Coronary microvascular adaptations distal to epicardial artery stenosis

Abstract

Until recently, epicardial coronary stenosis has been considered the primary outcome of coronary heart disease, and clinical interventions have been dedicated primarily to the identification and removal of flow-limiting stenoses. However, a growing body of literature indicates that both epicardial stenosis and microvascular dysfunction contribute to damaging myocardial ischemia. In this review, we discuss the coexistence of macro- and microvascular disease, and how the structure and function of the distal microcirculation is impacted by the hemodynamic consequences of an epicardial, flow-limiting stenosis. Mechanisms of endothelial dysfunction as well as alterations of smooth muscle function in the coronary microcirculation distal to stenosis are discussed. Risk factors including diabetes, metabolic syndrome, and aging exacerbate microvascular dysfunction in the myocardium distal to a stenosis, and our current understanding of the role of these factors in limiting collateralization and angiogenesis of the ischemic myocardium is presented. Importantly, exercise training has been shown to promote collateral growth and improve microvascular function distal to stenosis; thus, the current literature reporting the mechanisms that underlie the beneficial effects of exercise training in the microcirculation distal to epicardial stenosis is reviewed. We also discuss recent studies of therapeutic interventions designed to improve microvascular function and stimulate angiogenesis in clinically relevant animal models of epicardial stenosis and microvascular disease. Finally, microvascular adaptation to removal of epicardial stenosis is considered.

INTRODUCTION: FOCUS ON CORONARY MICROVASCULAR VERSUS MACROVASCULAR DISEASE

Cardiovascular disease is the leading cause of death worldwide, with myocardial ischemia being an important cause of morbidity and mortality. In the healthy heart, the coronary microvasculature plays an important role in the regulation of coronary blood flow to ensure adequate oxygen supply to all regions of the myocardium (1). Any increase in myocardial oxygen demand requires vasodilation of coronary small arteries and arterioles to facilitate an increase in blood flow. In addition, in the presence of clinically significant obstructive lesions in the large coronary arteries (i.e., epicardial stenosis), the increase in resistance in these conduit vessels is initially compensated by a decrease in resistance in the coronary microcirculation to maintain adequate coronary blood flow distal to the stenosis. Only with severe stenoses does the vasodilator capacity of the microvasculature fall short, resulting in myocardial ischemia.

Chilian and colleagues were some of the firsts to describe alterations in vascular reactivity in the microcirculation of atherosclerotic animal models (2, 3). Data from these studies revealed augmented vasoconstriction and impaired vasodilation in the microcirculation distal to atherosclerotic lesions. Specifically, vasoconstriction to serotonin and ergonovine was augmented in diseased coronary arteries as well as the microvasculature distal to atherosclerotic lesions in cynomolgus monkeys (2). Based on these findings, the authors concluded that because vasoconstrictor responses were potentiated in both the diseased arteries and the microcirculation distal to the atherosclerotic lesion, the pathophysiological consequences of atherosclerosis extend into the microcirculation. In a subsequent study, these investigators demonstrated significantly impaired endothelium-dependent dilation to ADP, histamine, and serotonin, as well as a loss of response to flow in atherosclerotic pigs (3), providing additional evidence for impaired function in the microvasculature distal to atherosclerotic lesions. Since these early studies, there is a vast amount of literature in which impaired vascular function has been reported both experimentally and clinically in the microcirculation of nonocclusive cardiovascular diseases, including hypercholesterolemia, metabolic syndrome, ischemia with nonobstructive coronary arteries, obesity, hypertension, and atherosclerosis.

Over the last decade, it has been increasingly recognized that epicardial stenoses are not the sole cause of myocardial ischemia, but that microvascular dysfunction also plays a role. Microvascular dysfunction has been clinically classified into four types (4). Isolated microvascular disease is referred to as type 1, whereas microvascular disease occurring in the presence of myocardial or valvular disease or in the presence of epicardial stenosis are referred to as type 2 and 3, respectively. Finally, microvascular dysfunction secondary to iatrogenic mechanisms such as reperfusion injury and intervention-related microembolization is categorized as type 4 (4).

As suggested by the above-mentioned classifications, micro- and macrovascular disease often coexist. This coexistence is associated with worse prognosis following percutaneous coronary intervention (PCI) (5). Risk factors, such as age, hypertension, and diabetes, impact both the macro- and microvasculature, and endothelial dysfunction is likely a common denominator contributing to the development of micro- and macrovascular disease. However, other factors, including the contribution of changing hemodynamics and microvascular rarefaction, that underlie the coexistence of macro- and microvascular disease are less well understood. Furthermore, changes in the coronary microvasculature are not limited to changes in microvascular function, but the structure of the microvasculature changes as well. On the one hand, reductions in pressure and/or flow have been shown to induce inward remodeling of microvessels (6, 7). On the other hand, episodes of myocardial ischemia induce ingrowth of new vessels from adjacent perfusion territories (angiogenesis) as well as outward remodeling of preexisting connecting vessels (arteriogenesis) in a process called collateral growth. This allows, at least in part, restoration of myocardial blood flow to areas distal to the stenosis.

In this review, we will discuss the reciprocal relationship between a stenosis and the microvasculature, with a stenosis contributing to development and/or aggravation of microvascular dysfunction while also inducing microvascular remodeling and angiogenesis. We will discuss 1) how changes in coronary hemodynamics induced by stenosis impact the distal coronary microvasculature, 2) the impact of risk factors on macro- and microvascular function as well as their interaction, 3) how a stenosis impacts microvascular structure and induces collateral formation, 4) how lifestyle changes, in particular exercise training, may alleviate some of these functional and structural changes in the coronary microvasculature, and finally 5) if and how removal of a stenosis can reverse some of the changes in the coronary microvasculature.

MICROVASCULAR FUNCTION DISTAL TO A STENOSIS

Resting coronary blood flow (CBF) to the left ventricular myocardium is ∼0.7–1.0 mL·min⁻¹ per gram of myocardium. CBF is determined by perfusion pressure and coronary vascular resistance. In the healthy heart, perfusion pressure equals aortic pressure and coronary vascular resistance is mainly determined by the coronary microvasculature, that is, vessels smaller than 200 μm in diameter, with negligible contribution of the large epicardial coronary arteries. Indeed, previous findings (8, 9) have revealed that under control conditions ∼25% of total coronary resistance resided in arteries >170 µm in diameter, ∼68% between arterioles <170 µm and venules <150 µm, and ∼7% in the coronary veins. Administration of dipyridamole redistributed coronary resistance away from the microcirculation and into small arteries and veins, such that resistance increased to ∼42% in arteries >170 µm and decreased to ∼27% in the microvasculature, with 31% redistributed to the veins >150 µm. Thus, under ischemic conditions, where the microcirculation undergoes vasodilation to enhance blood flow into the compromised myocardium, a shift in resistance and thus control of coronary blood flow occurs toward the small arteries. Any increase in myocardial oxygen demand is met by vasodilation of the coronary microcirculation and results in an increase in CBF (1). The ratio between resting flow and maximal flow, referred to as coronary flow reserve (CFR), is 4–5 in a healthy heart. Yet, there are regional differences in CFR. CFR is generally lower in the subendocardium than the subepicardium. This lower subendocardial CFR is due to cardiac contraction, which compresses intramyocardial vessels, thereby increasing intravascular pressure and impeding blood flow to the subendocardium, in particular (10). To compensate for the increased extravascular compression, vascular density is slightly higher in the subendocardium (1), although this is not an unanimous finding (7).

Hemodynamic Consequences of a Stenosis: Autoregulation and Steal

Stenosis in an epicardial coronary artery creates a site of increased resistance to coronary blood flow and, depending on the stenosis severity, induces a decrease in pressure distal to the stenosis, and thereby, in perfusion pressure of the distal myocardium (Fig. 1). Severity of stenosis is clinically determined using the ratio between pressure distal and proximal of the stenosis under maximal vasodilation, which is called the fractional flow reserve (FFR). A stenosis is considered flow limiting with an FFR below 0.75, and/or when exceeding a 50% reduction in diameter. However, vasodilation in the coronary microcirculation results in a decrease in microvascular resistance in an attempt to maintain myocardial perfusion. This process, that maintains coronary blood flow, and hence myocardial perfusion, constant over a wide range of perfusion pressures is called autoregulation (1). In dogs and humans, autoregulation can maintain a constant coronary blood flow until perfusion pressure distal to a stenosis drops below 40 mmHg (1). In terms of diameter, a 90% reduction of diameter is required for a reduction in CBF under resting conditions. At this point, autoregulatory reserve is exhausted and resting flow decreases. Autoregulatory reserve is not equally distributed across the myocardium; flow reserve of the subendocardium is slightly lower than that of the subepicardium. Hence, at a perfusion pressure of 40 mmHg, flow to the subendocardium begins to decrease, whereas flow to the subepicardium is maintained. This explains why myocardial ischemia, as well as myocardial infarction, is usually more pronounced in the subendocardium than in the subepicardium.

