Oscillatory Shear Stress Stimulates Endothelial Production of O 2 from p 47 phox-dependent NAD ( P ) H Oxidases , Leading to Monocyte Adhesion *

Arterial regions exposed to oscillatory shear (OS) in branched arteries are lesion-prone sites of atherosclerosis, whereas those of laminar shear (LS) are relatively well protected. Here, we examined the hypothesis that OS and LS differentially regulate production of O2 from the endothelial NAD(P)H oxidase, which, in turn, is responsible for their opposite effects on a critical atherogenic event, monocyte adhesion. We used aortic endothelial cells obtained from C57BL/6 (MAE-C57) and p47 / (MAE-p47 / ) mice, which lack a component of NAD(P)H oxidase. O2 production was determined by dihydroethidium staining and an electron spin resonance using an electron spin trap methoxycarbonyl2,2,5,5-tetramethyl-pyrrolidine. Chronic exposure (18 h) to an arterial level of OS ( 5 dynes/cm) increased O2 (2-fold) and monocyte adhesion (3-fold) in MAE-C57 cells, whereas chronic LS (15 dynes/cm, 18 h) significantly decreased both monocyte adhesion and O2 compared with static conditions. In contrast, neither LS nor OS were able to induce O2 production and monocyte adhesion to MAE-p47 / . Treating MAE-C57 with a cellpermeable superoxide dismutase compound, polyethylene glycol-superoxide dismutase, also inhibited OS-induced monocyte adhesion. In addition, over-expressing p47 in MAE-p47 / restored OS-induced O2 production and monocyte adhesion. These results suggest that chronic exposure of endothelial cells to OS stimulates O2 and/or its derivatives produced from p47-dependent NAD(P)H oxidase, which, in turn, leads to monocyte adhesion, an early and critical atherogenic event.

(2-fold) and monocyte adhesion (3-fold) in MAE-C57 cells, whereas chronic LS (15 dynes/cm 2 , 18 h) significantly decreased both monocyte adhesion and O 2 ؊ compared with static conditions.In contrast, neither LS nor OS were able to induce O 2 ؊ production and monocyte adhesion to MAE-p47 ؊/؊ .Treating MAE-C57 with a cellpermeable superoxide dismutase compound, polyethylene glycol-superoxide dismutase, also inhibited OS-induced monocyte adhesion.In addition, over-expressing p47 phox in MAE-p47 ؊/؊ restored OS-induced O 2 ؊ production and monocyte adhesion.These results suggest that chronic exposure of endothelial cells to OS stimulates O 2 ؊ and/or its derivatives produced from p47 phox -dependent NAD(P)H oxidase, which, in turn, leads to monocyte adhesion, an early and critical atherogenic event.
Fluid shear stress, the frictional force generated by blood flow over the vascular endothelium, is a major factor in atherogenesis.The importance of shear stress in vascular biology and pathophysiology has been highlighted by the focal development patterns of atherosclerosis in hemodynamically defined regions.For example, regions of branched and curved arteries experience disturbed blood flow patterns, including oscillatory shear stress (OS), 1 typically ranging Ϯ 5 dynes/cm 2 (Ϯ indicates changes in flow directions) (1,2).These disturbed shear regions correspond to "lesion-prone" areas that develop early forms of atherosclerotic lesions (1)(2)(3).In contrast, relatively straight arteries, which are exposed to steady uni-directional laminar shear stress (typically ranging from 5-25 dynes/cm 2 ), are usually protected from early atherosclerotic plaque development (1,2).The mechanisms by which laminar shear (LS) acts as atheroprotective force whereas OS initiates or contributes to atherogenesis have been the subject of intense investigation by many researchers.
The vascular endothelium is in direct contact with blood flow and acts as a mechanotransducer by sensing and transducing the changes in local mechanical forces into cellular signals.Endothelial function, shape, physiology, and pathophysiology are greatly regulated by the types (uni-directional laminar or disturbed flow conditions) and magnitudes (high or low) of the shear stress imparted upon them (4,5).
LS exerts atheroprotective effects, including regulation of vascular tone, diameter and vessel wall remodeling, prevention of apoptosis, and monocyte adhesion (4,5).These functions are mediated by regulating the production of several vasoactive factors, including NO, and by the expression of mechanosensitive and atheroprotective genes in endothelial cells (4 -6).In contrast, the exposure of endothelial cells to disturbed flow conditions such as OS has been shown to induce several inflammatory and pro-atherogenic responses, including monocyte adhesion as observed in human arteries, animal models, and cultured endothelial cells (4,7,8,34).The induction of inflammatory adhesion molecules, including E-selectin, vascular cell adhesion molecule-1 (VCAM-1), and intercellular adhesion molecule-1 (ICAM-1) and the consequent monocyte adhesion to the lesion-prone areas seem to be the earliest visible markers leading to atherosclerotic plaque development (3, 7, 8, 34 -36).
