Reactive oxygen species (ROS) are a class of oxygen-derived molecules including superoxide anion and hydrogen peroxide, both of which are modulators of vascular tone. Vascular smooth muscle cells, endothelium and perivascular adipose tissue are known to contain ROS 18 .

Superoxide anions can induce vasoconstriction through Ca²⁺ sensitization pathways, although it is not clear whether they act directly or via conversion to hydrogen peroxide 19 . Furthermore, contraction in response to perivascular nerve activation by electrical field stimulation is enhanced by superoxide anions from perivascular adipose tissue 18 .

Hydrogen peroxide is a more likely paracrine ROS because hydrogen peroxide is not a free radical and therefore more stable and less reactive with other tissue radicals 20 . Hydrogen peroxide is known to induce both vasorelaxation and vasoconstriction, depending on species, type of vascular bed, concentration, membrane potential and degree of obesity 20 21 22 . Vasorelaxation is possibly induced by endothelium-dependent mechanisms involving the release of vasodilating cyclooxygenase metabolites 23 and NO 24 , as well as endothelium-independent mechanisms 21 mediated by the activation of different potassium channels on smooth muscle cells 23 25 26 . On the other hand, vasoconstriction by hydrogen peroxide is likely induced in a Ca²⁺-dependent way, although Ca²⁺ sensitization and Ca²⁺-independent pathways have also been reported 20 24 27 . Furthermore, hydroxyl radicals, cyclooxygenase metabolites, protein kinase C, phospholipase A₂ phospholipase C and tyrosine kinase appear to play a role in hydrogen peroxide-induced contractions 27 .

Oxidative stress occurs when the production of ROS exceeds the cell's capacity to detoxify these potentially injurious oxidants using antioxidant defense systems 28 . In general, superoxide and hydrogen peroxide production in adipose tissue is increased in obese mice, which promotes endothelial dysfunction. Superoxide anions impair endothelium-dependent vasorelaxation by decreasing NO bioavailability via the formation of peroxynitrite, which is in turn another ROS 29 . Furthermore, ROS contributes to endothelial dysfunction by upregulating the expression of adhesion and chemotactic molecules in endothelial cells, which promote monocyte adhesion and migration to the vessel wall 28 . The adhesion of these circulating blood cells to vascular endothelium is a key element in the development of inflammation and thrombosis within the vasculature in vascular diseases associated with oxidative stress, such as atherosclerosis 28 .

Leptin is almost exclusively secreted by white and brown adipocytes 30 . Under normal conditions, leptin contributes to blood pressure homeostasis by its vasorelaxing and vasocontractile effects 31 32 . While the contractile effect of leptin is attributed to sympathetic nervous system activation 31 , various mechanisms seem to be responsible for leptin-induced vasorelaxation. This latter effect can be endothelium-dependent, either through the release of NO 33 or by other mechanisms 32 34 . The involvement of the endothelium-derived hyperpolarizing factor (EDHF) in leptin-induced vasorelaxation remains controversial 32 35 . It has been postulated that epoxyeicosatrienoic acids (EETs) and/or EDHF-dependent vasorelaxation in vivo might act as a backup in case of reduced NO availability 36 . On the other hand, EETs are able to activate endothelial NO synthase and subsequently release NO to influence arterial tone 37 . There is also evidence that leptin affects vascular tone without endothelial involvement 38 . A study on endothelium-denuded rat aortic rings showed that leptin attenuated angiotensin II (Ang II)-induced contraction by inhibiting Ca²⁺ release from the intracellular stores in vascular smooth muscle cells 39 .

Leptin levels are markedly increased in obesity 22 40 . Hyperleptinemia in obesity is believed to dysregulate blood pressure, resulting in hypertension. Significant associations have been found between plasma leptin levels and hypertension in both males and females, which makes leptin a potential predictor of hypertension 41 42 . In obesity, endothelium-dependent vasorelaxation is likely to become less effective, as sustained hyperleptinemia leads to endothelial dysfunction 43 . This might be the result of a leptin-induced increase of vasoconstrictor endothelin-1 44 , a leptin-induced expression of endothelin type A receptors in vascular smooth muscle cells 45 , a leptin-induced depletion of NO and an increase of cytotoxic ROS 46 . Leptin also promotes smooth muscle cell proliferation, contributing to the increased peripheral vascular resistance 47 . Furthermore, it stimulates the release of proinflammatory cytokines from macrophages, which may further elevate blood pressure and exacerbate the inflammatory process 48 .

