Inflammatory Monocytes Determine Endothelial Nitric-oxide Synthase Uncoupling and Nitro-oxidative Stress Induced by Angiotensin II*

Background: Inflammatory monocytes are drivers of vascular injury and disease. Results: Depletion of lysozyme M-positive monocytes prevents eNOS uncoupling and iNOS-derived nitro-oxidative stress. Conclusion: Monocytes determine eNOS and iNOS function by directly modulating tetrahydrobiopterin bioavailability. Significance: Understanding the impact of inflammation on endothelial function in detail is essential to identify tailored therapeutic strategies. Endothelial nitric-oxide synthase (eNOS) uncoupling and increased inducible NOS (iNOS) activity amplify vascular oxidative stress. The role of inflammatory myelomonocytic cells as mediators of these processes and their impact on tetrahydrobiopterin availability and function have not yet been defined. Angiotensin II (ATII, 1 mg/kg/day for 7 days) increased Ly6Chigh and CD11b+/iNOShigh leukocytes and up-regulated levels of eNOS glutathionylation in aortas of C57BL/6 mice. Vascular iNOS-dependent NO formation was increased, whereas eNOS-dependent NO formation was decreased in aortas of ATII-infused mice as assessed by electron paramagnetic resonance (EPR) spectroscopy. Diphtheria toxin-mediated ablation of lysozyme M-positive (LysM+) monocytes in ATII-infused LysMiDTR transgenic mice prevented eNOS glutathionylation and eNOS-derived Nω-nitro-l-arginine methyl ester-sensitive superoxide formation in the endothelial layer. ATII increased vascular guanosine triphosphate cyclohydrolase I expression and biopterin synthesis in parallel, which was reduced in monocyte-depleted LysMiDTR mice. Vascular tetrahydrobiopterin was increased by ATII infusion but was even higher in monocyte-depleted ATII-infused mice, which was paralleled by a strong up-regulation of dihydrofolate reductase expression. EPR spectroscopy revealed that both vascular iNOS- and eNOS-dependent NO formation were normalized in ATII-infused mice following monocyte depletion. Additionally, deletion as well as pharmacologic inhibition of iNOS prevented ATII-induced endothelial dysfunction. In summary, ATII induces an inflammatory cell-dependent increase of iNOS, guanosine triphosphate cyclohydrolase I, tetrahydrobiopterin, NO formation, and nitro-oxidative stress as well as eNOS uncoupling in the vessel wall, which can be prevented by ablation of LysM+ monocytes.

triphosphate cyclohydrolase I, tetrahydrobiopterin, NO formation, and nitro-oxidative stress as well as eNOS uncoupling in the vessel wall, which can be prevented by ablation of LysM ؉ monocytes.
Angiotensin II (ATII) 3 -induced vascular dysfunction is characterized by increased vascular oxidative stress and loss of NO bioavailabilty. As predominant sources of superoxide (O 2 . ), the vascular NADPH oxidase, an uncoupled endothelial nitric-oxide synthase (eNOS), xanthine oxidase, and mitochondria have been identified (1)(2)(3)(4)(5). The uncoupled eNOS has gained growing attention, because a dysfunctional eNOS leads to a decreased nitric oxide (NO ⅐ ) bioactivity in the vasculature and turns into a source of O 2 . by transferring electrons to molecular oxygen in the uncoupled state, thereby shifting the O 2 . /NO ⅐ equilibrium toward O 2 . (6). The uncoupling reaction of eNOS is triggered by oxidation (e.g. by peroxynitrite, ONOO Ϫ ) of the eNOS cofactor tetrahydrobiopterin (BH 4 ) leading to the formation of the BH 3 radical and subsequently to dihydrobiopterin (BH 2 ) (7). Intracellular depletion of BH 4 is counteracted mainly by the activity of the biopterin-synthesizing enzyme guanosine triphosphate cyclohydrolase I (GTPCH) and the BH 2 -reducing enzyme dihydrofolate reductase (DHFR). Importantly, the cascade of eNOS uncoupling depends on excess production of O 2 . , which avidly reacts with NO ⅐ to form ONOO Ϫ . Therefore, superoxide anion acts as a so-called "kindling radical" to fuel eNOS uncoupling, making identification and control of the source of O 2 . a promising therapeutic target. In parallel, GTPCH as the rate-limiting enzyme of BH 4 synthesis has classically been defined as constitutively active in macrophages and can be induced by proinflammatory cytokines like tumor necrosis factor ␣ and interferon ␥ (IFN-␥) (8). In inflammatory cells, it is necessary to supply BH 4 to the inducible NO synthase (iNOS), a signature enzyme of inflammatory cells like proinflammatory monocytes or macrophages. Equipped with this machinery and combined with their nicotine amide dinucleotide phosphate (NADPH) oxidase activity, inflammatory cells are capable of performing their innate immune functions like cytotoxic microbial killing but also of exerting harmful nitro-oxidative stress in inflammatory diseases like atherosclerosis. Monocytes that express high levels of lymphocyte antigen 6 complex locus C1 (Ly6C) have been shown to dominate monocytosis in atherosclerosis, to adhere to endothelium, and to give rise to inflammatory macrophages in atheromata (9).
