Patch Clamping

Much of the understanding of the molecular mechanisms behind the action potential has derived from the development and implementation of three techniques: patch clamping, a technique that allows recording of ion flow through individual channel molecules, voltage clamping of isolated cardiac cells, and cloning and heterologous expression of ion channel genes. Comparison of ionic currents recorded in isolated myocytes with currents recorded from cells expressing ion channel genes has resulted in the identification of many of the ion channel molecules that underlie the cardiac action potential.

The development of the voltage clamp technique in the early 1950s and its application to multicellular preparations of cardiac muscle allowed identification of the major ionic currents that underlie the cardiac action potential. The whole-cell currents recorded by the technique were smooth waveforms derived from summation of the activity of thousands of ion channels, and many different patterns of events at the single-channel level, arising from more than one molecular specie, can summate to produce identical whole-cell current waveforms. Patch clamping allows resolution of events at the single-channel level. In this technique, small patches of cell membrane (< 1 μm²) are isolated electrically and physically in the tip of a glass micropipette.² Single-channel events can then be resolved because there are only a few ion channel molecules present in the patch. Current flowing across the patch typically jumps between well-defined values corresponding to sudden opening or closing of the ion-conducting pore (Figure 7-1A). The whole-cell current will be the sum of the currents through all of the individual channels in the cell membrane; summation of the current flowing through a single channel during repeated stimuli will reproduce the macroscopic whole-cell current (see Figures 7-1A and B). As channel opening and closing in response to a stimulus are a stochastic phenomena, the regulation of ion flow, whether resulting from a change in membrane potential or from interaction with regulator molecules, is usually achieved by increasing the probability that the channel will be open. Thus, in Figure 7-1, which shows records from human cardiac muscle Na⁺ channels, the channels open (activate) a few milliseconds after depolarization because the probability that the channel will be open increases. Similarly, as the channels spontaneously close (inactivate), the open probability decreases.

Figure 7-1 Voltage clamp recordings from muscle Na⁺ channels.

A, Patch-clamp recordings. Individual Na⁺ channel openings are recorded as short-lived downward deflections. Note that openings are concentrated in the first part of the trace. The top five traces represent a single depolarizing clamp pulse. The sum of 100 such pulses (bottom trace) recreates the whole-cell current (B).

An ion current with distinct electrical and pharmacologic properties indicates the presence of a population of identical ion channel molecules. Application of molecular techniques has allowed the identification of many ion channel molecules, and thus a better understanding of the currents that underlie the cardiac action potential. An as-yet unrealized dream would be tailoring of pharmacologic agents that interact with specific channel types to shape the action potential.

Electrical Events Underlying the Cardiac Action Potential

Figure 7-2 shows a diagram of a cardiac action potential with a summary of the ionic currents flowing during each phase. As far as has been determined, the probable molecular identity of the ion channels that underlie these currents is also given. This section examines the biophysical properties of these currents; subsequent sections focus on possible molecular mechanisms underlying the biophysical phenomena.

Figure 7-2 Ionic currents underlying the cardiac action potential.

Currents are listed on the left, and the ion channel genes encoding the currents are listed on the right. 4-AP, 4-aminopyridine; CFTR, cystic fibrosis transmembrane regulator.

(From Roden DM, Lazzarra R, Rosen R, et al, for the SADS Foundation Task Force on LQTS: Multiple mechanisms in the long QT syndrome: Current knowledge, gaps and future directions. Circulation 94:1996, 1996.)

Phase 1: Early Rapid Repolarization

The early rapid repolarization phase of the action potential, which follows immediately after phase 0, results both from rapid inactivation of the majority of the Na⁺ current and from activation of a transient outward current (I_TO), carried mainly by K⁺ ions. On depolarization of the membrane, I_TO opens rapidly, over about 20 msec, before spontaneously inactivating. I_TO comprises two separate currents: the rapidly inactivating I_TO1, which is activated by depolarization and blocked by 4-aminopyridine, and the slowly inactivating I_TO2, which is activated by elevated intracellular Ca⁺⁺ (possibly explaining the observation that action potential duration tends to decrease with rapid heart rates and hypercalcemia)^6,7 (Figure 7-3).

Figure 7-3 Voltage clamp recordings from an atrial myocyte showing pharmacologic separation of I_TO1 (ito1) and I_TO2 (ito2).

In the presence of cobalt ions, only I_TO1 is seen. When cobalt is omitted and 4-aminopyridine (4-AP) is added, I_TO2 is revealed. Caffeine (Caf) eliminates I_TO2, leaving the underlying inward calcium current. In the presence of both 4-AP and cobalt ions, all outward currents are inhibited.

(From Wang Z, Fermini B, Nattel S: Delayed rectifier outward current and repolarization in human atrial myocytes. Circ Res 73:276, 1993.)

In addition to its effect on phase 1, I_TO, in combination with the delayed rectifier potassium currents (I_Kr and I_Ks) and I_K1, also contributes to membrane repolarization. Arrhythmogenic prolongation of the action potential in myocardial cells recovered from patients with myocardial hypertrophy,⁸ congestive cardiomyopathy,⁶ and from the border zone of myocardial infarction in animals⁹ appears to result from depression of I_TO.

Phases 2 and 3: Plateau Phase and Final Rapid Repolarization

The action potential plateau and final rapid repolarization are mediated by a balance between the slow inward current and outward, predominantly K⁺ current. During the plateau phase, membrane conductance to all ions falls and very little current flows. Potassium conductance is low because of inward rectification of I_K1 (i.e., inward current passes more easily than outward current), so little outward current flows despite the large outward electrochemical gradient for K⁺ ions and the delayed onset of the outwardly rectifying K⁺ currents (I_Ks, I_Kr, and I_Kur). The resulting small outward current is balanced by inward current, predominantly through L-type Ca⁺⁺ channels (I_Ca-L), but also via a slowly inactivating population of Na⁺ channels, and a small inward flux of chloride (Cl^–) ions, possibly carried by the cardiac variant of the ATP-dependent channel (abnormalities of which underlie cystic fibrosis).^10,11 Phase 3, regenerative rapid repolarization, results from time-dependent inactivation of L-type Ca⁺⁺ current and increasing outward current through delayed rectifier K⁺ channels. The net membrane current becomes outward and the cell repolarizes.