The cytokine TNFα is a potent, time-dependent vasoconstrictor 49 50 and vasodilator 51 52 53 54 . Besides time dependency, it is unclear what underlies the differential regulation of arterial contractility by TNFα. Vasoregulatory actions of TNFα may be vascular bed-specific. Also, differences in experimental protocols used may explain the diversity of observations reported in various studies.

A source of TNFα that has recently been identified is perivascular adipose tissue 55 . This implies that TNFα is produced in the direct vicinity of the vascular endothelium. TNFα-mediated vasoregulation can occur through both endothelium-dependent 52 and endothelium-independent mechanisms 53 . Some studies have suggested that TNFα promotes vasorelaxation by an increase of NO and prostaglandin production 52 54 , while another study has suggested the involvement of hydrogen peroxide 56 .

On the other hand, TNFα is able to induce vasoconstriction by increasing endothelin-1 57 and angiotensinogen levels 58 . In addition, TNFα impairs endothelium-dependent vasorelaxation in various vascular beds as a result of a decrease in endothelial NO release or an increase in NO scavengers such as ROS 50 . Moreover, a recent study has shown a reduced vasorelaxing effect of perivascular adipose tissue in response to TNFα and IL-6, which upregulate ROS 59 .

Increased adipose tissue expression of TNFα mRNA has been reported in different rodent models of obesity as well as in clinical studies involving obese patients 40 . TNFα is considered a molecule that links inflammation to obesity 40 . Moreover, the infiltration of macrophages in adipose tissue during obesity contributes to increased TNFα production 60 . The increase in TNFα expression induces the production of ROS, resulting in endothelial dysfunction in obesity and obesity-related disorders such as hypertension, atherosclerosis and type 2 diabetes 61 . Furthermore, TNFα decreases adiponectin expression 62 and stimulates the secretion of proinflammatory proteins (for example, IL-6), which contribute to the maintenance of the chronic inflammatory state of adipose tissue in obesity 63 .

A sustained increase in proinflammatory cytokine IL-6 plasma levels is associated with high blood pressure 64 65 . On the other hand, acute exposure of IL-6 in vitro relaxes the aorta 66 . This vasorelaxing effect is likely regulated by an endothelium-independent pathway involving an increase in prostacyclin in vascular smooth muscle cells. IL-6 also relaxes skeletal muscle resistance vessels. However, this occurs only in vivo, suggesting that IL-6 interacts with parenchymal or intravascular factors to elicit vasorelaxation 67 .

In obesity, an increase in cytokine IL-6 has been observed at the mRNA and protein levels in white adipose tissue 68 69 . IL-6 has been shown to be a predictor of future myocardial infarction 65 and is highly associated with cardiovascular mortality 70 . IL-6 induces the induction of hepatic C-reactive protein (CRP) production, which is now known to be an independent major risk factor for cardiovascular complications 40 . Some studies have suggested that IL-6 is rather an indirect marker of vascular dysfunction, while others have suggested a more active role in vascular dysfunction 71 . Long-term elevation of IL-6 in mice has been shown to impair endothelial function by increasing angiotensin II-stimulated production of ROS as well as by reducing endothelial NO synthase mRNA expression 72 . In addition, IL-6 enhances vascular smooth muscle cell proliferation 73 , which is a key event in the genesis of atherosclerotic lesions.

Genetic deletion of IL-6 attenuates angiotensin II-induced hypertension in mice 64 , suggesting that elevated IL-6 in obesity might contribute to hypertension via Ang II. In addition, IL-6 inhibits adiponectin gene expression in cultured adipocytes 68 , which may exacerbate obesity-related hypertension.

