NF-κB signaling pathway has been known to play a major role in the pathological process of atherogenesis. Unlike high shear stress, in which the NF-κB activity is transient, our earlier studies have demonstrated a persistent activation of NF-κB in response to low shear stress in human aortic endothelial cells. These findings partially explained why low shear regions that exist at bifurcations of arteries are prone to atherosclerosis, unlike the relatively atheroprotective high shear regions. In the present study, we further investigated 1) the role of NF-κB signaling kinases (IKKα and β) that may be responsible for the sustained activation of NF-κB in low shear stress and 2) the regulation of these kinases by reactive oxygen species (ROS). Our results demonstrate that not only is a significant proportion of low shear-induced-kinase activity is contributed by IKKβ, but it is also persistently induced for a prolonged time frame. The IKK activity (both α and β) is blocked by apocynin (400 μM), a specific NADPH oxidase inhibitor, and diphenyleneiodonium chloride (DPI; 10 μM), an inhibitor of flavin-containing oxidases like NADPH oxidases. Determination of ROS also demonstrated an increased generation in low shear stress that could be blocked by DPI. These results suggest that the source of ROS generation in endothelial cells in response to low shear stress is NADPH oxidase. The DPI-inhibitable component of ROS is the primary regulator of specific upstream kinases that determine the persistent NF-κB activation selectively in low shear-induced endothelial cells.
- upstream κB kinases
- laminar shear stress
- oxidative stress
- reactive oxygen species
endothelial cells lining the vascular wall respond differentially to different flow shear regimen. Studies from the past have clearly demonstrated that regions of artery that experience low shear stress and reversing flow patterns such as arterial bifurcations and curved segments of the large elastic artery are highly predisposed to the development of atherosclerotic lesions, unlike high shear regions that are relatively resistant to the disease (2, 15). Our earlier in vitro studies (17–19) and in vivo reports from other laboratories (9) have indicated the possible involvement of NF-κB signaling mechanisms in low shear lesion-prone areas. The NF-κB family of transcription factors (p65 or RelA, p50, p52, Bcl-3, c-Rel and RelB) regulate many proinflammatory endothelial signals that are activated in the early stages of atheroma formation (29). NF-κB family members form homo- or heterodimeric complexes that are retained in the cytoplasm in quiescent (unstimulated) cells by associating with a family of inhibitor proteins, called IκBs, which include IκBα and IκBβ. Activation of NF-κB by a variety of stimuli is dependent on the phosphorylation of site-specific serine residues such as Ser32 and Ser36 of IκBα and Ser19 and Ser23 of IκBβ, followed by the subsequent degradation of the IκB proteins, which allows the translocation of the active form of NF-κB into the nucleus to activate downstream target genes. Phosphorylation of IκB is mediated by a macromolecular complex called the signalsome with a molecular mass of ∼900 kDa that is composed of three major proteins, IKKα, IKKβ, and IKKγ/NEMO (NF-κB essential modulator). Both IKKα and IKKβ are serine/threonine kinases, which exert kinase activity upon stimulation, whereas IKKγ lacks kinase activity but is required for the activation of IKKα and IKKβ in vivo (14). IKKα and IKKβ are highly homologous with the characteristic structure of an NH2-terminal protein kinase domain, which contains a canonical mitogen-activated protein kinase kinase activation loop motif of the sequence Ser-X-X-X-Ser. Phosphorylation of the site-specific serine residues in the activation loop, i.e., Ser176 and Ser180 in IKKα and Ser 177 and Ser181 in IKKβ, is thought to be important for the regulation of their kinase activity. Replacement of these serine residues to alanine or mutation of these residues inactivates their kinase activity.
The possible involvement of certain components of the NF-κB signaling pathway that could be influenced by flow shear stress has been reported (4, 10). However, these reports have focused on the effect of physiological high shear stress of 12 and 15 dyn/cm2. Studying the nature of these upstream signaling elements involved in NF-κB regulation under conditions of low shear stress is important, because this environment is more relevant to the atherogenic process observed at sites of branching arteries. Therefore, the present study has investigated the IKK regulation in low shear flow-adapted human aortic endothelial cells. Our results indicate that even though both IKKα and IKKβ activity are turned on by low shear stress, IKKβ activity contributes to a major proportion of total IKK activity. NF-κB-inducing kinase (NIK), the upstream member of this signaling pathway, also contributes to in vitro IκBα phosphorylation. The IKK activity by low shear stress is significantly downregulated by inhibitors of NADPH oxidase, indicating that the reactive oxygen species (ROS) generated by low shear stress through this enzyme could be responsible for activating the NF-κB signaling mechanism. The contribution of ROS from xanthine oxidase, however, may not influence the low shear-induced NF-κB signaling mechanism. These results identify new specific targets that may help to control the atherogenic process that is augmented by low shear stress.
