Endothelial dysfunction in insulin resistance and type 2 diabetes

Authors

  • P.-A. Jansson

    1. From the Lundberg Laboratory for Diabetes Research, Department of Molecular and Clinical Medicine/Diabetes, The Sahlgrenska Academy at Göteborg University, Blå Stråket 5, Sahlgrenska University Hospital, Sahlgrenska S-413 45, Göteborg, Sweden
    Search for more papers by this author

Per-Anders Jansson, MD, PhD, The Lundberg Laboratory for Diabetes Research, Department of Molecular and Clinical Medicine/Diabetes, The Sahlgrenska Academy at Göteborg University, Blå Stråket 5, Sahlgrenska University Hospital, Sahlgrenska S-413 45 Göteborg, Sweden.
(fax: 46 31 410573; e-mail: Per-anders.jansson@medic.gu.se).

Abstract.

Macrovascular disease is the number one killer in type 2 diabetes patients. The cluster of risk factors in the insulin resistance syndrome (IRS) partly explains this notion. Insulin action in muscle, liver or adipose tissue has been thoroughly described in the literature, whilst this has been less described for the endothelium. Insulin stimulates nitric oxide (NO) production in the endothelium and reduced bioavailability of NO is usually defined as endothelial dysfunction. This impairment might be related to defective insulin signalling in the endothelial cell. Therefore, insulin resistance mechanisms in the endothelial cell will be emphasized in this review.

Imbalance between the vasodilating agent NO and the vasoconstrictor endothelin-1 (ET-1) contributes to endothelial dysfunction. Different methods and circulating markers to assess endothelial function will be outlined. Circulating markers of an activated endothelium appear long before type 2 diabetes develops suggesting a unique role of the endothelium in the pathophysiology of the disease.

Hampered blood flow in nutritive capillaries due to endothelial dysfunction is coupled with decreased glucose uptake and hyperglycemia. The forearm model combined with muscle microdialysis enables us to measure interstitial glucose and an index for capillary recruitment, the permeability surface area (PS). Available data from this method suggest that capillary recruitment in response of insulin is impaired in insulin resistant human subjects.

Endothelial dysfunction

Basic concepts

Endothelial dysfunction refers to an imbalance in the release of vasodilating factors such as NO and prostacyclin (PGI2), and vasoconstricting factors such as ET-1 and angiotensin-II (AT-II) [1, 2]. When this balance is changed, it predisposes the endothelium towards an atherogenic milieu. By that time, this will result in rolling of leukocytes, smooth muscle growth, impaired coagulation, vascular inflammation, atherosclerosis and thrombosis [3]. In atherosclerotic coronary arteries vasodilation is impaired and even a paradoxical constriction has been reported, suggestive of endothelial dysfunction [4].

Either loss of NO biological activity or biosynthesis of NO is a central mechanism of endothelial dysfunction but the molecular basis of this condition is far from understood [2]. In a setting of low concentrations of cofactors or substrate, endothelial nitric oxide synthase (eNOS) may become uncoupled, resulting in the release of reactive oxygen species (ROS) as exemplified by superoxide anion and hydrogen peroxide. In a situation like this, defined as oxidative stress, there is an exaggerated generation of ROS, normally scavenged by different intra- and extracellular mechanisms. In the presence of high concentrations of ROS this system is insufficient, and NO may react with some ROS species to form peroxynitrite, in turn increasing the oxidative stress inside the cell. Several underlying conditions like hyperglycemia, dyslipidaemia, cigarette smoking, inflammation and insulin resistance may induce oxidative stress [5].

Endothelial dysfunction and type 2 diabetes

Endothelial dysfunction is a common finding in type 2 diabetic patients. Several independent factors, e.g. insulin resistance, hyperglycemia, hypertension, dyslipidaemia, abdominal obesity and low-grade inflammation, have all been associated with this condition in subjects with type 2 diabetes [6].

