The immediate cause of pro-arrhythmic events involves actions of, and interactions between, cell membrane ion channels within individual or between adjacent cardiomyocytes. However, these are influenced by a hierarchy of mechanisms. These extend over a 12-decade, logarithmic timescale. This ranges from ns/μs (∼10^–6 s) timescales shown by molecular events, through cellular events extending over ms/s within each or multiple cardiac cycles, to hours/days or even weeks/months (∼10⁶ s) shown by remodeling processes (Figure 5A). Thus, there are (Figure 5B) (a) cell surface membrane ion channels and transporters underlying automaticity and AP excitation, propagation, and recovery and the (b) cellular feedforward and feedback effects of excitation–contraction coupling and its triggering by Ca²⁺. These latter overlap over microsecond/millisecond time scales. In contrast, the (c) systems-level G-protein signaling-dependent autonomic inputs and their related central nervous system circadian rhythms extend over second/minute/hour time scales. Finally, (d) longer-term upstream mechanisms related to metabolic feedback and other upstream modulators extend over days/weeks/months. Nevertheless, each of these processes can be grouped (Figure 5C) by their participating biomolecules (Figure 5D) (Lei et al., 2018). The scheme complements and prompts extensions of existing modeling attempts that extend to full electromechanical coupling at the systems level. These had integrated local and transmural ventricular myocyte surface electrical (ten Tusscher et al., 2004) with the cytosolic and SR Ca²⁺ transport and storage properties outlined below. They also modeled the consequent Ca²⁺–troponin binding to cross-bridge cycling activity and myofilament mechanics (Rice et al., 2008). These finally extended to mechanical activity in realistic two- and three-dimensional ventricular cardiac tissue models (Pathmanathan and Whiteley, 2009; Adeniran et al., 2013; Huang, 2015).

Many clarifications of these relationships arise from monogenically modified murine models permitting single cardiomyocyte, tissue, whole organ, and systems-level experimental study (Huang, 2017). Murine hearts translate to human hearts in their two-sided atrial and ventricular circulations, pacing, or conducting SAN, AVN, and atrioventricular (AV) components. They differ in size, heart rates, some ion channel types (Figures 2Ba, b), and consequent detailed AP waveforms (Figures 2c, d). Nevertheless, they share major, depolarization and conduction, AP features and regional heterogeneities (Huang, 2017). Particular genetic variants recapitulate corresponding human clinical arrhythmic and pharmacological phenotypes. Genetically induced pluripotent stem cell-derived cardiomyocyte (iPSC-CM) platforms also show promise as cellular rather than systems models in not recapitulating in vivo conducting, Purkinje, and contractile tissue organization. However, many human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) show immature embryonic-like rather than adult atrial/ventricular phenotypes (Lee et al., 2012). Nevertheless, recent reports describe hiPSC-CMs with atrial AP properties and acetylcholine-activated K⁺ current expression (Ahmad et al., 2023). There are recent iPSC models for normal and disease changes in ion channel expression, Ca²⁺ homeostasis, neurocardiac interactions, and hypertrophy (Li et al., 2023; Chen et al., 2023; Langa et al., 2023; Zhou et al., 2023). Theoretical reconstructions remain useful in predicting physiological end effects (Alrabghi et al., 2023; Hancox et al., 2023). Ultimate translatability of such results still requires direct human clinical electrophysiological studies with recently innovated electrocardiographic (Van Der Werf and Lambiase, 2021) and electrical mapping capacities (Taggart et al., 2022).

A number of associated proteins modify function and expression of surface membrane ion channels in short local feedback loops. First, auxiliary β-subunits modulate function in their various associated Na⁺, Ca²⁺, and K⁺ channel subtypes. Na⁺ channel β-subunits (Navβ) each comprise an amino-terminal immunoglobulin (Ig) connected to a single transmembrane domain and a short largely disordered intracellular region. They enhance channel trafficking to the surface membrane modifying peak I _Na, modulate steady-state and kinetic inactivation (Namadurai et al., 2015), and ion channel clustering in the plasma membrane (Salvage et al., 2020).

The Navβ₁ and Navβ₃ isoforms associate non-covalently with α-subunits, mostly with the Ig domain binding to the domain III voltage-sensing module (VSM); the Navβ₂ and Navβ₄ Ig domains covalently associate via a disulfide bond to an extracellular loop region of domain II (Salvage et al., 2020a). However, Nav1.5 constitutes an interesting exception to this rule. Here, the Navβ₁ and Navβ₃ subunit Ig domains cannot bind to domain III VSM because the binding site is likely blocked by glycosylation. Similarly, Navβ₂ and Navβ₄ Ig domains cannot form disulfide bonds because the necessary free cysteine residues are absent in the Nav1.5 α-subunit. Consequently, Nav1.5-associated β-subunit Ig domains may be free to form extended cis and trans cross-links. Here, Navβ₁-subunits may interact in trans influencing cell proximity. In contrast, cis-Navβ₃ interactions may enhance local clustering of Nav1.5 (Salvage et al., 2020a). These may enhance ephaptic cell-to-cell AP conduction at intercalated disc perinexal membranes (Salvage et al., 2020a; Salvage et al., 2023). Genetic β-subunit mutations produce pro-arrhythmic phenotypes resembling conduction defects shown by Scn5a ± arrhythmic models (Hakim et al., 2008; Hakim et al., 2010).

