Role of endothelial nitric oxide synthase in endothelial activation: insights from eNOS knockout endothelial cells

Peter J. Kuhlencordt, Eva Rosel, Robert E. Gerszten, Manuel Morales-Ruiz, David Dombkowski, William J. Atkinson, Fred Han, Frederic Preffer, Anthony Rosenzweig, William C. Sessa, Michael A. Gimbrone Jr., Georg Ertl, Paul L. Huang


The objective of this study was to determine whether absence of endothelial nitric oxide synthase (eNOS) affects the expression of cell surface adhesion molecules in endothelial cells. Murine lung endothelial cells (MLECs) were prepared by immunomagnetic bead selection from wild-type and eNOS knockout mice. Wild-type cells expressed eNOS, but eNOS knockout cells did not. Expression of neuronal NOS and inducible NOS was not detectable in cells of either genotype. Upon stimulation, confluent wild-type MLECs produced significant amounts of NO compared with Nω-monomethyl-l-arginine-treated wild-type cells. eNOS knockout and wild-type cells showed no difference in the expression of E-selectin, P-selectin, intracellular adhesion molecule-1, and vascular cell adhesion molecule-1 as measured by flow cytometry on the surface of platelet endothelial cell adhesion molecule-1 (PECAM-1/CD31)-positive cells. Both eNOS knockout and wild-type cells displayed the characteristics of resting endothelium. Adhesion studies in a parallel plate laminar flow chamber showed no difference in leukocyte-endothelial cell interactions between the two genotypes. Cytokine treatment induced endothelial cell adhesion molecule expression and increased leukocyte-endothelial cell interactions in both genotypes. We conclude that in resting murine endothelial cells, absence of endothelial production of NO by itself does not initiate endothelial cell activation or promote leukocyte-endothelial cell interactions. We propose that eNOS derived NO does not chronically suppress endothelial cell activation in an autocrine fashion but serves to counterbalance signals that mediate activation.

  • vascular biology
  • atherosclerosis
  • mouse models

the changes in the morphology and gene expression pattern of endothelial cells at sites of inflammation and cell-mediated immune responses have been described as the process of endothelial cell activation (11, 12, 33). The activated state is characterized by an increase in the surface expression of adhesion molecules, which regulate leukocyte-endothelial cell interactions. Nitric oxide (NO) is a homeostatic regulator of vascular tone, and pharmacological inhibition of NO synthesis or disruption of the endothelial NO synthase (eNOS) gene significantly increases blood pressure (15, 27). Moreover, NO inhibits platelet aggregation in vitro and modulates leukocyte adhesion in the microcirculation (17, 27).

The mechanisms by which NO modulates leukocyte-endothelial cell interactions are not known. One possibility is a direct effect of NO on the regulation of expression of adhesion molecules and cytokines by the transcription factor NF-κB (39). NO induces transcription of IκBα, an inhibitor of NF-κB, thus stabilizing the inhibitory NF-κB/IκBα complex in the cytosol (35). A second possibility is that NO may protect cells from oxidative stress because it interacts rapidly with superoxide, which acts as a proadhesive molecule (3, 7, 9, 36, 37). Hence, NO may indirectly influence leukocyte-endothelial cell interactions by counterbalancing oxygen free radicals, the latter being the actual effector molecules that initiate vascular pathology.

NO donors and NOS inhibitors have been employed to study the importance of NO in regulation of endothelial gene expression in vitro. NO donors, which release NO independently of NOS enzymes, decrease cytokine-induced endothelial cell activation in human umbilical vein endothelial cells (HUVECs) (5). NO also modulates NF-κB activation by IκBα. These effects require the expression of inducible NOS (iNOS) in mononuclear cells (39), or high concentrations of NO donors not likely to be produced by eNOS under physiological conditions (35).

