| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Diabetes Unit (R.M., M.J.Q.), National Center for Complementary and Alternative Medicine, National Institutes of Health, Bethesda, Maryland 20892; Department of Pharmacology and Human Physiology (M.M.), Section of Pharmacology, University of Bari Medical School, 70124 Bari, Italy; and Division of Cardiology (K.K.K.), Gil Heart Center, Gachon Medical School, Incheon 405760, Korea
Correspondence: Address all correspondence and requests for reprints to: Michael J. Quon, M.D., Ph.D., Chief, Diabetes Unit, National Center for Complementary and Alternative Medicine, National Institutes of Health, 10 Center Drive, Building 10, Room 6C-205, Bethesda, Maryland 20892-1632. E-mail: quonm{at}nih.gov
| Abstract |
|---|
|
|
|---|
| I. Introduction |
|---|
|
|
|---|
| II. Insulin-Signaling Pathways Regulating Cardiovascular Physiology |
|---|
|
|
|---|
, p110ß, and p110
catalytic subunit isoforms that bind to the major regulator p85 isoforms, p85
, p55
, and p50
(11). When SH2 domains of p85 subunit bind to tyrosine-phosphorylated motifs on IRS-1, this allosterically activates the preassociated p110 catalytic subunit to generate the lipid product phosphatidylinositol 3,4,5-trisphosphate [PI(3,4,5)P3] from the substrate phosphatidylinositol 4,5-bisphosphate (12). PI(3,4,5)P3 binds to the pleckstrin-homology domain in 3-phosphoinositide-dependent protein kinase-1 (PDK-1), resulting in its phosphorylation and activation to subsequently phosphorylate and activate other downstream serine-threonine kinases including Akt and atypical protein kinase C (PKC) isoforms (13, 14). A phosphorylation cascade of serine-threonine kinases downstream from PDK-1 in this PI3K-dependent branch of the insulin-signaling pathway culminates in many of the metabolic actions of insulin (Fig. 1
|
1. IR tyrosine kinase.
In human endothelial cells, IRs are expressed on the cell surface at a level approximately 10-fold less than the related IGF-I receptor (IGF-IR) (
40,000 IR and
400,000 IGF-IR per cell) (22). IGF-IRs and hybrid receptors (IR/IGF-IR) composed of heterodimers containing an
ß-chain of the IR associated with an
ß-chain of the IGF-IR have low affinity for insulin (30). Physiological concentrations of insulin (100–500 pM) selectively activate IR and downstream signaling pathways leading to activation of eNOS, whereas supraphysiological concentrations of insulin activate IGF-IR and hybrid receptors (22, 31). The first study directly demonstrating that insulin stimulates production of NO from endothelial cells used an NO-selective electrode to show that levels of NO produced in human umbilical vein endothelial cells (HUVEC) in response to a maximally stimulating concentration of insulin are approximately twice those that can be elicited by IGF-I stimulation (22). Overexpression of wild-type IRs in HUVEC causes a 3-fold increase in insulin-stimulated production of NO that is not seen in HUVEC transfected with mutant kinase-deficient IRs (28). These data suggest an essential role for IR tyrosine kinase activity in activation of eNOS in response to insulin.
2. IRSs.
IRS-1 is a major substrate for the IR tyrosine kinase. Overexpression of wild-type IRS-1 in bovine aortic endothelial cells (BAEC) results in a 3-fold increase in NO production in response to a maximally stimulating concentration of insulin (determined using the NO-specific fluorescent dye DAF-2) (29). These results suggest that IRS-1 is capable of mediating insulin-stimulated activation of eNOS. The relative contribution of IRS-2 in insulin-mediated NO production in endothelial cells is not known. However, overexpression of IRS-2 in rat fibroblasts cotransfected with eNOS significantly increases insulin-stimulated NO production (29). Expression of a mutant IRS-1 (IRS1-F6) that is unable to bind to the p85 subunit of PI3K inhibits both insulin-stimulated PI3K activity and NO production in endothelial cells (29). Introduction of an IRS-1 antisense ribozyme into endothelial cells substantially reduces insulin-stimulated production of NO (29). Taken together, these results suggest that IRS-1 is a necessary component (and the predominant IRS family member) of the insulin-signaling pathway leading to activation of PI3K that then enhances activation of eNOS.
3. PI3K.
Downstream from IRS-1, the essential role of PI3K in mediating insulin-stimulated production of NO is demonstrated in studies where preincubation of HUVEC with wortmannin (PI3K inhibitor) blocks NO production in response to insulin (22, 27). More importantly, overexpression of a dominant inhibitory mutant of the p85 regulatory subunit of PI3K significantly and substantially inhibits insulin-mediated production of NO in transfected HUVEC (28). In addition, PI3K mediates effects of insulin to stimulate increased expression of eNOS in endothelial cells (32, 33). Interestingly, inhibition of MAPK-dependent insulin-signaling pathways may enhance the PI3K-dependent vascular actions of insulin on eNOS (33, 34).
4. PDK-1.
PDK-1 is immediately downstream from PI3K. Overexpression of wild-type PDK-1 in BAEC results in a 2-fold increase in insulin responsiveness with respect to production of NO. Expression of a kinase-deficient PDK-1 mutant significantly blocks insulin-mediated production of NO (29). These data suggest that PDK-1 is an essential component of the insulin-signaling pathway leading to production of NO in vascular endothelial cells.
5. Akt.
Akt is activated in response to phosphorylation by PDK-1. Akt directly phosphorylates human eNOS at Ser1177 (equivalent to Ser1179 in bovine eNOS), resulting in enhanced eNOS activity (35). With respect to insulin signaling, the Akt phosphorylation site on eNOS is absolutely essential for activation of eNOS because cells expressing a mutant eNOS with a disrupted Akt phosphorylation site (alanine substituted for serine at position 1179) are unable to produce NO in response to insulin (25). Overexpression of dominant inhibitory mutant Akt proteins in HUVEC nearly completely inhibits production of NO in response to insulin (28). Akt1 is the predominant isoform in the vasculature, and endothelial cells from Akt1 knockout mice have significantly low levels of active eNOS (36). Therefore, it is most likely that Akt1 isoform mediates insulin-induced activation of eNOS. Moreover, pretreatment of cells with the calcium chelator BAPTA does not inhibit the ability of insulin to stimulate phosphorylation of eNOS at Ser1179 or enhance eNOS activity (25). In addition, insulin treatment does not alter intracellular calcium levels in endothelial cells (27). This suggests that insulin-stimulated production of NO is calcium-independent and mediated by activation of Akt1.
