Electroimmunology and cardiac arrhythmia

Abstract

Conduction disorders and arrhythmias remain difficult to treat and are increasingly prevalent owing to the increasing age and body mass of the general population, because both are risk factors for arrhythmia. Many of the underlying conditions that give rise to arrhythmia — including atrial fibrillation and ventricular arrhythmia, which frequently occur in patients with acute myocardial ischaemia or heart failure — can have an inflammatory component. In the past, inflammation was viewed mostly as an epiphenomenon associated with arrhythmia; however, the recently discovered inflammatory and non-canonical functions of cardiac immune cells indicate that leukocytes can be arrhythmogenic either by altering tissue composition or by interacting with cardiomyocytes; for example, by changing their phenotype or perhaps even by directly interfering with conduction. In this Review, we discuss the electrophysiological properties of leukocytes and how these cells relate to conduction in the heart. Given the thematic parallels, we also summarize the interactions between immune cells and neural systems that influence information transfer, extrapolating findings from the field of neuroscience to the heart and defining common themes. We aim to bridge the knowledge gap between electrophysiology and immunology, to promote conceptual connections between these two fields and to explore promising opportunities for future research.

The clinical presentation of patients with heart rhythm abnormalities ranges from a lack of symptoms to sudden cardiac arrest, which can result in sudden cardiac death (SCD). SCD is a major public health problem, accounting for 50–60% of deaths in patients with coronary artery disease. With an incidence between 50 and 100 per 100,000 in the general population, SCD is a leading cause of death worldwide^1–5. Even when cardiac arrest occurs in a hospital, survival rates are only 3–10%, highlighting an unmet clinical need for risk prediction, prevention and adequate treatment of arrhythmias^6–8.

Classic electrophysiological concepts provide mechanistic insights into the pathophysiology underlying arrhythmogenesis, including abnormal automaticity, triggered activity and re-entry (FIG. 1). Although these mechanisms are well characterized, the prevalence of sudden cardiac arrest remains high. Effective treatment options are mostly limited to implantation of a defibrillator, and rates of SCD remain high, perhaps owing to the lack of preventive therapies. Expanding our understanding of the pathobiology of the arrhythmia beyond electrophysiology might help to develop new approaches to prevention and treatment.

Fig. 1 |. Electrophysiological concepts of conduction and arrhythmogenesis.

a | The physiological cardiac action potential can be divided into different phases. Depolarization due to rapid Na⁺ influx (phase 0). Transient K⁺ channels open and cause K⁺ efflux (phase 1). The plateau phase is facilitated by Ca²⁺ influx and counterbalanced by K⁺ efflux (phase 2). Closing of the Ca²⁺ channels causes rapid repolarization, but K⁺ channels remain open until the membrane potential returns to −90 mV (phase 3). Resting membrane potential, with Na⁺ and Ca²⁺ channels closed, but K⁺ channels open, holding the membrane potential at −90 mV (phase 4). b | Triggered activity usually results from prematurely activated cardiac tissue, causing early or late afterdepolarizations. c | Increased automaticity from dominant pacemaker cells is usually triggered by abnormal impulse function and can lead to tachyarrhythmias by increasing the rate of action potential discharge due to sympathetic stimulation, whereas decreased pacemaker rates slow the heart rate. d | A depolarizing impulse (red arrow) encounters an anatomical obstacle and circles around it. The electrical impulse constantly ‘runs around’ the block, along the excitable tissue gap. If the wavefront never reaches the refractory tail, a re-entrant loop forms, causing sustained arrhythmia. ECG, electrocardiogram; LV, left ventricular.

New insights into leukocyte function suggest that immune cell populations might have a role in normal heart rhythm via crosstalk with cardiomyocytes and hypothetically also in the pathophysiology of arrhythmogenesis. Re-entrant arrhythmias, in which anatomical obstacles are arrhythmogenic, are particularly associated with leukocyte recruitment, which, we hypothesize, shapes the composition of the arrhythmogenic substrate. Re-entry is the underlying mechanism for most sustained cardiac arrhythmias and is thought to be based on two components³: a trigger, which is often a mild form of arrhythmia such as a premature ventricular complex, and a substrate, leading to sustained and prolonged arrhythmia, including structural remodelling such as scar formation after acute myocardial infarction (MI) or patchy fibrosis in the setting of hypertension and hypertrophy. The underlying structural heart diseases are diverse and range from acute events to chronic processes, each with unique electrophysiological properties. Re-entrant arrhythmias can also cause atrial fibrillation (AF)⁹, a common type of arrhythmia^10,11 (TABLE 1). Treating re-entrant arrhythmias often begins with resolving the underlying disease, such as alleviating ischaemia or repairing a valve. However, understanding the specific electrophysiological characteristics of structural heart disease is important for the development of targeted treatment options.

Table 1 |.

Definitions of clinically important arrhythmias

Arrhythmia	Example ECg	ECg description	Underlying pathologies and their inflammatory signatures
Ventricular arrhythmias
Ventricular tachycardia		More than three consecutive QRS complexes originating in the ventricles at a rate of >100 bpm. Regular QRS complex shape missing; subtypes include sustained (>30 s), unsustained (more than three beats, terminating spontaneously), monomorphic (symmetrical QRS morphology from beat to beat) and polymorphic (unstable and multiform QRS morphology from beat to beat)	Myocarditis: massive recruitment of macrophages and T lymphocytes133,139,146,278 Myocardial infarction: neutrophil and monocyte recruitment, tissue-resident macrophage death and release of inflammatory cytokines and chemokines, such as CC-chemokine ligand 2, IL-1, IL-6 and tumour necrosis factor149–151,153,157,279
Ventricular fibrillation		Rapid and disorganized electrical activity, with pronounced variability in electrocardiographic waveform; ventricular rate usually >300 bpm; irregular shape of the QRS complex
Ventricular flutter		Regular arrhythmia with ~300 bpm and sinusoidal, monomorphic appearance; owing to the high speed, usually no isoelectric interval between successive QRS complexes
Ventricular tachycardia or ventricular fibrillation storm		A state of severe electrical instability defined as more than three episodes of sustained ventricular tachycardia or fibrillation within 24 h
Sudden cardiac arrest		Sudden cessation of cardiac activity, without QRS complexes; affected patients become unresponsive, with no signs of blood circulation (such as a perceptible pulse)
Atrial arrhythmias
Atrial fibrillation		Absent P wave combined with an irregular ventricular rate; the isoelectric line can be characterized by either fibrillatory waves (F waves) or just minute oscillations; subtypes include paroxysmal (terminates spontaneously or with intervention within 7 days of onset), persistent (continuous atrial fibrillation that is sustained for 7 days) and longstanding persistent (continuous atrial fibrillation for >12 months)	Heart failure: recruitment of macrophages, inflammasome activation and fibrosis280–283 Sepsis: increased blood leukocyte count, systemic release of pro-inflammatory cytokines, and changes in cardiac macrophage number and phenotype114–117,218
Atrial flutter		Classic sawtooth ECG sign with atria firing at a rate of 200–350 bpm; R-R intervals are regular; the ratio of P waves to QRS complexes can range from 2:1 to 4:1
Conduction delay
Sinus node dysfunction (sinoatrial block)		Conduction and sinus rate are usually slower than physiological sinus rhythm because the dominant pacemaker (the sinoatrial node) is disturbed; subtypes include sinus bradycardia (sinus rate <50 bpm), sinus pause (sinus node depolarizes >3 s after the last atrial depolarization) and sinus node arrest (no sinus node depolarization)	Structural abnormalities: progressive, degenerative fibrosis associated with inflammageing284–286 Autoimmune disorders: autoantibodies against cardiac ion channels or adrenergic receptors264–266,270
Atrioventricular block	First-degree atrioventricular block	Disturbed conduction from the atria to the ventricles owing to improper function of the atrioventricular node; subtypes include first-degree atrioventricular block (P waves associated with 1:1 atrioventricular conduction, but delayed), second-degree atrioventricular block (P waves with a constant rate in which atrioventricular conduction is present but not at a ratio of 1:1) and third-degree atrioventricular block (complete heart block lacking atrioventricular conduction)
	Second-degree atrioventricular block
	Third-degree atrioventricular block
Conduction tissue disease	Right bundle branch block	Complete or partial interruption of the electrical pathways in the ventricles, mostly affecting the anatomical structures in the bundle of His and Purkinje fibres; subtypes include right bundle branch block (mainly affecting the right ventricle; typical ECG signs are a wide QRS complex with large, positive, dominating R waves) and left bundle branch block (mainly affecting the left ventricle, with a broad notched or slurred R wave and a dominating S wave)
	Left bundle branch block

Definitions are in accordance with clinical guidelines^{3,284,287,288}. ECG, electrocardiogram.

