|Home | About | Journals | Submit | Contact Us | Français|
Cardiovascular disease is a major complication of diabetes mellitus, and improved strategies for prevention and treatment are needed. Endothelial dysfunction contributes to the pathogenesis and clinical expression of atherosclerosis in diabetes mellitus. This article reviews the evidence linking endothelial dysfunction to human diabetes mellitus and experimental studies that investigated the responsible mechanisms. We then discuss the implications of these studies for current management and for new approaches for the prevention and treatment of cardiovascular disease in patients with diabetes mellitus.
Type 2 diabetes mellitus is a growing public health problem  and a major cause of cardiovascular disease in the United States . Type 2 diabetes is associated with systemic insulin resistance, which promotes hyperglycemia and dyslipidemia , and it has been proposed that these metabolic abnormalities account for increased cardiovascular risk. Endothelial dysfunction contributes to the pathogenesis and clinical expression of atherosclerosis and has been linked to Type 2 diabetes mellitus and insulin resistance in experimental and clinical studies . This article will review the concept of endothelial dysfunction and the evidence linking it to human diabetes mellitus. We will then review mechanisms of endothelial dysfunction in diabetes mellitus and the implications of this work for the management of patients with diabetes mellitus.
Once thought to be simply a passive lining for blood vessels, it is now recognized that the vascular endothelium is a key determinant of vascular health. Broadly speaking, the term “endothelial dysfunction” refers to an impairment of the ability of the endothelium to properly maintain vascular homeostasis . Although the term is often used in reference to a loss of bioavailable nitric oxide (NO), endothelial dysfunction also reflects increased production of vasoconstrictors and disturbed regulation of inflammation, thrombosis, and cell growth in the vascular wall [5, 6]. Numerous studies have linked endothelial dysfunction and resultant atherosclerosis with insulin resistant states such as obesity and diabetes [7–10].
The endothelium plays a key role in the regulation of arterial tone and blood flow. In this regard, the endothelium orchestrates the production of vasodilator molecules such as NO, prostacyclin, and endothelium-derived hyperpolarizing factor (EDHF), and vasoconstrictors, including endothelin-1 (ET-1) and angiotensin II . Stimuli for production of endothelium-dependent vasodilators include physiologically relevant factors such as acetylcholine, thrombin, serotonin, angiotensin II, and alpha adrenergic agonists. In general, these factors also promote vasoconstriction via direct effects on vascular smooth muscle. Endothelium-derived NO and other vasodilators oppose such vasoconstrictor effects, and thus act in a homeostatic fashion to maintain normal arterial patency and compliance despite local production of vasoconstrictors. When the endothelium is dysfunctional, the vasoconstrictor effects are unopposed and arterial tone is increased. In addition, pathological states are associated with increased endothelial production of endothelin-1 and other endothelium-derived vasoconstrictors that may further promote vasospasm and increase arterial stiffness.
Another key stimulus for endothelial production of NO is increased shear stress, which is the frictional force at the endothelial surface produced by flowing blood . Shear stress relates directly to arterial flow and inversely to the third power of the arterial diameter, and thus for a given level of flow, a small change in diameter has a large effect on local shear stress. In healthy conduit arteries, an increase in arterial flow stimulates “flow-mediated dilation”. The resultant increase in lumen diameter acts in a homeostatic manner to limit the increase in shear stress that results from increased flow. This response can be measured in humans using non-invasive methods, including ultrasound imaging of the brachial artery diameter during reactive hyperemia . Flow-mediated dilation has emerged as an important indicator of endothelial function in clinical studies of diabetes mellitus and vascular disease [5, 12].
In addition to responding to acute changes in flow by stimulating vasodilation, the endothelium also is responsible for chronic changes in arterial structure and lumen dimension that are produced by chronic changes in blood flow [13, 14]. Like flow-mediated dilation, flow-induced remodeling is stimulated by changes in shear stress. The remodeling response to flow involves a complex interplay between vasodilator factors, local inflammation, and factors that modify the intercellular matrix [13, 15–17]. Impaired remodeling has been observed in animal models of diabetes  and may, in part, explain why atherosclerosis tends to produce diffusely narrowed and small caliber arteries in patients with diabetes .
