Myocardial oedema: pathophysiological basis and implications for the failing heart

Abstract

Myocardial fluid homeostasis relies on a complex interplay between microvascular filtration, interstitial hydration, cardiomyocyte water uptake and lymphatic removal. Dysregulation of one or more of these mechanisms may result in myocardial oedema. Interstitial and intracellular fluid accumulation disrupts myocardial architecture, intercellular communication, and metabolic pathways, decreasing contractility and increasing myocardial stiffness. The widespread use of cardiac magnetic resonance enabled the identification of myocardial oedema as a clinically relevant imaging finding with prognostic implications in several types of heart failure. Furthermore, growing experimental evidence has contributed to a better understanding of the physical and molecular interactions in the microvascular barrier, myocardial interstitium and lymphatics and how they might be disrupted in heart failure. In this review, we summarize current knowledge on the factors controlling myocardial water balance in the healthy and failing heart and pinpoint the new potential therapeutic avenues.

Introduction

The adequate compartmentalization of water in the myocardium is essential to maintain normal cardiac function. ¹ Despite several mechanisms known to regulate cardiomyocyte and interstitial volume, ² the myocardium remains particularly susceptible to oedema formation due to its dense microvascular network and high interstitial flow rate.

Myocardial oedema (MO), defined by the accumulation of cardiac water in interstitial and/or intracellular compartments, has been shown to induce cardiomyocyte injury, dysfunction ³ , ⁴ , ⁵ , ⁶ and remodelling. ³ , ⁴

The recent introduction of magnetic resonance imaging (MRI) techniques (e.g. myocardial T1 and T2 mapping) has enabled the non‐invasive assessment of the extracellular component, namely, the myocardial water content, suggesting that MO negatively affects the prognosis across acute and chronic heart failure (HF). ⁷ , ⁸ , ⁹ Moreover, advances in the understanding of the myocardial microvascular barrier and lymphatics suggest that myocardial fluid balance disturbances are key determinants of the extent and duration of myocardial injury. These aspects may recast MO as a therapeutic target yet to explore in clinical practice.

The present review aims to summarize the current knowledge on the pathophysiological mechanisms of MO formation and their contribution to the disruption of cardiac homeostasis in the failing heart, also discussing future perspectives on therapeutic targeting of MO.

Basic concepts

Myocardial fluid balance and myocardial oedema

To maintain fluid homeostasis, microvascular fluid filtration into the myocardium must be matched by its removal rate via myocardial lymphatic vessels. Microvascular fluid exchange is governed by the Starling principle, expertly reviewed elsewhere, ¹ , ¹⁰ summarized by the revised Starling equation:

$$J_{\mathrm{V}}=L_{\mathrm{P}}S\left[\left(P_{\mathrm{C}}-P_{\mathrm{I}}\right)-\sigma \left({\mathrm{\Pi }}_{\mathrm{C}}-{\mathrm{\Pi }}_{\mathrm{G}}\right)\right]$$

where L _P is the hydraulic conductivity, S is the filtration surface area, P _C and P _I are the intracapillary (C) and interstitial (I) hydrostatic pressures, σ is the protein reflection coefficient and Π_C and Π_G are the intracapillary and subglycocalyx (G) colloid osmotic pressures, respectively (Figure 1 ). In order to keep a stable interstitial volume (V _I) and defend against oedema formation, several physiological mechanisms counteract primary perturbations in P _C, Π_C and endothelial barrier function—oedema safety factors. ²

Figure 1.

The disruption of myocardial fluid balance in the failing heart. Multiple mechanisms can contribute for oedema formation in the failing heart and are differentially observed in several types of acute and chronic heart failure. Myocardial ischaemia, inflammation and volume overload negatively impact on microvascular barrier function by promoting the glycocalyx degradation and pericyte detachment, resulting in excessive fluid filtration. The resulting increase in interstitial volume and pressure disrupt the extracellular matrix (ECM) architecture, pulling cardiomyocytes away from capillaries and increasing oxygen diffusion distance. Moreover, ECM degradation and high central venous pressure impair lymphatic recruitment and drainage, leading to the accumulation of inflammatory cells, cytokines and metabolic waste products in the myocardial interstitium. Collectively, these mechanisms can impair myocardial contractility and bioenergetics, increase myocardial stiffness and promote cardiomyocyte apoptosis.

Myocardial oedema develops when fluid filtration rate exceeds lymphatic fluid removal and can be generated by increased ΔP or S, decreased ΔΠ or alterations of microvascular membrane properties (increased L _P or decreased σ) (Figure 1 ). Increased P _C can be driven by high pre‐capillary pressure in the setting of acute ¹¹ and chronic ¹² arterial hypertension, or high post‐capillary pressure in coronary sinus occlusion, ⁴ , ¹³ , ¹⁴ pulmonary hypertension ¹⁵ or in acute HF with increased central venous pressure. ¹⁶ Moreover, increased S, caused by increased capillary recruitment or vasodilation, promotes MO formation and is particularly relevant in inflammatory HF aetiologies (e.g. myocarditis and sepsis). Finally, as albumin is the major determinant of Π_C, states of hypoalbuminemia facilitate fluid filtration and global interstitial oedema. ⁶ This is particularly relevant in crystalloid coronary perfusion during cardiac surgery ¹⁷ and in shock management (e.g. septic and cardiogenic) in which, excessive fluid resuscitation worsens prognosis. ¹⁸ , ¹⁹

Myocardial oedema dramatically reduces energetic efficiency, impairing both contraction and relaxation. ¹ , ¹⁹ , ²⁰ However, increased V _I and P _I have been shown to primarily affect myocardial viscoelastic properties, resulting in higher diastolic stiffness. ⁶ , ²² Due to its low interstitial compliance, small interstitial volume expansions create high interstitial pressures, making the myocardium particularly sensitive to oedema formation. The experimental increase in myocardial water content by 3.5% was associated with a 40% drop in cardiac output. ³ In addition, MO directly opposes filtration, by decreasing ΔP and physically compressing the capillaries and disrupting nutrient and oxygen delivery. ²³ The disruption of the extracellular matrix structure, increased oxygen diffusion distance and accumulation of metabolic waste products are additional proposed mechanisms of MO‐associated functional deterioration (Figure 2 ). ¹ , ²⁴

Figure 2.