6. Role of HSP90.
Association of heat shock protein (HSP) 90 with eNOS is critically important for eNOS-mediated NO production (24). Although insulin-induced eNOS activation is calcium-independent, insulin stimulates calmodulin binding to eNOS (26). This requires HSP90 binding to eNOS, which facilitates insulin-stimulated activation of eNOS mediated by phosphorylation of eNOS at Ser1177 by Akt.
Akt is a necessary signaling molecule for insulin-stimulated activation of eNOS. However, activation of Akt per se is not sufficient for activation of eNOS. For example, treatment of endothelial cells with either insulin or platelet-derived growth factor (PDGF) results in comparable phosphorylation and activation of endogenous Akt. Nevertheless, only insulin (but not PDGF) treatment results in phosphorylation and activation of eNOS at the Akt phosphorylation site Ser1179 with consequent production of NO (22, 25, 26). One potential mechanism underlying this specificity may be that insulin (but not PDGF) elicits the formation of a ternary eNOS-HSP90-Akt complex (26).
7. Role of protein phosphatases.
eNOS activity is enhanced by phosphorylation at Ser1177 and decreased by phosphorylation at Thr495 (37). The temporal dynamics of phosphorylation at these regulatory sites involves both kinases and Ser/Thr phosphatases. For example, PP-2A may specifically dephosphorylate eNOS at Ser1177, whereas PP-1 has been implicated in dephosphorylation of eNOS at Thr495 (24, 38). In endothelial cells, insulin stimulates an acute decrease in phosphorylation of eNOS at Thr495 while simultaneously increasing phosphorylation of eNOS at Ser1177 (39). It is possible that insulin is activating a phosphatase targeting Thr495 while inhibiting a phosphatase targeting Ser1177.
8. Other vasodilators.
Prostacyclin (PGI2), a metabolite of arachidonic acid produced by cyclooxygenase-1 in endothelial cells, is another endothelial-derived vasodilator (40). Insulin acutely stimulates production of PGI2 from vascular endothelium (41, 42). NO can directly suppress activity of cyclooxygenase-1 and decrease both basal and stimulated release of PGI2 (40, 43). However, inhibition of insulin-stimulated NO production using N (G)-nitro-L-arginine methyl ester does not prevent a PGI2 production in endothelial cells (41). This suggests that insulin has direct actions to stimulate PGI2 production in an NO-independent fashion. Insulin-signaling pathways regulating PGI2 production are yet to be elucidated.
C. Insulin signaling in vascular endothelium regulating production of ET-1 and adhesion molecules
Endothelin-1 (ET-1) is a vasoconstrictor secreted by endothelial cells that opposes vasodilator actions of NO (44). Recent studies in BAEC and in mesenteric vascular beds have demonstrated that insulin and other hormones acutely stimulate the secretion of ET-1 using MAPK-dependent (but not PI3K-dependent) signaling pathways (45, 46, 47).
Endothelial expression of cellular adhesion molecules including intercellular adhesion molecule-1, vascular cell adhesion molecule (VCAM-1), and E-selectin is critical in modulating cell-cell interactions between circulating inflammatory cells and vascular endothelium. Insulin stimulates increased expression of VCAM-1 and E-selectin on endothelium using MAPK-dependent, but not PI3K-dependent, signaling pathways (33). Blockade of PI3K-dependent pathways enhances the effects of insulin or vascular endothelial growth factor (VEGF) to increase expression of these adhesion molecules (33).
D. Insulin signaling in vascular smooth muscle
1. Production of NO in vascular smooth muscle cells (VSMC).
In the vasculature, bioavailable NO originates mostly from the endothelium. Endothelial-derived NO diffuses into VSMC where it activates guanylate cyclase to increase cGMP levels that evoke vasorelaxation. However, expression of eNOS, inducible NOS (iNOS), and neuronal NOS mRNA and protein has been detected in VSMCs in certain contexts (48, 49, 50, 51, 52, 53, 54, 55). This raises the possibility that NO production in VSMCs may act in an autocrine fashion to regulate vasodilator functions. In VSMC, insulin increases NOS (eNOS and iNOS) activity and NO-dependent GMP production (48, 51, 56, 57). VSMCs express both IR and IGF-IR (58). Physiological concentrations of insulin stimulate IR autophosphorylation in VSMC and result in a rapid increase in cGMP levels by activating eNOS in human VSMCs (51, 56, 59, 60). Likewise, stimulation of IR/IGF-IR evokes a rapid release of NO in VSMC (as assessed using an NO-selective electrode) (61). Genistein (tyrosine kinase inhibitor) and wortmannin (PI3K inhibitor) both block effects of insulin to stimulate activation of NO in VSMC, suggesting that the IR tyrosine kinase and subsequent activation of PI3K are both necessary for regulation of eNOS or iNOS by insulin in VSMC (48, 56).
2. VSMC contractility.
Insulin attenuates VSMC contractility by regulating agonist-induced increases in cytosolic calcium through voltage-sensitive calcium channels and altering the activity of myosin light chain phosphatases (62, 63, 64). This may be mediated by signaling molecules including Rho kinase and PKC (65, 66, 67). RhoA, a small GTP binding protein, plays a key role in agonist-induced VSMC contraction (68). Rho activation and membrane localization are regulated by geranylgeranylation and phosphorylation of Rho (68). Active RhoA recruits and stimulates ROK-
, which then phosphorylates and inhibits myosin light chain (MLC) phosphatase leading to an increase in levels of phosphorylated MLC and heightened vascular tone (69, 70). In VSMCs, insulin acutely inhibits geranyl-geranyl transferase and decreases membrane levels of RhoA (66). In addition, insulin stimulates phosphorylation of RhoA at Ser188, which prevents it from binding to and activating ROK-
(66). These inhibitory actions of insulin are NO/cGMP-dependent, and may be mediated by cGMP-dependent protein kinase 1 (cGKI)
(66, 71). Moreover, insulin-activated cGKI
interacts with and activates MLC phosphatase (72). In VSMC, wortmannin (PI3K inhibitor) (48) and small interfering RNA against Akt (73) abrogates the effects of insulin on the RhoA/ROK/MLC phosphatase pathway, whereas expression of constitutively active Akt up-regulates cGKI
, ROK-
, and MLC phosphatase activities (73). Thus, the PI3K/Akt insulin-signaling pathway in VSMC is likely to mediate decreased contractility.