Another important function of the endothelium is the regulation of tone in smaller resistance vessels that control blood flow and maintain the balance between blood supply and tissue demand. This complex process depends in large part on the production of non-endothelium-dependent vasodilators, such as adenosine, that dilate resistance vessels and increase tissue blood flow in response to increased oxygen demand [19, 20]. However, endothelium-derived NO also contributes to ischemia-mediated vasodilation and the hyperemic response to exercise . By this mechanism, endothelial dysfunction may impair the regulation of blood flow and contribute to impaired exercise capacity in certain pathological states, including heart failure and peripheral arterial disease.
The importance of inflammation for the pathogenesis of atherosclerosis is well recognized . Under physiological conditions, NO prevents leukocyte adhesion and maintains the endothelium in a quiescent, anti-inflammatory state . In the presence of risk factors, the endothelium can be activated to express adhesion molecules, such as vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1), which are required for the adhesion of leukocytes to the endothelial surface . The activated endothelium also expresses chemotactic factors, such as monocyte chemoattractant protein-1, and other proinflammatory cytokines, like macrophage colony-stimulating factor, and tumor necrosis factor-beta (TNF-β) . Endothelial expression of these factors contributes to the development of inflammation within the arterial wall and promotes atherogenesis .
In addition to regulating vessel wall inflammation, the vascular endothelium produces a host of other molecules that affect blood fluidity and thrombosis [5, 24, 25]. Endothelial production of pro-thrombotic molecules such as plasminogen activator inhibitor-1 (PAI-1), thromboxane, tissue factor, and von Willibrand’s factor (vWF), is balanced by the production of antithrombotic molecules such as NO, heparans, prostacyclin, tissue plasminogen activator, and thrombomodulin. Risk factors, including diabetes mellitus, are associated with a shift in this balance toward a pro-thrombotic, anti-fibrinolytic state.
Thus, the dysfunctional endothelium contributes to the pathogenesis of atherosclerotic vascular disease by promoting inflammation, thrombosis, arterial stiffness, and impaired regulation of arterial tone and flow. Several lines of evidence support the relevance of endothelial dysfunction to human disease, and most of these studies have focused on endothelium-dependent vasodilation as a surrogate measure of endothelial health. First, endothelium-dependent dilation is impaired in patients with documented coronary artery disease [26, 27] and in patients with classical and with more recently-recognized risk factors [28, 29]. Indeed, such abnormalities are detectable very early in the course of the disease before measurable lesions or clinical symptoms . Second, lifestyle changes and drugs proven to reduce cardiovascular risk have been shown to reverse endothelial dysfunction, including exercise, smoking cessation, weight loss, cholesterol-lowering drugs, and angiotensin converting enzyme inhibitors . Finally, endothelial dysfunction in the coronary or peripheral circulations predicts increased risk for cardiovascular events in patients with risk factors and in patients with established atherosclerosis [5, 31–38]. Overall, it is clear that endothelial dysfunction relates to the pathogenesis of cardiovascular disease. Since diabetes mellitus is a major cardiovascular disease risk factor, it is not surprising that clinical studies have confirmed a relationship between diabetes mellitus and endothelial dysfunction.
A large body of evidence links endothelial dysfunction to human diabetes mellitus. As has been reported for other cardiovascular risk factors, cross-sectional studies show reduced endothelium-dependent vasodilation in coronary and peripheral arteries of patients with Type 1 [39, 40] and Type 2 diabetes mellitus [41–44]. Endothelial dysfunction is also observed in conditions associated with Type 2 diabetes including obesity [29, 45, 46], sedentary lifestyle, and the metabolic syndrome [47–49]. In addition to impaired vasodilator function, diabetes is also associated with increased circulating levels of endothelium-derived adhesion molecules and plasminogen activator inhibitor-1 [50–52], reflecting a pro-inflammatory and pro-thrombotic endothelial phenotype (Table 1).
