Cellular Structure of the Heart

The myocardium is composed of cardiac myocytes, which are highly differentiated and specialized cells responsible for the conduction of the electrical impulse and the heart’s contractile behavior, and nonmyocytes, which serve a number of functions. Myocytes occupy two thirds of the structural space of the heart; however, they represent only one third of all cells. Nonmyocytes include fibroblasts responsible for the turnover of extracellular matrix that predominantly consists of fibrillar collagen types I and III. The collagen scaffolding provides for myocyte alignment and coordinated transmission of contractile force to the ventricular chamber. Other nonmyocyte cells include the endothelial and smooth muscle cells of the intramural vasculature, neuronal elements (such as ganglia) and, under some conditions, inflammatory cells. The gross anatomic features of the heart, the extracellular matrix, and the intramural vasculature create both macroanatomic and microanatomic barriers that are central to both the normal electrophysiology of the heart and clinically important arrhythmias.

Cardiac myocytes are a family of structurally distinct cells, with a design commensurate with their function. Pacemaking cells such as those in the sinoatrial (SA) and atrioventricular (AV) nodes underlie the spontaneous electrical activity of the heart and contain relatively few contractile elements. In contrast, muscle cells are packed with actin and myosin filaments, which serve the main function of the heart—propulsion of blood through the vasculature. Contractile myocytes are rod-shaped cells of approximately 100 × 20 µm. The myocyte is enveloped by the cell membrane, a lipid bilayer 80 Å to 100 Å in thickness. This insulating bilayer permits little to no transport of ions and maintains a separation of charge established by active transporters that reside in the cell membrane. Ion channels are transmembrane proteins that serve as a conductive pathway between the inside and outside of the cell, allowing the flow of ions and, thus, charge (i.e., current). Also, no current flow exists between myocytes. However, unlike skeletal muscle, cardiac tissue is not a true syncytium; rather, cells are connected to each other by low-resistance communications called gap junctions that contain intercellular ion channels.

Membrane Potential and Conduction

Ion concentration and charge gradients across the cell membrane are responsible for the membrane potential of the cardiac myocyte. Transmembrane ionic and electrical gradients are maintained by a series of energy-requiring ion pumps and exchangers that perform the following functions: (1) Concentrate K⁺ inside the cell, (2) keep the intracellular sodium (Na⁺) low (<10 mM), and (3) tightly regulate intracellular calcium (Ca²⁺) concentrations (100–200 nmol/L). The pump responsible for establishing and maintaining most of the monovalent cation gradients is the Na⁺-K⁺ ATPase (sodium-potassium adenotriphosphatase), although other pumps and exchangers that transport Na⁺, Ca²⁺, and hydrogen (H⁺) play significant roles in the genesis of ionic concentration gradients and the membrane potential. Ion channels are passive but selective conduits for the flow of ions along electrical and chemical gradients established by active ion transport systems. The physicochemical basis of the membrane potential depends (as described below) on the Nernst potential (E_x):

(1)

where R is the gas constant, T is temperature, F is Faraday’s constant, and z is the valence of the ionic species. If the resting cardiac myocyte is assumed to have an intracellular [K⁺] of ~150 mmol/L and an extracellular [K⁺] of 4 mmol/L, then the Nernst potential for K⁺ (E_K) is roughly –90 mV. The Nernst potential represents the voltage at which the osmotic tendency for K⁺ to flow across the membrane is exactly balanced by the electrical tendency to flow in the opposite direction, which results in a net zero ionic flux and thus no net current flow. Accordingly, the K⁺ Nernst potential is sometimes referred to as the zero current potential, at which no K⁺ current would flow through open K channels. By contrast, the Nernst potential for Na⁺ in the resting cell is approximately +60 mV, indicating that if a pathway for Na⁺ to enter the cell were present, Na⁺ would enter the cell to move the membrane voltage (mV) toward the Nernst potential for Na⁺. Indeed, this is precisely what happens when Na channels are open and initiate phase 0 of the action potential. The Nernst potential for potassium is, in fact, very close to the resting membrane potential of the ventricular myocyte, indicating that the resting heart cell membrane is highly permeable to K⁺. Actual resting potentials are less negative than E_K due to small conductances of other ionic species with less negative Nernst potentials. In the course of the cardiac cycle, the cell membrane becomes permeable to different ionic species, and these changes in permeability determine time-dependent changes in the membrane potential, with each ion striving to move the membrane voltage to its Nernst potential and inscribing regionally specific action potentials (Figure 3-1).

Figure 3-1 Schematic of action potentials in different regions of the heart. The permeability of the cell membrane determines membrane voltage. At rest, cells in all regions of the heart are more permeable to potassium (K⁺) than to any other cation, hence the negative membrane potential, near the Nernst potential for K⁺. The characteristic shapes of action potentials are determined by the ionic currents that are active in each cell type during the cardiac cycle.

Passive Membrane Properties and Cable Theory

The cardiac cell membrane can be modeled as a circuit comprising variable resistors (ion channels) in parallel with a capacitor (lipid bilayer), an RC circuit (Figure 3-2). The flow of current across the membrane will alter the charge on the capacitor (and therefore the membrane potential) and change the membrane resistance. The flow of current occurs not only across but obviously along the inside and outside of the cell membrane from cell to cell as well, that is, current propagates.

Figure 3-2 Electrical and biologic representations of the cardiac cell membrane. The cardiac (or any excitable) membrane can be modeled as an RC circuit with variable resistors (ion channels) in parallel with a capacitor (cell membrane). The transmembrane voltage is established by a series of energy-requiring pumps that maintain ionic gradients across the membrane.

Cable theory, originally developed to understand current flow in trans-oceanic telegraphic cables, can be used to model passive current flow and propagation in a cardiac muscle fiber. In their simplest formulation, the cable equations define the distribution of voltage along a continuous, uniform cable of infinite length stimulated by a point source. The predictions of cable theory are that (1) a change in voltage exhibits a characteristic decay along the cable defined by the space constant (distance over which voltage decays to 1/e of the value at the site of injection) (Figure 3-3, A); and (2) there is an inverse relationship between resistance to current flow (both transmembrane and intercellular) and the cable diameter (Figure 3-3, B). That is, the space constant is directly proportional to the cable diameter, and thus, greater lengths of the cable are influenced by the same current injection into a thick, rather than a thin, cable.

Figure 3-3 Spread of current in an idealized cable. A, Injection of current into the cable at site I produces a transmembrane voltage change of smaller amplitude, and slower kinetics are recorded at more distal sites in the cable. In an excitable tissue such as the heart, a stimulus of sufficient amplitude elicits a regenerative response (i.e., an action potential) and can propagate along the length of the preparation. Transmission of the action potential is associated with current flow across and along the membrane. B, The characteristics of the cable determine the distance over which the voltage difference induced by the current injection decays (space constant). In a cable of larger diameter, the voltage difference falls over a larger distance; that is, the space constant (the distance over which the voltage difference falls to 1/e of the value at the site of injection) is proportional to the radius of the cable.

Substantial anatomic and biophysical limitations exist when applying the cable theory description of conduction to cardiac muscle. Anatomically, the shape of the heart is complex, and at any level of integration above a single muscle fiber, it does not resemble a cable. Conduction through the myocardium is not continuous; instead, myocardial cells are connected by gap junction channels that create a non-uniformity of intercellular resistance. Macroscopic discontinuities such as fibrous tissue and blood vessels also significantly perturb the cable view of conduction in the myocardium. Finally, the cardiac cell membrane does not just comprise RC circuits; instead, when stimulated to the threshold, it will generate action potentials (see Figure 3-3, A). Despite the limitations of cable theory, it serves as the foundation for several important concepts in impulse propagation and was used to demonstrate the electrical nature of conduction in the heart.¹

Depolarization of cardiac muscle results in the generation of an action potential at the site of excitation and, in doing so, sets up a voltage gradient between the excited cells and their nearest neighbors. The current generated by the action potential serves as an excitatory current for neighboring cells (source), which at their resting membrane potential (sink) are activated by the source. Impulse propagation depends on the balance between the magnitude of the currents in the source and the sink and the resistance along the fiber. Failure of conduction may result from alterations in the source current; examples include reduction of the source by drug blockade of Na (atrial or ventricular muscle) or Ca⁺ channels (nodal cells) or from changes in the characteristics of the sink, such as ischemia and activation of ATP-dependent potassium channels (I_K-ATP). In the latter case, propagation fails because the tissue with greatly increased repolarizing K current, due to activated I_K-ATP, acts as an infinite sink and cannot be sufficiently depolarized to reach the threshold for the generation of an action potential. Myocardial ischemia will produce changes in the intracellular environment such as decreased pH and increased intracellular Ca²⁺ that will serve to reduce gap junctional conductance and functionally uncouple myocardial cells, altering the relationship between the source and the sink and hindering impulse propagation.

The safety factor for conduction is the magnitude of the current provided by the source that is in excess of that required to activate the sink. The main factors influencing source current are the rate of rise of the upstroke and the amplitude of the action potential, which are the metrics that reflect the magnitude of inward currents. The factors that influence the current requirements of the sink are the membrane resistance and the difference between the resting and threshold potentials. One major reason for the mismatch between the source and the sink is an abrupt anatomic change, such as that which occurs at the Purkinje fiber–ventricular muscle junction. Orthodromic conduction over the Purkinje system results in the activation of a broad band of ventricular muscle by narrow strands of Purkinje fibers. Such an abrupt transition from an anatomically narrow source to a massive sink makes propagation tenuous such that small changes in the characteristics of the source or the sink are likely to produce failure of conduction. Antidromic conduction from ventricular muscle to Purkinje fibers produces just the opposite source-sink relationship and thus a higher safety factor for conduction.

While this discussion implicitly treats the activation wave in one dimension, the behavior of propagating waves in the heart is more complex. The complexity can be appreciated if one considers propagation in two dimensions. In this circumstance, the shape of the wavefront is a major determinant of the efficiency of propagation. A convex wavefront, as might be observed after point stimulation, creates a large sink around a smaller activating source. This mismatch reduces conduction velocity and the safety factor for propagation. Conversely, a concave activation front produces a source-sink mismatch that favors the source; this results in a high safety factor and more rapid impulse transmission. Thus, not only do source-sink characteristics influence propagation, they also influence the curvature of the wave front (see Figure 3-4).

Figure 3-4 Relationship of current sources and sinks. A, Compared with the steady-state conduction velocity for the planar wavefront in a homogeneous medium (θ₀), the conduction velocity of a convex wavefront is slower, and a concave wavefront is faster with lower and higher safety margins for conduction, respectively. In situations of structural inhomogeneity, as when a Purkinje fiber activates ventricular muscle (B), the conduction velocity is slower, with a lower safety margin at the structural discontinuity. C, In the case of a larger mass of tissue activating a smaller mass, the conduction velocity is fast, with a high safety margin.

