The notion that regular aerobic exercise reduces cardiovascular morbidity and mortality in the general population as well as in patients with coronary artery disease is strongly supported by evidence derived from epidemiologic studies (0005,0010,0015). Physically active people also experience fewer clinical manifestations of coronary artery disease than do less active men and women (0020). By contrast, sedentary life-style has been identified as a risk factor for development of coronary artery disease, and there is a strong correlation between physical inactivity and cardiovascular mortality ((0015),0025). Although vigorous physical exertion is a precipitating factor for myocardial infarction, this adverse outcome is usually incurred by persons who otherwise lead a sedentary existence ((0030),0035). Consequently, daily physical aerobic activity is considered an effective component of both primary and secondary prevention of cardiovascular events ((0040),0045). The purpose of this review is to provide an overview of the mechanisms by which exercise exerts its salutary effect on cardiovascular function and structure. We will focus on the role of exercise-induced increases in blood flow to modulate vascular tone and structure.

In addition to changes in vascular structure, changes in vascular tone may be induced by long-term regular exercise. Haskell and colleagues (0135) used quantitative coronary angiography to study vascular reactivity in ultra-distance runners and sedentary age- and gender-matched men and women. Although there were no differences between groups in basal diameter of the epicardial coronary arteries, when sublingual nitroglycerin was administered, the arteries of the marathoners showed a 200% greater increase in vasodilation than did those of the sedentary group. These data (0135) provide evidence for exercise-induced alterations of the structure or reactivity, or both, of the vessel wall. Exercise may also induce changes in lumen diameter in patients after coronary angioplasty. Indeed, patients randomized to a 12-week intervention program consisting of daily exercise after balloon angioplasty enjoyed a significantly lower rate of restenosis than patients in the control group (0170).

Collectively, these findings support the hypothesis that regular aerobic exercise induces salutary changes in vessel structure and reactivity. However, these studies leave unanswered many questions regarding the molecular mechanisms of exercise-induced vascular remodeling and reactivity. Recent investigations have provided some clues into the cellular and molecular mechanisms by which alterations in blood flow and shear stress may affect vascular structure and reactivity.

The two principal forces acting on the blood vessel are pulsatile stretch and shear stress. Pulsatile stretch is determined by fluctuation in arterial pressure and is a force exerted at a vector that is perpendicular to the longitudinal axis of the vessel. Shear stress is determined by blood flow and is a tractive force exerted at a vector that is parallel to the long axis of the vessel. The preponderance of scientific data suggests that exercise-induced increases in endothelial shear stress has beneficial effects on vascular structure and reactivity. Our focus is on the mechanism by which exercise-induced changes in flow may alter vascular reactivity and structure.

Physical exercise increases intracoronary blood flow, resulting in a vasodilation of the epicardial coronary arteries that is largely dependent on the integrity of the endothelium (0175,0180,0185,0190). The essential role of the endothelium in exercise-induced flow-mediated vasodilation has been confirmed in exercising dogs (0195). Whereas exercise-induced epicardial coronary vasodilation was observed during treadmill running, marked vasoconstriction was found during treadmill exercise up to 6 days after balloon denudation of the endothelium. This flow-mediated vasodilation was proportional to the shear stress induced by blood flow and was independent of changes in pressure within the lumen ((0200),0205).

Flow-mediated vasodilation is largely due to the release of endothelium-derived relaxing factor (EDRF) ((0190),(0210),0215), although in some circulations prostacyclin (0215,0220,0225) or an endothelium-derived hyperpolarizing factor may contribute (0230). Endothelial cells produce bursts of prostacyclin in response to sudden increases in shear stress. This action may indicate that the shear stress gradient, rather than the absolute level of shear stress, may be the more relevant physiologic stimulus (0235). Of the factors released by the endothelium in response to flow, EDRF is the most potent vasodilator. EDRF was first described in 1980 by Furchgott and Zawadzki (0240) and is now known to be nitric oxide (NO) ((0245),0250). NO is derived from the metabolism of L-arginine to L-citrulline by NO synthase (0255). Like other nitrovasodilators, NO exerts its effect on vascular smooth muscle by activating soluble guanylate cyclase to produce cyclic guanosine monophosphate. In addition, NO may stimulate calcium (Ca²⁺) activated potassium channels, thereby inducing hyperpolarization and relaxation of vascular smooth muscle (0260).