3. Calcium flux in VSMC.
Treatment of VSMC with insulin impairs agonist-evoked increases in intracellular calcium and accelerates the rate of calcium decline by inhibiting calcium influx and stimulating calcium efflux (61, 62, 74, 75, 76, 77, 78, 79). Insulin treatment results in hyperpolarization of membrane potential (80), stimulation of sodium pumps (81), and activation of Ca2+-dependent K+ channels (82), which all tend to reduce Ca2+ influx via voltage-operated channels (74). Insulin also activates Ca2+-pumps at the plasma membrane and sarcoplasmic reticulum to enhance cytoplasmic Ca2+ efflux (79). Treatment of VSMC with genistein and tyrophostin A-23 (tyrosine kinase inhibitors) attenuates these effects, suggesting that these actions of insulin are mediated by the IR (83). Inhibition of NOS also inhibits these effects of insulin on intracellular calcium (60). NO/cGMP is known to increase activity of sodium pumps (60) and activate Ca2+-dependent K+ channels (84). Moreover, IR/IGF-IR regulates sodium pumps through signaling by PI3K and atypical PKC-
(85). Thus, insulin-stimulated pathways involving PI3K/Akt/NO may help to regulate decreases in intracellular calcium in VSMC that result in decreased vasoconstrictor tone.
E. Insulin signaling in heart
Insulin regulates metabolism in the heart by modulating glucose transport, glycolysis, glycogen synthesis, lipid metabolism, protein synthesis, growth, contractility, and apoptosis in cardiomyocytes (86, 87, 88). In addition, vasodilator actions of insulin in coronary vasculature augment myocardial perfusion (89). IRs are expressed at levels of about 10,000 to 100,000 receptors per cardiomyocyte. Oxidation of fatty acids supplies approximately 70% of the hearts energy needs, however glucose and lactate may account for up to 30% of total ATP production. Insulin-stimulated glucose uptake in cardiomyocytes is mediated primarily by the insulin-responsive GLUT4. However, in addition to the basal cardiac glucose uptake mediated by GLUT1, contraction-mediated GLUT4 translocation to the sarcolemma may contribute significantly to myocardial glucose uptake (90). As in other insulin-sensitive tissues, insulin signaling via PI3K/Akt pathways plays a key role in cardiac glucose uptake. Insulin-stimulated activation of Akt also promotes cardiac glycogen accumulation by simultaneously inhibiting activity of both glycogen synthase kinase 3 and AMP-activated protein kinase (86, 91). Moreover, in the heart, insulin-stimulated Akt phosphorylates the transcription factor FOXO-1, which is known to affect glucose and lipid metabolism (92).
1. Cardiac contractility.
Insulin enhances cardiac contractility in vivo in humans as well as in isolated cardiac muscle (86, 87, 93, 94, 95). Myocardial excitation is associated with trans-membrane movement of extracellular Ca2+ into cardiac myocytes through activated Ca2+ channels and reverse Na2+/Ca2+ exchange. This influx of Ca2+ stimulates additional release of Ca2+ from the sarcoplasmic reticulum via ryanodine receptors, which results in myofilament activation and contraction. Studies in isolated human cardiac myocytes suggest that insulin enhances Ca2+ influx through activation of L-type Ca2+ channels and reverse-mode Na2+/Ca2+ exchange (94, 95). The PI3K inhibitors wortmannin or LY294002 inhibit the inotropic actions of insulin (94, 95, 96). The role of Akt in inotropic actions of insulin has not been directly assessed. Overexpressing Akt in cardiac myocytes is associated with increased cytoplasmic Ca2+ due to enhanced influx through L-type Ca2+ channels and release from sarcoplasmic reticulum (20, 97). Insulin also enhances myofilament Ca2+ sensitivity (95). Moreover, insulin increases cardiac NO production through the PI3K/Akt/eNOS pathway (98), and this may contribute to inotropic effects of insulin (99). Chronic overexpression of myocardial Akt (
15-fold) leads to cardiac dysfunction and heart failure (100), whereas a smaller (
2-fold) increase in Akt activity/expression associated with exercise (for 4 wk) is not associated with impaired contractility (101). These studies suggest that the effects of Akt activation on cardiac contractility may be dependent on the magnitude and duration of Akt activation (102).
2. Cardiac growth.
The PI3K/PDK-1/Akt branch of insulin-signaling pathways also plays an important role in developmental and physiological growth of the heart (19, 87, 88, 103). Downstream from Akt, activation of mammalian target of rapamycin promotes cardiac growth, whereas suppression of GSK3ß and FOXO helps to regulate cardiomyocyte size (104, 105). Constitutive overexpression of Akt leads to cardiac hypertrophy and dysfunction (20, 105). Thus, beneficial effects of Akt on cardiac growth may depend on temporal patterns of Akt activation as well as subcellular localization of Akt (20, 97, 104, 105). Pathological cardiomyocyte hypertrophy may be regulated by a distinct subset of the insulin-signaling pathways involving MAPK, p38 MAPK, and small G proteins Rho and Ras in addition to other signaling pathways. including PI3K/Akt pathway, calcineurin-nuclear factor of activated T cell pathway, kinases regulating histone deacetylases, cyclin-dependent kinase-7 and -9, PKC, and calmodulin kinase (104, 106).