Reactive oxygen species (ROS), including superoxide (O 2 Ϫ ), hydrogen peroxide (H 2 O 2 ), and peroxynitrite (ONOO Ϫ ), have been implicated in atherogenesis (9,10).Although the phagocytic NAD(P)H oxidase is the major source of O 2 Ϫ in the circulatory system, vascular endothelial cells, smooth muscle cells, and fibroblasts also express functional leukocyte-type NAD(P)H oxidases (9).The endothelial NAD(P)H oxidase is comprised of five major components, namely gp91 phox (and/or its homologues) and p22 phox in the membrane and p47 phox , p67 phox , and Rac in the cytosol, and it has been suggested to be the major source of O 2 Ϫ in these cells (9).p47 phox in endothelial cells has been shown to play an essential role in activation of NAD(P)H oxidase and the production of O 2 Ϫ (11, 12).Recent results have shown that both LS and OS can induce ROS production from endothelial cells (13,14).De Keulenaer et al. suggested that an NAD(P)H-dependent oxidase was activated by LS and OS (14); however, the molecular identity of this oxidase remains to be determined.Moreover, the identity of its source as well as its pathophysiological roles in cells remain unclear due, in large part, to the lack of specific molecular approaches, thus relying heavily on the use of inhibitors/ scavengers of ROS.For example, chronic exposure to OS has been shown to activate endothelial cells to stimulate monocyte adhesion, which can be blocked by treating with a nonspecific antioxidant, N-acetylcysteine, during shear (8).This result suggests a role for ROS in this inflammatory response.
Here, we hypothesized that the ROS produced by shear stress is O 2 Ϫ derived from NAD(P)H oxidase.We further hypothesized that LS and OS differentially regulate production of O 2 Ϫ , which, in turn, is responsible for regulating monocyte adhesion.To examine these hypotheses, we used aortic endothelial cells (MAE cells) obtained from either normal (C57BL/6) or knockout mice (p47 phoxϪ/Ϫ ) that lack a key component of NAD(P)H oxidases, the p47 phox gene (15).Our results demonstrate that the chronic exposure of endothelial cells to OS stimulates monocyte adhesion by a mechanism dependent on O 2 Ϫ derived from NAD(P)H oxidase, whereas LS induces transient O 2 Ϫ production followed by a substantial inhibition in both O 2 Ϫ production and monocyte adhesion.
Isolation and Culture of Mouse Thoracic Aorta Endothelial Cells-Pathogen-free male C57BL/6 mice were purchased from the Jackson Laboratories, and p47 phoxϪ/Ϫ mice were described previously (15).MAE cells were isolated by modifying a previously described method (16).Briefly, the thoracic aorta was dissected free of adventitia and cut into small pieces (ϳ2 ϫ 2 mm).The explants were cultured intimal side down on Matrigel (BD Biosciences)-coated 6-well plates in an initial growth medium (Dulbecco's modified Eagle's medium containing 20% fetal bovine serum, 100 g/ml endothelial cell growth supplement (Sigma), 10 units/ml heparin, and 1% penicillin/streptomycin).Initially, a very small volume of medium was added onto the outside of the piece of aorta, which was just enough to keep the Matrigel wet but not enough to cover the outer side of the aorta; this inhibited fibroblast growth.One drop of medium was added each day.Two to three days post-isolation, endothelial cells began to grow out of the aortic intima.Aortic pieces were removed from the wells showing capillary-like tubes within 4 -5 days, a characteristic of endothelial cells on Matrigel matrix.Adherent cells were allowed to grow in 2 ml of medium.After ϳ10 days of culture, cells grown out of the explants were trypsinized and seeded on collagencoated 6-well tissue culture plates.
Cell Sorting and Characterization-To isolate pure MAE cells, wells containing cells reaching confluence with apparent cobblestone shapes were loaded with acetylated low density lipoprotein conjugated to 1,1Јdioctadecyl-3,3,3,3Ј-tetramethyl-indocarbocyanine perchlorate (DiI-Ac-LDL) and sorted by fluorescence-activated cell sorting.Briefly, cells were incubated with 10 g/ml DiI-Ac-LDL (Biomedical Technologies) in the initial growth medium for 4 h at 37 °C.The cells were rinsed with the initial growth media, trypsinized, and centrifuged for 5 min at 1,000 rpm.Cell pellets were resuspended into a single cell suspension in fresh medium and sorted by florescence-activated cell sorting (488-nm excitation and 525-nm emission wavelength).Rat vascular smooth muscle cells and bovine and mouse aortic endothelial cells were used as negative and positive controls, respectively.The sorted cells were grown on collagen-coated tissue culture plates and observed by fluorescent microscopy.Only when all of the observed cells were labeled with DiI-Ac-LDL were the cells used for further study.
To further ascertain the purity of endothelial cells, passaged endothelial cells were stained with a polyclonal von Willebrand factor antibody (Dako, 1:100 dilution) for 30 min, followed by rhodamine-conjugated goat anti-rabbit IgG antibody (Molecular Probes, 1:1000 dilution) for 30 min.Subsequently, cells were mounted with an anti-fade (Molecular Probes) and photographed with a Zeiss fluorescent microscope.