Insulin resistance is the defining mechanism of Type 2 diabetes mellitus, and several studies have examined the relationship between insulin resistance and endothelial dysfunction. In this regard, degree of insulin sensitivity, as determined by euglycemic clamp, correlates with increases in skin blood flow induced by acetylcholine in obese women  and with brachial artery flow-mediated dilation in non-diabetic subjects [54, 55]. A higher insulin response during oral glucose tolerance test, is likewise associated with endothelial dysfunction in the coronary [56, 57] and forearm circulations [58, 59]. Higher plasma insulin levels correlate inversely with brachial artery flow-mediated dilation . Finally, homeostasis model assessment-insulin resistance (HOMA-IR) correlates inversely with the leg blood flow response to methacholine in non-diabetic subjects with varying levels of body mass index  and with skin blood flow responses to acetylcholine in diabetic compared to non-diabetic subjects . In Framingham Heart Study participants, brachial artery flow-mediated dilation correlated inversely with HOMA-IR, although this relation was lost after adjusting for factors associated with the metabolic syndrome .
Interestingly, endothelial dysfunction may precede the development of diabetes mellitus. In this regard, healthy non-diabetic subjects who have a first degree relative with Type 2 diabetes mellitus display impaired endothelium-dependent vasodilation as well as increased plasma markers of endothelial cell activation [54, 61, 62]. Blood markers of endothelial activation and systemic inflammation are also elevated in non-diabetic offspring with evidence of insulin resistance by glucose tolerance test . In addition to these cross-sectional studies, prospective studies have shown that blood markers of endothelial activation predict incident Type 2 diabetes mellitus after adjusting for other risk factors, including body mass index, level of physical activity, lipids, family history of diabetes mellitus, and glucose tolerance [52, 63]. Similarly, impaired flow-mediated dilation  and polymorphisms of endothelial NO synthase (eNOS) are multivariable predictors of incident Type 2 diabetes mellitus . The occurrence of endothelial dysfunction prior to the development of Type 2 diabetes mellitus suggests that there are common pathophysiological mechanisms and raises that possibility of a causal link between insulin resistance and endothelial dysfunction.
Additional clinical evidence for a link between endothelial dysfunction and insulin resistance is provided by intervention studies that demonstrate improved endothelial function following treatments that improve insulin sensitivity. For example, rosiglitazone  and troglitazone  improve endothelial function in forearm microvessels of patients with Type 2 diabetes mellitus. Metformin improves endothelium-dependent dilation in patients with Type 2 diabetes mellitus  and in patients with the metabolic syndrome . Pioglitazone improves endothelial function in non-diabetic patients with hypertension or hypercholesterolemia with or without insulin resistance at baseline , and rosiglitazone improves flow-mediated dilation and decreases markers of inflammation in completely healthy subjects without obesity, risk factors, or family history of diabetes . Less specific interventions associated with improved insulin sensitivity and or decreased risk of diabetes mellitus also improve endothelial function including weight loss, exercise, and inhibitors of the renin-angiotensin system [45, 46, 72, 73].
Overall, it is clear that human diabetes mellitus is associated with abnormalities of endothelial function in humans, which may explain, in part, increased risk for cardiovascular disease in diabetic patients. An improved understanding of the mechanisms of endothelial dysfunction in this setting could provide new approaches for patient management.
One important mechanism of endothelial dysfunction in diabetes mellitus is alteration of the signaling pathways that lead to eNOS activation in the endothelium, a process that has been extensively studied. Briefly, NO production in endothelial cells depends on the enzymatic conversion of Larginine to NO and citrulline by eNOS. The enzyme is constitutively expressed in endothelial cells and is localized to caveolae, which are specialized invaginations of the plasma membrane that are rich in specific lipids and proteins, including caveolin-1 . eNOS has a low level of basal activity because of its association with caveolin-1, but it can be activated within seconds following stimulation of endothelial cells with receptor-dependent agonists such as acetylcholine and serotonin that increase intracellular calcium and promote calcium/calmodulin-dependent-displacement of caveolin-1 and activation of the enzyme . Activation of eNOS is further enhanced by other protein-protein interactions, including association with heat shock protein 90 . Alternatively, eNOS can be activated by bradykinin, estrogen, and shear stress via activation of the phosphoinositide-3 kinase (PI3 kinase)/Akt system. Akt phosphorylates eNOS at serine 1177, which directly increases activity and also enhances the response to other activators [77–80]. Once produced, eNOS-derived NO diffuses locally in the arterial wall and activates guanylyl cyclase in vascular smooth muscle cells, platelets, and endothelial cells to induce its biological effects .