Directionally different conduction velocity is a characteristic feature of cardiac muscle known as anisotropic conduction. Anisotropic conduction has its basis in the structure of the myocyte and cardiac tissue; myocytes are rod shaped and are organized in bundles that are oriented along the long axis of the cell. Transmurally, the axes of these bundles undergo significant changes in orientation through the ventricular wall (~120 degrees maximal deviation).² Communication among myocytes occurs via gap junction channels that are non-uniformly distributed over the surface of the heart cell, with larger numbers of channel poised to propagate the impulse longitudinally, rather than transversely, to the long axis of the muscle fiber.³ The implications of anisotropy for conduction in the longitudinal and transverse direction under pathologic conditions are controversial. In the context of uniform depression of conduction, as might exist with antiarrhythmic drug treatment, propagation in the transverse direction is preserved compared with conduction in the longitudinal direction. However, when cellular uncoupling occurs, such as in ischemia, longitudinal propagation may exhibit a higher safety factor than does transverse conduction.

Major Breakthroughs: Voltage Clamp, Molecular Cloning

A stimulus of sufficient magnitude applied to a myocyte (or any excitable cell) elicits a typical change in membrane potential known as the action potential. The ionic current basis of the action potential was confirmed and quantitatively studied using the voltage clamp developed in the middle of the twentieth century.⁴ Voltage clamping is a technique whereby the experimenter controls the transmembrane voltage and measures the current at that defined voltage. Much of what we know about ionic currents in myocytes comes from voltage clamp experiments and a more recently developed type of voltage clamp called the patch clamp.⁵ A variant of the patch clamp technique permits the measurement of ionic currents through single ion channels.

Typically, in a voltage clamp experiment, the membrane voltage (V) is held near the resting membrane potential, (approximately −80 mV for ventricular myocytes) and then stepped to more positive voltages. This voltage step induces two components of the membrane current (I_M) flow; at the instant of the voltage change, ions (I_C) flow to charge the membrane capacitance (C_M), after which the current reflects the movement of ions through the ion channels (I_i).

(2)

Capacitive current is generally small and transient and can usually be electronically compensated, so the voltage clamp provides a robust measure of current flow through ion channels, thereby permitting the study of the detailed biophysics and pharmacology of ionic currents and channels. The major limitation of such experiments in myocytes is the existence of many currents that are simultaneously active in response to the voltage step. Experimental conditions can be altered to isolate a current of interest; however, this often requires the presence of drugs, toxins, or highly nonphysiologic conditions. An alternative to the study of ionic currents in native cells was afforded with the molecular cloning and heterologous expression of ion channel genes. Expression of an ion channel gene in a non-excitable cell without other overlapping currents permits the study of the ionic current of interest under more physiologic conditions. The fundamental limitation of heterologous expression is that the ion channel is removed from its native cellular background, which may change the behavior of the channel.

The combination of highly sensitive electrophysiological methods, such as patch-clamp recording, and deoxyribonucleic acid (DNA) cloning, heralded the era of understanding of the molecular basis of cardiac excitability.

Ion Channels and Transporters: Molecular Building Blocks of the Action Potential

Currents that underlie the action potential are carried by complex, multi-subunit transmembrane glycoproteins called ion channels (Table 3-1). These channels open and close in response to a number of biologic stimuli, including a change in voltage, ligand binding (directly to the channel or to a G-protein–coupled receptor), and mechanical deformation. Other ion-motive transmembrane proteins such as exchangers and transporters make important contributions to cellular excitability in the heart. Ion pumps establish and maintain the ionic gradients across the cell membrane that permit current flow through ion channels. If pumps, transporters, or exchangers are not electrically neutral (e.g., 3 Na⁺ for 1 Ca²⁺), they are termed electrogenic and can further influence electrical signaling in the heart.

The most abundant superfamily of ion channels expressed in the heart consists of voltage-gated ion channels. Various structural themes are common to all voltage-dependent ion channels. First, the architecture is modular, consisting either of four homologous subunits (in K channels, see Figures 3-7 and 3-8) or of four internally homologous domains (in Na and Ca channels) (see Figures 3-5 and 3-6). Second, proteins wrap around a central pore (see Figure 3-8). The pore-lining (P segment) regions exhibit exquisite conservation within a given channel family of like selectivity (e.g., jellyfish, eel, fruit fly, and human Na channels have very similar P segments), but not among families with different selectivities. Third, the general strategy for activation gating (opening and closing in response to changes in the membrane voltage) is highly conserved: The fourth transmembrane segment (S4), typically studded with positively charged residues, lies within the membrane field and moves in response to depolarization, thus opening the channel. Fourth, most ion channel complexes include not only the pore-forming proteins (α-subunits) but also auxiliary subunits (e.g., β-subunits) that modify channel function.

Figure 3-5 Models of the sodium (Na) channel. Top, Topographic illustration of the Na channel α-subunit with four pseudohomologous domains (I to IV). One or more transmembrane β-subunits coassemble with the α-subunit to form the intact channel. The fourth membrane-spanning repeat (S4) is charged and serves as the voltage sensor for channel activation. The segments or linkers between S5 and S6 in each domain, called permeation segments, or P segments, form the outer pore mouth and the selectivity filter. The linker between the third and fourth domains underlies fast inactivation and contains one of the mutations that underlies the chromosome 3–linked form of long QT syndrome. The S6 segment of the fourth domain contains residues critical for the binding of the local anesthetic to the channel. Bottom, Illustration of the caniliculus through which the S4 segment slides during channel activation. An outward movement of the S4 during channel activation is proposed.

Figure 3-6 Top, Subunit structure of the cardiac calcium (Ca) channel. The α₁1.2 (α_1C) subunit forms the pore and contains drug-binding sites. β₂ and α₂– to δ-subunits coassemble with the α₁ subunit. The intact cardiac L-type Ca channel containing the α₁1.2-subunit is referred to as Ca_V1.2. Bottom, The transmembrane topology of the Ca channel is similar to the sodium (Na) channel with four homologous domains, each containing six membrane-spanning repeats. The S4 segments (dark gray) are generally conserved, but the molecular basis of inactivation is distinct from Na channels. The Ca channel exhibits both voltage-dependent inactivation and Ca²⁺-dependent inactivation. Calmodulin (CaM), which is required for Ca²⁺-induced inactivation of the channel, is tethered to the channel in the carboxyl terminus; with binding of Ca²⁺, a conformation change occurs in this region of the channel that requires CaM, the CaM binding motif (IQ), and the EF hand motif, and together these constitute the Ca²⁺ inactivation region of the channel.

Sodium Channels

Sodium (Na) channels have been highly conserved through evolution and exist in all species, from the jellyfish to humans; they are nature’s solution to the conundrum of coordination and communication within large organisms, particularly when speed is of the essence. Thus, Na channels are richly concentrated in axons and muscle, where they are often the most plentiful ion channels. A mammalian heart cell, for example, typically expresses more than 100,000 Na channels but only 20,000 or so L-type (large and long-lasting) Ca channels and fewer copies of each family of voltage-dependent K channels.

Na channels were the first ion channels to be cloned and have their sequence determined.⁷ In humans, more than 10 distinct Na channel genes have been cloned from excitable tissues, with striking homology to the complementary DNA (cDNA) cloned from eel electroplax. The cardiac Na channel gene (SCN5A) resides on the short arm of chromosome 3 (3p21) (see Table 3-1). The Na channel complex is composed of several subunits, but only the α-subunit is required for function. Figure 3-5 shows that the α-subunit consists of four internally homologous domains (labeled I to IV), each of which contains six transmembrane segments. The four domains fold together so as to create a central pore, whose structural constituents determine the selectivity and conductance properties of the Na channel.

Peptide linkers between the fifth (S5) and sixth (S6) membrane-spanning repeats in each domain, referred to as the P segments, come together to form the pore. The primary structure of the S5-S6 linkers of Na channels in each domain is unique. Thus, the structural basis of the permeation of Na channels differs fundamentally from that of K channels, in which four identical P segments can come together to form a K⁺-selective pore (see below).

One of the seminal contributions of Hodgkin and Huxley was the notion that Na channels occupy several “states” (which are now viewed as different conformations of the protein) in the process of opening (activation); yet another set of conformations is entered when the channels close during maintained depolarization (inactivation).⁴ The m gates that underlie activation and the h gate that mediates inactivation were postulated to have intrinsic voltage dependence and to function independently.⁸ While some of the implicit structural predictions of that formulation have withstood the test of time, others have not. For example, the four S4 segments are now widely acknowledged to serve as activation voltage “sensors.” In the process of activation, several charged residues in each S4 segment physically traverse the membrane (see Figure 3-5, bottom panel). The contributions of each S4 segment to activation are markedly asymmetrical; some of the charged residues play a much more prominent role than do others in “homologous” positions.⁹ Other studies have revealed that activation is coupled with inactivation. Indeed, the time course of current decay during maintained depolarization predominantly reflects the voltage dependence of activation, although single-channel inactivation itself does vary with voltage (particularly in cardiac Na channels). If the S4s are the sensors, where are the activation “gates” themselves? This crucial question still remains unresolved. However, according to experimental evidence, S6 is the leading contender for the physical activation gate.

Inactivation of Na channels is as arcane a process as activation. Not only is there loose coupling to activation, but there are multiple inactivation processes. One common approach to distinguishing inactivated states is to determine the rate at which they recover the ability to activate: Repriming from the traditional “fast” inactivation occurs over tens of milliseconds, while recovery from “slow” inactivation may need tens of seconds or longer. Fast inactivation is at least partly mediated by the cytoplasmic linker between domains III and IV (the crucial residues are labeled IFM in Figure 3-5), which may function as a hinged lid, docking onto a receptor formed by amino acids in the S4-S5 linkers of domains III and IV. This notion is consistent with observations that fast inactivation can be disrupted by internal proteases. Nevertheless, it is increasingly clear that mutations scattered widely throughout the channel affect inactivation gating. The structural determinants of slow inactivation are less well localized than those of fast inactivation. Mutations in the P region of domain I affect both activation gating and slow inactivation, while various widely scattered disease mutations identified in paramyotonia congenita and other skeletal myopathies suppress slow inactivation of the Na channel.

The S6 segment of domain IV has been proposed to contain the receptor for local anesthetics that block Na channels in a voltage-dependent manner. The homologous domains on Ca and K channels are also loci for drug binding. Na current blockade is enhanced at depolarized potentials and/or with repetitive pulsing. These observations are consistent with the idea that local anesthetics act as allosteric effectors of the inactivation gating mechanism: When they bind to the channel, they facilitate inactivation. It is clear that gating interacts with local anesthetic blockade so profoundly that it is difficult to interpret the localization of a “receptor” to S6. Mutations in S6 at putative receptor sites alter gating, independent of superimposed drug effects. Further, mutations in distant parts of the molecule can also dramatically alter the phenotype of local anesthetic blockade. Despite these caveats, the S6 segments appear to play a special role in the effects of drugs in all of the voltage-gated ion channels.

Pharmacologic competition studies and mutagenesis have defined a number of neurotoxin binding sites on the Na channel. Among these, tetrodotoxin (TTX)—a guanidinium-containing blocker—has contributed the most to our understanding of Na channel structure and function. Externally applied TTX blocks neural and skeletal muscle Na channel isoforms potently (in the nM range) but blockade of cardiac channels requires much higher concentrations (~10⁻⁵ M). The identity of one particular residue in the P region of domain I accounts for most of the isoform-specific TTX sensitivity: An aromatic residue at this position (373 in the human heart Na channel) confers high affinity, while its absence renders the channel TTX resistant. Many other residues in the outer mouth of the channel contribute to the binding of TTX and the related divalent guanidinium toxin saxitoxin (STX), suggesting that the toxin has a large footprint on the external surface of the channel.