Insulin resistance.  A number of different altered metabolic states as exemplified by glucose, lipid and cytokine metabolism can lead to peripheral insulin resistance. The ensuing metabolic dysregulation that occurs as a consequence of insulin resistance further exacerbates its progression. Obesity and insulin resistance, independent of other risk factors, are associated with endothelial dysfunction. In the insulin resistant state the normal suppression of free fatty acids (FFA) release from adipose tissue is impaired contributing to diabetic dyslipidaemia i.e. VLDL-hypertriglyceridemia, low HDL-cholesterol concentrations and elevation of FFA [7]. Elevated circulating FFA and transient hypertriglyceridemia induce endothelial dysfunction in healthy subjects [8]. In insulin resistance and type 2 diabetes a state of pro-atherogenesis and low-grade-inflammation occurs (e.g. reflected by plasminogen activator inhibitor-1 (PAI-1), hematocrit, family history of type 2 diabetes, glucagon like peptide-1 (GLP-1), increased TNF-alpha, IL-6, and CRP-levels), all of which have been associated with endothelial dysfunction [6, 9–12].

Hyperglycemia.  Hyperglycemia is an important factor inducing endothelial dysfunction and several theories have emerged to explain the adverse effects of hyperglycemia on the endothelium, including the aldose reductase hypothesis, the advanced glycation end product hypothesis, hexosamine hypothesis, protein kinase C hypothesis, the hyperperfusion hypothesis and the oxidative stress hypothesis. All these hypotheses overlap each other but the common disturbance due to hyperglycemia most likely is oxidative stress [13]. Interestingly, hyperglycemia also induces endothelial dysfunction in healthy subjects. This was observed as an attenuated response to metacholine, but not to calcium channel blocker infusion, indicating that the deficit involves endothelium-derived NO [14]. Due to this observation, studies were initiated to find out whether L-arginine and/or tetrahydrobiopterin (BH4) may be deficient in various conditions associated with endothelial dysfunction. In fact, treatment with BH4 has been reported to augment endothelium-dependent vasodilation in different dysmetabolic phenotypes such as in smokers [15]. Also, L-arginine restores haemodynamic changes during acute hyperglycemia, suggesting that hyperglycemia may reduce NO availability [16]. Finally, asymmetric dimethylarginine (ADMA), a competitive inhibitor of eNOS, may also be of importance causing endothelial dysfunction in the oxidative stress milieu characterizing type 2 diabetes [17].

Inflammation.  Type 2 diabetes may in part be precipitated or accelerated by an acute phase reaction as part of the innate immune response, in which large amounts of cytokines are released from adipose tissue, creating an inflammatory milieu [18]. It is now well established that adipose tissue constitutes a highly active endocrine organ, releasing a variety of secretory products, e.g. hormones, cytokines, and enzymes with the propensity to impair insulin sensitivity [19–21]. These cytokines (TNF-alpha, IL-6, PAI-1) and adipokines (adiponectin and leptin) have been suggested to be associated with inflammation, insulin resistance and cardiovascular disease [12]. Adiponectin is an anti-inflammatory agent and is inversely related to BMI, insulin resistance and atherosclerosis [22–24]. Also, the recently described protein resistin released from adipocytes in animal studies [25], but not from mature adipocytes in humans [26], may have a role in the atherosclerotic process, as such that plasma levels of resistin correlate with inflammation and is an independent predictor of coronary artery disease (CAD) [27]. Keeping in mind that cytokines are involved in the atherosclerotic process, it has been hypothesized that obesity and insulin resistance, fuelled by the cytokines TNF-alpha, IL-6 and PAI-1, might sustain endothelial inflammation [12, 28]. It is possible that visceral obesity in particular, leads to insulin resistance and endothelial dysfunction through a cascade of released pro-inflammatory cytokines.

Methods and markers of endothelial function

The notion that severity of endothelial dysfunction relates to the risk for an initial or recurrent cardiovascular event has increased the interest to use measurements of endothelial function in atherosclerosis research [29–31]. Previous studies include both the coronary and brachial vascular beds and it seems that endothelial dysfunction is a systemic process that can be identified in vascular beds remote from the coronary and cerebral circulation where events occur [32]. Techniques have been developed to evaluate the release of NO in the coronary as well as in the systemic circulation. Most endothelial function tests pertain to abnormalities in the regulation of the lumen of the vessels. The action of endothelial cells may affect one or several functions, either simultaneously or in a temporal sequence. Therefore, it is hard to define any method for measurement of endothelial function superior to another. Instead, different techniques seem to be complementary to one another. Methods employed include the thermo- or dye-dilution, based on the Fick principle, as well as positron emission tomography (PET) scan, laser Doppler flowmetry, plethysmography and Doppler ultrasound. These methods have in common their ability to monitor the capacity of the endothelium to synthesize and release vasodilating and vasoconstricting compounds.