Figure 1.

Hemodynamic changes and microvascular adaptations in the coronary circulation distal to epicardial stenosis. Ach, acetylcholine; ADP, adenosine diphosphate; bFGF, basic fibroblast growth factor; BK, bradykinin; Ca, calcium; CFR, coronary flow reserve; FFR, fractional flow reserve; IMR, index of microvascular resistance; KCl, potassium chloride; NO, nitric oxide; P, pressure; Pa, pressure proximal to stenosis; Pd, pressure distal to stenosis; Q, flow; SNP, sodium nitroprusside; VEGF, vascular endothelium growth factor. Adapted images from Servier Medical Art: Creative Commons License; Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License.

When intracoronary infusion of a vasodilator is used to measure stenosis severity and corresponding flow reserve, vasodilation occurs in the territories with residual vasodilator reserve. Hence, the vessels in the subepicardium dilate, whereas vasodilator reserve in the vessels in the subendocardium is exhausted. The subsequent increase in flow results in further pressure drop across the stenosis (or along a gradually narrowed coronary artery) and hence in a decrease in perfusion pressure, resulting in a decreased flow into the subendocardium. Thus, flow to the subepicardium increases at the expense of flow to the subendocardium, a phenomenon called coronary “steal”(1, 11). This steal phenomenon may be aggravated under conditions that further impair subendocardial perfusion, such as an increase in extravascular compressive forces either due to a decrease in diastolic time and/or to increased diastolic left ventricular pressure (11).

Hemodynamic Assessment of the Microvasculature

Myocardial blood flow can be assessed non-invasively using positron emission tomography (PET), which can quantify regional absolute blood flow. However, in the presence of a stenosis, a combination of several invasive measurements, that can be performed in the cath laboratory, is required to appropriately assess the coronary microcirculation. First, coronary angiography is performed to assess the presence or absence of a stenosis, using intracoronary administration of contrast agent. Semiquantitative analysis of the passage of contrast agent through the microvasculature using blush score as well as the TIMI (thrombolysis in myocardial infarction) frame score can be used as a measure of myocardial perfusion in the absence of a stenosis. A corrected TIMI frame count >27 (images acquired at 30 frames/s) suggests microvascular dysfunction with impaired resting flow. However, the presence of a stenosis precludes such analysis. Microvascular resistance in the presence of a stenosis can be assessed using a diagnostic guide wire procedure. A wire combining pressure and flow measurements is required to assess both stenosis severity and microvascular resistance. Two types of wires are available, one that combines pressure measurements with flow velocity, and one that combines pressure and thermodilution.

The FFR, i.e., the ratio of pressures distal and proximal to the stenosis (Pd/Pa) under maximal vasodilation is an index of stenosis severity, whereas CFR, the ratio of flow with maximal vasodilation to baseline flow, is an indicator of total coronary vascular resistance (Fig. 1). Using the thermodilution method, CFR is calculated as the ratio of mean transit time at baseline and maximal vasodilation (12, 13). Validation studies in a porcine model have shown that CFR measurements derived from transit times reflect invasively measured CFR better than flow velocity measurements (14). In addition, using a special catheter with side holes to inject saline for the thermodilution measurements also results in maximal vasodilation to the same extent as adenosine that is conventionally used to obtain maximal vasodilation (15). To delineate the contribution of the stenosis and microvascular alterations to a reduction in flow reserve, the index of microvascular resistance (IMR) is calculated as the product of distal pressure and mean transit time during vasodilation (Fig. 1). An IMR >25 is considered to reflect microvascular dysfunction (5, 16).

One pitfall in IMR measurements is that IMR may be falsely elevated in the presence of collateral flow. This flow is not measured by the thermodilution method, as it comes from adjacent coronary arteries. Correction for wedge pressure eliminates this artificially high IMR (5, 16). As indicated by the name, IMR and CFR measure vascular resistance, which is a composite measure of microvascular structure, function and extravascular compression. In addition to these clinical measurements, direct measurement of microvascular function is possible in animal models that combine in vivo with in vitro measurements (17). Assessment of coronary microvascular function in vivo requires measurement of the relation between myocardial oxygen supply and myocardial oxygen demand, as interventions modulating coronary microvascular function often also impact systemic vascular resistance and hence blood pressure and/or heart rate. Furthermore, coronary small arteries and arterioles can be isolated and their function examined using pressure or wire myography. This allows delineation of endothelial and smooth muscle cell dysfunction in the absence of confounding changes in hemodynamics, neural input, or metabolic changes in the surrounding myocardium.

Role of Endothelial Dysfunction

The healthy endothelium responds to mechanical and chemical signals with production of numerous factors that regulate vascular tone, permeability, cellular adhesion, angiogenesis, platelet aggregation, smooth muscle cell proliferation, and vascular inflammation. In normal conditions, endothelial cells display an antithrombotic, anti-inflammatory phenotype with a balance of vasodilators and vasoconstrictors that allows for vascular reactivity to successfully meet the metabolic demands of tissues. Endothelial dysfunction results from several abnormalities in vascular function that tend to promote cardiovascular pathologies. In pathological conditions, the endothelium shifts to a prothrombotic, proinflammatory phenotype, and the vasculature displays reduced vasodilatory responses by a change in a variety of factors that are produced and released by the endothelium (Fig. 1). Thus, it is not uncommon for endothelial dysfunction to result from a cardiovascular pathologic state and then in turn exacerbate the underlying pathology.

Endothelial dysfunction is predictive of coronary artery disease and typically progresses as the severity of disease advances. Furthermore, endothelial dysfunction impairs the ability of the microvasculature to dilate making the patient with disease more vulnerable to impaired blood flow distal to stenosis. Microvascular dysfunction has more often been viewed as a consequence of acute coronary syndromes that arise from the presence of epicardial disease and resultant plaque rupture, whether spontaneous or iatrogenic; however, emerging evidence indicates that the presence of microvascular dysfunction influences large artery function and plaque stability. In patients with normal or minimally diseased coronary arteries, cardiovascular event rate was significantly (threefold) higher in patients with the lowest CFR, suggesting that microvascular dysfunction predisposed these patients to coronary events (18). Indeed, a recent study by Usui et al. (19) demonstrated that a higher IMR, measured downstream of coronary arteries with intermediate-to-severe lesions, was associated with increased prevalence of thin-cap fibroatheroma, larger lipid volume of epicardial lesions, and higher prevalence of subclinical plaque rupture. It is important to note that the coexistence of micro- and macrovascular disease may be attributable to factors acting on the endothelial cells in the micro- and macrovasculature alike. Endothelial dysfunction in the coronary microvasculature and the consequent impaired coronary microvascular vasodilation may be the first signs of coronary vascular disease. Such impaired vasodilator response limits the increase in coronary blood flow and shear stress in the coronary arteries, thereby predisposing them to development of endothelial dysfunction (20).

In the initial phase of coronary atherosclerosis, plaque formation is accompanied by outward remodeling of the coronary arteries, maintaining the lumen area constant (21, 22). This so-called Glagov phenomenon has been shown to be able to maintain lumen area until the percent stenosis exceeds nearly 40%. However, ∼40% of vessels fail to show outward remodeling. Low shear stress and high baseline plaque area are risk factors for plaque progression (23) in patients with stable angina pectoris, whereas high shear stress is associated with plaque regression (22). In the presence of nonobstructive coronary artery disease, microvascular function is the main determinant of maximal blood flow. Hence, it is likely that plaque progression is associated with microvascular dysfunction. Taken together, these findings in both animal models and clinical studies suggest that the existence of microvascular dysfunction, and endothelial dysfunction in particular, contribute to the progression of epicardial lesion development and plaque rupture. However, our understanding of the influence of coronary microvascular dysfunction on the development of epicardial lesions and progression toward acute coronary syndromes is limited by a lack of animal models that encompass all of the factors that are present in human disease. Most animal models of epicardial stenosis are not able to recapitulate the presence of slowly developing microvascular disease in an aging myocardium influenced by the presence of comorbidities that are common in patients with epicardial stenosis.