| III. Cardiovascular Actions of Insulin |
|---|
|
|
|---|
|
1. Skeletal muscle capillary recruitment.
The microvascular unit, a group of capillaries fed by a single terminal arteriole, is the smallest functional unit for control of blood flow in skeletal muscle (117). Dilation of terminal arterioles can produce "maximal" opening (i.e., recruitment) of downstream dependent capillaries without initially changing total blood flow (118). Animal studies using isolated hindlimb preparations and human studies using limb or tissue balance techniques support the concept that insulin significantly modulates microvascular perfusion through capillary recruitment (93, 111, 115, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133). Insulin-stimulated capillary recruitment was first studied by Rattigan et al. (122) in rat hindlimb by measuring endothelial metabolism of exogenously infused 1-methylxanthine. Recently, a highly sensitive and specific, noninvasive technique using ultrasound imaging of skeletal muscle during microbubble contrast infusion has allowed for real-time assessment of capillary recruitment in response to insulin (111, 115, 127, 128, 130). In the hindlimb of anesthetized rats, insulin infusion (steady-state plasma insulin levels of
600 pM) increases microvascular volume by 1.5-fold after 5–10 min and is maximal (2.5-fold) after 20 min of insulin infusion (130). Insulin-stimulated capillary recruitment occurs well before changes in total limb blood flow that peak after 2 h of insulin infusion (130). Upon cessation of insulin infusion, increases in microvascular volume persist for 15–30 min after insulin concentrations return to basal levels (131). Pretreatment with N (G)-nitro-L-arginine methyl ester (NOS inhibitor) attenuates insulin-enhanced capillary volume by 50 to 70%, suggesting that these effects are partially NO-dependent (129, 130). In addition to effects on endothelium-derived NO, direct actions of insulin on VSMC, release of other vasoactive factors, changes in muscle metabolism, and alterations in sympathetic activity may contribute to insulin-stimulated capillary recruitment. Consistent with studies in rats, in deep flexor muscles of the human forearm, local intraarterial infusion of insulin (arterial plasma levels of
300 pM) results in a 25% increase in muscle capillary blood volume (127). Similarly, 1 h after a mixed meal, microvascular volume in human forearm increases by approximately 45% (111). Thus, physiological concentrations of insulin in both animals and humans rapidly enhance skeletal muscle capillary recruitment.
2. Limb blood flow.
Intravenous infusion of insulin increases total limb blood flow in a majority (93, 108, 121, 127, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150) but not all (151, 152, 153, 154, 155, 156) studies in humans. Increases in bulk muscle blood flow have been demonstrated at both physiological (108, 132, 134, 135, 137, 140, 143) and supraphysiological (123, 134, 140, 142, 150, 157) concentrations of insulin. Some controversy exists over whether physiological concentrations of insulin cause significant increases in total limb flow (147, 158). This may be the result of differences in subject selection as well as differences in physical fitness, muscularity, endothelial function, and capillary density of study subjects. Technical limitations or differences in sensitivity of various experimental approaches for estimating limb blood flow (e.g., plethysmography, thermodilution, positron emission tomography, dye dilution, Doppler ultrasound, and ultrasound measurements of brachial or femoral artery diameter) may also contribute to conflicting reports (140, 159, 160). Nevertheless, the preponderance of experimental evidence in humans suggests that physiological concentrations of insulin increase total limb blood flow, albeit with a slower time course than capillary recruitment. It remains unclear whether changes in capillary recruitment and total blood flow are independent or functionally coupled.
Insulin-induced vasodilation does not depend on concomitant changes in carbohydrate metabolism (108). Scherrer et al. (139) and Steinberg et al. (107) were among the first to provide compelling evidence that NO mediates the vasodilator actions of insulin. Coinfusion of L-NMMA (NOS inhibitor) during steady-state hyperinsulinemia under euglycemic conditions abrogates insulin-induced increases in blood flow in the leg (107) and forearm (139). Insulin also increases NOx (nitrate/nitrite, stable oxidative end-products of NO) release from the leg (161). Furthermore, a significant 2-fold increase in NOS activity (without significant changes in NOS protein content) is observed in human skeletal muscle in response to insulin stimulation (determined by biopsy of vastus lateralis) (162).
B. Vasoconstrictor actions of insulin
In addition to vasodilator actions of insulin discussed above, opposing hemodynamic actions of insulin include activation of the sympathetic nervous system and stimulation of secretion of the vasoconstrictor ET-1 from vascular endothelium.
1. Role of sympathetic nervous system.
In humans, the role of the sympathetic nervous system to mediate vasoconstriction in the integrated hemodynamic response to insulin has been reviewed in detail previously (163). In healthy lean individuals, physiological concentrations of insulin increase venous catecholamine levels and sympathetic nerve activity (135, 156, 164). In addition, insulin infusion augments centrally mediated sympathetic outflow to skeletal muscle in humans and rats (163, 165). The classical observation that insulin decreases arterial pressure in patients with autonomic failure highlights the role of insulin-induced sympathetic vasoconstriction in normal individuals (166). Indeed, in people who have undergone regional sympathectomy, NO-dependent vasodilation in response to insulin in the denervated limb occurs more quickly than in the innervated limb (167). This suggests that heightened sympathetic vasoconstrictor tone stimulated by insulin opposes the vasodilator actions of insulin mediated by NO. In animal models, the cholinergic system may also be involved in mediating vasoactive actions of insulin (168). However, in humans, neither cholinergic nor ß-adrenergic pathways seem to be involved in modulating vasodilator actions of insulin (138). Under conditions of the euglycemic hyperinsulinemic glucose clamp, the rise and fall in peak muscle sympathetic activity temporally lags behind the rise in plasma insulin concentrations. This may reflect the time required for insulin to redistribute and cross the blood-brain barrier as well as the activation/inactivation kinetics of cellular events that trigger an increase in nerve activity (135). Of note, in rats, differential hypothalamic activation of PI3K and MAPK has been demonstrated in the regional sympathetic responses to insulin (165). Insulin-induced release of NO may oppose sympathetically mediated vasoconstriction at the level of the myocyte as well as in the central nervous system. In proximal and distal arterioles regulating total blood flow and capillary recruitment, respectively, smaller vessels are associated with increased sensitivity to insulin-mediated vasodilation. However, in the face of elevated sympathetic nerve activity, distal arterioles vasodilate in response to insulin, whereas proximal arterioles undergo sustained vasoconstriction (169). Thus, various parts of the vascular tree have a differential response to insulin (149, 170) and sympathetic nerve activity (169).