Pathophysiological pathways contributing to myocardial oedema in heart failure.

In summary, myocardial fluid balance is largely dependent on microcirculation dynamics, microvascular barrier, interstitial architecture and lymphatic drainage. Disruption of any of these components may disturb myocardial fluid homeostasis. In this review, each factor will be addressed in detail regarding its physiological role and how it may be disrupted in the failing heart.

Coronary microcirculation

The healthy myocardium is one of the most densely vascularized tissues in human body, possessing a high density capillary network (3.000–4.000/mm²) closely disposed around cardiomyocytes ²⁵ , ²⁶ (Figure 2 ). Such proximity between cardiomyocytes and capillaries is of utmost importance to maintain a short diffusion distance not only for oxygen, but also for potentially toxic byproducts of cellular metabolism. ²³ Moreover, the high metabolic rate of the myocardium, which primarily depends on oxidative phosphorylation, translates in an elevated oxygen demand that is matched by a very high oxygen extraction rate (70%–80% in resting conditions). ²⁷ , ²⁸ , ²⁹ Consequently, in stress conditions, additional increments in oxygen demand are predominantly met by parallel increases in myocardial blood flow (MBF). ²⁷ This metabolic contribution to MBF autoregulation is made possible by the close contact between cardiac muscle and vasculature, enabling cardiomyocyte‐derived mediators (CO₂ ³⁰ and lactate ³¹ ) and microenvironmental factors (pH ³² and extracellular K^{+ 33} ) to modulate local vasomotor tone and haemoglobin dissociation curve. Therefore, pathological conditions limiting the close communication between cardiomyocytes and vasculature, namely, the expansion of the interstitial space due to oedema or fibrosis, as well as arteriolar and capillary rarefaction, common features in chronic HF, ³⁴ impair diffusional transport and MBF autoregulation, contributing to oxygen supply/demand mismatch (Figure 2 ).

Another consequence of the proximity between coronary microvasculature and cardiomyocytes is their mechanical interaction. Previously considered an important modulator of contractility (i.e. Gregg phenomenon), the effect of coronary perfusion was later shown to be negligible within the autoregulatory pressure‐flow range. ³⁵ , ³⁶ Extravascular forces are not uniform across the ventricular wall and a gradual increase in interstitial pressure and vascular compression is observed from the subepicardium to the subendocardium. ³⁷ , ³⁸ This is partly compensated by a higher arteriolar density at the subendocardium so that, in physiological conditions, MBF is similar in both myocardial layers. ³⁹ , ⁴⁰ Yet, the distinct mechanical cross‐talk between different myocardial layers, makes arterio‐venous pressure gradient (i.e. perfusion pressure) at the subendocardium about half of that of subepicardium. ⁴¹ Consequently, in the setting of decreased coronary pressure (e.g. coronary artery disease), the subendocardial perfusion is predominantly affected. ⁴² , ⁴³ , ⁴⁴ This intricate relation between microcirculation and perfusion may underly, at least partially, the existence of clearly distinct patterns of MO distribution associated with different kinds of myocardial injury: in acute inflammatory conditions (e.g. viral myocarditis, sepsis) oedema is generally evident in the subepicardial layers whereas in acute ischaemia, the oedema is transmural or predominantly affects the subendocardium. ⁷ , ⁴⁵ , ⁴⁶ , ⁴⁷

Coronary vasculature also influences myocardial tissue properties. Higher coronary perfusion pressure is associated with increased myocardial stiffness, shifting diastolic pressure‐volume relationship left and upwards, even in the absence of oedema formation. ⁴⁸ , ⁴⁹ , ⁵⁰ The underlying mechanism resides in the fact that cardiomyocyte contraction increases the cell diameter, which happens at the expense of coronary vascular diameter, contributing to the abovementioned systolic vascular compression. ³⁶ , ⁵¹ Accordingly, higher intravascular volume and pressure, caused by increased coronary perfusion or venous outflow pressure, oppose intravascular fluid displacement, and therefore impair muscle contraction and relaxation. This has been shown to be especially relevant in the setting of increased coronary sinus pressure, seen in acute and chronic HF, where increased intravascular and interstitial volume act cooperatively to impair systolic function and diastolic compliance. ⁵² , ⁵³ , ⁵⁴

Coronary microvascular barrier

Overlooked in the past, the cardiac microvascular barrier became increasingly recognized as highly active and complex structure, composed of a continuous non‐fenestrated endothelial cell monolayer, which is internally coated with a negatively charged gel‐like mesh (i.e. glycocalyx) and externally covered by pericytes and basement membrane (Figure 3 ).

Figure 3.

Molecular interactions in myocardial fluid balance. The myocardium is composed by cardiomyocytes, microvascular capillaries enclosed by pericytes and lymphatic capillaries: fluid is filtrated in microvascular capillaries, through the endothelial surface layer and interendothelial junctions. In the myocardial interstitium, fluid entry is limited by type I and type III collagen fibres and GAGs, extracellular matrix components that act as a buffer for Na⁺ and water. Interstitial and intracellular water are in delicate balance, maintained by cardiomyocyte volume regulators. Interstitial fluid (IF) and solutes are collected by initial lymphatic capillaries, enabling a continuous IF renovation, which is returned ultimately to the venous circulation. (A). Cardiomyocyte ionic transporters: cardiomyocytes closely regulate intracellular water entry and extrusion. Water enters through aquaporins or passively diffuses through the cell membrane, according to osmotic gradients established by ionic and solute concentrations. (B). Endothelial cell–pericyte interaction: these cells establish close paracrine and physical (N‐cadherin) interactions regulating microvascular stability. Endothelial cells secrete PDGF‐BB that binds to PDGFR‐β, promoting pericyte recruitment and microvascular integrity, whereas pericytes secrete angiopoietin 1 (Ang‐1), which acts on Tie‐2 and stabilizes endothelial cells. (C). Endothelial surface layer and interendothelial junction: the endothelial surface layer is composed by endoluminal glycocalyx, which binds plasma proteins and protects endothelial cells. Furthermore, endothelial cells establish varied connections, maintaining cohesiveness and cell survival. (D). Lymph drainage in initial lymphatic capillary: fluid enters the lymphatic vasculature via lymphatic capillaries, which are blunt‐ended vessels attached to the extracellular matrix by anchoring filaments. Lymphatic endothelial cells overlap, creating valve‐like structures that promote unidirectional lymph flow. These vessels converge progressively from the subendocardium to the subepicardium, forming epicardial lymphatic collectors. ALK‐1 and ‐5, anaplastic lymphoma kinase‐1 and 5; Ang‐1, angiopoietin‐1; AngII, angiotensin II; Aqp, aquaporins; GAG, glycosaminoglycans; HA, hyaluronic acid; JAMs, junctional adhesion molecules; NBS, Na⁺/HCO₃ ⁻ Symporter; NCX, Na⁺/Ca²⁺ exchanger; NHE, Na⁺/H⁺ exchanger; PDGF‐BB, platelet‐derived growth factor BB; PDGFR‐β, PDGF receptor β; TGF‐β, transforming growth factor β; TGFR‐β2, TGF receptor β2; Tie‐2, angiopoietin‐1 receptor.