Plethysmography is a commonly used method based on venous occlusion plethysmography for studying endothelium dependent and independent vasodilation in peripheral circulation, e.g. dose-response relationships of endothelial agonists and antagonists. Usually, this technique uses infusion of acetylcholine (Ach) or other muscarinic receptor agonists in the brachial artery and determines the vasodilator responses over the forearm resistance vessels. Sodium nitroprusside (SNP) is used as a control substance in many of these studies to evaluate endothelium-independent vasodilation. The net change in blood flow, contributed by the whole limb vessel portion, provides a measure of endothelial function.

The reference method to assess endothelial function in the coronary circulation is the quantitative coronary angiography with an intra-coronary ultrasound and Doppler transducer. Following a stepwise infusion of Ach and SNP, the endothelium dependent and independent vasodilation can be assessed. Unfortunately, this technique is complicated and invasive. Therefore, a simple non-invasive method suitable for repeated studies also for evaluating large groups of patients has been developed – the flow-mediated vasodilation (FMD) [33]. FMD correlates to the endothelial function in the coronary circulation and has become a popular method in clinical studies [32, 34, 35]. Some limitations of the method have been discussed, e.g. great variation in absolute values between study centres and less reliable data in elderly subjects [34, 35]. FMD measurements are based upon the shear stress theory, whereby a short period of arterial occlusion increases flow in an artery, and this stimulus induces the endothelium to release the vasodilator NO. Oral nitroglycerin is given to assess the endothelium-independent vasodilation. The receptor signalling pathways activated by shear stress and subsequent signal transduction to modulate vasomotor tone are not fully understood. There seems to be a redundancy in the system and several endothelial mediators other than NO may act as signals between the endothelium and vascular smooth muscle [36, 37].

Measurements of endothelial biochemical markers may be the simplest method to monitor endothelial function indirectly. An increased expression of biochemical markers has been detected when endothelial cells undergo inflammatory activation in the presence of endothelial dysfunction. There is a number of circulating markers linked to endothelial dysfunction, including adhesion molecules, selectins, integrins, cytokines, fibrinolytic molecules, which all promote the adherence of monocytes and hence accelerate atherogenesis [38, 39]. Another more indirect method might be to analyse number and function of circulating endothelial progenitor cells known to be inversely related to endothelial dysfunction [40].

Insulin resistance mechanisms in endothelial cells

Clinical research of the in vivo insulin action has been focused on glucose, lipid and protein metabolism. However, it has now become evident that insulin also is a vasoactive hormone [41]. Intravenously administered insulin enhances blood flow and vasodilation in a NO dependent manner. The relevance of insulin as a modulator of blood flow has been under debate, amongst other things because the increase in blood flow evoked by insulin is much less than the classical endothelium dependent agonist Ach [42, 43]. Moreover, the increase in blood flow elicited by insulin differs between different types of vessels, e.g. capillary and resistance vessels. Insulin recruits and rapidly increases blood flow in capillaries [44–46], but on the other hand, it seems that insulin action on resistance vessels is slower in onset and requires at least several hours for a maximal effect [42]. Although insulin effects on blood flow regulation in muscle seem to have physiological relevance, the interaction of insulin and NO may also be of interest. Stimulation of NO production by insulin is mediated by signalling pathways involving activation of PI-3 kinase leading to phosphorylation of eNOS. Previous results elegantly show that the metabolic and vascular actions by insulin share the same signalling pathway, i.e. via PI3-kinase. It should be kept in mind, however, that insulin may also activate the pro-atherogenic mitogen-activated protein (MAP)-kinase pathway in endothelial cells [12, 47], (Fig. 1).

Figure 1.

 Mechanisms for the contribution of insulin resistance to atherosclerosis. In vitro model of metabolic insulin resistance with compensatory hyperinsulinemia in vascular endothelium. Modified with permission from Kim et al. Circulation 2006; 113: 1888. © Lippincott Williams & Wilkins.