Of K⁺-channel-associated β-subunits, KCNE2-encoded minK-related protein 1 (MiRP1) is associated with many K⁺ channel α-subunits. It modifies steady-state and kinetic activation and deactivation properties of HERG K⁺ channel-mediated I _Kr (Lu et al., 2003). Similarly, KCNE1-encoded β-subunits are associated with I_Ks channels. Their genetic ablation causes abnormal AP recovery, increasing APDs and inverting transmural APD gradients (Thomas et al., 2007; Hothi et al., 2009), and pro-arrhythmic phenotypes including early afterdepolarizations (EADs) and spontaneous VT in murine hearts recapitulating human Jervell and Lange–Nielsen syndrome (LQTS5) (Balasubramaniam et al., 2003).

Second, ion channels interact with cell adhesion, signal transduction, and cytoskeletal anchoring proteins. Nav1.5 interacting proteins may also function in specific subcellular Nav1.5 localizations producing regional variations in Nav1.5 expression, I _Na, density, and kinetics between subcellular microdomains. At lateral cardiomyocyte membranes, Nav1.5 interacts with dystrophin–syntrophin complexes, calcium/calmodulin (CaM)-dependent serine protein kinase, and caveolin-3. At intercalated disks, Nav1.5 associates with N-cadherin, connexin-43, β_IV-spectrin, and desmosomal proteins including plakophilin-2 and desmoglein-2. Gene mutations involving some of these proteins are associated with arrhythmic disorders suggesting Nav1.5 dysfunction (Shy et al., 2013; Remme, 2023).

Excitation–contraction coupling refers to the feedforward events connecting Nav1.5-mediated surface and transverse (T-) tubular membrane depolarization by AP initiation and propagation to initiation of contraction. Tubular Cav1.2 channels sense triggering surface membrane voltage changes, transducing these into ryanodine receptor (RyR2)-mediated SR Ca²⁺ store release by Ca²⁺-induced Cav1.2-RyR2 coupling (Endo, 2009). This likely occurs at T-SR junctional regions where the relevant membranes assume geometrically close ∼100–400 nm parallel alignment while remaining electrically separate (Martin et al., 2003; Dulhunty, 2006). Nav1.5, Cav1.2, and RyR2 come into close proximity within the resulting restricted diffusion spaces permitting local ion accumulation or depletion (Bardsley et al., 2021). Local Ca²⁺ release might then significantly elevate [Ca²⁺]_TSR (Saucerman and Bers, 2012; Despa et al., 2014; Sanchez et al., 2021), in turn elevating bulk cytosolic [Ca²⁺]i-activating troponin, thereby triggering myofilament action and cardiomyocyte contraction. The background cytosolic [Ca²⁺]_i is ultimately restored by SR Ca²⁺-ATPase (SERCA2), which mediates the Ca²⁺ return to its SR store and expulsion into the extracellular space by sarcolemmal NCX and, to a lesser extent, sarcolemmal Ca²⁺-ATPase. Further intracellular organelles, including lysosomes and mitochondria, are often regulated by further signaling molecules, such as cADP-ribose, nicotinic acid adenine dinucleotide phosphate (NAADP), and inositol tris-phosphate (IP₃), which also modulate cellular Ca²⁺ homeostasis (Terrar, 2023).

Feedback actions of consequently altered intracellular Ca²⁺ fit its strategic second messenger function through widespread cellular processes and cell types (Figure 6A) (Huang, 2017). Released Ca²⁺ itself exerts a short-loop slow RyR inactivation (Laver and Curtis, 1996). Further longer-loop feedback modulation may modify surface membrane channel or transporter activity with pro- or anti-arrhythmic effects (Figures 6B–D), acting on either arrhythmic triggering or substrate (Figure 6E). Of voltage-sensitive channels mediating inward depolarizing current, Na⁺ and Ca²⁺ channel C-terminal domains contain both direct and indirect Ca²⁺ regulatory binding sites in the form of one or more EF hand motifs and an isoleucine–glutamine (IQ) CaM-binding domain, respectively. Their III–IV loops also contain Ca²⁺-CaM-binding sites. The channels additionally possess multiple kinase, including CaM kinase II (CaMKII) and phosphokinase C (PKC), and phosphorylatable regulatory sites. One or more of these sites have been implicated in Ca²⁺-mediated feedback inhibition of Nav1.5 in intact cardiac myocytes and implicated in arrhythmogenic phenotypes in both pharmacological (King et al., 2013b; Valli et al., 2018b) and genetic disease models (King et al., 2013c; Zhang et al., 2013; Salvage et al., 2021; Salvage et al., 2023). Similar mechanisms may operate for skeletal muscle Nav1.4 (Matthews et al., 2019; Sarbjit-Singh et al., 2020; Liu et al., 2021). Ca²⁺, likely through CaMKII, conversely increases I _NaL with potential pro-arrhythmic effects (Liu et al., 2023).

Of ion exchange processes, NCX stoichiometry produces inward, potentially pro-arrhythmic I _NCX under conditions of elevated [Ca²⁺]_i in polarized membranes (Shattock and Bers, 1989).