2. Role of ET-1.
In addition to production of NO, insulin regulates synthesis and secretion of ET-1 from vascular endothelium. Consequently, in mice with targeted deletion of the IR in vascular endothelium [vascular endothelium IR knockout (VENIRKO) mice], expression of both eNOS and ET-1 is significantly diminished (171). In humans, the effects of insulin to change circulating levels of ET-1 are unclear (172, 173, 174). However, because ET-1 is a paracrine factor, plasma concentrations are less relevant than local concentrations and do not predict ET-1 activity in the vascular milieu (44). Supporting this concept, vasodilator actions of insulin are potentiated by ET-1 receptor blockade in animals (175) and humans (176). Consistent with the MAPK dependence of insulin-stimulated secretion of ET-1 in vascular endothelium, inhibition of MAPK blocks vasoconstrictor effects of insulin in rat skeletal muscle arterioles (177). A shift in balance between vasoconstrictor and vasodilator actions of insulin mediated by pathway-specific impairment in PI3K signaling may be an important factor in the vascular pathophysiology of insulin resistance and endothelial dysfunction.
C. Effects of insulin on blood pressure
As discussed above, insulin has opposing vasodilator and vasoconstrictor actions such that the net hemodynamic effect of insulin on blood pressure is minimal in healthy humans. Indeed, short-term insulin infusion under isoglycemic conditions modestly decreases (157) or has no effect on arterial blood pressure (108, 119, 135, 178, 179, 180). In these studies, iv insulin infusion significantly increases heart rate and cardiac output and decreases total peripheral resistance. However, the fall in systemic vascular resistance is modest (
15%) when compared with the reduction in leg vascular resistance (
40%), suggesting a differential and specific effect of insulin to dilate skeletal muscle vasculature (119). By contrast, with iv insulin infusion, intraarterial infusion of insulin does not change (150, 155, 176, 181, 182, 183) or minimally increases limb blood flow (123, 146, 184). This is most likely because insulin simultaneously stimulates both NO production and ET-1 secretion. In the presence of ET-1 receptor blockade, intraarterial insulin infusion causes measurable vasodilation (176).
Acute infusion of insulin also promotes sodium retention by enhancing distal tubular sodium reabsorption in normal and insulin-resistant individuals (185, 186). However, prolonged hyperinsulinemia during euglycemic glucose clamp conditions results in compensatory natriuresis due to diminished proximal tubular sodium reabsorption and increases in renal plasma flow and glomerular filtration rate in healthy individuals (187, 188). Due to the small effects on sodium retention and compensatory natriuresis, it is unlikely that renal actions of insulin play an important role in modulating blood pressure acutely.
D. Regulation of cardiac function by insulin
Due to high basal oxygen extraction, cardiac oxygen demand is a dominant determinant of myocardial blood flow (MBF). The coronary microcirculation is a major contributor to coronary vascular resistance and MBF. In the setting of increased myocardial oxygen consumption, myocardial hyperemia is initially associated with increased capillary blood flow velocity and followed by capillary recruitment (189). As previously discussed, insulin action in the endothelium and the vascular wall modulates hemodynamics through changes in both flow and capillary recruitment. Consistent with this, accumulating evidence suggests that insulin enhances MBF in the heart (89, 190, 191, 192, 193, 194, 195). In addition, insulin increases cardiac contractility resulting in increased myocardial work and oxygen consumption (93). Because MBF and myocardial oxygen consumption are tightly coupled and regulated, it is difficult to evaluate direct actions of insulin on the coronary vasculature in vivo. One approach used to address this issue is to evaluate effects of insulin under conditions in which MBF and oxygen consumption are uncoupled by simultaneous infusion of adenosine. Under these circumstances, physiological concentrations of insulin enhance adenosine-stimulated MBF and coronary flow reserve in humans (89, 190, 191). Similarly, physiological hyperinsulinemia in healthy subjects increases MBF, specifically in areas of the myocardium associated with high rates of glucose uptake (192). This suggests coupling between metabolic and vascular actions of insulin in the heart. As observed in human skeletal muscle, ingestion of a mixed meal enhances capillary recruitment in the heart (196). However, meals evoke a complex neuroendocrine response in addition to changes in plasma insulin that may independently influence cardiac hemodynamics. In fact, ingestion of a mixed meal is associated with an increase in left ventricular ejection fraction and contractility that may augment MBF independent from the effects of insulin (197). Therefore, meal-induced increases in MBF cannot be solely attributed to effects of insulin. Nevertheless, concomitant infusion of insulin along with a meal further enhances myocardial capillary recruitment, suggesting that insulin does have some direct effects to increase flow in the myocardial capillary bed (194).
E. Role of insulin to couple hemodynamic and metabolic physiology
Studies in animals and humans suggest that insulin-stimulated increases in skeletal muscle capillary recruitment and blood flow play an important physiological role in augmenting the delivery of insulin and glucose to skeletal muscle. Glucose delivery to skeletal muscle is dependent on muscle blood flow and vascular capillary surface area and permeability. When capillary surface area and permeability are small, increasing blood flow per se has a minimal effect on net glucose uptake (198, 199). However, after a mixed meal or an oral glucose load, recruitment of capillaries expands the capillary surface area and increases muscle blood flow, which together substantially increase glucose and insulin delivery (111, 148). Changes in insulin-mediated capillary recruitment are positively correlated with changes in insulin-stimulated glucose disposal (130). The time course for insulin-stimulated capillary recruitment approximates the time course for insulin-mediated glucose uptake in skeletal muscle (130). Moreover, inhibitors of NOS that block insulin-mediated capillary recruitment cause a concomitant 40% reduction in glucose disposal (129, 130). In human studies, under conditions of high glucose extraction, insulin stimulates parallel increases in leg glucose disposal and blood flow in a dose-dependent manner (109, 136, 200, 201). Although the time course of increases in leg blood flow during physiological hyperinsulinemia is slower than that for glucose uptake, it generally follows leg glucose uptake. Infusion of the competitive NOS inhibitor, N (G)-nitro-L-arginine methyl ester, completely blocks the effect of insulin to increase flow and partially blocks insulin-stimulated leg glucose uptake (134, 136). Thus, in addition to direct PI3K-dependent metabolic actions of insulin to promote glucose uptake in skeletal muscle through stimulating translocation of insulin-responsive GLUTs, the PI3K-dependent vascular actions of insulin to increase blood flow and capillary recruitment substantially contribute to promoting glucose disposal under healthy conditions and help to couple metabolic and hemodynamic homeostasis (Figs. 1
and 2
).
