− mice than in ApoE^−/− mice fed a Western diet⁴⁶ and indicates that iNOS contributes to lesion size under these dietary conditions (possibly via protein nitration). Cholesterol, cholesteryl esters, and lipoperoxides (a marker for oxidative stress) have been found to be significantly diminished in ApoE^−/−iNOS^−/− mice compared to ApoE^−/− mice.^45,46 These results indicate that iNOS contributes to the size of atherosclerotic lesions in ApoE^−/− mice, perhaps by promoting inflammation and oxidative stress at the site of the lesion.

Finally, we show that smooth muscle cells cultured from ApoE^−/−iNOS^−/− mice express significantly higher levels of PGHS-1 and PGHS-2, as well as increased levels of PGE₂, compared to those cultured from ApoE^−/− mice. Cells challenged with LPS and IFN-γ, immunoactivators known to induce iNOS as well as PGHS-2, did not boost PGE₂ levels in ApoE^−/− cells to the high levels obtained from ApoE^−/−iNOS^−/− cells. Northern blot analysis of PGHS-1 and PGHS-2 mRNA demonstrated that both PGHS-1and PGHS-2transcription are elevated in ApoE^−/−iNOS^−/− compared to ApoE^−/− smooth muscle cells. The effect of iNOS deletion on eicosanoid production has previously been studied in mice.³⁰ It was found that PGE₂ production was decreased in peritoneal macrophages and in urine obtained from iNOS^−/− mice compared to wild-type mice. These observations are consistent with NO or NO-derived species having a stimulatory effect on PGHS in inflammatory cells and in the kidney. Indeed, ONOO⁻ is known to stimulate purified PGHS and PGHS in cells^16,17 and could explain why higher levels of urinary and macrophage eicosanoids are observed in wild-type mice. Our current study investigated the effect of iNOS deletion in a mouse model of atherosclerosis (ie, using ApoE^−/− mice). Previous investigations have demonstrated that urinary eicosanoid levels are elevated in human and murine atherosclerosis (using either ApoE^−/− or low-density lipoprotein receptor knockout, LDLR^−/−, mice).^9–11 Although the measurement of urinary eicosanoids represents a noninvasive method, the levels are not specific to the tissue of origin of these compounds. We chose, therefore, to investigate levels of eicosanoids produced by smooth muscle cells cultured from both ApoE^−/− and ApoE^−/−iNOS^−/− mice. Surprisingly, we have observed that iNOS deletion in ApoE^−/− mice has a stimulatory effect on PGHS-1 and PGHS-2 pretranslational levels. Others have demonstrated that in the absence of eNOS, other vasodilator pathways compensate to maintain near-normal coronary arterial function.³⁹ In this regard, coronary arteries from wild-type mice responded to acetylcholine by a mechanism involving NO from eNOS; however, coronary arteries from eNOS^−/− mice dilated by a mechanism dependent on PGHS. These observations suggest a compensatory ability by PGHS in response to chronic loss of NOS, which has also been reported by others.^34–38 It is conceivable that mechanisms may exist that compensate for the lack of iNOS. Because PGHS-mediated PGE₂ production is increased in cultured ApoE^−/−iNOS^−/− versus ApoE^−/− (iNOS-expressing) mouse aortic smooth muscle cells, we infer that iNOS can markedly suppress PGHS activity in vascular cells. The extent to which iNOS-dependent nitration of Tyr in PGHS-1 contributes to perturbed eicosanoid biosynthesis in human atherosclerotic lesions awaits future investigations.