Of voltage-sensitive channels mediating outward hyperpolarizing current, Ca²⁺-activated K⁺ channel opening contrastingly hyperpolarizes, thereby reducing membrane excitability. In addition to membrane depolarization, large conductance Big K⁺ (BK) channel opening requires μM local [Ca²⁺] binding to two distinct high-affinity intracellular C-terminal domain Ca²⁺-binding sites (Zeng et al., 2005; Yang et al., 2015; Sancho and Kyle, 2021), modifying sino-atrial node firing rate and cardiac pacing (Meredith et al., 2014; Pineda et al., 2021). Voltage-insensitive small conductance (SK) SK1, SK2, and SK3 K⁺ channels (Adelman et al., 2012; Weisbrod, 2020) respond to RyR-induced [Ca²⁺]_i through a regulatory C-terminal CaM-binding domain (Neelands et al., 2001; Li et al., 2023). SK2 channels, in particular (Schmitt et al., 2014), are clinically implicated in atrial fibrillation (Healey et al., 2005) and arrhythmia and hypertrophy accompanying clinical and experimental ventricular failure (Terentyev et al., 2014; Gu et al., 2018). SK2 modified AP waveform in pro-hypertrophic Pak-1-deficient murine experimental platforms predisposing to ventricular hypertrophy and failure following angiotensin II challenge (Yang et al., 2021). Finally, of anion channels, Ca²⁺-activated Cl⁻, TMEM16A, channels are [Ca²⁺] sensitive (Yang and Colecraft, 2016).

Autonomic, sympathetic and parasympathetic, modification of pacing, ion current activation, and excitation–contraction coupling involves transmitter and co-transmitter binding to G-protein-coupled receptors (GPCRs). G-protein-linked activation cascades show significant amplification through their constituent signaling molecules exerting multiple inotropic, chronotropic, and lusitropic effects on cardiac function (Bers, 2002). Both levels of and balance between sympathetic and parasympathetic activity may affect arrhythmogenicity: autonomic dysregulation is involved in multiple cardiac arrhythmias. Thus, both increased sympathetic and vagal rebound activity amplify ventricular arrhythmic risk. Its modulation may offer novel therapeutic options.

The sympathetic transmitters adrenaline and noradrenaline are, respectively, released into the circulation by adrenal medullary release and locally by widely distributed cardiac sympathetic nerve terminals. They, respectively, preferentially bind surface membrane β₁- and β₂-adrenergic receptors. Cardiomyocytes express β₁-adrenergic receptors, activating the stimulatory G_s protein. Its G_α subunit then binds guanosine triphosphate (GTP), dissociates from the GPCR and its βγ-subunit, and activates adenylyl cyclase, enhancing cellular cyclic 3′,5’ adenosine monophosphate (cAMP) levels, with diverse compartmented cellular actions (Zaccolo, 2023). Of nerve–cardiomyocyte communication mechanisms under current study, ventricular Dbh ⁺ catecholaminergic cardiomyocytes (Dbh ⁺ Cate-CMs) expressing dopamine β-hydroxylase occur in close structural relationship with the sympathetic innervation. Their catecholaminergic-type vesicles suggest endocrine functions. They may also be involved in embryonic development and maturation of the cardiac conduction system (Wang et al., 2017).

At the membrane level, sympathetic activity increases heart rate and modifies AP generation and waveform. cAMP binds to and opens SAN HCN, increasing I _f and heart rates. It also activates protein kinase A (PKA) whose widespread phosphorylation actions excite Nav1.5, Kv11.1, Kv7.1, and Cav1.2, increasing rapid inward I _Na and slow outward I _Kr and I _Ks and I_CaL, increasing ventricular AP amplitudes but shortening the AP plateau durations. It also accelerates SAN pacemaker potentials (Huang and Lei, 2023). In excitation–contraction coupling, sympathetic activity increases I _CaL and net cellular Ca²⁺ entry increasing rates and force of subsequent muscle contraction. PKA-mediated RyR2 phosphorylation also reduces regulatory FKBP12 ligand binding otherwise stabilizing RyR2 closure, thereby increasing RyR2 Ca²⁺ sensitivity and Ca²⁺-induced Ca²⁺ release. It also phosphorylates phospholamban (PLN), relieving its SERCA2 inhibition, enhancing diastolic SR Ca²⁺ store.

The Epac2 isoform of cAMP-dependent exchange protein activates CaMKII (Takla et al., 2020; Terrar, 2023) and, therefore, RyR2-mediated SR Ca²⁺ release (Hothi et al., 2008). At further upstream levels discussed below, Epac1 activates Ca²⁺-dependent calcineurin signaling, driving hypertrophic, morphological, and cytoskeletal changes. Additionally, excitatory postganglionic sympathetic co-transmitters, adenine nucleotides, act on metabotropic P2Y receptors. Adenosine (A₁) receptor activation activates phosphokinase C (PKC), which also acts on voltage-gated Na⁺ and K⁺ channels, L-type Ca²⁺ channels, and RyR2. Finally, sympathetic responses can differ among cardiomyocyte types, exemplified by the distinct pulmonary vein (PV) and left atrial cardiomyocyte responses to noradrenaline implicated in atrial ectopy (Bond et al., 2020).

