OUP user menu

Editor in chief

Roberto Ferrari

Associate Editors

Francisco Fernández-AvilésJeroen BaxMichael BöhmThomas F. LüscherFrank Ruschitzka


Published on behalf of

Perspectives: Coronary microvascular dysfunction in post-infarct remodelled myocardium

Daphne Merkus, Dirk J. Duncker
DOI: http://dx.doi.org/10.1093/eurheartj/sut016 A74-A79 First published online: 10 January 2014


Myocardial infarction (MI) results in alterations in cardiac structure and function that not only impact the infarcted area, but also result in changes in the remote, remodelled myocardium. These changes also encompass structural changes in the coronary vascular tree and an increase in extravascular compressive forces acting on the microcirculation, thereby blunting myocardial flow reserve. Moreover, regulation of microvascular tone in the remote coronary vasculature is altered after MI. The alterations in regulation of microvascular tone appear to be principally the result of a loss of NO bioavailability in conjunction with an increased reactive oxygen species production. Interestingly, a reduced influence of a number of vasoconstrictors, including PDE5 and endothelin, serves to blunt the abnormalities in myocardial oxygen balance in remodelled myocardium after MI. Nevertheless, the overall changes in control of microvascular tone perturb the myocardial perfusion, resulting in impaired myocardial O2 delivery to the non-infarcted regions. The accompanying cardiomyocyte contractile dysfunction and/or enhanced apoptosis may contribute to the progression of cardiac dysfunction after MI.

  • Myocardial infarction
  • Cardiac remodeling
  • Coronary circulation
  • Coronary microvascular dysfunction


The incidence and prevalence of heart failure are increasing, principally due to an increased survival of acute myocardial infarction (MI) in conjunction with an ageing population. Particularly, in patients with a large MI, left ventricular (LV) function continues to deteriorate over time. The mechanisms that contribute to the transition of LV dysfunction towards overt heart failure remain incompletely understood, but may involve (i) loss of cardiomyocytes through apoptosis,1 (ii) a primary reduction in contractile function of the surviving myocardium,2 and/or (iii) alterations in extracellular matrix leading to progressive LV dilation.3 These alterations in LV structure and function not only impact the infarcted area, but also result in changes in the remote, remodelled myocardium. Thus, myocardial blood flow (MBF) abnormalities, resulting in impaired myocardial O2 delivery to the non-infarcted regions (leading to secondary contractile dysfunction and/or enhanced apoptosis), have been suggested to contribute to the progression of LV dysfunction after MI.4 A significant reduction in MBF reserve has been shown in the surviving remodelled LV myocardium in patients with overt heart failure.5 Furthermore, even in patients with only asymptomatic LV dysfunction, flow reserve is reduced in the non-stenotic myocardial regions.6 In line with these clinical observations, we observed in a porcine model of post-infarct remodelling that during increased O2-demand induced by exercise, the increase in coronary blood flow was impaired resulting in perturbations in oxygen delivery (Figure 1).7,8

Figure 1

Schematic drawing of the coronary microvasculature and the alterations in control mechanisms of microvascular tone in post-infarct remodelled myocardium. In red are shown the mechanisms that contribute to microvascular dysfunction; in green are shown the counteracting mechanisms that blunt the increased microvascular tone; in black are shown mechanisms that are unchanged as compared with normal hearts. See text for further explanation.

Mechanisms of flow perturbations in post-myocardial infarction remodelled myocardium

The reduction in flow reserve and the associated perturbation in oxygen delivery during exercise are likely caused by the physical interaction between the coronary vasculature and the surrounding myocardium, in conjunction with structural and functional changes in the coronary microvasculature. It is evident that the coronary vasculature is compressed by the surrounding myocardium during systole.9,10 Hence, diastolic time fraction is an important determinant of myocardial perfusion.11,12 However, the coronary microvasculature is particularly sensitive for the compression that occurs during diastole, as a result of changes in LV diastolic function, i.e. elevated LV diastolic pressures7 and changes in the rate of relaxation.13 Structural changes of the remote coronary vasculature involve both inward, hypertrophic remodelling of the existing vessels14 as well as insufficient adaptations of the vasculature to keep up with the post-MI eccentric myocardial hypertrophy. The latter results in increased length of the coronary vascular tree in the presence of a maintained arteriolar density and even a decreased capillary density in the remote remodelled myocardium.7

