Insulin resistance and endothelial dysfunction: the road map to cardiovascular diseases

Abstract

Cardiovascular disease affects approximately 60% of the adult population over the age of 65 and represents the number one cause of death in the United States. Coronary atherosclerosis is responsible for the vast majority of the cardiovascular events, and a number of cardiovascular risk factors have been identified. In recent years, it has become clear that insulin resistance and endothelial dysfunction play a central role in the pathogenesis of atherosclerosis. Much evidence supports the presence of insulin resistance as the fundamental pathophysiologic disturbance responsible for the cluster of metabolic and cardiovascular disorders, known collectively as the metabolic syndrome. Endothelial dysfunction is an important component of the metabolic or insulin resistance syndrome and this is demonstrated by inadequate vasodilation and/or paradoxical vasoconstriction in coronary and peripheral arteries in response to stimuli that release nitric oxide (NO). Deficiency of endothelial-derived NO is believed to be the primary defect that links insulin resistance and endothelial dysfunction. NO deficiency results from decreased synthesis and/or release, in combination with exaggerated consumption in tissues by high levels of reactive oxygen (ROS) and nitrogen (RNS) species, which are produced by cellular disturbances in glucose and lipid metabolism.

Endothelial dysfunction contributes to impaired insulin action, by altering the transcapillary passage of insulin to target tissues. Reduced expansion of the capillary network, with attenuation of microcirculatory blood flow to metabolically active tissues, contributes to the impairment of insulin-stimulated glucose and lipid metabolism. This establishes a reverberating negative feedback cycle in which progressive endothelial dysfunction and disturbances in glucose and lipid metabolism develop secondary to the insulin resistance. Vascular damage, which results from lipid deposition and oxidative stress to the vessel wall, triggers an inflammatory reaction, and the release of chemoattractants and cytokines worsens the insulin resistance and endothelial dysfunction.

From the clinical standpoint, much experimental evidence supports the concept that therapies that improve insulin resistance and endothelial dysfunction reduce cardiovascular morbidity and mortality. Moreover, interventional strategies that reduce insulin resistance ameliorate endothelial dysfunction, while interventions that improve tissue sensitivity to insulin enhance vascular endothelial function. There is general agreement that aggressive therapy aimed simultaneously at improving insulin-mediated glucose/lipid metabolism and endothelial dysfunction represents an important strategy in preventing/delaying the appearance of atherosclerosis. Interventions that 1 correct carbohydrate and lipid metabolism, 2 improve insulin resistance, 3 reduce blood pressure and restore vascular reactivity, and 4 attenuate procoagulant and inflammatory responses in adults with a high risk of developing cardiovascular disease reduce cardiovascular morbidity and mortality. Whether these benefits hold when the same prevention strategies are applied to younger, high-risk individuals remains to be determined. Copyright © 2006 John Wiley & Sons, Ltd.

Introduction

Cardiovascular disease affects approximately 60% of the adult population over the age of 65 and represents the number one cause of death in the United States 1. Coronary atherosclerosis is responsible for the vast majority of cardiovascular events, which occur with increased frequency in individuals with hypertension, dyslipidemia, obesity, diabetes, renal disease, and family history of coronary artery disease, and in persons who are sedentary and who smoke 2. A cluster of cardiovascular risk factors, including glucose intolerance, hypertension, hypertriglyceridemia, and visceral obesity, have been collectively termed the metabolic syndrome, and they occur in approximately one out of four adults over the age of 40 3. The metabolic syndrome is typically associated with insulin resistance and endothelial dysfunction 4–6. Insulin resistance is defined as impairment of insulin-stimulated glucose and/or lipid metabolism, compared to the response in healthy subjects, while endothelial dysfunction is defined as paradoxical or inadequate endothelial-mediated vasodilation. The presence of insulin resistance and endothelial dysfunction represents early events in individuals at high risk for developing cardiovascular disease. The metabolic syndrome identifies a group of subjects who are at very high risk (>20% for those with diabetes and 10–20% for those with two or more risk factors) to experience a cardiovascular event 3. Consequently, aggressive therapy, aimed at normalizing each individual risk factor, is indicated 7. The earlier the intervention is instituted, the greater will be the benefit with regard to cardiovascular protection. Since insulin resistance and endothelial dysfunction are common early findings in most asymptomatic people with the potential for developing ‘metabolic syndrome’ 8, and since both abnormalities are present in subjects with documented atherosclerosis or in individuals at high risk of atherosclerosis 9, a strong argument can be made for the institution of preventive interventions that simultaneously reduce insulin resistance and improve endothelial dysfunction.

In the present review, we will examine the relationship between insulin resistance, endothelial dysfunction, and atherosclerotic cardiovascular disease. Differences in the vascular response to ischemia—a test for endothelial function—between adults and children will be explored and their implications for the development and prevention of atherosclerotic cardiovascular disease will be discussed.

Insulin resistance

Insulin is a powerful anabolic hormone that exerts its metabolic effects primarily on the liver, the adipose tissue, and the skeletal muscle 10. Insulin enhances fuel storage by (1) promoting glycogen synthesis in liver and muscle, (2) augmenting triglyceride synthesis and deposition in adipose tissue, and (3) increasing protein synthesis and inhibiting proteolysis. Insulin also promotes glucose oxidation, providing an important energy source in the form of ATP. Under certain physiologic (growth, adolescence, pregnancy) and pathophysiologic (obesity, type 2 diabetes, stress, acute illness) conditions, the cellular action of insulin becomes impaired and insulin resistance develops. In order to compensate for the increasing resistance of tissues to insulin's metabolic action, the pancreas augments its secretion of insulin, resulting in systemic hyperinsulinemia 11.