Slow Inward Ca⁺⁺ Current

The slow inward current (I_Ca-L) is activated by depolarization of the cell to potentials less negative than −40 to −50 mV. In ventricular and atrial myocytes and in Purkinje fibers, I_Ca-L is activated by the regenerative depolarization caused by I_Na during phase 0 of the action potential. I_Ca-L does not contribute significantly to phase 0 because, in comparison with I_Na, it activates much more slowly (over about 10 msec) and is smaller in amplitude. I_Ca-L also inactivates slowly and, therefore, contributes the major inward current during the plateau of the action potential. I_Ca-L flows through L-type (long-lasting) Ca⁺⁺ channels, which are sensitive to block by dihydropyridines (e.g., nifedipine), and activation of contraction is related to the magnitude of the resulting calcium influx.¹² Gating of I_Ca-L is generally similar to I_Na in that channel opening and closing are dependent on membrane potential and time. Ca⁺⁺ channels are also, importantly, dynamically regulated by the autonomic nervous system.¹³ β agonists activate I_Ca-L (and hence increase myocardial contractility) indirectly by activating adenylyl cyclase via a guanosine triphosphate (GTP)–binding protein, G_s (Figure 7-4). The resulting increase in intracellular cyclic adenosine monophosphate (cAMP) activates protein kinase A (PKA), which phosphorylates the Ca⁺⁺ channel. Phosphorylated channels open in response to membrane depolarization; nonphosphorylated channels do not, so the effect of β-adrenergic stimulation is to increase the number of functional channels. The electrophysiologic effect of this is illustrated in Figure 7-5, which shows enhancement of the slow inward current by increase of adenosine monophosphate (AMP) level in single-channel Ca⁺⁺ channels and intact cells. β-Adrenergic effects on I_Ca-L are antagonized by acetylcholine, which, in myocardial cells, activates M₂ muscarinic receptors and inhibits adenylyl cyclase through activation of the GTP-binding protein G_i.

Figure 7-4 Autonomic regulation of ion currents.

ACh, acetylcholine; 5′AMP, adenosine 5′ monophosphate; β-AR, β-adrenergic receptor; ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; Cyclase, adenylyl cyclase; Gi and Gs, guanosine triphosphate–binding proteins; ISO, isoproterenol; mAChR, M2 muscarinic receptor; Pi, pyrophosphate; R and C, regulatory and catalytic subunits, respectively, of protein kinase A.

Figure 7-5 Effects of elevated cyclic adenosine monophosphate (cAMP) on cardiac calcium channels.

A, Single-channel recordings; B, whole-cell currents. Left panels are control recordings; right panels show the effect of elevation of intracellular cAMP (in this case, induced by exposure to parathyroid hormone). Note the increased probability of channel opening and resulting increase in whole-cell current in the presence of increased cAMP.

(A, B, Modified from Rampe D, Lacerda AE, Dage RC, Brown AM: Parathyroid hormone: An endogenous modulator of cardiac calcium channels. Am J Physiol 261[6 Pt 2]:H1945, 1991.)

In the relatively depolarized pacemaker cells, which lack I_K1, I_Na is inactivated and the slow inward current is solely responsible for the upstroke of the action potential. I_Ca-L also can generate slowly propagated action potentials in diseased or damaged myocardial cells in which I_Na has been inactivated by depolarization. These slow responses, which may occur in the border zone of myocardial infarcts, are important because they may cause the slow conduction that can lead to reentrant arrhythmias.

Delayed Rectifier K⁺ Currents

Delayed rectifier K⁺ channels are present in all cardiac myocytes. They open slowly (over 200 to 300 msec) after depolarization of the membrane to the plateau level (−10 mV and greater), producing a K⁺-selective outward current, I_K. I_K does not inactivate on prolonged depolarization (unlike I_Na and I_Ca-L), and the channels close on repolarization of the membrane. Unlike I_K1, I_K displays outward rectification; that is, it passes outward current more easily than inward current. This is the expected behavior for a K⁺-selective channel because both the concentration and electrical gradients for potassium are outward. Thus, for any depolarizing displacement of membrane potential from E_K, the driving force will be larger in an outward direction. Similar to I_Ca-L, I_K is under autonomic control (see Figure 7-4). β-Adrenergic stimulation enhances I_K by a mechanism similar to that of the enhancement of I_Ca-L, thus ensuring repolarization of the cell in the face of increased inward Ca⁺⁺ current.¹⁴

Three components of I_K, carried by different channel molecules, can be distinguished. A rapidly activating component, I_Kr, is blocked by the compound E4031 (a Class III antiarrhythmic agent), which leaves a slower activating component, I_Ks, unaffected.¹⁵ This is illustrated in Figure 7-6, which also emphasizes the importance of I_K in the regulation of repolarization, and hence of action potential duration. A third component, the ultra-rapidly activated delayed rectifier, I_Kur, can be distinguished in atrial (but not ventricular) myocytes.¹⁶ This additional repolarizing current explains, in part, the enhancement of repolarization in atrial myocardium when compared with ventricle and Purkinje fibers.

Figure 7-6 A, Effect on action potential duration of blockade of I_Kr by E4031 in isolated guinea pig ventricular myocytes. After application of E4031, outward repolarizing currents are inhibited and action potential duration is increased. B, Voltage clamp records showing separation of I_Kr and I_Ks by E4031. The control delayed rectifier K⁺ current (C) is partially inhibited by E4031 (E), and the difference between the two currents (DIF) represents I_Ks.

(A, B, From Sanguinetti MC, Jurkiewicz NK: Two components of cardiac delayed rectifier K⁺ current. Differential sensitivity to block by class III antiarrhythmic agents. J Gen Physiol 96:195, 1990. Reproduced from The Journal of General Physiology by copyright permission of The Rockefeller University Press.)

Repolarization in Different Cardiac Tissue Types

Phase 3 repolarization in atrium and pacemaker tissues, but not in ventricular myocardium, is also enhanced by the presence of a large outward repolarizing K⁺ current (I_K[ACh]).^17,18 This potential independent current is activated indirectly by stimulation of muscarinic (M2-type) receptors by acetylcholine or purinergic (A-type) receptors by adenosine.¹⁹ This channel is potential independent and is activated via binding of an activated, membrane-bound GTP-binding protein (G_i), as discussed later.²⁰

There is variability in action potential duration between cells in normal ventricle.²¹ A gradient in action potential duration exists across the myocardium (from epicardium to endocardium), and specialized midmyocardial cells (M cells) have been identified, which exhibit prolongation of action potential duration at slow stimulation rates, possibly as a result of a decrease in I_Ks.

Phase 4: Diastolic Depolarization and I_f

Phase 4 diastolic depolarization, or normal automaticity, is a normal feature of cardiac cells in the sinus and atrioventricular (AV) nodes, but subsidiary pacemaker activity is also observed in the His-Purkinje system and in some specialized atrial and ventricular myocardial cells (see Chapter 4). Pacemaker discharge from the sinus node normally predominates because the rate of diastolic depolarization in the sinoatrial (SA) node is faster than in other pacemaker tissues. Pacemaker activity results from a slow net gain of positive charge, which depolarizes the cell from its maximal diastolic potential to threshold.

Pacemaker cells in the sinus node are relatively depolarized, with a maximal diastolic potential of −60 to −70 mV and a threshold potential of −40 mV. Rapid regenerative depolarization (phase 0) is dependent on opening of T-type and then L-type Ca⁺⁺ channels. Repolarization is dependent on activation of delayed rectifier K⁺ channels, and the maximum diastolic potential is around −80 mV. Pacemaker channels are activated by hyperpolarization to this potential and produce a slow inward Na⁺ current, I_f. This flows against slowly inactivating delayed rectifier K⁺ currents and results in diastolic depolarization.²² Because the current is nonselective among cations, its reversal potential lies between E_K and E_Na, at around −10 mV, and activation of I_f will tend to depolarize the cell toward this value. Similar to I_Ca-L, I_f is under autonomic control (see Figure 7-4) through GTP-dependent binding proteins G_s and G_i, which regulate cAMP production by adenylyl cyclase.^23,24 β-Adrenergic stimulation shifts the voltage dependence of activation of I_f to more depolarized potentials, so for any hyperpolarizing stimulus, more I_f will be activated and diastolic depolarization will be enhanced. Acetylcholine has the opposite effect (Figure 7-7).