Cell Culture and Shear Exposure-MAE cells obtained from C57 (MAE-C57) and p47 phoxϪ/Ϫ (MAE-p47 Ϫ/Ϫ ) mouse aortas were maintained at 37 °C and 5% CO 2 in a growth medium (the same as the initial growth medium except for the heparin concentration, 2.5 units/ml).Cells used in this study were between passages 4 and 10.For shear experiments, cells were plated on 0.5% gelatin-coated 100-mm plates (Falcon).One day prior to shear experiments, the media were changed to endothelial cell growth supplement-free growth media.Monocytes from a human monocytic leukemia cell line, THP-1, were obtained from the American Type Culture Collection and maintained at 37 °C and 5% CO 2 in a growth medium (RPMI 1640; Invitrogen) containing 10% fetal bovine serum, 2 mM L-glutamine, 50 g/ml streptomycin, and 50 units/ml penicillin.Monocytes were maintained at a density of 5 ϫ 10 5 cells/ml.
Confluent MAE monolayers grown in 100-mm tissue culture dishes were exposed to an arterial level of uni-directional laminar shear stress (15 dynes/cm 2 ) in the growth medium by rotating a Teflon cone (0.5°c one angle) as we have described previously (17).To mimic unstable shear conditions in vivo, cells were exposed to oscillatory shear stress with directional changes of flow at 1 Hz cycle (Ϯ 5 dynes/cm 2 ) by rotating the cone back-and-forth using a stepping motor (Servo Motor) controlled by a computer program (DC Motor Company) (7).
Determination of Protein Expression by Western Blot Analysis-Cell lysates (20 g protein each) were resolved by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane (Millipore), and probed with a primary antibody specific to p47 phox (BD Transduction Laboratories).IgG conjugated with alkaline phosphatases (Bio-Rad) were used as secondary antibodies, and the membrane was developed using chemiluminescence detection (18).The intensities of the immunoreactive bands in the Western blots were quantified using the NIH IMAGE program.
Superoxide Measurement by Dihydroethidium-Intracellular O 2 Ϫ production was measured using the cell-permeable dye dihydroethidium (DHE; Molecular Probes), which binds to nuclear DNA when oxidized by O 2 Ϫ and emits red fluorescence (19,20).Briefly, following shear exposure in the presence or absence of antioxidants, cells were immediately incubated with 10 M DHE for 10 min at 37 °C.Cells were then washed in ice-cold fresh DHE-free medium, and the fluorescence intensity was semi-quantified by Bio-Rad laser scanning fluorescence microscopy (514-nm excitation line and 580 Ϯ 30 nm band pass filter) (19).O 2 Ϫ production was estimated by determining the intensity of the fluorescence images taken randomly at a minimum of five different locations of each dish using the NIH IMAGE program (Scion Image Beta 4.02).
Superoxide Measurement by ESR-As a second approach, O 2 Ϫ was measured by ESR spectroscopy using the spin probe, methoxycarbonyl-2,2,5,5-tetramethyl-pyrrolidine (CMH), as described previously (21).Each ESR assay was carried out using cells treated with or without polyethylene glycol-superoxide dismutase (PEG-SOD; 50 units/ml) to quantify the SOD-inhibitable formation of CM ⅐ .Following experimental treatments (shear or static), MAE cells were rinsed with ice-cold Krebs-Hepes buffer (5.6 mM D-glucose, 2.5 mM CaCl 2 , 99 mM NaCl, 4.69 mM KCl, 1.2 mM MgSO 4 , 25 mM NaHCO 3 , 1 mM KH 2 PO 4 , and 20 mM sodium-HEPES; pH 7.4) and scraped from the tissue culture dishes.To inhibit iron-catalyzed reactions, 25 M desferrioxamine and 5 M diethylenetriamine-pentaacetic acid were added to all samples.After centrifugation at 2,000 rpm for 1 min, the cells were resuspended in 200 l of Krebs-Hepes buffer and stored on ice.Cells were counted, and an equal number of cells containing 1 mM CMH was transferred to a 50-l capillary tube to scan in a super high Q microwave cavity of an EMX ESR spectrometer (Bruker BioSpin).The ESR settings were as follows: field sweep, 50 G; microwave frequency, 9.78 GHz; microwave power, 20 milliwatts; modulation amplitude, 2 G; conversion time, 656 ms; time constant, 656 ms; resolution, 512 points; and receiver gain, 1 ϫ 10 5 .The kinetics were recorded by using a 1312-ms conversion time and 5248-ms time constant and monitoring the ESR amplitude of the low field component of the ESR spectrum of 3-methoxycarbonyl-proxyl.
p47 phox Transfection into MAE-p47 Ϫ/Ϫ -MAE-p47 Ϫ/Ϫ cells were grown to 90% confluency and transfected with the full-length p47 phox cDNA inserted in EbopLPP vector (kindly provided by Dr. B M. Babior) or the empty vector alone using LipofectAMINE as we have described previously (7).Following 36 h of incubation, the medium was changed to a fresh growth medium.Transfection efficiency (30 -40%) was estimated by determining green fluorescent protein expression (pEGFP vector, Clontech) by fluorescent microscopy.
Statistical Analysis-For all the experiments, statistical analysis was performed by using the Student's t test.A significance of p Ͻ 0.05 from three or more independent experiments was considered statistically significant.

Characterization of MAE-C57 and MAE-p47
Ϫ/Ϫ Cells-Genotypes of C57BL/6 and p47 phoxϪ/Ϫ mice were confirmed by PCR with genomic DNA obtained from tail tissues using two p47 phox -specific primers and neomycin-resistant primers as described (15).As expected, the wild type mice showed a 650-bp band (Fig. 1A, lane 1), whereas p47 phox knockout mice showed a 900-bp band because it contains the neomycin cassette inserted to generate the knockout (Fig. 1A, lane 2).Fig. 1A, lane 3 was a negative control for PCR containing no DNA.