Recently, we identified infiltrating lysozyme M-positive (LysM ϩ ) CD11b ϩ /Gr-1 ϩ monocytes and vascular CD11b ϩ /F4/ 80 ϩ macrophages as critical mediators of ATII-induced vascular dysfunction and arterial hypertension. In that study, we found that both eNOS-dependent vascular relaxation and overall aortic expression of iNOS can be normalized by depletion of inflammatory monocytes. In addition, we found that iNOS-derived oxidative stress is largely determined by the mutual activation of proinflammatory natural killer cells and monocytes and intact IFN-␥/interleukin-12 signaling in the vasculature of ATII-infused mice (10,11).
We therefore wanted to investigate the end point that inflammatory monocytes determine eNOS uncoupling via controlling BH 4 availability and function. We set out to test the hypothesis that depletion of inflammatory LysM ϩ myelomonocytic cells would prevent both iNOS activity and eNOS uncoupling and thereby preserve NO bioavailabilty in ATII-induced vascular dysfunction. We found that ATII induces a Ly6C high inflammatory cell-dependent increase of iNOS, GTPCH, BH 4 , NO ⅐ formation, and nitro-oxidative stress as well as eNOS uncoupling in the vessel wall, which can be prevented by ablation of LysM ϩ monocytes.
Flow Cytometric Analysis of Aortic Lysates-Prior to analysis, mouse aortas were lysed by 6.5 units/ml liberase TM (Roche Applied Science) for 20 min at 37°C. To block nonspecific Fc receptor-mediated binding, single cell suspensions were preincubated with unlabeled mAb against CD16/CD32. Cells were stained for 20 min with CD45 APC-efluour-780, CD11b PE, Gr-1 Horizon, and Ly6c PerCP-Cy5.5 and for outgating of dead cells additionally with Fixable Viability Dye eFluor506 (CD11b from Pharmingen; all other antibodies from eBioscience, San Diego). iNOS FITC was stained after fixation with Fixation/ Permeabilization solution from Pharmingen. A minimum of 100,000 events was acquired using the FACS Canto II (BD Biosciences), and viable cells were analyzed with FACSDiva software. For analysis of iNOS expression, isolated cells were stimulated with 10 ng/ml mIFN-␥ and 1 g/ml LPS in the presence of 10 g/ml brefeldin A overnight. After stimulation and surface staining, cells were fixed with the Cytofix/Cytoperm kit as indicated in the manufacturer's instructions (Pharmingen) and stained for intracellular iNOS FITC (Pharmingen) or matched isotype control. Additionally, iNOS-positive cells A549 cells were used as a positive control (14).
Reconstitution of Depleted Mice with Monocytes and Neutrophils-CD11b ϩ Gr-1 ϩ monocytes were prepared from venous blood of C57BL/6, Agrt1 Ϫ/Ϫ , and gp91 phox Ϫ/y mice by negative selection using magnetic activated cell sorting after discarding granulocytes following Histopaque 1083 gradient. LysM iDTR mice were monocyte-depleted by diphtheria toxin injections according to our protocol (diphtheria toxin solved in PBS once daily, 25 ng/g from days 1 to 3 then 5 ng/g thereafter) and ATII-infused (1 mg/kg/day for 7 days, starting at day 4 of depletion protocol) and were reconstituted in vivo by single i.v. injection with 1.5 ϫ 10 6 monocytes at day of the depletion protocol.
Oxidative Fluorescent Microtopography-Isolated aorta was cut into 3-mm rings, incubated in Krebs/Hepes solution for 15 min at 37°C in the presence or absence of the eNOS inhibitor N -nitro-L-arginine methyl ester (L-NAME, 10 M), embedded in aluminum cups of about 1 ml of a polymeric resin (Tissue Tek, Sakura Finetek, Alphen aan den Rijn, Netherlands), and frozen in liquid nitrogen. Cryosections (6 m) were stained with the superoxide-sensitive dye dihydroethidium (DHE, 1 M in PBS) and incubated for 30 min at 37°C. Green and red fluorescence was detected using a Zeiss Axiovert 40 CFL camera (Zeiss, Oberkochen, Germany). Sections of all four study arms were analyzed in parallel with identical imaging parameters.