Figure 7-7 Effect of acetylcholine on spontaneous pacemaker activity in isolated sinoatrial (SA) node cells. Pacemaker activity in control solution is compared with activity in increasing doses of acetylcholine (ACh). Note the slowing of diastolic depolarization and resultant slowing of heart rate.

(From DiFrancesco D: Current I and the neuronal modulation of heart rate. In Zipes DP, Jaliffe J [eds]: Cardiac electrophysiology from cell to bedside. Philadelphia: WB Saunders Company, 1990, p 28.)

Molecular Biology of Ion Channels

The preceding sections have focused on the electrical events that underlie cardiac electrical excitability, and on the identification of cardiac ionic currents on the basis of their biophysical properties. Here the molecular structures behind these electrical phenomena are reviewed. The first step in understanding the molecular physiology of cardiac electrical excitability is to identify the ion channel proteins responsible for the ionic currents. Figure 7-2 gives the current classification of the ion channel responsible for each of the cardiac ionic currents. There are firm molecular candidates for voltage-gated Na⁺ and L-type Ca⁺⁺ channels. Similarly, channel molecules with properties similar to delayed rectifier K⁺ channels, the 4-aminopyridine–sensitive component of I_TO, the inward rectifier I_K1, the ligand-gated K⁺ channel I_K(ACh), and the pacemaker current I_f have been cloned. Figure 7-8 shows diagrams of the predicted membrane topology of some of these channels. Voltage-gated Na⁺, Ca⁺⁺, and K⁺ channels exist as conglomerates of molecules, consisting of a large α subunit and several accessory subunits (labeled β, δ, and γ in Figure 7-8). The α subunit alone is usually sufficient to induce channel activity in biologic membranes, but its activity is modulated by the presence of the accessory subunits. The diagrams in Figure 7-8 were deduced from hydrophobicity analysis of the primary structure of the major channel polypeptides. Regions of the polypeptides predicted to span the membrane are those that contain a high concentration of hydrophobic amino acids, whereas peptides linking these transmembrane sections are hydrophilic. Similarities among the various channels strongly suggest a common evolutionary ancestry. Na⁺ and Ca⁺⁺ channel α subunits (see Figures 7-8A and B) are strikingly similar, each consisting of four homologous transmembrane domains (labeled I to IV), linked by cytoplasmic peptides. Each homologous domain contains six linked membrane-spanning segments (labeled S1 to S6). These large polypeptides, containing more than 2000 amino acids, form a tetrameric structure and generate Na⁺ or Ca⁺⁺ channel activity in biologic membranes. Voltage-gated K⁺ channel α subunits, in contrast, are much smaller (see Figure 7-8C) and consist of a single transmembrane domain with six membrane-spanning segments, an arrangement similar to one of the individual domains of Na⁺ and Ca⁺⁺ channels. Four molecules are noncovalently linked in the membrane to produce a tetrameric structure, similar to a Na⁺ or Ca⁺⁺ channel, to produce K⁺ channel activity. The structure of the inwardly rectifying K⁺ channel molecules, I_K1 and I_K(ACh), is dissimilar to other K⁺ channels (see Figure 7-8C). The molecules are much less complex, having only two membrane-spanning segments, although these segments share considerable homology with the S5 and S6 segments of the classic voltage-gated K⁺ channel.

Figure 7-8 Diagrams of ion channel molecular structure.

A, Na⁺ channel. B, Ca⁺⁺ channel. C, K⁺ channels. ATP, adenosine triphosphate.

Voltage-gated ion channel activity requires that the channel molecule should sense and respond to changes in membrane potential, form an ion-selective membrane pore, and (in some cases) inactivate despite continuing depolarization. The molecular mechanisms for these phenomena are examined in separate sections later.

Molecular Mechanisms

The Voltage Sensor

Channel proteins respond to changes in electrical potential across the cell membrane by conformational changes (gating), which result from electrostatic interactions between charged portions of the molecule and the membrane electric field (Box 7-3). Gating of the channel is associated with a measurable flow of electrical charge through the membrane lipid bilayer (called gating current), as a zone of the molecule rich in electrical charge moves within the membrane.²⁵ This charge movement is linked to opening of the channel pore. The voltage sensor of voltage-dependent ion channels resides in the mobile S4 membrane-spanning segments, α-helical structures unusually rich in positively charged amino acids.²⁶ At rest, each of the positive charges in the S4 segment is balanced by fixed negative charges in other segments of the molecule. The resting membrane potential (negative inside) forces the (mobile) positive charges inward and the fixed negative charges outward. This dynamic equilibrium holds the channel pore closed. On depolarization, the force pulling the positive charge inward is relieved; positive charges (the S4 segments) are repelled outward and assume new partners with the fixed negative membrane charges. This charge movement comprises the gating current. If the depolarizing stimulus is short, repolarization of the membrane is followed by a gating current of equal and opposite magnitude as the S4 segment relaxes to its original position. If the depolarizing stimulus is prolonged, however, the movement of the S4 segments induces a conformational change in the channel molecule, which prohibits easy return to baseline. This conformational change in the channel molecule is manifested as activation (or channel opening), and this is closely coupled to channel closing (or inactivation; see later) in channels that inactivate. Thus, small changes in the membrane electric field cause conformational changes in the channel molecule, which result in opening (and closing) of the channel pore. An S4 segment rich in positive charge is a remarkably consistent feature of voltage-gated ion channels from a variety of different species and with a variety of ion selectivities. The dependence of channel activation on membrane potential is proportional to the density of positive charge in the S4 segment.

BOX 7-3 Molecular Mechanisms of Ion Channels

• Voltage sensor

• Gating mechanism (activation and inactivation)

• Ion pore

• Selectivity filter

Ion Channel Pore and Selectivity Filter

The presence of four homologous domains in voltage-gated Na⁺ and Ca⁺⁺ channels suggests that basic ion channel architecture consists of a transmembrane pore surrounded by the four homologous domains arranged symmetrically (see Figure 7-8). The membrane-spanning segments each form an α helix so that the walls of the pore will be derived from α-helical segments from each of the four domains. A pore formed from four such α helices would have limiting dimensions of 3 by 5 Angstrom units, similar to the size inferred for the Na⁺ channel pore by measurement of the permeability of cations of different sizes.^27,28

The selectivity filter is formed by the S5 and S6 membrane-spanning segments of each domain together with their peptide linker.²⁹ As emphasized in Figure 7-8, unlike the hydrophilic extracellular linkers between other membrane-spanning segments, the S5/S6 linker is sufficiently hydrophobic to place it (at least partially) within the membrane lipid bilayer. The channel pore is lined both by the S5/S6 linker and the S5 and S6 membrane-spanning segments. Point mutations in the S5/S6 linker have dramatic effects on channel ion selectivity and reduce channel conductance to its primary ion. Extensive site-directed mutagenesis experiments of the S5/S6 linkers from a variety of channels suggest that they form a funnel that allows the passage of a specific ion into the pore. In Na⁺ channels, selectivity is imposed by two rings of negatively charged amino acids at the outer mouth of the funnel, which collect Na⁺ ions for transmission into the cell.²⁷

