Introduction

Over the last 2 decades it has become evident that the endothelium is not an inert single-cell lining covering the internal surface of blood vessels, but plays a crucial role in vascular homeostasis, i.e., by regulating vascular tone and structure. Furthermore, a healthy endothelium inhibits platelet and leukocyte adhesion to the vascular surface and keeps a balance between profibrinolytic and prothrombotic activity (Fig. 1) [1]. Common conditions predisposing to atherosclerosis, such as hypercholesterolemia, hypertension, diabetes, and smoking, are associated with endothelial dysfunction leading to a proinflammatory and prothrombotic phenotype of the endothelium. Endothelial dysfunction may therefore play a pivotal role in the development and progression of atherosclerosis and its clinical complications [1–3].

Fig. 1
figure 1

The healthy endothelium not only mediates endothelium-dependent vasodilation, but also actively suppresses thrombosis, vascular inflammation, and hypertrophy. Nitric oxide (NO) is a particularly important mediator of endothelium-dependent vasodilation as well as anti-inflammatory and antithrombotic effects of the endothelium. Endothelium-dependent vasomotion is therefore thought to represent a surrogate marker of other important functions of the endothelium

Full size image

Endothelial function has mostly been assessed as endothelium-dependent vasomotion, largely based on the assumption that impaired endothelium-dependent vasomotion also reflects alterations of other functions of the endothelium. An important rationale for this approach has been the observation that endothelium-derived nitric oxide (NO), synthesized by the endothelial NO synthase (eNOS) from the precursor L-arginine, not only mediates endothelium-dependent vasodilation, but is also critically involved in the regulation of other protective properties of the healthy endothelium [4–7]. This may, at least in part, underlie the consistent observation of recent clinical studies, demonstrating that the degree of impairment of endothelial function (assessed as endothelium-dependent vasomotion) represents a strong and independent predictor of cardiovascular events in patients with coronary disease and its risk factors [8–11].

Vascular oxidant stress: a major cause of reduced endothelial NO availability

The impairment of endothelium-dependent, NO-mediated vasodilation in experimental models of cardiovascular disease, such as atherosclerosis and hypertension, is reversed after administration of superoxide dismutase (SOD) or other superoxide scavengers, suggesting a pivotal role of increased vascular superoxide production for endothelial dysfunction [12–15].

Numerous clinical studies have demonstrated that the acute administration of structurally unrelated antioxidants (i.e., high local dose of ascorbic acid or N-acetylcysteine) improves endothelium-dependent vasodilation in patients with coronary disease or cardiovascular risk factors, supporting the concept that increased vascular reactive oxygen species (ROS) formation plays an important role for endothelial dysfunction in patients at risk for cardiovascular events [10, 16–20]. An example from our laboratory is shown in Fig. 2a [18], indicating that administration of a high local dose of the antioxidant vitamin C reverses impaired NO-mediated vasodilation in patients with coronary disease. Importantly, Heitzer et al. [10] have recently demonstrated that the beneficial effect of vitamin C (high local dose) on endothelial dysfunction has profound and independent prognostic implications in patients with coronary disease (Fig. 2b). In a study on 179 patients with coronary disease it was shown that patients with a substantial improvement of endothelium-dependent vasodilation after vitamin C had a markedly higher cardiovascular event rate as compared to patients with limited effect of vitamin C on endothelial function [10], suggesting that increased vascular oxidant stress is associated with an impaired prognosis. These findings have further stimulated an interest in understanding the mechanisms causing increased vascular superoxide production and consequently a reduced endothelial NO bioavailability.