| IV. Reciprocal Relationships between Insulin Resistance and Endothelial Dysfunction |
|---|
|
|
|---|
|
|
1. Impaired insulin signaling.
Treatment of vascular endothelial cells with FFA reduces basal and insulin-stimulated eNOS activity and NO production (211). Moreover, FFA treatment impairs insulin-stimulated activation of PI3K, PDK1, Akt, and eNOS (211). These effects of FFA are specific to insulin because FFA treatment does not alter the ability of VEGF to stimulate the PI3K/Akt/eNOS pathway. This impairment in insulin signaling in the endothelium caused by FFA treatment is similar to that observed in skeletal muscle. Exposure to FFA increases cellular levels of diacylglycerols, ceramide, and long-chain fatty acyl coenzyme A (CoA). These lipid metabolites activate serine kinases such as PKC and inhibitory kB kinase (IKKß) that regulate activation of nuclear factor-
B (NF-
B), a transcription factor associated with inflammation (216). Interestingly, activation of PKCß1 and -ß2 isoforms results in increased serine phosphorylation of IRS-1 that leads to reduced insulin-stimulated Akt and eNOS activities (217). Likewise, palmitate activates IKKß and Jun N-terminal kinase (JNK), which increases serine phosphorylation of IRS-1 and decreases insulin-stimulated production of NO (218). Inhibitory effects of FFA treatment on insulin signaling and NO production in endothelial cells can be blocked by overexpression of a dominant inhibitory mutant of IKKß. Moreover, deleterious effects of FFA treatment on endothelial cells are recapitulated by overexpression of wild-type IKKß. Treatment of endothelial cells with FFAs up-regulates expression of the lipid phosphatase PTEN, a negative regulator of PI3K-dependent signaling (211). These same cellular signaling pathways are impacted by FFA in the heart (219). In cardiomyocytes, FFA treatment is associated with reduced PI3K/Akt activity, resulting in diminished insulin-stimulated glucose uptake (220), eNOS activation, and contractile function (92). In addition to effects on insulin signaling, long-chain acyl-CoA esters directly stimulate opening of ATP-sensitive potassium (KATP) channels leading to K+ efflux, shortened action potentials, reduced Ca2+ influx, and decreased contractile force (221). In coculture studies of adipocytes and cardiomyocytes, adipocyte-derived factors directly depress intracellular systolic Ca2+ peaks and cardiac contraction (222). Moreover, intracellular fatty acid accumulation is associated with local generation of TNF-
, which is known to inhibit cardiac contraction (219).
2. Oxidative stress.
FFAs increase reactive oxygen species (ROS) production in the vasculature. Two primary sources of ROS in the vasculature are nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (213) and the mitochondrial electron transport chain (212). Insulin treatment decreases FFA-induced ROS production in a PI3K-dependent manner. This suggests that specific impairment of PI3K signaling may accentuate FFA-evoked oxidative stress (212). Reduction in eNOS activity due to FFA is reversed by preventing fatty acid oxidation, uncoupling oxidative phosphorylation, or scavenging locally increased superoxide production. In addition, FFA-associated ROS production enhances PKC activity, activates the hexosamine biosynthetic pathway, and increases formation of advanced glycation end-products (AGEs). All of these mechanisms may independently contribute to inhibition of insulin-stimulated NO production in vascular endothelium (212). The increased mitochondrial superoxide generation was shown to inhibit PGI2 activity by nitration. Consistent with these in vitro studies, in vivo exposure to high concentrations of FFA reduced eNOS and PGI2 activity in the aorta. Administration of antilipolytic agents or ROS scavengers (superoxide dismutase mimetics) to rodents with PI3K pathway-specific insulin resistance normalized the impaired eNOS and PGI2 activity (212). In addition, PKC-dependent activation of NADPH oxidase has been demonstrated in endothelial and VSMCs exposed to FFA (213). Superoxide may react with NO, generating peroxynitrite and reducing NO bioavailability. Peroxynitrite in turn can oxidize tetrahydrobiopterin, "uncouple" eNOS, nitrate and inhibit PGI2, and alter NO/cGMP signaling in VSMCs by altering the expression of soluble guanylyl cyclase (sGC) and by inhibiting the activity of the sGC and cGKI (223). Furthermore, arterial content of eNOS dimers is significantly reduced in a rodent model of high-fat-diet-induced obesity and diabetes (224). This reduction in eNOS dimerization was associated with an increase in nitrotyrosine content, suggesting that oxidative stress may have contributed to eNOS disruption and endothelial dysfunction. Interestingly, insulin-stimulated eNOS phosphorylation was unaltered and does not appear to contribute to endothelial dysfunction in this model of diabetes. However, in a similar rodent model of high-fat-induced diabetes, insulin-mediated vasorelaxation is significantly reduced (225). Indeed, a high-fat meal is associated with elevated circulating levels of nitrotyrosine in insulin-resistant individuals (226). ROS production due to "uncoupled" eNOS in circulating endothelial progenitor cells is significantly increased in patients with diabetes (227). In summary, impaired phosphorylation, activity, expression, and "uncoupling" of eNOS may all act in concert to mediate FFA-provoked endothelial dysfunction.
3. Altered sympathetic activity and the RAS.
Elevated FFAs secondary to lipid infusion in healthy individuals are known to increase sympathetic nerve activity, plasma catecholamine concentrations (228, 229), vascular
-adrenergic reactivity (215), and local RAS activity (214). These changes may counteract vasodilator actions of insulin and potentiate vasoconstriction, resulting in enhanced vascular tone in insulin-resistant states.