A dynamic balance between endothelium-derived vasodilating (e.g., adenosine triphosphate [ATP], substance P) and vasoconstricting (e.g., endothelin-1) factors contributes to the regulation of vascular tone by shear stress. Shear stress induces the release of ATP or substance P, or both, from endothelial cells of numerous vascular beds ((0265),0270). Both ATP and substance P stimulate specific receptors on the endothelium to release NO. In addition to inducing the release of ATP, increases in flow deliver more circulating ATP to the endothelial surface, overwhelming ectopyrases that degrade the nucleotide. The result is that more ATP reaches its purinergic receptors to stimulate endothelial release of NO ((0275),0280).

Physiologic levels of flow tend to inhibit the release of endothelin-1 (0285,0290,0295,0300,0305), a potent peptide vasoconstrictor. The release of endothelin-1 is known to be inhibited by NO (0310). However, the effect of shear stress on the release of endothelin-1 from endothelial cells is complex. Physiologic levels of shear stress induce a transient increase in endothelin-1 messenger ribonucleic acid (mRNA) expression and peptide release, which is followed by a significant suppression ((0295),(0305),0315). In human umbilical vein endothelial cells, low levels of shear stress stimulate the release of endothelin-1, whereas high levels inhibit its release (0285).

In addition to its vasoconstrictor effects, endothelin-1 is a potent mitogenic agent (0320). This effect of endothelin-1 is further potentiated by other growth stimulating substances such as angiotensin II, platelet-derived growth factor beta (PDGF-beta) and insulin ((0325),0330). Therefore endothelin-1 may critically contribute to the induction of coronary vasospasm as well as to the proliferative processes involved in vascular pathology.

Bicycle ergometry performed by healthy subjects did not lead to any significant changes in endothelin-1 levels, but plasma endothelin-1 concentrations increased in patients with coronary atherosclerosis (0335). When these patients received supplementation of NO with oral nitrates during the 72 h before exercise testing, the exercise-induced increase in endothelin-1 was attenuated. This observation indicates that the reduced activity of endothelium-derived NO in atherosclerotic blood vessels may play an important role in the exercise-induced increase of endothelin-1 secretion in humans. An imbalance of these endogenous vasoactive agents may explain in part the association between myocardial infarction and preceding vigorous physical exercise in subjects at risk who did not exercise on a regular basis ((0030),0035).

Recent investigations have shed light on the mechanisms by which the endothelium senses and transduces the stimulus of flow. An endothelial potassium channel is known to be activated by flow (0340). The mechanisms by which ion channels are activated by flow remain obscure, but they probably involve shear stress-mediated deformation of cytoskeletal elements (0345). The activity of an endothelial potassium channel also appears to be required for the flow-induced (but not receptor-mediated) release of NO from vascular segments or cultured endothelial cells ((0190),0350). This unique signal transduction pathway is coupled to a pertussis toxin-sensitive G protein (0355). Intriguingly, this same signal transduction pathway is involved in the flow-induced transcription of transforming growth factor-beta₁ (TGF-beta₁) (0360). Thus, the initial response to an increase in flow (vasodilation) and the later response (change in vascular structure) may share a common mechanotransducer.

These observations dovetail nicely with work from other groups indicating that flow increases Ca²⁺ influx ((0275),(0280),0365) in endothelial cells, an effect that is necessary for the synthesis and release of NO. By hyperpolarizing the endothelial cell, the flow-activated potassium channel maintains the electrochemical gradient for Ca²⁺ entry. The flow-induced changes in Ca²⁺ flux may also be modulated by vasoactive agents such as ATP (0280). Flow-induced changes in ion channel activity may also be regulated by the activity of kinases (e.g., mitogen-activated protein (MAP) kinase) that are associated with cytoskeletal elements responding to mechanical deformation. MAP kinases are members of a well characterized protein kinase system, and they have been shown to mediate cell responses to physical forces such as osmotic stress (0370) and stretch ((0375),0380). However, the pathways leading to activation of MAP kinase by physical forces are not well understood.