The parasympathetic, inhibitory, transmitter, acetylcholine (ACh), acts on muscarinic (M₂) receptors activating the G_i2 protein in the SAN, AVN or atria, in both the presence and absence of, pre-existing adrenergic challenge. It does so in ventricular tissue in the presence of such challenge. Its G_α subunit binds GTP and splits off from the receptor and its Gβγ-subunit. The latter opens GIRK1 and GIRK4 channel components, increasing inward rectifying I_KACh or I_KAdo, particularly in supraventricular tissue (Schmitt et al., 2014; Dascal and Kahanovitch, 2015; Lerman, 2015). The dissociated G_iα inhibits adenylate cyclase (AC) and, therefore, cAMP production (DiFrancesco, 1993), increasing I _CaL and I _f in pacemaker cells. G_i activation also increases protein phosphatase (PP2A) activity through cell division control protein 42 homolog (Cdc42)/Ras-related C3 botulinum toxin substrate 2 (rac2) and the cardioprotective p21-activated kinase PAK1. PP2A dephosphorylates PKA-phosphorylated proteins at the same serine/threonine phosphorylation sites, reversing their L-type Ca²⁺ channel, RyR2 and PLN effects. Parasympathetic action, thus, slows heart rates and decreases contractile force.

PAK1 may be cardioprotective through increasing PP2A activity (Ke et al., 2009; Wang et al., 2015) and its remodeling actions (Liu et al., 2011; Wang et al., 2014; He et al., 2023; Jung et al., 2023). It may also modify Cav1.2/Cav1.3 (I _CaL)-mediated Ca²⁺ entry, RyR2-mediated SR Ca²⁺ release, and CaMKII-mediated SERCA2a and NCX transcriptional regulation (He et al., 2023) and promote SERCA activity (Ke et al., 2009; Wang et al., 2014; Wang et al., 2015). PAK1 deficiency may promote atrial arrhythmogenesis under adrenergic stress conditions through posttranslational and transcriptional modifications of key molecules including RyR2 and CaMKII (Jung et al., 2023).

Over longer systems-level timescales, autonomic innervation provides effector pathways for central nervous responses to environmental inputs. Normal circadian ion channel remodeling cycles explain SAN pacemaking changes, resulting in higher background heart rates during wakefulness than during periods of sleep (Jesus et al., 2021). These are likely driven by sympathetic, as opposed to parasympathetic, actions coupled to suprachiasmatic nuclear circadian rhythms. Rather than variations of autonomic transmitter activity (Black et al., 2019), recent evidence attributes these to periodic transcriptional cardiac remodeling cycles varying ion channel abundances and their consequent ionic current densities. The SAN ion channel transcriptome for many important cardiac ion channels, particularly HCN, displays circadian rhythms (Anderson et al., 2023; Black et al., 2019; D’Souza et al., 2021; Wang et al., 2021), sensitive to chronic but not acute pharmacological autonomic blockade (Pitzalis et al., 1996; Black et al., 2019). It possibly involves cAMP response element action on key Per1 and Per2.18 clock genes (Travnickova-Bendova et al., 2002). Critically needed are future studies of autonomic plasticity and age-related properties (Chadda et al., 2018; Takla et al., 2023).

Longer-term pro-arrhythmic upstream cellular energetic and tissue remodeling can accompany metabolic pathology in various conditions, including acute hypoxic/ischemia–reperfusion (Liu et al., 2020), chronic obesity, insulin resistance and type 2 diabetic conditions (Vianna et al., 2006; Sato et al., 2009; Kucharska-Newton et al., 2010), and cardiac failure, accompanying primary hypertrophic or fibrotic changes (Terentyev et al., 2008; Grivennikova et al., 2010; Belevych et al., 2012; Chouchani et al., 2014). First, maintenance of ionic gradients and Ca²⁺ cycling processes represents ∼30–40% of cardiomyocyte ATP utilization. Of the latter, ∼90% is replenished by their extensive mitochondrial network (Leone and Kelly, 2011; Nguyen et al., 2019; Wang et al., 2019).

Mitochondrial mass and function are regulated by transcriptional coactivators (Sonoda et al., 2007): peroxisome proliferator-activated receptor (PPAR) γ coactivator-1 (PGC-1) family members are highly expressed in oxidative tissues including the heart. Of these, mitochondrial promoter PGC-1α or PGC-1β expression increases with upstream signals including cold exposure, fasting, and exercise, likely signaling anticipated cellular energy requirements (Figure 7A) (Huang, 2017). PGC-1 proteins exert matching multi-level regulation of cellular mitochondrial function and metabolism. Of their activated nuclear receptor targets, PPARα regulates genes involved in mitochondrial fatty acid oxidation; estrogen-related receptor alpha (ERRα) regulates mitochondrial fatty acid oxidative and oxidative phosphorylative energy transduction pathways (Vega et al., 2000). Additionally, PCG-1α coactivation of nuclear respiratory factor-1 (NRF-1) and -2 (NRF-2) (Wu et al., 1999) modulates nuclear-encoded transcription factor Tfam expression required in replicating, maintaining, and transcribing mitochondrial DNA (Garesse and Vallejo, 2001) and mitochondrial biogenesis (Vega et al., 2000; Huss et al., 2004). These mechanisms may underlie the induction of nuclear genes encoding mitochondrial proteins in energetic pathways, including the tricarboxylic acid cycle, and nuclear and mitochondrial genes encoding electron transport chain and oxidative phosphorylation complex components in cultured cardiomyocytes with forced PGC-1 expression (Lehman et al., 2000). PCG-1s fall in obesity, insulin resistance, type II diabetes mellitus, and aging in parallel with mitochondrial dysfunction (Leone and Kelly, 2011; Dillon et al., 2012).

