Tissue-resident macrophages have been observed in various organs, including the heart, brain, liver, and lung, and originate from the embryonic lineage, which is different from that of monocytic progenitors (Ginhoux et al., 2010; Davies et al., 2013; Moore et al., 2013). Cardiac macrophages maintain a homeostatic population through their self-proliferative properties and are independent of blood monocyte-derived macrophages. Recently, in steady-state conditions, two resident cardiac macrophage subsets, MHC-IIlowCCR2- and MHC-IIhighCCR2- cells, were identified by gene fate-mapping techniques. Under cardiac injury conditions, a third cardiac macrophage population, MHC-IIhighCCR2+ cells, was identified in the heart (Epelman et al., 2014; Honold and Nahrendorf, 2014). Cardiac CCR2− macrophages originate from the primitive yolk sac and are replenished through local proliferation, whereas CCR2+ macrophages originate from bone marrow-derived monocytes and repopulate through monocyte recruitment and proliferation. Resident CCR2− macrophages are involved in angiogenesis and cardiomyocyte proliferation. Bajpai et al. (2018) demonstrated that the human myocardium is populated by distinct subsets of CCR2− macrophages, CCR2+ macrophages, and CCR2+ monocytes. Subsequently, Bajpai et al. (2019) showed that the depletion of resident cardiac CCR2- macrophages in a murine model of myocardial infarction increased the infarct area, reduced left ventricular (LV) systolic function, and aggravated LV remodeling. Interestingly, a recent study published in Cell demonstrated that resident macrophages in the steady-state heart facilitated electrical conduction, thus highlighting a novel concept regarding the potential role of cardiac macrophages in modulating cardiac electrical function (Hulsmans et al., 2017).

Under myocardial inflammatory conditions, monocyte-derived macrophages are recruited to the heart and characterized as MHC-IIhighCCR2+ cells. Mouse blood monocytes have been divided into Ly6C^hi and Ly6C^low (Robbins et al., 2012; Yap et al., 2019). Ly6C^hi monocytes induce excessive monocytosis, accumulate in the injured area, and differentiate into macrophages. Ly6C^low monocytes, derived from pro-inflammatory Ly6C^High cells, are less recruited than their Ly6C^High counterparts following MI. Ly6C^low monocytes are responsible for patrolling and tissue injury repair. However, human peripheral blood monocytes have 3 phenotypes: proinflammatory CD14⁺⁺CD16^– monocytes, anti-inflammatory CD14⁺CD16⁺⁺ monocytes with a function similar to Ly6C^low monocytes, and proinflammatory CD14⁺⁺CD16⁺ monocytes secreting TNF-α (Heidt et al., 2014; van der Laan et al., 2014).

Macrophages are extremely heterogeneous and show adaptation of the phenotype and functions according to the surrounding microenvironment and aging (Gosselin et al., 2014; Pinto et al., 2014). Macrophage activation produces distinct functional phenotypes that are most commonly categorized as classically “inflammatory” macrophages and alternatively “anti-inflammatory” macrophages. This is a basic delineation of macrophage phenotypes but an overly simplistic view of macrophage behavior. Classically activated macrophages have proinflammatory properties, whereas alternatively activated macrophages are linked to cell proliferation and tissue repair (Nahrendorf et al., 2007; Biswas and Mantovani, 2010; Italiani and Boraschi, 2014). Classically activated macrophages secrete proinflammatory cytokines such as interferin-1β (IL-1β), IL-6, IL-12, tumor necrosis factor-α (TNF-α), and matrix metalloproteinase (MMP) and chemokines, which play a key role in host defense. Alternative activated macrophages produce cytokines, including transforming growth factor-β (TGF-β), IL-10, enzyme arginase-1 in mice (ARG 1) and chemokines, which are involved in collagen formation and tissue repair.

After myocardial infarction (MI), macrophages are abundant in the infarcted area. Ly-6C^High monocytes from the bone marrow and spleen are recruited to the infarcted zone and then differentiate into macrophages. The levels of this inflammatory subtype reach a peak approximately 3 days after injury (Heidt et al., 2014). Between days 5 and 7, macrophage populations are at their maximum within the infarct (Nian et al., 2004; Libby, 2013). Classically activated macrophages dominate the cell population of the infarcted zone. Classically activated macrophages produce proinflammatory cytokines, enhance the proinflammatory response and facilitate the breakdown of collagen. The reparative phase is characterized by phenotypic transition from inflammatory monocytes and macrophages (Ly-6C^High monocytes and classically activated macrophages) to the anti-inflammatory subtypes (Ly-6C^low monocytes and alternatively activated macrophages) (Nahrendorf et al., 2007). The pool of cardiac macrophages is replenished as Ly-6C^low monocytes are extensively recruited to the infarcted area. Ly-6C^low monocytes mainly derive from pro-inflammatory Ly-6C^High monocytes (Hilgendorf et al., 2014). After accumulation, Ly-6C^low monocytes are thereby differentiated into alternatively activated macrophages. Alternatively activated macrophages release IL-10, which inhibits the proinflammatory effects of classically activated macrophages, and TGF-β, which promotes tissue remodeling and angiogenesis (Lavine et al., 2014). Fei et al. (2019) demonstrated that activated macrophages directly connected to cardiomyocytes, thereby prolonging the action potential duration (APD), and ultimately led to APD heterogeneity and post-MI arrhythmias via gap junctions. This finding suggests that macrophages directly participate in ventricular arrhythmias after myocardial injury.