Channel Inactivation

Inactivation gating is the process by which ion channels close in the face of continuing depolarization. Inactivation is characteristic of voltage-gated Na⁺ and Ca⁺⁺ channels, as well as the K⁺ channels underlying I_TO. Inactivation begins after activation gating, as a second, slower conformational change in the molecule that halts the ion flux through the channel. Inactivation gating is thus closely coupled to activation gating, and ionic current flows only while both the activation and inactivation gates are open simultaneously. In Na⁺ channels, the inactivation gate is formed by the intracellular peptide linker between homologous domains III and IV (Figure 7-9A).³⁰ This peptide is postulated to act as a hinged lid, which moves upward to plug the ion pore (and thus halt current flow) shortly after membrane depolarization.²⁶ For the channel to recover from inactivation (i.e., to be ready to open in response to a new depolarizing stimulus), the III/IV linker peptide must resume its resting position, a process that requires hyperpolarization of the membrane to the resting potential for a finite period. Site-directed mutagenesis of the III/IV linker peptide has revealed a trio of hydrophobic amino acid residues (isoleucine, I; phenylalanine, F; and methionine, M), near to the domain III end of the peptide, which are crucial for normal channel inactivation. Replacement of just one of these residues (the phenylalanine) almost completely removes inactivation. These residues are postulated to latch on to a receptor in the channel pore to close the channel. The molecular basis of inactivation in K⁺ channels is rather different from Na⁺ channels. Because the four domains of K⁺ channels are formed by noncovalently linked molecules, there are no interdomain linkers to plug the channel pore. In K⁺ channels, a picture of an N-terminal ball-and-chain mechanism has emerged (see Figure 7-9B).³¹ The terminal 20 or so amino acids are very hydrophobic and are postulated to swing up and attach to the open pore. The next few amino acids contain a number of positively charged residues that draw the whole N-terminal end up to the membrane. These two domains act as a ball. The remaining amino acids, up to the beginning of the transmembrane S1 segment, act as a chain. If the chain is made longer, inactivation is slower, and vice versa.

Figure 7-9 Mechanism of inactivation of Na⁺ and K⁺ channels.

A, Hinged-lid mechanism of Na⁺-channel inactivation. B, Ball-and-chain mechanism of K⁺-channel inactivation. F, phenylalanine; I, isoleucine; M, methionine.

(A, B, From Catterall WA: Structure and function of voltage-gated ion channels. Annu Rev Biochem 64:493, 1995, by permission of the Annual Review of Biochemistry, Volume 64, ©1995, by Annual Reviews, Inc.)

Clinical Correlates

Receptors

Receptors are grouped in several broad classes, the protein tyrosine kinase receptors and the G-protein–coupled receptors (GPCRs) being the most important ones. The protein tyrosine kinase receptors are large molecular complexes that incorporate phosphorylating enzyme activity in the intracellular segment. Ligand binding induces activation of this enzyme activity. Because phosphorylation is one of the major mechanisms of cellular regulation (see, for example, the phosphorylation of the Ca⁺⁺ channel described earlier), such receptors can have a variety of cellular effects (Box 7-4). GPCRs are much smaller than protein tyrosine kinase receptors. Ligand binding results in activation of an associated protein (G protein) that subsequently influences cellular processes. The receptors discussed in this section all belong to the GPCR superfamily, and the properties of this class are discussed in some detail.

The number of GPCRs is large. For more than 100 receptors, the function has been defined. In addition, the olfactory epithelium expresses hundreds of GPCRs, which are thought to mediate the sense of smell, and another large group, with unknown function, is expressed on sperm cells. Taken together, the superfamily has more than 1000 members.

All of these have similar molecular characteristics. They are generally several hundred to 1000 amino acids in length and contain 7 stretches of 20 to 25 hydrophobic amino acids. These hydrophobic domains are thought to form α helices and traverse the membrane, thus anchoring the receptor to the cell (Figure 7-10). For this reason, the family is often referred to as the “seven transmembrane” family. Although crystallographic data are not yet available for the clinically relevant GPCRs, it is thought that the seven-transmembrane domains arrange in a funnel-like structure, the inside of which forms the ligand-binding domain. The intracellular domains, particularly the third intracellular loop and the C terminus, bind to the G protein.

Figure 7-10 Model of G-protein–coupled receptor.

A, Linear model. Seven hydrophobic stretches of approximately 20 amino acids are present, presumably forming α helices that pass through the cell membrane, thus forming seven transmembrane domains. (t1–t7). Extracellularly, the amino terminus (N) and three outside loops (o1 through o3) are found; intracellularly there are similarly three loops (i1 through i3) and the carboxy-terminus (C). B, Top-down view. Although in (A), the molecule is pictured as a linear complex, the transmembrane domains are thought to be in close proximity, forming an ellipse with a central ligand-binding cavity (dashed circle). Asp and Tyr refer to two amino acids important for ligand interaction. G-protein binding takes place at the i3 loop and the carboxy terminus.

The heart and blood vessels express a variety of GPCRs, the most important of which are discussed later. The β-adrenergic and muscarinic acetylcholine receptors are those most important for regulation of cardiac functioning, but a number of others play relevant modulatory roles. These include the α-adrenergic, adenosine A₁, ATP, histamine H₂, vasoactive intestinal peptide (VIP), and angiotensin II receptors.

G Proteins

As do the receptors, G proteins (GTP-binding proteins) also come in two families: the small (cytoplasmic) G proteins and the heterotrimeric (membrane) G proteins. Common to both groups is their mechanism of function. In their resting state, they bind a molecule of guanosine diphosphate (GDP). When activated by a GPCR (in the case of heterotrimeric G proteins) or by an intracellular messenger (in the case of cytoplasmic G proteins), this GDP is exchanged for a GTP molecule. The activated G protein can now perform functions within the cell, as discussed later, until it is inactivated when an intrinsic enzyme activity hydrolyzes the GTP to a GDP. The critical point about this hydrolytic activity is that it is (molecularly speaking) very slow, on the order of seconds. As a result, brief activation of a receptor (on the order of milliseconds) can lead to more prolonged activation of the intracellular signaling machinery.

GPCRs bind to heterotrimeric G proteins, so called because they consist of three subunits: α, β, and γ. Of these, the β and γ subunits are so tightly associated that, for practical purposes, they can be viewed as a single unit, often termed the βγ unit. The α subunit contains both the GDP-GTP binding domain and the hydrolytic activity, and it was classically thought to be the “business end” of the molecule, with the βγ unit roaming freely and inactively, serving as an anchor and a sink of free α units. This turns out not to be the case; the βγ subunit has activating functions as well, as is discussed later in relation to the muscarinic K⁺ channel.