Studies using a porcine model of chronic coronary artery occlusion have reported functional adaptations in the endothelium of coronary arterioles (∼70–170 µm luminal diameter) distal to occlusion (Fig. 1). These studies revealed that bradykinin-mediated, endothelium-dependent vasodilation was significantly impaired in arterioles isolated from the collateral-dependent region of chronically occluded hearts compared with those from a corresponding nonoccluded (control) region in the same heart (24–26) or compared with vessels from hearts that were not occluded (25). Similar findings were observed with substance P (27) and adenosine diphosphate (ADP) (25, 27–30). Dilation to the receptor-independent calcium ionophore, A23187, was not different in collateral-dependent and arterioles isolated from control hearts (25). Furthermore, reactivity to the endothelium-independent nitric oxide donor, nitroprusside, have shown no difference compared with arterioles isolated from the nonoccluded region (24, 27–29). These vascular reactivity studies reveal that collateral-dependent arterioles exhibit impaired endothelium-dependent vasodilation that may be attributed to reduced endothelium-derived vasodilators in arterioles from the ischemic myocardial region. It is also important to note that some of these studies were completed 4–7 wk (25) following surgical placement of the ameroid occluder and others 14–16 wk (24, 26) postoperatively, suggesting that impaired endothelium-dependent dilation in the collateral-dependent region is persistent in this animal model. Consistent with impaired receptor-dependent, endothelium-mediated vasodilation in the porcine model, a canine model of chronic coronary artery occlusion also revealed that vasodilation to the endothelium receptor-dependent agonists ADP and acetylcholine was significantly attenuated in collateral-dependent coronary microvessels (∼100–220 µm luminal diameter), whereas dilation to the receptor-independent calcium ionophore, A23187, was not different from microvessels isolated from control dogs (31). Taken together, these data suggest that microvasculature that is dependent on the collateral circulation for blood flow demonstrate impaired production or enhanced degradation of endothelium-derived vasodilating mediators with no impairment in the response of smooth muscle to dilate to exogenous nitric oxide donors.

Additional studies sought to determine the underlying mechanisms of impaired endothelium-derived vasodilation in collateral-dependent microvasculature. The first of these supported a role for impaired nitric oxide bioavailability in collateral-dependent arterioles (Fig. 1) in that the nitic oxide synthase inhibitor, N^G-monomethyl-l-arginine, abolished the difference in bradykinin-mediated dilation between arterioles of the collateral-dependent and nonoccluded myocardial regions (24). In the same studies, investigators also reported the eNOS mRNA expression was significantly lower in collateral-dependent compared with nonoccluded arterioles (24). Subsequent studies demonstrated a potential role for impaired nitric oxide signaling in collateral-dependent arterioles at lower and potentially more physiological concentrations of the endothelium-derived bradykinin concentrations, but these studies demonstrated that both total eNOS and phospho-eNOS (Ser1179) protein levels were statistically increased in arterioles from the collateral-dependent compared with nonoccluded regions (32). Interestingly, these findings suggested that an increase in eNOS protein and presumably increased eNOS activity via increased phosphorylation of the protein did not translate to increased contribution of nitric oxide to endothelium-dependent dilation in collateral-dependent arterioles (32). It is important to note that NOS can produce both NO and superoxide. When electron flow from the oxidase domain to the reductase domain of NOS is diverted to molecular oxygen rather than l–arginine, superoxide is produced and NOS is said to be uncoupled (33). Thus, the observed increase in eNOS protein without increased contribution of nitric oxide to dilation may suggest eNOS has become uncoupled in the collateral-dependent vasculature.

In addition to reduced bioavailability of nitric oxide, diminished activity of the nitric oxide-soluble guanylyl cyclase (sGC)-cGMP signaling pathway is also recognized as a major contributor to impaired dilation and reduced angiogenic responses in vascular disease states. A decline in the contribution of this pathway has been attributed to not only reduced nitric oxide availability, as discussed above, but also increased oxidation of sGC, which makes this enzyme unresponsive to nitric oxide. Indeed, increases in reactive oxygen species that are often involved in the development of cardiovascular diseases can drive increased protein oxidation as well as limit nitric oxide bioavailability. Evidence from the literature demonstrates that sGC is the downstream mediator for VEGF-nitric oxide-stimulated increases in vascular permeability and angiogenesis (34). Consistent with a substantial role of sGC in angiogenesis and vasodilation and ineffective treatment with nitric oxide donors due to nitrate tolerance, several activators and stimulators of sGC, with distinct mechanisms of action related to the state of the sGC heme iron, have been or are in the process of being developed as potential therapeutic agents (35).

Vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) also have been shown to have vasoactive properties in the microcirculation, in addition to angiogenic properties and structural adaptations brought on by these growth factors. Evidence in a porcine model of chronic coronary artery occlusion has shown that vasodilation elicited by VEGF and bFGF are increased in collateral-dependent compared with nonoccluded arterioles (Fig. 1) and dilation was mediated through receptor tyrosine kinase and nitric oxide production (30). Consistent with increased dilation in collateral-dependent arterioles, Northern blot analysis of both the collateral-dependent and nonoccluded myocardium of the left ventricle revealed that expression of VEGFR-1, VEGFR-2, and FGFR-1 mRNA was significantly increased the collateral-dependent region. Taken together with previous findings of impaired endothelium-dependent dilation in collateral-dependent arterioles, these data suggest that the coupling of nitric oxide through tyrosine kinase receptors that are upregulated in ischemic myocardium is sufficient to overcome impaired endothelium-dependent dilation that is observed through G-protein coupled receptors. However, contrary to these data, others have shown no difference in vasodilation of collateral-dependent compared with nonoccluded arterioles as well as no alteration in VEGFR-1 or VEGFR-2 protein levels as measured by immunoblot (36). Thus, further studies are necessary to elucidate the contributions of growth factor signaling in vasodilation distal to chronic occlusion or stenosis.

Role of Smooth Muscle Dysfunction

Although investigation into microvascular smooth muscle adaptations to chronic ischemia are not as prevalent as evaluation of endothelial function, there have been a number of studies that have explored responsiveness to smooth muscle-specific vasoactive agents in the collateral-dependent vasculature (Fig. 1). Studies using a porcine model of chronic coronary artery occlusion demonstrated that vasodilation to the β-adrenergic agonist, isoproterenol (29, 37), as well as the adenylyl cyclase activator, forskolin, were significantly attenuated (37) in arterioles (∼75–150 µm luminal diameter) isolated from the collateral-dependent compared with nonoccluded myocardial region. Yet, in other studies, vasodilation to forskolin (29) and adenosine (38) were not different in arterioles isolated from the collateral-dependent compared with nonoccluded myocardial regions. Thus, our understanding of cAMP-mediated dilation in vasculature distal to chronic occlusion remains ambiguous.

Similarly, conflicting findings have also been reported regarding contractile responses to endothelin-1 in collateral-dependent arterioles in that there is evidence of significantly increased contractile response to endothelin-1 compared with arterioles from the nonoccluded region (6, 28) and no difference between collateral-dependent and nonoccluded arterioles (39) (Fig. 1). In a canine model of chronic coronary artery occlusion, vasoconstriction to vasopressin was significantly increased in collateral-dependent coronary microvessels (∼100–220 µm luminal diameter) compared with microvessels isolated from control dogs (31). Recent studies have reported significantly increased contractile responses to high KCl in collateral-dependent arteries compared with those from the nonoccluded region in a porcine model of chronic coronary artery occlusion (40). The enhanced contractile response to KCl was attributed to increased calcium sensitization of the contractile elements and no change in voltage-gated calcium channel currents (40). Beyond these findings, there is a paucity of information regarding mechanisms that underlie adaptations in smooth muscle in the microcirculation distal to stenosis since most literature is focused on endothelial dysfunction, a key component of cardiovascular disease rather than alterations in smooth muscle function.

Contribution of Emboli Formation to Microvascular Dysfunction

Coronary microembolization resulting from either plaque rupture or percutaneous coronary intervention can lead to coronary microvascular embolization and persistence of reduced microvascular flow even after recovery of epicardial flow. Early autopsy studies of patients who died after acute coronary syndromes provided evidence that angioplasty or thrombolytic therapy led to distal embolization of the microcirculation and microinfarctions (41, 42). In early work in dogs (43), infusion of microspheres to produce microvascular emboli demonstrated a progressive decline of regional myocardial function in the absence of a decrease in regional myocardial blood flow. In both pigs (44) and dogs (43), evidence of invasion by leukocytes and macrophages indicates that inflammation at the site of coronary microembolization resulted in decline of tissue function and localized necrosis. Indeed, two recent studies suggest that myocardial infarction in swine is larger, and cardiac dysfunction is more severe when ischemia-reperfusion is combined with embolization of the distal microvasculature either with microspheres (45) or with autologous thrombi (46). Microemboli may be a source of procoagulants and inflammation, initiating a signaling cascade through nitric oxide and TNFα, resulting in myocyte damage and contractile dysfunction, even in the absence of changes in myocardial flow (47). Investigation of coronary aspirate obtained during stent procedures has shown that debris produced by plaque rupture can also be a source of vasoconstrictors and prothrombotic agents including serotonin, thromboxane, and TNFα (48, 49). For further discussion of the consequences of coronary microvascular emboli and viable treatment options, the reader is referred to excellent reviews by Heusch (50) and Heusch et al. (51). Thus, it has become increasingly clear that treatment aimed at restoring blood flow to the ischemic myocardium distal to a stenosis or thrombotic occlusion must target the microcirculation in addition to reopening of the epicardial vessel (52). After occlusion of an epicardial artery and following reperfusion, multifactorial processes within the microvasculature and in surrounding myocytes, initiated by microemboli, reactive oxygen species and inflammatory factors, contribute to progressive microvascular injury through obstruction by the microemboli in combination with endothelial swelling, thereby aggravating myocardial damage. Interventions targeted at mitigating this damage will need to take an integrated approach that considers inflammation of the microcirculation and the surrounding myocytes at the site of microemboli.