Oxidative Burst of Whole Blood-Venous blood was drawn into 0.1 volume of 3.8% sodium citrate. The blood was kept at room temperature and diluted 1:50 in Dulbecco's PBS (without Mg 2ϩ , Ca 2ϩ , and bicarbonate). The L-012 (100 M)-enhanced chemiluminescence (ECL) signal was counted in 0.5-ml samples in the absence or presence of phorbol 12,13-dibutyrate (10 M) at intervals of 30 s for 10 min. ECL was expressed as counts/ min/ml after incubation for 10 min.
Protein and mRNA Expression-Protein expression was assessed using SDS-PAGE and Western blotting as described previously. mRNA expression was analyzed by quantitative real time RT-PCR as described previously (11). Aortas were cleaned of adhesive adipose tissue, rinsed, and snap-frozen. For protein expression analyses, protein suspensions from homogenized aortic tissue were submitted to SDS-PAGE and immunoblotting (Bio-Rad), using antibodies against ␣-actinin (mouse, monoclonal, Sigma), eNOS (mouse, monoclonal, 1:1000, BD Biosciences), heme oxygenase-1 (HO-1, mouse, monoclonal, StressGen, San Diego, GTPCH (mouse, monoclonal, Abnova, Heidelberg, Germany), dihydrofolate reductase (DHFR, mouse, monoclonal, Abnova, Heidelberg, Germany), and Nox4 (rabbit, polyclonal, Novus Biologicals, Littleton, CO) followed by peroxidase-labeled secondary antibody against mouse or rabbit IgG (Vector Laboratories, Burlingame, CA). Immunodetection was accomplished with either SuperSignal Substrate (Pierce) or ECL Reagent (Amersham Biosciences). Bands were evaluated by densitometry. mRNA expression was analyzed by quantitative real time RT-PCR using a 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). Briefly, total RNA from mouse aorta was isolated according to the manufacturer's protocol of the RNeasy fibrous tissue mini kit (Qiagen, Hilden, Germany). 0.5 g of total RNA was used for real time RT-PCR analysis with the QuantiTect TM Probe RT-PCR kit (Qiagen, Hilden, Germany). TaqMan Gene Expression assays (Applied Biosystems, Foster City, CA) for TATA boxbinding protein (TBP; Mm00446973_m1) and inducible nitricoxide synthase (iNOS, Nos2; Mm00440485_m1) were purchased as probe and primer sets.
The comparative ⌬Ct method was used for relative mRNA quantification (15). Gene expression was normalized to the endogenous control, TBP mRNA, and the amount of target gene mRNA expression in each sample was expressed relative to that of control. eNOS S-Glutathionylation-Briefly, M-280 sheep anti-mouse IgG-coated beads from Invitrogen were used along with a monoclonal mouse eNOS antibody (BD Transduction Laboratory). The beads were loaded with the eNOS antibody and cross-linked according to the manufacturer's instructions. Next, the aortic homogenates were incubated with the eNOS antibody beads, precipitated with a magnet, washed, transferred to gel, and subjected to SDS-PAGE followed by a standard Western blot procedure using a monoclonal mouse antibody against S-glutathionylated proteins from Virogen (Watertown, MA) at a dilution of 1:1000 under nonreducing conditions. Disappearance of the signal on incubation with 2-mercaptoethanol served as a control. After the membrane was stripped, the bands were stained for eNOS to allow normalization of the signals. All signals were normalized on the eNOS staining of the same sample.