Insulin resistance in adipocytes leads to accelerated lipolysis with subsequent elevation in circulating free fatty acids (FFA). Increased transport of plasma FFAs into muscle, liver, and pancreatic β cells further impairs insulin action and insulin secretion, the two characteristic metabolic disturbances in type 2 diabetes mellitus 11. In addition, insulin resistance is associated with mitochondrial defects 12 that may limit fatty acid oxidation and favor the accumulation of intracellular fatty acyl-CoA and other toxic products of lipid metabolism (ceramides, diacyl-glycerol, etc.). In the presence of hyperglycemia, intracellular malonyl-CoA accumulates, further reducing fatty acid oxidation by inhibiting the carnitine palmitoyl-transport (CPT-1) system. The latter is required for the transport of long-chain fatty acids from the cytosol into the mitochondria, where β oxidation of long-chain fatty acyl-CoAs takes place. The accumulation of intracellular fatty acyl-CoA impairs insulin action in skeletal muscle and leads to worsening of insulin resistance 10, 121.

Several methods have been designed over the years to evaluate insulin resistance, but the hyperinsulinemic euglycemic clamp remains the ‘gold-standard’ technique for evaluating insulin-stimulated glucose metabolism 13. Obesity, which is characterized by moderate-to-severe insulin resistance 14 has become a formidable nutritional problem of modern civilization, affecting nearly 65 million Americans 15. The obesity epidemic has turned into a global problem and largely accounts for the marked increase in the prevalence of type 2 diabetes worldwide (WHO report 2004). This represents a major challenge for all health professionals today.

In addition to its classical metabolic effects to promote fuel storage and stimulate glucose oxidation, insulin has important nonmetabolic hemodynamic actions. Insulin promotes capillary recruitment, causes peripheral vasodilatation, and increases regional blood flow 16–18. These hemodynamic effects of insulin complement the hormone's metabolic action by enhancing glucose delivery and exposing previous underperfused tissues, i.e. skeletal muscle, to insulin 19. It has been proposed that as much as 25% of insulin's stimulatory effect on muscle glucose uptake is related to its hemodynamic actions 20. Insulin exerts its vascular effects primarily by augmenting the availability of endothelium-derived nitric oxide (NO), a potent vasodilator. In experimental conditions mimicking the state of insulin resistance, endothelial-mediated vasodilation is impaired, partly because of insulin's inability to stimulate the activity of the key enzyme responsible for NO synthesis, nitric oxide synthase (NOS), in vascular endothelial cells 21–23. In addition, enhanced NO consumption results from increased levels of ‘reactive oxygen species’ (ROS) and ‘reactive nitrogen species’ (RNS), which are believed to be generated in excess by intracellular defects in insulin-mediated glucose and lipid metabolism. Whether this further contributes to reduce NO availability is currently under investigation.

Endothelial dysfunction

The endothelium represents a single layer of cells that line all the vessels in the body, including the conduit vessels, the resistance vessels, precapillary arterioles and capillaries 24. By virtue of its direct contact with the circulating blood, the endothelial layer provides a critical interface between the elements of blood and the tissues. The function of each vessel and the role of its respective endothelium vary according to its location in the body. In the conduit vessels, which include larger arteries such as the aorta and the carotid/coronary/brachial/femoral arteries, a healthy endothelium provides a smooth, quiescent surface that limits the activation of clotting and proinflammatory factors, blocks the transfer of Apo-B 100-containing atherogenic lipid particles into the arterial wall, inhibits the release of chemokines/cytokines/growth factors, and prevents adhesion of platelets and monocytes to the vascular endothelium. In the resistance vessels, endothelial cells help regulate blood flow and systemic blood pressure, whereas in the precapillary arterioles, they play a role in the transport and distribution of nutrients and hormones, including glucose, fat, and insulin, and for the adequate disposal of waste products of metabolism. Vascular endothelial dysfunction, therefore, may occur at any or all levels in the arterial system and contributes to the development and progression of atherosclerosis by favoring coagulation, cell adhesion, and inflammation, by promoting inappropriate vasoconstriction and/or vasodilation, and by enhancing transendothelial transport of atherogenic lipoproteins. Many of these disturbances in endothelial function become apparent very early in subjects who are at high risk for developing atherosclerosis, and all of these abnormalities are present in varying degrees in the later stages of the disease 6–8, 25, 26.

Endothelial function can be assessed in vivo by documenting the response of conduit and resistance arteries to physiologic stimuli 22. A commonly used method to evaluate endothelial function employs the measurement of flow-mediated vasodilation following a period of transient ischemia. This vasodilatory response is dependent upon a series of neurologic, myogenic and chemical intermediates, including the release and adequate availability of NO 24, and all of these processes can be examined noninvasively in accessible peripheral vessels. Systemic or regional (intra-arterial or subcutaneous) administration of cholinergic agents, such as acetylcholine or methacholine, can be used to evaluate the endothelial-dependent component of the vasodilatory response 27–29. Vasodilators, such as nitrates or sodium nitroprusside, directly induce vascular smooth muscle (VSM) cell relaxation, independent of the endothelium and, more specifically, in the absence of NO. Their use allows one to distinguish between endothelium-dependent and endothelium-independent vasodilatory responses 25. Additionally, measurement of circulating levels of oxidative products and markers of endothelial damage and inflammation 30, such as vascular and intercellular adhesion molecules (VCAM and ICAM) 31, C-reactive protein, tumor necrosis factor alpha (TNF-α) 119, interleukin-6 (IL-6), endothelin-1 32–34, to name a few, can provide an indirect assessment of the activity and degree of involvement of the vascular endothelium and its surrounding tissues in the atherosclerotic process.