Mitochondrial dysfunction, whether intrinsic or arising from excessive energetic demand, compromised vascular oxygen supply or pathological energetic disorders, destabilizes inner membrane potentials, driving their electron transport chain. At the cell membrane level, the consequently compromised ATP and/or rising adenosine diphosphate (ADP) increase sarcolemmal K-ATP (sarcKATP) channel opening probabilities (Akar and O’Rourke, 2011). This shortens APDs and ERPs and compromises cell excitability and AP propagation (Akar and O’Rourke, 2011). These events predispose to re-entrant arrhythmia (Fosset et al., 1988; Faivre and Findlay, 1990).

Increased reactive oxygen species (ROS) production accompanying oxidative stress also influences cardiomyocyte excitability (Figures 7B–D). The resulting atrial and ventricular arrhythmic tendency may result from reduced Nav1.5 expression (Liu et al., 2010), increased I _NaL(Liu et al., 2023), and reduced connexin-43 (Cx43) trafficking and function (Akar and O’Rourke, 2011; Yang et al., 2014; Yang et al., 2015) and cell–cell coupling (Smyth et al., 2010) and inhibited voltage-dependent I _K (Wang et al., 2004), sarcolemmal K_ATP (Weiss et al., 1989), and I _Ca. These effects could affect AP conduction (Liu et al., 2010) and underlie the atrial ERP shortening and AF initiation with rapid pacing (Figure 7E) (Carnes et al., 2001; Dudley et al., 2005). Right atrial appendages of AF patients show increased oxidative stress markers (Emelyanova et al., 2016). Accompanying changes in reduced (NADH) or oxidized nicotinamide adenine dinucleotide (NAD⁺), reflecting cell oxidative state, also, respectively, inhibit and enhance activity even with normally expressed Nav1.5 (Liu et al., 2010).

At the excitation–contraction coupling level, increased ROS oxidize RyR2, increasing SR Ca²⁺, and reduce SERCA-mediated Ca²⁺ reuptake, increasing cytosolic [Ca²⁺]_I (Kukreja et al., 1988). CaMKII oxidation may induce a kinase activity similar to auto-phosphorylated CaMKII (Erickson et al., 2008). ROS also oxidize and activate PKA (Eager and Dulhunty, 1998). ROS, thus, altered intracellular Ca²⁺ cycling (Terentyev et al., 2008; Brown and O’Rourke, 2010; Bovo et al., 2012) in aging rabbit ventricular myocytes, effects reversed by mitochondrial specific ROS scavengers (Terentyev et al., 2008). At the level of further upstream targets, ROS are implicated in fibroblast activation and transforming growth factor-β (TGF-β) production, leading to cardiac fibrosis (Friedrichs et al., 2012).

Mice deficient in either but not both PCG-1α or PGC-1β survive to adulthood. Pgc-1α−/− hearts had normal baseline function but failure with increased afterload (Arany et al., 2005). Pgc-1β−/− hearts showed normal background features but blunted heart rate responses to adrenergic challenge (Lelliott et al., 2006), increased arrhythmic propensity, and APD alternans and VT episodes (Gurung et al., 2011). Single cells showed altered ion channel expression and spontaneous diastolic Ca²⁺ transients. Chronic studies in young (12–16 weeks) and aged (>52 weeks) Pgc-1β−/− mice suggested sinus node and AVN dysfunction in intact animals following β₁-adrenergic challenge (Ahmad et al., 2018b). The atria and ventricles of Langendorff-perfused Pgc-1β−/− hearts showed increased arrhythmic inducibility with age, accompanied by reduced (dV/dt)_max, increased AP latencies, and reduced APD, together reducing λ and increasing arrhythmogenicity (Ahmad et al., 2017; Valli et al., 2017a; Ahmad et al., 2018a). These findings correlated with direct demonstrations of reduced I _Na but not I _K (Valli et al., 2018a; Ahmad et al., 2019). Pgc-1β−/− hearts also showed accelerated age-related fibrotic change, discussed further in the next section (Ahmad et al., 2017; Valli et al., 2017c).

Second, longer-term inflammatory and structural, fibrotic and hypertrophic, changes can pro-arrhythmically add to and influence the physiological processes above (Nattel et al., 2007; Dobrev and Nattel, 2010; Iwasaki et al., 2011). They involve inflammatory cell recruitment, cellular proliferation, angiogenesis, and extracellular matrix accumulation (Murphy et al., 2015). Accumulating evidence implicates inflammatory processes involving tumor necrosis factor (TNF)-α, macrophage migration inhibitory factor, interleukin (IL)-1β, IL-6, and JUN N-terminal kinases (JNKs) with pro-inflammatory M1 macrophages participating particularly in AF. Additionally, pro-inflammatory (M2) macrophages may enhance atrial fibroblast proliferation and differentiation. TNF-α, in turn, activates profibrotic transforming growth factor-β (TGF-β) signaling and matrix metalloproteinase secretion. TGF family members, particularly TGF-β1, regulate tissue homeostasis and repair, immune and inflammatory responses, extracellular matrix (ECM) deposition, and cell differentiation and growth (Yang et al., 2010). TGF-β1 is a key contributor to fibrosis (Dobaczewski et al., 2011). TGF-β1 signaling is itself activated by angiotensin II (Ang II) action on local cardiac renin–angiotensin system (RAS) angiotensin receptor type 1 ATR₁s, which themselves directly activate TGF-β production.