In this Review, we aim to integrate electrophysiology and immunology by focusing on conceptual bridges between these two fields. We summarize basic principles of leukocyte biology, focusing on those that are relevant to conduction in the heart and also in the brain, where resident macrophages (microglia) modulate information transfer by neurons. We discuss the clinical association between leukocyte expansion and arrhythmias and explore emerging mechanistic data on the causal role of leukocytes in arrhythmogenesis. Furthermore, we examine inflammation as a potential therapeutic target and highlight technical developments in immunology and electrophysiology that will facilitate future investigation into the interactions between leukocytes and conducting tissue. We are promoting the case for a more comprehensive consideration of the roles of leukocytes in cardiac conduction and arrhythmia; however, we do not discount the undisputed contributions by other heart cells that have been reviewed previously^12–14.

Charging leukocytes for defence

The innate immune system

Innate immunity refers to immune responses that are inborn and not learned, as opposed to lymphocyte-based responses that form the adaptive immune system¹⁵. The innate immune response — a crucial step in the first-line host defence against infectious agents and tissue damage — is mounted by the epithelial and endothelial barriers, neutrophils, macrophages and monocytes, dendritic cells, natural killer cells and mast cells¹⁶. The innate immune system is activated when Toll-like receptors (TLRs), NOD-like receptors or RIG-I-like receptors recognize pathogen-associated molecular patterns¹⁷. Some of these receptors also recognize self-derived host molecules called damage-associated molecular patterns, which arise from tissue injury and inflammation¹⁸. The binding of ligands to these receptors releases cytokines and chemokines, which recruit and activate additional immune cells.

Leukocytes

Leukocytes are the cellular protagonists of immune defence, but also have non-immunological functions that are dependent on the organ in which they reside. In this section, we briefly summarize the functional properties of leukocytes and their subsets, focusing on the heart (BOX 1). Monocytes and macrophages are the most numerous leukocytes in the heart and have pivotal roles in cardiovascular pathophysiology. Monocytes are produced in the bone marrow and circulate in the bloodstream before migrating to tissues where they differentiate into macrophages or dendritic cells¹⁹. In mice, mature monocytes can be identified by the expression of CD11b and CD115 and distinguished by the surface marker Ly6C. Ly6C^high monocytes preferentially accumulate at sites of inflammation and differentiate into macrophages. Ly6C^low monocytes patrol the vasculature, remove damaged endothelial cells, and maintain vascular integrity and homeostasis^20,21. Macrophages reside in almost all tissues. In the murine heart, macrophages constitute 7–8% of non-cardiomyocyte cells^22,23, and single-cell RNA-sequencing data indicate that 5–10% of cells in the human heart are leukocytes²⁴.

Box 1 | Leukocytes and their subsets in heart homeostasis

Macrophages comprise 75% of total cardiac leukocytes, and their subsets and functions in cardiac health are described in the main text, but how other leukocytes present in the heart contribute to heart homeostasis and conduction in steady state is mostly unknown.

Neutrophils

Neutrophils are polymorphonuclear, phagocytic leukocytes that act as the first line of host defence against pathogens but also mediate inflammation-induced injury. Rarely found in the healthy myocardium, neutrophils are terminally differentiated cells with phenotypic heterogeneity and functional versatility^290,291. The granulocyte population in the heart increases with age²⁹². Myeloperoxidase, which is a haem enzyme abundantly expressed by neutrophils, promotes atrial fibrosis, leading to an increased vulnerability to atrial fibrillation²³¹.

Dendritic cells

Dendritic cells, which are the major antigen-presenting cells, have an essential role in adaptive immunity and are an important link between the innate and adaptive immune systems²⁹³. The healthy heart contains resident cardiac dendritic cells, abundant in the aortic valve and sinus, that constitute about 1% of total cardiac leukocytes. These dendritic cells comprise CD103⁺CD11b⁻ and CD103⁻CD11b⁺ subsets^26,294; CD103⁺ cardiac dendritic cells proliferate more than CD11b⁺ cardiac dendritic cells²⁹⁵. CD26 expression is higher in cardiac dendritic cells than in cardiac macrophages²⁹⁵. Resident dendritic cells isolated from the aorta and valves present antigens to CD8⁺ T cells in vitro and in vivo²⁹⁴.

B lymphocytes

B lymphocytes produce antibodies, present antigens and secrete factors involved in immune regulation. Prevalent in the healthy murine heart, B lymphocytes make up about 9% of cardiac leukocytes^23,296,297. Myocardial B cells can be grouped into three subpopulations: CD19⁺CD11b⁻, CD19⁺CD11b⁺IgM⁺CD5⁺ and CD19⁺CD11b⁺IgM⁺CD5⁻ (REF.²⁹⁶). Most myocardial B cells reside in the microvasculature and patrol along the endothelium, whereas some cross the endothelium into myocardial tissue²⁹⁸. B cell-deficient mice have fewer myocardial Ly6C⁺ monocytes and more myocardial CD4⁺ and CD8⁺ T cells²⁹⁸, suggesting that B cells modulate the composition of the myocardial leukocyte pool.

T lymphocytes

T lymphocytes elicit cell-mediated immunity and are divided into T helper cells, cytotoxic T cells and regulatory T cells. In the healthy heart, T lymphocytes comprise about 3% of total cardiac leukocytes^23,297. Aged hearts have more CD8⁺ lymphocytes; CD4 to CD8 ratios decrease with age, but the functional implications of this change are unclear²⁹². After myocardial infarction, activated CD4⁺ T lymphocytes are found in the heart-draining lymph nodes. Elderly MHC class II-deficient mice have preserved left ventricular fractional shortening and end-diastolic dimension²⁹². Heart samples from CD4-deficient or MHC class II-deficient mice showed reduced expression levels of pro-inflammatory cytokines compared with heart samples from wild-type mice²⁹². These data indicate that CD4⁺ T lymphocytes mediate inflammation and mild dysfunction in aged hearts.

Mast cells

Mast cells are well known as primary effector cells of allergic and anaphylactic reactions. Mouse^28,299 and human^24,300 hearts contain a fairly low number of mast cells. Most of the heart’s mast cells are located in the myocardium and in the epicardium²⁸. Mast cell-derived renin cleaved angiotensin I and promoted local angiotensin II formation and arrhythmias in guinea-pig isolated hearts subjected to ischaemia³⁰¹. In addition, cardiac mast cells accumulated in the left and right atria of mice with pressure overload, where they promoted atrial fibrosis through the secretion of platelet-derived growth factor A³⁰².

Cardiac macrophage subsets are classified by their surface expression of CC-chemokine receptor 2 (CCR2) and major histocompatibility complex (MHC) class II molecules. Human and mouse hearts contain at least three subpopulations of macrophages: CCR2⁻MHCII^low, CCR2⁻MHCII^high and CCR2^highMHCII^high (REF.²⁵). The two CCR2⁻ macrophage subsets have embryonic origins and are maintained by local proliferation in the myocardium, independent of blood monocytes²⁶. CCR2⁻ macrophages reside predominantly in the myocardial wall and participate in coronary vasculature development²⁷. The CCR2⁺ macrophage subset (5–15%) is the smallest in healthy hearts and derives from blood monocytes. Single-cell RNA-sequencing studies have revealed additional cardiac macrophage subsets^28–30, but their functional relevance remains to be determined.