To assess the purity of wild type MAE and MAE-p47 Ϫ/Ϫ cells isolated from the thoracic aortas of C57BL/6 and p47 phoxϪ/Ϫ mice, cells were labeled with the endothelial marker DiI-Ac-LDL (Fig 1, B and C).The identity and purity of MAE cells were further confirmed by the positive staining of all sorted cells using an additional endothelial cell marker, the von Willebrand factor (Fig. 1C).Immunostaining with antibodies specific to the platelet endothelial cell adhesion molecule (PECAM-1) and endothelial nitric oxide synthase (eNOS) further confirmed endothelial cell purity (data not shown).
To ensure that the p47 phox gene product was indeed absent in MAE-p47 Ϫ/Ϫ cells, cell lysates were analyzed by Western blot using an antibody specific for p47 phox .As expected, MAE-p47 Ϫ/Ϫ cells did not express p47 phox protein in more than four independent preparations from four different mouse thoracic aortas, whereas its expression was clearly identified in more than four independent MAE-C57 cell preparations (Fig. 1E; see also Fig. 9C).
Alignment of MAE Cells by Shear Stress-MAE-C57 and MAE-p47 Ϫ/Ϫ cells grown under static (no flow) culture conditions were not distinguishable from each other and showed typical cobblestone shapes (Fig. 2).Although we do not use cells beyond 10 passages for any of our studies, both cell types could be cultured for at least 10 passages without showing any detectable phenotypic changes.To examine whether these cells respond to shear stress in an expected manner, confluent MAE-C57 and MAE-p47 Ϫ/Ϫ monolayers were exposed to arterial levels of LS or OS by using a cone-and-plate device for 18 h.
Both MAE-C57 and MAE-p47 Ϫ/Ϫ cells aligned to the direction of imposed laminar shear (Fig. 2).Additional comparisons between the two MAE cell types showed that their exposure to LS stimulated phosphorylation of p42/p44 mitogen-activated protein kinases (extracellular signal-regulated kinase 1/2) and Akt (two well known shear-dependent signaling proteins) in a similar time course pattern with similar fold stimulations (data not shown).Moreover, as expected, neither MAE cell type displayed any signs of alignment when exposed to oscillatory shear (Fig. 2).These results suggest that both MAE cells respond in an expected manner to align to the imposed shear conditions and that the p47 phox gene product does not alter this morphological response.
Differential Effects of LS and OS on O 2 Ϫ Production from NAD(P)H Oxidase-Here, we examined whether relatively short (1-2 h) or long-term (18 -24 h) exposure of endothelial cells to LS or OS regulates O 2 Ϫ production from NAD(P)H oxidases.We initially used an immunocytochemical method using DHE as an indicator of intracellular O 2 Ϫ production.Acute exposure of MAE-C57 cells to either LS or OS for a short period (1 h) significantly increased O 2 Ϫ production by 2.5-fold above static controls (Fig. 3).This result is consistent with previous findings (14).To ascertain the specificity of O 2 Ϫ production, MAE-C57 cells were treated with a cell-permeable SOD (PEG-SOD) or a SOD mimetic (manganese-TBAP).Both reagents substantially prevented O 2 Ϫ production in response to shear (Fig. 3).These results suggest that acute exposure of MAE-C57 cells to either LS or OS stimulated O 2 Ϫ production to a similar degree.We next examined whether this ROS was produced from NAD(P)H oxidases.To this end, we exposed MAE-p47 Ϫ/Ϫ cells to LS or OS for 1 h and determined O 2 Ϫ production by DHE staining.In static conditions, MAE-p47 Ϫ/Ϫ cells produced approximately half the amount of O 2 Ϫ compared with that of MAE-C57 cells (Fig. 3).More importantly, neither LS nor OS exposure was able to stimulate O 2 Ϫ production in MAE-p47 Ϫ/Ϫ (Fig. 3).These results clearly suggest that shortterm exposure of MAE-C57 cells to LS or OS stimulate ROS production, most likely O 2 Ϫ , from p47 phox -dependent NAD(P)H oxidases.
To assess shear stress-stimulated production of O 2 Ϫ in a more specific and quantitative manner, we used ESR.CMH was used as an O 2 Ϫ -specific spin trap, and PEG-SOD-inhibitable formation of CM ⅐ was determined.As shown above using the DHE staining assay, acute exposure (1 h) of MAE-C57 cells to LS or OS again stimulated O 2 Ϫ production by ϳ70% over that of static controls (Fig 4B).We also confirmed that MAE-p47 Ϫ/Ϫ cells produced Ͻ50% of O 2 Ϫ in static cells and that both LS and OS failed to acutely stimulate these cells to produce more O 2 Ϫ (Fig. 4B).
Next, we examined the long-term effects on O 2 Ϫ production of exposing endothelial cells to LS or OS.MAE-C57 cells exposed to prolonged OS continued to produce ϳ75% more O 2 Ϫ than static conditions, whereas chronic LS exposure substantially inhibited O 2 Ϫ production by 70% below the static control (Fig. 4C).The chronic effect of OS on O 2 Ϫ production becomes even more impressive if we compare it to that of LS at 18 h.Chronic OS induced almost 6-fold increase in O 2 Ϫ production compared with that of LS (Fig. 4C).O 2 Ϫ production in MAE-p47 Ϫ/Ϫ cells again showed no significant change in response to either LS or OS (Fig. 4C).