4. Effects of lipotoxicity on cardiovascular function.
FFAs concomitantly reduce insulins vasodilator and metabolic effects. Infusion of intralipids in rats raises circulating FFA levels causing significant impairment in skeletal muscle capillary recruitment with a concomitant 40% decrease in glucose disposal during a glucose clamp procedure with steady-state plasma insulin levels of approximately 600 pM (230). Likewise, the effects of insulin on capillary recruitment and glucose uptake are impaired when FFA levels are increased in healthy lean women (231). Moreover, when FFA levels are lowered in obese women, vasodilator actions of insulin are improved, suggesting that insulins microvascular and metabolic effects may be coupled during changes in exposure to FFAs. In fact, changes in capillary recruitment account for 30% of the association between changes in FFA levels and changes in insulin-mediated glucose uptake (231). Infusion of a lipid emulsion in conjunction with heparin to elevate circulating FFA concentrations simultaneously decreases glucose uptake and attenuates insulin-induced increases in leg blood flow and NO flux (161, 208), with significant correlations between FFA-induced changes in glucose uptake and FFA-induced decreases in leg blood flow. Thus, vascular and metabolic actions of insulin are tightly coupled such that impairment of the PI3K/Akt pathway by FFAs in the vascular wall contributes to insulin resistance in skeletal muscle. FFAs also induce cardiac insulin resistance. Six weeks of high-fat feeding in dogs induces systemic insulin resistance and decreased coronary hyperemia in response to exercise (232). This suggests direct lipotoxic effect(s) on cellular mechanisms that mediate metabolic coronary vasodilation and may explain diminished cardiac flow reserve in insulin-resistant individuals with dyslipidemia.
C. Glucotoxicity
Long-term glycemic control is an important predictor of both micro- and macrovascular disease (233, 234, 235, 236). Hyperglycemia impairs both metabolic and vascular actions of insulin by multiple biochemical and cellular mechanisms (237, 238, 239). These include elevated oxidative stress, increased flux through polyol and hexosamine biosynthetic pathways, formation of AGEs, and activation of diacylglycerol and PKC.
1. Oxidative stress.
Hyperglycemia increases the production of ROS. In endothelial cells exposed to high glucose concentrations, insulin-stimulated activation of Akt and eNOS is significantly reduced (240, 241). Overexpression of uncoupling protein-1 or manganese superoxide dismutase prevents these inhibitory effects of glucose and restores vasodilator actions of insulin (240). In addition to impairing insulin-signaling pathways, ROS decreases NO bioavailability, reduces cellular tetrahydrobiopterin levels, and promotes generation of superoxide by eNOS. ROS also activates PKC-
, PKC-ß, and PKC-
, leading to decreased expression of eNOS and increased expression of ET-1, VEGF, and TGF-ß in endothelial cells (237).
2. Hexosamine biosynthetic pathway.
Increased flux through the hexosamine biosynthetic pathway is another mechanism by which hyperglycemia may impair both metabolic and vascular actions of insulin (237, 242). Glutamine:fructose-6-phosphate amidotransferase (GFAT) is the rate-limiting enzyme for this pathway. Overexpression of GFAT in transgenic mice causes insulin resistance (243). The hexosamine biosynthetic pathway may function as a nutrient sensor that plays a role in insulin resistance and vascular complications by causing reversible O-GlcNAc modifications at regulatory serine/threonine phosphorylation sites on proteins involved with insulin signaling. For example, increased O-GlcN acylation of IRS-1 may lead to reduced insulin-stimulated translocation of GLUT4 and decreased glucose uptake (244, 245). In endothelial cells, hyperglycemia increases O-GlcN acylation of eNOS at the Akt phosphorylation site at Ser1179, resulting in impaired eNOS activity. These defects are reversed by decreasing GFAT expression (240).
3. AGEs.
AGEs are proteins or lipids that become nonenzymatically glycated and oxidized after contact with aldose sugars (246, 247). Enhanced AGE formation associated with hyperglycemia and oxidative stress inhibits both vascular and metabolic actions of insulin (248). Human glycated end-products inhibit insulin-stimulated tyrosine phosphorylation of IRS-1 and IRS-2 leading to impaired activation of PI3K and Akt (249). Moreover, AGE produces ROS and increases oxidative stress by activation of NADPH oxidase through specific receptors for AGE (250). In endothelial cells, AGEs decrease NO bioavailability and eNOS expression by accelerating eNOS mRNA degradation (251, 252, 253, 254). AGEs also enhance expression of ET-1 in endothelial cells through the activation of NF-
B (255). Thus, AGEs alter the balance of NO and ET-1 to favor vasoconstriction and endothelial dysfunction.
In cardiomyocytes, increased AGE, hexosamine and polyol flux, oxidative stress, and PKC activation have negative effects on function by prolonging action potentials, reducing relaxation kinetics, and altering myofilament Ca2+ sensitivity through changes in expression and function of various ion channels, receptors (ryanodine receptor, ß-adrenergic receptor), and ion pumps (SERCA, sodium-pump) (for review, see Refs. 219 and 220). These changes are predicted to result in ventricular stiffening and impaired ventricular filling (diastolic dysfunction), characteristics frequently observed in rodent and human models of diabetes.
4. Effects of glucotoxicity on cardiovascular function.
In animal studies, acute hyperglycemia impairs endothelial function in both macro- and microvascular beds (256, 257). Local hyperglycemia achieved by infusing concentrated glucose directly into the brachial artery of healthy humans diminishes agonist-induced vasodilation, an effect prevented by antioxidants (258, 259, 260). Similarly, moderate hyperglycemia after an oral glucose load is associated with reduced flow-mediated vasodilation in healthy individuals (261). Acute hyperglycemia consistently impairs endothelial function in individuals with insulin resistance or type 2 diabetes (262, 263). Glucosamine, a product of the hexosamine biosynthetic pathway, impairs insulin stimulated glucose uptake in skeletal muscle and production of NO in endothelium in vitro (242, 264). In vivo, acute iv glucosamine administration causes metabolic insulin resistance (265, 266) and impairs insulin-mediated increases in femoral arterial blood flow (266, 267) and capillary recruitment (266). Collectively, these data suggest that hyperglycemia impairs insulin action in skeletal and cardiac muscle as well as in vascular endothelium.
