sickle cell disease (SCD) is an inherited condition involving the β-hemoglobin subunit in red blood cells (RBC), where a single β6-amino acid mutation results in glutamic acid being switched to valine. Under hypoxic and acidotic conditions, sickle hemoglobin (HbS) molecules polymerize and sickled RBCs are formed. These structurally altered RBCs are unable to traverse distal blood vessels, causing microvascular obstruction and subsequent tissue hypoxia and ischemia. The clinical presentation of this process manifests as a painful vaso-occlusive crisis (VOC), infarction, or acute chest syndrome (ACS), which has multifactorial inciting events that remain an area of investigation and debate. Although these processes are certainly important in SCD, the roles of inflammatory mediators and vascular dysfunction are also now recognized as significant in the pathophysiology of SCD. As evidence of this vascular inflammatory state, cell adhesion molecules on RBC and endothelial surfaces, such as VCAM-1 and ICAM-1, are upregulated (4, 35, 37, 39). Cytokines and reactive inflammatory mediators (e.g., free radicals, oxidants) are also produced in elevated amounts (15, 30, 38).
Nitric oxide (NO) is a multifaceted mediator of vascular homeostasis. NO inhibits the upregulation of cell adhesion molecules, platelet aggregation (1), and monocyte adhesion to endothelial cells (5). NO also mediates vasodilation through cGMP-dependent pathways and inhibits endothelin-1-induced vasoconstriction (12, 13). For lipid-derived radicals, NO acts as a beneficial free radical scavenger, and, upon reaction with the oxygen free radical superoxide (), the potent oxidizing and nitrating species peroxynitrite (ONOO-) is formed. In SCD, NO-mediated inhibition of adhesion molecule expression is impaired (39), as is NO-mediated vasodilation (2). Interestingly, many patients with SCD are also hypotensive (14, 29). Xanthine oxidase (XO) has been shown to be responsible for catalytic inactivation of NO signaling in a variety of vascular diseases, most recently SCD, by playing a role in the dysregulation of both NO-dependent vascular relaxation (2) and neutrophil function (41). XO is a rich source of , and hydrogen peroxide (H2O2) is particularly abundant in splanchnic tissues and endothelium and is released into the circulation following a variety of pathogenic events, including tissue ischemia/reperfusion (28). Once in the circulation, XO binds to and is transcytosed by vascular endothelium to anatomic locations where XO-derived reactive oxygen species contribute to impaired NO signaling via direct reaction with NO or following H2O2-mediated, peroxidase-dependent NO consumption (2, 6, 12). In support of this precept, XO has been shown to be increased in vessel walls and the plasma both clinically and in a transgenic, knockout murine model of SCD that expresses exclusively human HbS (2).
Diverse approaches have been employed in quantifying and modifying various aspects of NO signaling in SCD. Basal levels of the amino acid precursor of NO, l-arginine, are decreased in SCD and are suppressed even more so during complications, such as VOC and ACS (20, 23–25, 34). Both inducible and constitutive nitric oxide synthase (NOS) isoforms convert l-arginine to NO. Subsequent tissue reactions of NO predominantly yield nitrite () and nitrate (), with low, but potentially significant levels of nitroso (RNO) and nitro (RNO2) derivatives produced as well. Neuronal NOS is located throughout the lung and remains constitutively activated (7, 21, 42). In SCD, neuronal NOS displays polymorphisms that have been associated with ACS (40). Inducible NOS (iNOS) is located in most vascular cells, with its expression induced by the spectrum of inflammatory mediators associated with SCD. Although differing levels of iNOS expression and activity have been reported in SCD (3, 11, 27), our data support the view that iNOS is generally “upregulated” in multiple organs and vascular beds in SCD (3). Endothelial NOS (eNOS) is located in the vascular endothelium and manifests both increased expression and activity in SCD (11, 16). It is critical to understand that NO derived from iNOS and eNOS displays frequently unique but sometimes overlapping reactivities due to differences in their mechanisms of activation and intracellular compartmentalization. iNOS is induced by cytokines and functions independently of changes in calcium homeostasis. Because of this, it does not appear to play a role in receptor-dependent vasodilation but, rather, in more chronic inflammatory activation and hypotensive states (17). eNOS, on the other hand, is activated by calcium influxes and effects vasodilation (11, 17, 43). Plasma levels of the NO metabolites (NOx) and have been used as a measure of endothelium-dependent NO signaling, but a significant limitation becomes operative upon inflammatory induction of iNOS expression and activity (as in SCD). At this point, and levels become less reliable as a measure of the more salutary endothelium-dependent NO that is produced in response to vasodilatory stimuli and can actually reflect iNOS-derived NO and decomposition products of inflammatory oxidants such as nitrogen dioxide (·NO2) and ONOO- that can contribute to impaired vascular function (22). Nonetheless, the plasma levels of and are comparable in HbA (control) and HbS (SCD) patients under basal conditions but are decreased in SCD patients during VOC and ACS (18–20, 24, 26, 39) and are inversely related to pain scores during VOC (18).
