New insights into eNOS signaling in microvascular permeability

Sarah Y. Yuan

since the original discovery that nitric oxide (NO) is the endothelium-derived relaxing factor, extensive investigation has been carried out to study the role of this signaling molecule in control of vascular homeostasis. We now know that endothelial NO synthase (eNOS) is the primary physiological source of NO in the vascular system, where exposure of the endothelium to physical stress or circulating factors causes eNOS activation and NO production. Such responses often occur via a calcium/calmodulin-dependent transduction pathway, and the downstream effects involve the formation of cGMP (21). The posttranslational regulation of eNOS is mediated by fatty acid modification and phosphorylation of the enzyme, as well as protein-protein interactions with other effector molecules (28). Furthermore, the intracellular localization of eNOS may be an important determinant of its activity (29). Mechanisms of eNOS trafficking between different subcellular compartments represent an active area of investigation.

When compared with the vasodilatory response, the effect of eNOS on microvascular permeability is less characterized. Tremendous controversy revolves around the question as to whether eNOS-derived NO acts as a permeability-increasing or -decreasing factor under physiological conditions and in disease states. The paradox emerged from the observations that manipulation of endothelial NO production could be both beneficial and detrimental to microvascular barrier properties. Back in the early 1990s, Hughes et al. (12) reported that intradermal injection of substance P caused skin edema that was inhibited by concurrent administration of the eNOS inhibitor NG-nitro-l-arginine methyl ester (l-NAME) or NG-monomethyl-l-arginine (l-NMMA), contributing to initial evidence for the possibility that endothelial NO acts as an edemagenic factor in vivo. The physiological importance of NO/cGMP formation in enhanced microvascular permeability became evident when Huxley's laboratory (18, 25) showed that a cGMP-dependent vasodilator increased the hydraulic conductivity of capillaries. Following this, Yuan and colleagues (38) demonstrated that flow was able to modulate coronary venular permeability via an eNOS-dependent mechanism. Subsequently, many studies have shown that exogenous NO donors induce leakage of water and macromolecules across the vascular endothelium, whereas inhibition of eNOS with pharmacological agents protects the barrier function against inflammatory injuries caused by histamine, bradykinin, platelet-activating factor (PAF), substance P, oxidants, calcium ionophore, ATP, and cytokines (3, 10, 13, 16, 22, 24, 39). In line with these results, studies by Granger's group (9) suggested that eNOS mediated VEGF-induced angiogenesis, an important microvascular process initiated with endothelial hyperpermeability. Further studies by this group and others elucidated a biphasic upregulation of eNOS by VEGF and receptor-mediated signaling cascades involving phospholipase C, calcium, and MAP kinases in NO-dependent hyperpermeability (2, 11, 3537). More recently, advances in transgene technologies resulted in the development of eNOS knockout mice; characterization of their phenotypes revealed an impaired vascular leakage response to systemic inflammatory and angiogenic stimuli (4, 7). On the basis of these findings, it was postulated that eNOS activation and subsequent NO production contributed to the physiological and pathophysiological regulation of microvascular transport as a permeability-increasing factor. However, this paradigm has been challenged by results that have led to the opposite conclusion that endothelial NO is a barrier-tightening factor. For example, superfusion of the mesenteric microcirculation with NO donors significantly reduced microvascular leakage during ischemia-reperfusion injury (15), and administration of either l-NAME or l-NMMA increased fluid and albumin leakage in microvessels (1, 14). Consistently, the same eNOS inhibitors were found to attenuate the barrier disruption effect of thrombin, bradykinin, and cytokines (5, 30, 34). In an attempt to characterize the ultrastructural feature of NO-dependent endothelial barrier properties, Malik's laboratory (23) conducted electron microscopic analyses of multiple vascular beds, including lung microvasculature, in eNOS null mice. They observed endothelial morphology characteristic of intercellular junction opening in the absence of eNOS expression and during treatment with eNOS inhibitors (23). These results provide compelling evidence supporting the argument that eNOS-derived NO is a permeability-decreasing factor that protects the microvascular barrier against inflammatory injuries.

Possible explanations for the discrepant effects of NO on microvascular permeability include the heterogeneity of microvasculature, different approaches used to measure barrier function, and varied experimental conditions (2, 19). Adding to the complexity is the multifunctional characteristics of eNOS signaling and its diverse cellular responses in the circulatory system. In addition to directly acting on the barrier structure, NO alters endothelial function by modulating platelet aggregation, neutrophil adhesion, and cytokine production (14, 31). It also interacts with the oxidative pathway, contributing to both the generation and scavenging of reactive oxygen species (33). All of these cellular and molecular responses can directly or indirectly affect the microcirculatory exchange process (17, 19). Furthermore, the NO-induced hemodynamic changes may confound the interpretation of data based on the indicator dilution technique or measurement of tracer transvascular flux, which can be affected by changes in the exchange surface area or microvascular pressure (2). As in the case of treatment with vasoactive agents, dilating upstream of the microcirculation would enhance its downstream perfusion with an increased hydrostatic pressure in exchange microvessels, promoting the transvascular flux of fluid and macromolecules. This is often interpreted as microvascular leakage; however, it does not necessarily reflect endothelial barrier dysfunction. Thus the hemodynamic effect of NO imposes a great challenge to discerning its specific contribution to microvascular permeability regulation.