The concept that endothelial dysfunction is an essential part of the atherosclerotic process derives from the original observation by Ludmer et al.26 on patients undergoing diagnostic coronary angiography. These investigators demonstrated that prestenotic and stenotic segments of coronary arteries exhibited paradoxical vasoconstriction in response to intra-arterial infusion of acetylcholine. In the absence of the opposing vasodilatory action mediated via endothelial NO and other factors, arterial vasodilation due to direct effects on muscarinic receptors prevails. A normal vasodilatory response to nitroglycerin infusion demonstrated that the defect was not in the VSM cells, but most likely was related to the deficiency of some endothelium-derived relaxing factor(s). The idea that impaired release of endothelium-derived NO was responsible for the paradoxical coronary vasoconstriction has been supported by many subsequent studies, which confirmed that the abnormal vascular reactivity in peripheral arteries is due primarily to insufficient availability of NO 35–39. The determination of the brachial/femoral arterial response to vasodilator stimuli has become the preferential method for noninvasive assessment of vascular endothelial function, since results obtained in peripheral vessels correlate very closely with those in coronary arteries 40, 120. Numerous experiments using pharmacological blockade of NO synthesis have reproduced the abnormal vascular response described in individuals with coronary artery disease, in individuals with multiple cardiovascular risk factors, in patients with type 2 diabetes mellitus (a cardiovascular disease equivalent), and in ‘prediabetics’ 41–43. Inappropriate endothelial-mediated vasodilation in the forearm skin microcirculation also has been reported in diabetics 44 and in normoglycemic, first-degree relatives of type 2 diabetic patients and individuals with impaired glucose tolerance 25. It should be emphasized that the vast majority (>90%) of diabetic and prediabetic individuals, who are at very high risk for the development of coronary atherosclerosis and who have been shown to have endothelial dysfunction, also manifest moderate-to-severe insulin resistance 45. These two groups represent the ‘prototype’ of those people at risk for the development of atherosclerotic cardiovascular disease and illustrate very clearly that insulin resistance and endothelial dysfunction both precede the appearance of clinically significant vascular disease.

Insulin and the endothelium: from precapillary arterioles to large conduit vessels

Transendothelial insulin transport

In order to reach its target receptor on the cell membrane, insulin must first cross the endothelial barrier. Insulin is thought to be internalized by the vascular endothelial cells via a receptor-mediated process, which has not been fully elucidated, before it reaches the perivascular space 46. This represents a potential rate-limiting step in peripheral insulin action and has been shown to contribute to the delay in insulin-stimulated glucose metabolism in some conditions of insulin resistance 14, 47–50. During intravenous glucose tolerance tests performed in conscious dogs, plasma and lymph glucose concentrations were very similar, but the lymph insulin concentration was about one-half of that in plasma. Moreover, while it took only 10 min for glucose to reach peak levels in lymph, it took nearly 21 min for insulin to reach its peak level. If one accepts that insulin transport in lymph reflects that in the interstitial space, these data suggest that there must be a substantial delay in the transendothelial transport of insulin and this could represent an important rate-limiting step in the action of insulin to stimulate muscle glucose uptake and metabolism. Consistent with this hypothesis, previous investigations in normal 14, obese 48, and type 2 diabetic 49, 51 subjects have documented a significant retardation in insulin-mediated peripheral glucose disposal when compared to lean nondiabetic individuals (Figure 1).

Figure 1. Muscle glycogen synthesis (measured by ¹³C-NMR) and leg glucose uptake (measured by the leg catheterization technique) (adapted from DeFronzo, Ref. 51)

Sjostrand et al.50 have provided additional support for the important role of transendothelial insulin transport in the regulation of glucose transport in man. Using microdialysis probes inserted into the subcutaneous tissue, these investigators demonstrated that the time needed for insulin to achieve half-maximal concentration in the interstitial fluid (72 versus 45 min) and the time for the onset of the metabolic action of insulin (70 versus 45 min) was prolonged in insulin-resistant, obese subjects as compared to insulin-sensitive lean subjects. These observations are consistent with the concept that the delay in peripheral insulin action is, in part, due to impaired insulin diffusion across the capillary bed. This defect could result from a reduced capillary density, inadequate capillary recruitment, or the increased diffusion distance present in the capillary bed in obese individuals. The authors 50 hypothesized that the major factor that determines the rate of insulin transport across the capillary bed must be the size of the capillary bed itself or the inability of insulin to recruit previously closed capillary beds. In addition to insulin, adipocytokines and other endothelial products that are capable of altering capillary permeability 52, 53, and thus could also interfere with insulin's metabolic action, have been identified in insulin-resistant obese individuals.

Hemodynamic action of insulin

Many investigators have attempted to integrate the metabolic actions of insulin with its effects on regional blood flow. Insulin infusion causes peripheral vasodilation and capillary recruitment and these effects are collectively referred to as the hemodynamic action of insulin. Baron and colleagues 20, 54, 55 have demonstrated a consistent increase in leg blood flow during euglycemic physiologic hyperinsulinemic clamp studies performed on lean, obese and diabetic individuals. When insulin was infused at the rate of 40 and 120 mU/m² (steady-state plasma insulin concentration = 69 ± 3 and 210 ± 13 µU/mL, respectively), after 240 min leg blood flow increased by approximately 100% in both the low and high-dose insulin infusion (Figure 2). Others also have demonstrated increases in forearm and leg blood flow in humans at even lower, more physiologic insulin infusion rates 56, 57.