NO Measurement by Electron Paramagnetic Resonance Spectroscopy-Aortic NO ⅐ formation was measured using EPR-based spin trapping with iron-diethyldithiocarbamate (Fe(DETC) 2 ) colloid as described in general previously (5,16). Whole aortas were cleaned of fat and connective tissue and cut into 3-mm rings. For iNOS stimulation, aortic rings were incubated in RPMI 1640 medium, 10% FCS ϩ 1% penicillin/streptomycin ϩ 10 g/ml LPS (from Escherichia coli, Sigma) for 19 -24 h at 37°C, 5% CO 2 . LPS-pretreated or freshly prepared rings were transferred in 1 ml of Krebs/Hepes buffer on a 24-well plate. In selected experiments, 10 M iNOS inhibitor N- [3(aminomethyl)benzyl]acetamidine, dihydrochloride (1400W) was added. For eNOS stimulation, samples were incubated with 10 M calcium ionophore (A23187) for 2 min on ice before the colloid solution was added. NaDETC (5.4 mg) and FeSO 4 ⅐7H 2 O (3.4 mg) were separately dissolved under argon gas bubbling in two 15-ml volumes of ice-cold PBS with Ca 2ϩ /Mg 2ϩ . These solutions were rapidly mixed to obtain a colloid Fe(DETC) 2 solution (0.4 mM), which was added immediately to the rings (1 ml). After 60 min of incubation at 37°C, aortic rings were placed at a fixed position in a 1-ml syringe with the top removed in Krebs/Hepes buffer and frozen in liquid nitrogen (so that the entire aortic sample was placed within a 100-l volume of the syringe). For measurement, the frozen cylinder with the aortic sample was pressed out of the syringe and placed in a special Dewar vessel (Magnettech, Berlin, Germany) filled with liquid nitrogen. The localization of the aortic sample was adjusted to the middle of the resonator. EPR conditions were as follows: B0 ϭ 3274 G, sweep ϭ 110 G, sweep time ϭ 60 s, modulation ϭ 7000 mG, power ϭ 10 milliwatts, using a Miniscope MS400 (Magnettech, Berlin, Germany).
Measurement of Aortic Levels of BH 4 , BH 2 , and Biopterin-Whole mouse aorta was homogenized in ice-cold HCl (0.1 N). BH 4 , BH 2 , and biopterin content were assessed by high performance liquid chromatography using sequential electrochemical and fluorescence detection. Isocratic elution was performed (flow rate 0.85 ml/min) using a Nucleosil 100 -5C18 column (25 cm, 4.6 mm, 5 M, Supelco) and mobile phase (octyl sulfate sodium salt 0.6 mM, EDTA 0.5 mM, diethylenetriaminepentaacetic acid 0.25 mM, DL-dithiothreitol 1.25 mM, NaH 2 PO 4 50 mM, pH 2.8), containing 2% acetonitrile (v/v). The electrochemical detection system for measuring BH 4 consisted of an ESA Coulochem III detector equipped with a boron-doped diamond electrochemical cell (model 5040) set at a potential of ϩ450 mV and a guard cell set at a potential of ϩ800 mV to oxidize contaminants in the mobile phase. BH 2 was converted to a fluorescent form by post-column electrochemical oxidation. An ESA Coulochem II with model 5011 analytical cell set at a potential of ϩ800 mV was used. The electrochemical electrodes were followed in series by a Jasco FP-2020 Plus Intelligent fluorescence detector. The excitation and emission wavelengths were set at 348 and 444 nm, respectively. Data acquisition and analysis were performed using the Chrome-leon7 software (Dionex). BH 4 , BH 2 , and biopterin concentrations were expressed as picomoles/g protein.
Vascular Relaxation Studies-Vascular relaxation of isolated aortic rings of C57BL/6 and iNOS Ϫ/Ϫ mice subjected to ATII infusion or sham was assessed. Isolated aortas were cut into 4-mm segments and mounted on force transducers (Kent Scientific Corp., Torrington, CT, and PowerLab, AD Instruments, Spechbach, Germany) in organ chambers filled with Krebs-Henseleit. In selected experiments, 10 M iNOS inhibitor 1400W was added to the organ bath. To test for vasorelaxation in response to acetylcholine (ACh), aortic segments were stretched gradually over 1 h to reach a resting tension of 3.0 g. Following preconstriction with prostaglandin F 2 (3 nM), cumulative concentration-relaxation curves were recorded in response to increasing concentrations of ACh (10 Ϫ9 to 10 Ϫ5.5 M).
Blood Pressure Recordings-Systolic blood pressure was measured using the tail-cuff method in conscious trained mice (tail-cuff plethysmography, Kent Scientific CODA STD). Three measurements were taken for each mouse and averaged to yield one data point.
Statistical Analysis-Data are expressed as mean Ϯ S.E. Statistical calculations were performed with GraphPad Prism 5 (GraphPad Software Inc, San Diego). Mann-Whitney test, paired or unpaired t test, one-way ANOVA, or Kruskal-Wallis test with post hoc Dunn test or Bonferroni test was used as appropriate. One asterisk indicates p values Ͻ0.05; two asterisks indicate p Ͻ 0.01, and three asterisks indicate p Ͻ 0.001, considered to be statistically significant.

