The increase in myocardial oxygen consumption of the left ventricle, as required during exercise or stress, is principally met by an increase in oxygen delivery and thus in CBF, as the myocardium already has a high oxygen extraction at rest (>70%). The increase in CBF can amount up to 4–5 times the resting flow in the healthy heart (Duncker and Bache, 2008), and is mainly achieved by a reduction in vascular resistance. Under normal circumstances, proximal epicardial arteries (>400 μm in diameter) serve as conduit vessels as they contribute minimally (<5%) to overall coronary vascular resistance (Goodwill et al., 2017). The major loci of coronary vascular resistance are the coronary small arteries and arterioles (20–400 μm in diameter), responding to changes in physical forces (wall stress and shear stress), as well as metabolic needs of the tissue, while less than 20% of the resistance resides in the capillaries and venules (Chilian et al., 1986; Goodwill et al., 2017). The acute and/or chronic modulation of coronary vascular resistance in response to changes in myocardial oxygen demand, involves both active (vascular tone) and passive (extravascular compressive and intravascular distending mechanical forces) changes in vessel diameter.

An optimal level of vascular diameter is thus achieved by a tightly regulated balance between a variety of vasoactive mechanisms, including metabolic, endothelial, neurohumoral and mechanical factors. The latter includes passive vascular responses to changes in the mechanical environment produced by myocardial compression and intravascular distending pressure, but also active responses contributing to vascular tone, such as myogenic response and flow-mediated response induced by changes in the perfusion pressure and shear stress. These mechanisms exert specific influences on different segments of the microvasculature (Jones et al., 1993a; Muller et al., 1996; Laughlin et al., 2012; Duncker et al., 2015) with the distal, smallest arterioles (<100 μm diameter) being most sensitive to myocyte-derived metabolic stimuli, whereas wall stress-induced myogenic mechanisms are dominant in the intermediate larger arterioles and small arteries (100–200 μm) and flow-mediated dilation dominates the vasomotor tone of small arteries (200–400 μm). The most important vasoactive mechanisms regulating vascular tone are individually described below and summarized in Figure 1.

Vasodilator reserve of the microvasculature is computed as the ratio between maximal hyperemic flow and basal flow. The maximal flow is strongly dependent on the vascular architecture, the biomechanical characteristics of the vascular segments and their number (arterial and capillary density). Biomechanical properties of the vasculature play an important role in determining the minimal and maximal vascular resistance by limiting the vasoconstrictor and vasodilator reserve of the coronary vasculature. The vasoactive responses are limited by the thickness and structure of the vascular layers, in particular the organization of the VSMC layers and the extracellular matrix. The vascular structure is controlled by local forces and it is thought that shear stress, induced by the local flow profile and the wall stress related to the blood pressure, play a major role (Jones et al., 1993a,b). Additionally, chronic alterations in vascular tone induced by vasoactive substances can also result in changes in the vascular wall structure.

Alterations in the vascular structure are termed “remodeling” of the vessel wall and have been classified (Mulvany et al., 1996) in either eutrophic, hypotrophic or hypertrophic remodeling, depending on the changes in the wall cross-sectional area (Figure 2). Eutrophic remodeling describes vascular remodeling with preserved wall cross-sectional area around a smaller or larger lumen, hypotrophic remodeling involves loss of vascular wall components, while in hypertrophic remodeling the cross-sectional area of the vascular wall increases. Furthermore, depending on the changes in lumen diameter, vascular remodeling can be categorized as inward or outward remodeling, where a decrease in vascular lumen denotes inward remodeling while the opposite is called outward remodeling (Figure 2). Vascular remodeling is thus related to the amount of tissue in the vascular wall, and its organization around the luminal diameter, affecting the biomechanical properties and limiting the distensibility and thus the luminal diameter of that specific vascular segment. However, there is evidence that even under normal circumstances, the vessel wall is not quiescent, as vessels undergo continuous turnover of wall components, thereby contributing to homeostasis (Van Den Akker et al., 2010).

Most studies assessing small artery remodeling have been performed in vitro in isolated vessels or based on histological examination (measurement of cross-sectional areas) of vascular segments fixated under pressure (Bakker et al., 2004, 2005; Van Den Akker et al., 2010; Tuna et al., 2013). While the latter approach might seem more physiologically relevant, since the vessels are not removed from their surrounding tissue, a direct comparison with the “pre-diseased” state is difficult and one-to-one comparison to similar healthy vessels from the exact same location is impossible. The in vitro approach (based on diameter lumen measurements in conjunction with optical/histological assessment of the wall/lumen ratio) has been used by many research groups, and enables the study of the vascular segment prior to and following exposure to the mechanical stimulus needed to induce the remodeling response. Passive pressure-diameter curves performed at different points in time allow for the observation of the process dynamics and a more complete characterization of the remodeling type. Additionally, the in vitro set-up is ideal to study the mechanisms underlying vascular remodeling. Conversely, removal of the vessel from the myocardium completely abolishes the natural contribution of any metabolic stimuli to this process.