TGF-β likely acts through Smad and synergistically with extracellular signal-regulated protein kinase ½ (ERK1/2), c-Jun NH₂-terminal kinase (JNK), p38 mitogen-activated protein kinases (p38MAPK), all mitogen-activated protein kinases (MAPK) (Dobaczewski et al., 2011). It stimulates myofibroblast differentiation, key to the fibrotic process, and ECM protein synthesis (Desmouliere et al., 1993) through inhibiting matrix metalloproteinase (MMP) and inducing tissue inhibitor metalloproteinase (TIMP) synthesis (Mauviel, 2005). Ang II similarly acts in this, both itself, in synergy with TGF-β1 (Schnee and Hsueh, 2000; Murphy et al., 2015), and itself activating TGF-β1 signaling (Schultz et al., 2002; Dobaczewski et al., 2011). TGF-β1 administration-induced fibroblast proliferation causes atrial fibrosis (Dzeshka et al., 2015; Hu et al., 2015) and AF (Lifei et al., 2020).

ATR₁s also activate G protein, Gα_q/11, Gα_12/13, and G_βy, and non-G protein-related pathways, modulating multiple oxidase and kinase signaling pathways. Among serine/threonine kinases, Ang II- or ROS-mediated CaMKII activation enhances phosphorylation of excitation–contraction coupling-related proteins and cell survival and transcription factors, driving hypertrophic and inflammatory gene expression (Swaminathan et al., 2012; Rusciano et al., 2019). PKC activates NAD(P) H oxidase, thence increasing ROS production and consequent cardiomyocyte hypertrophy (Nakamura et al., 1998; Sugden and Clerk, 1998). Activation of the MAPKs, including ERK1/2, JNK, and p38MAPK, is implicated in cell growth and hypertrophy (Wollert and Drexler, 1999), as well as cardiac fibrosis, through increased procollagen I, procollagen III, fibronectin, and TGF-β gene transcription. Epidermal growth factor receptor (EGFR) and platelet-derived growth factor (PDGF) insulin receptor signaling can also occur. Of non-receptor tyrosine kinases, Src, janus kinase/signal transducer and activator of transcription IL (JAK/STAT), and focal adhesion kinase (FAK) (Mehta and Griendling, 2007), JAK/STAT may promote pressure overload and ischemia-related hypertrophy (Booz et al., 2002). Finally, recent reports implicate upregulated Notch signaling in hypertrophic and dilated cardiomyopathy (Langa et al., 2023).

Cardiac remodeling first affects surface membrane ion channel function, excitation–contraction coupling, and conduction of excitation. Of inflammatory cytokines, TNF-α may reduce I _Ca,L, while increasing I _to, systolic Ca²⁺ transients, diastolic intracellular [Ca²⁺], and inward NCX and decreasing SR ATPase expression. These changes accentuate DAD-induced triggered activity (Lee et al., 2007). It may also alter Cx40 and Cx43 expression and distribution. JNK2 activates CaMKII, increasing diastolic SR Ca²⁺ leak, and downregulates Cx43 expression, compromising cell-to-cell communication and atrial AP conduction (Yan et al., 2018) Inflammasome NLRP3 activation upregulates RyR2 expression and CaMKII-dependent hyperphosphorylation and consequent pro-arrhythmic diastolic SR Ca²⁺ release, similarly inducing DADs and ectopic firing (Yao et al., 2018). It also increases I _Kur and I _KACh and increases atrial hypertrophy and fibrosis (Heijman et al., 2020). PAK1 signaling may exert protective signaling actions through Cav1.2/Cav1.3 (I _CaL)-mediated Ca²⁺ entry, RyR2-mediated SR Ca²⁺ release, and CaMKII-mediated regulation of SERCA2a and NCX transcription (He et al., 2023). Its deficiency may modify posttranslational and transcriptional processing in the key Ca²⁺ homeostatic molecules RyR2 and CaMKII with atrial pro-arrhythmic effects under adrenergic stress (Jung et al., 2023).

Second, fibrotic changes may contribute to both AF and ventricular arrhythmias. Increased fibrosis-related gene expression accompanying altered surface and Ca²⁺ homeostasis ion channel gene expression may contribute to SAN dysfunction in experimental rat pulmonary arterial hypertension (Logantha et al., 2023). Disruption of myocardial bundle continuity and gap junction formation due to increased extracellular matrix associated with fibrotic and/or hypertrophic change reduces and accentuates heterogeneities in AP conduction velocity, disrupting AP propagation wavefronts and promoting re-entry (Davies et al., 2014). Connexin-mediated electrotonic coupling between cardiomyocytes and inexcitable fibroblasts with relatively depolarized (∼-30 mV) resting potentials may depolarize cardiomyocyte resting membrane potentials altering myocyte excitability, slow conduction, shorten APD, and induce spontaneous phase 4 depolarization (Nattel, 2018). Additionally, Ca²⁺ signaling in fibroblasts employs stretch-sensitive short TRPC3 or melastatin-type TRPM7 transient receptor potential (TRP) channels, both upregulated in AF fibroblasts (Yue et al., 2015).

Other downstream processes may feed forward to modify these changes. Both genetic Nav1.5 (Jeevaratnam et al., 2012; Jeevaratnam et al., 2016) and Pgc1-β-related metabolic deficiencies (Ahmad et al., 2017; Valli et al., 2017b; Ahmad et al., 2018a; Ahmad et al., 2018b) were associated with accentuated TGF-β1-mediated fibrosis, possibly driving observed electrophysiological remodeling underlying age-related SND dysfunction (Hao et al., 2011) and pro-arrhythmic atrial and ventricular fibrosis. The former non-canonical roles for Nav1.5 in cellular biological processes extend to associations with morphological changes in dilated and arrhythmic cardiomyopathies (Remme, 2023).