The increased extravascular compression and structural changes in the coronary vasculature result in an increased minimal coronary microvascular resistance and a decreased myocardial perfusion, particularly in the subendocardial layers as the maximal subendocardial blood flow was blunted by 40% in anaesthetized swine with heart failure 3 weeks after an MI.15 This impairment of flow reserve is particularly important during exercise, when myocardial oxygen demand and extravascular compression increase simultaneously. Although MBF increases during exercise in the remote coronary vasculature after MI, this increase in flow did not completely match the increase in myocardial oxygen demand, as evidenced by an increase in myocardial oxygen extraction. Moreover, MBF was redistributed away from the subendocardium in favour of the subepicardium in post-MI as compared with normal swine,7 suggesting that the increase in myocardial oxygen extraction occurred predominantly at the subendocardium. Impaired perfusion of the subendocardium is likely to be, at least in part, responsible for the deterioration of LV function because local interstitial oedema and disruption of collagen fibres resemble the ultrastructural changes that occur with recurrent ischaemia.16,17 Furthermore, a decrease in maximal force development was observed in cardiomyocytes isolated from the subendocardium but not the subepicardium.18

In addition to changes in myocardial compressive forces and structural changes in the coronary vasculature, we have studied neurohumoral and metabolic alterations in the regulation of coronary vasomotor tone after MI as reviewed by Duncker et al.19 In the remainder of this review, we will focus on novel insights into the role of endothelial dysfunction in the altered regulation of coronary blood flow in the remote remodelled myocardium after MI (Figure 1).

Alterations in endothelial function in post-myocardial infarction remodelled myocardium

A healthy endothelium is a key factor in vascular homeostasis. The endothelium not only constitutes the inner lining of the vasculature, but is also a highly active endocrine and paracrine organ that releases a variety of factors that are anti-inflammatory, anti-coagulant, and/or involved in regulation of vascular tone. Thus, the endothelium releases vasoactive factors such as nitric oxide (NO), prostacyclin, endothelin (ET), thromboxane, reactive oxygen species (ROS), and the recently discovered factor uridine adenosine tetra-phosphate (Up4A).10,20 The balance between endothelium-derived vasodilators and vasoconstrictors is continuously modulated by haemodynamic factors such as shear stress, locally secreted inflammatory mediators, as well as circulating neurohormones such as norepinephrine and angiotensin II, thereby influencing vasomotor tone. Numerous studies have shown that endothelial dysfunction, i.e. an imbalance between endothelium-derived vasodilators and constrictors contributes to the aetiology cardiovascular disease.