MATERIALS AND METHODS
Human aortic endothelial cells (HAEC; Clonetics, San Diego, CA) were cultured in MCDB-131 medium (Sigma, St. Louis, MO) containing 10% bovine calf serum (BCS; Hyclone, Kansas City, KS) and enriched with 250 ng/ml fibroblast growth factor (PeproTech, Rocky Hill, NJ), 1 mg/ml epidermal growth factor (PeproTech), 1 mg/ml hydrocortisone (Sigma), 100 U/ml penicillin, and 100 mg/ml streptomycin (Media-tech, Herndon, VA). Cells from passages 4 to 7 were used in all the experiments.
HAEC maintained in MCDB-131 medium supplemented with 10% BCS were incubated in reduced serum (2%)-containing medium for 20 h before treatments. Cells induced by TNF-α (Pharmingen, San Diego, CA) at a concentration of 10 ng/ml were used as positive controls. Cells incubated with 400 μM apocynin, 10 μM diphenyleneiodonium chloride (DPI; Sigma), 100 μM pyrrolidine dithiocarbamate (PDTC), 200 μM allopurinol, and 0.5 μM carbonyl cyanide m-chlorophenylhydrazone (CCCP; Sigma) were used for blocking studies. For spectrofluorometric studies, cells were incubated with 10 μg of SOD mimetic [Mn(III) tetrakis(4-benzoic acid) porphyrin chloride; Sigma] for 45 min before shear stress. Concentrations of these inhibitors were chosen on the basis of studies reported previously (5, 11, 24). Cell viability determined using the standard Trypan blue dye exclusion method at these concentrations after 24 or 48h was >98%.
Shear stress experiments.
Polyester film (10 × 19-cm Mylar sheets; Regal Plastics, San Antonio, TX) precoated with 2% gelatin was seeded with HAEC that were grown to near confluence. The cells were then incubated in MCDB-131 medium supplemented with 2% serum without growth factors and hydrocortisone for 20 h before the initiation of the flow shear to avoid any influence of growth factors on the induced responses. Flow experiments were performed using the closed-loop flow system as described previously (17–19). Cells were subjected to low (2 dyn/cm2) or high shear stress (16 dyn/cm2) by adjusting the height of separation between the parallel plates. Cells seeded on slips and grown as static cultures at 37°C were used as negative controls.
Expression vectors and transient transfections.
Mammalian expression vectors encoding untagged NIK (wild or mutant), Flag-tagged IKKα and IKKβ (wild type), the mutant constructs of Flag-tagged IKKα-KM (K44A) and IKKβ-KM (K44A), and the substrate glutathione S-transferase fusion protein of IκBα with 1–100 amino acids (GST-IκBα) were used in this study as described previously (21). Transfection of these constructs was performed as reported earlier (17). In brief, HAEC seeded on plastic slips for shear experiments were incubated in MCDB-131 medium containing 2% serum devoid of growth factors for 24 h. The expression vectors (full-length cDNA, 2 μg) were transfected using the Effectene transfection reagent (Qiagen, Valencia, CA) following the manufacturer's protocol. The DNA-lipid complex was added to the cells slowly and incubated in a limited volume (8–10 ml) of medium for 24 h and then subjected to shear stress. For confirmation studies, parallel experiments were carried out using bovine aortic endothelial cells (BAEC) to test limitations due to reduced transfection efficiencies of human endothelial cells.
NF-κB-dependent reporter assays.
Endothelial cells were transfected with 100 ng of the NF-κB luciferase reporter system (p4xNF-kB Luc; Mercury pathway profiling luciferase system, Clontech Laboratories) along with each wild-type (IKKα and IKKβ) expression construct. The total DNA was kept constant at 2 μg/plate. Transfected cells were subjected to shear stress for 6 h, and the luciferase activity was measured as per the manufacturer's instructions (Promega, Madison, WI) using a TD-20/20 luminometer (Turner Design, Sunnyvale, CA).
In vitro phosphorylation assays.
Cells induced by shear stress were washed with ice-cold PBS and lysed for 30 min on ice in 1 ml of lysis buffer containing 1% Nonidet P-40, 50 mM HEPES (pH 7.3), 150 mM NaCl, 2 mM EDTA, 1 μg /ml aprotinin, 1 μg/ml leupeptin, 1 μg/ml pepstatin, 1 mM phenylmethylsulfonyl fluoride, 0.1 mM sodium orthovanadate, and 1 mM NaF. The cell lysates were precleared by incubating with protein A-agarose 4B (BD Pharmingen, San Diego, CA) for 1 h at 4°C in the cold room. After centrifugation, the cleared lysates were incubated with anti-Flag monoclonal antibody (1 μg/ml)-tagged protein A-agarose for 2 h at 4°C. The immunoprecipitates were washed three times with the lysis buffer and twice in kinase buffer containing 20 mM HEPES (pH 7.3), 20 mM MgCl2, 20 mM MnCl2, 1 mM EDTA, 1 mM NaF, 0.1 mM sodium orthovanadate, and 1 mM DTT. The immunoprecipitates were then incubated with 1 μg of GST-IκBα(1–100) and [γ-32P]ATP (10 μCi) in the kinase buffer for 20 min at 30°C. The reaction was stopped by the addition of the Laemmli's sample buffer. The eluted proteins were subjected to SDS-PAGE, and the autoradiograms were visualized using a PhosphorImager (Bio-Rad). In all cases, expression of the transfected proteins was verified by immunoblotting of aliquots of cell lysates. GST-IκBs in the reaction mixtures were also verified by Coomassie blue staining to confirm equal loading.