Several classes of heterotrimeric G proteins exist, indicated by subscripts (Box 7-5). The classic types are G_s and G_i, which stimulate and inhibit, respectively, the enzyme adenylate cyclase, thereby leading to changes in cytoplasmic cAMP concentrations. G_q proteins (and G_o in brain) activate phospholipase C (PLC) and thereby induce the generation of inositol-1,4,5-triphosphate (IP₃) and diacylglycerol (DAG) from phosphatidylinositol bisphosphate (PIP₂). IP₃ acts on its own receptor/channel complex on intracellular Ca⁺⁺ storage sites and induces release of Ca⁺⁺ from these sites, thereby increasing intracellular Ca⁺⁺ concentrations. DAG activates protein kinase C (PKC), leading to phosphorylation of a variety of targets (including the receptors that initiated the cascade). In recent years, cloning efforts have shown each of these classes of G proteins to consist of a number of members, but their functional differences are as yet incompletely defined.

Adrenergic Receptors and Signaling Pathways

Muscarinic Receptors and Signaling Pathways

Regulation of G-Protein Functioning

In view of the profound changes in GPCR expression that occur in various disease states, the expression and function of G proteins in cardiovascular disease have been studied with interest. G_s proteins appear unchanged in cardiac failure, both in expression level and in function. With G_i proteins, the situation is more interesting, because G_iα is considered to have a secondary role in addition to its inhibition of adenylate cyclase. Under normal conditions, G_i is present in greater amounts than G_s. Activation of receptors coupled to G_i would, therefore, lead to the release of a large number of free βγ subunits. These could combine with any free G_sα, thereby making it unavailable for activation of adenylate cyclase.⁵² In addition, these βγ units can enhance the phosphorylation of β receptors by β-ARK.⁵³ In failing human cardiac tissue, the amount of G_i (as assessed by ADP-ribosylation by pertussis toxin) is increased.⁵⁴ Although this would be expected to make muscarinic signaling more efficacious (thereby helping counteract the adrenergic overstimulation), it has been difficult to correlate these changes with alterations in adenylate cyclase functioning. It is similarly unclear whether the reported increases in G_i levels are a result of increased mRNA expression or increased stability of the G protein itself.

The catalytic subunit of adenylate cyclase appears little influenced by cardiac disease. Pressure overload is the only situation in which a consistent decrease in its activity has been observed.⁵⁵

Other Receptors

As stated earlier, the heart and vasculature express a large variety of GPCRs apart from the adrenergic and muscarinic receptors. A few examples are mentioned here. Angiotensin receptors mediate hormonal vasoconstriction in the vascular tree and are also present in the heart, although their function there is not fully defined. Receptors for several purinergic compounds are expressed in the heart as well and are the subject of intense investigation, as discussed in the next section. In addition, histamine H₂ and VIP receptors are present, the former mediating the inotropic action of histamine.

Although less is understood about the role and regulation of these receptors than about their adrenergic and cholinergic counterparts, it is clear that some are affected by cardiovascular disease. For example, VIP receptors are downregulated by 70% in idiopathic dilated cardiomyopathy (DCM), whereas histamine receptors are unaffected.⁵⁶ In general, receptors coupled to G_i show little alteration in expression and function during disease states, whereas G_s-coupled receptors are affected more profoundly.

Clinical Correlates

Understanding of the role of adenosine in cardiac regulation has expanded significantly over the past years. Its established use as an antiarrhythmic compound and its probable role in cardiac preconditioning are two examples of clinical advances resulting from this increase in understanding. Adenosine acts through a GPCR, activating several intracellular signaling systems. This section discusses the molecular aspects of adenosine signaling, as well as their clinical implications. More detailed recent reviews on the topic are available.^57,58

Adenosine Signaling

Although adenosine can be generated by several pathways, in the heart, it is usually found as a dephosphorylation product of AMP.⁵⁹ Because AMP accumulation is a sign of a low cellular energy charge, an increased adenosine concentration is a marker of unbalanced energy demand and supply; thus, ischemia, hypoxemia, and increased catecholamine concentrations are all associated with increased adenosine release.⁶⁰ Adenosine is rapidly degraded by various pathways, both intracellularly and extracellularly. As a result, its half-life is extremely short, on the order of 1 second.⁶¹ Therefore, it is not only a marker of a cardiac “energy crisis,” but its concentrations will fluctuate virtually instantly with the energy balance of the heart; it provides a real-time indication of the cellular energy situation.

Adenosine signals through GPCR of the purinergic receptor family. Two subclasses of purinoceptors exist: P₁ (high affinity for adenosine and AMP) and P₂ (high affinity for ATP and ADP). The P₁ receptor class can be divided into two main receptor subtypes: A₁ and A₂. A₁ receptors are present mostly in the heart and, when activated, inhibit adenylate cyclase; A₂ receptors are present in the vasculature and, when activated, stimulate adenylate cyclase. The A₂ receptors mediate the vasodilatory actions of adenosine. The A₁ receptors mediate its complex cardiac effects, and they are the topic of the remainder of this section.

The A₁ adenosine receptor couples to (at least) two intracellular signaling systems. Both of these actions are mediated by G proteins of the G_i class. The first intracellular system is one already encountered: the K_ACh channel. Presumably, through the same G_K protein, adenosine activates this channel in the same way as does M2 muscarinic stimulation, and the cardiac electrophysiologic effects of acetylcholine and adenosine are therefore quite similar. The specific effect of adenosine depends on the cardiac tissue studied because K_ACh expression varies with location. As has been discussed, whereas the channel is present in large amounts in the atrial conduction system and atrial myocardium, it is virtually absent in the ventricle. Therefore, in the unstimulated heart, adenosine shortens the atrial action potential, decreases atrial refractoriness and decreases atrial contractile force, but it is almost without effect on the ventricle.⁶²

The second intracellular signaling system activated is a G_i protein that inhibits adenylate cyclase. As cAMP levels are quite low under resting conditions, this mechanism plays little role until cAMP concentrations are increased by adrenergic stimulation of the heart. Therefore, cAMP-mediated cardiac actions of adenosine are observed only under conditions of adrenergic drive. Because the adenylate cyclase system is present throughout the heart, its effects are widespread; L-type Ca⁺⁺ channel functioning is diminished (by inhibiting cAMP-induced phosphorylation of the channel) in atrium as well as ventricle, resulting in decreased inotropy and shortening of the action potential.⁶³

Interactions with Channels: Ca⁺⁺ Channels

Of the variety of ion channels present in the heart, those most likely to be significantly affected by anesthetics in the clinical setting are the voltage-gated Ca⁺⁺ channels. Although interactions with other cardiac channel types may also be relevant (particularly with K⁺ channels, such as the K_ATP channel, described earlier), they have not yet received as much research attention, and it is not yet possible to draw firm conclusions about the clinical relevance of these interactions. Hence, in this section, the focus is on Ca⁺⁺ channels only.

Anesthetic actions on cardiac Ca⁺⁺ channels have been studied in a variety of models. The original observations that halothane blocked Ca⁺⁺ flux into heart cells date back approximately 30 years,⁷⁴ and much specific information has been gained since. In particular, voltage-clamp and patch-clamp studies have contributed significantly to the understanding of the interactions between anesthetics and Ca⁺⁺ channels, and have elegantly described the effects of anesthetics on electrophysiologic behavior. However, it is not straightforward to assign the observed electrophysiologic effects to a molecular substrate. This issue is only beginning to be addressed with the use of recombinant technology.