Using in vitro culture of isolated small arteries, it has been shown that chronic vasoconstriction was sufficient to induce vascular inward remodeling, while vasodilation did not affect the vascular diameter or even resulted in outward remodeling (Bakker et al., 2002, 2003; Pistea et al., 2005, 2008; Sorop et al., 2006). Additionally, even in the absence of active constriction, vessels cultured at low pressures maintaining a small lumen diameter remodeled in the course of a couple of days, whereas pharmacological vasodilation significantly attenuated the inward remodeling (Sorop et al., 2006). Importantly, vessels cultured under flow remained more dilated than vessels cultured without flow, and flow inhibited the inward remodeling (Pistea et al., 2005). Bakker et al. studied the mechanisms involved in wall stress-induced vascular remodeling in more detail (Bakker et al., 2005), demonstrating tissue transglutaminase (TG2) as one of the crucial mediators of crosslinking between the extracellular matrix proteins, limiting vascular distensibility. Several factors have subsequently been shown to be involved in modulating TG2 activity, including NO, Ca²⁺ and GTP/GDP concentrations, but also the redox state in the micro-environment (Van Den Akker et al., 2010; Del Campo et al., 2013; Huelsz-Prince et al., 2013). Bakker et al., (2005) postulated that the mechanical force exerted by VSMCs during contraction, directly activates TG2, which further results in crosslinking of extracellular matrix components, explaining the link between smooth muscle activation and inward remodeling (Bakker et al., 2005). Additionally, in vitro studies also indicate that matrix metalloproteinases (MMPs) are activated in VSMCs subjected to either stationary stretch (Meng et al., 1999; Asanuma et al., 2003; Lehoux et al., 2004) or cyclic stretch (Grote et al., 2003), or in arteries exposed to longitudinal tension (Jackson et al., 2002), possibly contributing to a continuous turnover of matrix elements and counterbalancing the activation of TG2. Distal to small arteries and arterioles, and strongly dependent on the activation and remodeling status of these proximal vessels, the capillary network is involved in the oxygen delivery to the cardiomyocytes. Although its contribution to the total vascular resistance is below 10%, (Chilian et al., 1986), alterations in capillary density, especially in the subendocardium, have important consequences for local myocardial function. In order to compensate for a higher oxygen demand and lower perfusion pressures, the subendocardial capillary to fiber ratio is higher (~1) than in the subepicardial layer, (~0.8); however, these ratios have been shown to be strongly affected by different pathologies (Duncker and Bache, 2008). Loss of microvascular density, i.e., vascular rarefaction, is a key component of various pathologies. Factors contributing to capillary rarefaction include removal of angiogenic stimuli or generation of anti-angiogenic substances, flow discontinuation, disruption of endothelial-pericyte association and endothelial dysfunction (Goligorsky, 2010). It is important to realize that different disease entities, particularly metabolic syndrome, hypertension and aging, as well as the presence of a proximal obstruction, resulting in metabolic and/or hemodynamic alterations, can have a large impact on both function and structure of the coronary microvasculature, and hence on vascular resistance and myocardial perfusion. The impact of these diseases on the microvascular structure and architecture will be discussed below.

Chronic increases or decreases in shear stress and wall tension have been shown to induce vessel remodeling to maintain proper tissue perfusion. Changes in shear stress lead to proportional changes in vascular diameter to the extent that the original levels of shear stress are restored (Langille, 1996; Tuttle et al., 2001; Baeyens and Schwartz, 2016), suggesting shear stress to be a set variable controlled by endothelial-mediated dilation and remodeling. Given the important influence of mechanical forces on vascular function and structure, the understanding of mechanosensing mechanisms by which ECs and VSMCs convert physical stimuli to biological responses has been an active field of research, although data exclusively for the coronary circulation are limited and to date these mechanisms have been studied mainly in vitro. Below, we will discuss data obtained in different vascular beds, on the main molecular contributors to mechanosensing. The pathways discussed are summarized in Figure 3.

Metabolic dysregulation includes conditions such as metabolic syndrome, obesity, insulin resistance, diabetes mellitus, hypercholesterolemia and hypertriglyceridemia, that either alone or in combination contribute to coronary artery disease and stroke (Cho et al., 2018).

Clinical and experimental studies have shown that metabolic dysregulation is associated with perturbations in CBF control during increased metabolic demand (Berwick et al., 2012; Paneni et al., 2013; Crea et al., 2014; Duncker et al., 2015; Badimon et al., 2017; Van De Wouw et al., 2020). Indeed, metabolic syndrome and obesity are associated with impaired coronary flow reserve (CFR; Di Carli et al., 2003; Pirat et al., 2008; Zorach et al., 2018), which also worsens with the onset of type 2 diabetes (Kondo et al., 2001; Schindler et al., 2006). These studies are supported by observations in dogs (Setty et al., 2003) and swine with co-morbidities (Bender et al., 2016; Van De Wouw et al., 2020), demonstrating progressive impairment of myocardial oxygen delivery during graded treadmill exercise, suggesting that the mechanisms responsible include both microvascular dysfunction and remodeling. Indeed, acute hyperglycemia in young subjects impaired adenosine-mediated increase in CBF (Di Carli et al., 2003). Additionally, swine subjected to 2.5 months of hyperglycemia and hypercholesterolemia showed impaired endothelial function of isolated coronary small arteries, mediated via loss of NO, despite preserved VSMC function (Van Den Heuvel et al., 2012). Indeed, hyperglycemia and hypercholesterolemia can induce a state of oxidative stress, which is a state in which the formation of reactive oxygen species (ROS), including superoxide anion, hydroxyl anion or H₂O₂ exceeds the antioxidant defense mechanisms. Superoxide anions can directly interact with NO, thereby limiting NO bioavailability and resulting in the formation of the pro-inflammatory peroxynitrite. In the same animal model studied 15 months after induction of hyperglycemia and hypercholesterolemia (Sorop et al., 2016), increased vasoconstrictor response to ET-1 was observed, which was ET_B-mediated. Surprisingly, the endothelium-dependent vasodilation to bradykinin was no longer reduced as compared to control, although the contribution of EDH to the bradykinin-induced dilation was reduced (Sorop et al., 2016). Interestingly, at this stage, these microvascular alterations were also observed in non-diabetic, hypercholesterolemic swine. Similarly, obesity has also been shown to increase coronary microvascular sensitivity to vasoconstrictors (ET-1, prostaglandin H2 or thromboxane A₂) in animal models (Berwick et al., 2012) and humans (Barton et al., 2012; Campia et al., 2012). In addition to endothelial dysfunction, VSMC function is also affected by metabolic dysregulation, as in obese Ossabaw swine with metabolic syndrome, 16 weeks of high fat diet resulted in increased coronary vasoconstriction mediated by altered electromechanical coupling between K_V and Ca_V1.2 channels in VSMCs (Berwick et al., 2013).

