flow-originated unidirectional laminar shear stress is an essential physiological protector of vascular integrity. It exerts beneficial effects via an array of cellular responses, such as upregulation of endothelial nitric oxide synthase (eNOS) gene expression and nitric oxide formation (23), upregulation of the antioxidant enzyme superoxide dismutase (6, 11), prevention of endothelial cell apoptosis (22, 24), and attenuation of inflammatory protein expression (25). Under physiological conditions, low ambient concentrations of reactive oxygen species are required for growth signaling (9). Flow-dependent physiological production of hydrogen peroxide from mitochondrion mediates coronary vasodilatation (19). On the other hand, overly produced reactive oxygen species from further activated NAD(P)H oxidases and secondary oxidase systems generally promote inflammation, apoptosis, and uncoupling of eNOS (2, 5, 8). These observations seem to suggest that the central functional core of shear stress is to prevent reactive oxygen species overproduction. Indeed, a recent study by Yamawaki et al. (25) demonstrated that shear stress inhibition of vascular inflammation is mediated by downregulation of endothelial thioredoxin-interacting protein, a protein that is negatively involved in thioredoxin-mediated protection against oxidant stress. Nevertheless, the entire molecular mechanisms underlying shear stress control of redox signaling and the identity of the potential mechanotranducers involved remain to be fully elucidated.
Milovanova et al. (Ref. 18, see p. C66 in the present issue) report that cessation of flow activates endothelial NAD(P)H oxidases by inactivating Kir6.2-containing ATP-sensitive K+ (KATP) channels, resulting in enhanced reactive oxygen species production in pulmonary endothelial cells and subsequent cell proliferation. The authors found that flow cessation-induced cell proliferation is absent in endothelial cells isolated from Kir6.2 and gp91phox-deficient mice. In addition, activation of NAD(P)H oxidase did not occur in Kir6.2-deficient endothelial cells, indicating an upstream role of the KATP channels. Although they used mouse pulmonary microvascular endothelial cells in their study, the same group recently demonstrated that blockade of KATP channels activates NAD(P)H oxidases in large vessel (bovine aortic) endothelial cells (16). These data indicate that KATP channels could function as mechanotransducers in both large and small vessel endothelial cells. In contrast to lack of flow, normal shear stress could thus prevent NAD(P)H oxidase-dependent overproduction of reactive oxygen species by maintaining normal endothelial KATP channel activity. Activation of vascular NAD(P)H oxidases underlies development of atherosclerotic vascular diseases (2, 3). Likewise, NAD(P)H oxidase-derived reactive oxygen species contribute to chronic pulmonary hypertension by augmenting vasoconstriction and vascular remodeling (1, 15). Thus the observation that KATP channels play an important role in modulating endothelial redox balance could represent a novel mechanism whereby flow-mediated shear stress is vasoprotective in both large and small vessels.
In addition, this KATP channel-dependent activation of endothelial NAD(P)H oxidases appears specific for loss of flow. The authors demonstrated that KATP antagonist glyburid mimicked flow cessation-induced activation of NAD(P)H oxidases, and that this response was absent in Kir6.2-deficient cells. Angiotensin II, however, had similar effects in activating endothelial NAD(P)H oxidases in both wild-type and Kir6.2-deficient cells, strongly suggesting a flow-specific role of KATP channels in activation of endothelial NAD(P)H oxidases. Thus these data characterized an agonist-specific modulation of gp91phox (Nox2)-containing endothelial NAD(P)H oxidases. Whether similar differential modulations occur with other isoforms of NAD(P)H oxidases, i.e., Nox1, Nox4, and Nox5 (all exist in endothelial cells) (2, 7), remain to be examined. This is an interesting question to address because Nox proteins are activated in an agonist-specific fashion that seems linked to their intracellular localizations (2, 7). Activation of Nox1 promotes endothelial cell proliferation (12). Then, besides Nox2, is endothelial Nox1 activated by flow cessation as well? If so, does it require KATP channels similarly or differently from other agonists such as angiotensin II?