Almost all volatile anesthetics inhibit L-type Ca⁺⁺ channels.^75,76 Inhibition is modest, approximately 25% to 30% at 1 MAC anesthetic, but certainly sufficient to account for the physiologic changes induced by the anesthetics. Volatile anesthetics decrease peak current and, in addition, tend to increase the rate of inactivation.⁷⁷ Hence maximal Ca⁺⁺ current is depressed, and duration of Ca⁺⁺ current is shortened. Together, these actions significantly limit the Ca⁺⁺ influx into the cardiac myocyte. However, some specific actions may depend on the particular compound studied. In the presence of β-adrenergic stimulation, halothane, but not sevoflurane, is associated with a long-lasting enhancement of Ca⁺⁺ channel function that may contribute to its proarrhythmic effects.⁷⁸ Xenon is without effect on cardiac Ca⁺⁺ channels, explaining, in large part, its lack of effect on myocardial contractility. Other types of Ca⁺⁺ channels have different sensitivities. Neuronal (N-type) channels have been shown to be quite resistant to volatile anesthetics. T-type channels in general tend to be much more sensitive than L-type channels; at clinical concentrations of most volatile anesthetics, T currents are inhibited 50% or more.

The effects of volatile anesthetics on cardiac Ca⁺⁺ channels can be modulated greatly by concurrent interactions of the compounds on other cardiac signaling systems. As discussed later, volatile anesthetics inhibit function of several types of muscarinic acetylcholine receptor systems. Because Ca channel function can be inhibited by muscarinic signaling, as discussed earlier, clinicians would anticipate additional interactions when this system is exposed to volatile anesthetics, and such is the case. Halothane and isoflurane further inhibited currents through Ca⁺⁺ channels when either volatile anesthetic was applied after inhibition produced by prior muscarinic stimulation. However, when muscarinic receptors were stimulated after administration of volatile anesthetic, its effect was reduced. Thus, whereas volatile anesthetics directly inhibit L-type channels, they also interfere with channel modulation by GPCR.⁷⁹

Not only volatile but also injected anesthetics have been reported to inhibit cardiac L-type Ca⁺⁺ channels in some models. However, the concentrations used generally exceed those used in clinical practice. Thiopental and methohexital block L-type Ca⁺⁺ currents.^76,80 Similarly, propofol has been reported to inhibit these channels, but at concentrations well beyond the clinical range.

Interactions with Receptors: Muscarinic Receptors

As with cardiac channels, anesthetic interactions with a variety of GPCRs may be of potential relevance to cardiovascular side effects of anesthetic compounds. However, most of these potential interactions have not been described in significant detail. The two main control systems of myocardial functioning, the muscarinic and adrenergic systems, have been studied to some extent, and a generalizing conclusion can be that muscarinic receptors are and adrenergic receptors are not sensitive to many anesthetics. Hence the focus is on muscarinic receptors in this section.

Determination of anesthetic effects on GPCR is in some ways more complex than determination of the effects on channels. Although the receptors are smaller and consist of a single subunit only, they are only one part of very complex signaling pathways. Therefore, it is not sufficient to determine the effects on the receptor itself; actions on intracellular signaling should be investigated as well.

There is no doubt that at least some volatile anesthetics interfere with muscarinic signaling. Unfortunately, most of these investigations have not studied the heart, or even looked specifically at the M2 receptor, the only subtype expressed in cardiac tissue.

Aronstam et al⁸¹ published a series of articles reporting the effect of various anesthetics on muscarinic receptor binding. A summary of their findings has appeared in print as well. Several studies investigated the effect of halothane on agonist and antagonist binding.^82,83 The conclusions drawn from this work were: (1) the anesthetic enhanced antagonist binding by slowing the rate of ligand dissociation, and (2) the anesthetic inhibited agonist binding (by 48%, using 10% halothane).

The site of action of these effects was studied in more detail by investigating the interactions of anesthetics with G-protein functioning.⁸⁴ As discussed earlier, G proteins are activated by GDP-GTP exchange. An intrinsic, but remarkably slow, enzyme activity hydrolyzes GTP back to GDP, thereby inactivating the G protein after several seconds. While active, the G protein is no longer able to couple to the receptor, and uncoupled receptors exhibit a decreased affinity for their agonist. This decrease in activity can be induced in the experimental setting by including a nonhydrolyzable analog of GTP [such as Gpp(NH)p] in the reaction mixture, resulting in irreversibly activated G proteins. This effect is known as the GTP shift in receptor affinity. Halothane shifts the Gpp(NH)p concentration–response relation to the right. In other words, a greater concentration of the GTP analog was necessary to induce a similar decrease in agonist binding. In the absence and presence of 5% halothane, the half-maximal inhibitory concentration (IC₅₀) values for the inhibitory effect on agonist binding of Gpp(NH)p were 0.7 and 83 μM, respectively—a 100-fold difference. These findings were interpreted as an ability of halothane to stabilize high-affinity G-protein–receptor complexes.

Therefore, it appears that halothane affects receptor binding, as well as receptor–G-protein interaction. In a subsequent study, the effect of the anesthetic on G-protein functioning—its ability to hydrolyze bound GTP—was investigated. Halothane was found not to inhibit binding of radiolabeled GTP analog to G proteins.⁸² However, the anesthetic completely blocked the stimulation of G-protein GTPase activity induced by acetylcholine, with a half-maximal effect at the clinically relevant halothane concentration of 0.3 mM. The site of this effect was not determined. Largely similar findings were obtained with other anesthetics. Halothane also has been shown to interfere with G-protein–mediated Ca⁺⁺ sensitization in airway smooth muscle by inhibiting G proteins.⁸⁵ More recently, however, the interaction between halothane and purified, recombinant G_i was investigated. In contrast with the findings described earlier, no effect of the anesthetic on G_i protein function was found.⁸⁶ Hence the earlier results might have been contaminated by interactions with other G-protein subtypes that are less relevant to muscarinic M2-receptor signaling. The conclusion to be drawn from these studies is that anesthetics may interfere with several components of the muscarinic signaling pathway. However, the following should be kept in mind: (1) most times a mixture of muscarinic receptor subtypes was studied; (2) the anesthetics were administered in relatively high, and clinically unequal, concentrations; and (3) anesthetic effects on functional properties of the receptor–G-protein unit were not specifically addressed. Magyar and Szabo,⁸⁷ using the patch-clamp technique, have investigated the effects of halothane (0.9 mM) and isoflurane (0.8 mM) on acetylcholine (10 μM)-induced activation of the muscarinic K⁺ channel in frog atrial myocytes. They found that if anesthetic and agonist were administered at the same time, a reduction in the peak K⁺ current was observed, which was greater with halothane than with isoflurane. However, pretreatment with the anesthetic was found to have a significant, time-dependent, additional effect: 25-minute exposure to halothane restored the peak and significantly increased the steady-state current, whereas exposure to isoflurane decreased both. As equilibration of the anesthetic with the direct signaling pathway should be complete within milliseconds to seconds, the prolonged time course of these effects appears to indicate that additional intracellular pathways (such as PKC-mediated phosphorylation) are involved. Exposure of the membrane patches to the nonhydrolyzable GTP analog GTPγS, thereby irreversibly activating G_K, prolonged the current. When single-channel measurements were performed, halothane was found to enhance the frequency of channel opening without significantly affecting the single-channel conductance. Isoflurane was without effect. Therefore, halothane affects the signaling pathway downstream of the muscarinic receptor in this model. Whether this action is on the G protein or on the channel itself cannot be conclusively determined from these studies. When the time considerations are taken into account, it appears likely that the initial action of halothane on the system is inhibiting, whereas that of isoflurane is less so. After prolonged exposure, halothane is found to enhance signaling, whereas isoflurane inhibits it. These results are probably due to various effects of the anesthetics on intracellular systems that modify the signaling properties of the muscarinic pathway. In addition, halothane, but not isoflurane, has a direct activating effect downstream of the muscarinic receptor. The latter is consistent with findings of halothane effect on Ca⁺⁺ channels mentioned earlier.