Inhibition of IKKα and IKKβ activity by siRNA transfection.
siGenome smartpool small interference (si)RNA reagents specific for IKKα and IKKβ were purchased from Dharmacon (Lafayette, CO). The nonspecific control siRNA-transfected cells and untransfected cells were used as controls and to determine the efficiency of transfection. Transfections were performed using Hiperfect transfection reagent (Qiagen). To determine the efficiency of siRNA in blocking IKK activity, we harvested transfected cells after 72 h to measure protein levels of IKKα and IKKβ expression by immunoblotting. Furthermore, cells transfected with siRNA for IKKα and IKKβ either individually or together were subjected to low shear stress for 6 h or kept at static culture conditions. Upon completion of shear stress, cells were washed with HBSS, fixed with 3.7% paraformaldehyde for 10 min, and permeabilized using 0.2% Triton X-100 for 5 min. Cells were blocked with 1% goat serum for 30 min at 37°C and further incubated with rabbit anti-p65 antibody to determine the translocation of p65 subunit into the nucleus. Alexa Fluor 350-tagged goat anti-rabbit was used as secondary antibody. Nuclear translocation of p65 subunit of NF-κB induced by low shear stress was examined under a Nikon ES-T-2000u epifluorescence microscope.
Measurement of intracellular ROS induced by shear stress.
ROS generation was measured in endothelial cells using dihydrorhodamine 123 (DHR 123; Molecular Probes, Eugene, OR) as described previously (23, 30, 31). Cells were preincubated with 10 μM DHR 123 in PBS for 30 min. DHR 123-loaded cells were incubated with or without ROS inhibitors for specific enzyme sources that were shown to inhibit kinase activity. To ensure the specificity of ROS measurements, we used the general ROS inhibitors PDTC and SOD mimetic. Treated cells were exposed to 1 or 3 h of low or high shear stress. Intracellular levels of generated ROS in the cell lysates were measured using a spectrofluorometer (490-nm excitation, 530-nm emission; Jobin Yvon-SPEX Fluorolog FL-3; Edison, NJ) to quantitate the fluorescence level of control versus shear samples.
Each experiment was repeated three times independently. All results were expressed as means ± SD, and P <0.05 was used for significance. Student's t-test was used to determine statistical significance. In some experiments (independent IKKα, IKKβ, and NIK activity and kinase activity with wild-type and mutant constructs in combination), data were analyzed using the randomized block analysis of variance (ANOVA), treating experiments as blocks and treatments as repeated measures. Comparisons of the treatments with each other and with the static control condition were done with means from the ANOVA and t-tests.
Both IKKα and β activity are independently induced by low shear stress.
Our earlier reports (18) have demonstrated the low shear (2 dyn/cm2) stress-induced persistent increase of NF-κB DNA binding activity with the associated phosphorylation and degradation of inhibitor subunit IκBα in HAEC. To further understand whether or not, and to what extent, the upstream kinases in the κB signaling cascade are responsible for this persistent activation of NF-κB, we investigated the independent contribution of IKKα and IKKβ in response to low shear stress. Since the components of NF-κB upstream kinases, namely, IKKα, IKKβ, and NIK, are very tightly associated complexes of a 900-kDa signalsome, the contribution of kinase activities by independent components of this signalsome is difficult to determine. Therefore, to better understand their independent regulation, we used the approach of forced expression of Flag-tagged IKKα and IKKβ components to determine the ectopic kinase activity after shear stress exposure. The transfected cells subjected to high shear (16 dyn/cm2) did not alter the independent activity of either IKKα or IKKβ. In contrast, cells subjected to low shear (2 dyn/cm2) showed a 55 ± 7.1% (P < 0.01) and 101 ± 19% (P < 0.001) increase in the IKKα and IKKβ kinase activity, respectively, compared with static control (Figs. 1 and 2A). The minimum difference in GST-IκBα staining and immunoblotting of Flag-tagged IKKα and IKKβ expression clearly indicated the increase in the kinase activity induced by low shear stress was not due to differences in the transfection efficiency and the quantity of substrate added in each lane. In addition, in both the cells transfected with kinase mutant constructs for IKKα or IKKβ (plasmids encoding IKKα-KM and IKKβ-KM), the IKKα and IKKβ activity remained at the basal level similar to that in the static controls, thus conferring the specificity of these assays. These results, therefore, showed that both IKKα and IKKβ are independently activated by low shear stress, which is not the case in high shear stress (Figs. 1 and 2A). As demonstrated in Fig. 2B, the independent overexpression of wild-type IKKα and IKKβ resulted in enhanced downstream target gene expression when the IKK wild-type constructs were cotransfected with 100 ng of p4xNF-κB Luc reporter system. Cells transfected with p4xNF-κB Luc alone exhibited a 1.5-fold increase over control (P < 0.05). Both wild-type IKKα and IKKβ cotransfection with NF-κB Luc reporter system significantly increased luciferase activity independently (8.1 ± 0.2-fold increase with IKKα and 30.01 ± 1.7-fold increase with IKKβ; P < 0.01 compared with NF-κB Luc-transfected cells). Luciferase activity was downregulated when cells were cotransfected with IKKα mutant (1.2 ± 0.45-fold; P < 0.01). Cotransfection with IKKβ mutant resulted in an unaltered luciferase activity (P = 0.68), thus indicating the specificity of IKKα- or IKKβ-dependent downstream luciferase activity.