Figure 2. Leg blood flow during euglycemic insulin clamp studies performed in lean, obese and type 2 diabetic subjects (adapted from Baron et al., Ref. 58)

However, the concept that increases in regional blood flow and stimulation of peripheral glucose metabolism are part of the combined physiological action of insulin continues to be a matter of intense debate 59, 119. The controversy stems from the fact that insulin-induced peripheral vasodilatation with increased blood flow is not a universal finding and has not been reproduced by all investigators under comparable experimental conditions 59. Moreover, the increment in blood flow has been more clearly documented after prolonged tissue exposure and at high pharmacological doses of insulin. Furthermore, studies in which peripheral vasodilation or vasoconstriction was created by regional infusion of vasoactive agents, including adenosine 60, bradykinin 61, angiotensin II 62, or methacholine 58, failed to demonstrate any change in peripheral glucose uptake 58, 60–62. In the only study that demonstrated an increase in peripheral glucose uptake in response to vasodilatation, a simultaneous elevation in plasma insulin concentration was present 36. These studies indicate that vasodilation per se does not increase muscle glucose uptake. However, when vasodilation occurs concomitantly with the recruitment of new capillary beds, as brought about by insulin 19, muscle glucose uptake is enhanced.

The preceding discussion emphasizes the problem associated with the measurement of regional blood flow with conventional techniques, since redistribution of flow cannot be detected 63. Because there is no discernible change in total blood flow in most conditions, more refined techniques that allow evaluation of the microcirculation are necessary. Two examples of conditions in which alterations in blood distribution within the microcirculation is of important physiological relevance are the postprandial state and exercise. After a meal, there is a recruitment of capillary blood toward microvessels in the peripheral circulation and, during exertion, blood flow is directed preferentially to the exercising muscles, prior to any detectable increment in regional blood flow 52. Moreover, it is well known that under resting conditions only 50% of the capillary bed in the skeletal muscle is perfused with whole blood, whereas the other 50% is only plasma-filled 64. Thus, the metabolic action of insulin is enhanced by its effect on the recruitment of new capillary beds and the redirection of capillary blood flow more toward insulin-sensitive tissues (muscle and adipocytes), and away from insulin-independent tissues such as bone, tendon, and skin. Importantly, this effect of insulin is independent of changes in total regional blood flow. By redistributing the blood flow and opening precapillary sphincters, insulin can gain access to a larger area surrounding the target tissues, and its own metabolic action is augmented.

To further investigate whether insulin exerts an effect on capillary recruitment, Coggins et al.19 employed contrast-enhanced ultrasonography with exogenously administered 1-methyl-xanthine to provide a measure of capillary endothelial exposure to blood in vivo. In healthy volunteers intra-arterial insulin infusion augmented the size of the capillary bed distal to the infusion site by promoting capillary recruitment. Despite clear evidence of capillary bed expansion, no change in regional blood flow was observed using either plethysmography or ultrasonography simultaneously (Figure 3). On the basis of these and other confirmatory reports 65, 66, the current hypothesis is that there is a time-dependent effect of insulin on regional blood flow distribution that is integrated with the hormone's metabolic action. Upon contact with the endothelial barrier, insulin initially triggers capillary dilation and presphincter tone relaxation. As a result, more microvessels are recruited within the metabolically active area in conjunction with increased microvascular perfusion. Insulin then diffuses 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 (Figure 4). The physiological relevance, thus, of insulin-induced vasodilation remains unclear.

Figure 3. The effect of intra-arterial insulin infusion on muscle capillary recruitment and regional blood flow (adapted from Coggins et al., Ref 19)

Figure 4. Hemodynamic and metabolic actions of insulin. The time-course of insulin's hemodynamic action is closely integrated with its metabolic effects. Following its passage through the endothelial barrier, insulin promotes precapillary sphincter tone relaxation with capillary dilatation. As a result, more microvessels are recruited, the capillary network is expanded, and peripheral microvascular perfusion increases. Insulin then diffuses into the interstitial fluid more readily, and the exposure of the target tissues to insulin is magnified, resulting in an increase in insulin-mediated glucose metabolism

Knowledge about the biochemical/molecular mechanisms responsible for the hemodynamic effects of insulin has advanced rapidly. It is now well established that the vascular endothelium must be intact and that the endothelial-derived vasorelaxation factor, NO, plays a critical role in mediating the vascular actions of insulin 67–69. In the intact vascular endothelium, arginine is converted to NO by the enzyme nitric oxide synthetase (NOS). The activity of NOS is increased severalfold by cytokines such as IL-1β, IL-6, TNF-α, interferon-γ, and adenosine 70. There is also evidence that NOS responds to pharmacological amounts of insulin in endothelial cell preparations in vitro68. NO is a highly reactive free gas with easy tissue penetration and passive movement across all cell membranes, including VSM cells. Once NO diffuses into VSM cells (Figure 5), it binds to the heme moiety of guanylate cyclase, leading to a reduction in the local concentration of cyclic GMP. In turn, this activates myosin light chain phosphatase and opens K-_ATP channels via cGMP-dependent protein kinase pathways. A second mechanism by which NO causes vasorelaxation is by direct nitrosylation of the K-_ATP channel, resulting in hyperpolarization of the VSM cell membrane and inhibition of calcium entry 70, 71. The presence of NO in the VSM cells, therefore, is critical in regulating the balance between vasodilation and vasoconstriction 72. Lack of NO, for whatever reason, leads to unopposed VSM contraction and can cause paradoxical vasoconstriction.