Relevant to endothelial dysfunction in diabetes, it is interesting that insulin itself is an important stimulus for eNOS activation (Fig. 1). Binding of insulin to its receptor on endothelial cells leads to phosphorylation of insulin receptor substrate-1 (IRS-1) and subsequent phosphorylation and activation of eNOS via PI3 kinase/Akt [8, 82]. Support for this mechanism is derived from studies showing that insulin-stimulated NO production is blocked by inhibitors of PI3 kinase or Akt [77, 83]. In addition, mutations in IRS-1 decrease insulin-stimulated eNOS phosphorylation and eNOS gene expression in cultured endothelial cells . Furthermore, mice with endothelial specific knockout of the insulin receptor display decreased eNOS expression and impaired endothelial vasodilator function [85–87]. Animal models of insulin resistance, including the obese Zucker rat display defects in the PI3 kinase/Akt system and impaired NO bioavailability .
Insulin activates other cellular signaling pathways in addition to PI3 kinase/Akt. For example, insulin activates the mitogen-activated protein kinases (MAPK) via the small GTPase Ras . The Ras/MAPK insulin-signaling pathway generally leads to cellular growth and proliferation. In endothelial cells, activation of this pathway has been linked to the expression of endothelin-1, which is a potent vasoconstrictor and mitogen, and to the expression of pro-inflammatory adhesion molecules such as ICAM-1 . In the setting of diabetes and insulin resistance, insulin-mediated activation of eNOS via PI3 kinase/Akt is inhibited, while the adverse effects of insulin remain unopposed, which may promote vascular disease [90, 91]. Flow and other physiological stimuli activate eNOS via PI3 kinase/Akt, and thus, an impairment of this signaling mechanism in diabetes may have broad implications for vascular dysfunction. Furthermore, study of flow-mediated dilation may have particular relevance in studies of the mechanisms of vascular disease in human diabetes mellitus.
Human studies support the relevance of these mechanisms. In healthy humans, insulin infusion stimulates vasodilation and increases blood flow in peripheral tissues, and this effect is blunted in patients with diabetes mellitus and insulin resistance [92, 93]. In insulin clamp experiments, the time course and magnitude of insulin-mediated vasodilation appear to be closely linked to the rate of glucose uptake in the limb [94–96]. These findings suggest that the actions of insulin in tissues and the vasculature are tightly coupled and that an impaired ability to dilate in response to insulin may contribute to impaired glucose metabolism .
Additional human studies further support the relevance of altered insulin signaling to endothelial dysfunction. Akt phosphorylation and eNOS expression are impaired in arterial tissue collected from diabetic patients at the time of coronary bypass surgery . As observed in experimental systems, insulin infusion increases circulating levels of endothelin-1 in humans, a response that may be increased in the setting of insulin resistance . Thus, selective impairment of PI3 kinase/Akt signaling characterizes human insulin resistance and may contribute to endothelial dysfunction and vascular disease in Type 2 diabetes mellitus .
Diabetes mellitus alters a number of other mechanisms related to eNOS activation. For example, diabetes and obesity are associated with increased expression of caveolin-1 in the endothelium in experimental animals [99, 100] and in adipose tissue of diabetic patients . This effect is associated with impaired activation of eNOS and impaired vasodilator function. Diabetes also impairs the interaction between eNOS and heat shock protein 90, which also impairs NO production . Finally, diabetes mellitus is associated with increased levels of endogenous inhibitors of eNOS, such as asymmetric dimethyl arginine (ADMA) . Thus, diabetes has multiple effects on the cell signaling and enzyme activity that result in an impaired ability to produce NO in response to physiological stimuli.
Increased oxidative stress in the vasculature is another important mechanism of endothelial dysfunction in diabetes mellitus and associated conditions. Exposure of arterial tissue to increased glucose or free fatty acid concentrations induces superoxide production and impairs NO bioavailability in the vascular wall, while antioxidant treatment acts to restore endothelial function under these conditions[104, 105]. Increased oxidative stress has the potential to impair NO bioavailability in several ways [106, 107]. First, superoxide anion may react with NO to form peroxynitrite and eliminate the biological activity of NO. Peroxynitrite is a highly reactive oxidant that may alter the function of a variety of cellular enzymes . In particular, peroxynitrite can alter the catalytic activity of eNOS in endothelial cells and guanylyl cyclase in vascular smooth muscle cells. As a result, peroxynitrite reduces both the production of NO and the responsiveness of target tissues to NO [109, 110]. Increased production of reactive oxygen species may also influence the redox status of critical eNOS co-factors, including tetrahydrobiopterin. Loss of tetrahydrobiopterin uncouples eNOS and promotes production of superoxide, rather than NO . Reactive oxygen species may increase lipid peroxidation products that interfere with receptor dependent activation of eNOS, inactivate NO, and decrease the responsiveness of target tissues [112, 113].