For almost a decade, our laboratory has examined the clinical phenotype of individuals characterized by normal or low insulin receptor substrate 1 (IRS-1) protein expression in fat cells [48]. IRS-1 is one of the key proteins downstream of the insulin receptor for signalling to metabolic effects, e.g. glucose uptake in fat cells and NO-production in endothelial cells. In contrast to the commonly used risk marker, known heredity for diabetes, low cellular IRS-1 identified individuals who were markedly insulin resistant, had high proinsulin and insulin levels, and exhibited evidence of early atherosclerosis measured as increased intima media thickness in the carotid artery bulb. Circulating levels of adiponectin were also significantly reduced. In a recent study, we were also able to show that low adipocyte IRS-1 expression may be a marker not only for insulin resistance but also for arterial stiffness [49]. The data implicate that shared stressors, such as hyperglycemia and dyslipidaemia, cause oxidative stress and down-regulation of IRS-1 in fat cells and endothelial cells leading to insulin resistance and endothelial dysfunction. On the other hand, patients carrying a point mutation in IRS-1 that has been implicated in insulin resistance also show evidence of genetically based endothelial dysfunction [50].

Vascular inflammation, coronary artery disease and type 2 diabetes

Coronary heart disease, dyslipidaemias and atherosclerosis are cardiovascular disorders with endothelial dysfunction that are often seen with increased concentrations of inflammatory markers in the circulation [51]. Inflammatory cytokines like TNF-alpha and IL-1beta, signal through their receptors to activate JNK and IKKbeta in turn leading to activation of NF-kappaB [52]. This signalling inhibits insulin-stimulated activation of eNOS and expression of eNOS [53]. Moreover, NF-kappaB stimulates the expression of adhesion molecules, including intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), and E-selectin, known to enhance vascular pathology. Interestingly, NO has anti-inflammatory actions in the endothelium to inhibit NF-kappaB activity and reduce the expression of adhesion molecules VCAM-1, ICAM-1 and E-selectin [54]. Consequently, reduced bioavailability of NO under basal conditions caused by insulin resistance may be an additional pathogenic factor in chronic diseases (such as atherosclerosis, hypertension and diabetes) known to have inflammatory components. TNF-alpha is a strong inducer of the expression of several inflammatory proteins, including CRP and IL-6. CRP is a crucial marker of vascular inflammation whose plasma levels also correlate with risk of cardiovascular disease [55]. CRP has the ability to accelerate cardiovascular disease directly by stimulating the expression of proinflammatory cytokines in endothelium [56]. CRP decreases eNOS expression and increases angiotensin receptor type 1 expression in the vessel wall [57, 58]. Moreover, CRP increases the expression of endothelial ICAM-1, VCAM-1, E-selectin and MCP-1 and increases the secretion of ET-1 [59, 60]. Thus, CRP is an inflammatory marker that may be a player in atherogenesis and endothelial dysfunction [61].

A key feature of insulin resistance is that it is characterized by specific impairment in PI-3 kinase-dependent signalling pathways, whereas other insulin-signalling branches, including Ras/MAP-kinase-dependent pathways, remain unaffected [62, 63]. Accordingly, hyperinsulinemia, in the vasculature and elsewhere, will overdrive unaffected MAP-kinase-dependent pathways, leading to an imbalance between PI-3 kinase and MAP-kinase-dependent functions of insulin. Interestingly, metabolic unfavourable effects of insulin to promote secretion of ET-1, activate cation pumps, and increase expression of VCAM-1 and E-Selectin are also under the control of MAP-kinase signalling pathways [47]. In endothelium, decreased PI 3-kinase signalling and increased MAP-kinase signalling in response to insulin may lead to decreased production of NO and increased secretion of ET-1, characteristic of endothelial dysfunction. ET-1 increases serine phosphorylation of IRS-1, leading to decreased PI 3-kinase activity in vascular smooth muscle cells and may also impair insulin-stimulated translocation of GLUT4 in adipocytes [64, 65]. Thus, it is possible that decreased production of NO in endothelium mediated by insulin resistance also contributes to accelerated atherosclerosis by mediating vasoconstriction, inflammation and thrombosis.

Adipose tissue secretes a variety of hormones that can modulate endothelial function. In addition, components of the renin-angiotensin system (RAS), including angiotensin II, are present in adipose tissue [66]. Angiotensin-converting enzyme (ACE) inhibitors and angiotensin II type 1 receptor blockers (ARBs) may decrease the incidence of diabetes in patients with cardiovascular disease [67]. This observation may be due to inhibition of cross-talk between angiotensin II signalling and insulin signalling in adipose and vascular tissues [68]. In fact, addition of angiotensin II to endothelial cells activates JNK and MAP-kinase pathways, leading to increased serine phosphorylation of IRS-1, impaired PI 3-kinase activity, and endothelial dysfunction. Activation of angiotensin II type 1 receptor by angiotensin II has also been shown to stimulate the production of ROS via NADPH oxidase, increase expression of ICAM-1 and increase ET-1 release from endothelium [69–71]. It is therefore interesting that endothelial dysfunction mediated by angiotensin II is abolished by ARBs [72].