Changes in Microvascular Permeability

Changes in microvascular permeability lead to microvascular fluid filtration, myocardial edema formation, and increased myocardial interstitial fluid pressure. Myocardial function is markedly compromised with small changes in interstitial fluid volume, demonstrating high sensitivity to increases in microvascular permeability that occur in many disease states as well as clinical interventions such as cardiothoracic surgical procedures (53). Indeed, evidence reveals that myocardial edema can develop during myocardial ischemia before the onset of irreversible myocardial injury and thus, may serve as a diagnostic marker (54). Growth factors, such as VEGF, directly stimulate microvascular permeability allowing for filtration into the interstitium, which contributes in a regulated manner to angiogenesis, but also is more marked during disease states potentially resulting in edema and diminished cardiac function (55, 56).

IMPACT OF RISK FACTORS AND LIFESTYLE ON FUNCTIONAL ADAPTATION AND REMODELING OF THE MICROVASCULATURE DISTAL TO A STENOSIS

Diabetes

In a significant proportion of patients with epicardial coronary artery disease, percutaneous coronary intervention reestablishes coronary artery patency, but not myocardial reperfusion (57–59), indicating that microvascular dysfunction is a critical contributor to ischemia distal to an epicardial stenosis. Although the coronary microvasculature does not develop atheroma, microvascular dysfunction distal to a stenosis is more prevalent in patients with risk factors, including diabetes (60–62) (Fig. 2). Murthy et al. (60) reported that mortality from cardiovascular events was higher in patients with epicardial stenosis and diabetes compared with patients with epicardial stenosis without diabetes. In patients who are nondiabetic, risk for cardiovascular-related death was also elevated if epicardial stenosis was accompanied by significantly reduced CFR. These findings suggest that, in patients with diabetes, epicardial stenosis and microvascular dysfunction combine to elevate the risk for major cardiac events and cardiac-related death. CFR is reduced in patients with diabetes compared with controls, even in the absence of angiographically identified cardiovascular disease (62). In the diabetic heart, altered myocardial energy metabolism can lead to reduced microvascular resistance at rest (63), and eventual reduction in hyperemic flow due to impaired vasodilatory function, both of which contribute to impairment of CFR (64). Microvascular dysfunction in the diabetic heart is characterized by inappropriate vasoconstriction (65–67), loss of endothelium-dependent vasodilation (68, 69), impaired K channel-mediated dilation (70), and structural remodeling of coronary arterioles (66, 71–73). Capillaries are also altered in the diabetic myocardium (74); their diameter decreases and they become more tortuous, likely contributing to reduced resting perfusion. These findings in both humans and animal models underscore the significant role that microvascular dysfunction plays in the development of acute coronary syndromes. Results of these studies also emphasize the need for treatment of risk factors, especially diabetes, following restoration of macrovascular flow by percutaneous intervention or thrombolytic treatment.

Figure 2.

Potential mechanisms that influence adaptation of the coronary microcirculation distal to epicardial stenosis. Adapted images from Servier Medical Art: Creative Commons License; Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License.

Animal models have allowed for mechanistic study of microvascular function distal to epicardial stenosis in the diabetic myocardium. In a porcine model of chronic ischemia, induction of diabetes resulted in mild impairment of relaxation of coronary arterioles to ADP and substance P, and marked impairment of relaxation to VEGF (75). In contrast, endothelium-independent relaxation to sodium nitroprusside was not altered. Microcirculatory rarefaction, induced by chronic regional ischemia, was greater in diabetic porcine hearts as compared with control hearts (76). Thus, although the presence of microvascular disease results in pronounced elevation of microvascular resistance distal to epicardial stenosis in the diabetic heart, a temporal relationship between epicardial lesion development, metabolic changes in the diabetic myocardium, and altered microvascular reactivity has not been clearly established in humans. The time-dependent contribution of these factors during development and remediation of stenosis will need to be identified in animal models to develop the most effective therapies for restoration of microvascular function in the ischemic diabetic myocardium.

Metabolic Syndrome

Metabolic syndrome is a condition encompassing hypertension, obesity, hyperlipidemia, and glucose intolerance/type II diabetes. Metabolic syndrome (77) and hypercholesterolemia (78) are associated with diminished collateral formation and endothelial dysfunction (79) (Fig. 2). Metabolic syndrome significantly increases the risk of cardiovascular disease and mortality (80). Patients with components of metabolic syndrome experience significantly greater mortality after angioplasty (81), suggesting that metabolic syndrome contributes to microvascular dysfunction distal to epicardial stenosis. In the coronary circulation, the metabolic syndrome enhances vasoconstrictor tone, disrupting the balance between myocardial metabolism and coronary oxygen delivery (82). In a study designed to assess the angiogenic effects of VEGF in the ischemic myocardium of pigs fed a hypercholesterolemic diet, arteriolar vasodilation to ADP and VEGF were impaired in the ischemic region (83) and the angiogenic response to VEGF was also significantly impaired in the ischemic region. When hypertension is induced in a pig model of coronary artery stenosis, cardiac dysfunction and myocardial fibrosis are aggravated; however, somewhat paradoxically, hypertension attenuates the impairment of coronary vasodilation to adenosine induced by coronary artery stenosis (84). Similarly, Lassaletta et al. (85) reported that microvascular responsiveness to sodium nitroprusside increased in coronary arterioles from young Ossabaw swine fed a hypercaloric, high-fat/cholesterol diet despite the development of dyslipidemia, hypertension, and glucose intolerance. In addition, myocardial perfusion and capillary and arteriolar density were maintained in the overfed swine. Metformin administration ameliorated diet-induced hypertension and glucose intolerance, but metformin did not alter hemodynamics or perfusion in the ischemic myocardium, suggesting that at a relatively early stage of metabolic syndrome, compensatory mechanisms provide vascular protection despite cardiac metabolic dysfunction and onset of hypertension. The use of animal models to better understand the timeframe by which microvascular dysfunction develops in response to the metabolic syndrome may provide insight into the interaction between the development of epicardial disease, microvascular dysfunction, and myocardial dysfunction. In turn, a better understanding of these interactions and the role of microvascular dysfunction in acute coronary syndromes and recovery after an acute coronary event in the setting of metabolic syndrome, could contribute to development of adjuvant therapies following a revascularization procedure or even novel compounds for drug-eluting stents. Certainly, the recognition that the metabolic syndrome profoundly affects the function of the microcirculation distal to epicardial stenosis emphasizes the urgent need to monitor and control risk factors after a revascularization procedure.

Several lines of evidence indicate that coronary microvascular dysfunction that occurs with myocardial ischemia in the setting of metabolic syndrome is linked to oxidative stress. Dhawan et al. (86) reported that plasma glutathione levels were an independent predictor of microvascular dysfunction in patients with components of metabolic syndrome undergoing cardiac catheterization. Coronary collateral growth, stimulated by repetitive ischemia, is severely compromised in obese Zucker rats, a rodent model of metabolic syndrome. Gene therapy with VEGF is only effective in stimulating coronary collateral growth in obese Zucker rats when administered in combination with gene therapy to increase extracellular superoxide dismutase and alleviate oxidative stress (77). Similarly, Pung et al. (87) reported that treatments which reduce mitochondrial oxidative stress restore coronary collateral growth in obese Zucker rats. In a porcine model of chronic ischemia, 20 wk of hypercholesterolemic diet resulted in increased cardiac tissue oxidative stress, reduced VEGF expression, reduced capillarity, endothelial dysfunction, and reduced perfusion of the ischemic myocardium (78). Despite the lipid-lowering effects of statins, in swine models of chronic ischemia with hypercholesterolemia (88) or metabolic syndrome (89), atorvastatin treatment was associated with increased myocardial oxidative stress that possibly contributed to the lack of improvement in collateral-dependent perfusion in the ischemic region of the myocardium. When considered together, these results suggest that the metabolic syndrome impairs homeostatic regulation of oxidative stress in the ischemic myocardium, contributing to reduction of angiogenesis and NO-mediated vascular function.

Aging

Aging is a primary risk factor for the development of cardiovascular dysfunction and disease including coronary artery disease, ischemic heart disease, and heart failure (90) (Fig. 2). Studies of the interaction between epicardial stenosis and coronary microvascular dysfunction occur primarily in older adults (91–93); thus, aging is a default risk factor in these studies. In animal studies, aging alters contractile function (94–98) and impairs vasodilatory function (99–102) of coronary arterioles. Reduction of endothelium-dependent vasodilation of coronary arterioles is related to decreased NO bioavailability (99–103) and increased endothelial inflammation (104).