ATII-induced Aortic Infiltration of Inflammatory Myelomonocytic Cells Impairs eNOS Function and Amplifies iNOS-
derived NO ⅐ Formation-In C57BL/6 mice, ATII infusion induced a drastic expansion of CD45 ϩ leukocytes in the aorta (Fig. 1A) with a strong and selective increase of CD11b ϩ Ly6C high monocytes, although the number and percentage of CD11b ϩ Ly6C Ϫ monocytes was not significantly altered (Fig. 1B). Importantly, ATII sharply increased iNOS (Nos2) mRNA expression in whole aortic lysates in parallel with the number of iNOS-positive CD11b ϩ inflammatory cells in the aortic wall (Fig. 1, C and  D). Levels of eNOS protein were increased in the aorta of ATIIinfused mice. However, eNOS protein isolated from the aorta of ATII-infused mice showed a significantly higher rate of S-glutathionylation compatible with eNOS uncoupling (Fig. 1, E and  F). Indeed, when we assessed endothelial reactive oxygen species formation with the O 2 . -sensitive dye DHE, we observed an increase in vessel sections obtain from ATII-infused mice that was inhibitable by L-NAME, indicating eNOS uncoupling and eNOS-derived O 2 . formation in ATII-induced hypertension and vascular inflammation (Fig. 1G). These observations were accompanied by a decrease in calcium ionophore enhanced NO ⅐ formation (Fig. 1H) and a drastic increase in lipopolysaccharide (LPS) enhanced NO ⅐ formation (Fig. 1I), indicating impaired eNOS function and increased iNOS activity in vessels of ATII-infused mice.

Depletion of LysM ϩ Inflammatory Cells Attenuates ATII-induced Vascular eNOS Uncoupling, GTPCH Up-regulation, and
Nitro-oxidative Stress-ATII infusion increased the O 2 . formation detectable in the DHE staining in the endothelial cell layer of mouse aortic cryosections, which could be blocked by the eNOS inhibitor L-NAME (10 M). O 2 . formation could be increased by L-NAME in the endothelium of sham-infused controls. Therefore, this ROS formation can be attributed to eNOS uncoupling. Depletion of LysM ϩ cells attenuated the eNOSderived O 2 . formation detectable in the DHE staining ( Fig. 2A).
eNOS-derived O 2 . formation was re-established by adoptively transferred monocytes but not when we reconstituted depleted LysM iDTR with Nox2-deficient or Agtr1-deficient monocytes (Fig. 2B). Accordingly, the increase of eNOS glutathionylation observed in ATII-infused control mice was prevented in LysM ϩ cell-depleted LysM iDTR mice indicating that inflammatory cells are crucial mediators of eNOS uncoupling (Fig. 2C). ATII-induced vascular oxidative stress was characterized not only by Nox2, Nox1, p67 phox , and p47 phox up-regulation (10) but also by an increase in Nox4 protein expression (data not shown), which was recently implicated in peroxynitrite-dependent eNOS uncoupling in response to ATII (17). Increased NADPH oxidase expression was accompanied by heme oxygenase-1 (HO-1) up-regulation known to counterbalance ATIIinduced oxidative stress. Like Nox4, HO-1 protein expression was normalized by LysM ϩ cell depletion (Fig. 2D).
Depletion of LysM ϩ Cells Attenuates iNOS-dependent NO ⅐ Formation in the Aortas of ATII-infused Mice-ATII in vivo increased aortic protein expression levels of GTPCH, the biopterin synthase, and the rate-limiting enzyme of BH 4 synthesis. In parallel, ATII augmented the aortic levels of total biopterin, although the ratio of BH 2 /biopterin was unaffected. Depletion of LysM ϩ cells prevented both GTPCH and biopterin increase in mouse aorta (Fig. 3, A and B, and data not shown). Vascular levels of BH 4 were not only determined by de novo biosynthesis but also by recovery via tetrahydrofolate and the DHFR (salvage pathway). Interestingly, ATII-infused mice had significantly increased levels of DHFR only when inflammatory cells were depleted (Fig. 3C). At the same time, vascular content of BH 4 was significantly increased in both LysM and LysM iDTR in response to ATII infusion in vivo, but this increase was more pronounced in the depleted mice (Fig. 3, D and F). When normalized on biopterin content, BH 4 was drastically increased only in monocyte-depleted ATII-infused mice (Fig. 3E), which accounts for the reversal of eNOS uncoupling. As a functional consequence, depletion of LysM ϩ cells attenuated the decline in eNOS-derived NO ⅐ formation and strongly reduced iNOS activity in aortas of ATII-infused mice (Fig. 3, G and H), leaving only a background of LPS-inducible NO formation, which can be related to iNOS expressed by vascular smooth muscle cells or endothelial cells. These findings indicate the following: (i) ATII causes an increased demand for BH 4 in the vasculature, presumably to feed iNOS; and (ii) depletion of LysM ϩ inflammatory cells has a differential and beneficial impact on eNOS and iNOS function via handling BH 4 bioavailability.