Figure 5. Mechanism of vascular smooth muscle cell relaxation by nitric oxide. Endothelium-derived nitric oxide induces vascular smooth cell relaxation by two distinct mechanisms: (i) activation of cyclic GMP and (ii) direct hyperpolarization of the cell membrane. The resting membrane potential of vascular smooth muscle ranges from −30 to −60 mV. A more positive potential (depolarization) opens the voltage-gated calcium channels, increases cytosolic calcium, and promotes vasoconstriction. Conversely, hyperpolarization closes these channels, decreases the cytosolic calcium, and promotes relaxation. Nitric oxide induces vasodilation by activating guanylate cyclase, leading to an increase in intracellular cGMP levels. Stimulation of myosin phosphatase dephosphorylates myosin, thereby inhibiting muscle contraction. In addition, nitric oxide activates K-_ATP channels in vascular smooth cells preventing the rise in cytosolic calcium and antagonizing the effect of vasoconstrictors, such as angiotensin II and norepinephrine

Metabolic disturbances and nitric oxide

Numerous studies have demonstrated that conditions of insulin resistance are commonly associated with endothelial dysfunction and that the exposure of vascular endothelium to high circulating levels of lipids and glucose is accompanied by reduced NO availability 46, 73–76. These observations have given rise to the theory that endothelial dysfunction is both a cause and a consequence of the metabolic disturbances observed in states of insulin resistance. Impaired suppression of adipose tissue lipolysis and mild transient postprandial hyperglycemia 10, which characterizes insulin resistance in the prediabetic state, favor FFA utilization and oxidation in insulin-dependent tissues and increase glucose flux into insulin-independent tissues. Hypertriglyceridemia and elevated small dense LDL cholesterol particles in the circulation, also typical of insulin resistance states, contribute to vascular damage and trigger the inflammatory response, resulting in adhesion of monocytes/lymphocytes to endothelial cells. Increased flux of glucose and FFA into VSM cells and surrounding inflammatory cells may lead to an excess formation of ROS and RNS (superoxide, peroxide, peroxynitrites, etc.), and further hemodynamic and metabolic deterioration. Within the mitochondria the abundant free radicals saturate the oxy-reduction enzymatic capacity and NO is diverted into nitrite formation, impairing the electron transfer chain and promoting cell apoptosis 77. Mitochondrial NOS is also impaired and the synthesis of NO is further compromised. The decreased synthesis and release of NO, combined with its accelerated consumption during the neutralization of the oxidative stress, makes NO less available for its normal vascular functions, namely, vasorelaxation and inhibition of vasoconstriction. Thus, in the face of insulin resistance with its associated metabolic disturbances in glucose and lipid metabolism decreased NO availability leads to impaired endothelial-mediated vasodilation. The disruption of normal vascular endothelial function, particularly in the arterioles and capillaries, further impairs the metabolic action of insulin, thus establishing a self-perpetuating negative feedback cycle. The relationship between insulin resistance and abnormal vascular reactivity has been demonstrated in a variety of clinical conditions, and there is evidence that a reduction in insulin resistance is accompanied by improved endothelial function and vice versa 78–80. Although the molecular mechanisms responsible for the metabolic and vascular abnormalities associated with insulin resistance have yet to be entirely elucidated, deficient NO production appears to play an important role.

Baron et al.20 measured leg glucose uptake in healthy individuals using the femoral arterial–venous catheterization technique. Infusion of L-N mono-methyl arginine (L-NMMA), a specific competitive inhibitor of NOS, into the femoral artery caused a sustained 25% reduction in insulin-stimulated leg glucose uptake, suggesting that NO plays an important role in the regulation of muscle glucose utilization by insulin. Additional evidence that endothelial-derived NO participates in insulin-stimulated glucose utilization comes from similar experiments in rats 18, 81. Thus, infusion of L-Nitro Arginine methyl-ester (L-Name) an inhibitor of NOS, reduced insulin-stimulated glucose uptake by 40% (4.4 versus 7.0 mg/kg min). This study also demonstrated that, in the presence of the NOS inhibitor L-NAME, insulin-induced capillary recruitment was reduced by 30% at 10 min with a further decrease to 80% at 30 min. The reduction in capillary recruitment was associated with a ∼50% decrease in glucose fractional extraction from 9 to ∼4.5%. These results provide strong evidence 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.

Collectively, the data reviewed above support the contention that the state of insulin resistance encompasses abnormalities in both the metabolic and vascular hemodynamic actions of the hormone. Defective insulin-stimulated endothelial release of NO appears to be responsible, in part, for the impaired capillary network expansion and the inability of insulin to redirect blood flow in the microcirculation toward metabolically active tissues. As a result, the diffusion of insulin and its metabolic substrates is delayed and diminished, further aggravating the underlying insulin resistance. Characterization of the molecular mechanisms responsible for the impairment in insulin action and the endothelial-mediated vascular responses to the metabolic derangements induced by insulin resistance is of considerable clinical interest in order to better understand the interrelationship between insulin resistance and endothelial dysfunction in the initiation and progression of atherosclerosis.