There is recent evidence, that flow may activate dual signal-transduction pathways in endothelial cells. As described earlier, one pathway is Ca²⁺ dependent and involves increases in intracellular Ca²⁺ and activation of phospholipase C ((0385),0390). A Ca²⁺-independent pathway may also contribute to flow-mediated activation of NO synthase. Ayajiki et al. (0395) found that flow-stimulated NO release from the perfused rabbit iliac artery has a biphasic nature. An initial Ca²⁺-dependent phase is followed by a phase of sustained NO release that is Ca²⁺ independent. This pathway appears to require the activity of the sodium-hydrogen ion exchanger as well as activation of protein kinase C and tyrosine kinase.

A kininergic mechanism may contribute to flow-mediated vasodilation. Flow may induce the release of bradykinin from endothelial cells (0400). Subsequently bradykinin may stimulate its respective receptors on the endothelial cell to induce the release of NO. Indeed, Groves et al. (0405) recently addressed this issue using quantitative coronary angiography to study the flow-mediated vasodilation in the epicardial coronary arteries of subjects without coronary artery disease. They found that the bradykinin receptor antagonist Hoe 140 significantly inhibited flow-mediated vasodilation but did not affect receptor-mediated endothelium-dependent vasodilation in these subjects. The study indicates that bradykinin may play an autocrine role in flow-mediated vasodilation.

Can regular exercise alter vascular reactivity so as to enhance vasodilation? Evidence is accumulating that regular exercise can exert beneficial effects on vascular reactivity, and that these salutary changes are due to exercise-induced increases in blood flow. Long-term changes in flow exert their effects on endothelium-dependent vasodilation by modulating the expression of NO synthase. The expression of mRNA for NO synthase is up-regulated in cultured endothelial cells exposed to laminar shear stress (0410). Similarly, the magnitude and nature of shear stress have a major impact on endothelial cell NO synthesis (0415). In cultured human umbilical vein endothelial cells, laminar shear stress dose dependently up-regulates NO synthesis, whereas turbulent shear stress has no effect on the NO synthase pathway.

The changes in endothelial reactivity and NO synthase expression seen with sustained increases in flow are similar to changes in endothelial behavior in vessels exposed for long periods to intermittent increases in flow (as with exercise). Exercise-induced changes in vascular reactivity have been observed in the coronary arteries of dogs after 8 weeks of exercise (0420). In patients with congestive heart failure, there is an impaired flow-mediated vasodilation of the brachial artery; exercise training enhances this NO-mediated vasodilation (0425). These studies suggest that exercise-induced changes in blood flow alter vascular responsiveness. These findings are consistent with recent observations from animal studies in which short-term (0430) and long-term ((0185),(0215),0430) exercise increased the mRNA expression of NO synthase, augmented NO activity and enhanced endothelium-dependent vasodilation in coronary arteries.

The mechanism by which flow augments the expression of NO synthase involves shear stress responsive elements (SSRE) within the promoter region of the gene. The consensus sequence for the SSRE in the NO synthase promoter is GAGACC, a putative transcription factor binding site that is common to the promoter regions of many endothelial genes that are responsive to shear stress including tissue plasminogen activator, intercellular adhesion molecule, TGF-beta₁, PDGF-beta, and endothelin-1 ((0435),0440). There are almost certainly other cisacting transcriptional regulatory elements and proteins responsive to shear stress.

Chronic increases in blood flow may also effect the elaboration of prostacyclin. This may be particularly important in circulations (such as skeletal muscle microvasculature) where this prostanoid plays a larger role in flow-mediated vasodilation. Koller et al. (0215) demonstrated in a rat model that the sensitivity of gracilis muscle arterioles to wall shear stress was up-regulated after short-term daily exercise, which resulted in an augmented dilator response, probably due to increased release of both endothelium-derived NO and prostacyclin. Hecker et al. (0445) observed continuous release of NO and prostacyclin from the endothelium that was enhanced by an increase in shear stress; the release of both vasodilators counteracts neurogenic and myogenic vasoconstriction in vivo. However, in human conduit arteries the contribution of prostacyclin to flow-mediated vasodilation is negligible, as NO plays the major role (0450).