Because eNOS expression and activity is increased in SCD, this enzyme has gained significant interest in the pathogenesis of SCD complications, such as ACS. Of special significance and the subject of a current article in focus (Ref. 33, see p. L705 in this issue), eNOS binds heat shock protein 90 (HSP90), leading to conformational changes in eNOS and enzyme activation (8, 9, 31, 32, 36). However, if these proteins interact without appropriate conformational changes, electron transfer reactions of eNOS become uncoupled, to yield eNOS-derived and H2O2, which concomitantly “inactivate” NO and generate secondary oxides of nitrogen such as ·NO2 and ONOO- (31, 32). Despite the important regulatory interactions of HSPs with eNOS, the biochemical and functional actions of HSP90 in SCD have not been previously evaluated.
To address this issue, a team of prominent vascular biology investigators led by Kirkwood Pritchard reveal novel insight into the actions of HSP90 and eNOS in acute lung injury in a murine model of SCD (33). This investigation first tested the hypothesis that the “oxidative stress” caused by hypoxia-induced increases in XO activity in pulmonary microvascular endothelial cells (PMVEC) leads to decreased association of HSP90 and eNOS. This in turn would impair NO production by PMVEC and, by extension of this concept to the clinical scenario, could contribute to the acute lung injury seen in SCD during ACS. The mice studied included a SCD model (with HbS, as well as evidence of mild to moderate β-thalassemia), a heterozygous SCD mouse (representing sickle trait but expresses mild pathological changes unlike sickle trait in human subjects), and wild-type control mice. These mouse models demonstrated vaso-occlusive events, as evidenced by increased vascular congestion, with these events observed in both SCD and SCD trait mice, but not in wild-type controls.
Several measurements revealed enhanced lung injury in this model of SCD. Quantitative immunohistochemical analysis of XO activity showed increased pulmonary XO present in SCD mice but not in trait or control mice under normoxic conditions. After hypoxic exposure, pulmonary XO was increased twofold in SCD mice, only minimally in trait mice, and was unchanged in controls. Pulmonary 3-nitrotyrosine (NO2Tyr) content, an index of the abrogation of salutary signaling actions of NO [e.g., NO-dependent oxidative inflammatory reactions (10)], was also evaluated by immunohistochemical staining, with NO2Tyr levels in SCD mice 2.5 times greater than trait or control mice. Hypoxic conditions led to ∼50% increases in lung NO2Tyr content in SCD mice, compared with only modest increases in trait and control mice. Finally, the investigators evaluated guanylate cyclase-dependent NO signaling activity by measuring lung tissue cGMP content, an index that was increased under basal conditions only in SCD mice. Interestingly, following hypoxic exposure of mice, cGMP content was significantly decreased from basal conditions in all mice to the same level.
From these results, the authors concluded that hypoxic conditions in SCD induce oxidant-mediated lung injury and impaired NO signaling, as evidenced by increased XO activity, lung NO2Tyr content, and the suppression of cGMP levels under hypoxic conditions. The role of XO in the pathogenesis of SCD is thus gaining increased relevance, especially in light of the recent expanding literature supporting the contributions of XO and the benefits of allopurinol (an XO inhibitor) in a variety of clinical vasculopathies including heart failure, hypertension, and atherosclerosis. The conclusion that pulmonary vascular NO production and signaling have been impaired seems reasonable but is complicated by 1) the similar cGMP contents measured in the posthypoxic lung tissue of the different animal groups and 2) the discrepant clinical measurements of plasma and levels in HbA and HbS patients and the question of how to interpret the meaning of these measurements in terms of guanylate cyclase activation. In any event, the authors' measurement of increased lung NO2Tyr content affirms increased NO production and/or its secondary oxidative reactions. In the author's defense, the similar cGMP levels observed in the different posthypoxia experimental groups may also be due to NO-independent effects on cGMP levels or changes in the content of this mediator in nonvascular tissues.