Previously, several clinical studies provided evidence that increased inflammatory cytokines are closely associated with cardiac arrhythmias (Lewek et al., 2014). These results suggested that inflammation affects the initiation and progression of VAs. The proarrhythmic effects involve substrate-triggered cardiac electrical and structural remodeling. Inflammation contributes to the occurrence of ectopic-triggered activity and re-entry (Wakili et al., 2011; Dobrev et al., 2012; Aulin et al., 2015).

Electrophysiological changes (ion channel disturbance, early and late afterdepolarizations), as well as gap junction remodeling and enhanced myocardial fibrosis, are immune-related mechanisms responsible for cardiac arrhythmias. Macrophage-dependent and macrophage-independent inflammation, including cytokine processes, serves as the basis of proinflammatory-induced VAs. Inflammation in the heart can also directly result in fluctuations in membrane potential. VAs can be triggered by early afterdepolarizations (EADs) and delayed afterdepolarizations (DADs). EAD results from the reduced function of potassium channels or the increased function of calcium or sodium channels. Abnormal intracellular Ca²⁺ handling, such as sarcoplasmic reticulum (SR) overload and uncontrolled Ca²⁺ leak, contributes to DAD (Kao et al., 2010; Chilukoti et al., 2013). Furthermore, gap junctions (GPs) are cell-to-cell pathways mediating electrical and chemical signal exchange between adjacent myocytes. GPs can transmit an orderly wave of electrical excitation. In the ventricles, gap junction remodeling, including connexin43 (Cx43) reduction, Cx43 dephosphorylation and Cx43 lateralization in pathological conditions, produces arrhythmia substrates (Duffy, 2012). The inflammatory response in the local heart area may play a crucial role in GP remodeling. Finally, excessive fibrosis or cardiac sarcoid infiltrated with abundant macrophages produces VAs not only by the mechanism of triggered substrate but also by re-entry (Okada et al., 2018). On the one hand, Haider et al. (2019) showed that cardiac macrophages can develop a fibroblast-like phenotype and directly contribute to the formation of fibrosis after myocardial infarction. On the other hand, macrophage-derived cytokines activate fibroblasts and produce cardiac fibrosis (Jung et al., 2017; Shimodaira et al., 2018; Abe et al., 2019). Cardiac fibrosis complicates electrical impulse propagation, slows conduction velocity, and forms unidirectional conduction blocks (Rohr et al., 1997; Rohr, 2012). Inflammation is also related to tissue repair after injury. Tissue repair is accompanied by parenchymal cell regeneration and finally fibrous tissue formation, namely, scar formation (Klein et al., 2000; Said et al., 2011; Gorenek et al., 2014; Vonderlin et al., 2019).

Sympathetic overactivity and structural remodeling play a critical role in ventricular arrhythmogenesis (Podrid et al., 1990; Shen and Zipes, 2014; Kalla et al., 2016). In an experimental study, sympathetic nerve stimulation caused a change in ventricular electrophysiology, reduced the ventricular fibrillation threshold and triggered Vas (Mantravadi et al., 2007; Ng et al., 2009). Although it is difficult to induce VAs in mice, norepinephrine (NE) injected into the epicardial tissue of guinea pigs elicits triggered automaticity, and computational modeling identified a Ca²⁺ overload mechanism in cardiac ventricular electrophysiology. This finding supports the hypothesis that heterogeneity or gradients of sympathetic activation are proarrhythmic (Vaseghi et al., 2012). A growing body of work has demonstrated that macrophages can contribute to sympathetic hyperactivity. Levick et al. (2010) showed that substance P released by sympathetic afferent fibers could bind to the neurokinin-1 receptor of macrophages to induce the production of macrophage-derived angiotensin II. Angiotensin II could further stimulate the terminus of sympathetic efferent fibers and then increase the production of norepinephrine. Thus, macrophages play a critical role in mediating VAs relevant to sympathetic activity by enhancing the production of angiotensin II (Levick et al., 2010). Furthermore, peripheral proinflammatory factors produced by macrophages, such as IL-1β, IL-6, and TNF-α, can transmit signals to the brain by the circulation or through afferent fibers, which in turn activate the sympathetic nervous system (SNS).