Apelin, of which different isoforms exist, acts through the binding to a specific G protein-coupled receptor named APJ 74 , which is present on endothelial cells, vascular smooth muscle cells and cardiomyocytes 75 . Apelin causes NO-dependent vasorelaxation of human arteries both in vitro and in vivo 76 77 . In vivo exogenous apelin administration has been shown to cause a rapid NO-dependent fall in blood pressure in a rodent model, confirming its powerful vasorelaxing effect 78 . However, some reports have associated apelin with an increase in arterial pressure 79 . It has been proposed that apelin-induced changes in blood pressure (that is, an increase or decrease) are both dose- and time-dependent 74 . Furthermore, it is also possible that the observed bioactivity of apelin varies depending on species and/or vascular bed. Other data also suggest that apelin has vasoconstrictive potential by acting directly on vascular smooth muscle cells. In endothelium-denuded isolated human veins, apelin has been shown to be a potent vasoconstrictor with nanomolar potency and a maximum response comparable to that of Ang II 80 . In the presence of functional endothelium, this vasoconstrictive effect may be counterbalanced or even masked by activation of APJ receptors on vascular endothelial cells, resulting in the release of endothelial vasodilator substances such as NO 81 . All of these data taken together suggest a role for the apelin-APJ system as a regulator of vascular tone.

Apelin production in adipose tissue is strongly upregulated by insulin, and plasma concentrations are increased in obese and hyperinsulinemic mice and humans 82 . In contrast to acute exposure, long-term exposure of apelin does not affect blood pressure 83 , which might be explained by resistance to its hypotensive effect. This is in contrast to a study in which high apelin levels were found to increase blood pressure in obesity via stimulation of sympathetic outflow in the central nervous system when crossing the blood-brain barrier 84 .

In atherosclerosis, apelin might have beneficial effects, as apelin has been shown to stimulate endothelial NO production and antagonize the Ang II-induced formation of atherosclerotic lesions and aortic aneurysms in a murine model of atherosclerosis 85 .

Adiponectin is mainly released by both brown 86 and white 69 adipocytes and is the most abundant adipokine in the circulation 87 . Adiponectin has been considered an anti-inflammatory and antioxidative adipokine that protects against cardiovascular disease 88 . Adiponectin inhibits TNFα production and other inflammatory pathways in adipocytes and macrophages 40 88 . Plasma adiponectin has been correlated with endothelium-dependent vasorelaxation in humans 89 . These results were confirmed by other studies that have shown an increase in NO production as well as NO-mediated and potassium channel-mediated (that is, voltage-dependent) vasorelaxation in rats by adiponectin 59 90 91 . NO release from the endothelium is likely stimulated by adiponectin's binding to either the adiponectin type 2 receptor or T-cadherin on the endothelial surface 59 . Increased NO production inhibits platelet aggregation, leucocyte adhesion to endothelial cells and vascular smooth muscle cell proliferation. Furthermore, it reduces oxidative stress by decreasing ROS production in endothelial cells. All of these effects protect the vascular system against endothelial dysfunction 88 .

The use of an adiponectin receptor 1-blocking peptide has been found to abolish the vasorelaxing effect of human perivascular adipose tissue 59 . However, vasorelaxation induced by perivascular adipose tissue remained unchanged in adiponectin gene-deficient mice 91 . It is possible that this vasorelaxing effect of perivascular adipose tissue in the adiponectin gene-deficient mice might be the result of an endothelium-independent pathway 92 . Despite the latter findings, adiponectin remains an important vasoactive regulator.