iNOS Inhibition Abolishes ATII-induced Nitro-oxidative Stress and Vascular Dysfunction-To test the functional implication of iNOS for vascular dysfunction with a pharmacologic approach, we used 1400W, a specific irreversible inhibitor of iNOS (18). Acute ex vivo blockade of iNOS completely abolished the vascular NO ⅐ signal provoked by LPS incubation (Fig.  4A) and significantly attenuated ATII-induced vascular endothelial dysfunction (assessed by the acetylcholine concentration-relaxation curve, Fig. 4B and Table 1). In contrast to C57BL/6 mice (see Fig. 1G), ATII-infused iNOS Ϫ/Ϫ mice had increased endothelial O 2 . formation that was not significantly different from iNOS Ϫ/Ϫ controls (Fig. 4C). Oxidative burst in whole blood and endothelial dysfunction was prevented in ATII-infused iNOS Ϫ/Ϫ mice (Fig. 4, E and F, and had an increase in systolic blood pressure comparable with ATII-infused C57BL/6 mice (Fig. 4D). These results suggest that nitro-oxidative stress mediated by inflammatory cells in the vasculature is iNOS-derived that the majority of ATII induces iNOS expression and activity can be attributed to inflammatory cells (compare Figs. 3H and 4A), and that iNOS-derived nitro-oxidative stress causes vascular endothelial dysfunction evoked by ATII in vivo.

DISCUSSION
We present here novel data showing that CD11b ϩ Ly6C high iNOS ϩ monocytes containing a functional Nox2 and the angiotensin II receptor type 1 (AT 1 R) are mediators of eNOS uncoupling and dysfunction in response to ATII. ATII-induced iNOS activity and expression in the vasculature critically depends on inflammatory cells and can be normalized by LysM ϩ cell depletion.
The rate-limiting enzyme of BH 4 synthesis, the GTPCH, is induced in the vessel wall of ATII-infused mice as are vascular BH 4 levels; these processes are determined by the presence of LysM ϩ inflammatory cells that contain iNOS. We also provide evidence that ablation of LysM ϩ cells is beneficial for eNOS uncoupling. This results from a net effect of three synergistic processes as follows: 1st, prevention of oxidative decay of BH 4 by eliminating the leading source of nitro-oxidative stress from the system; 2nd, prevention of inflammation-induced up-regulation of the pacemaker enzyme of BH 4 , the GTPCH; and 3rd, activation of the salvage pathway of BH 4 by up-regulation of vascular DHFR.
The NADPH oxidase has been implicated in eNOS uncoupling because of the fact that it is one of the  (6) presumably caused by BH 4 oxidation mediated by radicals, which may originate from phagocyte type NADPH oxidases. S-Glutathionylation of eNOS, a novel and well accepted surrogate parameter of eNOS uncoupling (5,22), was increased by ATII in the vasculature and normalized by depletion of LysM ϩ cells. Because we had previously shown that ATII-induced vascular dysfunction critically depends on monocytes harboring both a functional AT 1 R and Nox2 (10), we performed adoptive transfers with monocytes from wild type, gp91 phox Ϫ/y , and Agtr1 Ϫ/Ϫ mice. We concluded from these reconstitution experiments that monocytes require Nox2 and Agtr1 to mediate eNOS uncoupling. Interestingly, Nox2and Agtr1-deficient monocytes express lower levels of Gtpch mRNA as compared with C57BL/6 controls (data not shown).
These observations indicate that inflammatory Agtr1 ϩ monocytes that are capable of forming ROS, iNOS-derived NO ⅐ , and ONOO Ϫ are indispensable for ATII-induced eNOS uncoupling.