A proarrhythmic substrate is usually formed by regional myocardial remodeling, as well as heterogeneity of sympathetic innervation (Barber et al., 1983; Tomaselli and Zipes, 2004). The heterogeneity of sympathetic innervation is called nerve sprouting (Cao et al., 2000; Chen et al., 2001). Emerging data have indicated that cardiac sympathetic sprouting and cardiac electrical remodeling are involved in the post-MI remodeling process (Zipes and Rubart, 2006; Wang et al., 2012). Sympathetic nerve sprouting may produce electrical and structural remodeling following AMI, resulting in electrophysiological instabilities and finally inducing VAs. In fact, studies have shown that the inhibition of sympathetic nerve sprouting induced by MI can exert antiarrhythmogenic effects (Wernli et al., 2009; Yang et al., 2016; Yin et al., 2016; Hu et al., 2019). Sympathetic nerve remodeling is a complex pathophysiological process, and sympathetic nerve sprouting is closely associated with the inflammatory reaction and is primarily present at the infarct border zone, where abundant macrophages and macrophage-derived cytokines are observed. Macrophages could promote sympathetic hyperinnervation via the regulation of nerve growth factor (NGF) expression (Yin et al., 2016). Atorvastatin could effectively improve cardiac sympathetic nerve remodeling by modulating macrophage polarization (Yang et al., 2016). It was recently demonstrated that inhibiting miR-155 can downregulate NGF expression by decreasing M1 macrophage polarization, subsequently impairing sympathetic nerve remodeling and VAs induced by acute myocardial infarction (Wernli et al., 2009; Hu et al., 2019).

Sympathetic nervous system can affect macrophages at both the systemic and regional levels (Chen et al., 2016). SNS fibers innervate the primary and second lymphoid organs (bone marrow, thymus, spleen, and lymph nodes) and are capable of modulating immune functions. The SNS contributes to the differentiation, maturation, recruitment, and regulation of macrophages. Moreover, the SNS directly innervates the target lesion, plays a proinflammatory role and promotes M1 polarization. Norepinephrine released from sympathetic nerve endings could bind to α- or β-adrenergic receptors expressed on immune cells (T cells, B cells, natural killer cells, and macrophages). This response gives rise to a cascade of events, including the production of proinflammatory cytokines and recruitment of leukocytes. Therefore, these changes might form a vicious circle between SNS activity and M1 polarization, both contributing to ventricular arrhythmogenesis.

Macrophages secrete multiple cytokines. Proinflammatory factors produced by macrophages, such as IL-1β, IL-6, TNF-α, and matrix metalloproteins (MMPs), can regulate cardiac SNS activity, form a proarrhythmic substrate and directly affect myocardial electrophysiology (Hirayama et al., 2017). VAs might be induced by cytokines derived from macrophages in both acute and chronic diseases. It was shown that cytokines derived from macrophages following acute MI could target cardiac myocytes and induce electrophysiological remodeling, including a reduction in repolarizing K⁺ currents, Cx43 expression and intracellular Ca²⁺ mishandling (Pinto and Boyden, 1999; Francis Stuart et al., 2016). These changes could provide the trigger and substrate for ventricular arrhythmias (De Jesus et al., 2017). In the case of chronic inflammation, a growing body of clinical evidence has demonstrated that the serum concentration of macrophage-derived cytokines was significantly higher in post-MI patients with ventricular arrhythmias than in post-MI patients without ventricular arrhythmias (Streitner et al., 2007, 2009; Francis Stuart et al., 2016). Furthermore, Monnerat et al. (2016) experimentally demonstrated that IL-β derived from cardiac macrophages could trigger ventricular arrhythmias in mice. Even in the absence of cardiac injury, systemic inflammation was found to be related to an increased risk for ventricular arrhythmias. Extracardiac injury was also shown to enhance macrophage-related inflammation in the heart. VAs are also one of the most critical complications after renal or brain injury. Renal ischemia reperfusion increased susceptibility to ventricular arrhythmias depending on activation of NLRP3-CASP1-IL-1β; accordingly, this effect was inhibited by macrophage depletion (Alarcon et al., 2019). Rheumatoid arthritis is a chronic inflammatory disease, and increased serum concentrations of macrophage-derived cytokines were detected. Increased susceptibility to VAs was also found in patients with rheumatoid arthritis (Lazzerini et al., 2017a). The role of individual cytokines on VAs is listed below in detail (Figure 1).