IKKβ activity contributes to majority of the total IKK activity in HAEC exposed to low shear stress.
To determine whether both IKKα and IKKβ overexpression together may yield an additive increase in the total IKK activity under conditions of low shear stress, we transiently transfected endothelial cells with both Flag-tagged IKKα and IKKβ constructs and exposed them to shear stress for 30 min or left them as static controls. As demonstrated in Fig. 3, both IKKα and IKKβ wild-type constructs together resulted in 53.66 ± 4.5% (P < 0.01) in total IKK activity over (transfected/no-shear control) basal activity (100%). Both the mutant constructs exhibited 5% IKK activity over no-shear control (P > 0.1). Next, to validate the contribution of each IKK isoform on low shear stress-induced total kinase activity, we transfected cells with a combination of either wild-type IKKβ with ΔIKKα or wild-type IKKα with ΔIKKβ. Cells that were overexpressed with IKK β wild type in the presence of IKKα mutant showed a 30 ± 4.9% increase in activity over static control (P < 0.05), whereas in the presence of IKKβ mutant, the total IKK activity was only 3.6 ± 2.5% over control.
Forced expression of NIK upregulates in vitro IκBα phosphorylation.
To examine the contribution of NIK in low shear-induced IKK, cells were forcefully coexpressed with untagged NIK wild or mutant expression vector with Flag-tagged IKKα and IKKβ expression plasmid. Since the IKK activation results in the phosphorylation of IκBα, we directly measured IκBα phosphorylation that is responsible for the translocation of active NF-κB complex into the nucleus. Cells overexpressed with untagged wild-type NIK demonstrated an increase in IκBα phosphorylation (55 ± 10% over control; P < 0.01) under conditions of low shear stress (Fig. 4). This increase was downregulated to 8 ± 15% when the cells were overexpressed with mutant NIK construct, confirming the activity observed was totally mediated by NIK. In contrast, overexpression of NIK in cells exposed to high shear stress slightly increased (10 ± 14% over control) IκB α phosphorylation, and the level of activity remained the same in ΔNIK overexpressed cells, indicating that high shear stress fails to induce this pathway.
Kinetics of IKKβ activity under low versus high shear stress.
Since IKKβ activity was enhanced and found to be dominant in low shear-exposed cells, we next investigated the time-dependent activation of IKKβ under conditions of low shear stress. In vitro kinase assays were performed in cells overexpressed with Flag-tagged IKKβ and subjected to low shear stress for various time intervals starting from 30 min to 2 h. Cells induced with TNF-α for 2 and 5 min were used as positive controls (data not shown). As presented in Fig. 5A, IKKβ activity reached a maximum activity at 30 min (101 ± 19% increase over control; P < 0.05). The levels started decreasing (40 ± 28% over control) at 60 min and again increased at 120 min (66 ± 24% over control), the longest time point examined in this study. In the case of high shear stress, after a slight rise (10.7 ± 8.9% over control) at 30 min, the IKKβ activity significantly decreased to the basal level (less than no-shear control) at both 60 and 120 min. To determine the responses of other upstream kinases at earlier time periods, we determined IKKα, IKKβ, and NIK activities after 15-min onset of shear. IKKβ activity was not significantly induced in both low and high shear stress compared with static control. IKKα was significantly reduced in high shear-exposed cells (93 ± 1.4%; P < 0.05), whereas low shear-induced activity did not achieve statistical significance compared with control (138 ± 18%; P = 0.19). NIK activity was significantly increased under conditions of low shear stress (36 ± 0.6%; P < 0.05), whereas the level of activity in high shear stress exhibited insignificant changes (95 ± 14%; P = 0.7) compared with static control.