In vivo, leukocyte-endothelial cell interactions have been studied in splanchnic ischemia-reperfusion with the use of intravital microscopy. Superfusion with NO donors reduces neutrophil-endothelial cell interactions, possibly by acutely scavenging oxygen radicals (10). NOS inhibitors increase leukocyte-endothelial cell interactions due to degranulation of mast cells and increased superoxide production (16). In addition, intravital microscopy of the mesenteric circulation of eNOS knockout mice has shown enhanced leukocyte adhesion to the vascular endothelium associated with increased surface expression of P-selectin (10, 22). Cultured endothelial cells exposed to ischemia-reperfusion rapidly express P-selectin, which leads to neutrophil adherence (32, 34). This condition is reversible by administration of superoxide dismutase, a scavenger of superoxide radicals (13, 40). Thus the increased adhesion seen in eNOS knockout mice could be caused by increased superoxide production not counterbalanced by eNOS-dependent production of the superoxide scavenger NO. In a different study by Sanz et al. (38), also employing in vivo microscopy, baseline adhesion of leucocytes to the microvasculature of eNOS knockout animals was unchanged. These investigators found that neuronal NOS (nNOS) compensates for the loss of eNOS at baseline. However, leukocyte recruitment elicited by oxidative stress was more pronounced in eNOS knockout than in wild-type animals, suggesting that nNOS did not completely compensate for eNOS deficiency.

Reduction of endothelium-derived, eNOS-dependent NO production has been reported after ischemia-reperfusion injury, in sepsis, in hyperlipidemia, and in atherosclerosis (8, 19, 20, 2325, 31). We previously showed that genetic deficiency of eNOS increases intimal proliferation in response to vessel injury and increases atherosclerosis in hypercholesterolemic apolipoprotein E (apoE) knockout mice (2, 18, 29). To date, ample evidence suggests that NO modulates leukocyte-endothelial cell interactions and vascular pathology. However, uncertainty remains as to whether NO is required to continuously suppress endothelial cell activation. In this study, we applied a reliable novel cell culture technique to isolate endothelial cells from eNOS knockout mice to study the effects of complete NO deprivation on endothelial homeostasis. Using this strategy, we have avoided problems with pharmacological inhibition of NOS and interference with other cell types “in vivo.” Our major finding is that endothelial activation does not result from eNOS deficiency in and of itself, raising the possibility that eNOS-derived NO counterbalances signals mediating endothelial activation.


All procedures performed conform with Massachusetts General Hospital policies and the National Institutes of Health guidelines for the care and use of laboratory animals.

Materials. Flow cytometry antibodies biotin-conjugated CD31, clone MEC 13.3; R-PE-CD54, clone 3E2; FITC-CD62P, clone RB40.34; FITC-CD106, clone 429; and R-PE-CD62E, clone 10E9.6 were from Pharmingen. Streptavidin PerCP was from Becton Dickinson. Amplex red was from Molecular Probes.

Generation of eNOS knockout mice. eNOS-deficient mice were generated by targeted deletion in our laboratory (15). The mice were backcrossed for 10 generations to C57BL6, and the latter strain served as the wild-type control.

Cell culture. Microvascular endothelial cells were isolated from lungs of animals 3–4 mo old. For each experiment, primary cultures of both genotypes were started simultaneously. Animals were killed by cervical dislocation, and lungs were collected in ice-cold Dulbecco's modified Eagle's medium (DMEM). Peripheral lung tissue was minced and digested for 1 h at 37°C in 0.1% collagenase-A (Boehringer Mannheim). The digest was passed through a blunt 14-gauge needle and filtered through a 130-μm steel mesh. Cells were pelleted at 300 g and resuspended in murine lung endothelial cell (MLEC) medium (37°C) containing 20% FBS, 35% DMEM, 35% F-12, 50 μg/ml endothelial mitogen (Biomedical Technologies), 2 mM l-glutamine, 100 μg/ml heparin, and 100 U/100 μg/ml penicillin-streptomycin and plated in 0.1% gelatin-coated T75 flasks. Cells were washed after 24 h and cultured for 2–4 days. Magnetic beads were coated with anti-mouse CD102 (clone 3C4; Pharmingen) antibody (5 μg/4×106 beads, Dynabeads M-450; Dynal). Per flask, 4×106 beads were added and incubated for 1 h at 4°C. Cells were trypsinized and selected in a magnetic field for 10 min. Cultures were grown to confluence and selected twice before being plated for experiments. By following this procedure, cells used in the experiments were, on average, 10 days in culture.