In this issue of the American Journal of Physiology-Heart and Circulatory Physiology, Sánchez et al. (26) from Durán's laboratory provided new insights into the distinctive mechanisms of eNOS signaling for microvascular dilation and hyperpermeability. They used ECV-304 cells that were stably transfected with eNOS-GFP as a model to follow eNOS trafficking among different subcellular compartments in response to different vasoactive stimuli. A unique feature of the experimental design lies in the intelligent selection of ACh and PAF for comparative analysis of eNOS signaling. Both agents are known to activate eNOS and stimulate NO production; however, ACh induces vasodilation with a minimal effect on endothelial permeability, whereas PAF is a potent permeability-enhancing vasoconstrictor. Thus a comparison between the two eNOS stimuli enables differentiation of its hemodynamic and hyperpermeability effects. An interesting aspect of the results is that ACh translocated eNOS to the Golgi network, whereas PAF induced a preferential localization of the enzyme in the cytosol. The agonist-induced eNOS activation was confirmed by simultaneous Ser1177 phosphorylation and Thr497 dephosphorylation, consistent with the notion that eNOS activity is coordinately regulated by phosphorylation and dephosphorylation at the same residues (29). The effect appeared to be highly specific because inhibiting the muscarinic ACh receptor and the PAF receptor blocked the phosphorylation response to the respective stimulus. Interestingly, the unique pattern of eNOS compartmentalization did not seem to depend on the status of its phosphorylation, as both ACh and PAF modulated phosphorylation at the same residues to a similar extent. In addition, the in vivo evidence for PAF-stimulated eNOS cytosolic translocation in the hamster cheek pouch is of particular physiological relevance. Note that PAF has been shown to cause macromolecular leakage independent of vasomotor activities in the hamster cheek pouch microvessel (24).

What is the functional implication of differential eNOS translocation with respect to its distinctive contribution to vaosodilation and hyperpermeability? In endothelial cells under nonstimulated conditions, the majority of eNOS is associated with caveolae and lipid rafts in the plasma membrane. Caveolin-1 is a principal component of caveolae membrane that binds eNOS and inhibits its catalytic activity. Certain growth factors, lipid metabolites, and vasoactive agents can modulate eNOS activity by displacing the enzyme from caveolae to the nuclei or intracellular organelles, such as mitochondria and the Golgi apparatus (6, 21, 29, 40). Within this context, the study by Durán and colleagues (26) provides the first line of evidence for the ability of ACh and PAF to chaperone eNOS from lipid raft domains to the Golgi and cytosol, respectively. While the exact molecular consequences of such subcellular movement remain to be determined, one may reason that it is advantageous to remove the enzyme from its inhibitory calveolin-binding conformation and place it in close proximity to its downstream effectors. In this regard, a potential downstream target of NO signaling is the cytoskeleton-junction complex in endothelial cells. It has been established that physical stress and inflammatory agonists compromise endothelial barrier properties by inducing cytoskeletal contraction and junction disorganization (8, 17, 19, 37). The hyperpermeability response is characterized by sequestration of junctional molecules (e.g., vascular endothelial-cadherin and catenins) from the plasma membrane at cell-cell contacts and, in some cases, internalization to the cytosolic region (13, 32). Several cascades of signaling reactions have been described as underlying molecular events, including the activation of protein kinase C, protein kinase G, myosin light chain kinase, and MAP kinases (35, 37). Interestingly, all of these molecules are recognized as intracellular molecules capable of interacting with NO. Therefore, it is possible that translocation of eNOS to the cytosol represents an initial step of multiple intracellular signal transduction pathways that ultimately lead to cell contraction and junction opening. Alternatively, eNOS compartmentalization may have a direct effect on the structural components of the endothelial barrier. This is supported by recent electron micrographic evidence of defects in lung endothelial cell-cell junctions coupled with increased paracellular permeability in caveolin-1-deficient mice, and treatment with eNOS inhibitors, such as l-NAME, reversed the microvascular hyperpermeability caused by caveolin-1 deficiency (20, 27). Indeed, colocalization of eNOS with junctional proteins was documented in the study by Durán and colleagues (26). The authors suggest that the importance of eNOS location resides with the need for high local concentrations of NO to provide the correct signaling to the specific acceptor proteins. Thus eNOS translocation to the cytosol may serve the role of optimizing the paracrine effect of NO on the cytoskeletal and/or junctional structures to cause their conformational changes, conferring the hyperpermeability response to PAF. On the other hand, localization to the Golgi network, as occurs on ACh stimulation, implicates a different subcellular targeting unique to the vasodilatory action of eNOS.

In summary, eNOS-derived NO is involved in diverse cellular responses in the circulatory system. The effect of eNOS activation on microvascular transport varies depending on the type of stimulation, site of action, hemodynamic condition, and contribution of blood cells. Shear stress and certain eNOS activators can directly affect endothelial barrier properties, leading to microvascular hyperpermeability. This response is mediated by complex mechanisms involving protein-protein interactions in the endothelium at the subcellular level. Differential translocation of eNOS to specific subcellular compartments may serve as an important mechanism encoding for the distinctive microvascular responses to different vasoactive stimuli. The study by Durán and colleagues (26) demonstrates interesting data that have the potential to lead to a new explanation for eNOS signaling in the differential regulation of PAF-promoted vasopermeability and ACh-induced vasodilation. Further understanding of the molecular events dictating specific endothelial responses to particular stimuli may hold the key to successful treatment of various vascular diseases and injuries.