Low shear-induced nuclear translocation of NF-κB/p65 subunit is blocked by the inhibition of endogenous IKK activity with IKKα- and β-specific siRNA.
To test whether inhibition of endogenous IKKα and IKKβ activity would abolish the nuclear translocation of NF-κB p65 subunit, we transfected endothelial cells with siRNA specific for IKKα and IKKβ, either independently or together. As shown in Fig. 6A, after 72 h of transfection, siRNA significantly downregulated (>95%) the endogenous IKKα and IKKβ (protein) expression in endothelial cells. In parallel experiments, siRNA-transfected cells were kept as static controls (Fig. 6A) or subjected to low shear stress (2 dyn/cm2) for 6 h (Fig. 6B). Low shear stress induced nuclear translocation of NF-κB/p65 subunit in siRNA control transfected (Fig. 6B) and untransfected cells (data not shown). The IKKα and IKKβ siRNA transfection significantly blocked the nuclear translocation of low shear-induced NF-κB/p65 subunit (Fig. 6, C and D).
Specific inhibitors of ROS downregulate IKK activity induced by low shear stress.
To identify the upstream mechanism and the mediator that regulates the persistent activity of NF-κB signaling kinases in low shear stress-induced HAEC, we examined the contribution of ROS that are known to induce NF-κB-mediated signaling mechanisms. HAEC transiently transfected with Flag-tagged IKKα and IKKβ were incubated with both specific and general (known) ROS inhibitors for 45 min before the cells were exposed to low shear stress or kept as no-shear controls. As shown in Fig. 7, increased kinase activity observed at 30 min after low shear stress was not downregulated by CCCP (0.5 μM), allopurinol (200 μM), or PDTC (100 μM). However, apocynin (400 μM), a specific inhibitor of NADPH oxidases, and DPI (10 μM), an inhibitor of flavin-containing oxidases like NADPH oxidase, downregulated the low shear-induced IKK activity by 88.4 ± 17.5 and 76 ± 10.5% (P < 0.05 compared with kinase activity in low shear stress in the absence of the inhibitors), respectively. The basal kinase activity in the presence of inhibitors under no-shear control was not influenced.
Direct measurement of ROS in HAEC exposed to low versus high shear stress.
The generation of ROS in low versus high shear stress-exposed endothelial cells was determined using DHR 123. As demonstrated in Fig. 8, low shear exhibited significant increase (P < 0.05) in ROS generation at both 1 h (154.6 ± 4.1%) and 3 h (152 ± 7.55%) after shear stress exposure compared with no-shear control, wherein fluorescence intensity (indicating ROS generation) was considered 100%. ROS generation induced by 3 h of low shear stress was downregulated significantly by preincubating the cells with 10 μg of SOD mimetic agent (105.7 ± 3.05%; P < 0.05 compared with ROS generation in cells exposed to 3 h of low shear), indicating specificity of ROS measurements. When cells were preincubated with inhibitors of specific ROS-generating enzyme sources, intracellular ROS levels were significantly reduced. Cells incubated with DPI (10 μM), an inhibitor of flavin-containing oxidases like NADPH oxidase, exhibited significantly reduced ROS generation by low shear stress (119.7 ± 0.6%; P < 0.05 compared with low shear at 3 h). HAEC exposed to high shear stress for both 1 and 3 h exhibited a significant decrease (P < 0.05 compared with ROS generation in conditions of low shear) in fluorescence intensity. The levels at 1 and 3 h were 81.3 ± 7.8 and 87 ± 25.2%, respectively.
In this report we have demonstrated the persistent increase in IKK activity, and in particular, IKKβ activity, after laminar low shear stress in HAEC. Our results indicated the possibility that the low shear-induced increase in IKK activity is regulated by ROS generated by NADPH oxidases. This ROS-mediated persistent IKK activity may be responsible for the prolonged NF-κB DNA binding activity and the associated endothelium-monocyte adhesion that we observed after exposure to low shear stress in our earlier studies (18, 19). The IKK-dependent translocation of the p65 subunit of NF-κB into the nucleus has been shown in endothelial cells that are adapted to uniform (both spatially and temporally) laminar flow (4, 10). These studies were carried out after unidirectional high shear stress of 15 or 12 dyn/cm2 with human umbilical vein endothelial cells (HUVEC) in which the endogenous activity of total IKKα plus IKKβ was measured or in BAEC in which forcefully expressed total IKK activity was measured using IKKα and IKKβ expression vectors. In our studies, we used HAEC, expressing IKKα and IKKβ independently or together, exposed to uniform laminar flow generating a unidirectional low shear stress of 2 dyn/cm2. HAEC exposed to unidirectional high shear stress of 16 dyn/cm2 were used as a standard to evaluate the low shear response. Static cultures were used as negative control. Although results reported by Bhullar et al. (4) showed total kinase activity, we sought to measure the activity of individual kinases (i.e., IKKα or IKKβ activity) by using GST-IκBα(1–100) as a substrate. Interestingly, in contrast to the findings reported previously (4, 10), in our studies high shear stress failed to induce independent IKKα or IKKβ activity. Time-course experiments showed a significant downregulation of IKK activity at increasing time points compared with no-shear control. However, total kinase activity after high shear stress has been reported to be transiently induced at 5 (10) and 30 min (4). Consistent with these results, we demonstrated previously that 30 min of high shear stress induced a transient NF-κB DNA binding activity that decreased after 2 h and showed a small second phase of activity after 16 h (18). Together, these results clearly revealed a distinct pattern of kinase activity in the same cell type from different sites of the vascular bed or from different sources under similar flow shear regimen. Alternatively, measuring the low levels of endogenous total kinase activity in HAEC versus independent kinase activity (IKKα or IKKβ activity) that are forcefully expressed with the individual expression vectors may account for those differences described above.