D. Proinflammatory signaling and adipocytokines
Insulin resistance and endothelial dysfunction are pathological states that are both characterized by increased circulating markers of inflammation (268). Visceral fat accumulation may play a key role in development of the systemic proinflammatory state associated with insulin resistance (268, 269, 270). Adipose tissue (and infiltrated resident macrophages) secretes a plethora of peptide hormones including leptin, adiponectin, TNF-
, IL-6, resistin, angiotensinogen, and plasminogen activator inhibitor-1 that play crucial roles in metabolic and vascular homeostasis.
1. Cytokines.
A number of potential biochemical mechanisms may explain the contribution of proinflammatory signaling to insulin resistance. The most extensively studied proinflammatory cytokine implicated in insulin resistance is TNF-
. FFAs are important determinants of adipose tissue TNF-
activity and expression (271). Increased ROS in response to FFA activates NF-
B, which further stimulates production of other proinflammatory cytokines including TNF-
and IL-6 (272, 273, 274, 275). TNF-
activates a variety of serine kinases including JNK, IKKß, and IL-1ß receptor-associated kinase (276, 277, 278, 279) that directly or indirectly increase serine phosphorylation of IRS-1/2, leading to decreased binding and activity of PI3K in response to insulin stimulation. For example, TNF-
-activates JNK, resulting in increased phosphorylation of IRS-1 at Ser307 (280). This reduces insulin-stimulated activation of PI3K/Akt/eNOS in endothelial cells (281). In addition to modulating eNOS activity, JNK and IKKß (through activation of activator protein-1 and NF-
B) also inhibit insulin-stimulated expression of eNOS (282). Thus, insulin resistance reduces bioavailability of NO under basal conditions, and this may be an additional pathogenic factor in chronic diseases including atherosclerosis, hypertension, and diabetes with inflammatory components. Furthermore, suppressors of cytokine-signaling proteins are induced by treatment of cells with TNF-
, IL-1ß, or IL-6. Increased expression of suppressors of cytokine-signaling proteins interferes with interaction of the IR and IRS-1 and enhances proteasomal degradation of IRS-1 in adipose tissue (283). TNF-
also stimulates expression of other inflammatory proteins including C-reactive protein (CRP) and IL-6. CRP is an important marker of vascular inflammation whose plasma levels are correlated with risk of cardiovascular disease. In addition to being a marker of inflammation, CRP may have biological actions to inhibit insulin-evoked NO production in endothelial cells through mechanisms involving phosphorylation of IRS-1 at Ser307 (mediated by syk, RhoA, and JNK) (284) and decreased expression of eNOS (285). In addition, CRP simultaneously increases ET-1 production and may directly promote cardiovascular disease by modulating expression of proinflammatory cytokines (286), up-regulating angiotensin receptor type 1 expression (287), and increasing expression of intercellular adhesion molecule, VCAM, E-selectin, and monocyte chemoattractant protein-1 in vascular endothelium (288, 289).
In animal models, administration of TNF-
induces insulin resistance (290), whereas neutralization of TNF-
improves insulin sensitivity (291). By contrast, neutralization of circulating TNF-
in patients with type 2 diabetes fails to alter insulin sensitivity significantly (292). Although this argues against a systemic role for TNF-
in mediating insulin resistance, the autocrine/paracrine nature of TNF-
action may have important pathophysiological significance. Indeed, systemic infusion of high doses of TNF-
results in loss of insulin-induced increases in glucose uptake, limb blood flow, and capillary recruitment in rat hind limb (293). This inhibitory action of TNF-
is specific to insulin-mediated, but not to exercise-mediated, hemodynamic and metabolic changes (293). TNF-
specifically down-regulates the insulin-dependent PI3K/Akt/eNOS vasodilator pathway while simultaneously augmenting ET-1-mediated vasoconstriction in skeletal muscle arterioles (280). In humans, high local concentrations of TNF-
achieved by intraarterial infusion simultaneously inhibit both insulin-stimulated glucose uptake (294) and endothelium-dependent vasodilation in the forearm (294, 295). Similarly, systemic infusion of TNF-
to achieve circulating concentrations slightly higher than those observed in chronic inflammation such as diabetes reduced glucose disposal in healthy individuals (296). Thus, proinflammatory cytokines may contribute to coupling of metabolic and vascular insulin resistance manifested by impaired insulin signaling and endothelial dysfunction.
2. Adipokines and related peptide hormones.
Adipocyte-derived hormones such as leptin and adiponectin have both metabolic and vascular actions. Leptin, a key regulator of appetite, body weight, and energy balance in the central nervous system acts directly on the vasculature. Similar to insulin, leptin induces endothelium-dependent vasodilation (297, 298) through a PI3K/Akt/eNOS pathway (299). Insulin enhances leptin-induced eNOS activation, NO production, and vasorelaxation suggesting cross-talk between the insulin and leptin signaling pathways (300). Like insulin, leptin-evoked vasodilation is opposed by sympathetically induced vasoconstriction (301). Leptin replacement partially ameliorates cardiac contractile dysfunction that is present in hypoleptinemic (ob/ob) mice (302). Similarly, administration of leptin improves cardiac dysfunction in transgenic mice with cardiac-restricted steatosis caused by overexpression of acyl-CoA synthase (303). This salutary effect of leptin may be mediated by activation of AMP-activated protein kinase and may be secondary to mobilization of myocardial lipid. In contrast to these potentially beneficial actions of leptin, in a large prospective study, the West of Scotland Coronary Prevention Study (WOSCOPS), leptin was an independent risk factor for coronary artery disease (304). This suggests that hyperleptinemia and/or leptin resistance may have deleterious vascular and metabolic effects. In support of this concept, leptin enhances cytokine (TNF-
and IL-6) (305) and ROS production (306), up-regulates expression of angiotensinogen and ET-1 (307, 308), and negatively regulates insulin signaling (309) and glucose uptake by increasing serine phosphorylation of IRS-1 (310). Interestingly, leptin also suppresses cardiac contractile function in ventricular myocytes by an ET-1-dependent pathway (302). Angiotensin II increases leptin secretion from cultured human fat cells (311). Leptin may potentiate pressor effects of hyperinsulinemia in insulin-resistant states. Therefore, interactions between angiotensin II and insulin wi