Acetylated LDL labeling. Labeling was done according to the manufacturer's protocol.

Measurement of NO release. We analyzed the release of NO using chemiluminescent detection for nitrite (Math) according to a previously published protocol (28). Net NO per microgram of protein was calculated after subtracting background levels of NO found in the media.

Measurement of NO release after exposure to fluid flow. Confluent monolayers of MLECs were placed in serum-free medium for 1 h and then exposed to static conditions or flow. As previously reported, flow was induced by placing a confluent 60-mm culture dish on a mixing table and rotating at 120 rpm for 8 and 12 h (42). Cells were kept at 37°C, 5% CO2 in a cell culture incubator throughout the procedure.

Flow cytometric analysis. Cells were trypsinized (37°C) and washed with PBS-1% BSA, pelleted, and resuspended at 106 cells/100 μl. Primary monoclonal antibodies were incubated in the dark at 1 μg/106 cells for 15 min on ice. Cells were washed, pelleted, resuspended, and stained with streptavidin PerCP for 15 min at 4°C. Cells were washed with PBS-1% BSA, resuspended in PBS-1% paraformaldehyde, and kept at 4°C in the dark until analyzed. Unstained cells, cells stained with isotype-matched antibodies, and second-step reagent alone served as controls. The fluorescence of 5 × 104 cells was measured on a FACScan (Becton Dickinson).

Detection of cell viability. Cells were washed in PBS and resuspended in 100 μl of 20 μg/ml propidium iodide (PI) in PBS. Cells were incubated for 15 min at room temperature. The cells were washed, and 5×104 cells were analyzed on a FACScan.

Cell cycle analysis. Cells were fixed overnight in 70% ethanol. Fixed cells were washed in PBS, pelleted, resuspended in 20 μg/ml PI-PBS containing 1 mg/ml RNase (type IIA; Sigma), and subjected to analysis on a FACScan.

Leukocyte isolation. Human monocytes were purified from healthy human donors by centrifugation on Ficoll-Hypaque density gradient at 15°C (LSM; Organon Teknika, Durham, NC) followed by magnetic bead purification (Miltenyi Biotech). Monocyte suspensions were >92% pure as determined by light scatter and cell surface antigen analysis.

Adhesion assay under flow. MLECs were plated at confluence in 0.5-cm2 chambers immobilized with agarose on plastic tissue culture slides and cultured for 48 h. The monolayers were placed in a parallel plate laminar flow chamber (Immunetics, Cambridge, MA), and leukocytes were perfused across the endothelial surface over a range of laminar shear stress from 0.5 to 2.0 dyn/cm2. The experiments were performed on cells sheared roughly 30 min before experiments began and within 90 min after the cells were introduced into the flow chamber. A video recorder equipped with a time-date generator and a millisecond clock was used to count firmly adherent leukocytes in three to five randomly chosen high-power fields.

Western blotting. Electrophoresis and immunoblotting were performed as previously described (15) by using Transduction Laboratory antibodies for nNOS (polyclonal), iNOS (clone 2), and eNOS (clone 3).

RT-PCR. RT-PCR was performed according to previously published protocols (14). GAPDH was amplified under the same conditions and with primer 1, 5′-GGGGAGCCAAAAGGGTCATCATCT-3′, and primer 2, 5′-GACGCCTGCTTCACCACCTTCTTG-3′.

P-selectin translocation. The rapid translocation of P-selectin to the endothelial cell surface was measured discriminately by using a modified ELISA with nonpermeable monolayers as previously described (41). Menadione and thrombin (Sigma) solutions were freshly prepared from stock solutions on the day of the experiment.