Four different β-subunits (Na_Vβ₁ to 4, SCN1B-SCN4B) have been isolated, and all appear to be single membrane–spanning domain (type I topology) proteins with a large extracellular V-shaped immunoglobulin (Ig) fold often found in cell adhesion molecules and small carboxyl terminal cytoplasmic domain.¹⁰ The effects of the particular β-subunit on the kinetics and voltage dependence of the α-subunit vary depending on the particular β-subunit and the cell expression system. Despite the expression of β₁ mRNA, using subtype-specific antisera, the functional role of β-subunits in cardiac Na currents is still being debated. Several pieces of evidence are consistent with a role for β-subunit(s) in heart cells. First, β-subunits are found in cardiac myocytes, although no Na_Vβ1 is found in association with the α-subunit protein from the rat heart. Second, Na_Vβ1 variably modulates the function of Na_V1.5 in heterologous expression systems, including the sensitivity of the channel to blockade by antiarrhythmic drugs and free fatty acids. Finally, mutations in β-subunits have been directly or indirectly implicated in heritable cardiac arrhythmias.

Regulation of the Na channel by serine/threonine phosphorylation is a complex process. Isoforms of the Na channel α-subunit fall into one of two groups, long (neuronal and cardiac) and short (skeletal muscle and eel). Neuronal isoforms have a substantially larger intracellular linker between domains I and II, which contains five consensus sites for cyclic adenosine monophosphate (cAMP)–dependent protein kinase (PKA) phosphorylation. In fact, PKA modulates the function of expressed neuronal and cardiac Na channels. The cardiac Na channel has eight candidate consensus PKA phosphorylation sites in the I-II linker, all of which are distinct from neuronal channels. In vitro studies of the expressed cardiac Na channel demonstrate cAMP-dependent phosphorylation on only two of these serines. Interestingly, when the cardiac channel is phosphorylated by PKA, whole-cell conductance increases, suggesting that the specific pattern of phosphorylation is responsible for the functional effect and may involve changes in the trafficking of the channel to the cell membrane.

In contrast to PKA, protein kinase C (PKC) alters the function of all of the mammalian Na channel isoforms. The PKC effect is largely attributable to phosphorylation of a highly conserved serine in the III-IV linker (see Figure 3-5). Conventional PKC isoforms reduce the maximal conductance of the channels and alter gating in an isoform-specific fashion. The macroscopic current decay of neuronal channels is uniformly slowed by PKC, which suggests destabilization of the inactivated state. Cardiac channels exhibit a hyperpolarizing shift in the steady-state availability curve consistent with an enhancement of inactivation from closed states.

Overexpression of calmodulin kinase (CaMKIIδC) in transgenic murine heart alters cardiac I_Na function, stabilizing inactivation and increasing persistent current; however, this finding is complicated by the presence of heart failure in this murine model. Subacute expression after adenoviral infection of adult ventricular myocytes has a similar effect on I_Na. In contrast, in acutely isolated guinea pig ventricular myocytes, enhanced CaMKII activity destabilizes inactivation gating and increases the persistent or late current, thus prolonging the action potential duration.

Alteration of ion channel function is an important pathophysiologic mechanism of various familial diseases of muscle and brain and of inherited arrhythmias. The Na channelopathies were among the first molecularly characterized human ion channel diseases.¹¹ Rare allelic variants in SCN5A have been linked to inherited ventricular arrhythmias,¹² conduction system disease, and sudden infant death syndrome. Complex electrophysiological phenotypes have been associated with mutations in both α- and β-subunits of Na channels. Rare variants or mutations have been associated with sudden death in women (in a population-based study) and with atrial fibrillation. More common acquired forms of long QT syndrome (LQTS), generally associated with drug ingestion, have been linked to common variants of SCN5A. Finally, disease-causing mutations in the Na channel have been associated with alterations in drug blockade. The highly variant phenotypes of these arrhythmic syndromes are, in part, explained by the variable effects of the mutations on channel subunit function or expression.

Calcium Channels: L-type

The pore-forming subunit (α₁) of the calcium (Ca) channel is built on the same structural framework as the Na channel.¹³ As is the case with the Na channel, a number of genes encode surface membrane Ca channel α₁-subunits. The predominant sarcolemmal Ca channels in the heart are the L-type (large and long-lasting) and T-type (tiny and transient) Ca channels (Table 3-2). The cardiac L-type Ca channel (Ca_V1.2) is a multi-subunit transmembrane protein composed of α_1C (α₁1.2) (165 kDa), β (55 kDa), and α₂ (130 kDa) to δ (32 kDa) subunits. Three genes are known to encode L-type Ca channel α₁-subunits (Ca_Vα₁1.x to α₁3.x), and the Ca_Vα₁1.2 is the gene expressed in the heart (see Table 3-1). Distinct splice variants of the Ca_Vα₁1.2 gene have been described, and these contribute to the diversity of the cardiac L-type Ca channel function. Similar to the α-subunit of the Na channel, the S5-S6 linkers (P segments) of the α₁-subunit of the Ca channel form the ion-selective pore (Figure 3-6). However, unlike Na channels, each P segment contributes a glutamic acid to a cluster that serve to bind Ca²⁺ in the channel pore. The β-subunit is completely cytoplasmic and noncovalently binds to the α_1C subunit, modifying its function and contributing to appropriate membrane trafficking of the channel complex. Although β_2a has been proposed to be the major L-type Ca channel β-subunit, splice variants of all β-subunits β₁ to β₄ are expressed in the mammalian heart in a complex spatial and temporal pattern. As many as five genes have been suggested to encode the α₂-δ subunit: These gene products undergo post-translational processing to produce the mature extracellular α₂-subunit linked by a disulfide bond to the transmembrane δ-subunit. In heterologous expression systems, α₂– to δ-subunits enhance expression of Ca channels and hasten current activation and deactivation in the presence of α₁– and β-subunits.

Table 3-2 Properties of Calcium Channels

Ba, Barium; Cd, cadmium; Ni, nickel.

The α₁-subunit of the Ca channel contains a highly basic S4 transmembrane segment in each homologous domain, which, by analogy, is thought to be the voltage sensor for channel activation. Activation of skeletal muscle and cardiac L-type Ca channels differ; skeletal muscle channels activate much more slowly than do their cardiac counterparts. Based on the properties of chimeric channels constructed from cardiac (Ca_Vα₁1.2) and skeletal muscle (Ca_Vα₁1.1) α₁-subunits, the difference in activation gating resides in the first homologous domain; however, it is unclear if the S4 membrane-spanning segment is the crucial structural motif.

Ca channels inactivate by both Ca²⁺-dependent inactivation (CDI) and voltage-dependent inactivation (VDI) processes. CDI and VDI are regulated by channel phosphorylation and β-subunits, which suggests shared structural mechanisms that perhaps involve the I-II domain linker. The C-terminus contains peptide sequences that bind calmodulin (CaM) and Ca²⁺; these are an IQ motif named for the signature amino acids (isoleucine and glutamine) in the sequence and a helix-loop-helix structural domain, referred to as an EF hand, that mediate CDI. CaM is permanently tethered to the channel complex and serves as a Ca²⁺ sensor for the L-type channel. The Ca²⁺-CaM complex facilitates the interaction of CaM with the IQ motif resulting in the occlusion of the inner mouth of the Ca channel pore and terminating the inward Ca²⁺ flux despite continued depolarization. As cytoplasmic Ca²⁺ concentration falls, calmodulin unbinds Ca²⁺ and the IQ motif, thus relieving Ca²⁺-dependent inactivation. Spatial discrimination of Ca²⁺ concentration by CaM may involve regions in the N-terminus of the channel. A Ca²⁺-binding EF hand motif in the carboxyl terminus of α₁ appears to be necessary to confer Ca²⁺-dependent inactivation to the Ca_Vα₁1.2-subunit, although not through a direct binding of Ca²⁺ (see Figure 3-6, bottom). The I-II interdomain linker has been demonstrated as the site of β-subunit binding as well as a critical structural determinant of VDI. Mutations in this linker have been associated with the highly arrhythmogenic Timothy syndrome.¹⁴

L-type Ca channels are found in all of the myocytes of the mammalian heart and have several important electrophysiological functions. In SA nodal tissue, both L- and T-type channels contribute to diastolic depolarization and, therefore, impulse formation. Modulation of L-type current by the autonomic nervous system is important in controlling the rate of sinus node discharge. Blockade of the L-type channel underlies the sinus node slowing, which is observed with some Ca channel antagonists. The AV node is the only place in the body where Ca channels (L-type) normally conduct excitatory impulses. Consequently, it is not surprising that modulators of the L-type current have profound effects on SA and AV conduction. In muscle tissue, the L-type Ca current is the major depolarizing current during the action potential plateau, and inhibition of this current reduces the voltage of the plateau and shortens the action potential duration. Downregulation of the L-type current, as seen in atrial myocytes isolated from patients with a history of atrial fibrillation, is thought to promote the maintenance of fibrillation.

Calcium Channels: T-type

The other major Ca channel present in the sarcolemma of heart (and prominently in vascular smooth muscle) cells is the T-type channel. The T-type channel has a biophysical fingerprint that is distinct from that of the L-type channel, opening at more negative voltages, inactivating more rapidly, and having a lower conductance than the L-type channel (for reviews, see references 15 and 16). The distribution of the T-type current is more restricted in the heart than is that of the L-type current. The T-type current has been recorded in the SA node, AV node, atrium, and Purkinje cells but not in the normal adult ventricle (see Table 3-2). The T-type current plays a prominent role in phase 4 diastolic depolarization and the action potential upstroke of pacemaking cells. T-type currents play a role in the developing heart and in pathologic remodeling in some species, but this current has not been detected in normal or diseased human ventricular myocytes. Three genes encoding the T-type Ca current α₁-subunits have been cloned. The Ca_Vα₁3.1 (CACNA1G) cDNA was isolated from neuronal tissue and was the first of this new class of Ca channels cloned. The gene resides on human chromosome 17q22 (see Table 3-1), with a predicted topology similar to that of other Ca channel cDNAs.¹⁶ The gene encoding Ca_Vα₁3.2 (I) expressed in human heart resides on chromosome 16p13.3. The gene encoding Ca_Vα3.3 (CACNA1I) on chromosome 22q13 first appeared as part of the output of the Human Genome Project.¹⁷ Expression of the Ca_Vα₁3 cDNAs produces currents with the biophysical hallmarks of the T-type Ca current. Interestingly, Ca_Vα₁3.2 lacks a consensus β-subunit–binding motif in the I-II linker of the channel and expresses robustly without the need for other subunits in contrast to L-type channels. The T-type channel genes lack an EF hand or IQ motifs, which suggests modes of inactivation distinct from the L-type channel.