Aging also promotes remodeling of coronary resistance vessels (72, 94). Tomanek et al. (105) reported that capillary numerical density was 16% lower in midmyocardium and 19% lower in endomyocardium of old beagles and that capillary length density (capillary volume) was reduced by 27% in the endomyocardium of old beagles. In the human heart, the angiogenic capacity of the coronary circulation decreases with age (106), resulting in a mismatch between angiogenesis and cardiac hypertrophy in the aged heart. The adaptive responses of the coronary microcirculation, distal to epicardial stenosis, have not been studied in the setting of aging, either alone or in conjunction with other comorbidities that are prevalent in aged humans (90). Faber et al. (107) found that aging causes collateral rarefaction in murine gastrocnemius muscle. Immediately following femoral artery ligation perfusion was lower in 16-, 24-, and 31-mo-old mice compared with young (3-mo-old) mice. Perfusion pressure to the hindlimb was not lower; thus, assuming that hypoxia induces maximal dilation distal to the ligation, lower flow in the hindlimb of the aged mice can be attributed to age-related reduction of perfusion through native collaterals. Recovery of blood flow following femoral ligation was also reduced in aged mice, indicating that ischemia-induced collateral growth and angiogenesis decline with age in the mouse hindlimb (108). Presumably, studies of the coronary microcirculation downstream of occlusion or atherosclerotic stenosis are lacking in old animal models because 1) the life span of large animal models that develop atherosclerosis is long enough to be cost-prohibitive for studies of this nature, 2) most rodent models do not develop atherosclerosis (rodent models that do develop atherosclerosis also have serious cardiovascular and metabolic disease with greatly shortened life spans), and 3) occlusion of rodent epicardial arteries, especially mice, is technically difficult and more likely to result in major infarction than a timely restriction of flow. Rabbits develop atherosclerosis when fed a hypercholesterolemic diet, and have a life span of 24–36 mo; thus, the rabbit could be developed as an animal model of coronary atherosclerosis in the aged myocardium (109). The influence of aging must be taken into consideration when studying the coronary microcirculation distal to a stenosis for two primary reasons: 1) aging alters the morphology and function of myocardium and its metabolic influence on the coronary microcirculation, and 2) the resiliency of the coronary microcirculation, both in terms of vasoactive function and angiogenic capacity, declines with age. Thus, although challenging, it is imperative that aging be incorporated into animal models when investigating adaptations of the coronary microcirculation distal to a stenosis.

Exercise Training

The impact of exercise training in the coronary microcirculation distal to chronic stenosis or occlusion is difficult to assess in vivo because one cannot delineate adaptations in the collateral circulation from those in the microcirculation distal to the stenosis and how these components individually contribute to changes in blood flow into the collateral-dependent myocardial region. Use of animal models allows for isolation of the microvasculature for study under in vitro conditions. Such studies are not possible in human subjects that have undergone exercise training regimens as it is impossible to acquire heart tissue samples from these individuals. Initial in vitro studies that assessed the effects of exercise training on coronary microcirculation distal to chronic coronary artery occlusion revealed that impaired endothelium-dependent dilation in arterioles from the collateral-dependent region of sedentary pigs was corrected after completion of a progressive treadmill exercise training regimen (5 days/wk for 16 wk) (24). Consistent with improved exercise training-induced endothelium-dependent dilation, these investigators also reported that eNOS mRNA expression which was significantly reduced in collateral-dependent arterioles of sedentary pigs was restored to control levels after exercise training (24).

Additional studies demonstrated that collateral-dependent arterioles from exercise-trained pigs displayed significantly increased VEGF₁₆₅-mediated dilation compared with those of sedentary animals as well as with arterioles from the nonoccluded region of sedentary and exercise-trained pigs (38). These findings suggested that both chronic occlusion and exercise training are necessary to increase sensitivity of the vasculature to VEGF₁₆₅-induced dilation. The enhanced dilation in the collateral-dependent arterioles was reversed by the inhibition of nitric oxide synthase, tyrosine kinase, as well as the interaction of VEGFR-2 with neuropilin (36, 38). The biological effects of VEGF in endothelial cells are mediated through two high-affinity tyrosine-kinase membrane receptors, VEGFR-1 and VEGFR-2. Furthermore, the nontyrosine-kinase, cell surface receptor, neuropilin-1, enhances the binding of VEGF₁₆₅ to VEGFR-2 in vascular endothelial cells, increasing the potency of VEGF. Immunoblot analysis in these studies revealed that collateral-dependent arterioles of exercise-trained animals exhibited significantly increased neuropilin-1, membrane and soluble forms of VEGFR-2, and VEGFR-1 protein levels (36). Taken together, these data suggest that enhanced VEGF₁₆₅-mediated vasodilatation in collateral-dependent arterioles from exercise-trained pigs may be mediated by increased receptor levels of VEGFR-1, VEGFR-2, and neuropilin-1. Furthermore, increased protein levels of VEGF receptors after exercise training in these animals may also contribute to arteriogenesis and collaterogenesis to support blood flow into the myocardial region distal to occlusion.

More recent reports have revealed that exercise training corrects impaired endothelium-dependent dilation in collateral-dependent arterioles through cellular adaptations that increase agonist-stimulated H₂O₂, contributing to enhanced dilation after exercise training (32). These data implicate a role for endothelium-derived vasodilators that are stimulated by exercise training, in addition to contributions of nitric oxide. Additional experiments in the same studies suggested that NADPH oxidase was a potential source of increased H₂O₂ in the collateral-dependent vasculature after exercise training (32). Further data support a role for BK_Ca channels as a downstream effector that contributed to enhanced dilation in the collateral-dependent arterioles (26). More studies are needed to confirm the source of H₂O₂ as well as the downstream signaling pathways that drive increased dilation in collateral-dependent arterioles after exercise training.

The effect of exercise training on mechanisms of basal active tone in the microvasculature distal to chronic occlusion has also been explored. These studies revealed that collateral-dependent arterioles of exercise-trained pigs displayed significantly enhanced Ca²⁺-dependent resting tone compared with arterioles from nonoccluded regions of exercise-trained pigs and arterioles from collateral-dependent and nonoccluded regions of sedentary pigs (110). These findings indicate that the combination of chronic coronary artery occlusion and exercise training produce a marked increase in Ca²⁺-dependent resting tone, despite a nominal effect of either occlusion or exercise training alone. In these same studies, arterioles from the collateral-dependent region of exercise-trained pigs displayed significantly enhanced constriction in response to treatment with the NOS inhibitor, l-NAME, as well as increased p-eNOS (Ser1179) and total eNOS protein levels, compared with arterioles from the other treatment groups (110). The greater contractile response to NOS inhibition in collateral-dependent arterioles of exercise-trained pigs suggested that enhanced nitric oxide activity may mask the increased Ca²⁺-dependent resting tone observed in these arterioles under basal conditions. Additional data from these studies also revealed that Kv channels contribute markedly to resting tone in arterioles of nonoccluded and collateral-dependent myocardial regions in both sedentary and exercise-trained animals. Intriguingly, the contractile response to Kv channel blockade (4-aminopyridine) was statistically greater in arterioles from the collateral-dependent region of exercise-trained pigs, indicating that Kv channels contribute to the regulation of basal tone to a greater extent in arterioles distal to occlusion after exercise training (110). Although seemingly contradictory, the enhanced Ca²⁺-dependent basal active tone, nitric oxide production, and K⁺ channel activity may generate an enhanced coronary flow reserve in addition to increased vasodilatory capacity in the collateral-dependent region after exercise training, and thereby a more precise regulation of blood flow into the at-risk myocardium. The stimulus for these adaptations may be most marked in the collateral-dependent region because of coupling of intermittent ischemia with mechanical forces on the vascular wall during each bout of exercise. Indeed, exercise-induced myocardial ischemia persists in this animal model despite improvements in coronary blood flow and myocardial contractile function after chronic exercise training (111).

Recent studies have also revealed increased contractile behavior of small coronary arteries (∼150–350 µm) from the collateral-dependent region of exercise-trained animals. Data from these studies demonstrate significantly enhanced endothelin-1-mediated contractile responses in collateral-dependent arteries compared with those from the nonoccluded region of exercise-trained animals and with both the collateral-dependent and nonoccluded regions of sedentary pigs (39). Based on findings of increased endothelin-1-mediated tension development at comparable intracellular Ca²⁺ levels in arteries from the collateral-dependent region of exercise-trained pigs, these studies revealed enhanced Ca²⁺ sensitization of the smooth muscle contractile apparatus as the mechanism underlying increased contraction, as well as upregulation of the protein kinase C signaling pathway (39). Interestingly, systemic and coronary endothelin-1-induced vasoconstriction have been shown to wane during dynamic exercise in pigs (112–115) and humans (116), suggesting that vasoconstrictors aid in the maintenance of resting blood flow, and potentially coronary flow reserve, and as blood flow demands increase during exercise, the influence of vasoconstrictors lessens.