Electrophysiology and leukocyte function

Like cardiomyocytes, neurons and many other cell types, leukocytes express various channels to regulate their membrane potential as well as signalling by Ca²⁺ and other ions. These cell features regulate gene expression, cytokine release, cell activation and migration³¹ and, consequently, inflammation^32,33. The resting membrane potential of leukocytes ranges from −30 mV to −60 mV (REFs^34–39) and is determined by the equilibrium potentials of ions such as K⁺, Na⁺ and Cl⁻, which traverse the cell membrane through specific channels.

In the 1970s, K⁺ currents in peritoneal macrophages were first described^34,35,40. K⁺ channels in macrophages include the voltage-gated K⁺ channel K_v1.3 and the Ca²⁺-activated K⁺ channel K_Ca3.1, and these channels mainly promote a negative membrane potential. K_v1.3 is activated by membrane potential depolarization⁴¹. Although K_v1.3 is the predominant form of voltage-gated K⁺ channel, K_v1.5 is also present in myeloid cells and can form heterotetramers with K_v1.3. The ratio of K_v1.3 to K_v1.5 in the K_v complex can vary, leading to biophysically and pharmacologically distinct channels^42,43. Interestingly, cell activation with either the bacterial wall component lipopolysaccharide or tumour necrosis factor (TNF) increases K_v1.3 activity in macrophages^42,44. By contrast, anti-inflammatory signals such as glucocorticoids downregulate K_v1.3 and increase the presence of K_v1.5 in the heteromeric K_v channels^42,44. Unlike K_v1.3, K_Ca3.1 is voltage-independent and is activated by the binding of Ca²⁺ to calmodulin, which is associated with the C terminus of the K_Ca3.1 channel⁴⁵. Opening of K_Ca3.1 channels hyperpolarizes the membrane potential following an increase in the intracellular Ca²⁺ concentration.

In lymphocytes, the relative contributions of K_v1.3 and K_Ca3.1 to Ca²⁺ influx are determined by their expression levels, which depend on the lymphocyte subset and on cell activation^33,46,47. Under resting conditions, naive T cells predominantly express the K_v1.3 channel⁴⁶. Following activation, T cells transcriptionally upregulate K_Ca3.1, which sustains Ca²⁺ signalling⁴⁷. T helper 1 and T helper 2 lymphocytes predominantly express K_Ca3.1 and depend on K_Ca3.1 for cytokine production, whereas T helper 17 cells mainly express K_v1.3, which regulates IL-17 production³³. Taken together, activation of K_v1.3 by depolarization and of K_Ca3.1 by Ca²⁺ preserves the negative membrane potential required for Ca²⁺ entry.

Transient receptor potential cation channel subfamily M member 4 (TRPM4) is another essential regulator of the membrane potential of leukocytes and is much more permeable to Na⁺ than to Ca²⁺ (REF.⁴⁸). TRPM4 channels are activated by increases in intracellular Ca²⁺ concentration, resulting in Na⁺ influx and membrane depolarization, which then limits Ca²⁺ influx because of the reduction in the electrochemical gradient.

Ca²⁺ is a second messenger that regulates proliferation, gene expression and motility in leukocytes. Ca²⁺ release-activated Ca²⁺ (CRAC) channels (ORAI1, ORAI2 and ORAI3), known as store-operated Ca²⁺ channels, are present in most leukocytes. Activation of CRAC channels is mediated by stromal interaction molecule 1 (STIM1) and STIM2, which are located in the endoplasmic reticulum membrane. Following the depletion of endoplasmic reticulum Ca²⁺ stores, STIM1 and STIM2 activate, oligomerize and translocate to the junctions between the endoplasmic reticulum and the plasma membrane, into which ORAI1 is recruited, resulting in Ca²⁺ influx^49–51. The importance of ion channels for inflammation is exemplified by the observation that CD4⁺ and CD8⁺ T cells from mice and humans that are deficient in ORAI1 and/or STIM1 have defective production of many cytokines, including IL-2, IL-4, IL-17, IFNγ and TNF^52,53. Moreover, B cells from mice lacking either ORAI1 or STIM1 and STIM2 have less B cell receptor-induced proliferation⁵⁴. The crucial roles of CRAC channels in adaptive immune responses are well established, but little is known about their function in innate immunity.

Ionotropic P2X receptors are ligand-gated, non-selective cation channels that are activated by extracellular ATP and facilitate the influx of Na⁺, Ca²⁺ and other cations⁵⁵. P2X receptor opening causes Ca²⁺ influx and activates downstream signalling, leading to the proliferation of T cells and B cells^56,57. Ion and connexin channel expression in leukocytes based on data from the Immunological Genome Project⁵⁸ is summarized in TABLE 2, and their known functions are summarized in Supplementary Table 1. Taken together, these data indicate that leukocyte functions — and inflammation as a consequence^32,33 — are regulated by ion flux across the leukocyte plasma membrane and the resulting changes in membrane potential. Of note, leukocytes and cardiomyocytes share several ion channels, such as ORAI1 and its activator STIM1. Genetic variants in ORAI1 or STIM1 in humans that lead to the expression of non-functional ORAI1 or STIM1 protein or a complete lack of the protein are associated with a clinical phenotype characterized by immunodeficiency⁵⁹. Because individuals who lack ORAI1 or STIM1 have lethal immunodeficiency and die in their first years of life, the influence of a sustained lack of these proteins on the heart might not be apparent. However, cardiomyocyte-restricted Stim1-knockdown mice have a proclivity for arrhythmia, involving decreased conduction velocity and increased action potential duration (APD)⁶⁰. How cardiac immune cell functions are influenced by the regular depolarization and repolarization of the surrounding myocardium, electrolyte imbalance or even arrhythmia is so far unknown. Given that the heart has a sizeable population of resident immune cells that might affect cardiomyocyte health⁶¹ and extracellular matrix composition, this question merits further investigation.

Table 2 |.

Ion channel and connexin channel expression in leukocytes

Channel	Monocytes or macrophages	Resident macrophages	T lymphocytes	B lymphocytes	Dendritic cells
Ca2+ channels
ORAI1	+	+ (m, k, a)	+	+	+
STIM1, STIM2	+	+ (m, k, a)	+	+	+
TRPM2	+	+ (m, k)	+	+	+
TRPM7	+	+ (m, k, a)	+	+	+
TRPV2	+	+ (m, k, a)	+	+	+
P2X1	+	+ (m)	+	+	−
P2X4	+	+ (m, k, a)	+	+	+
P2X7	+	+ (m, k, a)	+	+	+
Cav1.3	+	−	+	+	+
Cav1.4	+	−	+	−	+
Cav2.1	+	−	+	−	−
Cav2.2	−	−	−	+	+
Cav2.3	−	−	+	+	+
Cav3.2	−	−	+	+	+
Cav3.3	−	−	−	+	−
K+ channels
Kv1.1	−	−	+	+	−
Kv1.2	+	+ (m, k, a)	+	+	+
Kv1.3	+	+ (m, k, a)	+	+	+
Kv2.1	−	+ (k)	−	−	−
Kv3.1	−	−	+	−	−
Kv3.4	−	−	+	−	+
Kv4.1	−	+ (m)	−	−	+
Kv7.1	−	+ (k)	+	−	+
Kv7.5	−	−	+	+	−
Kv11.1	+	−	+	−	−
KCa3.1	+	+ (m, k)	+	+	+
Kir2.1	+	+ (k)	−	−	−
Kir2.4	−	−	+	−	+
Kir3.3	+	−	−	−	−
Kir6.2	+	−	+	+	+
Na+ channels
Nav1.5	−	−	−	−	+
TRPM4	+	−	+	−	+
Gap junctions
Cx43	+	+ (k)	+	+	+
Cx30.2	+	+ (m, k, a)	+	+	+
Cx46	+	+ (m, k, a)	+	+	+
PANX1	+	+ (m, k, a)	+	+	+
PANX2	+	+ (m, k, a)	+	+	+

Data are based on Immunological Genome Project microarray gene expression data²⁸⁹; expression was deemed to be present (+) if the robust multiarray average normalized value was >120, otherwise expression was considered to be absent (−). a, alveolar macrophages; Ca_v, voltage-gated Ca²⁺ channel; Cx, connexin; k, Kupffer cells; K_Ca, Ca²⁺-activated K⁺ channel; K_ir, inward-rectifier K⁺ channel; K_v, voltage-gated K⁺ channel; m, microglia; Na_v, voltage-gated Na⁺ channel; PANX, pannexin; STIM, stromal interaction molecule; TRP, transient receptor potential.