To get further insight into why metabolic and cardiovascular diseases are often associated, it is speculated that multiple stressors independently cause insulin resistance and endothelial dysfunction. For example, hyperglycemia leads to increased oxidative stress, increased AGE, inflammation and increased flux through the polyol, hexosamine and protein kinase C pathways inside the cell, and elevated levels of FFA promote oxidative stress and inflammation [5, 13, 73]. Thus, in the insulin resistance syndrome, diabetes, obesity, dyslipidaemias and cardiovascular diseases, numerous stressors may simultaneously cause insulin resistance in metabolic tissues and endothelial dysfunction in vascular tissues [47], (Fig. 2).

Figure 2.

 Parallel PI 3-kinase-dependent insulin-signaling pathways in metabolic and vascular tissues synergistically couple metabolic and vascular physiology. Modified with permission from Kim et al. Circulation 2006; 113: 1888. © Lippincott Williams & Wilkins.

Capillary recruitment and insulin resistance

In a classical study, the protocol was designed to elucidate the relationship between insulin‘s effects to stimulate leg muscle glucose uptake and its components and the alteration of this relationship in human obesity [74]. Laakso et al. were able to show that the dose-response curve for insulin´s effect to increase skeletal muscle blood flow was right-shifted in obese subjects with an ED50 about fourfold than in lean subjects and, furthermore, differences in blood flow rates to skeletal muscle were most marked within the physiological insulin range and thus contributed significantly to lower leg and whole body glucose uptake in obese man. The authors attributed this novel mechanism of insulin resistance to capillary recruitment, i.e. opening of previously underperfused or non-perfused capillaries, thus directly exposing more muscle cells to insulin and glucose. This notion was consistent with findings by others of an effect of insulin to increase the volume distribution of glucose [75].

Insulin infusion causes peripheral vasodilation and capillary recruitment and these effects collectively are referred to as the haemodynamic action of insulin. Early studies, where peripheral vasodilation or vasoconstriction was created by regional infusion of vasoactive agents, failed to demonstrate any change in peripheral glucose uptake [76]. In fact, in the only report showing an increase in peripheral glucose uptake in response to vasodilation, an elevation in plasma insulin concentration was present [77]. These studies indicate that vasodilation per se does not increase muscle glucose uptake. However, muscle glucose uptake is increased when vasodilation occurs concomitantly with the recruitment of new capillary beds, as induced by insulin [44, 45, 78]. Taken together, these data suggest that impaired capillary recruitment can cause insulin resistance with respect to glucose disposal both when the microvascular endothelium is otherwise healthy but cannot react properly, endothelial dysfunction, or in the presence of reduced capillary density per volume of tissue [79]. These two vascular defects may decrease insulin-mediated glucose disposal by increasing the diffusion distance of glucose and insulin to glucose-metabolizing tissues, by impairing transendothelial insulin transport [79, 80]. In harmony with the foregoing discussion, capillary density and impaired capillary recruitment may in part explain why insulin resistance is associated with hypertension, atherogenic changes and dyslipidaemia. A reduced capillary endothelial surface area may also result in reduced access of triacylglycerol-rich lipoprotein particles to LPL [5].

The preceding discussion emphasizes the obvious need to get more refined techniques that allow evaluation of the microcirculation, because there is no discernable change in total blood flow in most metabolic conditions. The first method employed for assessment of microcirculation in vivo in man was the contrast-enhanced ultrasonography with exogenously administered 1-methyl-xanthine [78]. By using this method, intra-arterial insulin infusion promoted capillary recruitment, without any change in regional blood flow, in healthy volunteers. Based on this and other studies, it is suggested that there is a time-dependent effect of insulin on capillary recruitment that is integrated with insulins metabolic action. Accordingly, insulin initially induces capillary vasodilation and pre-sphincter tone relaxation, and as a result, more microvessels are recruited within the metabolic active area in conjunction with increased microvascular perfusion. Insulin then can be transported more readily into the interstitial fluid surrounding the target cells, and the exposure to insulin, as well as to other hormones and nutrients, is augmented.