After MI, oxidative stress is increased even in the remote myocardium,21,22 and this increased oxidative stress contributes to endothelial dysfunction not only in isolated large coronary arteries23 but also in the coronary microvasculature.24 Superoxide generation occurs in all layers of the vascular wall.25 Because it is highly reactive, non-membrane permeable, and because anti-oxidant defence mechanisms are abundantly present, the actions of superoxide are restricted to the sub-cellular compartment in the proximity of its source of generation.26 Superoxide influences vascular tone through several mechanisms. First, superoxide can oxidatively alter various K+ channels in vascular smooth muscle cells,27,28 thereby depolarizing the membrane potential and resulting in vasoconstriction. Second, it can react with the Sarco(Endo)Plasmic Reticulum Ca2+ pump in smooth muscle cells, which depletes the SR from Ca and thereby induces vasodilation.29,30 Finally, superoxide can react with NO to form ONOO, thereby limiting NO bioavailability and decreasing NO-mediated vasodilation.31 In healthy subjects, superoxide levels are tightly controlled by superoxide dismutase (SOD).32,33 SOD catalyses the reaction of superoxide into H2O2 and O2. H2O2, a potent vasodilator,3436 is not only more stable than superoxide, but also membrane-permeable. Superoxide, derived from the respiratory chain of the mitochondria in the cardiac myocytes, when converted to the vasodilator H2O2, can diffuse to the microvasculature and is responsible for the coupling of increased myocardial metabolism to vasodilation of the coronary microvessels.36,37 Administration of an SOD mimetic had no effect on coronary vasomotor tone in awake dogs either at rest or during exercise,38 indicating optimal efficiency of endogenous SOD in converting superoxide to H2O2. However, in the porcine coronary circulation, free radical scavenging with N-2-mercaptopropionyl glycine (MPG) did have a small but significant vasodilator effect both at rest and during exercise, indicating that the O2·− that is generated in the coronary microvasculature may not be (completely) converted to H2O2, and exerts a net vasoconstrictor effect on the coronary vasculature.24 These data are in accordance with the observation that the reaction between NO and superoxide to form ONOO (thereby causing vasoconstriction) is 3–4 times faster than the reduction of superoxide by SOD (which would cause vasodilation).39 This coronary vasodilator effect of ROS-scavenging with MPG,24 as well as the dihydroethidium staining of the remote myocardium,22 was increased in swine with MI as compared with normal swine, indicating an increased production of ROS even in the microvasculature supplying the remote, non-infarcted myocardium. These data are in accordance with findings that superoxide production is increased in the remote coronary arteries of rats with MI23 as well as in monocytes/macrophages within the intima, media, and adventitia in vessels with coronary artery disease (CAD).40

The increased oxidative stress in the porcine coronary microvasculature after MI was accompanied by uncoupling of eNOS.24 eNOS uncoupling occurs through oxidation of its co-factor BH4 and results in generation of superoxide by eNOS rather than NO,41,42 thereby further contributing to oxidative stress.24 Consistent with a role of eNOS in generation of superoxide, the vasodilator effect of MPG in swine with MI disappeared after eNOS-inhibition. However, despite eNOS uncoupling, eNOS inhibition still resulted in significant, albeit reduced,24,43,44 coronary vasoconstriction at rest and during exercise, indicating eNOS-mediated NO production was still dominant over ROS production. Moreover, agonists are still capable of inducing NO-dependent coronary vasodilation, although this vasodilation was impaired in the remote coronary vasculature after MI.45 The latter data are consistent with the blunted vasodilator responses to endothelium-dependent receptor-mediated vasodilators (particularly acetylcholine) in the microcirculation of the LV myocardium in clinical studies in patients with chronic heart failure.46 A causal relation between the loss of NO-mediated vasodilation and the progression of LV dysfunction to heart failure is suggested by studies in dogs with pacing-induced dilated cardiomyopathy, in which the loss of basal NO production in the LV myocardium coincided with the progression from LV dysfunction to overt heart failure.47,48

The vasodilator effect of NO on the vasculature is principally mediated by activation of soluble guanylyl cyclase (sGC) and formation of cGMP in vascular smooth muscle.49 cGMP signalling is terminated through breakdown of cGMP by phosphodiesterase 5 (PDE5) and to a lesser extent by PDE1.50 PDE5 is predominantly present in vascular smooth muscle and skeletal muscle51,52 although there are some reports of its presence in cardiomyocytes53 and endothelial cells.54 PDE5-inhibition results in vasodilation of the large coronary arteries in both humans55 and swine.56 In healthy swine, PDE5-inhibition also causes coronary microvascular vasodilation both at rest and during exercise that was reduced (but not abolished) after inhibition of eNOS, indicating that NO is indeed an important activator of cGMP production. In contrast to these findings in swine, human studies fail to show an increase in coronary blood flow in response to PDE5-inhibition under resting conditions,55,5759 potentially because all human studies involved patients with some degree of endothelial dysfunction in the coronary microcirculation. Hence, the lower levels of NO, and consequently cGMP production, could potentially explain the lack of effect of PDE5-inhibition in the coronary microvasculature of CAD patients. Indeed, even in patients with mild CAD who showed little effect of PDE5-inhibition under basal conditions, PDE5-inhibition enhanced the, NO-mediated, vasodilation by acetylcholine of both the large epicardial coronary arteries and the coronary microcirculation.55