Fig. 2
figure 2

a Effect of the antioxidant vitamin C on endothelium-dependent, NO-mediated vasodilation in patients with coronary artery disease (CAD) and healthy control subjects. A high local dose of the antioxidant vitamin C (250 mg i.a. in 10 min) was administered. Adapted and modified from [18]. b Proportion of patients with CAD without cardiovascular events dependent on their initial response to vitamin C on endothelium-dependent vasodilation. Adapted and modified from [10]

Full size image

Mechanisms of vascular oxidant stress in cardiovascular disease

Antioxidant enzyme systems

Vascular cells are equipped with three isoenzymes of SOD that dismute superoxide anions to oxygen and hydrogen peroxide. These isozymes have different cellular localizations: the cytosolic Cu,Zn-SOD, the mitochondrial Mn-SOD, and the extracellular SOD (ecSOD), which is bound to the cellular surface [21]. Inhibition of vascular SOD causes a profound impairment of NO-mediated vasodilation, suggesting that vascular SOD activity is critical for endothelial NO availability [22–24]. We therefore analyzed SOD isoenzyme activity in coronary arteries from patients with coronary disease as compared to nonatherosclerotic arteries. In coronary arteries from patients with coronary disease there was no change of the activity of the intracellular SOD isoenzymes (Cu,Zn-SOD, and Mn-SOD); however, there was a marked reduction of the extracellular SOD activity in atherosclerotic coronary arteries [18]. Extracellular SOD, which is expressed at approximately 100-fold higher levels in the vascular wall as compared to other tissues such as muscle or fat tissue, likely has a special function in the vessel [25, 26]. Observations from mice lacking ecSOD have shown that ecSOD deficiency is associated with an impaired endothelium-dependent, NO-mediated vasomotion and a markedly augmented endothelial dysfunction in response to hypertension [27]. In order to determine whether reduced ecSOD activity in patients with coronary disease is related to endothelium-dependent, NO-mediated vasodilation we analyzed endothelium-bound ecSOD activity (released from the endothelium to plasma by heparin bolus injection) in patients with coronary disease in vivo [18]. ecSOD is bound to heparan sulfate on the endothelial cell surface and can therefore be rapidly released to plasma by heparin bolus injection [28]. In line with the observations in coronary arteries, endothelium-bound ecSOD activity was markedly reduced in patients with coronary disease as compared to healthy subjects and was closely related to endothelium-dependent, NO-mediated vasodilation, suggesting that reduced ecSOD activity contributes to reduced vascular NO availability (Fig. 3; [18]). A mutation of the ecSOD heparin-binding domain, which impairs binding of the enzyme to the endothelial surface, has recently been identified. Carriers of this ecSOD mutation had a significantly higher risk of developing coronary disease, as was shown in the Copenhagen City Heart Study by analyzing more than 9,000 subjects, suggesting a protective function of vascular-bound ecSOD [29].

Fig. 3
figure 3

a Endothelium-bound extracellular SOD (ecSOD) activity (released from endothelium to plasma by heparin bolus injection) in patients with coronary artery disease (CAD) and healthy control subjects, suggesting a marked reduction of endothelium-bound ecSOD activity in patients with CAD. b Relation of ecSOD activity and endothelium-dependent, NO-mediated vasodilation in patients with CAD. Adapted and modified from [18]

Full size image

Oxidant enzyme systems

Although there are numerous potential enzymatic sources of ROS in vascular cells, including the mitochondrial electron transport chain, the arachidonic acid metabolizing enzymes lipoxygenase and cyclooxygenase, the cytochrome P450s, NADPH oxidase, xanthine oxidase, and “uncoupled” NO synthase, the last three enzymes have been studied rather extensively over the last several years and likely play a predominant role in cardiovascular disease [30].