The higher levels of IKK activities noticed in low shear stress confirm that insignificant activities observed in high shear stress are not due to decreased transfection efficiencies of endothelial cells from human origin. Our results (Figs. 1 and 2) demonstrate that IKKα and IKKβ are required together to enhance the transient total IKK activity in high shear stress, whereas low shear stress can independently activate significant levels of both IKKα and IKKβ activity. The results of triplicate experiments were consistent in showing that IKKβ activity was enhanced 101 ± 19%, whereas IKKα activity increased only 55 ± 7.1%, compared with control, indicating that IKKβ activity may be preferentially activated in low shear stress. This observation is supported by results of Meiler et al. (22) showing that dominant negative IKKβ overexpression significantly abolished the downstream events of both monocyte rolling and adherence in HUVEC exposed to laminar low shear stress of 2 dyn/cm2.
Induction of IKKs by cytokines, particularly TNF-α, has been shown to be directional. First, TNF-α triggers the activation of IKKα, which then stimulates IKKβ (33). Similar mechanisms could occur under conditions of low shear stress where unlike TNF-α, the preferential activation of IKKβ may regulate IKKα activity. To investigate whether such directional activation is possible under conditions of low shear stress, we performed cotransfection experiments to determine whether the activity of the heterodimeric IKKα and IKKβ complex is influenced by each other, as reported earlier in TNF-α-mediated activation of NF-κB (33), or remains independent. Whereas cells coexpressing both IKKα mutant and IKKβ wild type under conditions of low shear stress induced a 30 ± 4.9% increase in kinase activity, cells coexpressing IKKα wild type and IKKβ mutant showed no enhancement in the amount of total kinase activity. The activity in these cells was very similar to that in cells cotransfected with both IKKα and IKKβ mutant constructs. These cotransfection experiments may indicate 1) the increased functionality of IKKβ subunit, since it is known that IKKβ exhibits a 20- to 50-fold higher level of activity than IKKα (21, 33), and 2) IKKβ activity may be essential for IKKα activity. However, it is apparent from the results presented in Figs. 1 and 2B that low shear stress could induce the activity of both ectopically overexpressed IKKs in an independent manner, unlike TNF-α activity (33). Therefore, the possibility that IKKβ activity may be essential for IKKα kinase activity is implausible. A plausible model is that low shear stress could activate the upstream kinases like NIK or MEKK3 and then stimulate both IKKs. In our attempts to determine the contribution of MEKK3, low shear stress did not induce IKK activity when endothelial cells were overexpressed with Flag-tagged MEKK construct (unpublished results).
However, overexpression of NIK did have a positive influence over IκBα phosphorylation even at an earlier time point (15 min), indicating that NIK could be the only upstream regulator. These cotransfection experiments were also performed in BAEC to confirm the results and to determine whether transfection efficiency of wild and mutant constructs would influence the observed results. In our experience, the basal kinase activity (no-shear controls) in human endothelial cells seems to be higher than in the bovine cells. In addition, because of the limitations of transfection efficiencies, a maximum of 0.5- to 2-fold increase could only be achieved in kinase activity assays using human cells. However, this increase was sufficient to induce a significant downstream gene (luciferase expression as shown in Fig. 2B), thus confirming the functional consequence of the measured kinase activities.