We further examined whether p47 phox -dependent NAD(P)H oxidases were responsible for shear-regulated O 2 Ϫ production by using a highly specific ESR method.In this study, MAE-C57 cells were pretreated with apocynin (600 M), which inhibits a NAD(P)H oxidase by blocking the translocation of p47 phox to its membrane for 24 h prior to shear exposure (23), and O 2 Ϫ production was measured by ESR.Consistent with the DHE result shown in Fig. 4, chronic OS exposure increased O 2 Ϫ production by ϳ2-fold, which was virtually prevented by apocynin treatment (Fig. 5).This result further supports the notion that p47 phox -dependent endothelial NAD(P)H oxidases play a critical role in shear-induced O 2 Ϫ production.In addition, these results demonstrate that different types of shear stress determine the duration of O 2 Ϫ produced by endothelial NAD(P)H oxidases.LS exposure induced a transient stimulation of O 2 Ϫ production during the first 1-2 h, followed by a significantly blunted long-term response (18 -24 h).In contrast, OS exposure induced sustained levels of endothelial O 2 Ϫ production for the duration of the shear exposure.The inability of MAE-p47 Ϫ/Ϫ cells to produce O 2 Ϫ in response to shear demonstrates a critical role of p47 phox -dependent NAD(P)H oxidase as the mechanosensitive source of ROS production.FIG. 2. Cell alignment by LS, not by OS.Confluent MAE-C57 and MAE-p47 Ϫ/Ϫ monolayers grown in 100-mm plates were exposed to either static control, LS (15 dynes/cm 2 ), or OS (Ϯ 5 dynes/cm 2 ) using a cone-and-plate system for 24 h.Following shear exposure, cells were photographed under a phase microscope.Shown are typical cell shapes observed in Ͼ50 experiments each.Although shear exposure aligns both MAE-C57 and MAE-p47 Ϫ/Ϫ cells to the direction of laminar flow, this effect was absent in oscillatory sheared MAEC.A typical cobblestone pattern at confluence is shown in the static conditions of MAE-C57 and MAE-p47 Ϫ/Ϫ , respectively.

FIG. 3. Superoxide production by LS or OS can be inhibited by either PEG-SOD, mangnese-TBAP, or p47 phox gene knockout.
Confluent MAE-C57 monolayers were exposed to either laminar (15 dynes/cm 2 ) or oscillatory shear (Ϯ 5 dynes/cm 2 ) for 1 h.Some cells were pre-treated with cell-permeable superoxide dismutase PEG-SOD (300 units/ml, 16 h) or a SOD mimetic manganese-TBAP (100 M, 1 h) before shear onset.To examine whether superoxide was produced from NAD(P)H oxidase, MAE-p47 Ϫ/Ϫ monolayers were exposed to static, LS, or OS conditions as described above.Immediately following shear, cells were incubated with 10 M DHE for 10 min.After changing to fresh buffer, fluorescence images of cells at random positions representing O 2 Ϫ production were taken, and the fluorescence intensity was quantified (mean Ϯ S.E., n ϭ 3 to 4; *, p Ͻ 0.01).

NAD(P)H Oxidase-dependent Induction of Monocyte Adhesion by OS-Thus far our results demonstrate that shear stress stimulates O 2
Ϫ production from NAD(P)H oxidases.To determine whether the ROS produced from NAD(P)H oxidase in response to shear stress plays a critical role in endothelial function, we chose to study monocyte adhesion.First, we examined whether chronic exposure of endothelial cells to LS and OS affected monocyte binding.Chronic exposure of MAE-C57 cells to LS substantially inhibited monocyte binding by 50% compared with static conditions, whereas chronic OS significantly stimulated monocyte adhesion by almost 3-fold (Fig. 6).Treatment of MAE-C57 cells with an NAD(P)H oxidase inhibitor, apocynin (23), prevented OS-induced monocyte adhesion (Fig. 6), supporting a role for NAD(P)H oxidases in OS-induced monocyte binding.In contrast, apocynin did not alter the inhibitory effect of LS on monocyte adhesion, raising a speculation that p47 phox -dependent NAD(P)H oxidases play a role in OS-induced monocyte adhesion.
To further ascertain the role of p47 phox -dependent NAD(P)H oxidases in monocyte binding, we compared shear regulated monocyte adhesion of MAE-C57 to that of MAE-p47 Ϫ/Ϫ cells.In support of our hypothesis, we found that monocyte binding to MAE-p47 Ϫ/Ϫ cells under static conditions was ϳ50% lower than that to MAE-C57 cells (Fig. 7).Furthermore, unlike MAE-C57 cells, chronic exposure (18 h) of MAE-p47 Ϫ/Ϫ to LS or OS did not significantly affect monocyte adhesion (Fig. 7).This result is consistent with the apocynin effects shown in Fig. 5 and confirms the critical role of p47 phox -dependent NAD(P)H oxidase in shear-induced monocyte adhesion.It is interesting to note that O 2 Ϫ production in the MAE-p47 Ϫ/Ϫ cells is decreased by ϳ50% as well (Fig. 4).