The NADPH oxidase, a superoxide producing enzyme, has been found to be expressed in all vascular cells, i.e., endothelial, vascular smooth muscle cells, fibroblasts. This enzyme consists of two membrane-integrated subunits (nox and p22phox) and three cytosolic subunits (p47phox, p67phox, and the small GTPase Rac) (Fig. 4a). There are several Nox isoforms that vary in their tissue distribution, cellular localization, and requirement for cytosolic factors. In vascular homogenates of hypercholesterolemic or hypertensive animals NADPH oxidase activity is increased [12, 31]. By using a p22phox-antisense or vascular smooth muscle and endothelial cells genetically deficient in p47phox it was shown that superoxide production in response to important stimuli such as angiotensin II depends on the activation of the NADPH oxidase [32–34]. Furthermore, mice that are genetically deficient in the cytosolic subunit p47phox had no increase in vascular oxidant stress in response to angiotensin II (in contrast to wild-type mice), suggesting that activation of the NADPH oxidase is critically important for the development of vascular oxidant stress in vivo [35, 36]. Importantly, the blood pressure response to angiotensin II or deoxycorticosterone acetate (DOCA)-salt hypertension is significantly blunted in p47phox-deficient mice [35, 37], suggesting an important role of NADPH oxidase activation for the development and progression of hypertension, likely at least partly related to a reduced endothelial NO availability. In this respect, Li et al. [38] have shown that p47phox and the vascular NADPH oxidase are critically involved in the reduction of endothelial NO availability in response to angiotensin II.

Fig. 4
figure 4

a Structure of the NADPH oxidase enzyme system, which is composed of two membranous subunits (nox and p22phox) and three cytosolic subunits (p47phox, p67phox, and the small GTPase Rac) that translocate to the membrane upon activation of the enzyme. For the membranous component nox several homologues have now been identified in vascular cells with different cellular localization (i.e., nox1, 2, and 4). b NADPH oxidase activity as determined by electron spin resonance (ESR) spectroscopy in human atherosclerotic and nonatherosclerotic coronary arteries, suggesting activation of the enzyme in human coronary disease. A representative ESR scan of NADPH-stimulated superoxide production from an atherosclerotic artery (CAD) and nonatherosclerotic artery (Control) is shown. Adapted and modified from [40]

Full size image

In patients with coronary disease the vascular expression of the NADPH oxidase subunit p22phox is increased in all vascular cells [39]. By using electron spin resonance spectroscopy we have recently demonstrated that the activity of the NADPH oxidase is markedly increased in coronary arteries from patients with coronary disease as compared to nonatherosclerotic coronary arteries, suggesting that this enzyme is activated in human coronary disease (Fig. 4b; [40]).

Another important source of superoxide in the endothelium that has received increasing attention is the uncoupled eNOS. When eNOS is uncoupled, the enzyme produces superoxide upon activation in addition to NO [41, 42]. In particular, a deficiency of the cofactor tetrahydrobiopterin promotes “uncoupling” of the enzyme [41, 42]. In experimental hypertension we have observed a markedly increased superoxide production derived from eNOS [37]. Importantly, vascular superoxide production was lower in eNOS-deficient mice as compared to wild-type mice with DOCA-salt hypertension. eNOS uncoupling was not observed, however, in p47phox-deficient mice, suggesting a redox-sensitive process underlying eNOS uncoupling in vivo [37]. Of note, vascular levels of tetrahydrobiopterin were markedly reduced in hypertensive wild-type mice, but not in hypertensive p47phox-deficient mice, suggesting that at least one mechanism whereby eNOS becomes uncoupled in vivo relates to NADPH oxidase-dependent oxidation of tetrahydrobiopterin. This implies that activation of the NADPH oxidase can further promote vascular oxidant stress by causing eNOS uncoupling that increases superoxide production and reduces NO formation by eNOS in the endothelium. These observations may partly explain why Barry-Lane et al. [43] have observed a reduction of atherosclerotic lesion formation in the aorta of apoE-deficient mice lacking p47phox. Several studies in patients with coronary disease or conditions predisposing to atherosclerosis have now demonstrated a beneficial effect of tetrahydrobiopterin administration on endothelium-dependent vasomotion, suggesting that eNOS uncoupling is importantly involved in the impairment of endothelial NO availability in human cardiovascular disease [42, 44]. Of note, overexpression of eNOS has augmented atherosclerotic lesion formation in apoE-deficient mice, an effect that could be corrected by tetrahydrobiopterin administration, strongly suggesting that eNOS uncoupling has the potential to promote atherosclerotic lesion formation [45].