Pharmacological intervention underpins much clinical arrhythmia management. Both modern drug innovation and optimizing existing therapeutic strategies require systematically classified mechanisms of action at identified molecular physiological targets and their rational correlation with single cell, tissue, or organ tissue effects and, in turn, clinical indications and therapeutic actions. This requirement had been addressed by successive historic Singh–Vaughan Williams (VW) (Vaughan Williams, 1975) and updated European Society of Cardiology classifications (Task force of the working group on arrhythmias The European Society of Cardiology, 1991). The most recent modernized classification scheme (Lei et al., 2018) responding to the recent cardiac electrophysiological and pharmacological insights outlined here pragmatically retained but extended the original VW classes. It mapped actions on molecular targets within more recently characterized levels of surface membrane ion channel, excitation–contraction coupling, autonomic function, and longer-term energetic and remodeling changes outlined above. The resulting scheme for cardiac arrhythmic mechanisms thereby offered a basis for understanding existing explorations for novel arrhythmic therapies (Lei et al., 2018), and reviewing drug cardiac safety (Saadeh et al., 2022).

Of the introduced novel drug categories, class 0, bearing on cardiac automaticity, includes the use-dependent SAN inhibitor ivabradine. It reduces heart rate by acting on I _f, sparing myocardial contractility and vascular tone, actions distinct from the existing VW classes II and IV. It is clinically approved for reduced ejection fraction under circumstances of cardiac failure, adjunct therapy improving clinical outcomes by reducing heart rate. It is yet to be established for stable ischemic heart disease. It may also benefit patients with inappropriate sinus tachycardia intolerant of class II or IV agents (Koruth et al., 2017).

Of extended VW classes I–IV, class I now includes class Id drugs acting on I _NaL this complements the existing Class Ia–c, which reduce early I _Na thereby reducing AP phase 0 slopes and overshoots, and modify APD and ERP. Class Ia drugs classically bind to the Nav1.5 open state with τ ∼1–10 s dissociation time constants. They concomitantly block I _K, slow AV conduction, and increase ERP and APD. Classes Ib and Ic contrastingly bind preferentially to the inactivated state with rapid τ ∼0.1–1.0 s or slow τ > 10 s dissociation. This results in a use-independent and use-dependent channel block and a slowed AV conduction. These insights clarify some of flecainide’s paradoxical actions. Thus, flecainide exerts anti-arrhythmic actions in gain of Na⁺ channel, particularly I _NaL, function associated with LQTS3 and experimental murine Scn5a^+/Δkpq. In contrast, it shows pro-arrhythmic actions under post-infarct conditions and clinical and experimental Scn5a+/− Brugada syndrome (Martin et al., 2010; Martin et al., 2011a). Of drugs in the new class Id, ranolazine produces a frequency-dependent block of the pro-arrhythmic I _NaL. This shortens APD and increases refractoriness and repolarization reserve in LQTS3, pathological bradycardic and ischemic conditions, and cardiac failure (Belardinelli et al., 2015). Intracellular Na⁺ overload, otherwise arising from such I _NaL, increases reverse-mode NCX, promoting pro-arrhythmic intracellular Ca²⁺ overload and SR Ca²⁺ leak. Ranolazine may also reduce atrial peak I _Na and Kv11.1-mediated I _K (Burashnikov et al., 2007; Antzelevitch et al., 2011).

A broadened class II encompassed further modifiers of G-protein-mediated signaling besides the β-adrenergic antagonists originally used to slow down SAN pacing and AVN conduction. These included new selective β1-(atenolol and bisoprolol) and non-selective β-adrenergic antagonists (carvedilol), adenosine A₁-activators (adenosine), and cholinergic muscarinic M₂ receptor inhibitors (scopolamine) and activators (carbacholine and pilocarpine) (Singh, 2004). Future possibilities could arise from numerous (∼150) further orphan GPCRs (Lei et al., 2018). Finally, centrally acting sympathetic inhibitors, such as monoxidine, have been investigated in relation to AF management (Hu et al., 2023).

An expanded class III encompassed drugs directed at the large number of more recently discovered, sometimes atrial or ventricular-specific, voltage-sensitive and -insensitive K⁺ channel types. This added to the voltage-gated K⁺ channel blockers originally employed to delay AP phase 3 repolarization, lengthening ERPs. Exemplars, including investigated agents directed at atrial arrhythmia, included those directed at voltage-sensitive I _K such as the atrial I _Kur, including XEN-D0103 (Ford et al., 2016) and vernakalant, for rapid AF (Wettwer et al., 2013), with the latter producing some I _Na and I _NaL, and I _KACh in addition to I _Kur block. New I_Kr blockers under development include vanoxerine (Lacerda et al., 2010). Investigational agents directed at voltage-insensitive I _K include the SK channel inhibitor AP30663 (Bentzen et al., 2020) and the K2P blocker doxapram (Wiedmann et al., 2022).

Major advances in excitation–contraction coupling have yielded significant progress in class IV drugs beyond the L-type Ca²⁺ channel inhibitors introduced to reduce SAN and AVN rates and conduction (Vaughan Williams, 1975). First, investigational non-selective and phenylalkylamine and benzothiazepine inhibitors supplement dihydropyridine Cav1.2 and Cav1.3 surface membrane I _CaL and I_CaT inhibitors.