Next, we examined the time dependence of LS and OS effects on monocyte adhesion.Exposure of MAE-C57 cells to 4 h of LS was enough to inhibit monocyte adhesion (Fig. 7).However, 4 h of OS failed to stimulate monocyte adhesion in MAE-C57 cells (Fig. 7).As expected, LS or OS exposure of MAE-p47 Ϫ/Ϫ for 4 h had no significant effect on monocyte adhesion (Fig. 7).These findings indicate that OS-induced monocyte adhesion requires a chronic shear exposure longer than 18 h, whereas 4 h of LS exposure is enough to cause inhibition of monocyte adhesion as reported perviously (22).
O 2 Ϫ Produced from p47 phox -dependent NAD(P)H Oxidases Is Responsible for OS-induced Monocyte Adhesion to Endothelial Cells-To date, we have shown that OS induces monocyte adhesion in MAE-C57 cells.To confirm that these effects are a consequence of altering the superoxide level, we examined whether monocyte binding to MAE cells can be blocked by either treating cells with cell-permeable SOD or transfecting them with p47 phox cDNA.First, we treated MAE-C57 cells with a cell-permeable PEG-SOD to determine whether the removal of intracellular O 2 Ϫ would inhibit the OS-induced monocyte binding response.Treatment of MAE-C57 cells with PEG-SOD did indeed block monocyte adhesion to static control levels (Fig. 8).Immediately following shear, cells were scraped and centrifuged.An equal number of cells was incubated with a superoxide-specific spin trap CMH in the presence or absence of PEG-SOD (50 units/ml).Shown in panel A are representative superoxide spectra.For quantification, a time scan of CPH oxidation was obtained, and the slope was used as a measure of superoxide production.The bar graphs shown in panels C and D were calculated by subtracting SOD-inhibitable portions from total superoxide production (mean Ϯ S.E., n ϭ 3-4 each; #, p Ͻ 0.05; *, p Ͻ 0.01).Note that acute (1 h) exposure of MAE-C57 cells to either LS or OS increased the production of O 2 Ϫ to a similar degree.In contrast, chronic exposure of MAE-C57 to LS and OS inhibited and increased, respectively, O 2 Ϫ production.Unlike MAE-C57 cells, O 2 Ϫ production in MAE-p47 Ϫ/Ϫ was substantially inhibited in static conditions, and neither LS nor OS stimulated O 2 Ϫ production above the control level.

FIG. 5. Apocynin, an NAD(P)H oxidase inhibitor, blocks OS stimulated O 2
؊ production in MAE-C57 cells.Confluent MAE-C57 monolayers were exposed to 18 h of OS or static conditions as in Fig. 4 in the presence or absence of the NAD(P)H oxidase inhibitor apocynin (600 M).Following shear or static treatment, O 2 Ϫ production in MAEcells was measured by employing ESR in a manner described in the Fig. 4 legend.Note that chronic OS stimulated O 2 Ϫ production, which was inhibited by apocynin treatment.
Next, MAE-p47 Ϫ/Ϫ cells were transfected with the cDNA vector encoding p47 phox protein or an empty vector control.First, we verified that p47 phox transfection indeed resulted in over-expression of the 47-kDa protein by Western blot analysis by using a p47 phox -specific antibody (Fig 9A).Next, we found that the MAE-p47 Ϫ/Ϫ cells transfected with p47 phox responded to OS by producing O 2 Ϫ by Ͼ2-fold above the static cells (Fig. 9B).Finally, as shown in Fig. 9C, we were able to restore monocyte adhesion in response to chronic OS by over-expressing p47 phox protein in MAE-p47 Ϫ/Ϫ cells.In contrast, OS exposure of MAE-p47 Ϫ/Ϫ cells transfected with the empty vector did not increase monocyte adhesion nearly as much as that of p47 phox -transfected cells.The OS-induced monocyte adhesion was completely inhibited by the NAD(P)H oxidase inhibitor, apocynin (Fig. 9C), providing further evidence for the specific of effect of p47 phox .

DISCUSSION
The most significant and interesting findings of the current study are that chronic exposure of endothelial cells to OS stimulates O 2 Ϫ production and monocyte adhesion by p47 phoxdependent, NAD(P)H oxidase-derived mechanisms requiring O 2 Ϫ or its ROS derivatives.Previous studies have shown that shear stress stimulates production of ROS and reactive nitrogen species (13,14).Our current results strongly suggest that shear stress induces O 2 Ϫ production from endothelial cells.Several lines of evidence presented below support this conclusion.First, treatment of MAE-C57 cells with a cell-permeable SOD (PEG-SOD) and a SOD mimetic (manganese-TBAP) inhibited O 2 Ϫ produced by shear stress.Second, O 2 Ϫ production was determined by using two independent methods, DHE staining and ESR using CMH as a spin probe.In addition, ESR measurements were carried out in the presence or absence of PEG-SOD to determine the SOD-inhibitable portion of O 2 Ϫ production.This second approach provided clear specificity for O 2 Ϫ measurement.