Finally, xanthine oxidase has been suggested as an important oxidant enzyme system contributing to vascular oxidant stress in cardiovascular disease [46, 47]. In experimental atherosclerosis xanthine oxidase-derived superoxide has been found to be a major cause of endothelial dysfunction [48, 49]. In patients with coronary disease, coronary and endothelium-bound xanthine oxidase activity are increased and closely related to endothelium-dependent, NO-mediated vasodilation, suggesting that increased xanthine oxidase activity contributes to endothelial dysfunction in human cardiovascular disease (Fig. 5) [50]. Recent experimental data have suggested that endothelial expression of xanthine oxidase is regulated in a redox-sensitive way, dependent on the endothelial NADPH oxidase [50].

Fig. 5
figure 5

Potential mechanisms leading to increased vascular superoxide production in atherosclerosis. Increased vascular superoxide production likely relates to activation of the vascular NADPH oxidase enzyme system that can promote further superoxide production from uncoupled endothelial NO synthase (eNOS) and xanthine oxidase. In advanced atherosclerosis vascular extracellular SOD (ecSOD) activity is reduced, which may further enhance oxidant stress and thereby contribute to reduced vascular NO bioavailability

Full size image

Summary and conclusion

Accumulating evidence suggests that increased vascular oxidant stress represents a major cause of reduced endothelial NO bioavailability in experimental and clinical cardiovascular disease. The mechanisms leading to increased vascular production of ROS include activation of the vascular NADPH oxidase, and likely as a consequence increased superoxide formation from uncoupled eNOS and endothelium-bound xanthine oxidase (Fig. 5). In addition, in patients with advanced atherosclerosis there is a significant reduction of the vascular extracellular SOD activity that may further augment endothelial dysfunction. Increased oxidant stress, in particular superoxide production, may reduce endothelial NO availability not only by direct inactivation of endothelium-derived NO, but also by oxidation of tetrahydrobiopterin leading to eNOS uncoupling [37], and by inhibition of dimethylarginine dimethylaminodydrolase (DDAH), the enzyme that hydrolyzes the endogenous NOS inhibitor asymmetric dimethylarginine (ADMA) (Fig. 6) [51].

Fig. 6
figure 6

Mechanisms by means of which increased vascular reactive oxygen species (ROS) production may reduce the bioavailability of endothelium-derived NO. ROS may reduce endothelial NO availability at least via three pathways: (1) Direct inactivation of NO• by superoxide (O2 •−) resulting in loss of NO bioactivity. (2) Reduced eNOS activity due to increased levels of asymmetric dimethylarginine (ADMA), an endogenous eNOS inhibitor, resulting from redox-sensitive inhibition of dimethylarginine dimethylaminohydrolase (DDAH), the ADMA-metabolizing enzyme. (3) eNOS uncoupling at least in part due to increased oxidation of tetrahydrobiopterin (H4B), a critical cofactor of eNOS, resulting in reduced NO production and increased superoxide production from eNOS. These mechanisms may explain why changes of the endothelial redox state have a profound impact on endothelial NO availability

Full size image

Of note, whereas short-term administration of a high local dose of the antioxidant vitamin C improved endothelial function in numerous clinical studies, long-term oral administration of vitamin E/C appears not to be effective in improving coronary or peripheral endothelium-dependent vasomotion [52]. This observation may partly explain why administration of vitamin E in recent large-scale clinical trials had no beneficial effect in primary or secondary prevention [3].

In contrast, recent studies suggest that common treatment strategies that improve prognosis in patients with cardiovascular disease, in particular angiotensin-converting enzyme (ACE) inhibition and statin treatment, may have potent “antioxidant” effects by preventing activation of vascular oxidases, i.e., NADPH oxidase [31, 53–55], or by restoring antioxidant enzyme capacity, i.e., ecSOD [53, 56], that may importantly contribute to their therapeutic effect. A better understanding of the regulation and the pathophysiological importance of alterations of oxidant/antioxidant enzyme systems in cardiovascular disease may provide important novel insights into the pathophysiology of cardiovascular disease with a potential to improve prognosis, and may lead to novel therapeutic targets.