Baron et al. measured leg glucose uptake in healthy individuals by using the femoral arterial-venous catheterization technique [81]. Infusion of L-NMMA, an inhibitor of eNOS, into the femoral artery caused a 25% reduction in insulin-stimulated leg glucose uptake. Moreover, animal studies demonstrated that in the presence of the NOS-inhibitor L-NAME, insulin-induced capillary recruitment was reduced by 30% at 10 minutes. The reduction in capillary recruitment was associated with a 50% decrease in glucose fractional extraction. These results strongly suggest that endothelial-derived NO release, by enhancing capillary recruitment and microvascular blood flow in peripheral tissues, is required for the full metabolic action of insulin on muscle glucose uptake.

An alternative approach to examining the effects of insulin on blood flow distribution is to get access to interstitial fluid by the muscle microdialysis method. Our laboratory has used the Renkin equations to examine the effects of hyperinsulinemia on glucose and insulin that enters the interstitial fluid and what fraction that passes through to the venous blood. Measurements include blood flow through the deep muscles with plethysmography and measurements of interstitial concentrations of glucose and insulin in the forearm muscle of conscious man. Then, it is possible to calculate the permeability surface area (PS), a measure of the fraction of substance entering the tissue through the endothelium, during an oral glucose tolerance test or during a glucose clamp. PS was reported low under basal conditions but increased 10-fold during a glucose clamp [46]. The results from this study indicate that changes in blood flow distribution induced by insulin appear to mirror changes in glucose utilization. This suggests that insulin-stimulated alteration in blood flow patterns could be as important as direct signalling of cells by insulin in establishing the rate of glucose uptake in vivo. Another possibility to measure interstitial levels of insulin and glucose is to sample lymph [82]; however, it is difficult to sample lymph at frequent intervals in man. Therefore, we will continue the studies with muscle microdialysis to elucidate the role of the microvasculature in the pathogenesis of insulin resistance (Fig. 3). We have work in progress to find out means to facilitate glucose utilization in forearm muscle by administration of pharmacological agents known to enhance capillary recruitment.

Figure 3.

 Suggested pathways linking cardiovascular risk factors, insulin resistance and microvascular function. Reproduced with permission from Schalkwijk CG and Stehouwer CDA. Clinical Science 2005; 109: 143–159. © The Biochemical Society.

Endothelial dysfunction – a primary cause of type 2 diabetes?

Several lines of evidence support the hypothesis that endothelial dysfunction is a precursor of type 2 diabetes. In a prospective case-cohort study, the MONICA/KORA Study group was able to demonstrate that a sensitive marker of endothelial dysfunction, sE-selectin, was predictive of type 2 diabetes [83]. Elevated levels of sICAM-1 were also associated with an increased risk in men and women, but associations became non-significant after multivariable adjustment for other diabetes risk factors. In the Insulin Resistance Atherosclerosis Study, levels of log (PAI-1) increased the risk of incident type 2 diabetes, even after adjustment for major diabetes risk factors and directly measured insulin resistance [84]. In a recent study of postmenopausal women, each one-unit decrease in flow-mediated vasodilation of the brachial artery was associated with a significant ca 30% increase in the 4-year relative risk for incident diabetes after adjustment for most major diabetes risk factors [85]. Furthermore, in a community-based sample, plasma markers of endothelial dysfunction such as PAI-1 and the von Willebrand factor increased the risk of incident diabetes independent of other diabetes risk factors including obesity, insulin resistance and inflammation [86]. In the light of these observations, vascular endothelium in addition to fat, muscle, liver, and pancreas may be a tissue involved in the pathogenesis of type 2 diabetes (Fig. 4). Previous studies have shown endothelial dysfunction to be the key to pathogenesis of cardiovascular disease (CVD), supporting the idea of a common soil for developing type 2 diabetes and CVD. Taken together, available data support a role for endothelial dysfunction in the etiology of type 2 diabetes mellitus and provide new aspects for prevention of the disease. Interventions that improve endothelial function, including life style modifications and specific pharmacologic agents, could have beneficial effects on diabetes and cardiovascular risk, and should be instituted.

Figure 4.

 Proposed role of endothelial cells in the insulin resistant state.

Conflict of interest statement

No conflict of interest was declared.

Acknowledgements

This work was supported by Novo Nordisk Foundation, the Swedish Diabetes Association and the Swedish Medical Research Council.

Ancillary