With our study, we also shed new light on the impact of ATII-induced iNOS activity in vascular dysfunction. It is well accepted that iNOS marks an inflammatory phenotype of myelomonocytic cells (classically activated macrophages and proinflammatory monocytes, "M1") (23). iNOS is increased in myelomonocytic cells in an IFN-␥-dependent fashion (24) driving a proinflammatory monocyte-to-macrophage differentiation (25), and simultaneously, GTPCH is increased in these cells by the same cytokine as well (8). In fact, we show here for the first time that ATII selectively increases the number Ly6C high myelomonocytic cells expressing high levels of iNOS in the vascular wall. This finding supports the concept that ATII/AT 1 R signaling works as a fundamental proinflammatory stimulus in myelomonocytic cells in general (26,27). It is also in line with previous data showing that ATII-induced vascular dysfunction and inflammation are IFN-␥-dependent mechanisms and that mice defective in IFN-␥ formation are marked by a decreased iNOS expression and peroxynitrite burden in the vessel wall (11).
In this regard, it is conclusive that GTPCH is concordantly expressed with iNOS to supplement it with BH 4 . Oxidation of BH 4 to BH 2 with a consecutive drop of vascular BH 4 levels as anticipated in a disease state of increased vascular oxidative stress (6) obviously does not outweigh the augmentation of vascular BH 4 in ATII-infused mice. Here, vascular BH 4 levels are rather coupled with an increased need for this essential cofactor of NO ⅐ synthesis in a setting of an up-regulated iNOS-dependent NO ⅐ formation in response to ATII. Regarding iNOS expression, our findings also indicate that ATII is a particularly potent proinflammatory impulse (10,11), comparable with the power of classical iNOS inducers like LPS (24).
Targeted GTPCH overexpression in the endothelium has been shown to improve vascular endothelial dysfunction in diabetes and atherosclerosis (28,29). Supplementation of exogenous BH 4 or its precursors or stimulation of endogenous BH 4 synthesis has been shown to attenuate endothelial dysfunction in experimental and/or human arterial hyper- . formation appears in red. Right panel, quantification, three independent experiments; one-way ANOVA, Bonferroni's multiple comparison test are shown. H and I, aortic EPR spin trapping of NO ⅐ using Fe(DETC) 2 colloid. Freshly prepared aortas of sham-or ATII-treated C57BL/6 mice were stimulated with calcium ionophore A23187 (10 M) for 1 h, and eNOS activity was determined by measurement of NO-Fe(DETC) 2 EPR signals (G). Additionally, aortas were incubated for 19 h at 37°C in the presence of 10 g/ml LPS. Aortic iNOS activity was determined by measurement of Fe(DETC) 2 EPR signals (H). Data are mean Ϯ S.E. of n ϭ 8 animals per group; paired t test is shown. In control experiments, the NO-Fe(DETC) 2 EPR signal was measured in unstimulated murine aorta with a signal intensity that was close to the detection limit of the method. To show specificity of A23187 for eNOS, isolated aortas from eNOS Ϫ/Ϫ mice Ϯ ATII in vivo were incubated in the presence or absence of calcium ionophore A23187 (10 M), resulting in no detectable NO-Fe(DETC) 2 EPR signal (data not shown). n.s., not significant; *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001. tension (30), hyperglycemia (31), diabetes (32), atherosclerosis (33), or cigarette smoking (34). In our hands, depletion of inflammatory cells not only prevents GTPCH overexpression (which can be harmful because it feeds iNOS with BH 4 ) but also increases DHFR expression. This results in a net effect of elevated BH 4 levels in the vessel wall sufficient to recouple eNOS, thereby providing an endogenous source of BH 4 replenishment triggered by depletion of inflammatory cells. In a reversed approach, Gao et al. (35) had shown in a mouse model of eNOS uncoupling that ATII-induced abdominal aortic aneurysm formation hallmarked by vascular macrophage infiltration can be prevented by overexpression of DHFR. Together with our findings, this indicates a mutual repressive effect between DHFR expression and vascular inflammation.
Our findings strengthen the point that inflammatory cells are drivers of both eNOS uncoupling and iNOS-derived nitro-oxidative stress, which lead to the desensitization of the NO/sGC/ cGMP pathway known as endothelial dysfunction. ATII increases the iNOS of inflammatory cells and thereby the demand for BH 4 . Therefore, our findings help to interpret previous studies, in which increased systemic plasma levels of biopterins were positively correlated with C reactive protein as a marker of inflammation in patients with coronary artery disease (36).