A metabolic dysregulation-associated increase in sympathetic activity has also been documented both in patients and animal models, resulting in exaggerated alpha-adrenergic coronary vasoconstriction (Huggett et al., 2004; Dincer et al., 2006). Additionally, substantial evidence indicates that activation of the renin angiotensin aldosterone system (RAAS) associated with adipose tissue-derived angiotensinogen, as well as adipocyte-derived free fatty acids (FFAs) and leptin, results in microvascular dysfunction and impaired CFR, possibly due to increased angiotensin II-mediated vasoconstriction (Kachur et al., 2018). Indeed, perivascular adipose tissue-derived adipokines such as leptin, resistin, IL-6 and TNF-α are potent pro-inflammatory molecules promoting oxidative stress in the endothelium and altering endothelial function and NO bioavailability, either directly or via increased ET-1 production.

Furthermore, leptin derived from perivascular fat also promotes coronary arterial vasoconstriction and VSMC proliferation via Rho kinase signaling (Bagi et al., 2011; Noblet et al., 2016). Additionally, adipocyte-derived circulating FFA and hyperglycemia-induced advanced glycation end products (AGE) lead to increased oxidative stress, thereby limiting NO bioavailability, as well as increased production of vasoconstrictor factors such as thromboxane A2 and ET-1 (Creager et al., 2003).

The sustained vasoconstriction induced by the various mechanisms described above may contribute to structural vascular alterations. Indeed, not only changes in microvascular function have been documented in metabolic dysregulation, but microvascular structure is also affected. In Ossabaw swine, 16 weeks of metabolic syndrome induced by a high fat/high fructose diet resulted in impaired myocardial perfusion and blunted response to adenosine, associated with reduced microvascular density (Li et al., 2012). In the same animal model, 6 months of metabolic syndrome and high-fat diet resulted in impaired hyperemic flow associated with augmented coronary myogenic tone, hypertrophic inward remodeling of the coronary resistance arteries and capillary rarefaction (Trask et al., 2012).

Such alterations can be due to the deleterious effects of the metabolic dysregulation on the molecular mechanisms involved in mechanosensing. Thus, metabolic syndrome has previously been demonstrated to be detrimental to the endothelial glycocalyx, as evidenced by shedding-induced increase of glycocalyx components in the bloodstream of type 1 and type 2 diabetic patients, thereby reducing shear-induced eNOS activity and NO production (Nieuwdorp et al., 2006; Broekhuizen et al., 2010). Recent data in obese mice also indicate that breakdown of the glycocalyx induced by metabolic derangements impairs the function of the inward rectifying K⁺ channel K_ir2.1, thereby limiting the endothelial response to flow (Fancher et al., 2020). Additionally, metabolic syndrome-associated inflammation, FFAs and hyperglycemia have all been linked to activation of endothelial NADPH oxidase (NOX; Devallance et al., 2019), stimulating the production of ROS and uncoupling of eNOS. The consequent reduction in NO bioavailability not only affects vascular tone, but, over time, may also results in impaired cGMP-mediated inhibition of VSMC proliferation and hypertrophic remodeling (Tanner et al., 2000). Moreover, adipose tissue-derived pro-inflammatory adipokines potentiate VSMC proliferation, either directly via leptin (Oda et al., 2001), or indirectly via paracrine signaling through NF-κB-induced infiltration of macrophages (Zampetaki et al., 2005; Spescha et al., 2014). Leptin enhances the expression of collagen, fibronectin, transforming growth factor (TGF)-β and connective tissue growth factor (CTGF; Martinez-Martinez et al., 2014). In combination with decreased NO-mediated S-nitrosylation of TG2 (Jung et al., 2013), enhancing its ECM crosslinking activity, this elevated ECM transcription could contribute to the observed stiffening of small arteries in swine after 15 months of high fat diet in absence of vascular wall hypertrophy (Sorop et al., 2016).

Together with microvascular dysfunction and structural abnormalities of the vascular wall, alterations in vascular density may also contribute to the perturbations in CBF and oxygen delivery and even regional ischemia during exercise observed in humans (Camici and Crea, 2007; Bairey Merz et al., 2017) and animals with co-morbidities (Setty et al., 2003; Bender et al., 2016, van de Wouw et al., 2020). Indeed, in freshly explanted hearts of diabetic patients, lower capillary density and pericyte loss as compared to non-diabetic subjects were reported, accompanied by a lower angiopoietin 1/angiopoietin 2 ratio (Hinkel et al., 2017). Similarly, in the left ventricle of obese subjects, lower capillary densities were observed as compared to lean subjects (Campbell et al., 2013). These data were confirmed in different animal models, including obese Zucker rats (Toblli et al., 2004), obese Wistar–Kyoto rats with metabolic syndrome (Machado et al., 2017) and obese diabetic db/db mice (Gonzalez-Quesada et al., 2013). The pathways contributing to this capillary rarefaction were mostly associated with VEGF signaling, but inflammatory factors, certain miRNAs and even changes in ECM composition have also been shown to take part in metabolic derangement-induced capillary rarefaction (Paavonsalo et al., 2020).