ROS can be produced in the vasculature from several sources, including membrane and mitochondrial NAD(P)H oxidases, cytochrome P450, xanthine oxidase, and uncoupled nitric oxide synthase (9).Overwhelming evidence now suggests that non-mitochondrial NAD(P)H oxidases are a major source of ROS in vascular smooth muscle cells and endothelial cells (9).The endothelial NAD(P)H oxidase is believed to be similar to the leukocyte NAD(P)H oxidase, and several of its components have been identified, including p47 phox , p67 phox , Rac, p22 phox , and gp91 phox (9,24).Recently, several members of gp91 phox homologues (referred to as the "Nox" family) were identified (25,26).In endothelial cells, it appears that Nox 4 is the most abundant member, whereas Nox 1 and gp91 phox (currently designated as Nox 2) have also been detected at low levels (27).
Although it has been suggested that NAD(P)H-dependent oxidases are activated by shear stress ( 14), the extent of its involvement in shear-induced O 2 Ϫ production was not clear.Our current results strongly indicate that the major source of

endothelial O 2
Ϫ produced in response to shear stress is p47 phoxdependent endothelial NAD(P)H oxidase.This conclusion is based on the following observations.First, treatment of MAE-C57 cells with apocynin inhibited shear-dependent O 2 Ϫ production, suggesting a role for NAD(P)H oxidases activated by the translocation of cytosolic subunits.Second, MAE-p47 Ϫ/Ϫ cells failed to produce O 2 Ϫ in response to either chronic or acute LS and OS exposure.Third, over-expression of wild type p47 phox protein in MAE-p47 Ϫ/Ϫ cells restored OS-induced production of O 2 Ϫ , further demonstrating the critical role of p47 phox -dependent NAD(P)H oxidases as a major source of mechanosensitive ROS production.Although the role of a p47 phox -dependent NAD(P)H oxidase is clear, it is not known which Nox (Nox 1, 2 and 4) is responsible for shear-dependent O 2 Ϫ production in endothelial cells.
We found that O 2 Ϫ production in response to LS was transiently regulated, whereas OS induced continuous O 2 Ϫ production at higher (4-fold) levels compared with LS.Short-term exposure (1-2 h) to LS and OS resulted in similar levels of O 2 Ϫ production.The mechanisms responsible for this differential shear type-dependent and time-dependent O 2 Ϫ production are not clear at this point.Several known mechanisms, however, could explain this.First, consistent to previous findings (28 -30), we have found that chronic LS stimulated a significantly higher level of NO production in comparison to OS (data not shown).NO is a potent antioxidant and binds O 2 Ϫ , which would decrease its detectable levels (21).The shift in balance from "high NO and low O 2 Ϫ " in LS-conditioned cells to "low NO and high O 2 Ϫ " by OS may be an important factor contributing to a pro-atherogenic response such as monocyte adhesion.Second, LS and OS could differentially regulate expression of other antioxidant enzymes in a time-dependent manner.For example, LS but not turbulent or OS has been shown to induce numerous antioxidant enzymes such as copper/zinc-SOD (14,31), heme oxygenase-1 ( 14), manganese-SOD (32), and glutathione peroxidase (33).Third, it is possible that LS and OS may differentially regulate activity and expression levels of NAD(P)H oxidases.Consistent with this possibility, atherosclerotic lesions, which typically occur in arterial regions exposed to disturbed shear conditions, contain higher levels of ROS and NAD(P)H oxidase components, including Nox2, Nox4, and p22 phox (27).Taken together, we propose that the balance between O 2 Ϫ production and antioxidants is determined by the duration and type of shear imposed on endothelial cells.In the short term, other than endothelial nitric oxide synthase, the expression and activation of other anti-oxidant enzymes may not be sufficient to counter the O 2 Ϫ produced in response to LS and OS.On the other hand, chronic exposure to LS induces enough antioxidants to blunt O 2 Ϫ production, whereas OS may either decrease or have no effect on antioxidant levels.
We found that while prolonged (18 h) exposure of endothelial to OS induced monocyte adhesion, acute (4 h) OS did not (Fig. 7), raising a possibility that OS-dependent monocyte adhesion requires new gene synthesis.These results are consistent with the recent finding that chronic OS exposure of endothelial cells stimulates bone morphogenic protein-4 expression, which then leads to monocyte adhesion by inducing intercellular adhesion molecule-1 in a NFB-dependent manner (7).In addition, OS has been shown to stimulate monocyte adhesion in an N-acetylcysteine-inhibitable manner (8).Although this finding suggested a role for ROS in OS-induced monocyte adhesion, its identity and source were not known.Based on our current study, it is now clear that the shear-induced O 2 Ϫ and its derivatives being produced in endothelial cells are derived from NAD(P)H oxidases.Furthermore, we conclude that O 2 Ϫ and its derivatives produced from NAD(P)H oxidase are responsible for monocyte adhesion induced by OS.
It is not clear at this point, however, whether O 2 Ϫ and/or its derivatives (e.g.H 2 O 2 and ONOO Ϫ ) are responsible for the OS-dependent monocyte adhesion.Our results suggest that OS-induced monocyte adhesion may occur through multiple pathways acting in concert, one of which is O 2 Ϫ -dependent.Treatment of cells with PEG-SOD resulted in a significant inhibition of OS-mediated monocyte adhesion in MAE-C57 cells.