Genetic Cardiovascular Medicine

Over the past few decades, considerable progress has been made in the identification and understanding of the genetic basis of cardiovascular disease. More than 40 cardiovascular disorders are now known to be caused directly by gene defects. These disorders, spanning all aspects of cardiovascular disease and affecting all parts of the heart structure, can be divided into two groups (Box 7-6). Monogenic disorders are (usually rare) Mendelian disorders for which single gene changes are implicated in the disease process and which usually exhibit characteristic inheritance patterns. Examples include familial hypercholesterolemia, hypertrophic cardiomyopathies (HCM), DCMs, and the LQTSs.⁸⁸ More commonly, however, multiple genes influence the disease process by enhancing disease susceptibility or by augmenting the impact of environmental risk factors. The genetic component in those multigenic disorders comprises a collection of gene variants such as single nucleotide mutations, referred to as single nucleotide polymorphisms (SNPs). Each individual SNP may have a modest effect on the quantity or function of a translated protein product. However, when individual SNPs aggregate and interact with environmental risk factors, they may have a major impact on disease biology. Common diseases postulated to follow this paradigm include coronary artery disease (CAD), hypertension, and atherosclerosis.⁸⁹

This section discusses the current status of genetic diagnosis of monogenic and complex cardiovascular disorders, as well as the main techniques used to perform such diagnostic testing (Box 7-7).

Monogenic Cardiovascular Disorders

Great progress has been made in the identification and characterization of the disease genes specific to Mendelian cardiovascular diseases,^88,90 but several factors complicate these investigative efforts. Locus heterogeneity (many genes causing the same disease) is one of those.⁹¹ The channelopathies, encompassing the LQTS and related genetic arrhythmogenic disorders, are accounted for by more than 10 genes. Whereas HCM genes MYH7, MYBPC3, and TNNT2 (encoding the β-myosin heavy chain, cardiac myosin-binding protein C, and cardiac troponin T genes, respectively) account for 70% to 80% of HCM as diagnosed by molecular genetic approaches,⁹² an additional 19 genes have been implicated to cause the HCM phenotype.⁹³ In addition, more than 20 genes have been implicated in DCM.⁹¹ A second, related factor complicating genetic diagnosis is that not all disease loci have been identified, which diminishes the sensitivity of molecular testing. Whereas it has been estimated that a mutation causing LQTS can be identified in up to three quarters of index patients, this is not true for other cardiovascular single-gene disorders. For example, only 30% to 60% of the genetic causes of HCM⁹⁴ and only 20% to 30% of genetic causes for DCM have been identified.^90,91 Further complicating diagnostic strategies is allelic heterogeneity (many mutations within a single gene). As an example, MYH7 consists of 40 exons encoding the β-myosin heavy chain protein, and 194 mutations in this structure have been reported to be associated with HCM. For diagnostic purposes, this necessitates sequencing the entire coding sequence and intron/exon boundaries of each gene, making sequencing a time- and effort-intensive undertaking.

Methodologies for Identifying Mutations: Sequencing and Microarrays

Techniques for identifying gene sequences have evolved from the classic Southern blotting procedure into highly automated systems that can rapidly screen hundreds of thousands of gene sequences. Southern blotting, invented in the mid-1970s, is based on the transfer to nitrocellulose of DNA molecules separated on gels and their subsequent identification with DNA probes. It allows detection of small mutations, as well as large deletions, duplications, and gene rearrangements. The principle was expanded and automated in the development of DNA microarrays (Figure 7-11), which consist of series of thousands of microscopic spots of DNA oligonucleotides, each containing tiny amounts of a specific DNA sequence. The arrays are printed on a platform, usually a glass slide. A DNA probe, encoding a specific sequence, is then used to bind to and identify homologous nucleic acids, as detected by fluorophore-labeled targets whose signal is proportional to the relative abundance of nucleic acid sequences in the target. This makes it possible to examine the relative abundance of thousands of genes at any given time. Newer chip-based sequencing methods, designed to improve speed and efficiency, enable simultaneous sampling of 25,000 genes (Figure 7-12).

Figure 7-11 Schematic representation of a microarray experiment.

RNA is extracted from the tissue(s) of interest, labeled, and then hybridized (competitively, two color arrays; noncompetitively, one color arrays) to the DNA microarray where it binds complementary sequences. Bound sequences are identified by their position on the array. Signal intensities are normalized allowing interarray comparisons. Normalized datasets are filtered and subjected to computational analysis. Finally, differentially regulated transcripts should be prospectively validated and/or confirmed using independent methods.

(From Cook SA, Rosenzweig A: DNA microarrays: Implications for cardiovascular medicine. Circ Res 91: 559–564, 2002.)

Figure 7-12 Chip method of sequencing combines chromatin immunoprecipitation (ChIP) with massively parallel DNA sequencing, allowing improved speed and efficiency.

(From Central European Journal of Medicine, 1895–2058 (print) 1644–3640 (online), 2009, vol 4, No.1, pp 1–10.)

Clinical Applications

The ability of these techniques to identify diseases before they become clinically manifest allows preventive treatment as summarized in Table 7-2. For example, implantable cardioverters-defibrillators (ICDs) can prevent sudden cardiac death, improve quality of life, and prolong the time to cardiac transplantation (or avoid it altogether) in patients with genetic cardiomyopathies and arrhythmias^95–97 (Figure 7-13). Medical therapy may ameliorate the progression of genetic DCM. Prospective identification of those patients yet asymptomatic, but at greater risk for development of the disease, enables close surveillance and early intervention. Knowledge of the LQTS genotype may be used to tailor the treatment plan (Figure 7-14). For instance, LQT3 patients benefit less from β-blockers, so a lower threshold for ICD implantation is advised in LQT3 patients.⁹⁸ In addition, the triggers for malignant arrhythmias have been found to differ based on gene involved, and, for example, patients with LQT1 can be asked to avoid vigorous activities such as exercise and competitive sports.⁹⁹

TABLE 7-2 Clinical Applicability of Genetic Testing in Monogenic Cardiac Diseases

Figure 7-13 Proposed sequence of genetic testing for patients with hypertrophic cardiomyopathy.

HCM, hypertrophic cardiomyopathy; LV, left ventricular; SCD, sudden cardiac death. (From Keren A, Syrris P, McKenna WJ: Hypertrophic cardiomyopathy: The genetic determinants of clinical disease expression. Nat Clin Pract Cardiovasc Med 5:158–168, 2008.)