Measurement of reactive oxygen species generation. N-acetyl-3,7-dihydroxyphenoxazine (Amplex red) was used as an indicator reagent as previously described (41). Amplex red reacts with intracellular H2O2 to form an adduct that can be detected at 530-nm excitation and 590-nm emission.

Statistical analysis. Statistical analysis was performed by using StatView 4.51 (Abacus Concepts, Berkley, CA). Two-way ANOVA was used for repeated measures, followed by Scheffé's F-test. A probability value of P < 0.05 was considered significant.


Morphology and growth characteristics of MLECs. After two rounds of magnetic selection, CD102-positive cells from eNOS knockout and wild-type mice exhibited similar morphological characteristics. Both genotypes form a contact-inhibited monolayer. Virtually all cells demonstrate uptake of Dil-Ac-LDL, and typically >95% stain positive for platelet endothelial cell adhesion molecule (PECAM)-1 by flow cytometry (Fig. 1). Cell viability of unfixed trypsinized cells collected from confluent monolayers was assessed by PI staining. PI-positive cells, which were considered dead, comprised 5.78 ± 1.3% (n = 4) for eNOS wild-type and 7.69 ± 5.2% (n = 2) for eNOS knockout cells. Wild-type cells, but not eNOS knockout cells, demonstrated strong immunoreactivity for eNOS protein (Fig. 2). In three independent cultures of wild-type and eNOS knockout MLECs, nNOS expression was not detectable by Western blot. iNOS expression was not detected at baseline or after TNF-α stimulation for 4 or 6 h. Additionally, absence of iNOS and nNOS transcripts was confirmed by RT-PCR (data not shown). With the use of a bovine pituitary gland-derived endothelial mitogen, the eNOS knockout and wild-type cells demonstrated virtually identical growth characteristics. The distribution of near-confluent cells in the G1, S, and G2 + M phase of the cell cycle was unchanged between the two genotypes (Table 1).

Fig. 1.

Murine lung endothelial cells (MLECs). Top: phase-contrast images of C57BL6 wild-type (WT) and eNOS knockout (KO) cells. Bottom: Dil-Ac-LDL-labeled cells under rhodamine excitation and emission. The cobblestone pattern and Dil-Ac-LDL uptake are characteristic of endothelial cells.

Fig. 2.

Western blot analysis of samples from C57BL6 WT and eNOS KO MLECs. Endothelial nitric oxide synthase (eNOS) protein was present in primary cultures of WT cells and absent in eNOS KO cells (140 kDa). Neuronal NOS (nNOS; 155 kDa) was present in pituitary gland control lysates (ctr) but not in WT or KO MLECs. No inducible NOS (iNOS; 130 kDa) expression was present in either genotype at baseline or following stimulation with TNF-α (10 ng/ml) for 4 and 6 h. A cell lysate from activated macrophages served as a positive control.

View this table:
Table 1.

Cell cycle characteristics

Measurement of NO release. NO release was measured from six wild-type MLEC monolayers. NO release from resting endothelium was not significantly different from measured background values or Nω-monomethyl-l-arginine (l-NMMA)-treated cells. After stimulation with ionomycin (Fig. 3A; P < 0.0001), VEGF (P = 0.02), or exposure to shear stress (Fig. 3B; P = 0.079), wild-type MLECs produced significantly higher amounts of NO than unstimulated wild-type cells. Coincubation of ionomycin with l-NMMA (1 mM) or VEGF plus l-NMMA completely abrogated this effect (Fig. 3A), and shear exposure did not increase NO production in eNOS knockout MLECs (Fig. 3B). Western blot analysis shows that eNOS expression is upregulated after 12 h of shear exposure (Fig. 3C).

Fig. 3.