It is important to note that depletion of LysM ϩ cells is an effective measure to improve vascular endothelial dysfunction (10). To further address this topic, we inhibited iNOS acutely by 1400W and could show that ATII-induced endothelial dysfunction was strongly attenuated. In addition, iNOS Ϫ/Ϫ mice . formation in aortic tissue (endothelial scan) of sham-treated and ATII-treated monocyte-depleted LysM iDTR transgenic mice, reconstituted either with gp91 phoxϪ/y or Agtr1 Ϫ/Ϫ monocytes or buffer. Upper panel, representative DHE photomicrotopographs. Lower panel, quantification, n ϭ 4 independent experiments; paired t test is shown. C, Western blot analysis of eNOS S-glutathionylation in mouse aorta after eNOS immunoprecipitation, normalized to eNOS expression. Top, densitometry; bottom, representative original blot. Data are mean Ϯ S.E. of protein of aortic lysates sham-treated and ATII-treated monocyte-depleted LysM iDTR and LysM controls from two to three animals/group. One-way ANOVA, Bonferroni's multiple comparison test are shown. were partially protected from eNOS uncoupling, leukocyte-derived oxidative burst, and endothelial dysfunction in ATII-induced arterial hypertension. In earlier reports, iNOS inhibition could block ATII-induced aortic aneurysm formation in SMAD3 Ϫ/Ϫ mice (37) and restore endothelial function in models of vascular inflammation elicited by LPS, TNF␣, or IFN-␥ (38). Our data therefore indicate that iNOS induction sequelled by eNOS uncoupling substantially contributes to the pathophysiology of ATII-induced vascular injury.
In this study, we identify both inflammatory cell-determined iNOS activity and eNOS uncoupling as crucial factors for endothelial dysfunction and vascular inflammation in ATII-induced arterial hypertension. Depletion of inflammatory cells was able to restore iNOS-derived nitro-oxidative stress and to recouple eNOS by readjusting BH 4 synthesis and bioavailability. Targeting iNOS specifically on inflammatory Ly6C high monocytes inside the vascular wall might represent a therapeutic option to attenuate vascular inflammation in hypertension and cardiovascular disease without interfering with overall nitric oxide bioavailability or innate host defense. Identification of integrating pathways that involve adaptive immune cells like T cells might offer a possibility to regulate monocyte/macrophage responses in vascular inflammation (39). This approach is appealing, because . iNOS inhibition abolishes ATII-induced vascular dysfunction. A, EPR spectra for NO-Fe 2ϩ -DETC complex formed from isolated mouse aorta of sham-infused and ATII-treated C57BL/6 mice after 24 h of LPS stimulation (10 g/ml). Trap incubation was performed in the presence of 10 M iNOS inhibitor 1400W. Treatment of aorta with 1400W totally abolished NO-Fe 2ϩ -DETC signal for NO formation, indicating that the signal of LPS-stimulated ATII-treated aortas is specific for iNOS. Left, mean spectra; right, quantification of signal intensity differences. Data are mean Ϯ S.E. of n ϭ 4 animals per group; one-way ANOVA, Bonferroni's post-test are shown. B, relaxation of aortic segments of sham-infused and ATII-treated C57BL/6 mice in response to the endothelium-dependent vasodilator ACh was measured by isometric tension recordings. Two dose-response curves to ACh were performed consecutively. Before the second curve 10 M iNOS inhibitor 1400W was added to aortic segments. Data are mean Ϯ S.E. of n ϭ 7-10 animals per group; one-way ANOVA and Bonferroni's post hoc test for EC 50 and maximal relaxation (see also . formation in aortic tissue (endothelial scan) of sham-treated and ATII-treated iNOS Ϫ/Ϫ mice, incubated with L-NAME, or buffer. Left panel, representative DHE photomicrotopographs. O 2 . formation appears in red; right panel, quantification, three independent experiments are shown; one-way ANOVA, Bonferroni's multiple comparison test are shown. D, systolic blood pressure of sham-treated and ATII-treated C57BL/6 and iNOS Ϫ/Ϫ mice. Data are mean Ϯ S.E. of n ϭ 4 -6 animals per group; one-way ANOVA, Bonferroni's multiple comparison test are shown. E, L-012-derived chemiluminescence in whole blood of sham-treated and ATIItreated C57BL/6 and iNOS Ϫ/Ϫ mice. Data are mean Ϯ S.E. of n ϭ 4 -6 animals per group; one-way ANOVA, Bonferroni's multiple comparison test are shown. F, endothelium-dependent vasorelaxation (ACh concentration-relaxation curve) of sham-treated and ATII-treated C57BL/6 and iNOS Ϫ/Ϫ mice. Data are mean Ϯ S.E. of n ϭ 4 -6 animals per group; one-way ANOVA and Bonferroni's post hoc test for EC 50 and maximal relaxation are shown (see also Table 2). *, p Ͻ 0.05 versus C57BL/6; #, p Ͻ 0.05 versus C57BL/6 ϩ ATII.