Conductance and resistance vascular reactivity

Vascular reactivity is defined as the overall vasomotor response to physiologic or pharmacologic stimuli and requires intact and harmonious interactions between the vascular endothelium and smooth muscles cells. The presence of a defective endothelium in the arterial system can be unveiled by the demonstration of impaired vascular reactivity, such as paradoxical vasoconstriction and/or inappropriate vasodilation, following stimuli that require endothelium-derived vasorelaxation factors. Vascular reactivity can be tested by analyzing the changes in forearm or leg blood flow following a period of transient ischemia 28, 30, 31, i.e. ‘reactive hyperemia’, after the intra-arterial infusion of methacholine 25, acetylcholine 27, nitroprusside 27–29, or following the administration of oral or sublingual nitrates 27, 39. To identify the role of endothelium-derived NO, experiments are conducted in the presence and absence of a specific NOS inhibitor 22, 38, 42, 82. An alternative strategy is to examine vascular reactivity during exogenous administration of NO in the chemical form of nitrates or nitroprusside, and to compare the results with those obtained in similar experiments performed in the presence of vasodilatory agents/maneuvers that specifically stimulate endogenous NO release, such as acetylcholine, methacholine, or ischemia. An adequate vascular response to NO derived from an exogenous source usually indicates that, in the presence of sufficient NO, the VSM cells are intact and are capable of vasorelaxation. In contrast, if the response to exogenously administered NO is abnormal, the defect is localized to the VSM cells. Demonstration of an abnormal response to exogenous NO does not exclude the possibility that the vascular endothelium also may be dysfunctional. Vascular dynamic testing with evaluation of the vessel's response to agents that stimulate endogenous NO release (acetylcholine, methacholine) and to agents that provide exogenous NO (nitrates, nitroprusside) is the best strategy to help discern endothelial dysfunction from a more generalized vessel wall defect that includes VSM cells. With all of these approaches, it is presumed that the presence of an intact vascular endothelium and normal VSM cell function suffice to elicit an appropriate response. Inadequate vasodilation or paradoxical vasoconstriction detected in a specific vascular region is taken as evidence that either the endothelium or the VSM cells or both are defective. Although it is usually assumed that the defect can be generalized to the entire circulatory system, confirmatory evidence in another vascular bed is advisable before drawing definitive conclusions. Furthermore, since abnormal vascular reactivity may have diverse underlying primary and compensatory mechanisms that act in concert, these vascular tests provide limited insight into which one is predominant under different physiologic/pathophysiologic circumstances. Despite these assumptions and limitations, in the absence of a better method, the determination of vascular reactivity has been widely employed and accepted as a reasonable test for the evaluation of endothelial function in vivo.

The degree of arterial vasodilation can be estimated by measuring changes in regional blood flow or by direct visualization of the vessel's diameter. Venous occlusion plethysmography (VOP) has been used extensively to evaluate regional (forearm or leg) blood flow, and is considered to reflect the vascular response of resistance arteries 36–38. The application of high-resolution ultrasonography (HRU) allows the investigator to examine blood flow velocity and diameter in conductance vessels 25, 27–31. The arterial blood flow is then calculated using the equation derived from Poiseuille's Law 63. While the VOP provides quantitative data on changes in limb blood flow during dynamic testing, HRU, in addition to yielding information about blood flow velocity and arterial diameter, provides important information about the time-dependence and the entire response of the vessel to a specific stimulus. Testing vascular reactivity in conductance vessels with the HRU, therefore, can provide insight into the regulation and the adaptation of the vascular system to physiologic and pathologic conditions. This is illustrated by data obtained in the brachial artery of children and adults using HRU (Table 1). Vascular parameters were measured in the basal resting state and following 5 min of forearm ischemia (‘reactive hyperemia’). The percent increase in arterial blood flow (53%) and velocity (46%) 15 s after release of the forearm cuff in adults was greater (p < 0.01) than in children (by 35% and by 38%, respectively). The percent increase in brachial artery diameter 50–60 s following ischemia, an index of flow-mediated vasodilation, was also greater in adults than in children. The elapsed time to reach the peak posthyperemic blood flow was shorter in adults than in children (18 versus 48 s). As a result, the entire hyperemic response, estimated by the area under the curve over the 5-min period of observation required for arterial blood flow to return to baseline, was greater in adults than in children. The difference primarily was due to a larger percent increase in brachial artery diameter. These data indicate that the vascular response to ischemia is altered with growth and maturation. Thus, as individuals progress from childhood to adulthood there is a shortening of the length of time to reach the peak hyperemic flow, as well as an amplification of the peak hyperemic flow. Consequently, there is a greater increase in flow-mediated vasodilation. These differences most likely are due to the differences in the elastic properties of the vasculature in children versus adults, and to differences in the composition of the tissues (muscle, fat, skin, etc.) perfused by the artery in adults versus children. Greater proportional representation of skeletal muscle than skin and adipose tissue in the area of the arm perfused by the brachial artery could be partially responsible for a more rapid peak and a larger blood flow increment in response to ischemia.

Table 1. Brachial arterial response to 5 min of forearm ischemia (‘reactive hyperemia’) in healthy adults (n = 20) and children (n = 50) using high-resolution ultrasonography. The percent increase in the arterial blood flow (53%) and flow velocity (46%) 15 s after the release of forearm cuff in adults was significantly greater in adults than in children (35 and 38%, respectively). The percent increase in brachial artery diameter 50–60 s postischemia, an index of flow-mediated vasodilation, was also significantly greater in adults than in children. The time elapsed o reach the peak posthyperemic blood flow was shorter in adults than in children (18 versus 48 s). As a result, the hyperemic response, estimated by the area under the curve for the entire 5-m period required to return arterial blood flow to the basal level, was greater in adults than in children. This was primarily due to a larger percent increase in the brachial artery diameter. These data indicate that there is an adaptation of the vascular response to ischemia from childhood to adulthood. This adaptation involves shortening of the length of time to reach peak responses and amplification of the hyperemic peak flow with consequent larger flow-mediated vasodilation