Arterial hypertension, defined as a systolic blood pressure ≥ 140 mmHg and/or a diastolic blood pressure ≥ 90 mmHg (Unger et al., 2020), continues to increase in prevalence worldwide, especially in developing countries, mainly as a result of an unhealthy lifestyle. Arterial hypertension is an independent risk factor for coronary artery disease and stroke (Chobanian et al., 2003; Benjamin et al., 2019) and frequently co-exists with other aggravating pathologies, such as diabetes and chronic kidney disease, underlying their onset and progression. An important contributor to myocardial ischemia in patients with arterial hypertension is coronary microvascular dysfunction, (Modolo et al., 2015). Indeed, antihypertensive treatment has been able to improve both basal and hyperemic flow in patients (Neglia et al., 2011). In hypertensive rats, this effect was shown to be attributable to improved microvascular function and the reverse remodeling of intramural coronary arterioles (Neglia et al., 2011).

In the microvasculature, an increased blood pressure results in elevated circumferential stress (σ), which, according to the law of Laplace (σ = P × R/h, where P is transmural pressure, R is vessel radius, and h is wall thickness), can be normalized by lowering the luminal diameter or by increasing the wall thickness. These responses, aimed to normalize wall stress, are thought to be protective in the short term. However, the increase in vascular resistance and subsequent impairment in maximal dilation (reduced flow reserve) are detrimental over time. The molecular mechanisms responsible for these adaptations have been mostly studied in cell culture or in vitro experiments using isolated peripheral small arteries, while less data in the coronary vessels are available.

To orchestrate the acute reduction of the luminal diameter, VSMCs exposed to circumferential stretch undergo depolarization, leading to activation of L-type Ca²⁺ channel Ca_v1.2 and subsequent myogenic vasoconstriction. Several mechanosensitive ion channels, including TRPC6, TRPM4 and KCNQ (Welsh et al., 2002; Earley et al., 2007; Zhong et al., 2010) as well as ciliary PC1 and PC2 (Sharif-Naeini et al., 2009) contribute to this stretch-induced depolarization. Moreover, ligand-independent activation of the mechanosensitive GPCR AT_1aR has also been shown to induce myogenic vasoconstriction in response to elevated blood pressure in resistance vessels (Schleifenbaum et al., 2014). Additionally, hypertension-induced elevation of circumferential stretch results in activation of endothelial NOX and subsequent scavenging of NO by ROS (Spescha et al., 2014). In isolated ECs, it was demonstrated that this ROS production is initiated by stretch-induced activation of integrin α5β1, causing phosphorylation of JNK and p66^shc, eventually activating NOX (Spescha et al., 2014). The impairment in NO bioavailability is aggravated by hypertensive agents such as salt, ET-1, angiotensin II, renin and vasopressin (Neglia et al., 2011), which further contributes to enhanced vasoconstriction and adds to the increased vascular resistance (Lassegue and Griendling, 2004; Neglia et al., 2011).

In hypertensive animal models, medial hypertrophic remodeling, associated with VSMC proliferation, has been shown to occur in small arteries whereas eutrophic remodeling, marked by VSMC reorientation and ECM deposition, was the most common form of remodeling in the more distal arterioles (Owens et al., 1988; Stacy and Prewitt, 1989). An explanation for the different types of inward remodeling in vascular segments of different sizes is not readily found, but may be due to a difference in sensitivity to mechanical stimuli along the vasculature with flow-induced responses affecting the vascular tone mainly in proximal vessels and myogenic influences in more distal vessels. The variations in sensitivity to mechanical stimuli are accompanied by vessel size-related variations in lumen/wall (R/h) ratio, and therefore wall stress (σ). In arterioles these variables are both much smaller than in, for instance, the aorta leading to a different optimal response for a given increase in blood pressure (P) (Feihl et al., 2006). However, the responses in the proximal and distal vasculature may also be related, in that an increase in peripheral vascular resistance in patients with hypertension can result in reduced shear stress in large arteries by inducing low peak systolic blood flow velocity and a low shear rate (Khder et al., 1998), thereby resulting in impaired shear-induced NO synthesis and progression of endothelial dysfunction.