Leukocytes in cardiac conduction

Gap junctions, formed by the connexin protein family, allow cytoplasmic exchange of ions, nucleotides, metabolites and electrical signals between connected cells⁶². Connexins are present in almost all cells, including cardiomyocytes and leukocytes. In the healthy heart, gap junctions provide pathways for intercellular current flow, enabling coordinated propagation of the action potential^63,64. Direct macrophage–cardiomyocyte interaction through connexin 43 (Cx43)-containing gap junctions was found in mouse and human hearts³⁷. This contact enables electrical coupling between the two cell types. Patch-clamp experiments revealed that many macrophages rhythmically depolarize when they are coupled to cardiomyocytes. In vitro coupling rendered the resting membrane potential of cardiomyocytes more positive — a change that was reversed by pharmacological blockade of Cx43 (REF.³⁷). Computational modelling indicated that coupling of higher numbers of macrophages to one cardiomyocyte (a reasonable assumption given that cardiomyocytes are much larger than macrophages) lowers the action potential upstroke and overshoot, leading to earlier repolarization and a shorter refractory period³⁷.

Optical clearing of mouse and human heart samples followed by confocal fluorescence microscopy documented that macrophages are concentrated not only in the distal atrioventricular node, which electrically connects the atria and the ventricles, but also in the sinus node³⁷. In vivo, macrophage-specific genetic ablation of Cx43 (also known as Gja1) delayed conduction through the atrioventricular node, and depletion of myeloid cells in CD11bDTR mice resulted in progressive atrioventricular block³⁷. Collectively, these findings suggest that macrophages support normal electrical conduction and influence cardiac conduction through gap junctions. In addition to questions about whether this communication might hypothetically shape arrhythmias, which we discuss below, we speculate that immune cells can affect electrical communication between cardiomyocytes and influence conduction via currently unknown mechanisms, perhaps by shaping the development of the conduction system, regulating turnover of insulating extracellular matrix, interfering with gap junction communication between cardiomyocytes, or regulating ion channel expression and function in cardiomyocytes. The abundance of macrophages in cardiac conduction system structures³⁷ suggests that they have undiscovered roles. Single-cell RNA-sequencing analysis of mouse atrioventricular node samples³⁷ indicated that macrophages in the atrioventricular node cluster according to the typical subset markers MHC class II and CCR2 and express conduction-related genes, including those encoding ion channels, such as Cacna1c (Ca_v1.2), Kcnj2 (K_ir2.1), Kcnq1 (K_v7.1), Hcn2 (HCN2) and Kcnh2 (K_v11.1) (TABLE 2; Supplementary Table 1). This single-cell RNA-sequencing study contains data on 76 macrophages isolated by microdissection of the atrioventricular node but lacks a comparison with macrophages from a remote control region³⁷. Whether atrioventricular node macrophages differ from other macrophages located in the atria or ventricles remains unresolved. Another important question is which of these ion channels are functional in cardiac macrophages. Furthermore, whether other immune cells, although less numerous than macrophages, also couple to cardiomyocytes is worth exploring.

Leukocytes and neural function

Similar to the population of macrophages resident in the heart, the brain contains a large population of resident innate immune cells. Another similarity inviting a comparison of these organ systems is illustrated by electrophysiological studies showing that the behaviour of cardiac Purkinje fibres is electrically similar to that of nerve axons⁶⁵. Both the heart and the brain contain a massive number of electrically excitable cells (cardiomyocytes and neurons, respectively), which rely on an organized conduction system to function. In the brain, axonal projections from different neuronal subtypes propagate electrical impulses to specific anatomical locations. Microglia, the brain’s equivalent of tissue-resident macrophages, interact with almost all cell types in the central nervous system to mediate developmental programmes and maintain homeostasis⁶⁶.

Microglia have essential roles in excitatory networks in the brain. During development, microglia engulf synapse components and actively participate in synapse maturation^67,68, raising the question as to whether cardiac macrophages participate in the development of the cardiac conduction system. IL-33 produced by astrocytes in the brain promotes microglial synapse engulfment, indicating that astrocyte–microglia communication is required for synapse homeostasis during development⁶⁹. Moreover, microglia contact with dendrites induces filopodia formation in the developing somatosensory cortex⁷⁰. Microglia are also required for the development of oligodendrocytes, their progenitors and the subsequent myelination process⁷¹. In adults, microglia maintain the oligodendrocyte progenitor pool and modulate synaptic neurotransmission via microglia–astrocyte crosstalk. Activation of microglia by lipopolysaccharide induces a rapid increase in spontaneous excitatory post-synaptic currents in hippocampal slice preparations⁷². These activated microglia release small amounts of ATP, which targets P2Y1 receptors located on astrocytes, leading to glutamate release; subsequently, glutamate acts on neuronal receptors⁷².

Connexin expression by microglia depends heavily on the context of cell activation, which has not been explored in cardiac macrophages. Whereas Cx32 and Cx36 are expressed in the resting surveillance state^73,74, Cx43 expression is rarely detected in steady-state microglia. Intercellular electrical coupling was observed between cultured microglia and neurons through Cx36 gap junctions⁷³. Approximately one-third of co-cultured microglia and neuron pairs showed evidence of electrical coupling, with small unitary conductance and very low voltage sensitivity, which is consistent with the characteristics of Cx36 channels. This situation is reminiscent of the coupling between cardiac macrophages and cardiomyocytes via Cx43 (REF.³⁷). Likewise, Cx36 is involved in coupling between microglial pairs⁷³, raising the question as to whether cardiac macrophages couple to each other. Of note, >2 weeks of co-culture might have affected the native characteristics of microglia–neuron crosstalk, especially given that Cx36 expression is low in non-cultured naive microglia, as reported by the Immunological Genome Project⁵⁸. Nevertheless, the findings support the notion that brain tissue-resident macrophages couple with excitable cells and contribute to electrical conduction in the brain, just as in the heart.

Despite the obvious differences between the brain and the heart, both are organs in which the primary function directly depends on electrical activity; therefore, comparing them might be instructive. Direct electrical coupling between tissue-resident macrophages and cardiomyocytes³⁷ or neurons⁷³ has been observed in culture, but in vivo data on coupling are currently available only for the heart³⁷. Extrapolating the observation that microglia influence synaptic neuron-to-neuron communication leads us to wonder whether cardiac macrophages might have similar roles. We speculate that cardiac macrophages might modulate the excitation transfer between two neighbouring cardiomyocytes, perhaps by influencing gap junction abundance or deposition of insulating extracellular matrix by fibroblasts. One question arising from the microglia data on synaptic pruning^67,68 is whether macrophages influence the embryonic development of the conduction system. These conceptual connections underscore the value of neuroscience tools for research on cardiac conduction, as indicated by the use of optogenetics³⁷ and in vivo heart microscopy⁷⁵ in conjunction with ion and voltage reporters^76,77 (BOX 2).

Box 2 | Technological developments.