A number of enzymatic sources of superoxide in the diabetic vasculature have been identified. Nicotinamide adenine dinucleotide phosphate-oxidase (NADPH oxidase) is a membrane-associated multi-subunit complex that generates superoxide anion and is involved in the oxidative burst in inflammatory cells and normal cell signaling in endothelial cells . Under pathological conditions, including diabetes mellitus, NADPH oxidase activity and superoxide production is increased [114, 115]. Increased free fatty acid concentration activates NADPH oxidase and promotes activation of the pro-inflammatory transcription factor NFκB . NADPH oxidase expression is also upregulated by angiotensin II , and it is interesting that angiotensin converting enzymes inhibitors have favorable vascular effects in diabetes [118, 119].
Uncoupled eNOS is another important source of oxidative stress in the diabetic vasculature. Under physiological conditions, eNOS exists as a dimer and produces NO, however, the enzyme reduces oxygen to superoxide anion when there is decreased availability of the cofactor tetrahydrobiopterin . Diabetes is associated with eNOS uncoupling and decreased tetrahydrobiopterin levels. Since tetrahydrobiopterin is readily oxidized, a vicious cycle is established whereby oxidative stress leads to eNOS uncoupling and increased production of superoxide anion, which in turn may reduce the availability of tetrahydrobiopterin and promote further oxidative stress. Consistent with this mechanism, tetrahydrobiopterin supplementation improves NO production and endothelial function in experimental models [121, 122] and in human subjects with Type 2 diabetes mellitus . A number of other sources of reactive oxygen species have been identified. For example, hyperglycemia and diabetes are associated with increased production of superoxide anion by the aldose reductase system and, as discussed below, components of the mitochondrial electron transport chain .
Studies in humans support the importance of increased oxidative stress as a mechanism of endothelial dysfunction in diabetes mellitus and insulin resistance. For example, circulating markers of oxidative stress, including F2 isoprostanes and antibodies against oxidized low density lipoprotein, are increased in humans with diabetes, obesity, and insulin resistance [125, 126]. Clinical studies have shown improved endothelial function with antioxidant treatment. In this regard, infusion of ascorbic acid in concentrations sufficiently high to scavenge superoxide anion, improves endothelial function in the forearm of patients with diabetes mellitus . Alpha tocopherol treatment improved endothelial function in the coronary arteries of patients with Type 1 diabetes mellitus, although there was no benefit in Type 2 diabetes .
Although oxidative stress is clearly a contributing mechanism and antioxidants may improve endothelial function in certain situations, clinical trials have failed to show a benefit of alpha tocopherol therapy on outcomes in large-scale clinical trials [128, 129]. These findings may have important implications for prevention and management of vascular disease in diabetes mellitus. A strategy designed to inhibit enzymatic sources of reactive oxygen species is likely to provide greater benefit than treatment with scavengers of reactive oxygen species such as ascorbic acid, alpha tocopherol, and other antioxidant compounds. In support of this concept, angiotensin converting enzyme inhibitors and angiotensin receptor blockers limit angiotensin II-induced expression of NADPH oxidase and decrease superoxide anion production in the endothelium . These drugs have been shown to improve endothelial function and reduce cardiovascular risk in patients with diabetes mellitus .