This study is unique because the experimental strategy offers an oxygenated “ischemic” condition. Ischemia is often accompanied with hypoxia. The latter is also associated with modulation of redox signaling (20). A “pure” ischemic condition would be more effective in revealing the signaling events activated solely by the reduction in blood flow. In this study, the authors used a system to simulate ischemia by rerouting the flow from the luminal (endothelial cells cultured on capillary fibers to form circulation) to the abluminal compartment. This eliminates shear stress to endothelium while preserving oxygen and nutrient accessibility. Using this system, the authors clearly demonstrated that endothelial NAD(P)H oxidases are the sole dominant source for reactive oxygen species in response to ischemia/loss of flow. The gp91phox-deficient endothelial cells completely failed to produce reactive oxygen species in response to flow cessation. Kir6.2 knockout partially but significantly reduced reactive oxygen species production, indicating that Kir6.2-KATP is required for NAD(P)H oxidase activation. On the other hand, the partial effectiveness seems to suggest that other flow-insensitive endothelial potassium channels also contribute to activation of NAD(P)H oxidases. This hypothesis, however, requires further investigation.
In contrast to the vasoprotective unidirectional flow, disturbed or oscillatory flow at bifurcation and branching points of the vascular bed is atherogenic (13). Molecular studies further demonstrated that oscillatory shear stress provokes a marked and sustained increase in NAD(P)H oxidase-dependent reactive oxygen species production (6) in the endothelium and reactive oxygen species-dependent monocyte adhesion to endothelium (10). Importantly, NAD(P)H oxidase-derived hydrogen peroxide causes xanthine oxidase cleavage and activation in response to oscillatory shear (17). Oscillatory shear may also uncouple eNOS via hydrogen peroxide-dependent mechanisms (5), and under this condition, previously observed upregulation of eNOS (4) could be detrimental. Thus activation of NAD(P)H oxidases by oscillatory shear is a key step in activating secondary oxidase systems to sustain reactive oxygen species production. It would be very interesting to understand whether the same KATP channels involved in unidirectional shear attenuation of endothelial NAD(P)H oxidases are inhibited by oscillatory shear to allow activation of endothelial NAD(P)H oxidases. If true, one KATP channel agonist could augment beneficial effects of unidirectional shear at the mean time of diminishing pathological signaling initiated by oscillatory shear.
Finally, based on the observation of KATP channels as modulators of endothelial NAD(P)H oxidases, many questions remain as to how intracellular K+ ions mediate downstream signaling that ultimately affect NAD(P)H oxidase activity. Using angiotensin II as a potent agonist, Seshiah et al. (21) characterized a NAD(P)H oxidase-activating signaling cascade in vascular smooth muscle that involves sequential activation of PKC, Src, EGF receptor, phosphatidylinositol 3-kinase, and phosphatidylinositol 3,4,5-trisphosphate-dependent Rac-1 membrane translocation. Activation of endothelial NAD(P)H oxidases by angiotensin II requires p47phox phosphorylation and membrane translocation (14). The authors found the roles of KATP channels in activating endothelial NAD(P)H oxidases are specific for flow cessation. What are the signaling effectors downstream of intracellular K+ accumulation? Are none of the above-described signaling events involved in flow cessation-induced activation of NAD(P)H oxidases? Or if some are involved, at which levels do the signals diverge to allow agonist specificity? These signaling mechanisms are important as they would predict efficacy of K+ channel agonists in counterbalancing NAD(P)H oxidase activation by pathological stimuli other than loss of flow.
The author's work is supported by an American Heart Association Scientist Development Grant 0435189N, an American Diabetes Association Research Award 7-04-RA-16, a Career Development Award from the Schweppe Foundation, and National Heart, Lung, and Blood Institute Grants HL-077440 and HL-081571.
- Copyright © 2006 the American Physiological Society