Measurement of NO release in WT and KO MLECs. A: stimulation with ionomycin (3 μM) and VEGF (40 ng/ml) increased NO production in WT cells (P < 0.0001 and P = 0.002, respectively) compared with unstimulated WT MLECs. Coincubation with Nω-monomethyl-l-arginine (l-NMMA; 1 mM) completely inhibited the increase in NO production in ionomycin- and VEGF-treated cells (P < 0.0001 and P = 0.002, respectively). B: NO release from WT cells under static culture conditions and from WT and KO cells after 12 h of laminar shear stress (LSS). Shear stress increased NO production in WT cells (WT+) compared with WT cells under static culture conditions (WT-) (*P = 0.008) and KO cells exposed to shear stress (†P = 0.02). C: eNOS protein expression was increased in WT cells after 8 and 12 h of fluid flow.

Surface expression of resting endothelium. The expressions of intercellular adhesion molecule (ICAM)-1 (CD54), P-selectin (CD62P), E-selectin (CD62E), and vascular cell adhesion molecule (VCAM)-1 (CD106) were measured on the cell surface of resting, confluent, CD31-positive cells (Fig. 4, solid lines). The wild-type and eNOS knockout cells displayed similar high levels of ICAM-1 and low levels of P-selectin, E-selectin, and VCAM-1. The data from five separate primary cultures showed no significant difference in the percentage of positive-staining cells or the relative fluorescence intensity of any of the markers (Fig. 4, A and B). To test whether acute pharmacological inhibition of NOS in wild-type endothelial cells changes their activation state, we treated wild-type and eNOS knockout cells with Nω-nitro-l-arginine methyl ester (l-NAME; 1 mM) and its biologically inactive isomer, Nω-nitro-d-arginine methyl ester (d-NAME; 1 mM) for 24 h. There was no significant difference in surface expression among l-NAME, d-NAME, and untreated wild-type and eNOS knockout cells (data not shown).

Fig. 4.

Flow cytometry analysis. A: representative flow cytometric analysis of cell surface expression of intercellular adhesion molecule (ICAM)-1, P-selectin, E-selectin, and vascular cell adhesion molecule (VCAM)-1 on unstimulated WT and KO MLECs (solid lines). Shaded lines depict fluorochrome intensity for each marker after TNF-α/IL-1α (10 ng/ml each) stimulation for 12 h. The increase in fluorochrome intensity following cytokine treatment is shown as a rightward shift on the x-axis. B: percentages of CD31-positive cells that stain for E-selectin, P-selectin, VCAM-1, and ICAM-1 under standard culture conditions. Data represent means ± SE from 5 primary cultures of WT and KO cells. See text for detailed description of markers.

Surface expression of TNF-α-activated endothelium. TNF-α as well as TNF-α/IL-1α costimulation was used to document that MLECs can be activated and to examine the effects of eNOS deficiency on cytokine-stimulated adhesion molecule expression. Confluent monolayers were incubated with murine recombinant TNF-α for 12 h or with TNF-α/IL-1α for 12 h (10 ng/ml; Genzyme). Cell viability of unfixed, TNF-α-treated cells was assessed by PI staining. Analysis of PI-positive cells revealed a fraction of 9.46 ± 4.5% (n = 4) and 8.18 ± 2.7% (n = 2) of dead cells for wild-type and eNOS knockout cells, respectively. Data from three separate cultures showed that cytokine stimulation results in a marked increase in surface expression of CD54, CD62P, and CD106, but no change in CD31 expression, compared with unstimulated cells (Fig. 4A; shaded line). The increase in CD62E expression was less robust, but consistently observed, in three cultures. There was no significant difference in the relative fluorescence intensity (Fig. 4A) or the percentage of positive-staining cells (Fig. 4B) for any of the markers between the two genotypes.

Rapid regulation of P-selectin expression. We studied the rapid regulation of P-selectin expression using thrombin and menadione. Thrombin stimulates membrane-associated NADPH and xanthine oxidase (41), resulting in an increase in P-selectin expression. Menadione depletes intracellular glutathione and also increases P-selectin expression. Exogenous administration of thrombin and menadione resulted in a rapid and marked, dose-dependent increase in the surface expression of P-selectin on wild-type and eNOS knockout cells within 30 min of stimulation. There was no difference in the kinetics (data not displayed) and the degree of rapid regulation of P-selectin translocation between the two genotypes (Fig. 5).