The effects of flow on vascular reactivity may have clinical implications. It has been reported that individual patients with coronary artery disease who regularly participate in intensive physical exercise may show a significant increase in maximal rate-pressure product (an equivalent of myocardial oxygen consumption) ((0455),0460) or a decrease in exercise-induced myocardial ischemia (as documented by thallium-201 scintigraphy), or both, despite progression of coronary artery disease ((0085),0095). These observations may be explained in part by improvements in vascular reactivity.

Can regular exercise effect vascular structure and promote patency? The answer appears to be in the affirmative, and exercise-induced changes in blood flow probably contribute to this beneficial effect. Vessels are capable of significant remodeling in response to long-term changes in flow. In various animal models, vigorous endurance-type exercise training enlarges the diameter of coronary arteries, and this effect is independent of the exercise-induced hypertrophy of the end-organ supplied by the vessel ((0110),(0115),(0140),0465). In the canine carotid artery, increases in blood flow enlarge vessel diameter; this change is associated with an increased rate of protein turnover (0470), providing evidence that structural adaptation occurs in response to increased flow.

Just as flow-mediated vasodilation is endothelium dependent, so too is flow-mediated vascular remodeling. Langille and O’Donnell (0475) found that a long-term decrease in flow through the rat carotid artery causes the vessel caliber to diminish in response to structural modification of the arterial wall. This vascular remodeling is abolished by removal of the endothelium. Conversely, increasing blood flow or shear stress by an arteriovenous shunt causes structural changes that result in a larger vessel ((0470),0480). This increase in vessel diameter in the setting of increased flow reduces endothelial shear stress toward the preshunt value. This finding implies that the endothelium may be the transducer (and generate the paracrine effectors) of the hemodynamic signal that triggers vascular remodeling in response to flow. The endothelial mechanisms mediating these structural effects remain undefined. It appears plausible that the endothelium induces changes in vessel structure by producing mediators that regulate cell growth, extracellular matrix production and proteolysis. Flow modulates the endothelial elaboration of multiple paracrine factors that may affect vascular growth including NO (0190), PDGF-beta (0485), TGF-beta₁ (0360), tissue plasminogen activator (0490) prostacyclin (0215,0220,0225,0235,0495) and endothelin-1 (0285,0290,0295,0300,0305).

In the normal human coronary circulation, increased blood flow is associated with vasodilation of epicardial vessels ((0500),0505). By contrast, a paradoxic vasoconstriction to increased blood flow has been documented in atherosclerotic coronary arteries. The dilation of diseased vessels in response to nitroglycerin (endothelium-independent vasodilation) but not to increases in flow implicates endothelial dysfunction as the underlying cause of the altered vasomotion (0500). Even in subjects without angiographic evidence of atherosclerosis, exposure to coronary risk factors is associated with inhibition of endothelium-dependent vasodilation (0510). Progressive impairment of endothelial vasodilator function in human coronary arteries becomes manifest even at the early stages of atherosclerosis. In patients with hypercholesterolemia but an-giographically normal coronary arteries, there is selective impairment of acetylcholine-induced (but not flow-induced) endothelium-dependent vasodilation; with more evidence of disease, there is loss of flow responsiveness as well. A strong correlation has been reported between total serum cholesterol and the impairment of endothelium-mediated increases in coronary blood flow (0515).

Flow-mediated vasodilation of the brachial artery is often impaired before the onset of clinical evidence of atherosclerosis, and this impairment was manifested even in children as young as 7 years of age with hypercholesterolemia (0520). Subsequent studies (0525) revealed that loss of flow-mediated endothelium-dependent dilation in the brachial artery precedes the clinical onset of vascular disease; this dysfunction is associated with the same risk factors (hypercholesterolemia, hypertension, diabetes mellitus, tobacco exposure) known to predispose to atherosclerosis and its complications in later life.