Intriguingly, the exact nature of remodeling is also related to the pathologies underlying hypertension. Studies in small subcutaneous or omental vessels isolated from non-diabetic patients with essential hypertension indicated that these vessels undergo eutrophic remodeling (Schiffrin and Deng, 1996; Intengan et al., 1999; Schiffrin et al., 2000; Park and Schiffrin, 2001), whereas hypertension associated with diabetes or renovascular disease was shown to promote media hypertrophy in small arteries (Rizzoni et al., 1996; Schofield et al., 2002; Endemann et al., 2004). It could be speculated that the presence of co-morbidities such as diabetes and renovascular disease induces systemic low-grade inflammation and enhances oxidative stress, which in turn could directly interact with the outcome of mechanosignaling pathways. Lack of NO bioavailability, due to ROS scavenging in ECs, not only impairs NO-mediated NF-κB inactivation via S-nitrosylation (Kelleher et al., 2007), but also limits cGMP-mediated inhibition of VSMC proliferation. Wall stress-induced activation of NOX in ECs may subsequently result in ROS-mediated stimulation of AP-1 (Brandes et al., 2014). A similar activation of NF-κB and AP-1 has been observed in isolated coronary VSMCs in response to stretch (Hishikawa et al., 1997). Activation of these factors in both ECs and VSMCs, potentially coinciding with stretch-mediated nuclear accumulation of the focal adhesion protein Zyxin (Cattaruzza et al., 2004; Wojtowicz et al., 2010), results in transcription of inflammatory factors, including interleukin 6 (IL-6) and monocyte chemoattractant protein 1 (MCP-1), thereby stimulating monocyte extravasation (Zampetaki et al., 2005; Spescha et al., 2014). By secreting a variety of MMPs, these monocytes facilitate extracellular matrix reorganization and VSMC reorientation, thereby potentiating structural arterial remodeling (Intengan and Schiffrin, 2001; Wenzel, 2019). IL-6 also enhances VSMC motility (Wang and Newman, 2003), and like MCP-1, directly stimulates VSMC proliferation (Ikeda et al., 1991; Viedt et al., 2002). In addition, the stretch-activated Ca²⁺ channel Piezo1 also appears to be prominently involved in hypertension-dependent arterial remodeling. Piezo1 is highly expressed in VSMCs of murine small-diameter arteries (e.g., cutaneous caudal artery and cerebral arteries) and, while it is not involved in the myogenic response, VSMC-specific knockout attenuates hypertension-induced inward remodeling and VSMC hypertrophy in the cutaneous caudal resistance artery. The exact mechanism by which Piezo1 orchestrates arterial remodeling upon activation by wall stress is not completely understood, but Ca²⁺-dependent activation of TG2 may be involved (Retailleau et al., 2015).

A consistent finding in both clinical and experimental hypertension is capillary rarefaction, either structural (anatomic absence) or functional (non-perfusion), indicating that hypertension-induced vascular remodeling is not restricted to resistance arteries (Feihl et al., 2006). Experimental data in rat models of hypertension show reduced structural myocardial capillary density in adult hypertensive animals, but interestingly, this was age-dependent, as young animals showed normal capillary densities despite significantly hypertrophied ventricles (Tomanek et al., 1982). These animal findings are in agreement with data in young and adult patients with aortic stenosis-induced cardiac hypertrophy, also showing capillary rarefaction only at adult age (Rakusan et al., 1992). Interestingly, structural capillary rarefaction may be preceded by functional capillary rarefaction, as data in non-diabetic hypertensive patients show that functional capillary rarefaction parameters (percent capillary recruitment in nailfold skin) were lower in patients with modest hypertension and correlated with endothelial dysfunction, despite normal structural capillary density (Cheng et al., 2008). Functional capillary rarefaction may thus be the result of upstream arteriolar dysfunction resulting in perturbations in capillary perfusion.

Although the pathophysiology underlying the hypertension-induced rarefaction is still not fully understood, several mechanisms have been proposed, either related to loss of existing vasculature or insufficient growth of new vasculature (angiogenesis). In mesenteric and skeletal muscle microvasculature of spontaneously hypertensive rats, for instance, it has been shown that rarefaction induced by endothelial apoptosis was attenuated by systemic application of cell permeable superoxide scavengers Tempol and Tiron, indicating the involvement of excessive ROS formation (Kobayashi et al., 2005). Moreover, it has been demonstrated that low or absent mechanical shear, downstream of inward remodeled resistance arteries, could hamper KLF2-mediated inhibition of Smad2 phosphorylation, thereby enhancing endothelial to mesenchymal transition and subsequent vascular loss (Boon et al., 2007; Lee et al., 2017). Furthermore, altered shear could also lead to rarefaction via impaired NO synthesis, causing reduced NO-dependent production of vascular endothelial growth factor (VEGF) by VSMCs, which stimulates endothelial survival (Dulak et al., 2000). Besides its role in vasodilation and endothelial survival, NO stimulates endothelial proliferation and migration, important mediators of angiogenesis. This has been clearly illustrated in eNOS-deficient mice, in which impaired angiogenesis and arteriogenesis induced by hindlimb ischemia could be restored by adenoviral expression of constitutively active eNOS (Yu et al., 2005). Similarly, coronary occlusion in mongrel dogs triggered a myocardial ischemia-induced increase in capillary density, which could be blocked by L-NAME (Matsunaga et al., 2002). This study showed that, in the absence of NO, enhanced activity of MMP-2 and -9 results in degradation of plasminogen into the antiangiogenic factor angiostatin. With respect to hypertension-induced impairment of angiogenesis, there is also evidence that reduced numbers of circulating endothelial progenitor cells (EPCs) are involved (Hill et al., 2003; Imanishi et al., 2005). Although the cause of the reduced number of EPCs in hypertension is still unclear, the finding that lowering blood pressure by angiotensin receptor blockers (Bahlmann et al., 2005) or ACE inhibitors (Pirro et al., 2007) could restore their circulating numbers, in combination with their role in vascular repair, maintenance and angiogenesis, is consistent with the concept that shortage of these cells contributes to hypertension-induced rarefaction.

Advanced age is another independent risk factor for development and progression of coronary artery disease, IHD and heart failure, contributing to morbidity and mortality worldwide (Dhingra and Vasan, 2012). Data from both patient and animal studies indicate that aging induces changes in the functional and structural properties of the vascular system (Paneni et al., 2017). Vascular ageing comprises different aspects of arterial wall injury accumulated over a long period of time due to factors such as oxidative stress, low-grade inflammation, and activation of the sympathetic nervous system, all resulting in increased vascular stiffness (Laurent and Boutouyrie, 2021). In large arteries, augmented arterial stiffness results in an increase in the velocity of the pressure wave, which together with a stiffer peripheral vasculature results in augmentation of the systolic blood pressure and increased cardiac afterload.