Many studies on obesity-related diseases (for example, type 2 diabetes and hypertension) 40 93 , but not all 22 94 , have reported an overall decrease in adiponectin levels. Hypoadiponectinemia causes endothelial dysfunction by increasing superoxide anion production 95 by promoting the production of adhesion molecules in endothelial cells and the proliferation of smooth muscle cells 96 . Low adiponectin levels have recently emerged as an independent predictor of early atherosclerosis in obese patients 96 . However, after the establishment of atherosclerosis, this association may become weaker, especially in the presence of conditions inducing a hypercatabolic state (such as heart or renal failure) which are associated with increased plasma adiponectin, accelerated progression of atherosclerosis and worse clinical outcome 88 . In fact, several data show that high circulating adiponectin levels are associated with increased cardiovascular mortality in patients with coronary artery disease 88 . Therefore, hypoadiponectinemia may have clinical value at the early stages of atherogenesis, but at more advanced disease stages its role as a meaningful biomarker is questionable.

Although whether low levels of adiponectin predict hypertension remains controversial 42 97 and whether adiponectin levels in hypertension are decreased 87 98 , low adiponectin levels might contribute to the pathogenesis of obesity-related hypertension. Considering all the beneficial effects of adiponectin on the vascular system, an antihypertensive therapy which increases adiponectin levels could be of great value. It has already been demonstrated in obese adiponectin-knockout mice with hypertension that adiponectin replenishment lowers elevated blood pressure 99 . Existing drugs such as peroxisome proliferator-activated receptor γ agonists (thiazolidinediones), some angiotensin type 1 receptor blockers (telmisartan), angiotensin-converting enzyme inhibitors and cannabinoid type 1 receptor blockers (rimonabant and taranabant) have been shown to increase circulating adiponectin levels 88 . However, future strategies should focus on upregulation of adiponectin expression (and/or its receptors) or on targeting adiponectin receptors through the development of specific agonists.

Omentin is a recently identified adipose tissue-derived cytokine consisting of 313 amino acids and is mainly expressed in visceral rather than in subcutaneous adipose tissue 100 . Omentin consists of two isoforms in which omentin-1 appears to be the major isoform in human plasma 101 . Furthermore, higher plasma omentin-1 levels were detected in women compared with men 101 . In isolated rat aorta, omentin directly induces an endothelium-dependent relaxation which is mediated by NO. Omentin is even capable of inducing vasorelaxation in an endothelium-independent way. Omentin-induced relaxation is also observed in isolated rat mesenteric arteries, indicating the effectiveness of omentin in resistance vessels 100 . Since only in vitro studies on isolated blood vessels have been performed, in vivo studies are necessary to explore the influence of omentin on blood pressure and its chronic influence on vascular reactivity.

Very little is known about this novel protein in obesity. What is known is that omentin plasma levels and adipose tissue gene expression are decreased in obesity 101 and even more when overweight is combined with type 2 diabetes 102 . Furthermore, decreased omentin-1 levels are associated with low plasma adiponectin and high-density lipoprotein levels. In addition, omentin-1 levels are negatively correlated with leptin levels, waist circumference, body mass index and insulin resistance 101 . Like adiponectin, circulating omentin-1 concentrations increase after weight loss-induced improvement of insulin sensitivity 103 . Although further research is necessary, elevating omentin levels might be of interesting therapeutic value in obesity and obesity-related disorders.

Visfatin is another novel identified cytokine which is released from perivascular and visceral adipose tissue and which has an insulin-mimetic effect 104 105 . Visfatin has multiple functions in the vasculature. It stimulates growth of vascular smooth muscle cells 106 and endothelial angiogenesis via upregulating VEGF and matrix metalloproteinases 104 . Visfatin can also directly affect vascular contractility. Visfatin has been shown to induce endothelium-dependent vasorelaxation in rat isolated aorta through NO production. Also, in mesenteric arteries of rats, visfatin induces relaxation, suggesting that visfatin is effective in resistance vessels 107 . Because only acute effects of visfatin have been demonstrated, further studies are necessary to explore the chronic influence of visfatin on vascular reactivity.

Most studies, but not all, have shown an increase in visfatin levels in obesity 105 108 . A relationship of plasma visfatin levels was seen with body mass index and percentage of body fat, but not with abdominal circumference or visceral fat estimated on the basis of computed tomography 108 . It has been reported that the expression of visfatin is high at plaque rupture sites in patients with coronary artery disease 109 . Visfatin accelerates monocyte adhesion to endothelial cells by upregulating intercellular cell adhesion molecule-1 and vascular cell adhesion molecule (VCAM)-1 in vascular endothelial cells due to ROS overproduction, suggesting a possible role for visfatin in the development of atherosclerosis 110 . Further studies are necessary to clarify the atherogenic and vasoactive effects of visfatin and its potential clinical relevance.