Four chemical classes of compounds have been used to block Ca currents: (1) dihydropyridines, (2) phenylalkylamines, (3) benzothiazepines, and (4) tetralols. Ca channel blockers exhibit significant pharmacodynamic heterogeneity across classes and even within a given chemical class. Drugs of the phenylalkylamine (verapamil) and benzothiazepine (diltiazem) classes are effective antiarrhythmics primarily used to terminate some supraventricular arrhythmias, to control ventricular response in others, and in some forms of idiopathic ventricular tachycardia. Dihydropyridines are more potent vasodilators and are not useful as antiarrhythmic compounds. A number of mechanisms explain these clinical differences. The classes of drugs (dihydropyridines, phenylalkylamines, benzothiazepines) that block the L-type channel have distinct but overlapping binding sites on the Ca_Vα₁1.2-subunit. Vascular smooth muscle and cardiac muscle express different splice variants of Ca_Vα₁1.2, and the vascular variant is more sensitive to block by dihydropyridines. Perhaps more important than the intrinsic sensitivity of the specific Ca_Vα₁1.2 variant to a blocking compound are the voltage dependence and kinetics of channel blockade by the compound. Like Na channel–blocking local anesthetic antiarrhythmic drugs, Ca channel antagonists exhibit use-dependent and voltage-dependent blockade as a result of the preference of the drugs to bind to the inactivated states of the channel. The enhanced sensitivity of vascular Ca channels to blockade by dihydropyridines is predominantly due to the depolarized resting membrane potential of vascular smooth muscle cells (Vm approximately –40 mV), compared with that of cardiac myocytes, and the greater occupancy of the inactivated state. Differences among dihydropyridines, phenylalkylamines, and benzothiazepines in blocking cardiac Ca channels are significantly related to the kinetic of interaction of the drug and the channel. Phenylalkylamines dissociate from the Ca channel very slowly, dihydropyridines do so rapidly, and benzothiazepines recover with intermediate kinetics. T-type channels are less sensitive to blockade by dihydropyridines, phenylalkylamines, and benzothiazepines. Mibefradil is a tetralol Ca channel blocker that is relatively selective for T-type over L-type Ca channels. It was briefly marketed for the treatment of hypertension but was withdrawn because of a high incidence of adverse effects, often occurring as a result of drug interactions.

Potassium Channels

Currents through K channels are the major repolarizing currents in the heart, but the relative importance of any specific channel varies regionally in the heart. K channels are the most diverse subfamily of channel proteins comprising molecules with three distinct molecular architectures (Figure 3-7). The inward rectifier currents (I_K1, I_KACh, I_KAdo), designated Kir, are encoded by a K channel that is evolutionarily the most primitive and comprises only two membrane-spanning repeats (analogous to S5 and S6) and a pore, or P, segment. The latter contains the K channel signature sequence (TVGYGDM) that underlies the K⁺-selective permeability of the channel. The first K channel gene isolated was from Drosophilia melongaster. This mutant fruit fly was called Shaker because of its response to ether anesthesia. The gene that caused the Shaker phenotype was isolated by positional cloning and encoded a voltage-dependent K (Kv) channel.¹⁸ Since the original cloning of the Shaker K channel (Kv1.x), a number of K channel genes in the same or closely related gene families have been isolated (Kv1.x to Kv11.x). The voltage-dependent K channels that have been identified in the mammalian heart are shown in Figure 3-7. The voltage-dependent K channels, which are structurally similar to a single domain of the Na channel or the Ca channel, are composed of six membrane-spanning segments, including a highly basic S4 segment. The cytoplasmic half of the S6 membrane–spanning repeat appears to mediate drug blockade of voltage-gated K channels, analogous to regions of the Na channel that bind local anesthetics. Similar to Kir channels, Kv channels must tetramerize to form the intact channel and are typically associated with ancillary subunits. Within a subfamily of K channels (e.g., Kv1.x), subunits may hetero-multimerize, but it is believed that assembly does not occur across subfamilies. It seems likely that two rounds of gene duplication generated Ca and Na channels from the less complex Kv structure. It is possible that a more straightforward gene duplication of an inward rectifier channel produced the third type of K channel, the two-pore K⁺-selective channel (see Figure 3-7).

Figure 3-7 Evolutionary relationships of families of potassium (K) channels in the mammalian heart. The predicted transmembrane topology of the three subclasses are shown. Top, inward rectifier (Kir). Left bottom, twin pore K. Right bottom, voltage-gated (Kv) channels. Gene duplication events involving Kir channels likely produced the twin-pore and Kv channel gene products in this illustration. Examples of specific channels in each class that are expressed in the human heart are given.

The cDNAs that encode the α-subunits of the K channel are sufficient to generate K⁺-selective currents, but a number of ancillary subunits that modify channel function (Kvβ, KCNE, KChIP) have been identified. A family of related proteins (Kvβ1 to Kvβ3) modulates the function of Kv channels. The β-subunits bind to the amino terminus of Kv α-subunits that modify their function in an isoform-specific fashion (for review see reference 19). The crystal structure of Kvβ2 complexed with the amino terminus of Kv1.1 has been solved,²⁰ and it has been suggested that it functions as an oxidoreductase. Indeed, Kvβ1.2 has been shown to confer oxygen sensitivity to Kv4.2 channels. A recently described, unrelated family of proteins, KChIPs, that contain Ca²⁺-binding EF hand motifs modulate the function of the members of the Kv4 family, which suggests the possibility that more than one type of ancillary subunit can interact with Kv4 channels. The molecular details of the interaction of KChIP and Kv4 subunits are still being studied (Figure 3-8, left).

Figure 3-8 Potassium (K) channels are multi-subunit complexes. α-Subunits are the major, pore-forming subunits. Transmembrane segments are represented by cylinders; the red cylinders are the voltage sensors, and the yellow cylinders line the ion conductive pore. At least three types of accessory subunits are found in the heart. Kvβ associates with the amino terminal section of Kv1 α-subunits and the carboxyl terminus of Kv4 α-subunits. KChIPs are Ca²⁺-binding ancillary subunits that associate with the amino terminus of Kv4 α-subunits. Kvβ- and KChIP-subunits increase the current density when co-expressed with Kvα-subunits and modify gating. The gene products of KCNE1 (minK) and KCNE2 (MiRP-1) are believed to be transmembrane proteins. KCNE1 has been shown to influence the ion-conductive pathway of the I_Ks channel. Many other proteins may modulate K channel function.

Still other K channel ancillary subunits are transmembrane proteins that are predicted to have the ability to not only alter channel gating but, in some cases, also influence channel pore properties. Most relevant to the heart are the gene products of KCNE1 (minK) and KCNE2 (MiRP-1), which are thought to coassemble with KvLQT1 (KCNQ1)^21,22 and HERG (KCNH2) to form the two components of the delayed rectifier current I_Ks and I_Kr, respectively (see Figure 3-8, right, and see below). α-Subunits, which by themselves are not functional (e.g., Kv9.x), may modulate the function of other Kv α–encoded channels.

In response to a depolarizing voltage pulse, K channels, like other voltage-gated ion channels, undergo a series of conformational changes that alter function. The S4 membrane–spanning repeats are critical components of the activation gating machinery. Many K channels (like Na and Ca channels) close in the face of continued depolarization; that is, they inactivate. The molecular basis of inactivation, however, is mechanistically heterogeneous. The first type of inactivation to be understood in molecular detail in K channels validated a scheme, referred to as the ball-and-chain mechanism, proposed by Armstrong²³ In an elegant series of experiments, Aldrich and coworkers demonstrated the “ball” role of the amino terminus of the Shaker K channel in the inactivation process that they called N-type (because it involves the amino terminus).^24,25 After channel activation by a depolarizing stimulus, the amino terminus binds to and plugs the cytoplasmic mouth of the channel pore, thus terminating the K⁺ flux (see Figure 3-7, bottom). Channels that have the amino terminus removed fail to undergo this type of inactivation, but it can be restored if a peptide that resembles the amino terminal ball is added to the cytoplasm of the cell.²⁵ A second form of inactivation (carboxyl-terminal, or C-type) involves the outer mouth of the channel pore and amino acid residues in S6 and in the P segment. It has been suggested that C-type inactivation of the channel protein resembles the closing of a camera shutter; that is, it involves constriction of the outer pore of the channel.

K channels serve multiple roles in the maintenance of normal cardiac electrophysiology. Of the multiple subtypes (voltage-gated, inward-rectifier, twin-pore), voltage-dependent K channels underlie both the transient outward (sometimes called A-type) current and the delayed rectifier current in the heart. The transient outward K current activates and inactivates rapidly and is a critical determinant of phase 1 repolarization of the ventricular action potential (see Figure 3-1). The two components of the transient outward current in the heart are a Ca²⁺-independent K current (I_to1) and a Ca²⁺-dependent current (I_to2). The latter is a K current in some species and a chloride (Cl) current in others. The channels that encode cardiac I_to1 not only vary among species but may vary regionally in the ventricle as well. Kv1.4 is a minor, but important, component of I_to1 in some species, including humans. However, in the human ventricle, I_to1 is primarily encoded by Kv4.3, which recovers from inactivation much faster than do homomeric Kv1.4 channels. Indeed, the Kv1.4 channel recovers so slowly (2 to 3 seconds) that it cannot be a significant component of the cardiac I_to1 at physiologic heart rates. However, it is possible that hetero-multimerization of Kv1.4 with other Kv1 family genes and/or coassembly with β-subunits could alter the kinetics of the current. Another argument against Kv1.4 as the major component of cardiac I_to1 is the insensitivity of expressed Kv1.4 to blockade by low-dose flecainide (10 µM), whereas expressed Kv4 and native cardiac I_to1 are both flecainide sensitive. The Kv4 family of genes is expressed in relative abundance in the mammalian heart—Kv4.3 in larger mammalian ventricles, such as dog and human, and Kv4.2 in rodent ventricle. Data suggest that Kv1.4 mRNA and protein are also present in mammalian ventricular myocytes and that a physiologic correlate of Kv1.4-based I_to1 may be a slowly recovering transient outward current in the subendocardium of the human ventricle.

The ultrarapidly activating delayed rectifier current (I_Kur), which is primarily found in the atrium in human heart (and throughout rodent heart) is generated by Kv1.5, although other rapidly activating delayed rectifiers may be encoded by genes in the Kv3 family in the atria of some species. A close correspondence exists between the biophysical and pharmacologic properties of I_Kur in human atrial myocytes and Kv1.5. Furthermore, Kv1.5 protein and mRNA have been observed in human atrial and ventricular tissues, and Kv1.5-specific antisense oligonucleotides suppress I_Kur in atrial myocytes. The restricted expression of Kv1.5 in atrium makes it an attractive pharmacologic target for the treatment of supraventricular arrhythmias.

The delayed rectifier K current (I_K) plays a major role in terminating repolarization in the cells of large mammalian hearts. I_K is a composite current made up of a rapid component (I_Kr) and a slow component (I_Ks). Definition of the genetics of LQTS clarified the molecular basis of both components of the delayed rectifier. MinK (KCNE1) was initially considered a “minimal K channel” that encoded a current that resembled I_Ks. Subsequently, positional cloning identified the disease gene in chromosome-11 linked LQTS as KvLQT1²⁶ (Kv7.1), but the current encoded by KCNQ1 was a functional orphan, not resembling any known cardiac K current. However, the co-expression of KCNQ1 and KCNE1 generated a current with a much closer resemblance to native I_Ks than to either of the subunits expressed alone.^21,22 An alternatively spliced variant of KCNQ1 is expressed in the heart and exerts a dominant negative effect on I_Ks in vitro; thus, native I_Ks may be regulated, in part, by the extent of such alternative splicing.