Both PDE5 mRNA expression and vasodilation induced by PDE5-inhibition were reduced in the coronary microvasculature of the remote non-infarcted myocardium after MI.44 The reduced PDE5 gene expression and the blunted vasoconstrictor influence of PDE5 after MI suggest that PDE5 is down-regulated to prolong the biological availability of cGMP and to maintain the vasodilator influence of NO. This implies that the amount of NO required for this vasodilator influence is reduced. Moreover, our data suggest that, although inhibition of eNOS does result in vasoconstriction in swine with MI, the main vasodilator effect of NO was not mediated by increasing cGMP, as the effect of eNOS inhibition was similar in the presence and absence of PDE5 inhibition.44 It is therefore likely that the principal vasodilator effect of NO in the remote coronary circulation of swine with MI is mediated through limiting the vasoconstrictor effects of ET.43,60

Although circulating plasma levels ET are increased after MI, the vasoconstrictor influence of endogenous as well as exogenous ET on the remote coronary vasculature is reduced after MI in vivo.8 Paradoxically, in coronary small arteries isolated from the remote myocardium of swine with MI, ET-induced contraction was enhanced,8 suggesting that there are factors in vivo that modulate ET-sensitivity of the coronary vasculature. An ET-mediated vasoconstrictor influence was unmasked in MI swine after inhibition of either eNOS or cyclooxygenase (COX),43 confirming that NO, prostanoids and ET interact in regulation of coronary vasomotor tone.8 Indeed, it has been shown that both NO61,62 and prostanoids63 are capable of inhibiting ET production and release via a cGMP-dependent pathway. Moreover, NO reduces the ET receptor binding affinity in the human vasculature.64,65 Consistent with the reduced eNOS activity, the suppression of ET-mediated coronary vasoconstriction by NO was reduced in swine with a recent MI as compared with normal swine.43

In contrast to the reduced effect of NO, the suppression of ET by prostanoids was enhanced, while the overall vasodilator effect of prostanoids was unaltered in the remote coronary vasculature after MI.43 COX-1 is constitutively expressed in the healthy coronary vasculature,66,67 while COX-2 is mainly induced by shear stress and at sites of inflammation and could therefore become more important after MI. Indeed, the contribution of prostanoids to regulation of coronary vascular tone in humans increases with the progression of CAD. Distal to angiographically minimally diseased coronary arteries inhibition of prostanoid production induces mild vasoconstriction at rest that increases during exercise,68 whereas vasoconstriction is most pronounced in patients with CAD at rest6971 and during exercise.6971 COX-1 and COX-2 end-products influence the ET system at various levels, for instance, thromboxane A2 stimulates the production of pre-pro-ET72 the precursor of the vasoactive ET, while both PGE2 and PGI2 inhibit ET production as well as secretion.73 However, further experiments specifically designed to discern alterations in the role of the individual COX end-products after MI are required to further investigate their role in the regulation of coronary vascular tone in health and disease.


In conclusion, perturbations in coronary resistance vessel tone regulation in the post-MI remodelled remote myocardium appear to be principally the result of a loss of NO bioavailability in conjunction with an increased ROS production. Together with structural changes in the coronary vascular tree and the increase in extravascular compressive forces acting on the microcirculation, these alterations in resistance vessel tone contribute to the observed decreases in coronary flow reserve and myocardial oxygen supply. Interestingly, several studies have demonstrated a reduced influence of a number of vasoconstrictors, including PDE5 and ET, which serve to blunt the abnormalities in myocardial oxygen balance in remodelled myocardium after MI.


The present study was supported by Grants from the Netherlands Heart Foundation (2000T042 to D.M. 2000T038 to D.J.D.).

Conflict of interest: none declared.


View Abstract