More recent studies have also documented enhanced contractile response to KCl in small collateral-dependent coronary arteries (∼150–350 µm) that is further amplified with exercise training. This response was attributed to vascular smooth muscle cell Ca²⁺ sensitization but not to changes in Ca²⁺ channel current. Additional evidence implicated mediators of Ca²⁺ sensitization, Rho-kinase, and CaMKII pathways, as contributing to increased tension development in collateral-dependent arteries after exercise training. Increased contractile responses observed in collateral-dependent coronary arteries may serve to oppose the enhanced vasodilation responses previously reported following exercise training in this model of myocardial ischemia (24, 26, 117). An abundance of evidence documents that exercise training improves vascular endothelial function in individuals with coronary artery disease, ultimately leading to enhanced myocardial perfusion (118, 119). Increased contractile activity of the collateral-dependent vasculature may provide a balance to offset enhanced endothelial function and thereby maintain the distinctly close relationship between coronary blood flow and myocardial oxygen consumption normally observed in the heart, as well as contribute to maintenance of a functional coronary flow reserve.

COLLATERAL GROWTH AND ANGIOGENESIS

Collateral Growth

In adult animals, collateral growth occurs through expansion of a preexisting collateral network. Arteriogenesis and angiogenesis are two different processes of vascular remodeling that contribute to expansion of the collateral network and, relatedly, increased perfusion in the ischemic heart (120). Arteriogenesis is defined as remodeling of preexisting arterial vessels in a morphological process that involves a concurrent increase in intralumenal area and wall thickness. Angiogenesis is defined as new capillary growth that occurs when new vessels bud from preexisting capillaries. Fulton (121) presented initial evidence of arterial anastomoses that undergo outward remodeling and increase in diameter in the presence of epicardial stenosis. Collateral formation has been well documented in dog (122, 123) and swine models (31, 124, 125) of chronic coronary ischemia induced by placement of an ameroid occluder. Development of collaterals in response to occlusion of a coronary artery restores blood flow to ischemic areas of myocardium (Fig. 2). Stimulation of collateral growth represents a potential therapy for reestablishment of blood flow distal to an epicardial stenosis; however, much remains unknown about the process of collateral development in the ischemic human heart. For an excellent recent review of coronary collateral growth, the reader is referred to Jamaiyar et al. (126). The following section will focus on the effects of exercise training on collateral growth for two reasons: 1) the effects of exercise training on collateral growth were not covered in the review by Jamaiyar, and 2) cardiac rehabilitation, that is, exercise training, is arguably the best adjuvant therapy currently available to patients following treatment for epicardial stenosis in the cath laboratory or patients recovering from myocardial infarction.

Effects of Exercise Training on Collateral Growth

Historically, evidence of exercise training-induced collateral formation in human patients with coronary artery disease (CAD) is conflicting. However, many of these studies have used coronary angiography to assess collateral development, which lacks the sensitivity to detect small intramyocardial collateral vessels (<200 μm), and thereby may limit detection of vessels potentially responsible for improved blood flow into collateral-dependent myocardial regions during exercise.

In a randomized study of 113 patients with well-documented coronary artery atherosclerotic lesions, Niebauer et al. (127) examined the impact of an exercise training regimen (3 days/wk; >3 h weekly; 75% maximal heart rate) combined with a low-fat, low-cholesterol diet on collateral development over the period of a year. Collateral development was assessed by angiography and retrograde filling of the dependent coronary artery and revealed no significant difference between the exercise-trained and sedentary patients. Collateral growth was instead significantly related to changes in the degree of stenosis, in that progression of atherosclerotic lesions corresponded to an increase in collateral formation and regression of lesions corresponded to a decrease in collateral formation. The authors concluded that after 1 yr of exercise training and consumption of a low-fat diet, there was no significant difference in the number of collateral vessels or in the degree of opacification (marker of collateral filling) but also acknowledged that the inherent resolution limitations of angiographical assessment may have influenced the outcome of their studies. In a subsequent study (128), 90 of the original 113 patients were again assessed after 6 yr of intervention (n = 40 exercise plus diet; n = 50 sedentary). Data were consistent with that observed in the original 1-yr study in that no significant changes were observed in the total number of coronary collaterals within or between groups, thus providing further evidence that exercise training did not increase collateral vessel growth at a level detectable by angiography (128). Additional studies have also failed to observe collateral development using angiography in patients with recognized coronary artery disease (129, 130).

Despite a number of studies that have exhibited no detectable increase in collateralization with exercise training, a small number of studies have detected exercise training-enhanced collateral development as assessed by angiography. Belardinelli et al. (131) explored the effects of moderate exercise training (bicycle ergometry, 3 times/week for 8 wk; 60% V̇o_2max) on collateral development in patients with CAD with left ventricular systolic dysfunction. Contrary to the outcomes of other studies, the 8-wk exercise training regimen produced significant increases in collateral development as assessed by angiography and retrograde filling of the dependent coronary artery. Enhanced collateral development was associated with increased thallium activity and augmented contractile response of dysfunctional myocardium in response to low-dose dobutamine. These investigators also reported that patients with higher collateral scores at baseline displayed more remarkable development of collateral vessels with exercise training, suggesting a predictive relationship between collateral development and initial levels of collateral vessels (131). In a subsequent study using the same methodology and exercise training regimen, Belardinelli et al. (132) reported that dipyridamole (75 mg, 3 times/day for 8 wk) increased coronary collateralization, as well as myocardial thallium uptake and contractility in patients with CAD and that exercise training further potentiated the effects of dipyridamole.

Regardless of conflicting outcomes regarding increased coronary collateralization in response to exercise training in these angiography-based studies, the exercise-trained groups exhibited improvements in cardiorespiratory fitness measures, such as peak oxygen uptake, ischemic (angina) threshold, increased exercise capacity, and in some studies, increased myocardial perfusion (measured by thallium scintigraphy) in the collateral-dependent region. These findings indicate that exercise training improves the regional function of compromised myocardium, contributing to enhanced cardiac function as a whole, in spite of the lack of clarity regarding the contribution of exercise training-induced collateral growth and maturation.

In more recent years, Seiler and colleagues (133–135) have developed a novel technique to assess the functional presence of collateral vessels and adaptations in collateral flow with therapeutic interventions, such as exercise training, which they have termed collateral flow index. During angiography, a transient balloon occlusion of the stenotic vessel was performed and the coronary pressure (or flow velocity) beyond the occlusion was measured and expressed as a ratio to aortic pressure (or flow velocity), providing an index of collateral perfusion distal to the occlusion. In a small, nonrandomized study, patients underwent a 3-mo endurance exercise training program (3 days/week; 60 min/day; 80% of heart rate at V̇o_2max) and patients were divided into exercise training and sedentary groups based on whether they adhered to the prescribed exercise training regimen and showed evidence of increased V̇o_2max and exercise capacity. After completion of the 3-mo protocol, collateral flow index increased statistically in both stenosed and angiographically normal coronary arteries of exercise-trained but not sedentary patients with an apparent dose-response relationship between collateral flow index and exercise capacity gained (134). The observed increase in collateral flow index in coronary arteries without flow-limiting stenoses suggests that improved endothelial function in the coronary microcirculation (a well-documented phenomenon after exercise training) and the resultant reduction in vascular resistance may be a significant contributor to the measure of increased collateral flow index, independent of collateral expansion. Furthermore, the increase in pressure distal to the occlusive balloon fell into the range of ∼3–5 mmHg, and thus, the physiologic significance of this increase in pressure is ambiguous. Seiler et al. (135) have also demonstrated increased collateral flow index after exercise training in one healthy individual undergoing an ultramarathon training program. In this study, the pressure distal to the occlusive balloon increased ∼9 mmHg over the time that the subject went from an intermediate to a high fitness level. This case study is the first to demonstrate increased collateral flow in the absence of coronary artery disease and is in contrast to a number of studies in canine and porcine models that observed no increase in collateral blood flow after exercise training in the absence of occlusion or stenosis (136–138).

In a randomized study of 60 patients with significant coronary artery disease, patients were assigned to high-intensity or moderate-intensity exercise training regimen for 4 wk or a control group that did not exercise for the same duration (139). High-intensity training consisted of supervised exercise 4 times/day for 30 min, 5 days/wk at 70% of patients’ individual angina-free exercise threshold with some interval training up to 95% of threshold. The moderate intensity protocol consisted of six to eight supervised exercise sessions per day for 20 min each at 60% of threshold. After completion of the exercise training regimen, collateral flow index increased significantly and similarly in both moderate and high-intensity exercise-trained groups, whereas collateral flow index in the control group remained unchanged (139). Yet, as observed above in the study by Seiler and colleagues (134), the increase in pressure distal to the occlusive balloon was in the range of ∼2–5 mmHg (139), and thus, the physiologic and clinical significance of this increase to improvements in perfusion and myocardial function is unclear.