Figure 7-14 Genotype-phenotype relations seen in LQT1, LQT2, and LQT3.

The linear topologies for the three principal cardiac channels that account for two thirds of long QT syndromes (LQTSs) are superimposed on the cardiac action potential of a ventricular myocyte. The inset next to each channel summarizes some of the signature phenotypes associated with the three most common LQTS-causing genotypes. AV, atrioventricular.

(From Ackerman, Michael J: Genetic testing for risk stratification in hypertrophic cardiomyopathy and long QT syndrome: Fact or fiction? Curr Opin Cardiol 20:175–181, 2005.)

Multigenic Cardiovascular Disorders

Identifying the gene variants associated with the development and progression of multigenic diseases allows them to be used to assess treatment response or serve as targets for novel treatment strategies. The obvious challenge is to identify the genes and gene variants that collectively contribute to the disease. Several genomic technologies allow researchers to study the genetic components of multigenic disorders like CAD.

Perioperative Genomics in Cardiac Surgery

Despite advances in surgical, anesthetic, and cardioprotective strategies, the incidence of perioperative adverse events in cardiac surgery continues to be significant and is associated with reduced short- and long-term survival.¹¹³ Because all surgical patients are exposed to perturbations that potentially activate inflammation, coagulation, and other stress-related pathways, but only a subset experience adverse perioperative events (even after controlling for coexistent disease), a genetic component is likely to be involved^114–116 (Figure 7-15). Perioperative genomics is a new field that uses functional genomic approaches to discover underlying biologic mechanisms that explain why similar patients have dramatically different outcomes after surgery.¹¹⁷

Figure 7-15 Diagram illustrating how genetic factors in combination with perioperative insults may play a role in adverse postoperative outcomes. CPB, cardiopulmonary bypass; OR, operating room.

Genetic variants have been found for adverse events such as myocardial ischemia,¹¹⁸ postoperative arrhythmias,¹¹⁹ vein graft restenosis,¹¹⁸ renal compromise,¹²⁰ neurocognitive dysfunction,¹²¹ stroke,¹²² and death,¹¹⁸ as well as more systemic outcomes such as bleeding,¹²³ thrombosis,¹²⁴ inflammatory responses, and severe sepsis^125–128 (Figure 7-16). The Duke Perioperative Genomics investigative team initiated a prospective study (the Perioperative Genomics And Safety Study, US [PEGASUS]) in 2001. In recent years, broader multi-institutional perioperative genomics groups have been formed in the United States (PeriGReN) and internationally (iPEGASUS). The goal of the PEGASUS group is to use genetic variability to determine which individuals are at risk for adverse events after surgery. As an example of the power of perioperative genetic/genomic approaches, the PEGASUS group recently examined the perioperative myocardial injury/infarction (PMI) end point by genotyping 48 polymorphisms from 23 candidate genes in a prospective cohort of 434 patients undergoing elective cardiac surgery with cardiopulmonary bypass.¹²⁹ After adjusting for multiple comparisons and clinical risk factors, three polymorphisms were found as independent predictors of PMI. These represent SNPs in genes encoding IL-6 and 2 adhesion molecules, intercellular adhesion molecule 1 (ICAM-1) and E-selectin. In contrast, one mutation (SELE98G>T) decreased the risk for PMI in the study. The investigators concluded that functional genetic variants in cytokine and leukocyte-endothelial interaction pathways are independently associated with severity of myonecrosis after cardiac surgery.

Figure 7-16 Genetic variation linked to various postoperative outcomes. AF, atrial fibrillation; MACE, major adverse cardiac event.

(Reproduced with permission from http://anesthesia.mc.duke.edu/modules/anes_genes/index.php?id=5.)

Do these genetic/genomic studies result in practical information capable of facilitating therapeutic interventions designed to improve outcome? A study of the effects of a 5-lipoxygenase-activating protein (FLAP) inhibitor on biomarkers associated with increased risk for myocardial infarction demonstrated that, by defining at-risk patients using two genes in the leukotriene pathway, it can be predicted who will respond to targeted drug therapy.¹³⁰ Use of the FLAP inhibitor DG-031 in patients with these variants leads to significant and dose-dependent suppression of biomarkers associated with increased risk for myocardial infarction events. Although this study did not take place during surgery, similar principles would be expected to be operational in the perioperative period. These findings may soon translate into prospective risk assessment incorporating genomic profiling of markers important in inflammatory, thrombotic, vascular, and neurologic responses to perioperative stress, with implications ranging from individualized additional preoperative testing and physiologic optimization, to perioperative decision making, options for monitoring approaches, and critical care resource utilization.

Gene Therapy

Whereas molecular diagnosis is here, molecular therapy appears just tantalizingly out of reach—and has seemed that way for much of the past decade. The concepts are clear. The application, however, has been fraught with much more difficulty than at first appreciated.

Gene therapy attempts to modify expression of genetic material. Although a myriad of approaches have been suggested, those with the most promise are replacement of lacking or defective genes and selective inhibition of gene expression by antisense treatment. Another approach, further removed from clinical practice at this time, is the targeting of drug delivery to specific tissues using molecular techniques.

Replacement of genes is by now a routine technique in laboratory cell cultures, and can be performed with reasonable ease in animal models such as transgenic animals. Application in the clinical setting has been difficult, however. Obviously, it is most times not necessary to replace a gene in every cell of the body. In cystic fibrosis, for example, a disease in which gene therapy is probably closest to succeeding, the main target tissue will be the pulmonary system. The main issue then becomes the delivery of the functional gene to the target tissue in such a manner that it will be consistently expressed at adequate levels. A variety of experimental techniques exists. Cells can be removed from the patient’s body, changed in culture, and then be returned to the patient. Alternatively, the gene can be incorporated in an otherwise innocuous virus that then is used to infect the patient’s tissues. Both approaches show promise, but the technical difficulties indicate that this is still a considerable time away from routine clinical application. Nonetheless, several clinical trials are in progress.^131,132

Inhibition of gene expression by antisense treatment is similarly full of promise.^133–135 The technique is based on the fact that mRNA can no longer be translated efficiently into protein when bound to a complementary strand of nucleic acid, a so-called antisense strand. Thus, by introducing specific antisense DNA into cells, expression of a selected gene product can be inhibited. The problem here, again, is efficient targeting of the tissue of interest. Remarkably, cells are able to take up antisense material from the extracellular environment without degrading it, but the material still has to be brought in contact with the tissue of interest. A second technical problem involves the stability of the antisense DNA. Because most diseases for which this technique seems suitable require constant inhibition of gene expression over long periods, the antisense construct has to be highly stable, and this degree of stability has not been consistently achieved yet.

Thus, molecular therapy seems feasible but is not quite ready for the clinic yet. Several times clinicians have believed that a breakthrough would be achieved rapidly; each time they have been disappointed. It would, therefore, be dangerous to predict that these techniques will find practical clinical application in the next few years. Nonetheless, the concepts appear sound, and have been proved in animal models. It appears to be only a matter of time before patients will be treated with these methods.

Acknowledgments

The authors thank J. Paul Mounsey, MD, PhD, for his contributions to earlier editions of this chapter, and Logan Reeves, MD, for his assistance in preparing this edition.