Optogenetics

Optogenetics combines genetic and optical methods to cause or inhibit precisely defined events in specific cells of living tissue and animals³⁰³. There are two classes of optogenetic devices: actuators and light-emitting sensors³⁰⁴. Actuators, such as channelrhodopsin 2 (ChR2), transduce optical signals into physiological signals. Light exposure at a certain wavelength elicits action potentials in ChR2-expressing cells, such as neurons^305,306. ChR2 opens after absorption of a photon to generate a large permeability for cations, leading to depolarization³⁰⁷. Sensor proteins, such as voltage-sensitive fluorescent protein³⁰⁸ or green Ca²⁺ indicator protein (GCaMP)³⁰⁹, produce fluorescent signals in response to changes in membrane potential, intracellular Ca²⁺ concentration or synaptic transmission, which make such activity detectable. The first cardiac optogenetics studies used light to pace the hearts of transgenic mice³¹⁰ and to modulate pacemaker activity in the hearts of zebrafish³¹¹. Optogenetics can be used to alter the function of both cardiomyocytes³¹² and non-cardiomyocytes³⁷. Moreover, optogenetics can be used to terminate arrhythmias in rodent hearts^313–315.

Intravital microscopy

Intravital microscopy provides the ability to visualize and quantify events in the native tissue environment. Subcellular spatial and millisecond temporal resolution in the mouse heart in vivo were achieved by combining tissue stabilization, cardiac gating and image-processing algorithms to suppress motion artefacts^316,317. Fluorescent imaging probes can be used to assay physiology and cell–cell interaction in the heart by staining for specific cells, including leukocytes⁷⁵. Imaging of the genetically encoded, highly sensitive Ca²⁺ indicator GCaMP6 (REF.⁷⁶) in the beating heart resolves Ca²⁺ dynamics in single cardiomyocytes³¹⁸. In the future, cardiac intravital microscopy will be used quantitatively to measure not only cardiomyocyte contraction³¹⁷ but also membrane potential changes, Ca²⁺ signalling at the cellular level³¹⁸ and leukocyte trafficking³¹⁹.

Inflammation and arrhythmias

Clinical associations

AF is a prevalent condition, with a rising incidence in elderly individuals^10,78. The lack of organized atrial contraction causes blood stasis, which increases the risk of thrombus formation and stroke⁷⁹. Moreover, during AF, the ventricles might not fill adequately, reducing cardiac output and elevating the risk of heart failure (HF), cardiovascular death and SCD⁸⁰. Repetitive high-frequency ectopy and re-entry are thought to be the major AF-maintaining mechanisms in atria with vulnerable morphological substrates^81,82.

More than 20 years of research point to an association between inflammation and AF^83–85. Leukocyte recruitment occurs in the atria of patients with AF^86–88. Among immune cells, CD68⁺ macrophages are more numerous than adaptive immune cells such as CD3⁺ T lymphocytes in the atria of patients with AF⁸⁹. Leukocytes have an important role in local inflammation in releasing cytokines and chemokines, such as interleukins and TNF, and activating innate signalling pathways, such as the inflammasome⁹⁰. These actions contribute to structural, electrical and mechanical heterogeneity in the atrial myocardium and thereby facilitate the pathogenesis of arrhythmogenesis (see Inflammatory mechanisms below). Case–control studies have shown elevated circulating levels of inflammatory molecules (C-reactive protein (CRP), heat shock protein 27, IL-6 and TNF) in patients with AF^91–93. CRP has been linked to the development of AF^94,95: in the Cardiovascular Health Study⁹⁶ involving 5,806 participants, higher circulating CRP levels were associated with the presence and future development of AF. Anti-inflammatory therapy reduced the burden of AF in a dog model of sterile pericarditis and in patients after cardiac surgery^97–100. Profiling inflammatory biomarkers and other signalling molecules might help to predict the risk of AF in patients^101,102.

Multiple major cardiovascular risk factors and inflammation-associated conditions, including hypertension, coronary artery disease, HF and obesity, are linked to AF^103–105. Although various pathological mechanisms, including endothelial dysfunction and myocardial ischaemia, underlie these risk factors, they might converge on a final common pathway: leukocyte expansion in the atrial myocardium and consequent production of pro-inflammatory cytokines.

The interconnected development of HF and AF involves inflammatory components that probably interact with neurohumoral activation, morphological remodelling, fibrosis and disruption of cardiomyocyte energy metabolism; together, these actions alter the myocardial substrate, leading to arrhythmogenesis. In patients with AF, the arrhythmogenic substrate involves three electrophysiological mechanisms: abnormal automaticity¹⁰⁶, triggered activity^107,108 and re-entry¹⁰⁹. The most common underlying mechanisms for ventricular tachycardia (VT) or ventricular fibrillation (VF) are early or delayed afterdepolarizations, which can affect the cardiomyocyte APD. Several inflammatory mechanisms might contribute to ventricular dysfunction (reviewed previously¹¹⁰).

New-onset AF is common in patients with acute sepsis¹¹¹. The pathogenesis of sepsis includes the interaction between pro-inflammatory cytokines and stress hormones, autonomic dysfunction, extravascular volume shifts and cardiovascular compromise¹¹², all of which can beget AF. Signs of systemic inflammation consist of changes in leukocyte count, fever and elevated heart rate¹¹³. AF commonly occurs in patients with severe sepsis^114–116 and results in increased morbidity and mortality^117,118. Higher CRP levels occur before AF in patients with septic shock¹¹⁶, which suggests that systemic inflammation might trigger AF. In support of this hypothesis, hydrocortisone therapy was associated with a lower risk of AF and reduced production of pro-inflammatory cytokines in patients with septic shock^119,120.

In contrast to AF, third-degree atrioventricular block or severe arrhythmia affecting the ventricles can have acute and marked haemodynamic consequences, resulting in high mortality¹²¹ (TABLE 1). Patients presenting with ventricular tachyarrhythmias have a substantially increased mortality, regardless of the underlying disease^122–125. The electrophysiological origins of VT and VF include abnormal automaticity, triggered activity and re-entry^126–128. Ventricular arrhythmias have multifactorial origins and various combinations of derangements in ion channel distribution and expression patterns, intracellular ion dynamics, anatomical features and metabolic pathways (reviewed previously^129–131).

Myocarditis remains a major cause of SCD in young individuals¹³² and is usually caused by viral infections or autoimmunity leading to heterogeneous clinical manifestations, including life-threatening ventricular tachyarrhythmias¹³³. Acute myocarditis is associated with massive cardiac recruitment of macrophages and T lymphocytes, oedema and cell necrosis, all of which can contribute to electrical instability^134–137. Mild symptoms of infection, elevated leukocyte count and increased CRP levels often precede the onset of life-threatening ventricular arrhythmias and sudden cardiac arrest¹³⁸. Polymorphic and irregular ventricular arrhythmias are common during the active phase, whereas monomorphic, regular-shaped tachyarrhythmias are associated with healed myocarditis¹³⁹. The underlying mechanisms remain elusive, and hypotheses range from abnormal Ca²⁺ handling¹⁴⁰ to long-lasting, disturbed expression patterns of gap junctions¹⁴¹. A series of patients with viral myocarditis complicated by VF showed electrocardiographic features of Brugada syndrome, a genetic disorder caused by genetic variation in SCN5A, encoding the cardiac Na⁺ channel (Na_v1.5), and associated with ST-segment elevation and right bundle branch block¹²⁴. These patients showed electrocardiogram (ECG) changes that persisted after acute myocardial inflammation had subsided, suggesting structural and electrical remodelling of the ventricular substrate. For patients with Brugada syndrome, temperature-dependent alteration of ion-channel function might trigger ventricular arrhythmias during infection, indicating that fever might be arrhythmogenic¹⁴². This association aligns with previous findings linking febrile illness with infectious aetiology to increased vulnerability to VT and VF, not only in patients with Brugada syndrome^143–145 but also in patients without known repolarization abnormalities^146–148. Taken together, these data support the hypothesis that infection and fever might initiate life-threatening ventricular arrhythmias. Although no mechanistic data are available, we speculate that leukocytes recruited to the myocardium contribute to the occurrence of arrhythmia, especially in patients with myocarditis.