Functional adaptations in the heart as a result of exercise training are numerous and one must consider that exercise training-induced improvements in blood flow to the compromised myocardium are the combination of collateral growth and maturation and changes in resistance in the microcirculation of the collateral-dependent vasculature downstream of the stenosis or occlusion as well as potentially more efficient cardiac contractile cell function. Although recent studies reflect increased collateral development after exercise training, direct and unambiguous evidence that the observed degree of collateral growth has a major impact on increased blood flow into the collateral-dependent myocardium and subsequent clinical improvement in cardiac function remains equivocal (140, 141). Although it is apparent that exercise training generates improvements in coronary blood flow and myocardial function in patients with coronary artery disease, the contribution of collateral growth to these improvements remains unresolved.

Despite controversial findings in clinical studies that have evaluated the effects of exercise training on collateral vessel development in patients with coronary artery disease, animal models of chronic coronary artery occlusion typically provide experimental evidence for favorable effects of exercise training on collateral growth and development and improved perfusion and myocardial function in the compromised myocardium. Canine and porcine models of chronic coronary artery occlusion are commonly used experimentally for the induction of collateralization and are typically induced by fixed stenosis or gradual occlusion of a coronary artery with an ameroid constrictor. The dog heart possesses an abundance of innate epicardial collateral vessels and develops a significant collateral network after gradual coronary artery occlusion (142, 143). Numerous investigators have examined collateral development after coronary occlusion in the dog heart and have demonstrated that collateral blood flow to the myocardial region distal to occlusion increases to levels comparable to those present in the corresponding nonoccluded region of the same heart even during exercise stress within months after onset of occlusion (144–147).

In a seminal paper using a canine model of surgically induced chronic narrowing (fixed stenosis) of the left circumflex coronary artery, Eckstein (136) explored the effects of a 6- to 8-wk exercise training regimen on collateral development. Collateral blood flow was determined by measurement of retrograde flow into the distal left circumflex artery and was found to be proportional to the degree of arterial narrowing. Exercise training markedly increased collateralization relative to that observed in sedentary dogs. Furthermore, although mild arterial narrowing failed to stimulate collateralization in sedentary dogs, exercise training elicited collateral vessel development even with mild stenosis. These studies also demonstrated that myocardial ischemia, as well as development of a pressure gradient across innate collateral vessels, provided the underlying stimuli for collateral growth following arterial narrowing, mechanisms that are still thought to underlie collateral vessel growth in current literature (136). Although the study lacked baseline data, as measurements were not obtained before the onset of the sedentary and exercise protocols, the use of a relatively large number of dogs strengthened the conclusions drawn from this study (136).

A number of ensuing studies examined the effects of exercise training on collateral vessel growth, and subsequent regional perfusion and myocardial function in canine models of experimentally induced critical narrowing or those subjected to gradual left circumflex coronary artery occlusion with an ameroid constrictor (138, 148, 149). These studies, which have been reviewed in detail previously (124), confirmed the findings of Eckstein and demonstrated that exercise training leads to increased collateral vessel development as well as increased blood flow and contractile function of the ischemic myocardium that were not evident in sedentary dogs. In contrast to these studies, in a small number of dogs subjected to chronic occlusion with an ameroid occluder, exercise training did not appear to elicit collateral growth beyond that observed with occlusion alone, as determined using microspheres, angiography, and retrograde flow (150). However, typical exercise training induced adaptations, such as increased heart-to-body weight ratio and decreased heart rate at submaximal workloads, were absent suggesting that the exercise program was not sufficiently strenuous to elicit exercise training-related adaptations (150).

Collateral vessel development in the porcine model of chronic coronary artery occlusion more closely mimics that of human coronary artery disease patients. Specifically, both porcine and human hearts demonstrate few innate collateral vessels and the growth of new vessels in occlusive coronary artery disease usually occurs as an extensive network of functionally significant collaterals in the midmyocardial and endocardial layers (142, 151, 152). Furthermore, collateral growth and development is typically sufficient to provide normal blood flow to the collateral-dependent myocardium under resting conditions; however, blood flow during stress (e.g., exercise) remains compromised and unable to support function of the at-risk myocardium (111, 152–155). These characteristics are similar to patients with advanced coronary artery disease who often lack clinical symptoms at rest, but exhibit significant signs of myocardial ischemia (ECG abnormalities, angina) during periods of increased cardiac demand.

In a porcine model of critical stenosis of the left circumflex coronary artery that progressed to complete occlusion in all animals, Bloor et al. (156) explored the effect of exercise training on blood flow to both the proximal and distal portions of the stenosed artery. Exercise training consisted of treadmill running (5 days/wk, 35 km/wk, 6–8 km/h) for five mo and resulted in significant exercise training-induced adaptations, including lower heart rates at submaximal workloads and increased time to exhaustion during exercise stress test. Blood flow (microspheres) into the left circumflex artery proximal to the occlusion was similar to that determined in the control left anterior descending artery. Collateral-dependent flow into the distal left circumflex coronary artery after exercise training increased to 79% of flow in the proximal left circumflex, which was statistically greater than the increase to 63% of proximal blood flow in sedentary animals. Thus, collateral-dependent blood flow to the distal left circumflex artery was increased to a significantly greater degree after exercise training compared with the response to occlusion alone. It is important to note that blood flow into the collateral-dependent region was measured only under resting conditions in these studies and therefore, the effect of exercise training on collateral flow during increased myocardial demand was not assessed. Furthermore, myocardial function of the collateral-dependent region was not determined and thus, the clinical significance of exercise training-enhanced blood flow into the compromised myocardium was not assessed (156).

In a subsequent study, Bloor et al. (111) examined the effect of exercise training on coronary collateralization and regional myocardial function after gradual occlusion by placement of an ameroid occluder around the left circumflex coronary artery. Collateral blood flow (microspheres) and regional myocardial function (systolic wall thickening) were determined at rest and during both moderate (∼75% maximal heart rate) and severe (∼95% maximal heart rate) exercise bouts before and after completion of a 5-wk progressive treadmill exercise training (50 min/day, 5 days/wk, ∼75% maximum heart rate in the 5th week of training) or sedentary protocol. After completion of the exercise training protocol, the blood flow ratio of the collateral-dependent region to the normally perfused region was increased significantly in the subendocardium, midmyocardium, and subepicardium during severe exercise after the training protocol compared with pretraining levels and statistically increased only in the subendocardium during moderate exercise. No significant increases in blood flow ratios were observed in the sedentary group during severe exercise when the pretraining and posttraining exercise tests were compared. In the sedentary pigs, the blood flow ratio was increased significantly only to the subendocardium during moderate exercise and to a lesser extent than that observed in the exercise-trained pigs. Consistent with adaptations in regional blood flow, exercise training significantly improved myocardial function, as assessed by measures of systolic wall thickening, in the collateral-dependent region during severe exercise levels compared with pretraining levels. Thus, these studies revealed enhanced perfusion into the collateral-dependent region with associated improvements in regional myocardial function that were observed primarily under conditions of high-intensity exercise, although improvements in blood flow into the subendocardial region were also observed during moderate-intensity exercise. Thus, findings from these studies suggest that effects of exercise training, or any other therapeutic intervention, on collateral blood flow may be most apparent during periods of high myocardial demands where maximal collateral-dependent vasodilator reserve or maximal collateral conductance are achieved.

Angiogenesis Distal to Epicardial Stenosis

Angiogenesis distal to an epicardial stenosis, similar to angiogenesis in other vascular beds, is influenced by shear stress, ischemia, and metabolic signals from the surrounding myocardium. Favorable results were obtained from studies that attempted to induce angiogenesis in the ischemic myocardium using growth factors in healthy large animal models (157–160); however, clinical trials using growth factors were not successful (161, 162). It is now recognized that management of risk factors is critical to the recovery of patients following invasive procedures such as PCI, stent placement, and coronary artery bypass graft (163). The failure of clinical trials using growth factors in patients with cardiac ischemia led to development of animal models in which the mechanisms of angiogenesis could be studied in the ischemic myocardium in the presence of relevant risk factors and/or comorbidities. In particular, the pathophysiology associated with the presence of hyperlipidemia, hypertension, insulin resistance/type II diabetes can significantly alter the angiogenic capacity of the coronary circulation. In addition, as discussed in Aging and Exercise Training, habitual activity and age can also impact the angiogenic capacity of the ischemic myocardium.