Our time-course experiments clearly indicated that low shear stress exhibited IKK activity up to 120 min, unlike HAEC exposed to high shear stress (used as control) where IKK activity gradually declined to a level even less than that of no-shear controls with increasing hours of exposure to shear stress. Also, it is surprising that IKKβ and IKKα independently induced 101 ± 19 and 55 ± 7.1% increases in activity, respectively, but did not yield a cumulative increase when they were cotransfected together. The increase was only 54 ± 4.5% over control. This could be due to the amount of plasmid DNA that was used for transient transfection (with a limit of 2 μg of total plasmid DNA/30-cm2 area of slip, 1 μg of IKKα and 1 μg of IKKβ were used in cotransfection experiments, whereas 2 μg of IKKα or IKKβ plasmid DNA was used in independent transfections). The transient transfection of IKKα and IKKβ with p4xNF-κB Luc reporter system independently induced 8.1 ± 0.2-fold and 30.01 ± 1.7-fold luciferase activity, respectively, over luciferase activity in the absence of IKKα or IKKβ overexpression, indicating the increase in IKK activity is responsible for NF-κB-dependent downstream target gene expression.
We examined the NIK activity in HAEC overexpressed with untagged NIK expression construct with Flag-tagged IKKα and IKKβ plasmid. As presented in Fig. 4, NIK expression increased (55 ± 10% over control) IκBα phosphorylation in low shear stress. However, under conditions of high shear stress, the increase in IκBα phosphorylation (10 ± 14% over control) was insignificant compared with control. This observation in high shear stress is opposite to the report of Hay et al. (10), who demonstrated NIK involvement in high shear stress in NIK-overexpressed HUVEC, using nuclear translocation of p65 by immunofluorescence as the end point. This discrepancy again could reflect differences in methodology, controls, and cell type used. We performed in vitro kinase assays in untagged NIK-overexpressed HAEC that were cotransfected with Flag-tagged wild-type IKKα or IKKβ plasmids. Since the cell lysates were immunoprecipitated with anti-Flag antibody, the resulting kinase activity is due to IKK that is regulated by overexpressed NIK activity. The controls in our studies were HAEC overexpressed with NIK construct but not exposed to shear stress. Although NIK activity was slightly increased in high shear stress, this increase was statistically insignificant compared with the activity that resulted in cells expressing mutant NIK construct and the no-shear control. In the report by Hay et al. (10), it is ambiguous whether the results were compared with no-shear controls. In addition, endothelial cells derived from different vascular beds could have different signaling responses (in particular, differences between venous and arterial cells). Overall, these results strongly support the hypothesis presented in our earlier reports (17–19) stating that the activity of NF-κB signaling kinases is highly upregulated in endothelial cells under conditions of low shear stress typical of branching arteries, which are predisposed to atherosclerosis. In addition, results from these overexpression studies are further supported by the IKKα and IKKβ siRNA blocking experiments, as demonstrated in Fig. 6. IKKα or IKKβ siRNA transfected individually or combined together blocked low shear-induced nuclear translocation of p65 subunit in most cells. However, in IKKα siRNA-transfected cells, we could observe p65 translocation with limited fluorescence in some cells. This could probably be due to the endogenous IKKβ activity, which remains active even after the suppression of IKKα activity. Together, both overexpression and blocking studies support our hypothesis that persistent activity of upstream kinases regulates the prolonged NF-κB activity under conditions of low shear stress.
One of the earliest transcription factors recognized in eukaryotic cells as a redox-sensitive regulator is NF-κB. Accumulating evidence shows that activation of NF-κB and the subsequent upregulation of κB-dependent gene expression are significantly induced by the generation of ROS (1, 5, 27, 28). Stimulation of IKK activity by H2O2 (14) has provided direct evidence for the involvement of ROS in regulating upstream NF-κB signaling kinases. Use of reducing agents like N-acetylcysteine has been shown to inhibit IKKα and IKKβ activity, thus suggesting a pivotal role of IKKα and IKKβ in the redox regulation of NF-κB (26). Using aortic endothelial cells obtained from C57Bl/6 (MAE-C57) and p47phox−/− (MAE-p47−/−) mice, which lack a component of NAD(P)H oxidase, Hwang et al. (13) demonstrated the generation of superoxide produced from p47phox-dependent NADPH oxidase by oscillatory shear stress (±5 dyn/cm2), which in turn leads to monocyte adhesion. Acute exposure (1 h) of mouse aortic endothelial cells to laminar high shear stress (15 dyn/cm2) increased superoxide production, whereas chronic exposure (18 h) reduced it significantly. In both human and bovine aortic endothelial cells, we observed a significant decrease in dihydrorhodamine oxidation after acute (1 h) exposure to laminar high shear stress, indicating reduced generation of ROS. It is not clear at this time whether these discrepancies are due to differences in the genetic make up of the cell types used (mouse vs. human), the method of shear stress application (cone and plate model vs. parallel-plate method), and/or the type of fluorescein probes used [dihydroethidium (DHE) vs. DHR 123].