Second, of agents targeting intracellular membrane molecules mediating SR Ca²⁺ release and reuptake, the experimental drug dantrolene enhances CaM-RyR2 binding and, consequently, N-terminal-central domain interactions, stabilizing RyR2 closed states (Pabel et al., 2020). Recent reports implicate flecainide in class IV RyR2 open-state block reducing RyR2 open probabilities, in addition to its class I actions. Both experimental and clinical reports support its application in catecholaminergic polymorphic ventricular tachycardia (Watanabe et al., 2009; Hilliard et al., 2010; Hwang et al., 2011; Van Der Werf et al., 2011; Salvage et al., 2018). As expected from concentration dependences for the latter effect (Watanabe et al., 2009; Galimberti and Knollmann, 2011; Heath et al., 2011), flecainide (1 µM) was, respectively, pro- and paradoxically anti-arrhythmic in murine WT and homozygotic RyR2-P2328S atria. In WT, it reduced I _Na, slowing conduction velocity, sparing refractory periods. Contrastingly, in RyR2-P2328S, it restored normal WT I _Na values, similarly sparing refractory periods. Dantrolene challenge yielded similar results. These findings suggest its rescue of an arrhythmic phenotype in RyR2-P2328S involves relief of Nav1.5 inhibition by reducing RyR2-mediated Ca²⁺ release in general feedback effects of Ca²⁺ homeostasis on different Nav subtypes (Salvage et al., 2018; Salvage et al., 2022; Salvage et al., 2023). Higher flecainide concentrations (5 µM) reduced I _Na in both WT and RyR2-P2328S (Salvage et al., 2015).

Carvedilol has similar dual class IV in addition to β-adrenergic blocking class II actions: its antioxidant actions reduce RyR2 phosphorylation and oxidation, and its open-state channel block decreases RyR2-mediated Ca²⁺ release (Zhou et al., 2011). Finally, explorations of SERCA activation by istaroxime also have potential translational implications (Ferrandi et al., 2013; Huang, 2013).

Third, continued progress in understanding Ca²⁺ homeostasis demonstrates novel regulatory novel Ca²⁺ signaling molecules. In addition to CaMK-II inhibitors, exemplified by RA608 (Mustroph et al., 2020), and investigational PAK1-targeted agents (He et al., 2023) influencing phosphorylation of cytosolic handling proteins, further signaling molecules, including cADP-ribose, nicotinic acid adenine dinucleotide phosphate (NAADP) (Terrar, 2023), and Epac proteins (Li et al., 2017; Valli et al., 2018b), offer potential investigational targets.

The modernized scheme additionally introduced novel classes V–VII. Investigational drugs from class V blockers of mechanically sensitive TRP channels included N-(p-amylcinnamoyl) anthranilic acid, which may also target Cl⁻ channels (Harteneck et al., 2007), the TRPC3 inhibitor pyrazole-3 (Harada et al., 2012), and TRPM7 inhibitor, LBQ657 (Lifei et al., 2020). These may additionally potentially inhibit hypertrophic and fibrotic changes (Harada et al., 2012; Seo et al., 2014). Class VI targeting cell–cell electrotonic coupling includes the gap junction blocker carbenoxalone (De Groot et al., 2003) and Cx43 opening peptide rotigaptide (Hsieh et al., 2016). Class VII targeting upstream signaling and remodeling contains significant representatives for its component processes. Among inflammatory cytokines, these inhibit TNFα decoy receptor [etanercept (Aschar-Sobbi et al., 2015)], IL-1β [colchicine (Wu et al., 2020)], and IL1 [anakinra (De Jesus et al., 2017)]. Microtubule disruption by colchicine may prevent NLRP3 inflammasome assembly (Wu et al., 2020). Profibrotic modeling is targeted by the TGF-β1 inhibitor pirfenidone (Lee et al., 2006) and TGF-β1 receptor inhibitor GW788388 (de Oliveira et al., 2012). Agents targeting further signaling pathways include the ERK-phosphorylation inhibitor mesalazine (Künzel et al., 2021) and the HMG-CoA reductase inhibitors, statins (Adam et al., 2011). PAK1 activators may inhibit maladaptive, pro-arrhythmic, hypertrophic remodeling and progression in cardiac failure (Tsui et al., 2015; Wang et al., 2018; He et al., 2023) in addition to its role regulating ion channel activity, Ca²⁺ homeostasis, and cardiac contractility (Ke et al., 2009; Wang et al., 2014; Wang et al., 2015). Angiotensin-converting enzyme and angiotensin receptor blockers, aldosterone receptor antagonists, are further exploratory targets (Savelieva and Camm, 2008; Geng et al., 2020; Saljic et al., 2022; Hu et al., 2023).

Research in the physiological sciences has long involved work in successive, mutually reinforcing, interacting cycles between laboratory and clinic. Identified clinical or novel physiological phenomena initiate the development of representative experimental physiological models. These insights, in turn, prompt innovative clinical testing, management, and treatment, themselves prompting further iterations of experimental testing to identify novel physiological targets, investigational new drugs, and interventions (Huang et al., 2020; Huang and Lei, 2023). Future such cycles could add immunotherapeutic approaches, including antibody-based agents with further enhanced target specificities even targeting particular channel isoforms (Presta, 2008) and micro-RNAs/shRNAs knocking down protein expression at the mRNA level provided suitable delivery methods become available (Chakraborty et al., 2017). Within a framework of a modernized drug classification, these could themselves add to understanding arrhythmic events and their modification besides facilitating future clinical development.