Fig. 5.

Reactive oxygen species (ROS) generation. A: Amplex red assay detected a significant increase (*P < 0.05) in H2O2 generation in response to menadione (left) or thrombin stimulation (right) in WT MLECs, but both substances failed to increase H2O2 generation in KO MLECs. Levels of H2O2 in KO cells at baseline were higher than in WT cells (†P = 0.009 in thrombin group; †P = 0.025 in menadione group), which argues that NO in WT cells functions to scavenge ROS. ODU, optical density units. B: exogenous administration of thrombin and menadione resulted in a rapid and marked, dose-dependent increase in the surface expression of P-selectin on WT and eNOS KO cells within 30 min of stimulation.

Measurement of reactive oxygen species generation. The Amplex red assay detected a significant increase in reactive oxygen species (ROS) generation in response to menadione or thrombin stimulation in wild-type MLECs, but both substances failed to increase ROS generation in knockout MLECs. Absolute levels of ROS in knockout cells at baseline were higher than in wild-type cells (P = 0.009 and P = 0.025 in the thrombin and menadione groups, respectively; n = 8), which argues that NO in wild-type cells functions to scavenge ROS.

Adhesion studies under flow. To evaluate potential functional differences between wild-type and eNOS knockout endothelium, we performed adhesion studies under physiologically relevant flow conditions in a parallel plate laminar flow chamber. We tested the adhesion of leukocytes to resting and TNF-α-activated MLECs from eNOS knockout and wild-type mice at 1.0 dyn/cm2 (Fig. 6). There were no differences in firm adhesion or rolling of human monocytes on the two monolayers under resting and activated conditions. TNF-α treatment, however, increased firm adhesion to wild-type (P < 0.0001) and eNOS knockout (P < 0.0001) MLECs significantly compared with resting endothelium. Adhesion of purified mouse neutrophils and splenocytes was the same in knockout and wild-type MLECs in three experiments (P = nonsignificant; data not shown).

Fig. 6.

Leukocyte adhesion to WT and eNOS KO MLECs. Endothelial cells were plated at confluence on tissue culture slides and cultured for 48 h. Monolayers were stimulated with murine TNF-α (10 ng/ml) for 6 h. Firm adhesion of purified human monocytes to monolayers is reported for 1.0 dyn/cm2, because few interactions were noted above this level of shear stress. Firm adhesion to WT (*P < 0.0001) and eNOS KO MLECs (†P < 0.0001) increased significantly compared with resting endothelium.


Several experimental lines of evidence confirm a role for endothelium-derived NO as an autocrine regulator of endothelial adhesiveness and leukocyte-endothelial cell interactions. However, there are several potential molecular mechanisms by which NO could exert these effects. One possibility is that endothelial NO directly interacts with molecular targets, e.g., IκBα, that regulate adhesion molecule expression (35). For example, De Caterina et al. (5) found that human saphenous vein endothelial cells treated with l-NAME show increased transcription of VCAM-1, suggesting that endogenous NO may tonically inhibit VCAM-1 expression. Another hypothesis is that the role of NO lies in radical scavenging and maintaining low oxidative stress. In this respect, the balance between local superoxide and nitric oxide production may be a critical determinant in the etiology of disease including atherosclerosis and ischemia-reperfusion injury (4). Indeed, there are experimental data suggesting that the antiadhesive effects of NO are related to its ability to scavenge superoxide in vivo (9).

In this study, we used eNOS knockout mice as a unique tool to study the effects of chronic absence of eNOS on endothelial cell growth and gene regulation. We previously demonstrated that eNOS knockout mice are hypertensive and lack endothelium-dependent relaxation (15). In vivo studies using these animals showed that genetic deficiency of eNOS increases intimal proliferation in a model of perivascular injury (29) and atherosclerosis in hyperlipidemic apoE/eNOS double knockout mice (2, 18). Intravital microscopy studies in the eNOS knockout animals revealed conflicting results with regards to the role of eNOS in leukocyte-endothelial cell interactions (21, 38).