Atherosclerosis is recognized to be an inflammatory disease [21, 22] and the vascular endothelium is both affected by and contributes to the inflammatory process . Endothelial cells can be “activated” by pro-inflammatory factors, including tumor necrosis factor-alpha (TNF-α) and C-reactive protein to promote an atherogenic phenotype [132, 133]. The activated endothelium expresses adhesion molecules and other factors that accelerate the inflammatory process. An important consequence of endothelial activation is decreased expression of eNOS and loss of NO bioactivity. In this regard, inflammatory mediators including TNF-α decrease expression of eNOS in cultured endothelial cells . In humans, acute inflammatory states, such as infusion of low-dose lipo-polysaccharide or vaccination have been shown to impair endothelium-dependent vasodilation [135, 136]. On the basis of such studies, investigators argue that a broad alteration of endothelial function, including loss of NO under pro-inflammatory conditions, might be a critical mechanism that links systemic inflammation to atherosclerosis .
Diabetes is associated with a systemic inflammatory state that may impair endothelial function and contribute to atherosclerosis . In experimental systems, increased concentrations of glucose or free fatty acids activate the endothelium [116, 139, 140]. Patients with diabetes mellitus or obesity have increased circulating levels of inflammatory markers, including C-reactive protein, TNF-α, interleukin-6, and intercellular adhesion molecule-1 [50, 141–143]. Furthermore, increased levels of inflammatory markers predict cardiovascular risk in diabetic patients . Interestingly, increased levels of circulating inflammatory markers also relate to the incidence of new diabetes [52, 145–147].
The transcription factor NFκB is a key regulator of endothelial activation and, interestingly, has also been linked to the pathogenesis of insulin resistance [148, 149]. NFκB is activated by free fatty acids, inflammatory cytokines, and the receptor for advanced glycation end products (RAGE) [8, 150–152]. NFκB activation involves phosphorylation and subsequent degradation of the inhibitory subunit IκB by IκB kinase (IKK-β), which allows translocation of the regulatory subunits p50 and p65 to the nucleus, where they promote expression of inflammatory genes. In skeletal muscle, TNF-α or over expression of IKK-β produces insulin resistance . Conversely, genetic suppression or pharmacological inhibition of IKK-β with salicylates prevents insulin resistance [148, 154]. Studies in cultured endothelial cells and experimental animals support links between activation of NFκB, development of an inflammatory phenotype, insulin resistance, and impaired bioactivity of NO [155, 156].
Several recent human studies support the clinical relevance of these mechanisms. Treatment of obese human subjects with salsalate improved insulin sensitivity and reduced circulating markers of inflammation . Increased expression of p65 and decreased abundance of IκB, reflecting activation of NFκB has been observed in endothelial cells isolated from older and obese individuals, and these findings relate to impaired endothelium-dependent vasodilation [158, 159]. Salsalate treatment of obese human subjects reduced NFκB activation in freshly isolated endothelial cells . This effect was associated with improved insulin sensitivity and endothelium-dependent vasodilation . Such studies raise the interesting possibility that treatment of diabetics with salsalate or other anti-inflammatory drugs might improve glucose control and reduce the risk for vascular disease. Clinical trials are currently underway to address these possibilities (TINSAL-CVD and TINSAL-T2D).
Activation of protein kinase C beta (PKCβ) may explain the links between inflammation, endothelial dysfunction, and insulin resistance in diabetes mellitus [161, 162]. The PKC’s are a family of serine/threonine kinases that act at the plasma membrane in the regulation of signal transduction in a wide variety of cell types. PKCβ is an important isoform in endothelial cells and is activated by diacylglycerol under conditions of increased glucose and fatty acid concentrations [163, 164]. Interestingly, accumulation of diacylglycerol in this setting has been attributed to impaired mitochondrial substrate utilization . PKCβ inhibits PI3 kinase and Akt, thereby reducing eNOS phosphorylation [166, 167]. PKCβ also activates NFκB [168, 169]. Inhibiting PKCβ improves NO bioavailability and reduces inflammatory activation of the endothelium in experimental models [170–174]. In humans, treatment with a PKCβ inhibitor prevented the development of endothelial dysfunction following glucose infusion in the forearm of healthy volunteers  and improved brachial artery flow-mediated dilation in patients with diabetes mellitus .