In a seminal study, Voisine et al. (83) reported that the angiogenic response to VEGF was markedly reduced in swine fed a high cholesterol diet. Boodhwani et al. (78) investigated angiogenesis of the ischemic myocardium in pigs fed a high cholesterol diet for 13 wk, and found that, in comparison to the ischemic myocardium of pigs fed a normal diet, endothelium-dependent vasodilation and myocardial blood flow were significantly reduced. In a study using diabetic pigs with occlusion-mediated cardiac ischemia, Boodhwani et al. (75) reported endothelial dysfunction, reduced myocardial perfusion, and rarefaction of capillaries in diabetic pigs compared to control pigs. In a subsequent study of myocardial ischemia in Ossabaw pigs with early stage metabolic syndrome, myocardial perfusion, capillary and arteriolar density were maintained compared to lean Ossabaw pigs (85), and myocardial protein expression of VEGF, PPAR-α, γ, and δ was significantly increased. Hattan et al. (77) reported that coronary collateral growth, stimulated by episodic, repetitive ischemia, was impaired in obese Zucker rats as compared to lean Zucker rats. Proangiogenesis proteins notch2, notch4, jagged2, and angiopoietin 1 are all downregulated in the chronically ischemic myocardium in the setting of metabolic syndrome (164). Expression of anti-angiogenic proteins, endostatin and angiostatin, is increased in the ischemic myocardium of diabetic pigs (75). Endostatin expression was also significantly increased in the chronically ischemic myocardium of pigs fed a high cholesterol diet for 13 wk (78). When considered together, these studies in rat and porcine models of myocardial ischemia emphasize the complexity of the angiogenic response and the numerous signaling pathways that are impacted by the presence of ischemia and comorbidities that often accompany the development of myocardial ischemia in patients. Implementation of new models that incorporate risk factors that inhibit angiogenesis distal to epicardial stenosis will be needed to develop effective therapeutic strategies to promote cardiac angiogenesis after revascularization.

Therapeutic Strategies to Improve Collateral Formation and Angiogenesis

Clinical trials attempting to increase collateral growth and angiogenesis in patients with ischemia focused on the use of growth factors but were not successful (157, 161, 162, 165). In recent studies in animal models, therapeutic strategies have focused on improving lipid metabolism, reducing oxidant stress, and improving myocardial metabolism.

Statins.

In the ischemic myocardium of hypercholesterolemic pigs, high-dose atorvastatin reversed endothelial dysfunction without significant changes in perfusion in the collateral-dependent region of the myocardium (166). High-dose atorvastatin also increased expression of endostatin and decreased expression of VEGF in this pig model of cardiac ischemia with hypercholesterolemia. In a similar pig model of myocardial ischemia and hypercholesterolemia, low-dose atorvastatin treatment increased capillary and arteriolar density and upregulated eNOS and VEGF, but did not increase perfusion of the ischemic region of the myocardium (89); however, this low-dose atorvastatin treatment was associated with increased myocardial oxidative stress, which may have contributed to the lack of myocardial perfusion, despite the increase in angiogenic responses. A follow-up study confirmed that high-dose atorvastatin improved coronary vasodilatory responses and increased endothelial cell density in hypercholesterolemic swine with chronic myocardial ischemia; however, perfusion of the ischemic region of the myocardium was actually reduced in pigs treated with atorvastatin (88). In this follow-up study, high-dose atorvastatin increased myocardial protein oxidation and lipid peroxidation, and this increase in myocardial oxidant stress likely underlies the reduction in myocardial perfusion that occurred. These studies indicate that statins induce a bi-phasic effect in the ischemic heart. Statins improve endothelial function and increase expression of angiogenic proteins in the endothelium of the ischemic myocardium while simultaneously increasing oxidant stress in the myocardium. The effects of statins on myocardial metabolism may negatively affect metabolic regulation of coronary perfusion despite improving endothelium-dependent vasodilation and promoting angiogenesis.

Metformin.

Myocardial ischemia was induced with an ameroid occluder in pigs that were fed a control diet or a high-fat/cholesterol diet, with and without metformin supplementation (85). Treatment with metformin reversed the hypertension and glucose intolerance induced by the high-fat/cholesterol diet, but did not alter capillary and arteriolar density, myocardial protein oxidation, or perfusion of the ischemic region of the myocardium. Surprisingly, vasodilatory responses of coronary arterioles and expression of VEGF, PPAR-α, γ, and δ were significantly increased in pigs fed high-fat diet compared to control diet; however, metformin treatment did not alter these diet-induced changes in the coronary circulation. The observed cardioprotection in high-fat fed pigs may be secondary to changes in cardiac oxygen demand related to weight gain and/or development of hypertension. It is also possible that the protective effects of metformin are related to lipid-dependent upregulation of the PPAR pathway, which has been shown to affect both inflammation and myocardial fatty acid metabolism. A follow-up study indicated that metformin treatment of pigs with metabolic syndrome and myocardial ischemia upregulates insulin signaling in the ischemic myocardium (167), possibly conferring metabolic and vascular protection.

Antioxidant therapy.

Treatments that buffer oxidative stress in the ischemic myocardium may promote angiogenesis (168). Hattan et al. (77) reported that coronary collateral growth, stimulated by episodic, repetitive ischemia, was impaired in obese Zucker rats as compared to lean Zucker rats. In this study, treatment with VEGF restored the collateral growth in obese Zucker rats when coupled with delivery of extracellular superoxide dismutase to reduce oxidant stress. More recently, Pung et al. (87) reported that treatment with MitoQuinone and MitoTempol improved the metabolic profile of the myocardium of obese Zucker rats by reducing mitochondria-derived oxidant stress, inducing a concomitant restoration of coronary collateral growth to that seen in lean Zucker rats. Resveratrol, an activator of sirtuins, has been proposed to lower risk factors that influence cardiovascular disease progression (169) and to alter oxidant stress and improve vascular reactivity and blood flow in the ischemic myocardium (170). A recent study by Sabe et al. (171) reported that resveratrol supplementation reduced expression of several proteins involved in mitochondrial dysfunction in the ischemic myocardium. In pigs with metabolic syndrome and coronary ischemia created by ameroid occlusion, inhibition of either calpain or glycogen synthase kinase 3β, a kinase activated by calpain, increased coronary blood flow and microvascular density in the ischemic region of the myocardium, and altered expression of proteins in the insulin and WNT signaling pathways (172–174). Potz et al. (175) also showed that inhibition of glycogen synthase kinase 3β in pigs with metabolic syndrome and myocardial ischemia decreased oxidative stress in myocardial tissue. These studies suggest that treatments which reduce mitochondrial dysfunction and restore oxidant/antioxidant balance in the myocardium may represent therapeutic strategies that will promote angiogenesis in the ischemic myocardium.

Microvascular Adaptations to Removal of Stenosis

Removal of a stenosis in patients with acute coronary syndrome (ACS) is associated with recovery of myocardial blood flow. IMR was shown to normalize upon PCI in the culprit vessel (176) although a more recent study found IMR to remain higher in the culprit as compared to the nonculprit vessel shortly after the procedure (177). It should be noted however that the variation in post-PCI IMR is quite large and that a higher age is associated with a higher IMR (178), which in turn is associated with a worse prognosis, that is, an increased risk for death or rehospitalization in patients with ACS (179).

Histological measurements in both patients (180) and swine (7, 181) have shown an increase in capillary density and inward remodeling of coronary arterioles distal to a stenosis. Yet, 1 mo after removal of the stenosis in swine, capillary regression was present, but remodeling of the coronary arterioles persisted (7). Conversely, IMR was found to decrease over time after removal of the obstruction both in patients two months after opening of a chronic total occlusion (12, 13) as well as in patients 6 mo after PCI for acute coronary syndrome (177). This reduction in IMR over time was accompanied by a further regression of the stenosis and an increase in diameter of the epicardial vessels (12, 182) as well as by regression of collaterals (182), confirming the dynamic interplay between the coronary macro- and microvasculature.

Conclusions

A reciprocal relationship exists between epicardial stenosis and the microvasculature, with a stenosis contributing to development and/or aggravation of microvascular dysfunction while also inducing microvascular remodeling and angiogenesis. The integrated control of pressure and flow downstream of an epicardial stenosis requires adaptation of native coronary collaterals, the coronary resistance vasculature, and angiogenesis to maintain oxygen delivery to the distal myocardium. In healthy hearts, much of this adaptation can occur relatively rapidly in response to transient ischemia and hemodynamic signals, that is, changes in shear stress and transmural pressure. However, in many patients, epicardial stenosis occurs in the setting of multiple risk factors, and these risk factors impact the adaptive responses of the microcirculation, often increasing microvascular resistance and impeding angiogenesis. Future work will require the evolution of animal models that account for the impact of these risk factors on collateral development and microcirculatory adaptations distal to epicardial stenosis, before and after stenosis removal, as well as translation of findings in animal models into viable therapies that target the stenosis, collateral growth, and the distal coronary microcirculation.

GRANTS

This work was funded by National Heart, Lung, and Blood Institute Grant HL139903 (to C. L. Heaps), German Center for Cardiovascular Research (DZHK) Grant 81Z0600207 (to D. Merkus), and Dutch Cardiovascular Alliance (DCVA-Reconnext) (to D. Merkus).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.M., J.M., and C.L.H. prepared figures; D.M., J.M., and C.L.H. drafted manuscript; D.M., J.M., and C.L.H. edited and revised manuscript; D.M., J.M., and C.L.H. approved final version of manuscript.