Interendothelial junctions

Endothelial cells (EC) are tightly bonded by interendothelial junctions (IEJ), mostly comprised by tight (occludins, claudins and JAMs) and adherens junctions (VE‐cadherin), which define endothelial pore size and can be dynamically regulated at the expression level and through internalization, to finely tune endothelial permeability and regulate the passage of macromolecules and cells ⁵⁵ , ⁵⁶ , ⁵⁷ (Figure 3 ). Accumulating evidence suggests IEJ disruption as a potential pathophysiological mechanism in cardiac diseases. Importantly, endothelial expression of claudin‐5, a critical player in size‐selective barrier function, is reduced in human end‐stage HF hearts. ⁵⁸ This was also shown in experimental diastolic dysfunction induced by western diet, where claudin‐5 and occludin down‐regulation was associated with increased vascular permeability, ⁵⁹ an effect attenuated by amiloride, suggesting an important role for endothelial ENaC expression and sodium overload. Regarding adherens junctions, reduced VE‐cadherin/β‐catenin expression in dilated cardiomyopathy was associated with endothelial cell degeneration, ⁶⁰ whereas in post‐ischaemic MO, Src inhibition prevented VEGF‐mediated disruption of Flk/VE‐cadherin/β‐catenin complex and attenuated post‐ischaemic MO, fibrosis and mortality. ⁶¹ In addition, key risk factors for HF development and progression have been shown experimentally to promote endothelial hyperpermeability by disrupting EIJ, namely, renin‐angiotensin‐aldosterone system activation, ⁶² inflammation, ⁶³ , ⁶⁴ hypoxia, ⁶⁵ cardioplegic arrest, ⁶⁶ hyperglycaemia, ⁶⁷ oxidative stress, ⁶⁸ increased circulating LDL ⁶⁹ , ⁷⁰ and free fatty acid ⁷¹ levels.

The endothelial surface layer

The endothelial glycocalyx (eGC) covers the apical side of endothelial cells and consists of a complex meshwork of varied membrane‐associated macromolecules ⁷² , ⁷³ , ⁷⁴ (Figure 3 ). These include proteoglycans and glycoproteins, forming a backbone in which soluble proteins, plasma‐ or endothelial‐derived, are incorporated. eGC proteoglycans are constituted by linear core proteins, mostly Syndecan‐1, to which multiple glycosaminoglycans (GAGs) side chains can be covalently attached. GAGs are highly polyanionic compounds composed of disaccharide repeating units which can be non‐sulfated [hyaluronic acid (HA)] or sulfated (chondroitin sulfate, dermatan sulfate, keratan sulfate and heparan sulfate). Together, they form a negatively charged surface that will enable electrostatic interactions with plasma cations, mostly with divalent metal cations (e.g. Ca²⁺), but also with Na⁺ due to its high plasma concentration. ⁷⁵ , ⁷⁶ The resulting high cation concentration at the interface with the plasma enables negatively charged circulating proteins (albumin, antithrombin III and thrombomodulin), that would otherwise not be able to electrically interact with the glycocalyx, to approach and incorporate this layer, forming together the endothelial cell surface layer (ESL). ⁷⁷ , ⁷⁸ The ESL, measuring between 0.2 and 2.0 mm in vivo, is therefore a highly complex structure with critical functions in microvascular physiology by (i) physically shielding the underlying endothelium from luminal aggressions; (ii) regulating microvascular flow by transmitting shear‐stress forces; (iii) constituting a barrier for plasma proteins and ions, thereby maintaining intravascular oncotic pressure; (iv) avoiding platelet aggregation by accumulating platelet‐inhibitory factors (antithrombin III and thrombomodulin) and physically restricting its interaction with subendothelium at the endothelial gaps; (v) inhibiting endothelial proinflammatory activation (i.e. increased permeability and adhesiveness) by binding circulating cytokines; and (vi) limiting the access and adhesion of circulating immune cells to the EC surface. ⁷⁹ , ⁸⁰ , ⁸¹

The ESL structure is maintained by a fragile balance between flow and enzymatically mediated shedding, and de novo production of its components. ⁸² Not surprisingly, most pathological mechanisms shown to increase microvascular barrier permeability act concomitantly on IEJ and ESL, namely, in inflammation, ischaemia–reperfusion, ⁸³ hypoxia ⁸⁴ and hyperglycaemia. ⁸⁵ Importantly, the activity of glycocalyx‐degrading enzymes (i.e. hyaluronidases, heparanase and MMPs) is increased in the setting of inflammation, which, in combination with endothelial CAM overexpression, facilitates leukocyte adhesion and diapedesis. ⁸⁶ The importance of the permissive effect of ESL degradation on cardiac leukocyte infiltration has been shown in myocardial infarction, ⁸⁷ , ⁸⁸ viral myocarditis ⁸⁹ and sepsis, ⁹⁰ , ⁹¹ aggravating the myocardial inflammatory injury. Moreover, degradation of eGC components (hyaluronan ⁹² and heparan sulfate ⁹³ ) has been shown to promote MO by increasing microvascular permeability to water and proteins.