Recent studies have shed light on abnormalities of mitochondrial function as a proximate mechanism of increased oxidative stress and PKC activation in the diabetic vasculature. Although well-recognized for their role in the production of adenosine-triphosphate (ATP), mitochondria have many other cellular functions including a role in cell signaling via production of reactive oxygen species [177–179]. While most of the oxygen consumed by mitochondria relates to ATP generation, 1–2% is converted to superoxide anion under physiological conditions . Mitochondrial-derived reactive oxygen species affect cell growth, differentiation, and programmed cell death [177, 179, 181–185]. Relevant to diabetes and insulin resistance, mitochondrial-derived reactive oxygen species play a role in the activation of AMP kinase a central regular of cellular energy status[185, 186]. Interestingly, mitochondrial-derived hydrogen peroxide contributes to endothelium-dependent dilation in response to shear stress in specific vascular beds . While physiological levels of mitochondria-derived reactive oxygen species function in normal signaling, increased levels have pathological effects in diabetes . In support of this concept, mitochondria-directed antioxidants, including lipoic acid, reduce radical production, improve insulin sensitivity, increase Akt activation, and improve NO-mediated vasodilation [188–191].
There is growing recognition of the importance of mitochondrial biogenesis and dynamics for energetics and reactive oxygen species production in diabetes (Fig. 2) [192–194]. Formation of new mitochondria (biogenesis) is regulated by peroxisomal proliferator activator receptor gamma coactivator-1alpha (PGC-1α) and nuclear respiratory factor-1 (NRF-1). Interestingly, this process depends on eNOS and bioavailable NO [195–198]. Mitochondria have a life span lasting hours to days, and as part of their life cycle, undergo cycles of fusion to form complex networks and fission to form smaller individual mitochondria. The balance between these processes is referred to as mitochondrial dynamics [192, 199]. Fusion may be beneficial and allow for distribution of metabolites and DNA throughout the network. Toward the end of the life cycle, dysfunctional daughter mitochondrial produced by fission undergo autophagy eliminating them from the cell. For this reason, fission is an adaptive process that sequesters and targets damaged mitochondrial components for elimination . Under pathological conditions, however, there is a shift toward fission and inhibition of autophagy leading to the loss of mitochondrial networks and accumulation of dysfunctional mitochondria in the cell. These dysfunctional mitochondria have increased production of reactive oxygen species and impaired production of ATP [192, 200–202].
Diabetes and insulin resistance are strongly linked to abnormalities of mitochondrial function [203–205]. Decreased fatty acid oxidation by mitochondria and/or decreased mitochondrial mass may lead to increased diacylglycerol concentrations and activation of PKC, which, as described above, blocks insulin signaling and activates NFκB [148, 203]. Increased glucose concentrations increase mitochondrial membrane potential and radical production [124, 206]. Mitochondrial-derived reactive oxygen species may damage mitochondrial DNA (mtDNA), which lacks protective histones, leading to impaired expression of oxidative phosphorylation enzymes and decreased substrate utilization . Mitochondrial uncoupling proteins (UCP’s) act to prevent membrane hyperpolarization and limit superoxide production , and over-expression of uncoupling proteins in skeletal muscle prevents the development of diet-induced obesity and diabetes . Diabetic conditions impair mitochondrial biogenesis, mitochondrial fusion, and autophagy, leading to cells with decreased mitochondrial mass and predominance of fragmented and dysfunctional mitochondria [205, 208]. Finally, interventions that promote mitochondrial biogenesis, such as resveratrol and other activators of the histone deacetylase SIRT1, have been shown to improve insulin sensitivity .
Clinical studies have demonstrated impaired mitochondrial function in patients with diabetes. For example, diabetics  and offspring of diabetic patients have decreased oxidative phosphorylation  and decreased expression of oxidative phosphorylation genes in skeletal muscle [212, 213]. Skeletal muscle from patients with diabetes is characterized by smaller mitochondria, decreased mitochondrial mass, and decreased expression of genes related to mitochondrial biogenesis [204, 212, 214–216]. Anti-diabetic changes in lifestyle, including exercise and calorie restriction lead to an increase in mitochondrial biogenesis [204, 217]. Recent preliminary work demonstrated links between endothelial dysfunction and impaired mitochondrial biogenesis and increased mitochondrial superoxide production in arterioles isolated from patients with diabetes mellitus versus healthy controls . Thus, experimental studies have firmly established connections between mitochondrial dysfunction and diabetes, and these mechanisms appear to contribute to endothelial dysfunction in human diabetes mellitus.