Long QT genetics identified KCNQ1 as the gene underlying I_Ks, and HERG (Kv11.1) encoded by KCNH2 underlying I_Kr.²⁷ I_Kr exhibits a number of unusual physiologic properties, which when disrupted (by mutations in KCNH2, by hypokalemia, or by drug blockade) disrupt normal repolarization. With depolarizations to progressively more positive potentials, activating I_Kr actually decreases. This “inward rectification” is a manifestation of the very rapid inactivation that HERG channels undergo once they are open. The extent of this fast inactivation increases at positive potentials and with lower extracellular K⁺. The latter explains the decrease in I_Kr (causing action potential and QT prolongation) observed in hypokalemia. Further, when the action potential enters phase 3, channels recover from inactivation, transitioning rapidly to an open (conducting) state before closing relatively slowly. Thus, as the action potential begins to repolarize, I_Kr increases markedly, further accelerating repolarization. HERG channels are blocked by many drugs, including methansulfonanilide drugs such as dofetilide and sotalol. As with KVLQT1, HERG may coassemble with other proteins to produce native I_Kr. Database mining for homologs of KCNE1 uncovered a related gene, MiRP-1 (encoded by KCNE2), in the same locus on chromosome 21 that encodes a topologically similar, small polypeptide with an extracellular amino terminus, a single transmembrane domain, and a cytoplasmic carboxy tail (see Figure 3-8). When MiRP-1 is co-expressed with HERG voltage-dependent gating, single-channel conductance, regulation by K⁺_o, and biphasic blockade by methansulfonanilides are all modified. However, the role of MiRP-1 in native cardiac I_Kr remains uncertain. HERG exists in alternatively spliced forms, but the role that different splice variants play in generating the native current is uncertain. As with KCNH2 and KCNQ1, mutations in KCNE1 and KCNE2 have been linked to LQTS.

Another major class of K channel genes expressed in the heart encodes inwardly rectifying currents. The term inward rectification is used to describe the fact that these channels pass current more readily into cells than out of them (Figure 3-9). All inward rectifiers share a similar topology, with only two membrane-spanning repeats and a pore loop, and they must tetramerize to form the intact channel. In 1998, a major advance in ion channel biology occurred with the determination of the structure of a bacterial inward rectifier channel from Streptomyces lividans, called KcsA.²⁸ The structure is remarkable in that it accounts for a number of the physical principles that underlie K⁺-selective permeation.²⁹ The crystal structure demonstrated that the linker between the two membrane-spanning domains (P segments) form the outer mouth of the channel and that the K channel signature sequence forms the selectivity filter. High rates of ion flux are maintained despite the relatively avid binding of K⁺ due to the presence of the two K⁺ ions in the selectivity region that repel each other. The second membrane-spanning repeat, analogous to the S6 of Kv channels, forms much of the inner mouth of the channel, where antiarrhythmic drug binding is expected to occur (Figure 3-10). Models of other K channels have since been generated by structural or homology approaches. A KCNQ1 structural model has been proposed, based on homology to other K channel structures, and this, in turn, has been used to identify the key structural features of interactions between KCNQ1 and KCNE1.

Figure 3-9 Properties of ionic currents in the heart. A, Whole-cell sodium (Na) currents recorded from mammalian tissue culture cells transfected with the complementary DNA that encodes the human cardiac Na channel (NaV1.5). By convention, the current is inward (Na⁺ ions flowing into the cell) and therefore negative. The current activates rapidly on depolarization of the cell membrane and rapidly closes in the face of maintained depolarization of the cell membrane, a gating process referred to as inactivation. This channel passes current in both inward and outward directions, depending on transmembrane voltage. B, Whole-cell recording of the transient outward K current (I_to1) recorded from a human ventricular myocyte. The current activates rapidly with depolarization and then inactivates. The current flow is preferentially in the outward direction (positive current) and is referred to as outward rectification. C, Whole-cell current flow through inward rectifier potassium (K) current (I_K1). I_K1 channels are activated at rest and close with membrane depolarization. At voltages where the channel prefers to open (voltages negative to the Nernst potential for K⁺), there is little time-dependent current decay or inactivation. Current preferentially flows in the inward (negative) direction; thus, this channel and all of the channels in the Kir family are referred to as inward rectifiers.

Figure 3-10 Crystal structure of the KcsA bacterial inward rectifier potassium (K) channel. Left, Representation of the major features of the structure of the bacterial channel, KcsA. Each of the four channel subunits (only two are shown in this figure) contain two α-helical membrane-spanning repeats, a pore helix, and the K channel signature sequence that forms the K⁺-selective pore. Right, The structure of KcsA. The peptide backbone is rendered in a ribbon format. The features of the permeation pathway include the presence of two to three K⁺ ions in the pore with ion-ion repulsion, which facilitates high rates of ion transport and a large inner vestibule comprising the carboxyl terminal portion of the outer, or M2, helix. The M2 helix corresponds to the S6 domains of voltage-dependent K channels that mediate antiarrhythmic drug binding.

The inward rectifier family of cDNAs is designated Kir, and its members are part of the KCNJ superfamily. I_K1, the current that is important in maintaining the resting membrane potential and in facilitating terminal repolarization, is encoded by the Kir2.x subfamily (KCNJ2 and KCNJ4). It is likely that Kir2.1 encodes I_K1 in the human ventricle, but other Kir2 isoforms have been detected in the heart.

The other inward rectifiers in the heart exhibit specialized functions, as in response to neurohormones or metabolic stress. The Kir3 family of inward rectifier channels underlies the K current that is coupled to the M2 muscarinic (I_KACh) or A1 adenosine receptors (I_KAdo) in nodal cells and atria. I_KACh (I_KAdo) is a heteromultimer of the products of two different genes in the Kir3 family, initially referred to as GIRK (G-protein inwardly rectifying K channel, Kir3.1) and CIR (cardiac inward rectifier, Kir3.4) (KCNJ3 and KCNJ5). Kir3.1 and Kir3.4 tetramerize in a 2 : 2 ratio to form the I_KACh channel protein, which encodes a current that is directly activated by the β-subunit and γ-subunit of an inhibitory G-protein (Figure 3-11, left). I_KACh is the primary mediator of the negative chronotropic and dromotropic effects of parasympathetic activation in the heart.

Figure 3-11 Subunit structure of I_KACh and I_KATP. Left, IKACh is a glycoprotein-coupled channel that is activated by acetylcholine (ACh) binding to M2 cholinergic receptors in the heart. The channel is separate from the receptor and is formed by the heterotetramerization of two inward rectifier potassium (K) channel subunits, Kir3.1 (GIRK1) and Kir3.4 (CIR), in a 1 : 1 stoichiometry. Right, I_KATP is formed by the hetero-octomeric assembly of the adenosine triphosphate–binding cassette protein, the sulfonylurea receptor (SUR2A), and the inward rectifier Kir6.2.

Another inward rectifier, I_KATP, links electrical signaling to the metabolic state of the myocyte. Changes in the activity of I_KATP profoundly influence the electrophysiology of the heart in ischemia and play a key role in the endogenous cellular mechanism that limits the injurious effect of myocardial ischemia known as ischemic preconditioning.³⁰ I_KATP is believed to be a hetero-multimeric channel complex comprising a tetrameric assembly of Kir6.2 channels (KCNJ8 and KCNJ11) at its core surrounded by four sulfonylurea receptor subunits (SUR2A, Figure 3-11, right). SUR2A, encoded by ABCC9, is an ATP-binding cassette (ABC) protein that imparts sensitivity to sulfonylureas and K channel openers such as pinacidil and chromakalim to the channel complex.

A third structural class of K channels has been observed in the heart. These channels comprise four transmembrane segments and two pore loops. TASK (twin-pore acid-sensitive K channel) is a member of the twin-pore family of K channel genes that is highly expressed in the heart. The TASK channel exhibits little intrinsic voltage or time dependence and therefore most resembles a background current. The precise role for this channel and other members of the twin-pore family in cardiac myocytes is unknown.

I_f “Funny” or Pacemaker Current

I_f is a current that contributes to diastolic depolarization in pacemaking cells in the heart. The current is found in many cell types, but its features are variable. For example, I_f is present in ventricular myocytes, but its activation voltage is so negative that it is not likely to be of physiologic significance.³¹ I_f activates slowly on hyperpolarization and deactivates rapidly with depolarization. I_f supports a mixed monovalent cation (Na⁺ and K⁺) current with a reversal potential of –20 to –30 mV. The current is highly regulated: β-Adrenergic stimulation increases I_f and hastens diastolic depolarization. A family of genes topologically similar to voltage-dependent K channels and related to cyclic nucleotide–gated channels in photoreceptors in the retina appears to encode I_f. A number of hyperpolarization-activated cyclic nucleotide–gated channels (HA-CNG) have been cloned from the heart, and several exhibit the general features of I_f in cardiac pacemaking cells. It has been suggested that I_f itself is a composite current with fast and slow components encoded by HCN2 and HCN4, respectively. Support for I_f as the pacemaker current in the heart also comes from a genetic model of bradycardia in zebrafish with a dramatically reduced I_f.

Electrogenic Transporters

Na⁺-Ca²⁺ Exchanger

The Na⁺-Ca²⁺ exchanger is an electrogenic ion transporter that exchanges three Na⁺ ions for one Ca²⁺. The highest levels of exchange activity have been observed in the heart. The cardiac NCX, a transmembrane glycoprotein, was originally proposed to have 11 or 12 transmembrane repeats based on hydropathy analysis. More recent mutagenesis data challenge the original topologic models and instead suggest that there may be only nine transmembrane segments³² (Figure 3-12). NCX contains two membrane-spanning domains, with the first five transmembrane segments being separated from the remainder by a large cytoplasmic loop that makes up about half of the molecule. The intracellular loop contains domains that bind Ca²⁺ and the endogenous NCX inhibitory domain, XIP.

Figure 3-12 Subunit structure and transmembrane topology of the sodium-calcium (Na⁺-Ca²⁺) exchanger (NCX) and the sodium-potassium (Na⁺-K⁺) adenosine triphosphatase (ATPase) (Na pump). Top, Two alternative transmembrane topologies for the NCX. A large cytoplasmic loop is a crucial to physiologic regulation of the exchanger and contains Ca²⁺ and inhibitory peptide (XIP)–binding domains. Bottom, The Na⁺-K⁺-ATPase is a heteromeric assembly of a large α-subunit and a smaller single membrane–spanning repeat β-subunit.

Na⁺-Ca²⁺ exchange is an electrochemical process during which three Na ions are exchanged for one Ca ion. The exchange is thus electrogenic (i.e., generates a current). Ion exchange can occur in either direction. With each heart beat, cytosolic [Ca²⁺] is released from sarcoplasmic reticulum (SR) stores, primarily by the ryanodine release channel RyR2. [Ca²⁺]_i rises from the resting level of less than 100 nM to approximately 1 µM with each cardiac cycle. Under normal physiologic conditions, outward Ca²⁺ flux through the NCX (generating an inward current) and Ca²⁺ reuptake into the SR by the SR Ca²⁺-ATPase (SERCA) are the major mechanisms of restoration of normal diastolic [Ca²⁺]. NCX is sensitive to the cytoplasmic concentrations of Ca²⁺ and Na⁺, which determine the exchanger activity and the potential at which the exchange reverses direction. The NCX current is time independent and largely reflects changes in intracellular Ca²⁺ during the action potential. Thus, NCX has an important effect on membrane voltage both at rest and during activation of the myocyte. At highly depolarized potentials, reverse mode Na⁺-Ca²⁺ exchange (Ca²⁺ influx, net outward current) can occur; however, the role of reverse mode exchange in initiating SR Ca²⁺ release and contraction is uncertain.