Perhaps, the more striking association between eGC and HF is the fact natriuretic peptides (NP), mostly produced by cardiomyocyte stretching in the setting of hypervolemia and ventricular overload, have been repeatedly shown to promote eGC degradation. ⁹⁴ , ⁹⁵ , ⁹⁶ , ⁹⁷ , ⁹⁸ This effect seems to act concurrently with Na⁺ overload, which also leads to the destabilization and collapse of the eGC, mainly through loss of heparan sulfate residues, an effect attenuated by the use of spironolactone. ⁹⁹ This can be interpreted essentially as a compensatory mechanism, by enabling the escape of excessive intravascular fluid and sodium to the interstitium, which has a high Na⁺ buffering capacity due to its GAG content, ¹⁰⁰ and acting in conjunction with NP‐mediated venodilation to reduce cardiac overload. However, eGC degradation in the setting of myocardial functional impairment might also carry some drawbacks. In addition to eGC degradation being an inherently proinflammatory stimuli for EC, ¹⁰¹ , ¹⁰² , ¹⁰³ the impairment of glycocalyx Na⁺ buffering capacity may increase the amount of Na⁺ presented to the endothelium, promoting intracellular endothelial Na⁺ overload and increased transport to the interstitium, resulting in endothelial dysfunction and aggravated interstitial oedema, respectively. ¹⁰⁴ , ¹⁰⁵ , ¹⁰⁶ Furthermore, this combined effect of hypervolemia and Na⁺ overload also has important implications in the critical care setting (e.g. cardiogenic and septic shock), ¹⁰⁷ , ¹⁰⁸ where the frequently excessive crystalloid resuscitation might disrupt microvascular barrier function and complicate haemodynamic management and prognosis. Despite its proposed pathophysiological importance, a direct observation of ESL disruption in HF is still lacking.

Cardiac pericytes

Cardiac pericytes (CPc) are a highly heterogeneous population of perivascular contractile cells that ensheath and intimately interact with underlying endothelial cells, forming a microvascular syncytium. ¹⁰⁹ , ¹¹⁰ Despite conflicting reports, recent data suggest that CPc cover up to 99% of the length of the myocardial microvasculature. ¹¹¹ CPc share the basement membrane with EC and establish numerous physical interactions, ensuring an adequate control of microvascular permeability. Moreover, an intense reciprocal communication between CPc and EC takes place through gap junctions and paracrine factors, which has been shown to be especially relevant for angiogenesis and stabilization of newly formed vessels ¹⁰¹ (Figure 3 ). Importantly, multiple pericyte phenotypes with distinct cell‐surface marker signatures and variable expression of contractile proteins have been shown to be differentially distributed across the arteriolar, microvascular and venular sections of coronary vasculature. ⁹⁹ , ¹¹² Such diversity probably underlies distinct pathological roles attributed to pericytes in the context of myocardial injury and remodeling.

Extensive evidence supports a key role for CPc in the regulation of myocardial microvascular flow and permeability. Indeed, the disruption of key trophic and homeostatic pathways for CPc, namely, PDGF‐BB/PDGFR‐β, ¹¹³ , ¹¹⁴ Ang‐1/Tie2, ¹¹⁵ , ¹¹⁶ Sirtuin‐3 ¹¹⁷ , ¹¹⁸ and Notch3, ¹¹⁹ , ¹²⁰ has been shown to decrease CPc density and EC coverage, resulting in increased microvascular permeability in response to injury, MO and functional impairment. Importantly, common observations in genetic and drug‐induced CPc dysfunction are increased microvascular tortuosity and decreased coronary reserve in response to vasodilator challenge, with cardiac up‐regulation of hypoxia‐related genes. ¹²¹ In knockout mouse models, the genetic ablation of Notch3 ¹²² and Sirtuin‐3 ¹²³ impairs microvascular maturation and pericyte/EC interaction, exacerbating ischaemic injury and hindering post‐ischaemic functional recovery. Similar observations were made in experimental models of endotoxemia and diet‐induced obesity, in which Sirtuin‐3 has been shown to be down‐regulated. ¹²⁴ , ¹²⁵ Accordingly, in the setting of ischaemic injury, cardiomyocyte‐derived proNGF activates p75 neurotrophin receptor, causing pericyte process retraction, resulting in a lack of support of the microvascular endothelium and perivascular oedema. ¹²⁶ Moreover, Hypoxia‐Induced Endoplasmic Reticulum Stress Regulating (HypER) lncRNA, which promotes pericyte proliferation, viability and interactions with EC, is down‐regulated in human HF, ¹²⁷ supporting pericyte degeneration as a potentially important pathophysiological mechanism.

In line with the diversity of CPc phenotypes and functional roles in the setting of myocardial ischemia, CPc have also been implicated in the no‐reflow phenomenon. ¹²⁸ Importantly, some pericyte subpopulations express variable amounts of myosin and actin isoforms (α‐SMA and γ‐actin), having the ability to contract and relax in response to multiple paracrine factors (catecholamines and adenosine). ¹²⁹ , ¹³⁰ Being circumferentially disposed around capillaries, CPc contraction can decrease microvascular flow and theoretically reduce capillary luminal diameter enough to impede the passage of leukocytes. Indeed, in an ischaemia/reperfusion injury model, post‐ischaemic capillary blockage sites have been shown to be disproportionally close to pericytes, suggesting ischemic CPc contraction, probably mediated by an increase in intracellular Ca^{2+, 131} as an important mediator of impaired reoxygenation of ischemic tissue following myocardial revascularization. ¹³²