Vascular tone can also be regulated by an unknown ADRF which is released from perivascular adipose tissue. Soltis and Cassis 111 first described that the presence of perivascular adipose tissue reduced vascular contractions by norepinephrine in rat aorta, which was later confirmed by Löhn et al. 6 . Also, isolated adipose tissue and cultured rat adipocytes relaxed precontracted rat aorta previously cleaned of adherent adipose tissue. This modulatory effect was attributed to ADRF, which functions as a regulator of arterial tone by active antagonism of contraction 6 . A similar vasorelaxing effect of perivascular adipose tissue was observed in rat mesenteric arteries 112 , in mouse aorta 113 and in human internal thoracic arteries 114 . These data suggest a common pathway for arterial tone regulation in different species and different types of vascular structures. Verlohren et al. 112 even showed a positive correlation between the vasorelaxing influence of ADRF and the amount of perivascular adipose tissue. The observation that the resting membrane potential of vascular smooth muscle cells in arteries with adipose tissue is more hyperpolarized than in arteries without adipose tissue, further supports the idea that perivascular adipose tissue actively contributes to basal arterial tone 112 . Whether NO formation and endothelium are involved in the vasorelaxation effect of ADRF is still a matter of debate 6 92 112 . On the other hand, the vasorelaxing effect of ADRF is likely mediated by the opening of different K⁺ channels in vascular smooth muscle cells, depending on the tissue and species studied 6 92 112 114 115 . These divergent observations suggest a different distribution of K⁺ channels in different vessels and/or species or the existence of different ADRFs.

More and more evidence is accumulating in support of the existence of different ADRFs. Löhn et al. 6 first suggested that ADRF is a protein. Furthermore, analyses of adipose tissue secretion in a recent electrophoresis study resulted in the visualization of different protein bands with different molecular masses (13.8 to 74.0 kDa), which may include ADRF 116 . A possible candidate is peptide angiotensin 1 2 3 4 5 6 7 , which is a vasodilator located within adipose tissue surrounding rat aorta 117 . Blocking this particular peptide inhibits the vasorelaxing effect of perivascular adipose tissue surrounding rat aorta 117 . This hypothesis is contradicted, however, by the fact that certain ADRF-related potassium channels (K_ATP or K_v) 6 115 are not involved in this observed vasorelaxing effect. In addition to proteins, hydrogen peroxide produced from the NAD(P)H oxidase in adipocytes has been described as being involved in the endothelium-independent pathway of ADRF 92 . Also hydrogen sulphide has been proposed as a novel candidate of ADRF or at least as a mediator in the ADRF effect 115 118 , which is consistent with inactivation of ADRF by heating (65°C for 10 minutes) 6 . Hydrogen sulphide has recently been described as a gasotransmitter generated by cystathionine γ-lyase (CSE) in perivascular adipose tissue 119 120 . Blocking of CSE inhibits the vasorelaxing effect of perivascular adipose tissue in rat aorta and mouse mesenteric arteries 115 118 . Moreover, hydrogen sulphide-induced vasorelaxation of rat aorta was inhibited by a particular ADRF-related potassium channel (KCNQ) blocker 115 . However, hydrogen sulphide generation and CSE expression in the perivascular adipose tissue of stenotic aortas (but not in aortic tissue) have been shown to be increased in rat hypertension induced by abdominal aortic banding 118 , while the vasorelaxing effect of perivascular adipose tissue has been shown to be impaired in spontaneously hypertensive rats 121 . This might indicate that ADRFs other than hydrogen sulphide are impaired, resulting in a reduced vasorelaxing effect of adipose tissue. On the other hand, it is difficult to compare both studies, as they used different models of hypertension. Furthermore, the upregulation of CSE and hydrogen sulphide generation in perivascular adipose tissue of stenotic aortas may have developed independently of hypertension, as CSE-knockout mice have been shown to be hypertensive 120 .

