New Insights in (Inter)Cellular Mechanisms by Heart Failure with Preserved Ejection Fraction

The endothelium, the physical barrier between the blood and vascular wall, plays a pivotal role in cardiovascular homeostasis by regulating vasomotor tone, vascular permeability, and cardiac function [8]. The release of nitric oxide (NO) is crucial for the normal function of the endothelium: NO mediates vasodilatation, inhibits platelet aggregation, and protects the integrity of the endothelial layer via its anti-inflammatory [9‐12], anti-apoptotic [13], and pro-angiogenic [14] properties.

The HFpEF-associated comorbidities lead to endothelial dysfunction, a condition characterized by impaired endothelium-dependent vasodilatation and “endothelial activation,” referring to a state in which the endothelium loses its physiological properties and shifts towards a pro-inflammatory, pro-coagulatory, and vasoconstrictor state [15]. Particularly hyperglycemia, activation of the renin-angiotensin system, and tumor necrosis factor alpha (TNF-α) underlie decreased NO bioavailability, the main characteristic of endothelial dysfunction, involving (1) deprivation of the substrate L-arginine via impaired recycling of L-citrulline [16]; (2) increased expression/activity of the natural competitor of endothelial NO synthase (eNOS), arginase [17]; (3) reduced expression of eNOS [12, 18]; (4) a decreased ratio of the eNOS dimer to monomer [19]; (5) deprivation of the co-factor tetrabiopterin [20]; and/or (6) eNOS uncoupling, a phenomena by which eNOS produces superoxide (O₂•) rather than NO [19, 21]. Hyperglycemia- or dyslipidemia-induced nicotinamide adenine dinucleotide phosphate (NAD(P)H) oxidase (NOX) [22] leads to O₂• production and subsequent formation of peroxynitrite (ONOO–), which enhances eNOS uncoupling via reducing the affinity of eNOS for tetrabiopterin and via limiting the de novo synthesis of tetrabiopterin [23].

The cardiac endothelium comprises the endothelial cells of the coronary microvasculature, of the endocardium, and of the intramyocardial capillaries. Cardiomyocytes lie within the coronary microvascular network maximum 3 μm from endothelial cells. This specific anatomical arrangement allows not only adequate blood supply but also facilitates the bidirectional communication among those cells [24‐26]. The relevance of the cardiac endothelium in cardiac function follows from pioneer work from Brutsaert et al. [27] who demonstrated in vitro that the cardiac endocardial surface modulates the performance of cardiac muscle. Via coronary infusion of substance P leading to acute modulation of left ventricular (LV) function, Paulus et al. [28] provided later evidence that the cardiac endothelium is important for cardiac contractile function in humans potentially via the release of paracrine factors including NO, endothelin-1, natriuretic peptides, cytokines, and others. Both studies accentuate the significance of the cardiac endothelium for (acute) cardiomyocyte function. How the cardiac endothelium and consequently endothelial dysfunction or deterioration affects cardiomyocytes as well as cardiac fibroblasts and cardiac neurohumoral activation will further be addressed in the following paragraphs. The impact of the cardiac endothelium on those different cardiac cells along with complex (bidirectional) interactions accentuates the relevance of endothelial dysfunction [29, 30] in the pathogenesis of HFpEF.

The endocardial endothelial cells and the endothelial cells of intramyocardial capillaries regulate the contractile state of cardiomyocytes via autocrine and paracrine signaling involving NO and endothelin-1. Furthermore, co-culture of cardiac endothelial cells with cardiomyocytes protects those cardiomyocytes from hydrogen peroxide-induced apoptosis through neuregulin-erbB4 signaling [55]. This indicates that cardiac endothelial cells exert direct anti-apoptotic effects in addition to their role in oxygen supply, which is needed for cardiomyocyte survival. Cardiac endothelial cells also guide cardiomyocyte organization and promote physiological coupling of cardiomyocytes and their synchronized contraction via influencing the expression of the principal gap junction protein connexin 43 [56]. On the other hand, upon inflammation [57] and mechanical load [58, 59], endothelial cells may contribute to cardiomyocyte hypertrophy via the release of endothelin-1. Cardiomyocytes in turn can affect the coronary vasculature via multiple paracrine signals including endothelin-1 and fibroblast growth factor 2 [60, 61]. In addition, cardiomyocytes affect long-term growth and development of coronary arterial, venous, and lymphatic trees in which the paracrine release of VEGF-A by cardiomyocytes is of particular importance [60, 62]. This cardiomyocyte-vascular crosstalk plays a critical role in the vascular adaptation, which takes place during myocardial hypertrophy. An imbalance between vasculature and cardiomyocyte growth may lead to progressive cardiac dysfunction and heart failure [60]. In brief, the above findings indicate the delicate balance between endothelial cells and cardiomyocytes and imply that changes due to endothelial dysfunction or deterioration may contribute to cardiomyocyte hypertrophy and HFpEF.