Atherosclerotic lesions develop mainly near branches, bifurcations and bends, all areas of low shear stress, where there is separation from unidirectional laminar blood flow, reversal of flow and turbulence ((0480),0530,0535,0540,0545,0550,0555,0560,0565). Low shear stress is associated with increased wall thickening ((0480),(0550),0565,0570,0575) and accelerated lumen narrowing in moderately diseased human coronary arteries (0580). There is a continuous inverse relation between shear stress and rate of lumen narrowing (0580). Whereas low shear stress is associated with atherogenesis and disease progression, regions of moderate to high shear stress are relatively spared of intimal thickening as long as flow remains unidirectional and axially aligned (0565). Animal studies and clinical trials (0585,0590,0595) have documented a direct correlation between graft flow and long-term graft patency; reduced vascular runoff after percutaneous transluminal angioplasty is predictive of restenosis. Flow-induced release of paracrine modulators of vascular growth (e.g., NO and prostacyclin) may explain the effect of laminar flow to inhibit restenosis and atherogenesis.

Increased flow causes an increase in endothelial shear stress that is associated with enlargement of nonatherosclerotic arteries, as seen in arteriovenous fistulas ((0470),0480). The same phenomenon may occur in atherosclerosis. As plaque impinges on the vessel lumen, endothelial shear stress increases, given the same rate of flow. This might be expected to trigger the same remodeling response that occurs in other settings of increased shear stress. Indeed, in human coronary arteries lumen narrowing of up to 40% may be fully compensated by enlargement of the vessel, thus preventing the stenosis from becoming functionally relevant (0600). By contrast, areas of low shear stress or disturbed flow are vulnerable to atherogenesis. The mechanism by which low or disturbed flow accelerates atherogenesis and restenosis is probably multifactorial. In areas of disturbed flow, recirculation prolongs contact of atherogenic particles with the endothelium. In these areas of low shear stress, platelets and monocytes are more likely to adhere to the vessel wall (0605). These local differences in shear stress may explain why there is an increased rate of atherosclerosis progression among different vessels within the same patient (and even different segments within the same vessel) despite exposure to the same lipoprotein concentrations (0580).

Finally, areas of low shear stress may be more prone to intimal lesion formation, because in these areas the endothelium may produce less vasoprotective factors, such as NO and prostacyclin. Considering the beneficial effects of NO and prostacyclin, it could be speculated that the spatial distribution of atherosclerotic plaques in areas of low shear stress may in part be explained by a reduced elaboration of NO and prostacyclin at these sites. Indeed, endothelium-dependent vasodilation is reduced at bifurcations in human coronary arteries before angiographically observable disease develops(0610).

Animal studies ((0185),(0215),(0430),0450) indicate that an exercise intervention may reverse endothelial dysfunction. Preliminary studies from (0615) suggest that this observation also holds true for humans; young smoking army recruits manifested partial improvement in flow-mediated vasodilation of the brachial artery after several weeks of regular exercise as part of their military training. These alterations in flow-induced release of NO probably extend to the endothelial elaboration of other paracrine factors. The beneficial effects of exercise on endothelium-dependent vasodilation may have significant consequences for vascular structure as well as vascular reactivity. In addition to its effects on vasomotion, NO is known to antagonize key processes involved in atherogenesis, including monocyte adherence and chemotaxis (0620), platelet adherence and aggregation (0625,0630,0635) and vascular smooth muscle proliferation (0640). Enhancement of NO activity in hypercholesterolemic rabbits reduces monocyte adherence and accumulation, whereas inhibition of NO activity enhances endothelial adhesiveness for monocytes and increases the extent of lesion formation (0645,0650,0655).