In inflammatory conditions, CPc detachment from EC surface was associated with differentiation into myofibroblasts and increased production of ECM, potentially contributing to pathological myocardial remodelling. ¹³³ , ¹³⁴ In fact, galectin‐3, a well‐validated biomarker and mediator of cardiac fibrosis in HF patients, ¹³⁶ has been shown to stimulate pericyte proliferation and procollagen I secretion. ¹³⁷ This is in accordance with observations in angiotensin II‐induced myocardial hypertrophy model, in which Gli1⁺ cells were shown to consist in a subpopulation of pericytes that, in the setting of injury, differentiate into myofibroblasts and produce ECM in perivascular and interstitial spaces. ¹³⁸ Further supporting this role of CPc, in a clinically relevant rat model of HF with preserved ejection fraction (ZSF1 obese rats), decreased EC coverage was associated with subendocardial foci of CPc proliferation, which colocalized with ECM deposition and inflammatory cell infiltration. ¹³⁹ Consistently with this finding, pericytes have been shown to respond to proinflammatory stimuli with overexpression of cytokines, chemokines and CAMs, ¹⁴⁰ regulating immune cell diapedesis. ¹⁴¹ In the setting of experimental sepsis, inflammatory‐mediated CPc loss facilitates the infiltration of immune cells in cardiac interstitium. ¹⁴² These findings highlight the fact that, beyond being key determinants in the microvascular barrier, pericytes may detach from endothelial cells and promote interstitial remodelling in inflammatory injury.

Myocardial interstitium

The myocardial interstitium is a highly organized and compact structure, comprised by fibrillar collagen, non‐collagen matrix proteins, proteoglycans, GAGs and a wide array of bioactive signalling molecules ¹⁴³ (Figure 3 ). Cardiomyocytes are enclosed in a basement membrane, mostly constituted by integrins, laminin and fibronectin, behaving as anchoring points for fibrillar collagen and other matrix components (proteoglycans and GAG) attachment. Collagens (type I and III) are the predominant components of cardiac ECM, and their high tensile strength is assumed to be the main contributor for ECM structural integrity. ¹⁴⁴ Cardiac ECM architecture enables an effective force summation of individually contracting cardiomyocytes, allowing a coordinated myocardial tissue contraction, while at the same time maintaining adequate spatial relationships between cells, which prevents cardiomyocyte overstretching, preserves intercellular connections and opposes microcirculatory collapse.

Cardiac ECM composition is an important determinant of interstitial space volume and pressure. The interstitial space is densely crowded with intertwined components, which occupy the available physical space and limit the entrance of plasma proteins or cells, a phenomenon called steric interstitial exclusion. ² Given their polyanionic nature, interstitial GAGs futher contribute to limit the entrance of plasma proteins, while also binding free ions (mostly Na⁺) and annulling their osmotic force. ⁹³ Interestingly, changes in sulfated GAG conformation are associated with decreased Na⁺ buffering capacity and interstitial oedema. ⁹³ Moreover, the high stiffness of cardiac ECM not only preserves cardiomyocyte function by generating passive tension and avoiding tissue overstretching but also confers a low interstitial compliance to the myocardium and opposes interstitial space expansion. ¹⁴⁵ Consequently, in the setting of increased transcapillary filtration, interstitial fluid (IF) buildup stretches the ECM, causing a steep increase in interstitial pressure, which, in turn, forces IF into the lymphatic system. ²²

Alterations in ECM architecture or composition critically influence myocardial function. Increased ECM deposition, mainly in the form of collagen, has been recognized as an important mechanism of increased stiffness and diastolic dysfunction in most forms of chronic HF. ¹⁶ However, mechanical and enzymatic disruption of the ECM also significantly impairs myocardial systolic and diastolic function by compromising force transmission by displacing collagen struts from their anchoring points and breaking intercellular connections. ¹⁴⁶ , ¹⁴⁷ Moreover, inflammation‐driven up‐regulation of ECM‐degrading enzymes promotes both ECM and basement membrane degradation, decreasing interstitial exclusion effect and facilitating the interstitial passage of fluid, proteins and immune cells. ¹⁴⁸ ECM degradation has been shown in acute high‐grade myocardial inflammation, especially in experimental myocarditis ¹⁴⁹ and sepsis, ¹⁵⁰ where a significant acute decrease in total myocardial collagen content and collagen degradation were observed and associated with MO, systolic and diastolic dysfunction. Further supporting this experimental observation, post‐mortem evaluation of human septic myocardium found significant ECM disruption and interstitial oedema at the subepicardium, which colocalized with macrophage infiltration and cardiomyocyte apoptosis. ⁴⁵ Importantly, disruption of collagen struts may also increase coronary microvasculature susceptibility to external compression, which might compromise MBF in the setting of oedema‐associated increased interstitial pressure. ³⁶ , ¹²³

Interestingly, chronic oedematous states produced by increased microvascular filtration or decreased lymphatic drainage are associated with increased myocardial collagen deposition. ³ , ⁴ The interstitial remodelling may be interpreted as a compensatory mechanism, by decreasing interstitial compliance and preventing interstitial expansion, therefore minimizing the disruption of cardiac architecture. However, increased collagen deposition also causes long‐term detrimental effects on overall myocardial compliance and function. ¹⁵¹

Impaired turnover of non‐collagen ECM elements can also promote fibrosis and have detrimental effects on myocardial function. HA is observed in healthy cardiac ECM in its high‐molecular‐weight HA form and has a unique capacity to bind and retain water molecules. ¹⁵² Interestingly, while eGC HA degradation has been consistently associated with endothelial dysfunction, increased microvascular permeability and MO, ⁸⁵ , ¹⁵³ cardiac interstitial accumulation of HA, has similarly been shown to promote MO and structural remodelling. ¹⁵⁴ Cardiac interstitial accumulation of HA is normally associated with increased interstitial water content and MO, and is observed in myocardial infarction, ¹⁵⁵ hypertrophic cardiomyopathy, ¹⁵⁶ myocarditis ¹⁵⁷ and experimental cardiac transplant rejection. ¹⁵⁸ , ¹⁵⁹ Curiously, hyaluronidase treatment was able to decrease MO in rejected heterotopic transplants, ¹⁶⁰ whereas accumulation of low‐molecular‐weight HA (LMWHA) in hypertrophic cardiomyopathy is not associated with increased water content, ¹⁶¹ raising the possibility of distinct contributions of high‐molecular‐weight HA and LMWHA for oedema generation. Indeed, in the setting of inflammation and myocardial injury, production of LMWHA is preponderant and has been shown to stimulate TLR inflammatory signalling pathways. ¹²⁷ Collectively, these results underscore the importance of GAG structure, composition and regional distribution for IF balance.