Atherosclerosis is a focal, inflammatory disease preferentially occurring at lesion-prone areas exposed to unstable shear stress (3).Although LS has been shown to induce anti-inflammatory, atheroprotective responses, OS has been associated with pro-inflammatory and pro-atherogenic responses (4,5).Whereas these previous studies (8) have suggested a connection between OS, ROS, and monocyte adhesion, the identity and source of ROS and its functional significance have not been clearly demonstrated.Using MAE cells obtained from C57 and p47 Ϫ/Ϫ mice and employing ESR spectroscopy and monocyte FIG. 9. Chronic OS induces monocyte adhesion to MAE-p47 ؊/؊ cells transfected with p47 phox vector.MAE-p47 Ϫ/Ϫ cells were transfected with p47 phox vector or empty vector control (Vector) by using LipofectAMINE.A, 1 day after transfection, cell lysates were obtained and analyzed by Western blot using a p47-specific antibody.Cell lysates of MAE-C57 were used as controls.Shown is a representative of 3-6 different studies.B and C, 1 day after transfection, cells were exposed to OS or static control for 18 h.In some studies, apocynin (600 M) was added during shear.Then, O 2 Ϫ production was measured by ESR (B) as described in the Fig. 4 legend or monocyte adhesion was determined as described above (C).The bar graphs shown in panels B and C are the means Ϯ S.E., n ϭ 6; *, p Ͻ 0.05.adhesion assays, here we have determined the identity and source of ROS and its direct role in monocyte adhesion in response to shear stress.In summary, we have identified O 2 Ϫ as being produced from endothelial NAD(P)H oxidases in response to shear stress and have shown that this ROS and its derivatives are directly responsible for monocyte adhesion, a critical step in atherogenesis.

FIG. 1 .
FIG. 1. Isolation and characterization of MAE-C57 and MAE-p47؊/؊ cells.A, C57BL/6 (C57) and p47 Ϫ/Ϫ mice were genotyped by PCR using genomic DNA obtained from tail tissues.As a negative control, PCR was carried out in the absence of genomic DNA (lane 3).B and C, primary cultures obtained from thoracic aortas of C57Bl/6 and p47 Ϫ/Ϫ mice were loaded with the endothelial marker DiI-Ac-LDL and sorted by a fluorescence activated cell sorter.Shown are fluorescence micrographs (10ϫ magnification) of DiI-Ac-LDL-positive cells 2-3 h after plating on collagen-coated dishes (B) and cells photographed under a phase microscope (C).All cells shown are DiI-Ac-LDL-positive.D, some cells were stained with an endothelial cell marker, the von Willebrand factor (vWF), and a fluorescence-tagged secondary antibody and observed by fluorescence microscopy.E, cell lysates (20 g each) obtained from wild type MAE and MAE-p47 Ϫ/Ϫ were analyzed by immunoblot (IB) using a p47-specific antibody and actin (as a loading control).The middle blank lane contains molecular weight (MW) markers.Note that p47 phox is expressed only in MAE-C57 cells but not in MAE-p47 Ϫ/Ϫ cells.Shown are typical results of Ͼ10 separate isolation studies.

FIG. 4 .
FIG.4.Superoxide production by LS or OS determined by ESR spectroscopy in MAE-C57 and MAE-p47 ؊/؊ .Confluent MAE-C57 monolayers were exposed to laminar shear (15 dynes/cm2), oscillatory shear (Ϯ 5 dynes/cm2), or static conditions for 1 h (A and B) or 18 h (C).Immediately following shear, cells were scraped and centrifuged.An equal number of cells was incubated with a superoxide-specific spin trap CMH in the presence or absence of PEG-SOD (50 units/ml).Shown in panel A are representative superoxide spectra.For quantification, a time scan of CPH oxidation was obtained, and the slope was used as a measure of superoxide production.The bar graphs shown in panels C and D were calculated by subtracting SOD-inhibitable portions from total superoxide production (mean Ϯ S.E., n ϭ 3-4 each; #, p Ͻ 0.05; *, p Ͻ 0.01).Note that acute (1 h) exposure of MAE-C57 cells to either LS or OS increased the production of O 2 Ϫ to a similar degree.In contrast, chronic exposure of MAE-C57 to LS and OS inhibited and increased, respectively, O 2 Ϫ production.Unlike MAE-C57 cells, O 2 Ϫ production in MAE-p47 Ϫ/Ϫ was substantially inhibited in static conditions, and neither LS nor OS stimulated O 2 Ϫ production above the control level.

FIG. 6 .FIG. 8 .
FIG.6.Chronic exposure of MAE-C57 to LS and OS inhibits and stimulates monocyte adhesion in an apocynin-dependent manner.Confluent MAE-C57 monolayers were exposed to 18 h of LS, OS, or static conditions as in Fig.4in the presence or absence of a NAD(P)H oxidase inhibitor apocynin (600 M).Following shear or static treatment, MAE cells were washed and incubated with fluorescently labeled THP-1 monocytes.After 30 min, unbound cells were washed away, the bound monocytes were lysed, and the fluorescence intensity was determined by fluorescence spectroscopy.The result shown represents mean Ϯ S.E.(*, p Ͻ 0.05; n ϭ 6).Note that chronic LS inhibited and OS stimulated monocyte binding, which was blocked by apocynin treatment.