VF often occurs after MI^149,150. The onset of ventricular arrhythmias after MI follows certain temporal dynamics: the first 3 days constitute an acute phase, and a subsequent chronic phase is dominated by ventricular remodelling¹⁵¹. The ischaemia-induced cardiomyocyte death causes a rapid, massive infiltration of monocytes and neutrophils. The resulting increase in tissue heterogeneity might contribute to the markedly higher risk of ventricular arrhythmias early after MI^{122,152–155}, when the electrophysiological properties of the ventricular substrate change profoundly and in a layer-specific manner¹⁵⁶. The tissue heterogeneity in the peri-infarct region, which is particularly rich in leukocytes, favours arrhythmogenesis by forming re-entrant circuits with zones of slow conduction or complete block, leading to sustained monomorphic VTs^157,158. Very large, acute ischaemic zones might create the substrate for a transiently stable re-entrant circuit that can sustain a monomorphic re-entrant tachycardia¹⁵⁹. In subacute MI, the physiological substrate continues to change. Whereas macrophages and cardiomyocytes produce varying amounts of inflammatory cytokines and chemokines, such as IL-1, IL-6, TNF and CC-chemokine ligand 2 (REFs^160–162), cardiac fibroblasts release haematopoietic growth factors, such as granulocyte–macrophage colony-stimulating factor¹⁶³, and endothelial cells become activated, possibly also contributing to conduction heterogeneity in the ventricular substrate. During the subsequent development of myocardial scar tissue and inflammation-associated HF, ventricular arrhythmias and sudden cardiac arrest can occur¹⁶⁴. To address the increased risk of SCD, patients who experience VT or VF episodes >48 h after MI are implanted with defibrillators that effectively terminate life-threatening arrhythmias¹⁶⁵. Selection criteria for this treatment are weak, given that 80% of patients experiencing SCD do not meet the criteria for implantation of a defibrillator^166,167. Inflammatory biomarkers might improve risk prediction for ventricular arrhythmias and guide patient selection for device implantation, if further research documents a robust association between these biomarkers and SCD.

Inflammatory mechanisms

In accordance with the associations described above, emerging data suggest a causal relationship between inflammation and arrhythmia. We postulate that four distinct electroimmunological pathways can lead to arrhythmia: leukocyte release of cytokines that act on cardiomyocytes; altered electrotonic gap junction communication between conducting cells and leukocytes; leukocyte-instigated, insulating fibrosis; and autoimmune channelopathies.

Inflammatory mediators acting on cardiomyocytes.

The secretion of inflammatory cytokines by leukocytes can affect the capacity of cardiomyocytes to conduct properly. Clinical data associate arrhythmias with elevated levels of IL-1β^168,169, IL-6 (REFs^170–172), IL-8 (REF.¹⁷³), IL-10 (REF.¹⁷³), IL-17 (REFs^174–176) and TNF^168,169,177. Data obtained in mouse and dog models of AF indicate that macrophages, which are a major source of cytokines, have a causal role in the pathogenesis of AF; specifically, lipopolysaccharide-stimulated pro-inflammatory macrophages were shown to induce atrial electrical remodelling, increase AF inducibility and decrease atrial effective refractory period¹⁷⁸. Macrophage depletion with clodronate liposomes protected mouse atria from lipopolysaccharide-triggered electrical remodelling¹⁷⁸. Macrophage-specific Il1b knockout reversed lipopolysaccharide-mediated phenotypes in atrial myocytes¹⁷⁸. Macrophage-derived IL-1β also prolonged the APD and lowered K⁺ currents in diabetic mice, changes that increased the susceptibility to arrhythmia¹⁷⁹. IL-1β reduced L-type Ca²⁺ current density in neonatal mouse ventricular myocytes by activating protein kinase C¹⁸⁰. In agreement with these findings, IL-1 receptor inhibition raised conduction velocity and curtailed spontaneous and inducible ventricular arrhythmias in mice with acute MI¹⁸¹.

These preclinical data are particularly interesting when viewed together with the CANTOS trial¹⁸², which demonstrated that anti-inflammatory therapy targeting the IL-1β pathway with canakinumab led to significantly lower rates of recurrent cardiovascular events. IL-1β and IL-6, which is downstream of Il-1β, both affect cardiomyocyte Ca²⁺ regulation; although both cytokines prolonged the APD, IL-6 worsened Ca²⁺-mediated arrhythmia substrates more than IL-1β¹⁸³. Elevated circulating levels of IL-6 have been observed in both AF¹⁸⁴ and polymorphic VT¹⁸⁵. Data from humans suggest that increased circulating levels of IL-6 rapidly induce atrial electrical remodelling by downregulation of cardiac connexins¹⁸⁴. Anti-IL-6 treatment had beneficial effects in patients with rheumatoid arthritis and systemic inflammation, typically associated with long QT intervals and increased rates of SCD¹⁸⁶. Treatment with tocilizumab, an anti-IL-6 receptor antibody, shortened QTc intervals, which correlated with reduced CRP and TNF levels, suggesting a potential anti-arrhythmic effect of anti-IL-6 treatment.

Increased circulating levels of leukocyte-released IL-17A have been linked with AF¹⁷⁶. Moreover, IL-17A levels in the plasma were positively correlated with left atrial diameter, suggesting the functional relevance of IL-17 signalling in the pathogenesis of AF¹⁷⁴. This hypothesis is supported by preclinical studies demonstrating that IL-17 mediates inflammation by inducing increased cytokine release (IL-1β and IL-6) and collagen deposition, thereby shaping the arrhythmogenic substrate of the atria¹⁷⁵. IL-17 inhibition with an extract from Rhodiola crenulata suppressed atrial fibrosis, cardiomyocyte apoptosis and the incidence of ventricular arrhythmias in a dog model of HF¹⁸⁷, suggesting that targeting the IL-17 pathway has therapeutic potential.

Although we are only beginning to understand the multifaceted effects of leukocyte-released cytokines on conduction, these effects have already been partially shown for TNF, which alters the expression and distribution patterns of Cx40 and Cx43 in cardiomyocytes through the transforming growth factor-β (TGFβ)–SMAD signalling pathway, activates myofibroblasts, increases collagen deposition and ultimately shapes an arrhythmogenic substrate for AF^188,189. TNF might affect Ca²⁺ handling in cardiomyocytes by electrical remodelling, such as by decreasing sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase 2a expression^190–192. Finally, TNF induces cardiomyocyte apoptosis via cell death pathways¹⁹³, creating conduction heterogeneity (FIG. 2).

Fig. 2 |. Leukocyte-released cytokines shape the arrhythmogenic substrate.

Structural remodelling can be facilitated by leukocyte-released cytokines (such as tumour necrosis factor) by decreasing connexin (Cx) protein expression, hampering the intercellular conduction between cardiomyocytes and non-cardiomyocytes (such as leukocytes), ultimately affecting the cardiomyocyte action potential morphology, or by activating fibroblasts to become myofibroblasts, resulting in collagen deposition, shaping an arrhythmogenic substrate. Electrical remodelling by leukocyte-released cytokines usually refers to effects on ion channel expression (such as the sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase 2a (SERCA2a)), leading to abnormal Ca²⁺ handling in cardiomyocytes. Both structural and electrical remodelling increase conduction heterogeneity and arrhythmogenicity. PLN, phospholamban; RYR2, ryanodine receptor 2; SR, sarcoplasmic reticulum.

Upstream of cytokine release, nuclear factor-κB (NF-κB) signalling and the NLRP3 inflammasome cause cardiac arrhythmias. NF-κB facilitates transcription of IL-1, IL-6, IL-8 and TNF in response to damage-associated molecular patterns¹⁹⁴. A common mechanism for NF-κB signalling is the TLR4–MYD88 pathway^195,196. NF-κB, TLR4 and MYD88 protein expression increase in the atrial tissue of individuals with AF¹⁹⁷, while TLR4 activation also promotes cardiac arrhythmias by IRF3-dependent, but MYD88-independent, pathways¹⁹⁸. Moreover, NF-κB suppresses the transcription of genes encoding cardiac Na⁺ channels in response to oxidative stress, a hallmark of AF, and therefore also acts on electrical remodelling of the cardiac substrate¹⁹⁹.