Type 2 diabetes mellitus is a major risk factor for cardiovascular disease. This condition is associated with insulin resistance and related metabolic abnormalities, including hyperglycemia, hypertension, visceral adiposity, and dyslipidemia with low HDL and elevated triglycerides and free fatty acids [3, 219, 220]. Current efforts to reduce cardiovascular disease focus on risk factor reduction, but diabetics continue to have increased risk despite aggressive interventions [220–222]. Along these lines, intensive glucose control has disappointing effects on the incidence of cardiovascular events [223–226]. In addition, promising therapies to improve insulin sensitivity and glucose control have unexpectedly been associated with increased cardiovascular risk [227–229]. Thus, there is a need for new approaches to the prevention and management of cardiovascular risk in diabetes.
As we have reviewed, a key mechanism of diabetic vascular disease is the development of endothelial dysfunction, which is characterized by a loss of NO and development of a pro-inflammatory vascular phenotype that promotes atherosclerosis and cardiovascular events. Recent experimental work has shed light on the mechanisms of endothelial dysfunction, which include impaired cell signaling required for activation of eNOS, increased oxidative stress, activation of pro-inflammatory signaling mechanisms, and activation of PKCβ. A proximate mechanism that may unify many aspects of endothelial dysfunction is mitochondrial dysfunction.
To date, these insights have only partially been translated into useful management strategies. Studies of endothelial function in diabetes support several important recommendations about patient management (Table 2). In spite of the prominent role played by oxidative stress in diabetes and atherosclerosis, antioxidant therapy had no benefit in large randomized trials [128, 129, 230]. Therapy directed against enzymatic sources of superoxide anion appears to have greater utility for the prevention of vascular disease. In addition to favorable effects on the renal vasculature, decreased superoxide production by NADPH oxidase may explain the observed benefits of angiotensin converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARB’s) against cardiovascular disease. This mechanism may provide further rationale for the preferential use of these drugs rather than other hypertensive agents in the management of hypertension in diabetic patients. Improved endothelial function, reduced radical production, and inhibition of pro-inflammatory mechanisms may also explain the well-established benefits of statin therapy in diabetes.
The reviewed studies of the mechanisms of endothelial dysfunction in diabetes may point to new management strategies for the prevention of cardiovascular disease in diabetes. Anti-inflammatory drugs and PKCβ inhibitors appear to hold great promise and clinical studies are in progress, but these strategies have not yet been established as safe and effective. Recent work on mitochondrial dysfunction suggests possible new directions for therapy. Drugs that have favorable effects on excess mitochondrial superoxide production, biogenesis, dynamics, and/or autophagy might prove effective given the importance of these mechanisms for diabetes-associated inflammation, endothelial dysfunction, and insulin resistance. Resveratrol or more potent activators of SIRT1 have potential in this regard.
Finally, the reviewed work supports the concept that monitoring endothelial function might prove useful for management decisions in the care of patients with diabetes. Testing endothelial function might help with selecting types and intensity of therapy. Similarly, serial examination of endothelial function might be useful for monitoring the effectiveness of risk reduction interventions. Well-established invasive and non-invasive methods for measuring endothelial function have proven useful in the research setting. However, translation of this work to the clinic will require the development of standardized methods for measuring endothelial function in individual patients that can be applied to the day-to-day care of patients with diabetes mellitus or other risk factors for cardiovascular disease.
Drs. Tabit, Hamburg, and Vita received support from the NIH-sponsored Boston University Medical Center Leadership Program in Vascular Medicine (K12 HL083781). Dr. Vita’s work is supported by grants from the NIH (HL083801, HL081587, HL083269, and HL75795).
Corey E. Tabit, Evans Department of Medicine and Whitaker Cardiovascular Institute, Boston University School of Medicine, Boston, MA, USA.
William B. Chung, Evans Department of Medicine and Whitaker Cardiovascular Institute, Boston University School of Medicine, Boston, MA, USA.
Naomi M. Hamburg, Evans Department of Medicine and Whitaker Cardiovascular Institute, Boston University School of Medicine, Boston, MA, USA.
Joseph A. Vita, Evans Department of Medicine and Whitaker Cardiovascular Institute, Boston University School of Medicine, Boston, MA, USA. Section of Cardiology, Boston Medical Center, 88 East Newton Street, Boston, MA 02118, USA, Email: ude.ub@ativj.