Increases in intracellular Ca²⁺ shift the reversal potential of NCX in the positive direction and therefore increase the driving force for the inward exchanger current. The inward NCX current will depolarize the membrane toward the threshold for firing an action potential and thus is potentially arrhythmogenic. The NCX current is an important component of the inward current (transient inward current, I_TI) that underlies delayed after-depolarizations (DADs). DADs are spontaneous membrane depolarizations from rest after complete repolarization of the action potential. DADs are usually not present under physiologic conditions but are favored by conditions that increase the SR Ca²⁺ load, such as rapid firing rates, digitalis intoxication, or ischemia/reperfusion. Under these conditions, spontaneous SR Ca²⁺ release occurs, which then increases NCX and probably other Ca²⁺-dependent currents, which, in turn, results in membrane depolarization. DADs may produce arrhythmias in two ways. First, if DADs are of sufficient amplitude, they may trigger an action potential. Second, even if DADs are below the threshold for generation of an action potential, they may affect the excitability of the cell, slowing conduction in the myocardium.

Molecular Basis of Activation and Recovery of the Heart

In normal sinus rhythm, cardiac activation begins in the SA node, the specialized collection of pacemaking cells in the roof of the right atrium between the crista terminalis and the right atrial–superior vena cava (RA-SVC) junction. SA nodal cells undergo spontaneous depolarization, repetitively activating the rest of the heart. As a result of the lower density of the inwardly rectifying K current (I_K1) and the presence of a hyperpolarization-activated pacemaker current, I_f, the resting membrane potentials of SA and AV nodal cells are considerably less negative than those of atrial or ventricular muscle cells. The result is a continuous, slow depolarization of the membrane potential; thus, nodal cells do not have a true resting potential, but the maximum diastolic potential is never more negative than –60 mV.

The aggregate activity of I_f and diminished I_K1 slowly depolarizes the nodal cell until ~–40 mV when Ca currents are activated, hastening the rate of rise of the action potential. First, the transient T-type Ca current (I_Ca,T) is activated, driving the membrane potential toward E_Ca; then, the longer-lasting, dihydropyridine-sensitive L-type Ca current (I_Ca,L) is activated. Simultaneously, the more slowly activating outward K currents (delayed rectifier, I_K) are activated, hindering the movement of the membrane potential toward E_Ca. Ultimately, Ca currents inactivate, and the membrane potential moves back toward E_K, turning off I_K and activating I_f and then starting the cycle again. Current continues to flow through the electrogenic Na⁺-Ca²⁺ exchanger throughout the cycle, and the magnitude and direction of this current depend on the membrane potential as well as intracellular Ca²⁺ and Na⁺ concentrations. It has recently been demonstrated that cyclic variations of submembrane Ca²⁺ concentration drive the activation of the Na⁺-Ca²⁺ exchanger during diastole to act in concert with ion channels to confer pacemaking activity on SA node cells, a phenomenon known as the calcium clock.^33a

The synchronization of the somewhat diffuse pacemaking cells that comprise the sinus node is through gap junction channels comprising Cx-40 and Cx-43. The activity of pacemaking cells is synchronized by a process of mutual entrainment, whereby each of the cells in the nodal syncytium constantly modulates the discharge frequency of the other cells.

As in the case of the nodal cell action potential, the highly orchestrated activity of a number of ionic currents inscribes the muscle cell action potential. A prototypical action potential from atrial and ventricular myocytes with a schematic of the trajectory of the underlying ionic currents is shown in Figure 3-13. The action potential is divided into five phases: (1) Phase 0 is rapid upstroke; (2) phase 1 is early repolarization; (3) phase 2 is the plateau; (4) phase 3 is late repolarization; and (5) phase 4 is the resting potential or, in the case of a nodal action potential, diastolic depolarization (see Figure 3-1). Unlike in the case of nodal cells, a true resting potential can be defined in cardiac muscle cells, and it is ~90 mV, close to E_K; thus, at rest, cardiac muscle cells are mostly permeable to K⁺ due to the activity of I_K1.

Figure 3-13 Subunit structure of the gap junction channels. Left, Gap junction channels are intercellular ion channels composed of two hemichannels or connexons in adjacent cells. Each hemichannel is composed of six subunits or connexins composed of four highly conserved membrane-spanning repeats, two conserved extracelluar loops and a more divergent cytoplasmic loop, and amino and carboxyl termini. Right, Different subtypes of connexins may assemble to form channels that are homomeric (single type of connexin in each connexon), heterotypic (different connexins in each connexon), or heteromeric heterotypic, in which more than one type of connexin is present in each connexon.

Under normal conditions, muscle cells are stimulated by spontaneously occurring impulses generated in pacemaking tissue. When this stimulus moves the membrane voltage positive to threshold (~–65 mV), an action potential is initiated. Depolarization beyond the threshold explosively activates Na channels, producing an enormous (~400 pA/pF) but transient (1 to 2 ms) current, driving the membrane voltage toward E_Na (+65 to +70 mV). Although, Na channels are, by far, the most numerous in the myocyte cell membrane, their activity is fortunately short lived, as otherwise the transmembrane Na⁺ gradient would be quickly exhausted. The Na current quickly dissipates by inactivation, and the membrane must repolarize to its resting potential before Na channels recover from inactivation and again become available to activate. Thus, the time and the voltage dependence of the availability of the Na current is the basis of refractoriness in cardiac muscle.

The upstroke of an action potential falls short of E_Na because of the inactivation of the Na current and the activation of a K current and, additionally, in some cases a Ca²⁺-dependent Cl⁻ current (I_to2) that in concert produce rapid membrane repolarization to ~+10 mV (phase 1). The Ca²⁺-independent transient outward K current (I_to1) activates as Na channels inactivate. Activation of I_to1 is rapid (~10 ms), and this current decays over 30 to 40 ms at physiologic temperatures. The density of I_to1 is less than 5% of the Na current; thus, inactivation of the Na current is the main reason for early repolarization, while I_to1 is an important determinant of the membrane voltage at the end of phase 1. In canine ventricular myocytes, I_to2 is a prominent current during phase 1 of the action potential; however, its role in human myocytes is uncertain.

Depolarization of the membrane potential activates a number of other currents, albeit more slowly than in the case of the Na current and I_to1. In ventricular myocytes, I_Ca,L is activated and accounts for the major depolarizing current during the action potential plateau or phase 2. This current is the main route for Ca²⁺ influx and triggers Ca²⁺-induced Ca²⁺ release (CICR) from the SR to initiate contraction. I_Ca,L tends to depolarize the cell membrane, and delayed rectifier repolarizing K currents that are active during the plateau and phase 3 oppose this action. Activation of delayed rectifier K currents and inactivation of Ca currents serve to terminate the plateau phase and begin phase 3 or late repolarization. In atrial tissue, I_Kur, is a prominent delayed rectifier that is an important determinant of the plateau height and aids in the termination of the plateau. Delayed rectifiers (especially I_Kr) are important in terminating the plateau but are limited in their ability to restore the normal resting potential because they deactivate at voltages less than –40 mV. Final repolarization is mediated by the outward component of I_K1 even in atrial cells where the density of I_K1 is small compared with that of ventricular myocytes.

Excitability and Propagation

Many of the electrophysiological properties of the heart are direct consequences of ionic current activity during the action potential. Cardiac cells are excitable because the action potential, which is a typical, regenerative response, is elicited if the membrane potential exceeds a critical threshold. Action potentials are regenerative because they can be conducted over large distances without attenuation. Action potentials generated in the sinus node serve to excite adjacent atrial muscle and, thus, the remainder of the heart under normal conditions.

In atrial and ventricular muscle at rest, the membrane is most permeable to K⁺, which is the result of the activity of I_K1. Excitability in cardiac muscle is primarily determined by the availability of I_Na. In response to an external stimulus, either from adjacent cells or an artificial pacemaker, depolarization of muscle cells occurs. If the depolarization is sufficient and raises the membrane potential above a critical value, known as the threshold potential, Na channels open, depolarize the membrane, and initiate an action potential. In pacemaking tissues such as the sinoatrial or AV node, the Na current is absent, and excitability is mediated by activation of Ca currents. The consequence is a higher threshold for activation and a slower rate of rise (~1 to 10 V/sec versus hundreds of V/sec in muscle) of phase 0 of nodal cell action potentials.

Propagation of a wave of excitation in a homogeneous cable-like medium is continuous and obeys the laws of cable theory (see Passive Membrane Properties and Cable Theory). In such a preparation, the maximal upstroke velocity of the action potential (dV/dt)_max is an indirect measure of depolarizing ionic current and conduction velocity.³⁴ A continuous cable model is a structural oversimplification of all cardiac tissue with the possible exception of normal papillary muscles. Continuous propagation of excitation waves is not characteristic of cardiac tissue. Due to the structural and functional complexities of the myocardium, discontinuous conduction (see below) is the rule.

A feedback exists between network properties (cell-to-cell coupling via gap junctions) and active membrane properties (ionic currents) in propagation in cardiac tissue preparations.³ Under conditions of normal cellular coupling, fluctuations in local conduction velocity, action potential shape, and ionic current flow are small. However, with cellular uncoupling—such as that which accompanies ischemia—the interaction between intercellular conduction and active membrane properties assumes greater significance. In cardiac muscle, the Na current is the main determinant of membrane depolarization and local circuit current. When cells are uncoupled, discontinuity of conduction increases, the delay between activation of cells increases, and the Na current may sufficiently inactivate such that the currents active during the plateau (i.e., L-type Ca current) of the action potential become essential for driving excitatory current through gap junctions. In both experimental models and computer simulations, blocking the L-type Ca current was shown to reduce the safety factor for conduction and to lower the intercellular resistance that produces conduction blockade. Cellular uncoupling and discontinuous conduction has important implications for safety factors for the propagation of the impulse. With moderate cell-to-cell uncoupling in simple models of propagation, conduction is slower but has a higher safety factor. However, with more significant uncoupling, the transmitted current is so small that insufficient Na current is recruited to initiate an action potential.

The most important causes of discontinuous conduction in the heart are macroscopic discontinuities in cardiac tissue. Such anatomic discontinuities exist in all regions of the heart and are especially prominent in the trabeculated portions of the atria and ventricles, the layers of the left ventricular wall, and the Purkinje-muscle junction. Two-dimensional models of macroscopic discontinuities highlight the importance of the change in geometry and, consequently, the dispersion of the local circuit current in the characteristics of propagation and blockade at such sites. Analogous to the feedback between cellular coupling and ionic currents, a feedback occurs between the current to load mismatch produced by the tissue architecture and the ionic current flow. Small current-to-load mismatches (larger strand-to-sheet ratio) are associated with minor conduction delays across the tissue discontinuity. In contrast, tissue architecture characterized by a large current-to-load mismatch (narrow strand into a large sheet) is associated with significant conduction delay and blockade across the discontinuity that can be produced by either Na or Ca channel blockers. Thus, the L-type Ca current is essential for impulse propagation through cardiac tissue with structural discontinuities. Such structural discontinuities are present in the normal heart but may be much more prominent in the aged or diseased (e.g., hypertrophied or infarcted) myocardium.