In general, three fluorescein probes, namely, DHE, DHR 123, and dichlorofluorescein diacetate (H2DCFDA), have been described as useful tools for the measurement of ROS. The redox-sensitive fluoroprobe DHR 123 has been used widely in endothelial cells to measure the generation of ROS (23, 30, 31). In the present study, we preferred the use of DHR 123 to measure ROS generation over H2DCFDA because of the susceptibility of the latter agent to undergo autooxidation during shear stress (7). In addition, we experienced leakage of the dye from cells in a short span of time under shear in parallel-plate flow chambers. DHE has been suggested to be used as a qualitative indicator of superoxide anion but is not recommended to be used in quantitative measure, because it can catalyze the dismutation of superoxide and can be oxidized by cytochrome c within the cell (3). DHR 123 is reported to be more sensitive in detecting ROS than other dyes tested (32, 34). DHR 123 is membrane permeable. It is oxidized by ROS intracellularly to become a stable fluorescein (Rh123) derivative, which is membrane-impermeable and is pumped into mitochondria (6). Therefore, this dye does not leak out of cells after being oxidized. Although H2DCFDA and DHE are reported to specifically target H2O2 and superoxide anions, respectively, DHR 123 has been widely used for testing global ROS. Multiple ROS in particular peroxynitrites convert DHR 123 into a highly fluorescent form (16, 32, 34). As reported by Lopez-Ongil et al. (16), if DHR 123 is more sensitive to peroxynitrite, which is formed after the reaction of superoxide with nitric oxide, it is likely that high shear stress generates minimal peroxynitrite at the initiation of shear stress (there is more nitric oxide and less superoxide release) and thus could have exhibited insignificant DHR 123 fluorescence in our studies. This difference in the sensitivity of the fluorescent dye used in our study could have resulted in differences that we observed in the level of generation of ROS under conditions of high shear stress.
The generation of superoxide (O2−), an important ROS in the vasculature, is mediated by several enzyme systems, including NADPH oxidases, lipoxygenases, xanthine oxidases, and myeloperoxidases (8). Our attempts to identify whether one of these enzyme sources could be involved in the persistent IKK activity in laminar low shear stress-exposed endothelial cells clearly indicated that apocynin- and DPI-inhibitable ROS mediate the enhanced and persistent upstream IKK activity. PDTC, a metal chelator, which blocked the downstream NF-κB DNA binding activity as shown in our earlier report (19), did not alter the upstream IKK activity (data not shown), indicating that PDTC-inhibitable ROS does not influence upstream IKK activity. Our PDTC data support the earlier report of Hayakawa et al. (11), who reported that TNF-α-induced IKK activation was not significantly inhibited by PDTC. However, PDTC inhibits NF-κB DNA binding activity by preventing the downregulation of IκBα phosphorylation or directly blocks the ubiquitin ligase toward phosphorylated IκBα, therefore blocking nuclear translocation of NF-κB at the downstream end. These results raise the possibility that different ROS activate specific stages of NF-κB signaling mechanisms. Alternatively, oxidative radical stress may stimulate IKK through the suppression of protein phosphatase, such as PP2A, because these enzymes are known to be inactivated by oxidative modification (14). Therefore, thorough understanding of the contributing factors and their underlying mechanisms is required to identify potential drug targets to attenuate the enhanced NF-κB signaling that contributes to atherogenesis.
In addition, the involvement of ROS-generating enzyme sources could be cell type specific. Interleukin-1β has been shown to activate NF-κB through the production of ROS by 5-lipoxygenase in lymphoid cells and by NADPH oxidase in monocytic cells (5). Our results suggest that NADPH oxidases are involved in activating the NF-κB signaling pathway in endothelial cells exposed to laminar low shear stress. Furthermore, endothelial cells incubated with 2-deoxy glucose to block pentose shunt pathway for NADPH production significantly downregulated overexpressed IKK activity (data not shown). This result confirms the contribution of NADPH oxidase in activating IKK under conditions of low shear stress. Our study also suggests that specific species of reactive oxygen intermediates could only induce the upstream NF-κB signaling kinases. In summary, our study has identified potential targets that are influenced by hemodynamics of the blood flow, which contribute to the development of atherosclerotic lesions at localized regions (bifurcations) of large and muscular arteries that are characterized as low shear stress regions of the vasculature.
This study was supported by National Heart, Lung, and Blood Institute Grant 1R01 HL-63032-01A1 (to S. Mohan) and partial support as salary for a technician funded by the Office of Science (Biological and Environmental Research), U.S. Department of Energy, Grant DE-FG02-03ER63449. Spectrofluorometer facilities for ROS measurements were supported by an unrestricted grant to the Dept. of Ophthalmology from Research to Prevent Blindness, Inc. (to R. D. Glickman).
We thank Dr. Douglas R. Spitz, Free Radical and Radiation Biology Program, University of Iowa, and Dr. Anthony J. Valente, Department of Medicine, University of Texas Health Science Center at San Antonio, for support, discussion, and valuable suggestions.
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.
- Copyright © 2007 the American Physiological Society