To further our understanding of how eNOS may modulate leukocyte endothelial cell interactions and atherosclerosis development, we dissected the role of NO in the autoregulation of adhesion molecule expression and leukocyte-endothelial cell interactions, employing eNOS knockout and eNOS wild-type endothelial cells.

Here we have demonstrated that wild-type MLECs produce significant amounts of NO upon stimulation with VEGF or the calcium ionophore ionomycin or upon exposure to shear stress. Detection of NO production documents that eNOS function is preserved in wild-type cells and that the culture medium provides adequate amounts of substrate and cofactors. Given this evidence, we believe that we were not able to discern the difference in NO production between wild-type and knockout cells at baseline because of a lack of sensitivity of current NO detection methods (chemiluminescence being one of the most sensitive and robust methods available), rather than a lack of difference in NO generation between the cell lines itself. MLECs, deficient in eNOS, do not upregulate the expression of iNOS or nNOS to compensate for the loss of eNOS.

Despite the absence of NO production, eNOS knockout cells do not display the characteristics of activated endothelial cells, because they do not increase their surface expression of E-selectin, P-selectin, VCAM-1, or ICAM-1. Furthermore, treatment of unstimulated cells with l-NAME did not change the surface expression pattern of wild-type or knockout MLECs, indicating that acute pharmacological inhibition of eNOS did not lead to endothelial cell activation either. This finding suggests that the lack of activation in eNOS knockout endothelial cells is not due to compensatory changes in gene regulation secondary to eNOS deletion. Stimulation of the endothelial cells with TNF-α led to a significant increase in surface expression of P-selectin, VCAM-1, ICAM-1, and E-selectin, confirming the resting state of the endothelial cells and their responsiveness to an appropriate stimulus. This TNF-α stimulation of the MLECs, however, led to an equally strong increase in surface expression of the adhesion molecules between the two genotypes.

Aside from the transcriptional regulation of P-selectin expression within hours of stimulation, preformed P-selectin is also acutely released from Weibel-Palade bodies located inside the cells. To test whether this acute translocation of P-selectin differs between cells of the two genotypes, we performed cell surface ELISA minutes after stimulation with thrombin and menadione. Stimulation with menadione, which is known to deplete intracellular glutathione (26), the main intracellular antioxidant, led to a significant increase in surface P-selectin expression with no difference between the two genotypes. Stimulation with thrombin, which increases ROS generation by stimulation of the membrane-associated NADPH and xanthine oxidase (41), likewise led to a significant increase in surface P-selectin expression, with no difference between knockout and wild-type cells.

To test whether these results could be influenced by differences in ROS generation or peroxidase activity of the two genotypes, we measured ROS generation using N-acetyl-3,7-dihydroxyphenoxazine, which reacts with H2O2 in a 1:1 stoichiometry to produce a red fluorescent oxidation product. This assay was previously shown to be sensitive enough to allow detection of physiological levels of ROS released into the supernatant (41). Using this assay, we noted a marked and rapid increase in H2O2 generation following stimulation of wild-type cells with menadione and thrombin. In eNOS knockout cells, however, the two substances failed to significantly increase H2O2 generation. Interestingly, absolute levels of ROS in knockout cells at baseline were higher than in wild-type cells, consistent with the ability of wild-type cells to scavenge ROS through NO production. Because stimulation of the knockout cells with menadione and thrombin did not further increase H2O2 generation, signals other than H2O2 must be responsible for the increase in surface expression of P-selectin in knockout cells.

These results demonstrate that absence of endogenous eNOS does not by itself increase the surface expression of adhesion molecules. In agreement with these expression studies, studies of leukocyte-endothelial cell interactions under physiologically relevant laminar flow conditions revealed no differences in leukocyte rolling or firm adhesion to the two monolayers.