Cardiac lymphatic system

The cardiac lymphatic system is essential in maintaining myocardial fluid balance and immunological homeostasis. ¹³⁹ It represents the main route for the removal of cellular metabolites, allowing the continuous IF renewal while avoiding the buildup of interstitial volume and pressure. ² Additionally, an immunomodulatory role has also been attributed to cardiac lymphatics due to the washout of proinflammatory mediators and immune cells from the myocardial interstitium in the setting of myocardial injury. ¹⁶² , ¹⁶³

Lymphatic capillaries are highly specialized blind‐ended structures, composed by oak‐leaf shaped lymphatic endothelial cells (LEC), which mostly lack basement membrane and are connected by permeable flap‐like intercellular junctions that favour unidirectional passage of IF, solutes and immune cells ¹⁶⁴ , ¹⁶⁵ (Figure 3 ). Moreover, cardiac LEC are connected to the surrounding ECM and cardiomyocytes by structures designated as anchoring filaments, constituted by type VII collagen projections, integrins and focal adhesion kinases. Anchoring filaments maintain lymphatic patency by exerting tensile forces and opening the lumen of lymphatic capillaries, facilitating lymphatic flow. ¹⁶⁶ , ¹⁶⁷ Anatomically, the lymphatic capillary plexus progressively converges from the subendocardium to the subepicardium, suffering structural alterations along the way, namely, the appearance of a continuous basement membrane, intraluminal valves to promote unidirectional flow, tight junctions, and, in larger trunks outside the myocardium, an adventitial layer and surrounding smooth muscle cells to help pump lymph. ¹⁶⁸ , ¹⁶⁹ Subepicardial lymphatic pre‐collectors converge to form epicardial lymphatic collectors that transport cardiac lymph via lymph nodes towards thoracic ducts, ultimately draining into the superior vena cava. ¹⁷⁰

Several factors influence cardiac lymph flow, most of which known to be unique to the heart. A distinctive feature of the intramyocardial lymphatic system is the absence of smooth muscle in intramyocardial vessels. Therefore, lymph flow is highly dependent on external forces, namely, muscle contraction and deformation along the cardiac cycle, heart rate and contractility. ¹ However, factors not intrinsic the heart function also impact lymph drainage. By concentrating interstitial metabolic products and proteins, lymph oncotic pressure exceeds interstitial oncotic pressure, promoting water osmotic dragging and fluid drainage. ¹ Coronary venous pressure is also an important regulator of lymph flow. Experimental coronary sinus blockade increases capillary hydrostatic pressure and promotes fluid filtration upstream, which requires compensatory lymphatic dilation and increased lymph flow to maintain fluid homeostasis. ¹³ , ¹⁴ , ¹⁴⁷ On the other hand, downstream, because lymph is ultimately drained into the venous circulation, increased central venous pressure acts synergically with decreased contractility to impair lymph flow in acute HF, promoting MO. ¹⁶ , ⁹³

The frequent observation of MO in several aetiologies of HF suggests that cardiac lymphatic inability to respond to increased filtration is a rather common finding. Despite the recognized ability of the healthy heart to respond to an increased capillary filtration by increasing lymph drainage severalfold, ¹ multiple disease mechanisms may render the cardiac lymphatic system incapable to cope. In this setting, lymphatic dysfunction will not only promote accumulation of a protein‐rich IF, which contributes to microvascular and cardiomyocyte stress, but will also have a proinflammatory effect by decreasing the clearance of proinflammatory cytokines and immune cells. ¹⁷¹ Prolonged residence of cellular debris, inflammatory mediators and cells in myocardial interstitium will aggravate and prolong myocardial inflammation, especially in the setting of myocardial infarction and myocarditis. ¹⁷² Furthermore, the distortion of interstitial architecture mediated by oedema and the activation of collagen and GAG‐degrading enzymes may have a negative impact on anchoring filaments and initial lymphatics, further compromising lymphatic patency and function. Whereas acute lymphatic obstruction leads to oedema, chronic obstruction is associated with interstitial fibrosis and ECM remodelling. ¹²⁵ Moreover, given the close proximity of the lymphatic and electrical conduction system, lymphatic dysfunction has also been shown to be associated with electrical disturbances. ¹⁷³

Lymphangiogenesis, the process of producing new lymphatic vessels is known to be a dynamic process mainly regulated by VEGF‐C and VEGF‐D binding to lymphatic‐specific receptor VEGFR3, and to be affected by inflammation and other cardiovascular factors (diabetes and obesity). ¹⁴⁶ In acute inflammation ¹⁷⁴ and in myocardial infarction, ¹⁷⁵ higher fluid filtration increases the need for lymph drainage, with resulting up‐regulation of lymphangiogenic factors. However, this endogenous response appears to be insufficient and to result in deficient lymphangiogenesis, with a predominance of lymphatic capillaries and lack of pre‐collectors. In fact, in post‐infarct mouse models, stimulating lymphangiogenesis with exogenous VEGF‐C or adrenomedullin increases lymph flow, decreases MO, attenuates myocardial inflammation and fibrosis and improves cardiac function. ¹⁷⁶ , ¹⁷⁷ , ¹⁷⁸ , ¹⁷⁹ , ¹⁸⁰ Still, this promising therapeutic avenue has been recently questioned by the absent impact of genetic blockade of lymphangiogenesis on cardiac function after experimental myocardial infarction. ¹⁸¹

Cardiomyocyte volume regulation

The cardiomyocyte membrane is highly permeable to water, which moves passively according to osmotic gradients and directly sets cell volume. ¹⁸² Normal cell function requires a stable volume and excessive water entry may disrupt membrane and cytoskeleton integrity. To prevent abrupt cell volume alterations, intracellular osmolarity is highly controlled, either with active ionic fluxes or the synthesis/degradation of osmotically active solutes ³⁹ , ¹⁸³ (Figure 3 ).