Particularly, the contribution of cardiomyocyte stiffness in diastolic LV stiffness and HFpEF is established. Besides intracellular calcium dysregulation, involving dysfunction of the sarcoplasmic reticular adenosine triphosphate (ATP)-driven pump (SERCA), phospholamban (PLB), and/or ryanodine receptor (RYR) 2, cardiomyocyte stiffness is mainly regulated by the giant sarcomeric protein titin [63]. Titin spans the sarcomere from the Z disk to the M line and functions as a molecular spring supporting early diastolic recoil and late diastolic distensibility of cardiomyocytes [64]. Evidence that solely titin stiffness is sufficient to induce diastolic dysfunction and HFpEF, independent of extensive collagen deposition, follows from recent experimental studies. Chung et al. generated mice with a deletion of nine immunoglobulin-like domains from the proximal tandem immunoglobulin segment of the titin spring region, which resulted in overall titin stiffness [6•]. These knockout mice developed HFpEF despite unaltered myocardial collagen content. In addition, Hamdani et al. [65] recently demonstrated the importance of titin stiffness in the induction of diastolic dysfunction independently of cardiac fibrosis in a rat model of the metabolic syndrome. Besides shortage of titin length by experimental deletion of immunoglobulin-like domains [6•] or by oxidative stress-induced formation of disulfide bridges within the titin molecule [66], titin stiffness is also attributed to isoform shifts or posttranscriptional modifications like phosphorylation or oxidation [64]. Protein kinase Cα phosphorylates titin at its proline, glutamate, valine, and lysine (PEVK) titin region and raises the stiffness of cardiomyoctes from normal myocardium [67], but not from cardiomyocytes isolated from an animal model of HFpEF [68]. On the other hand, the protein kinase extracellular signal-regulated kinase 2 [69] and Ca²⁺/calmodulin-dependent protein kinase II [70] phosphorylate titin and lower cardiomyocyte stiffness, though their pathophysiological relevance for HFpEF needs to be further clarified. In HFpEF patients, titin at the stiff N2B isoform is hypophosphorylated [71]. Protein kinases A [72] and G [73, 74••] phosphorylate titin at its N2B segment, and lower cardiomyocyte stiffness, which implies that ß adrenergic stimulation, NO, and natriuretic peptides may decrease cardiomyocyte stiffness. With respect to Ca²⁺/calmodulin-dependent protein kinase II and its involvement in NO synthesis as a result of Ca²⁺-dependent activation of eNOS [75], it is tempting to speculate that Ca²⁺/calmodulin-dependent protein kinase II may further contribute to the relaxation of titin via triggering NO release. Due to the potential risk of arrhythmic death, the induction of protein kinase A via ß adrenergic stimulation is excluded as therapeutical option. Blocking the breakdown of the downstream NO target cGMP via sildenafil has not been successful in HFpEF patients despite promising experimental studies (see infra). In contrast, favorable effects in improving diastolic dysfunction in patients with HFpEF have recently been found by the use of the neprilysin inhibitor (LCZ696), which inhibits the breakdown of natriuretic peptides. Neprilysin activity is induced in obese patients [76]. The finding that IL-1ß induces neprilysin activity [77] may explain the increased neprilysin and B-type natriuretic peptide (BNP) paradox in obese HFpEF patients, which is associated with increased inflammation [78].

In summary, the importance of NO bioavailability and oxidative stress in titin relaxation (of the N2B segment) [74••] as well as the role of inflammation in the induction of neprilysin activity [77] and subsequent impaired natriuretic signaling further corroborates the impact of endothelial dysfunction on cardiomyocyte stiffness.

As it has already been broadly outlined before, HFpEF comorbidities provoke a systemic inflammatory state, which severely affects the coronary microvascular endothelium [36, 79]. The sum outcome of microvascular deterioration is neuronal dysfunction and death as well as the reduction of neurovascular perfusion, functional impairment, and cellular apoptosis [80]. Inflammatory and hyperglycemia-induced oxidative stress in vascular endothelial cells correlates with cardiac autonomic nerve dysfunction [81]. eNOS uncoupling leading to nitrosative stress and formation of peroxynitrite induces damage of neuronal axons, which are particularly prone to oxidative and nitrosative stress due to their high mitochondrial content [82]. This leads to a prominent abnormality in the levels of cardioactive neuropeptides including neuropeptide Y, substance P [83], calcitonin-generated peptide, natriuretic peptides (ANP, BNP, and C-type natriuretic peptide) [84] as well as the respective neuropeptide effectors [85], contributing to impaired endothelium-dependent vasodilation in HFpEF [86]. This limits the delivery and extraction of tissue oxygen, which subsequently results in nerve hypoxia and impaired nerve function [80]. The damage of the nerve fibers innervating the heart and blood vessels, common for cardiovascular autonomic neuropathy, results in abnormalities in heart rate control and vascular dynamics. Besides their role as vasodilatory proteins, the natriuretic peptides ANP and BNP exert anti-fibrotic [87] and anti-hypertrophic effects [88] and influence titin stiffness. Under metabolic conditions, BNP degradation seems to be paradoxically increased [89], contributing to cardiac fibrosis and cardiomyocyte stiffness.

Cardiovascular autonomic neuropathy in diabetes mellitus is associated with LV diastolic and systolic dysfunction, LV hypertrophy, and higher LV mass index [90]. Findings from studies of the autonomic nervous system indicate that it can modulate inflammatory reactions [91]. These findings suggest that dysregulation of the cardiac neurohumoral system not only directly modulates vascular dynamics and heart function but also activates cardiac inflammation and subsequent remodeling.