Vascular parameters	Adults	Children	P value
a Basal blood flow, flow velocity, and arterial diameter represent absolute values and are not corrected for arm size. b 15 s and 50–60 s postischemia represent results 15 s and 50 to 60 s after release of the arterial occlusion. c Absolute values are presented, followed in parenthesis by the percent increase (↑) above baseline. d AUC, brachial arterial blood flow incremental area under the curve above baseline over the 5-min period after release of the arterial occlusion.
Basal—blood flow (mL/min)a	310 ± 42	203 ± 14	p < 0.001
—Flow velocity (cm/s)a	62 ± 5	57 ± 2	p < 0.001
—Arterial diameter (cm)a	0.32 ± 0.03	0.27 ± 0.01	p < 0.01
15 s postischemiab
—Blood flow (%↑)c	459 ± 56 (53 ± 12)	263 ± 16 (35 ± 1)	p < 0.05
—Flow velocity (%↑)c	90 ± 6 (46 ± 8)	75 ± 2 (38 ± 5)	p < 0.05
—Arterial diameter (%↑)c	0.30 ± 0.02 (3 ± 4)	0.27 ± 0.01 (−3 ± 2)	NS
50–60 s postischemiab
—Blood flow (%↑)c	422 ± 64 (38 ± 13)	274 ± 15 (43 ± 4)	NS
—Flow velocity (%↑)c	70 ± 6 (13 ± 7)	62 ± 2 (17 ± 3)	NS
—Arterial diameter (%↑)c	0.36 ± 0.02 (13 ± 3)	0.30 ± 0.01 (10 ± 1)	p < 0.05
Elapsed time to peak posthyperemic blood flow (s)	18 ± 4	48 ± 5	p < 0.001
Integrated AUC (5 min)d
—Blood flow	123343 ± 14749	61157 ± 2573	p < 0.001
—Flow velocity	21798 ± 1477	21710 ± 4209	NS
—Arterial diameter	102 ± 7	81 ± 1	p < 0.001

Measurements of reactive hyperemia with either VOP or HRU are reproducible, noninvasive, and have been validated by direct comparison with results obtained in studies conducted with intrabrachial arterial infusion of vasoactive substances. Therefore, both techniques have been used widely to evaluate vascular reactivity and endothelial function in clinical research settings 25, 28–31. VOP and HRU often are employed concomitantly to evaluate vascular endothelial function during dynamic testing, and it is reassuring to know that comparable directional and quantitative changes have been observed with both techniques 83. The use of HRU in obese subjects can be technically very difficult and caution is recommended with the interpretation of the data. Also, since the repeated measurements of arterial diameter require precise definition and reproduction of the location of the arterial wall and anatomical landmarks, it is suggested that these determinations be performed by at least two independent observers 30. An additional method used to evaluate endothelial function is assessment of the skin microcirculation with laser Doppler technology 44. Measurement of changes in skin blood flow after the administration of either acetylcholine chloride (endothelium-dependent) or sodium nitroprusside (endothelium-independent) by iontophoresis has disclosed abnormalities in the microcirculation 25. This procedure does not provide absolute quantitation of the amount of blood flow through the skin area of interest; it only provides a qualitative assessment of the cutaneous hyperemic response to the administered stimulus. Although determination of skin blood flow using Doppler may be helpful as an adjunctive tool, the results are sometimes inconsistent 28 and must be interpreted with some caution. Moreover, the control mechanisms that regulate the cutaneous microcirculation differ from those that regulate the larger arterial capacitance 63, and observations made in the skin cannot be extrapolated to other microcirculatory systems in the body. Reliable measurement of skin blood flow using Doppler technology requires an experienced and well-trained operator, the selected skin area of interest must be precisely delineated and reproduced following pharmacological interventions, and room temperature must be kept constant during the entire experiment.

In summary, the application of techniques capable of measuring blood flow and vascular diameter with dynamic testing has allowed investigators to detect subtle and early vascular abnormalities in individuals at risk for the development of atherosclerosis. Impaired vascular reactivity is important in the pathogenesis and prognosis of cardiovascular disease. Although other noninvasive modalities to evaluate cardiovascular risk, such as electron beam and computerized tomography have been used to provide prognostic implications about atherosclerotic cardiovascular disease physiologic vascular testing of the endothelium is evolving into an important adjunctive tool for assessing cardiovascular risk and response to treatment.

Clinical considerations

Cardiovascular disease affects nearly 60% of adults over the age of 65. However, the process of atherosclerosis begins in childhood 84. The earliest documented histological signs are lipid deposits in the intimal layer of systemic arteries. These fatty streaks can be found in the aorta of children as early as 3 years of life and in the coronary arteries by adolescence 84. Endothelial dysfunction, which represents one of the earliest abnormalities in the development of atherosclerosis, precedes any histologic change and also contributes to the progression of the disease in later stages 40. With advancing age and progression of the vascular disease, the structural damage to the arterial tree, i.e. thickening of the intimal media layer and subendothelial lipid deposits further impairs endothelial function, leading to a progressive deleterious spiral of events, which leads to narrowing and eventual occlusion of major arteries. In vitro experiments have demonstrated that the vascular endothelium loses its natural protective barrier and anti-inflammatory properties early in the atherosclerotic process, prior to any detectable accumulation of subendothelial fatty streaks and long before the disease becomes clinically evident 85, 118. Moreover, abnormal endothelial-mediated vasodilation and paradoxical vasoconstriction contribute to impaired circulation and cause tissue ischemia 86. High shear stress (hypertension) and nicotine (smoking), as well as chronic exposure to elevated circulating levels of apo B-100-containing lipid particles, have been shown to alter endothelial function and accelerate the atherosclerosis process. In addition to altered capillary permeability, the increase in circulating levels of oxidized and glycosylated small dense LDL particles, common in conditions of insulin resistance, are of particular concern, because they more readily cross the endothelial cell layer and accumulate in the subendothelial region.