In the isotonic steady‐state, intracellular osmotic pressure exceeds extracellular osmotic pressure due to celullar concentration of organic phosphates and proteins, thus favouring passive water entry. To maintain the volume constant, the membrane Na⁺/K⁺ ATPase promotes the exit of 3 Na⁺ and entry of 2 K⁺ ions, a phenomenon known as the ‘Pump and Leak’ concept. Together with low Na⁺ membrane permeability, both mechanisms contribute to maintain a low intracellular [Na⁺] and a constant transmembrane gradient, on which many ionic transporters that regulate cell volume are highly dependent. In myocardial ischaemia, Na⁺/K⁺ ATPase dysfunction results in extracellular accumulation of K⁺, intracellular accumulation of lactate, Na⁺ and Cl^‐ and consequent cell swelling and membrane depolarization. ¹⁵⁹ , ¹⁶⁰ Furthermore, anaerobic metabolites accumulate in extracellular and intracellular spaces. Following reperfusion of the coronary vessels will re‐establish water delivery and wash out extracellular, but not intracellular, metabolic products, creating an osmotic gradient that promotes cell swelling. Highlighting the pathophysiological importance of cardiomyocyte oedema in ischaemia/reperfusion injury, reperfusion with a hypertonic solution limited MO and infarct size, when compared with isotonic solution. ¹⁸⁴ , ¹⁸⁵

Cell swelling depolymerizes actin filaments and disrupts cytoskeleton interactions with membrane proteins. ¹⁵⁹ Of note, cardiomyocyte swelling induced by ischaemia–reperfusion injury was associated with variable degrees of mitochondrial damage, cytoskeleton abnormalities and significant increases in sarcomere length, radial distance between myofibrils and distance between mitochondria and myofibrils, repercussing on maximal tension and calcium sensitivity. ¹⁸⁶ Accordingly, swelling of isolated cardiomyocytes induced by hypotonic medium was associated with lower contractility and activated NO/cGMP/PKG pathway. ¹⁵⁸

Despite most of myocardial water being confined to the intracellular compartment, few studies have addressed the pathophysiological role of cardiomyocyte swelling in HF.

Clinical perspective

In clinical research and practice, MRI stands out as the gold‐standard method for non‐invasive MO evaluation, based on its ability to identify the tissue ‘free’ water pool. ‘Free’ water molecules rotate very rapidly when subjected to a magnetic field and produce long T1 and T2 relaxation times, whereas ‘bound’ water molecules have their motion restricted due to hydrogen bonding with macromolecules, producing short T2 relaxation time values. The recent introduction of parametric mapping techniques—T1, T2 and extracellular volume, has enabled the detection of subtle changes in myocardial free water content and precise estimation of the interstitial fraction volume and composition. ¹⁸⁷

Making use of aforementioned MRI capabilities, evidence supporting the disruption of myocardial water balance has been shown in a broad range of cardiac and systemic diseases (Table 1 ). Overall, the increase in myocardial free water content is generally associated with depressed left ventricular function, increased NP plasma levels, disease progression and severity, and poor prognosis (Table 1 ). Nevertheless, due to the observational nature of these studies, a causal association between the presence of MO, LV dysfunction and cardiac prognosis could not yet be drawn.

Table 1.

Magnetic resonance imaging evidence of myocardial oedema in cardiac and systemic diseases

Disease	Myocardial oedema	CMR imaging	Analytical associations a	Clinical associations a	References
Acute heart failure	Global	T2 mapping	—	(+) PAWP (−) Decongestion	16 , 188
Myocardial infarction/Ischaemia–reperfusion	Focal	T2‐weighted imaging, T1, T2 and ECV mapping	(+) Troponin	(+) Infarct extension (+) MACE (+) LV dilatation (−) LV function	189 , 190
Aortic stenosis	Global	T1 and T2 mapping	—	—	191
Cardiomyopathies
Non‐ischaemic dilated cardiomyopathy	Global	T2‐weighted imaging, T2 mapping	—	(−) LVEF (+) Disease progression	192 , 193
Hypertrophic cardiomyopathy	Focal	T2‐weighted imaging, T2 mapping	(+) Troponin (+) BNP	(+) Risk of Syncope	194 , 195
Takotsubo cardiomyopathy	Focal	T1 and T2 mapping (USPIO enhancement)	(+) Myocardial macrophages	—	196 , 197
Peripartum cardiomyopathy	Global	T1 and T2 mapping	—	(−) LVEF	198
Infiltrative diseases
Cardiac amyloidosis	Global	T2 and ECV mapping	(+) NT‐proBNP	(+) Mortality (AL)	199
Cardiac sarcoidosis	Focal	T1 and T2 mapping	—	—	200 , 201
Fabry disease	Focal	T1 and T2 mapping	(+) Troponin	(+) ECG Changes (+) Clinical worsening	202
Infectious diseases
Viral myocarditis	Focal, subepicardial	T2‐weighted and LGE imaging, T2 mapping	(+) Troponin	(+) Arrhythmia (+) MACE (+) Death	9 , 203 , 204 , 205
COVID‐19	Focal	T2‐weighted and LGE imaging, T2 mapping	(+) Troponin (+) EMB macrophages	—	206 , 207 , 208 , 209
Sepsis	Focal	T2‐weighted imaging	—	—	45 , 210
HIV	Global	T2‐weighted and LGE imaging, T1 mapping	—	(+) Adverse cardiovascular events	211
Chagas disease	Focal	T2‐weighted and LGE imaging	—	(+) Disease severity	212
Inflammatory diseases
Rheumatoid arthritis	Focal	T1 and ECV mapping	—	(+) Disease activity (−) Circumferential Strain	213
ANCA‐associated vasculitides	Diffuse	T1 and T2 mapping	—	—	214 , 215
Systemic sclerosis	Focal	T1 and T2 mapping	—	(+) Cold pressor test (+) Disease activity (−) Circumferential Strain	216 , 217
Systemic lupus erythematosus	Focal	T1 and T2 mapping	—	(+) Disease activity	218 , 219
Endocrine diseases
Acromegaly	Global	T2 mapping	—	(+) Reversal of acromegalic cardiomyopathy	220
Hypothyroidism	Global	T2‐weighted and LGE imaging, T1 mapping	(−) FT3	(−) Stroke volume (−) Cardiac index	221
Cardiac surgery
Cardiac transplant	Global	T1 and T2 mapping	—	(+) Transplant rejection	222 , 223
Systemic diseases and others
Pulmonary arterial hypertension	Focal	T1 and ECV mapping	—	(−) RV function (+) RV dilatation	224
Chronic kidney disease	Global	T1 and T2 mapping	(+) Troponin T (+) NT‐proBNP	(+) Uremic Cardiomyopathy	225 , 226 , 227 , 228
Breast cancer chemotherapy (Anthracyclines/Trastuzumab)	Focal	T2 mapping	—	(+) Predictor of cardiotoxicity	229