A novel and important element of the results of Pritchard et al. (33) is that HSP90-eNOS interactions are diminished in SCD, compared with wild-type control mice, even under normoxic conditions. After a hypoxic episode, the SCD mice revealed an even greater decrease in pulmonary tissue HSP90-eNOS interactions, with no changes observed in wild-type controls. The investigators further explored mechanisms underlying altered HSP90-eNOS interactions by evaluating the impact of a stimulus of eNOS activity on PMVEC that had been subjected to oxidants derived from XO plus its substrate xanthine. In support of in vivo observations, there was an XO-dependent decrease in the association of HSP90 with eNOS, as well as decreased cell and production. The authors conclude that the oxidative stress induced by elevations in lung tissue XO impairs the ability of HSP90 and eNOS to associate, leading to a decrease in NO generation. We must be cautious when making this final inference, however, because of the difficulty in interpreting NOx levels as a measure of only eNOS-derived NO.
As with every seminal research advance, provocative new questions arise. A crucial aspect of this and other studies that needs further definition is the anatomic source of the elevated XO observed in the vasculature in SCD. Clinical and animal model studies of XO inhibition and cell biological studies such as those utilized in the present report will assist in resolving the clinical significance of the observations made by Pritchard and colleagues (33). It is possible that either enhanced lung parenchymal XO expression or the expression and/or remote organ release of XO, followed by the pulmonary uptake of circulating XO, is the source of this increased oxidative injury. Addressing the question of whether increases in other sources of oxidant production contribute to altered eNOS activity in SCD will also be illuminating and important from a therapeutic perspective. This increased inflammatory oxidant “burden” not only could contribute to altered interactions between eNOS and HSP but could also have an impact on other eNOS regulatory interactions (e.g., caveolar vs. cytoplasmic distribution). In addition, HSP90 also interacts with and enhances iNOS activity (44). It might be possible that HSP90 association with eNOS is decreased in part because of increased oxidant-induced HSP90 interactions with iNOS. By resolving the relative contributions of eNOS- and iNOS-derived NO to both vascular functional defects and inflammatory-oxidant-related actions (or some potentially regulable combination of both sequelae), we will gain better insight into the pathogenesis of SCD.
The authors and these editorialists are hopeful that this report will stimulate continued interest in the significance of oxidative stress and HSP90-eNOS interactions in SCD, especially as they apply to the clinical setting. Questions do arise regarding the applicability of some SCD mouse model-based findings to humans with SCD. For example, vaso-occlusive events were reported to occur in larger vessels in the murine SCD model, although in humans vaso-occlusion occurs in the microvascular environment. Also, the heterozygous sickle trait mouse showed evidence of vaso-occlusion, whereas humans with sickle cell trait are clinically asymptomatic. Regardless, this report sheds light on an important area of focus in developing new understanding regarding the pathogenesis of SCD. It is also entirely possible that new diagnostic and therapeutic strategies will emanate from the observations of Pritchard et. al. (33). Quantification of HSP90-eNOS interactions could be predictive of the severity of ACS or the potential onset of painful crises and the need for more aggressive clinical management. Could these interactions be affected by XO inhibitors, such as the pyrazolo derivatives oxypurinol and allopurinol, or an NO source, such as inhaled NO, l-arginine, or hydroxyurea, thus altering the clinical course of SCD?
In summary, Pritchard and colleagues have elegantly established that HSP90 displays decreased interactions with eNOS in SCD. This finding is significant in that HSP90 enhances eNOS function, and in SCD, NOS function and NO pathways are dysregulated. A better understanding of the altered interactions between HSP90 and eNOS in SCD may hold the key to successful predictive and therapeutic modalities for ACS.
- Copyright © 2004 the American Physiological Society