Although not all phases of the atherosclerotic process have been completely elucidated, much evidence indicates that metabolism, i.e. insulin resistance and hyperglycemia, vascular endothelial dysfunction, and inflammation play critical roles in the development of atherosclerosis, and that all these processes begin early in life. Autopsy findings in the Bogalusa Heart Study have demonstrated a close association between the severity of asymptomatic coronary artery disease and cardiovascular risk factors in children and young adults, ages 2 to 39 years 84. Other studies underscore the association between endothelial dysfunction and the number of cardiovascular risk factors 87–89. A close relationship between abnormal vascular reactivity in peripheral arteries and the presence of coronary artery disease has been demonstrated 90, 91. Endothelial dysfunction 89 and elevated circulating levels of inflammatory markers have been linked to cardiovascular morbidity and mortality 92, 93. On the basis of these data, efforts to prevent/slow the progression of atherosclerotic cardiovascular disease must focus on (1) correction of the disturbances in carbohydrate and lipid metabolism and amelioration of the insulin resistance, (2) restoration of normal vascular reactivity and endothelial function, and (3) reduction of tissue/cellular inflammatory responses. An important and challenging unanswered question centers on the time at which the intervention should be instituted. Can we achieve superior results by implementing in younger adults and children the same lifestyle modifications that are known to slow the progression of the process of atherosclerosis in high-risk adults?

Recent clinical trials 94–96 have shown the unequivocal benefits of aggressive cholesterol lowering in high-risk adults, and reduced cardiovascular morbidity and mortality has been documented with intensive insulin therapy and maintenance of near-normoglycemia in hospitalized patients 97, 98. In type 2 diabetic patients improvement in cardiovascular risk factors, decreased morbidity and mortality, regression of carotid intimal thickness, and prevention of stenosis of coronary stents have been demonstrated with ‘insulin sensitizing’ agents 79, 99–101, 117. A reduction in the incidence of diabetes mellitus with agents that inhibit the ‘renin-angiotensin-aldosterone system’ 102–104 also has been reported. Antihypertensive, lipid-lowering and antidiabetic 105–109 agents have been associated with improved vascular reactivity, diminished insulin resistance, and reduced circulating levels of inflammatory markers 110. The beneficial effects of physical activity and dietary modification on endothelial dysfunction and insulin resistance also have been documented 111–116. It is conceivable that many of these interventions may affect different mechanisms and pathways involved in the regulation of vascular reactivity and the inflammatory response. As demonstrated in the Steno II trial 95, the combined use of pharmacologic agents that target multiple risk factors and lifestyle modifications can produce striking cardiovascular benefits. While there is general agreement that restoration of normal vascular reactivity, reduction in insulin resistance and attenuation of the vascular inflammatory response lead to a reduction in cardiovascular morbidity and mortality in high-risk adult subjects, the ideal timing for implementation of these interventions remains to be determined. It is likely that the earlier the institution of an intervention, the greater will be the benefit of altering the natural course of the disease. Nevertheless, this remains to be tested, and clinical studies designed to promote healthy lifestyle, dietary modifications and pharmacologic interventions in younger, ‘unhealthy’ adults and children are needed.

Conclusion

In summary, endothelial dysfunction and insulin resistance commonly occur together and can be detected early in the pathogenesis of atherosclerosis. Insulin resistance can be inferred by the presence of a cluster of metabolic and cardiovascular abnormalities known collectively as the metabolic syndrome3 or by direct measurement of impaired insulin-stimulated glucose and lipid metabolism 10, 12. Endothelial dysfunction can be documented by the demonstration of inadequate vasodilation and/or paradoxical vasoconstriction in coronary and peripheral arteries. Lack of endothelial-derived NO may provide the link between insulin resistance and endothelial dysfunction 69. NO deficiency results from decreased synthesis and release, combined with an exaggerated consumption in tissues exposed to an overabundance of ROS and RNS, which accumulate secondary to disturbances in carbohydrate and lipid metabolism. Experimental evidence supports the notion that therapies that reduce cardiovascular morbidity and mortality are associated with improvements in insulin resistance and endothelial dysfunction. Moreover, a reduction in insulin resistance ameliorates endothelial dysfunction, and improved tissue sensitivity to insulin improves vascular endothelial function. However, most of these observations have been made in adult subjects, and information about the impact of these interventions in younger ‘healthy’ subjects is lacking. Aggressive therapy aimed simultaneously at improving insulin-stimulated glucose and lipid metabolism and endothelial dysfunction represents an important clinical option to delay the appearance and retard the progression of atherosclerosis. Current evidence indicates that interventions that (1) normalize carbohydrate and lipid metabolism, (2) ameliorate insulin resistance, (3) restore vascular reactivity, (4) reduce blood pressure, and (5) inhibit the procoagulant and inflammatory responses in adults at high risk of developing cardiovascular disease lead to reduced cardiovascular morbidity and mortality. Whether these benefits will be observed if the same prevention strategies are applied to younger, high-risk individuals remains to be investigated.

Acknowledgements

This work was supported in part by grants from the American Diabetes Association and the Kronkowsky Foundation.