NLRP3 inflammasome activation leads to atrial electrical and structural remodelling, frequent atrial ectopy and pacing-induced AF^200,201. Underlying mechanisms might include abnormal release of Ca²⁺ from the sarcoplasmic reticulum²⁰², shortening of the APD or atrial hypertrophy. Activation of the NLRP3 inflammasome and Ca²⁺/calmodulin-dependent protein kinase II signalling was evident in atrial cardiomyocytes from patients who subsequently developed postoperative AF compared with postoperative samples from patients without AF²⁰³. These inflammatory mediators sensitized cardiomyocytes to spontaneous release of Ca²⁺ from the sarcoplasmic reticulum and arrhythmogenic afterdepolarizations²⁰⁰.

Altered electrotonic gap junction communication between cardiomyocytes and leukocytes.

Similar to stromal cells, macrophages communicate with cardiomyocytes through Cx43-containing gap junctions^37,204,205. The functional relevance of heterocellular coupling has been indicated by synchronized electrical activity and electronic interactions between cardiomyocytes and stromal cells in the steady state^206–208 and in ventricular infarct border zones^209,210. In the diseased heart, Cx43 is expressed at lower levels and heterogeneously redistributed to the lateral sides of cardiomyocytes²¹¹. Reverse remodelling of reduced Cx43 expression can restore electrical stability²¹¹. In the steady state, macrophage coupling affects cardiomyocyte action potentials, producing a more positive resting membrane potential as well as a decrease in APD and refractory period³⁷. This contribution to physiological conduction might depend on macrophage number and phenotype, and we speculate that a change in these parameters might be arrhythmogenic. Data from mice with acute MI indicate that inflammation induced by lipopolysaccharide injections or atherosclerosis leads to inducible non-sustained VTs that were attributed to slower conduction²⁰⁵. In this model, macrophage expansion areas colocalized with regions of Cx43 degradation and decreased Cx43 expression²⁰⁵. In the setting of acute MI, resident cardiac macrophages, which stabilize physiological conduction³⁷, die alongside ischaemic cardiomyocytes²¹². We hypothesize that the loss of these cardiomyocyte-coupled resident macrophages might alter conduction. Recruited, phenotypically distinct monocyte-derived macrophages boost macrophage numbers by an order of magnitude in the ischaemic myocardium and simultaneously are likely to increase conduction heterogeneity.

In support of this hypothesis, a report suggested that cardiomyocyte–macrophage coupling occurs in MI border zones of mice and humans²¹³ (FIG. 3). Patients with arrhythmia after MI have more pro-inflammatory mononuclear cells in the circulation than patients with MI who do not have arrhythmia²¹³. In vitro macrophage co-culture with cardiomyocytes demonstrated that cardiomyocyte APD differs depending on the macrophage phenotype and their gap junction status²¹³. The concept of gap junction communication between cardiomyocytes and non-excitable cells has been promoted since the 1990s by Kohl and Noble^214,215 and is supported by a study testing whether lentivirus-mediated delivery of Cx43 into acute myocardial lesions increased electrical signal propagation²¹⁶. Gene therapy with Cx43 led to expression of Cx43 in (myo)fibroblasts and CD45⁺ leukocytes in the infarct region and prevented arrhythmia²¹⁶. Optical mapping of Cx43-overexpressing hearts revealed amplified conduction velocity within the scar, indicating that the therapy increased Cx43-mediated coupling between cardiomyocytes and non-cardiomyocytes²¹⁶.

Fig. 3 |. Macrophage–cardiomyocyte interactions.

a Macrophages are coupled to cardiomyocytes via connexin 43-containing gap junctions, which allow cation exchange that contributes to the steady-state cardiomyocyte action potential and improves atrioventricular node conduction. b | In the diseased heart, macrophages undergo phenotypic changes (for example, tissue-resident macrophages die and are replaced by monocyte-derived macrophages, which have distinct features) and are recruited in enormous numbers to the site of injury, affecting the cardiomyocyte action potential by modulating repolarization and conduction velocity and increasing conduction heterogeneity.

Taken together, an emerging body of evidence suggests that leukocytes directly participate in conduction in healthy and diseased hearts. In HF²⁶, myocarditis²¹⁷, sepsis²¹⁸ and acute MI¹⁵⁴, macrophage numbers increase and phenotypes are altered. Hypothetically, these changes could directly influence the action potential in cardiomyocytes, perhaps contributing to arrhythmia. Although we are only beginning to understand immune cell coupling to cardiomyocytes, we believe this field of research will reveal new therapeutic options for conduction disorders and arrhythmia.

Leukocyte-instigated, insulating fibrosis.

Leukocyte-mediated extracellular matrix deposition might disturb the capacity of cardiomyocytes to conduct properly because insulating extracellular matrix slows or even blocks conduction. Leukocytes orchestrate the formation of extracellular matrix in acute MI²¹⁹, HF²²⁰ and angiotensin II-induced cardiomyocyte hypertrophy²²¹. Macrophages modulate the extracellular matrix by releasing interleukins and TGFβ^222–224, thereby activating cardiac fibroblasts. The differentiation of fibroblasts into secretory and contractile myofibroblasts is a cellular hallmark of cardiac fibrosis, replacing dead cardiomyocytes with non-contractile scar tissue. Macrophages reside in close proximity to patchy fibrosis and myofibroblasts in ischaemic hearts, and their abundance increases in segments with contractile dysfunction^225,226. Indeed, the release of cytokines by macrophages might influence the activation and phenoconversion of cardiac fibroblasts into myofibroblasts^227,228, which deposit collagen to form scar tissue²²⁹. Myofibroblasts increase Cx43 expression in cardiac injury, which could have direct implications for heterocellular coupling between conducting cells²³⁰. Moreover, macrophages might themselves produce extracellular matrix²²⁹, although the relative importance of this direct contribution to fibrosis is unclear.

Only a few studies have addressed the possible causal relationship between leukocytes and AF via fibrosis. An experimental and clinical study of the haem enzyme myeloperoxidase (MPO), which is typically expressed by neutrophils, mechanistically demonstrated leukocyte involvement in the pathogenesis of atrial fibrosis and AF²³¹. Leukocyte-derived MPO generates hypochlorous acid, which modulates matrix metalloproteinase (MMP) activity and, therefore, extracellular matrix turnover²³². MPO-deficient mice were protected against the development of AF, which was reversed when MPO abundance was restored²³¹. Patients with AF had higher plasma concentrations of MPO and higher MPO levels in the right atria²³¹. Subsequent trials confirmed these results^233,234. Attenuated infiltration of neutrophils into the atria decreased atrial fibrosis and prevented AF episodes after angiotensin II infusion in CD11b-deficicent mice, providing further evidence for leukocyte-mediated fibrosis formation in AF²³⁵.

Additional research demonstrated a positive correlation between atrial fibrosis and higher numbers of macrophages and myofibroblasts in epicardial adipose tissue²³⁶, which produces inflammatory cytokines²³⁷ and has previously been linked to the pathogenesis of AF²³⁸. Left atrial collagen content correlated with the abundance of IL-6, MMPs and TNF in epicardial adipose tissue²³⁶. These data suggest that macrophages in the peri-atrial adipose tissue trigger atrial fibrosis and thereby shape the arrhythmogenic substrate²³⁶.

Macrophages also mediate the resolution of fibrosis by producing MMPs, removing apoptotic myofibroblasts or suppressing fibroblast activation^239–243. For instance, MMP7 is produced by macrophages²⁴⁴ and cardiomyocytes²⁴⁵ during the acute phase after MI, and this protease can cleave extracellular matrix substrates. Interestingly, deletion of Mmp7 in mice resulted in improved survival after MI and decreased the degradation of Cx43, which in turn influenced electrical remodelling of the ventricular substrate²⁴⁶. Although the mediating effects of leukocytes on fibrosis are well established in many settings, data on leukocyte-mediated fibrosis in arrhythmogenesis are mostly associative at this time, indicating a need for mechanistic studies.

Autoimmune channelopathies.