IL-1β contributes to electrical function not only through direct effects but also indirect effects on VAs. Wang et al. confirmed that IL-1β injection into the left stellate ganglion (LSG) could increase sympathetic activity and the occurrence of VA. IL-1β injection induced cardiac electrical remodeling, and this response was attenuated by IL-1Ra preinjection (Wang et al., 2017). Cardiac fibrotic substrate and Cx43 remodeling are important mediators responsible for the heterogeneity in ventricular conduction for reentry. Considerable experimental data have shown that excess fibrosis promotes ectopic triggers in the hearts of aged rats and rabbits (Bapat et al., 2012; Wang et al., 2017). This kind of proarrhythmic substrate produced early afterdepolarizations, triggered activity and reduced conduction velocity (Bapat et al., 2012). Previous study has shown that IL-1β plays an important role in the formation of fibrotic substrates and is implicated in Cx43 remodeling, induced cell–cell uncoupling, lateralization, and degradation (Baum et al., 2012).

Thus, macrophage-derived IL-1β during myocardial healing could induce deleterious electrophysiological consequences.

In addition to antigen presentation, phagocytosis and immunomodulation, cardiac macrophages have been proven to directly induce cardiac electrophysiological changes. This finding substantially changed our previous knowledge of macrophage function. Previous studies have shown that the transmembrane potential is −26 mV for mouse macrophages and −18 mV for guinea pig macrophages (Dos Reis et al., 1979). Macrophages present several different types of K⁺ and Cl^– channels in the membrane. Membrane potential and ionic conductance regulate many cell functions, including transmembrane signaling, phagocytosis, secretion, and motility (Udagawa, 2003; Pinto, 2017). However, previous data did not show that macrophages could directly modulate cardiac electrophysiology.

A recent study reported by Hulsmans et al. (2017) demonstrated that cardiac resident macrophages are integral for normal heart rhythm. These researchers found that cardiac macrophages are highly abundant at the atrioventricular node (AVN) in mice and humans. They found that macrophages connect to cardiomyocytes at the AVN via Cx43, forming punctate junctions between macrophages and cardiomyocytes. Cx43-mediated macrophage-CM coupling served as electrical coupling of the two cell types. In an in vitro study, cocultures of macrophages and mouse neonatal AVN CMs also established Cx43 coupling between the two cell types. The researchers further showed that depolarization of macrophages improved AVN conductance. However, depletion of macrophages resulted in AVN block, resulting in arrhythmia.

In the injured heart, the number of cardiac macrophages is significantly increased compared with that in the uninjured heart. Considering that resident macrophages might participate in cardiac electrophysiology through Cx43 between macrophages and myocytes, it would be interesting to investigate whether monocyte-derived macrophages recruited to the myocardium under inflammatory conditions also connect to myocytes by Cx43, affecting cardiomyocyte electrophysiology and resulting in VAs. Fei et al. (2019) found that proinflammatory macrophages formed gap junctions with cardiomyocytes and accumulated in MI border zones 3 days post-MI. These researchers further demonstrated that non-inflammatory macrophages connected with myocardiocytes could shorten APD₉₀, yet proinflammatory macrophages could prolong APD₉₀ in an in vitro coculture experiment. This finding indicates that the Cx43 connection between macrophages and cardiomyocytes leads to APD heterogeneity and post-MI arrhythmias. Therefore, targeting macrophage-CM coupling could be a potential useful target when treating inflammation-associated conduction abnormalities.

Macrophages appear to play direct and indirect roles in the occurrence of VAs. Activation of cardiac macrophages induces VAs through sympathetic nerve sprouting, proinflammatory cytokine production, and the direct influence of cardiac electrophysiology. Previously, inflammation has always been treated as an epiphenomenon and not suitable as a target for intervention. However, based on experimental and clinical evidence, anti-inflammatory therapy targeting the inflammatory cytokines contributed to a benefit from cardiovascular diseases (Ridker et al., 2017; Wang et al., 2017). Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS) has been provided the evidence that IL-1β treated as a target and reduced major cardiovascular events. It opens a landscape of cardiovascular diseases therapy (Ridker et al., 2017). Therefore, immunotherapy for therapeutic interventions, targeting these functional elements of macrophages and cytokines blockade, are likely to be a promising new research avenue and might be valuable for the purpose of developing new therapeutics to reduce VAs.

MC, XL, and SW participated in the study design and drafted the manuscript. SZ and MC responsible for writing the manuscript. LY, JT, and SZ participated in the overall editing and approval of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.