^a Positive (+) and negative (−) associations with myocardial oedema.

AL, amyloid light‐chain; ECV, extracellular volume; EMB, endomyocardial biopsy; FT3, free triiodothyroinine (T3); LV, left ventricule; LVEF, left ventricular ejection fraction; MACE, composite of total death, myocardial infarction, coronary revascularization, stroke and hospitalization; PAWP, pulmonary arterial wedge pressure; RV, right ventricule; USPIO, ultra‐small particles of iron oxide

Myocardial oedema has been particularly well‐studied in the acute setting of ischemic heart disease, in which it may have a role on early injury during reperfusion and also late tissue healing. ¹⁶⁵ , ¹⁶⁶ , ¹⁶⁷ , ¹⁶⁸ , ¹⁶⁹ , ¹⁷⁰ , ¹⁷¹ During the initial phase of reperfusion, MO may contribute to the pathophysiological process of microcirculation compression and perfusion defects underlying the ‘no‐reflow’ phenomenon. ¹⁷¹ , ¹⁷² MO is also detectable later, at the time of tissue healing and collagen deposition, ¹⁶⁶ which discloses the complex interplay between myocardial fluid balance and inflammation and underscores the need for a cautious interpretation of MRI assessment of infarcted and at‐risk myocardium. ¹⁷¹ , ¹⁷³ Interestingly, patient comorbidities might impact on the development of MO in a disease‐specific and somewhat unpredicted way, underscoring the lack of clinical knowledge on this topic. As an illustration, diabetes was shown to aggravate post‐ischaemic MO, ¹⁷⁴ , ¹⁷⁵ whereas the opposite effect may be present in Takotsubo cardiomyopathy. ¹⁷⁶

Myocardial oedema has not been evaluated as an endpoint in HF randomized clinical trials, and the effect of most drugs on myocardial fluid balance is currently unknown. However, pre‐clinical evidence supports the beneficial effect of spironolactone ⁹² and SGLT2 inhibitors ¹⁷⁷ by protecting endothelial glycocalyx. Interestingly, these two drug classes were shown to provide clinical benefit across a wide ejection fraction range in HF, ¹⁷⁸ , ¹⁷⁹ , ¹⁸⁰ , ¹⁸¹ supporting a possible role for myocardial fluid balance among their mechanisms of action. Other drugs have proved useful to protect microvascular barrier in distinct clinical scenarios and may oppose MO formation. Of note, aprotinin, a fibrinolysis inhibitor, preserves adherens junctions and reduces MO in experimental cardioplegic arrest, ⁶⁵ whereas in sepsis, hydrocortisone ¹⁸² and sulodexide, a mixture of GAGs (heparan and dermatan sulfates), ¹⁸³ may protect the glycocalyx and diminish oedema formation.

In contrast, some drugs may facilitate the development of MO by impacting on microvascular filtration and ESL preservation. NP are known disruptors of the ESL ⁸⁹ , ⁹¹ and BNP levels correlate with myocardial water content across several clinical scenarios (Table 1 ), an association not yet known to be causal. However, it is tempting to speculate that this effect might have contributed to the somewhat disappointing results of BNP analogue nesiritide in the setting of acute HF treatment. ¹⁸⁴ In line with this, neprilysin is a known regulator of microvascular permeability by increasing the half‐life of NP and bradykinin, ¹⁸⁵ suggesting that sacubitril may also perturb microvascular barrier function. ²³⁰ Preclinical evidence suggests that beta‐blockers ¹⁸⁶ and calcium channel blockers ¹⁸⁷ may increase microvascular permeability, an effect not yet observed in the myocardium.

Finally, experimental data suggest that stimulators of lymphangiogenesis (e.g. VEGF‐C and adrenomedullin) may accelerate oedema resolution after myocardial infarction, ¹⁵² , ²³¹ but clinical studies are needed before considering this therapeutic pathway in HF.

Conclusion

In the failing heart, myocardial fluid balance is disrupted due to alterations in microcirculation dynamics, microvascular barrier, extracellular matrix composition and lymphatic function. Experimental data suggest that MO significantly impairs cardiac performance, affecting systolic and diastolic properties and promoting long‐term adverse remodelling. In the last decade, CMR has been increasingly used for HF phenotyping and data suggest the increase in myocardial free water content as relevant pathophysiological mechanism of cardiac injury and dysfunction, also representing an important prognosticator across multiple cardiac and systemic diseases. The recent advances in the knowledge of microvascular barrier and lymphatic function open the prospect for novel therapeutics targeting myocardial fluid disturbances in HF.

Funding

This study was funded by national funds through FCT ‐ Portuguese Foundation for Science and Technology, under the scope of the Cardiovascular R&D Center – UnIC (UIDB/00051/2020 and UIDP/00051/2020).

Vasques‐Nóvoa, F. , Angélico‐Gonçalves, A. , Alvarenga, J. M. G. , Nobrega, J. , Cerqueira, R. J. , Mancio, J. , Leite‐Moreira, A. F. , and Roncon‐Albuquerque, R. Jr. (2022) Myocardial oedema: pathophysiological basis and implications for the failing heart. ESC Heart Failure, 9: 958–976. 10.1002/ehf2.13775.