The absence of morphological substrates and ion channel mutations in some patients with SCD led to the recognition of arrhythmia caused by autoantibodies^247,248 that target β₁-adrenergic receptors²⁴⁹ or Na⁺, K⁺ or Ca²⁺ channels²⁵⁰. Cardiac inflammation and its resolution are believed to trigger the release of proteins (such as ion channels) that can become ‘self antigens’, provoking immune system humoral responses that cause plasma cells to generate autoantibodies. Given that potential self antigens are present in virtually unlimited supply, the results can be chronic autoimmunity and cardiac inflammation.

Long QT syndrome, an ECG abnormality with a prolonged repolarization phase (the QT interval) and extended APD, predisposes individuals to life-threatening tachyarrhythmias²⁵¹. This channelopathy arises from either genetic variants²⁵² or drugs that prolong the QT interval²⁵³. Antibodies reacting with K⁺ channels (K_v1.4) can induce an autoimmune version of the long QT syndrome^254,255 (FIG. 4a). By directly binding to a channel subunit, antibodies limit K⁺ currents (the rapid component of the delayed rectifier K⁺ current (I_Kr) or the transient outward K⁺ current (I_to)), thereby prolonging the APD²⁵⁴. Injecting guinea pigs with a peptide corresponding to an extracellular subunit of the hERG K⁺ channel produced autoantibodies that inhibited I_Kr and lengthened the QT interval²⁵⁶.

Fig. 4 |. Autoimmune channelopathies.

a | Autoantibodies released by plasma cells can either inhibit ventricular ion channels such as hERG and K_v1.4, producing long QT syndrome, or activate ventricular ion channels such as K_v7.1, producing short QT syndrome. b | In pacemaker cells, blockade of Ca²⁺ channel subunits by autoantibodies can generate atrioventricular block or sinus bradycardia.

Autoantibodies causing short QT syndrome in patients with dilated cardiomyopathy activate K_v7.1 K⁺ channels, increase the slow component of the delayed rectifier K⁺ current (I_Ks) and shorten the APD, which is associated with ventricular arrhythmias and AF^257,258 (FIG. 4a). Cardiomyocytes isolated from rabbits injected with patient serum containing antibodies targeting K_v7.1 had a shortened APD, increased I_Ks and an increased likelihood of arrhythmia²⁵⁸. Autoantibodies specific for Ro proteins (also known as Sjögren syndrome-related antigen A; SSA)^259,260, which are associated with systemic lupus and Sjögren syndrome, were found to be present in 60% of patients with torsades de pointes²⁶¹ and can also trigger congenital heart block^262,263. Congenital heart block is associated with the transplacental transfer of maternal anti-Ro/SSA antibodies to the embryo as early as week 11 of gestation, leading to atrioventricular block, sinus bradycardia, inflammation, calcification and fibrosis (also referred to as neonatal lupus)²⁶⁴. The major arrhythmic manifestation of congenital heart block is a third-degree atrioventricular block^265,266; other ECG abnormalities, such as a long QT interval, sinus node dysfunction and atrial flutter, have occasionally been reported. These conduction abnormalities in the fetal heart result from autoantibodies targeting L-type and T-type voltage-gated Ca²⁺ channel α-subunits (Ca_v1.2, Ca_v1.3, Ca_v3.1 and Ca_v3.2), causing inhibitory effects on the L-type Ca²⁺ channel current (I_CaL) and T-type Ca²⁺ channel current (I_CaT)^267,268 (FIG. 4b). A small clinical trial demonstrated that autoantibody-related second-degree congenital heart block can be ameliorated to first-degree congenital heart block or even to normal atrioventricular conduction with the use of a combined anti-inflammatory therapy involving plasmapheresis, intravenous immunoglobulins and betamethasone²⁶⁹.

Some patients with idiopathic atrioventricular block have autoantibodies against Na_v1.5 (REF.²⁷⁰). These autoantibodies inhibit Na⁺ channel function and consequently decrease I_Na currents but can also lower protein expression levels of the Na_v1.5 channel. Profiling autoantibody signatures in patients with idiopathic cardiac arrest identified autoantibodies against the L-type Ca²⁺ channels as a biomarker for sudden cardiac arrest, whereas autoantibodies against K⁺ channel subfamily K member 2 (KCNK2; also known as TREK1) were found in both patients with ischaemic cardiac arrest and healthy control individuals²⁷¹. Autoantibody-induced arrhythmias have been reviewed in detail previously²⁵⁰.

Immune modulation and arrhythmia

Twentieth century research into electrophysiology gave rise to implantable defibrillators, anti-tachycardiac pacing terminating life-threatening VTs and VFs, catheter ablation of arrhythmic foci and antiarrhythmic drugs. However, arrhythmia-associated mortality remains high, and the use of some channel inhibitors was abandoned due to pro-arrhythmic and other adverse effects. We speculate that anti-inflammatory therapeutics might help to overcome these problems. Emerging data indicate that this approach is promising, such as the effects of low-dose hydrocortisone on AF in patients with sepsis¹¹⁹, IL-1 receptor inhibition in mice with acute MI¹⁸¹ or antibodies against KCNQ1 that suppress arrhythmic activities in ex vivo models of long QT syndrome²⁷². We concede that this development is still in its early stages, which is unsurprising given that cardiac immune cells were virtually unknown a decade ago. The next steps involve carefully dissecting the hypothetical inflammatory pathways that lead to arrhythmia, followed by selecting immune targets that are safe — that is, inhibiting them would not compromise host immune defence. The increasing availability of high-resolution data, especially single-cell RNA-sequencing data sets, on cardiac immune cells will aid this development. Fortunately, immunomodulatory therapy has been successful in many diseases, including cancer, leading to optimism that similar objectives can be accomplished for cardiovascular disease and cardiac arrhythmias. Particularly promising platforms include neutralizing antibodies against cells or secreted proteins¹⁸², nanoparticle drug delivery to phagocytic cells such as macrophages²⁷³ and liposome-enabled RNA interference²⁷⁴. Cell therapy might also be an option, as indicated by data on CD8⁺ T cells targeting fibroblast activation protein, which reduced cardiac fibrosis in mice²⁷⁵. On the basis of these and other studies, we believe understanding the role of immunity in physiological conduction and exploring electroimmunological contributions to arrhythmias will provide a foundation for the development of next-generation drugs to correct abnormal conduction. Of note, immune cell-targeted therapeutics can also elicit unwanted pro-arrhythmic adverse effects. For instance, immune checkpoint inhibitors can lead to myocarditis and arrhythmia^276,277.

Conclusions

Growing preclinical and clinical evidence demonstrates that immunity has a role in the pathophysiology of arrhythmia. Emerging findings obtained by understanding leukocyte electrophysiology provide insights into non-canonical leukocyte function associated with physiological and abnormal conduction. By extrapolating findings on the interaction between microglia and neurons, we suspect that cardiac macrophages might also have additional, currently unknown functions modulating the cardiac conduction system. Targeting inflammatory processes might be central to preventing and treating arrhythmias, but comprehensively understanding leukocyte heterogeneity, immune cell subtypes and their functions is necessary for the development of future therapeutic strategies for arrhythmia and SCD.

Key points.

• Immune cells express various ion channels that influence their phenotypes and functions.

• Numerous leukocytes reside in the normal myocardium; their numbers, phenotypes and electrophysiological properties change in pathologies that give rise to arrhythmia, including acute myocardial infarction, sepsis, heart failure and myocarditis.

• Macrophages are the most abundant leukocytes in the heart and electrotonically couple to cardiomyocytes via connexin 43-containing gap junctions; this sink–source relationship leads to rhythmic macrophage depolarization and modulates the resting membrane potential and action potential of cardiomyocytes.

• Leukocytes might contribute to rhythm disorders either directly through altered coupling or indirectly by influencing cardiomyocytes and their environment.

• Indirect pro-arrhythmic leukocyte actions include production of cytokines and antibodies, which act on cardiomyocytes and change tissue properties by instigating fibrosis.

• Immunotherapy is an emerging option for the treatment of rhythm disorders such as atrial fibrillation.