Repolarization and Refractory Periods

Refractoriness of tissue, a consequence of the long duration of the cardiac action potential, allows only gradual recovery of excitability. Refractoriness is essential to the normal mechanical function of the heart, as it permits relaxation of cardiac muscle prior to the next activation. Refractoriness of cardiac muscle is classified as either absolute or relative: The former occurs immediately after phase 0 and during the plateau and no stimulus, regardless of its strength, can re-excite the cell; the latter occurs during phase 3, when the cell is excitable but the stimulus strength for activation exceeds that at rest (Figure 3-14). The molecular basis of refractoriness is the lack of availability of depolarizing current (Na current in muscle) because repolarization to negative potentials is required for channels to recover from fast inactivation and thus be available to pass the excitatory current. The duration of refractoriness of any cardiac tissue thus depends on the complement of ion channels (and, in particular, depolarizing currents) expressed. When the depolarizing current becomes available to activate, outward currents (typically delayed rectifier K currents) increase the stimulus strength required to reach the threshold, making the tissue relatively refractory (compared with the rested state).

Figure 3-14 Absolute and relative refractory periods in the ventricle. Action potential recorded from a ventricular myocyte. The bars underneath the action potential delineate the periods of absolute refractoriness, where no stimulus, regardless of amplitude, can elicit another action potential, and relative refractoriness, where a subsequent action potential can be initiated with a high-strength stimulus. Under appropriate circumstances, during the period of relative refractoriness, the cell may exhibit supranormal excitability; that is, a stimulus that is normally subthreshold will elicit an action potential.

Under some conditions, some tissues, particularly Purkinje fibers, may exhibit supranormal excitability. This phenomenon occurs at the end of repolarization and is the result of reactivation of Na currents at a time when the membrane potential of the heart cell is closer to the threshold for reactivation than when the cell has fully returned to rest. Supranormal excitability is one contributor to the vulnerable period of the cardiac cycle; it contributes by increasing the likelihood of re-excitation during terminal repolarization (when heterogeneity of action potential durations are most likely to support re-entry).

Alterations in Impulse Initiation: Automaticity

Spontaneous (phase 4) diastolic depolarization underlies the property of automaticity, which is characteristic of cells in SA and AV nodes, the His-Purkinje system, the coronary sinus, and, possibly, pulmonary veins. Phase 4 depolarization results from the concerted action of a number of ionic currents, but the relative importance of these currents remains controversial (Figure 3-15). The inwardly rectifying K current (I_K1) maintains the resting membrane potential and resists depolarization; thus, the activity of other currents (e.g., Ca currents) or a reduction of I_K1 (and other K conductances) must occur to permit the cell to reach the threshold for the firing of an action potential. I_f may play a particularly prominent role in the normal automaticity of Purkinje fibers, although this hypothesis is not without controversy. Deactivation of I_K is another mechanism allowing depolarizing currents to move the membrane potential toward the threshold. Ca currents, both the T-type and the L-type, figure prominently in diastolic depolarization and in the upstroke of the action potential in nodal tissue and latent atrial pacemakers. A number of other time-independent currents may play a role in diastolic depolarization and pacemaking activity, including currents through the electrogenic Na⁺-K⁺-ATPase and the Na⁺-Ca²⁺ exchanger and background currents.

Figure 3-15 Nodal action potential and the currents that underlie phase 4 diastolic depolarization. Nodal cells exhibit phase 4 diastolic depolarization that spontaneously brings the cell to threshold, which results in the production of an action potential. Several currents play a role in phase 4, including calcium (Ca) currents (T- and L-types), I_f (or the pacemaker) current, and a reduction in the current flow through several K channels, including I_K, I_K1, and I_KACh. The rate of phase 4 diastolic depolarization is highly sensitive to the tone of the autonomic nervous system. Cholinergic agonists slow phase 4, whereas sympathomimetics hasten phase 4.

The rate of phase 4 depolarization and, therefore, the firing rate of pacemaker cells are dynamically regulated. Prominent among the factors that modulate phase 4 is the tone of the autonomic nervous system. The negative chronotropic effect of activation of the parasympathetic nervous system is the result of the release of acetylcholine that binds to muscarinic receptors, releasing G-protein βγ-subunits that activate a potassium current (I_KACh) in nodal and atrial cells (see Figure 3-11). The resultant increase in K⁺ conductance opposes membrane depolarization, slowing the rate of rise of phase 4 of the action potential. Agonist activation of muscarinic receptors also antagonizes activation of the sympathetic nervous system through inhibition of adenylyl cyclase, reducing cAMP and inhibiting protein kinase A (PKA). Conversely, augmentation of the tone of the sympathetic nervous system increases myocardial catecholamine concentrations, which activate both α- and β-receptors. The effect of β₁ adrenergic stimulation predominates in pacemaking cells, increasing the L-type Ca current and shifting the voltage dependence of I_f to more positive potentials, thus augmenting the slope of phase 4 and increasing the rate of SA node firing. L-type Ca current density is increased by PKA-mediated phosphorylation, which results in an increase in the rate of rise of phase 4 and the upstroke velocity of the action potential in nodal cells. The positive chronotropic effect of β-adrenergic stimulation has been related to increased subsarcolemmal Ca²⁺ release through the ryanodine receptor (RyR2) accelerating the rate of diastolic depolarization. Enhanced sympathetic nervous system activity can dramatically increase the rate of firing of SA nodal cells producing sinus tachycardia, with rates in excess of 200 beats/min. In contrast, the increased rate of firing of Purkinje cells is more limited, rarely producing ventricular tachyarrhythmia in excess of 120 beats/min.

Normal automaticity may be affected by a number of other factors associated with heart disease. Hypokalemia and ischemia may reduce the activity of the Na-K ATPase, thereby reducing the background repolarizing current and enhancing phase 4 diastolic depolarization. The end result would be an increase in the firing rate of pacemaking cells. Slightly increased extracellular K may render the maximum diastolic potential more positive, thus also increasing the firing rate of pacemaking cells. A greater increase in [K⁺]_o, however, renders the heart inexcitable by depolarizing the membrane potential and inactivating the Na current.

Sympathetic stimulation explains the normal response of the sinus node to stresses such as exercise, fever, and thyroid hormone excess. Normal or enhanced automaticity of subsidiary latent pacemakers produces escape rhythms in the setting of failure of more dominant pacemakers. Suppression of a pacemaker cell by a faster rhythm leads to an increased intracellular Na⁺ load (particularly in cells with a Na⁺-dependent action potential) and extrusion of Na⁺ from the cell by the Na⁺-K⁺-ATPase produces an increased background repolarizing current that slows phase 4 diastolic depolarization. At slower rates, the Na⁺_i load is decreased as is the activity of the Na⁺-K⁺-ATPase resulting in a progressively more rapid diastolic depolarization and warm-up. Overdrive suppression and warm-up may not be observed in all automatic tachycardias. For example, functional isolation of the pacemaker tissue from the rest of the heart (entrance blockade) may blunt or eliminate the phenomena of overdrive suppression and warm-up of automatic tissue.

Myocytes in the atrium and the ventricle may exhibit spontaneous activity under pathologic conditions associated with depolarization of the resting membrane potential to levels more positive than –60 mV. The mechanism of spontaneous depolarization in contractile cells is uncertain but is likely to involve the activity of a number of depolarizing and repolarizing currents that, on balance, favor membrane depolarization. Ventricular myocytes do express I_f, although the threshold for activation is well below the resting potential of the cell, so the functional significance of this current is uncertain. Currents that mediate the upstroke of the action potential of abnormally automatic cells depend on the diastolic potential. At more negative diastolic potentials, abnormal automaticity can be suppressed by Na channel–blocking drugs. At more positive diastolic potentials (>–50 mV), Na channel blockers are ineffective, whereas Ca channel blockers suppress abnormal automaticity, implicating the L-type Ca channel in the upstroke in this setting.

Abnormally automatic cells and tissues are less sensitive to overdrive suppression than are cells and tissues that are fully polarized with enhanced normal automaticity. However, in situations where cells may be sufficiently depolarized to inactivate the Na current and limit the intracellular Na⁺ load, overdrive suppression may still be observed due to increased intracellular Ca²⁺ loading. Such Ca²⁺ loading may activate Ca²⁺-dependent K conductances (favoring repolarization) and promote Ca²⁺ extrusion through the Na⁺-Ca²⁺ exchanger and Ca channel phosphorylation, increasing Na⁺ load and thus Na⁺-K⁺-ATPase activity. The increase in intracellular Ca²⁺ load may also reduce depolarizing L-type I_Ca by promoting Ca²⁺-induced inactivation of the Ca current.

Abnormal automaticity may underlie atrial tachycardia, accelerated idioventricular rhythms (IVRs), and ventricular tachycardia that is particularly associated with ischemia and reperfusion. It has also been suggested that injury currents at the borders of ischemic zones may depolarize adjacent nonischemic tissue causing predisposition to automatic ventricular tachycardia.

After-Depolarizations and Triggered Automaticity

Triggered automaticity or activity refers to impulse initiation that is dependent on after-depolarizations (Figure 3-16). After-depolarizations are membrane voltage oscillations that occur during EADs or following DADs an action potential.³⁵

Figure 3-16 Early and delayed after-depolarizations. Interruptions of repolarization before its completion are referred to as early after-depolarizations (EADs). Most EADs, especially phase 2 and early phase 3, are believed to result from reactivation of the L-type calcium (Ca) current and perhaps the sodium-calcium (Na-Ca) exchanger current. Later phase 3 EADs may also involve reactivation of Na currents. After-depolarizations that occur after the completion of repolarization are referred to as delayed after-depolarizations (DADs). The mechanism of DAD involves intracellular Ca²⁺ overload and oscillatory release of Ca²⁺ from the activation of a number of Ca²⁺-dependent conductances by the sarcoplasmic reticulum.

In the early 1970s, DADs were experimentally observed in Purkinje fibers exposed to toxic concentrations of digitalis glycosides. The cellular feature common to the induction of DADs is the presence of increased Ca²⁺ load in the cytosol and SR. Inhibition of the Na-K-ATPase by digitalis glycosides will increase Ca²⁺ load by increasing intracellular Na⁺, which is exchanged for Ca²⁺ by the Na⁺-Ca²⁺ exchanger. Increased [Ca²⁺]_i activates a transient inward current, I_TI, that depolarizes the cell. The ionic basis of I_TI is controversial but likely results from electrogenic currents through the Na⁺-Ca²⁺ exchanger and/or Ca²⁺-activated depolarizing currents.

Inhibition of the Na⁺-K⁺-ATPase by digitalis glycosides facilitates, but is not necessary for creating, the Ca²⁺ overload that predisposes to DADs. Catecholamines and ischemia sufficiently enhance Ca²⁺ loading to produce DADs. The presumed mechanism of cytosolic Ca²⁺ increase and DADs with catecholamine stimulation is an increase in transmembrane Ca²⁺ flux through L-type Ca²⁺ channels. Catecholamines may also enhance the activity of the Na⁺-Ca²⁺ exchanger, thus increasing the likelihood of DAD-mediated triggered activity. Elevations in intracellular Ca²⁺ in the ischemic myocardium are also associated with DADs and triggered arrhythmias. Accumulation of lysophosphoglycerides in ischemic myocardium with consequent Na⁺ and Ca²⁺ overload has been suggested as a mechanism for DADs and triggered automaticity. Cells from damaged areas or surviving the infarction may display spontaneous release of Ca from the SR, and this may generate “waves” of intracellular Ca²⁺ elevation and arrhythmias.