Our results demonstrate that eNOS-derived NO is not required to continuously suppress adhesion molecule expression in endothelial cells. Despite the absence of eNOS, MLECs from eNOS knockout mice maintain low levels of E-selectin and VCAM-1, hallmarks of resting endothelium. Furthermore, the increase in basal H2O2 production in eNOS knockout MLECs does not appear sufficient to alter basal adhesion molecule expression. Our findings are in agreement with several independent studies. Cartwright et al. (1) found no evidence that physiological levels of exogenous NO alter adhesion of lymphocytes to an IL-1β/TNF-α-stimulated HUVEC line when coincubated with superoxide dismutase. In addition, NO did not alter the endothelial expression of VCAM-1, ICAM-1, or E-selectin on these cells. Using the same eNOS knockout mice that we have used in the present study, Sanz et al. (38) found no difference in baseline leukocyte rolling or firm adhesion between wild-type and eNOS knockout mice in either the cremasteric or intestinal microcirculation. However, they did provide evidence that leukocyte adhesion in eNOS knockout animals is more sensitive to increases in oxidative stress (38).

Why then do some studies of cultured endothelial cells show an effect of l-NAME on the expression of surface adhesion molecules? The apparent differences may lie in differences between the cell populations studied, i.e., MLECs vs. human saphenous vein endothelial cells (5) and HUVECs (30), or in differences in oxidative stress in the cell culture systems studied. In this respect, we have taken great care to use endotoxin-free cell culture grade reagents. Studies in HUVECs suggest that the l-NAME effect on neutrophil adhesion is associated with increases in oxidative stress and can be blocked with oxygen radical scavengers (30). In addition, our culture system would reveal only effects of endogenous NO production within endothelial cells themselves and would not reflect the effects of interactions with nonendothelial cells such as circulating leukocytes or vascular smooth muscle cells.

The pathophysiological relevance of these results relates to the role of endothelial dysfunction and reduced endothelial NO production in atherogenesis. Atherosclerosis occurs in predictable lesion prone areas, areas of branch vessels with low or oscillatory flow. Data by De Keulenaer et al. (6) suggest that such flow conditions differentially upregulate NADH oxidase, a major source of ROS in vascular cells. Taken together, these findings suggest that deficiency of eNOS may exert its effects by altering the susceptibility of endothelial cells to oxidative stress and the balance between NO and oxygen free radicals.

In summary, we provide evidence that chronic absence of eNOS does not by itself activate cultured MLEC cells. These results argue against a requirement for eNOS to tonically suppress endothelial cell activation. Furthermore, chronic absence of endogenous eNOS-derived NO does not affect growth characteristics, viability, or morphology of cultured MLECs. We emphasize, however, that these results do not speak against an effect of eNOS-derived NO on endothelial activation; rather, they indicate that the mechanism for such effects cannot be tonic, direct suppression of endothelial activation by NO. In this respect, we provide evidence that NO acts to scavenge ROS in this cell culture system, an effect that would counter-balance signals that lead to endothelial cell activation.


We are grateful to Dr. Guillermo Garcia-Cardena for helpful discussions. We also thank Gabriele Riehl for excellent technical help. We are saddened by the death of William J. Atkinson, one of the coauthors of this work.


This work was supported by Deutsch-Forschungsgemeinschaft Grant KU-1206/1-2 (to P. Kuhlencordt), National Institute of Neurological Disorders and Stroke Grant NS-33335 and National Heart, Lung, and Blood Institute Grant HL-57818 (to P. L. Huang) and National Institutes of Health Grant P01–36028 (to M. A. Gimbrone, Jr.). P. L. Huang and A. Rosenzweig are Established Investigators of the American Heart Association.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • Deceased.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 23.
  23. 24.
  24. 25.
  25. 26.
  26. 27.
  27. 28.
  28. 29.
  29. 30.
  30. 31.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 39.
  39. 40.
  40. 41.
  41. 42.
View Abstract