Obesity is characterized by a decrease in the vasorelaxing effect of perivascular adipose tissue, leading to hypertension 22 59 91 122 . This might imply a decrease in ADRF release or an imbalance in adipose tissue-derived relaxing and vasocontractile factors during obesity. On the other hand, hypoxia, which develops within adipose tissue during obesity 12 , has recently been shown to enhance the release of vasorelaxing factors released from adipose tissue, which might implicate ADRF 123 . So, the release of ADRF in obesity warrants further research.

Brown and white adipocytes are rich sources of angiotensinogen, the precursor protein of a major vasocontractile peptide called Ang II 124 , and possess all the enzymes necessary to produce Ang II 125 . These findings suggest the existence of a local renin-angiotensin system in adipose tissue. Moreover, the amount of angiotensinogen mRNA in adipose tissue is 68% of that in the liver, supporting an important role for adipose angiotensinogen in Ang II production 126 . The importance of this angiotensinogen source in blood pressure regulation by the renin-angiotensin system was shown in wild-type and angiotensinogen-deficient mice in which adipocyte-derived angiotensinogen was overexpressed. When angiotensinogen expression was restricted to adipose tissue (in an angiotensinogen-deficient background), circulating angiotensinogen was detected and the mice were normotensive. On the other hand, wild-type mice were hypertensive because of the additional amount of angiotensinogen that developed as a result of overexpression of adipocyte-derived angiotensinogen 127 .

An important effect of Ang II is that this peptide enhances the metabolism of NO into oxygen free radicals, which damage the vascular tissue 128 . Therefore, an imbalance between Ang II and NO leads to endothelial dysfunction, resulting in a loss of vasodilator capacity. This results in an increased expression of adhesion molecules and proinflammatory cytokines in endothelial cells, which promotes monocyte and leukocyte adhesion and migration to the vessel wall 129 . Furthermore, Ang II exerts detrimental effects on the progression and destabilization of atherosclerotic plaque because of an increased release of PAI-1, causing thrombosis and increased expression of growth factors, which leads to smooth muscle cell proliferation and migration 129 . Most data support an elevation of angiotensinogen mRNA expression in adipose tissue during obesity 130 . Furthermore, several studies have highlighted a contribution of adipose tissue-derived angiotensinogen and/or angiotensin II to obesity-related hypertension 130 . High Ang II levels may deteriorate obesity-related hypertension because of an increased secretion of proinflammatory cytokines 131 , decreased adiponectin secretion 132 and increased leptin production in adipocytes 133 .

Resistin, which is expressed in brown and white adipose tissue, is a member of the family of cysteine-rich proteins called resistin-like molecules 86 134 . Resistin is secreted into the medium by cultured adipocytes and circulates in plasma, indicating that it is a secretory product of adipose tissue. However, circulating monocytes and macrophages in particular seem to be responsible for resistin production in humans 40 . Although resistin does not directly affect the contractility of isolated blood vessels 135 , coronary blood flow, mean arterial pressure or heart rate 136 , it has been associated with endothelial dysfunction and coronary artery disease 137 .

Initial findings have been reported regarding an association between obesity and elevated plasma resistin levels 138 139 . However, this finding was not confirmed by other investigators 140 141 . Resistin expression is stimulated by TNFα and IL-6, both of which are increased in obesity 142 , which offers an explanation for an increased level of resistin in obesity. Resistin augments endothelin-1 release, which causes endothelial dysfunction. Moreover, resistin impairs endothelial function with 143 or without 136 augmenting superoxide production, resulting in decreased expression of endothelial NO synthase and NO levels 144 . Resistin also augments the expression of VCAM-1 and MCP-1, both of which are involved in early atherosclerotic lesion formation 145 . It has also been shown that high plasma resistin levels are independently associated with an increased risk for hypertension among nondiabetic women 146 .