The duration of the action potential is a critical determinant of the presence of DADs. Longer action potentials associated with more transarcolemmal Ca²⁺ influx are more likely to be associated with DADs. If I_TI underlies at least part of the DAD, then the voltage dependence of the transient inward current should be reflected in the voltage dependence of DADs. Indeed, at membrane voltages where I_TI is near its maximum, DADs exhibit the largest amplitude. Importantly, stimulation of the experimental preparation at fast rates increases the size of the DAD and the presence of triggered activity, likely a function of frequency dependent loading of the SR with Ca²⁺.

Mutations in the cardiac ryanodine receptor (RYR2), the SR Ca²⁺ release channel in the heart, have been identified in kindreds with the syndrome of catecholamine-stimulated polymorphic ventricular tachycardia (CPVT) and ventricular fibrillation with short QT intervals. It seems likely that perturbed [Ca²⁺]_i changes in the relationship between SR Ca²⁺ load and the threshold for Ca²⁺ release, and thus perhaps DADs, contribute to the arrhythmias characteristic of this syndrome. Indeed, murine models of CPVT demonstrate adrenergically mediated arrhythmias, with DADs in vitro. It is likely that some ventricular tachycardias that complicate digitalis intoxication are initiated by triggered activity. It has also been suggested that DADs underlie some forms of idiopathic ventricular tachycardia, particularly from the right ventricular outflow tract (see Table 3-3). In a recent study, flecainide has been demonstrated to inhibit Ca²⁺ from RyRs in a murine model of CPVT, and preliminary data suggest efficacy in patients with refractory symptoms.

The other type of after-depolarizations, EADs, occur during the action potential and interrupt the orderly repolarization of the myocyte. They have been classified as phase 2 and phase 3, depending on when they occur, and the subclassification may have mechanistic implications. Recent experimental evidence suggests a previously unappreciated interrelationship between intracellular calcium loading and EADs. Cytosolic calcium may rise when action potentials are prolonged. This, in turn, appears to enhance the L-type Ca current (possibly via calcium-calmodulin kinase activation), further prolonging the duration of the action potential as well as providing the inward current driving EADs. Intracellular Ca loading by the prolongation of the action potential may also enhance the likelihood of DADs. The interrelationship among intracellular Ca²⁺ and delayed and early after-depolarizations may be one explanation for the susceptibility of hearts that are Ca loaded (e.g., in ischemia or congestive heart failure) to develop arrhythmias, particularly on exposure to drugs that prolong the action potential.

The plateau of the action potential is a time of high membrane resistance when there is little current flow. Consequently, small changes in either repolarizing or depolarizing currents can have profound effects on the duration and profile of the action potential. The ionic mechanisms of phase 2 and 3 EADs and the upstrokes of the action potentials they elicit may differ. At the depolarized membrane voltages of phase 2, the Na current is inactivated, and EADs can result from reactivation of the L-type Ca current. Although the available data are inadequate, it has been suggested that current through the Na-Ca exchanger and possibly the Na current may also participate in the inscription of phase 3 EADs. The upstrokes of the action potentials elicited by phase 2 and 3 EADs also differ. Phase 2 EAD-triggered action potential upstrokes are exclusively mediated by Ca currents; these may or may not propagate, but they can substantially exaggerate the heterogeneity of the time course of repolarization of the action potential (a key substrate for re-entry), since EADs occur much more readily in some regions (e.g., Purkinje, mid-myocardium) than others (epicardium or endocardium). Action potentials triggered by phase 3 EADs arise from more negative membrane voltages; the upstrokes may be due to both Na and Ca currents and are more likely to propagate.

EAD-triggered arrhythmias exhibit rate dependence. In general, the amplitude of an EAD is augmented at slow rates when action potentials are longer. Pacing-induced acceleration of the heart rate shortens the duration of the action potential and reduces EAD amplitude. Action potential shortening and the suppression of EADs with increased stimulation rate are likely the result of augmentation of delayed rectifier K currents and perhaps the hastening of Ca²⁺-induced inactivation of L-type Ca currents. Similarly, catecholamines increase the heart rate and decrease the duration of the action potential as well as EAD amplitude, despite the well-described effect of β-adrenergic stimulation to increase L-type Ca current.

A fundamental condition that underlies the development of EADs is action potential prolongation, which is manifest on surface electrocardiogram as QT prolongation. Hypokalemia, hypomagnesemia, bradycardia, and drugs can cause predisposition to the formation of EADs, with drugs being the most common cause.³⁶ Antiarrhythmics with class IA and III actions produce action potential and QT prolongation intended to be therapeutic but frequently cause proarrhythmia. Noncardiac drugs such as some phenothiazines, some nonsedating antihistamines, and some antibiotics can also prolong the duration of the action potential and cause predisposition to EAD-mediated triggered arrhythmias. Decreased [K⁺]_o paradoxically decreases some membrane K currents (particularly I_Kr) in the ventricular myocyte, which explains why hypokalemia causes prolongation of the action potential and EADs. Indeed, K infusions in patients with congenital LQTS and drug-induced QT prolongation have been shown to reduce the QT interval.^36a

EAD-mediated triggered activity likely contributes to the initiation of the characteristic polymorphic ventricular tachycardia—torsades de pointes—seen in patients with congenital and acquired forms of LQTS.^36b In addition, exaggerated dispersion of repolarization is also likely to play a role in both the initiation and the maintenance of torsades de pointes by generating a substrate for functional re-entry (see below). Acquired prolongation of the QT interval most often is the result of drug therapy or electrolyte disturbances, as noted previously. However, structural heart diseases such as cardiac hypertrophy and cardiac failure may also delay ventricular repolarization (so-called electrical remodeling) and cause predisposition to arrhythmias related to abnormalities of repolarization.^37,38 The abnormalities of repolarization in cardiac hypertrophy and cardiac failure are often magnified by concomitant drug therapy or electrolyte disturbances.

Abnormal Impulse Conduction: Re-entry

The most common arrhythmia mechanism is re-entry. Re-entry is as much a property of the networks of myocytes as it is a property of individual heart cells. Fundamentally, re-entry is circulation of an activation wave around an inexcitable obstacle. Thus, the requirements for re-entry are two electrophysiologically dissimilar pathways for impulse propagation around an inexcitable region such that unidirectional blockade occurs in one of the pathways and a region of excitable tissue exists at the head of the propagating wavefront.³⁹ Structural and electrophysiological properties of the heart may contribute to the development of the inexcitable obstacle and of unidirectional blockade. The complex geometry of muscle bundles in the heart and spatial heterogeneity of cellular coupling or other active membrane properties (i.e., ionic currents) appear to be critical.

At the macroscopic level, conduction through normal myocardial tissue is uniformly anisotropic; that is, propagation is continuous or “smooth” but is faster longitudinally than transversely. However, at higher spatial resolution, anisotropy is always non-uniform due to the irregularities of cell shape and gap junction distribution. The conversion of macroscopic anisotropy from being uniform to non-uniform is correlated with an increased predilection to arrhythmias. One well-studied example is the aged human atrial myocardium, in which non-uniform anisotropy, which manifests as highly fractionated electrograms, is associated with lateral uncoupling of myocytes and profound slowing of macroscopic transverse conduction, producing an ideal substrate for the re-entry that may underlie the very common development of atrial fibrillation in older adults.

Anatomically determined, excitable gap re-entry can explain several clinically important tachycardias such as AV re-entry, atrial flutter, and bundle branch re-entry tachycardia. Strong evidence suggests that arrhythmias such as atrial and ventricular fibrillations associated with more complex activation of the heart are re-entrant. However, this type of re-entry (“functional”) is mechanistically distinct from excitable gap re-entry.

Reflection is a type of re-entry that occurs in a linear segment of tissue (e.g., trabecula or Purkinje fiber) containing an area of conduction blockade with re-excitation occurring over the same segment of tissue. If the region of the segment proximal to the area of the blockade is excited, the wave will propagate and generate action potentials up to the area of conduction blockade. Assuming that the area of conduction blockade remains connected to the remainder of the tissue (by gap junctions), it can be electrotonically activated (i.e., by current flow without action potential induction). If the area of conduction blockade is short and the magnitude of the electrotonic current (source) is sufficiently large, the segment of tissue distal to the blocked area (sink) will be excited but with a significant delay. With the appropriate relationship of the electronic current transmitted through the inexcitable segment and the distal excitable tissue, not only can the distal segment be activated, but it can reactivate the proximal segment of muscle by electronic current flow from the distal segment to the proximal segment.

A key feature in classifying re-entrant arrhythmias, particularly for therapy, is the presence and size of an excitable gap. An excitable gap exists when the tachycardia circuit is longer than the tachycardia wavelength (λ = conduction velocity × refractory period), allowing appropriately timed stimuli to reset the propagation in the circuit. Re-entrant arrhythmias may exist in the heart in the absence of an excitable gap and with a tachycardia wavelength nearly the same size as the path length. In this case, the wavefront propagates through partially refractory tissue with no anatomic obstacle and no fully excitable gap. This is referred to as leading circle re-entry,⁴⁰ which is a form of functional re-entry (re-entry that depends on functional properties of the tissue). Unlike excitable gap re-entry, no fixed anatomic circuit exists in leading circle re-entry, and it may, therefore, not be possible to disrupt the tachycardia with pacing or destruction of a part of the circuit. Furthermore, the circuit in leading circle re-entry tends to be less stable than that in excitable gap re-entrant arrhythmias, with large variations in cycle length and predilection to termination. Atrial flutter represents an example of a re-entrant tachycardia with a large excitable gap not always due to an anatomic constraint but to functional blockade (reflecting the special properties of the crista terminalis discussed above). Experimental data and computer simulations have highlighted the shortcomings of tenets on leading circle re-entry and suggest that spiral waves may better explain some forms of functional re-entry.

Tissue anisotropy is another important determinant of functional re-entrant arrhythmias in ischemic heart disease. Changes in functional and anatomic anisotropy are characteristics of both acute and chronic ischemic heart disease. Within 30 minutes of the onset of myocardial ischemia significant increases in gap junction channel resistance and packing are observed. Further cellular uncoupling and a significant reduction in gap junction protein are observed with 60 minutes of ischemia; this coincides with irreversible cellular damage. These changes exaggerate anisotropic conduction in the ischemic zone.

Chronically ischemic, but not infarcted, myocardium also exhibits ~50% downregulation of gap junction protein (connexin43) with a significant change in the pattern or number of intercalated discs. The suggestion that a 50% reduction in gap junction protein influences anisotropic conduction is supported by measurements of conduction velocities in heterozygous connexin43 knockout mice. The border zones of infarcted myocardium exhibit not only functional alterations of ionic currents but remodeling of tissue and altered distribution of gap junctions in the human ventricle and infarction in canine heart. The alterations in gap junction expression in the context of macroscopic tissue alterations support a role for anisotropic conduction in re-entrant arrhythmias that complicate coronary artery disease. Altered expression of proteins at the intercalated disc is also seen in arrhythmogenic right ventricular dysplasia, a congenital arrhythmia syndrome, and similar mechanisms may thus underlie the development of monomorphic ventricular tachycardia in that entity.