Arterial stiffening involves changes in extracellular components of the arterial wall due to proteolytic degradation of elastin fibers shifting the load towards stiffer collagen fibres (Duca et al., 2016). VSMCs also undergo age-dependent changes such as alterations in the activity of the contractile filaments as well as in the molecular signaling pathways regulating actin polymerization, both directly contributing to the vascular stiffening (Kajuluri et al., 2021). Furthermore, adhesion to the extracellular matrix is increased (Zhu et al., 2012).

Further data point towards aging-associated endothelial dysfunction, showing that aging induces flattening and enlargement of ECs as well as cytoskeleton alterations affecting mobility and proliferation (Yeh et al., 2000; Shi et al., 2004). In the endothelium of aging rats, changes in gap-junction distribution and connexin expression were seen, which resulted in dysfunctional and leaky vessels (Yeh et al., 2000). Moreover, reduced NO bioavailability was documented in aging animals (Csiszar et al., 2002; Leblanc et al., 2008, 2009; Kang et al., 2009, 2011), related to endothelial inflammation (Csiszar et al., 2004), similarly to data obtained in aging healthy subjects, which indicate impairments in both NO and prostanoid pathways (Singh et al., 2002). This was confirmed in the human coronary circulation, where arteriolar flow-mediated dilation was shown to evolve from prostacyclin in the young, to NO in adulthood and to H₂O₂ later in life and/or with onset of coronary artery disease (Beyer et al., 2017). Furthermore, blunted flow-mediated dilation occurs with age, as shown in both humans (Drexler et al., 1989; Zeiher et al., 1993) and animal models (Csiszar et al., 2002; Kang et al., 2009) and could be improved by antioxidant therapy (Kang et al., 2009).

Moreover, increased expression and elevated plasma levels of pro-inflammatory ET-1, as observed in ECs isolated from old vs young healthy subjects (Donato et al., 2009), were also associated with endothelial dysfunction, possibly via ROS-induced reduction in eNOS expression and activity (Wedgwood and Black, 2005). In 23 months old Fisher 344 rats, ROS production was shown to originate from NOX (Adler et al., 2003), presumably in response to age-related low-grade systemic inflammation. In addition, vascular oxidative stress can result from endothelial and VSMC mitochondrial dysfunction, the efficiency of the mitochondrial respiratory chain diminishing with age, causing electron leakage and subsequent release of ROS (Ungvari et al., 2007). Accumulating evidence furthermore suggests that Nrf2 is involved in the age-related high vascular ROS levels. Besides its role in flow-mediated signaling, Nrf2 is a redox sensitive transcription factor stimulating the transcription of a variety of genes involved in the antioxidant response. In the aged vasculature, however, there is a markedly lower expression and activation of Nrf2 (Ungvari et al., 2011a,b), possibly due to an age-related impairment of mechanosensitive activation of Nrf2 (Ungvari et al., 2019). However, this proposed aging-induced impairment of shear-mediated Nrf2 activation cannot explain why reporter studies in statically cultured VSMCs isolated from aged macaques demonstrate lower H₂O₂-induced transcriptional activity of Nrf2 when compared with the activity in cells isolated from young animals (Ungvari et al., 2011a). Regardless of what causes Nrf2 dysfunction in the aging vasculature, it limits oxidative stress resilience and stimulates NF-κB-mediated inflammation.

In addition to scavenging of NO by ROS, impaired eNOS activation has also been linked to ageing, as different studies have illustrated that shear-induced eNOS activation is blunted in the endothelium of aged animals (Sun et al., 2004; Yang et al., 2009). This could be mediated by lower availability of the eNOS substrate L-arginine and eNOS dimerizing co-factor tetrahydrobiopterin (BH4; Yang et al., 2009), but also by age-related ECM remodeling and subsequent stiffness. Aging is, as mentioned above, associated with increased ECM crosslinking, which, in combination with impaired elastin synthesis and enhanced elastin fragmentation and calcification, results in arterial stiffness. In cultured ECs, exposure to pulsatile flow leads to PI3K-mediated activation of Akt and subsequent eNOS activation in distensible, but not in stiff tubules (Peng et al., 2003). Similarly, exposure of ECs to laminar flow when cultured on hydrogels with mechanical properties resembling those of young blood vessels has been shown to induce higher NO production when compared to ECs cultured on stiff hydrogels (Kohn et al., 2015). Recent studies in human skeletal muscle biopsies have also demonstrated that a lower eNOS phosphorylation in aged tissue in response to acute passive leg movement-induced flow was accompanied by lower phosphorylation of PECAM-1, suggesting this endothelial mechanosensor may be involved in reduced shear-stress responsiveness (Gliemann et al., 2018).

Besides functional alterations, aging also promotes structural vascular changes in coronary resistance vessels (Hanna et al., 2014; Mccallinhart et al., 2018). Data from the Framingham heart study indicate that, with ageing, systolic blood pressure slowly increases due to increased peripheral vascular resistance and arterial stiffness (Franklin et al., 1997). Additionally, aging-induced blunting of beta adrenergic-mediated vasodilation, as well as activation of the RAAS system in the aged arterial wall may also contribute to the rise in systolic blood pressure in the ageing population (Jiang et al., 2008). Animal studies showed that aging results in increase in angiotensin II and aortic MMP-2 activity, inducing arterial remodeling (Wang et al., 2003, 2005). Moreover, as discussed earlier, the increase in blood pressure resulting from the increased vascular resistance induces arterial inward remodeling, promoting a further increase in blood pressure, forming a vicious circle.