The mechanisms by which physiologic levels of shear stress inhibit atherogenesis are beginning to be understood. Exposure of endothelial cells to fluid flow in vitro reduces their adhesiveness for monocytes (0660,0665,0670,0675). This effect of flow is mediated by alterations in signal transduction of adhesion, as well as by changes in the expression and activation of endothelial adhesion molecules and chemokines (0660,0665,0670,0675,0680,0685). Flow-stimulated release of NO rapidly inhibits endothelial adhesiveness for monocytes, with a time course (in minutes) that is most consistent with an effect on adhesion signaling. However, with longer time courses, flow-induced NO also affects expression of adhesion molecules. Specifically, exposure of endothelial cells to fluid flow for 2 h suppresses lipoprotein- or cytokine-induced expression of vascular cell adhesion molecule-1 (VCAM-1) (0665). This effect is abrogated by NO synthase antagonists, indicating that endothelium-derived NO participates in the regulation by flow of VCAM-1 expression. NO modulates VCAM-1 expression by regulating redox-sensitive transcriptional pathways ((0680),0690). Oxidant stress activates transcriptional factors (e.g., nuclear factor kappa-B), leading to the expression of oxidant responsive genes such as VCAM-1 and monocyte chemotactic protein-1 (0695). NO may exert a counterregulatory role by inhibiting the local generation of oxygen-derived free radicals, thereby reducing oxidant stress. Preliminary evidence from our group (0665) reveals that fluid flow reduces the elaboration of superoxide anion by endothelial cells in an NO-dependent manner. This finding is consistent with previous observations that NO inhibits the activity of the oxidative enzyme NADPH oxidase (0700). Furthermore, fluid flow enhances the expression of superoxide dismutase, which scavenges superoxide anion (0705).

Flow-stimulated release of prostacyclin may also modulate processes that alter vascular architecture ((0215),(0225),(0235),0495). Prostacyclin, as well as its analogues, inhibits the uptake of cholesterol esters into macrophages and vascular smooth muscle cells ((0710),0715). Prostacyclin also inhibits the adherence of platelets and neutrophils to the endothelium, which may explain the observations that administration of exogenous prostacyclin analogues reduces the size of experimentally induced infarcts (0720,0725,0730,0735) and the extent of arrhythmias (0740). Accordingly, it is conceivable that exercise-induced increases in prostacyclin release may produce beneficial effects not only through vasodilation, but also by inhibiting platelet and neutrophil adherence. The elaboration of prostacyclin is reduced in atherosclerosis. Vessel wall segments involved by atheromatous plaques do not produce prostacyclin (0745). This finding is consistent with observations in animal models that reveal that prostacyclin production by the vessel wall is abolished by hypercholesterolemia and deendothelialization (0750). Disease-induced alterations in the elaboration of this factor may contribute to the impairment of vascular reactivity and remodeling in response to flow.

There is now abundant epidemiologic and experimental evidence indicating that physical exercise slows the progression of vascular disease and reduces cardiovascular morbidity and mortality. The mechanisms of this effect include beneficial changes in lipoprotein profile, rest blood pressure and heart rate, carbohydrate tolerance, neurohormonal activity and exercise-induced increases in blood flow. Exercise-induced increases in blood flow appear to have direct effects on vascular function and structure. Flow enhances endothelium-dependent vasodilation by increasing the vascular expression of NO synthase and by enhancing the release of NO and prostacyclin. NO and prostacyclin inhibit multiple processes involved in atherogenesis and restenosis (including generation of superoxide anion, adherence of monocytes, aggregation of platelets and proliferation of vascular smooth muscle). In addition to NO and prostacyclin, flow modulates the expression of a panoply of paracrine substances, including endothelial growth factors, matrix modulators, adhesion molecules, chemokines and regulators of blood fluidity, all of which may participate in the beneficial effects of exercise-induced vascular remodeling and reactivity. Several fundamental basic and clinical questions remain unanswered. Only a few elements of the endothelial signal transduction pathways mediating the response to shear stress have been identified. This is also true for the transcriptional mechanisms by which shear stress induces gene expression. A role for these cellular mechanisms in the clinical effects of exercise has not been established. It is not known how much exercise is needed to restore endothelial function, how long these effects might persist after exercise and to what degree these effects contribute to the salutary effects of exercise on cardiovascular health. The investigation of these questions is likely to lead to novel approaches to pharmaco-therapy and exercise rehabilitation for the patient with coronary artery disease.