Aging induces vascular alterations not only at the arterial level, but studies in dogs indicate that capillary density and length was also lower in the endocardium of old animals (Tomanek et al., 1991). This might be explained by the decreased angiogenic capacity of the aging heart as pathways central to vessel formation, such as hypoxia-inducible factor-1α (HIF-1α), PGC-1α, and eNOS, are affected by aging (Lahteenvuo and Rosenzweig, 2012), contributing to the mismatch between angiogenesis and cardiac hypertrophy in the aged heart. This impaired angiogenic response, besides the overall downregulation of angiogenic growth factors in aged tissue, seems to result from endothelial dysfunction and the associated low bioavailability of NO, which as described before, is an important mediator of angiogenesis (Matsunaga et al., 2002; Yu et al., 2005).

Intriguingly, vascular aging appears to progress differently in men and women (Dupont et al., 2019). The onset of menopause, which marks the end of a woman’s menstrual cycles, is associated with accelerated vascular aging, different from the gradual alterations in vascular function and structure as occur with chronological aging (Lakatta and Levy, 2003). The main mechanisms responsible relate to the hormonal changes inducing endothelial dysfunction, as it has been shown that reduced estrogen-mediated generation of NO likely underlies the progressive decline in endothelial function (Green et al., 2014). The reduced NO bioavailability is multifactorial, however, impaired availability of BH4 (Moreau et al., 2012), increased oxidative stress (Novella et al., 2012) and inflammation (Moreau et al., 2013) are likely contributors. Additionally, postmenopause has been associated with elevations in the vasoconstrictors ET-1 and norepinephrine (Green et al., 2014). Such imbalance between NO, ET-1 and norepinephrine contributes to the impaired vasodilation and sustained vasoconstriction observed in postmenopausal animal models (Knowlton and Lee, 2012), as well as to the increased arterial stiffness (Zaydun et al., 2006). Additionally, estrogen may not only enhance vascular relaxation via increased NO production, but may also promote angiogenesis. Estrogen receptor alpha gene knockout was associated with a decrease in VEGF levels and capillary rarefaction in the heart (Jesmin et al., 2010). In accordance with these findings, estrogen can inhibit TNFα-induced apoptosis by binding to its estrogen β-receptor, which induces Akt phosphorylation and Notch1 expression, thereby promoting vascular EC survival (Fortini et al., 2017). Altogether, these functional and structural alterations may explain, at least in part, the ischemic symptoms observed in postmenopausal women with chest pain, but no evidence of obstructive coronary artery disease.

Presence of a stenosis in an epicardial coronary artery has hemodynamic consequences for the distal vasculature. With severe stenosis, pressure and flow distal to the stenosis are compromised, resulting in impaired perfusion of the distal myocardium. An important clinical parameter used to define a stenosis as flow-limiting is the coronary pressure derived fractional flow reserve (FFR). This is the ratio between pressure distal and proximal to the stenosis during maximal coronary vasodilation, which is considered flow-limiting if below 0.8. Another parameter used for diagnostic and interventional purposes is CFR, which is defined as the maximum CBF, divided by the resting CBF and has a cut-off value of 2 (Stegehuis et al., 2018).

Due to the pressure drop across a flow-limiting stenosis, the perfusion pressure for the distal vasculature gradually decreases, lowering wall stress and thus the myogenic response due to blunted stimulation of stretch-activated mechanosensors (TRPs, KCNQ, etc.; Welsh et al., 2002; Earley et al., 2007; Zhong et al., 2010). The stenosis-induced impairment of perfusion and the resulting hypoxia may furthermore trigger red blood cell-dependent vasodilation via ATP release and subsequent activation of eNOS, S-nitrosohemoglobin-dependent bioactivity and NO synthesis via reduction of nitrite by deoxyhemoglobin (Kulandavelu et al., 2015). These autoregulatory mechanisms aim to lower vascular resistance in order to maintain flow and tissue oxygenation (Duncker et al., 2015). However, these mechanisms will be exhausted below perfusion pressures of about 40 mmHg, especially in the subendocardium, where the driving pressure is even lower due to the extra resistance of the transmural vessels and the increased extravascular compression, resulting in subendocardial ischemia.

Although microvascular dysfunction may be present already in the absence of a proximal coronary obstruction due to the presence of various comorbidities, hemodynamic changes induced by the stenosis in the distal vasculature can also directly cause alterations in both function and structure of the microvasculature, contributing to the reduced CFR in the myocardial area supplied by the stenotic artery. In a coronary stenosis model in dogs, a progressive increase in distal vascular resistance was documented in the first hours after the placement of the occluder, which was suggested to be mediated by withdrawal of the adenosine-induced vasodilation (Gorman et al., 1985). In a porcine model, 3 months after placing of an external occluder around the proximal LAD, increased vasoconstrictor response to ET-1 was present in subendocardial arterioles distal to the stenosis (Sorop et al., 2008). Vascular functional studies in chronically occluded hearts showed that endothelium-dependent relaxation in response to either bradykinin, substance P or adenosine diphosphate (ADP), was significantly impaired in arterioles isolated from the collateral-dependent region as compared with arterioles from the remote area in the same heart or compared with vessels from control, nonoccluded hearts (Sellke et al., 1992, 1996a,b; Park et al., 1996; Sodha et al., 2008). In contrast, reactivity to the endothelium-independent NO donor nitroprusside was not affected (Park et al., 1996; Griffin et al., 2001; Sodha et al., 2008). In Yucatan miniswine, 22 weeks after the placement of an ameroid occluder around the LCX, the bradykinin-mediated dilation in collateral-dependent arterioles was impaired, and could be improved by exercise-induced BK_Ca-channel activation, suggesting a possible role for H₂O₂ (Xie et al., 2013). The vasoconstrictor response to ET-1 also appeared to be significantly increased in collateral-dependent arterioles or arterioles from the stenotic area, as compared to remote arterioles, a response that was due to a loss of ET_B receptor-mediated vasodilation (Park et al., 1996; Sorop et al., 2008). Combined, these data suggest that, even in absence of comorbidities, hemodynamic changes in the microvasculature distal to a coronary stenosis or occlusion impair NO bioavailability. Paradoxically, both total levels of eNOS, as well as phosphorylation of eNOS were shown to be increased in Yucatan miniswine arterioles distal to a coronary occluder (Xie et al., 2012). It is in that light, however, important to note that uncoupling of the eNOS dimer, due to oxidation of co-factor BH4, results in eNOS-mediated production of superoxide rather than NO (Channon, 2004).

The functional alterations distal to a coronary stenosis or occlusion were associated with structural changes in vessels of different sizes. Thus, in both swine and rats, the presence of a gradual proximal occlusion resulted in inward remodeling of the resistance arteries, either hypertrophic or eutrophic (Mills et al., 1994; Sorop et al., 2008; Canty and Suzuki, 2012). Additionally, microvascular (200–500 μm diameter) and capillary rarefaction was documented both in the subepicardium and the subendocardium (Urbieta Caceres et al., 2011), all these changes contributing to a reduced CFR and an increased minimal microvascular resistance in the myocardium distal to the stenosis. These data contrast with observations in patients with a coronary artery stenosis (Verhoeff et al., 2005) in which, upon revascularization, a maintained or even slightly reduced minimal microvascular resistance was observed. Such discrepancy may have several explanations. On the one side, vascular remodeling may have been influenced by the medication of the patients. Indeed, inward remodeling of isolated arterioles at low traluminal pressure, as present distal to a significant proximal stenosis was prevented by incubation with Ca²⁺ antagonist amlodipine (Sorop et al., 2006). On the other side, such changes in vascular diameter and density (arterial or capillary) may vary in time during the progression but also after the removal of the stenosis. A recent study by Weil et al., in swine, has shown that distal to a critical stenosis subendocardial arterioles show inward remodeling with increased arteriolar wall thickness and a reduction in lumen area. Interestingly, this was compensated by an increase in arteriolar as well as capillary density. However, 1 month after revascularization, such compensatory adaptation was lost, as subendocardial arteriolar and capillary density normalized, but the arterial inward remodeling persisted (Weil et al., 2020). These findings can explain the reduced subendocardial flow reserve late after revascularization despite a near normal vasodilator reserve immediately after PCI as seen in the study by Verhoeff et al. (2005). Furthermore, the persistent coronary microvascular inward remodeling could explain the blunted vasodilator response to dobutamine observed one month following revascularization, in swine, during increased myocardial oxygen demand (Kelly et al., 2011).

Vascular remodeling at the arteriolar level may be related to alterations in the microvascular (endothelial) function, such as the increased vasoconstrictor response to ET-1 as shown in the arterioles distal to a chronic LAD stenosis (Sorop et al., 2008), but altered mechanical factors may also contribute to this response. As such, altered extravascular compression by the dysfunctional hibernating or stunned myocardium (Canty and Fallavollita, 2005; Sorop et al., 2008; Canty and Suzuki, 2012) may result in abnormal mechanosensing. In addition, the blunted myogenic response as shown in sub-endocardial arterioles distal to the stenosis (Sorop et al., 2008) in conjunction with the reduced perfusion pressure and flow may promote inward remodeling, as seen in resistance arteries cultured at 40 mmHg (Sorop et al., 2006). Similarly, reduction of blood flow in mesenteric arteries by means of artery ligation resulted in eutrophic inward remodeling four weeks after surgery (Pourageaud and De Mey, 1997). This inward remodeling could partially be explained by observations in isolated ECs, demonstrating that low wall stress results in transcriptional activation of NF-κB (Pedrigi et al., 2017), which promotes vascular remodeling via monocyte recruitment and VSMC proliferation. Activation of NF-κB, in combination with poor perfusion-associated hypoxia and subsequent activation of HIF1α, furthermore stimulates transcription of ET-1 (Yamashita et al., 2001; Bourque et al., 2011; Stow et al., 2011). Experiments in a murine endothelium-specific ET-1 overexpression model illustrate that, besides its vasoconstrictor effects, ET-1 enhances the media to lumen ratio in mesenteric resistance vessels (Amiri et al., 2004). In cultured coronary artery VSMCs (Wu et al., 2007) and mesenteric arteries (Amiri et al., 2004), ET-1 was also shown to stimulate ROS production, which is prominently involved in remodeling of resistance arteries (Pires et al., 2010; Martinez-Lemus et al., 2011). The intermediate filament Vimentin, on the other hand, seems to prevent low flow-induced medial hypertrophy, as illustrated by a substantially increased wall thickness and VSMCs displaying a proliferative synthetic phenotype in ligated carotid arteries of Vimentin knock-out mice when compared with ligated carotids of wildtype mice (Schiffers et al., 2000; Van Engeland et al., 2019).