|Home | About | Journals | Submit | Contact Us | Français|
Rapidly activating and inactivating cardiac transient outward K+ currents, Ito, are expressed in most mammalian cardiomyocytes, and contribute importantly to the early phase of action potential repolarization and to plateau potentials. The rapidly recovering (Ito,f) and slowly recovering (Ito,s) components are differentially expressed in the myocardium, contributing to regional heterogeneities in action potential waveforms. Consistent with the marked differences in biophysical properties, distinct pore-forming (α) subunits underlie the two Ito components: Kv4.3/Kv4.2 subunits encode Ito,f, whereas Kv1.4 encodes Ito,s, channels. It has also become increasingly clear that cardiac Ito channels function as components of macromolecular protein complexes, comprising (four) Kv α subunits and a variety of accessory subunits and regulatory proteins that influence channel expression, biophysical properties and interactions with the actin cytoskeleton, and contribute to the generation of normal cardiac rhythms. Derangements in the expression or the regulation of Ito channels in inherited or acquired cardiac diseases would be expected to increase the risk of potentially life-threatening cardiac arrhythmias. Indeed, a recently identified Brugada syndrome mutation in KCNE3 (MiRP2) has been suggested to result in increased Ito,f densities. Continued focus in this area seems certain to provide new and fundamentally important insights into the molecular determinants of functional Ito channels and into the molecular mechanisms involved in the dynamic regulation of Ito channel functioning in the normal and diseased myocardium.
The rapidly activating and inactivating “transient” outward current has been capturing the interest of cardiac electrophysiologists since it was first described as an early outward current in sheep Purkinje fibers [1,2], in which the activity of this current is evident in the prominent early phase of action potential repolarization. Although initially attributed to Cl− [1,2], subsequent work revealed that the major component of the transient outward current is blocked by 4-aminopyridine (4-AP) and carried by K+ [3,4]. The Cl− component, in contrast, is regulated by Ca2+ and is insensitive to 4-AP . The focus here is on the K+-dependent transient outward current, typically referred to as Ito or Ito1. Following the early finding in sheep Purkinje fibers, Ito has been identified in atrial and ventricular myocytes in most mammals [6–17], including humans [18–21], as well as in sinoatrial [22,23] and atrioventricular nodal [24,25] cells. Biophysical analyses further revealed the presence of two Ito components with distinct recovery kinetics: the fast (Ito,f) and slow (Ito,s) components [26,27]. In many species, Ito,f and Ito,s are differentially expressed and contribute to regional heterogeneities in action potential waveforms. In canine and human ventricles, for example, Ito,f is more prominent in epicardium and midmyocardium than in endocardium and underlies the “spike and dome” morphology of epicardial action potentials [28,29]. Markedly reduced Ito,f densities and pronounced alterations in action potential repolarization are observed in failing human hearts, as well as animal models [21,30], and in chronic atrial fibrillation [31–36]. Interestingly, dysregulation of Ito has also been postulated to play a pivotal role in the Brugada syndrome .
Considerable effort has been focused on defining the molecular correlates of Ito channels and on elucidating the mechanisms that regulate the functional expression of these channels. There is now consensus that pore-forming (α) subunits of the Kv4 subfamily, Kv4.2 (KCND2) and Kv4.3 (KCND3), encode Ito,f and that Kv1.4 (KCNA4) encodes Ito,s channels . A large number of Kv channel accessory (β) subunits have also been proposed to function in the generation of myocardial Ito channels. A critical role for the Kv channel interacting protein, KChIP2, was revealed with the demonstration that the targeted disruption of Kcnip2 eliminates Ito,f in mouse ventricular myocytes . In large mammals, KChIP2 appears to be the primary determinant of the transmural gradient of Ito,f [40–42]. Functional roles for Kvβ  and diaminopeptidyl transferase-like protein 6 (DPP6)  subunits in the generation of Kv4-encoded Ito,f channels have also been proposed. In addition, quite recently, members of the MinK related peptide (MiRP) subfamily, MiRP1 and MiRP2, have been suggested to function in the regulation of Ito,f [45–47] and, interestingly, a mutation in the gene (KCNE3) encoding MiRP2 has been identified in Brugada syndrome . Accumulating evidence also suggests that Kv channel pore-forming α and accessory β subunits interact (directly or indirectly) with components of the actin cytoskeleton and with a variety of regulatory and signaling molecules. In addition to summarizing the fundamental molecular physiology of cardiac Ito, discussed in previous comprehensive reviews [38,48,49], this review integrates recent findings suggesting novel physiological roles of accessory subunits and modulatory proteins, highlighting the dynamic and complex mechanisms that regulate the functioning of cardiac Ito channels.
Myocardial action potentials reflect the coordinated functioning of inward (Na+ and Ca2+) and outward (K+ and Cl−) current carrying channels that open and close in a voltage- and time-dependent manner. Of these, voltage-gated K+ (Kv) channels are the most diverse, and multiple Kv channels with distinct gating properties determine the heights and durations of myocardial action potentials . Cardiac Ito channels activate and inactivate rapidly on membrane depolarization, and in most mammalian cardiomyocytes, including rat [10,50], mouse , ferret , feline , rabbit [14,15], canine [17,28] and human [19–21] ventricular myocytes, as well as in atrial cells from rat , mouse , rabbit , canine  and human [18,19], Ito is a prominent repolarizing current. In large mammals, including human and canine, Ito partially repolarizes the membrane, shaping rapid (phase 1) repolarization  and setting the initial plateau (phase 2) potential. The activity of Ito channels influences the activation of voltage-gated Ca2+ channels and the balance of inward and outward currents during the plateau, thereby mediating the duration and the amplitude of phase 2. In rodents, high Ito densities dominate all phases of repolarization and account for the remarkably abbreviated action potentials, which lack a clear plateau phase, in these species . There are, however, exceptions. Macroscopic Ito is absent or very small in guinea pig [54,55] and porcine [56–58] atrial and ventricular myocytes. Although the removal of extracellular Ca2+ reportedly revealed the presence of fast activating and inactivating “Ito-like” currents in these cells [59,60], later pharmacological studies suggested that the measured currents were much more likely to be due to monovalent cation flux through L-type Ca2+ channels [61,62] or to be mediated by inward rectifier K+ channels [63,64].
Action potential waveforms are variable in different regions of the heart [38,65], reflecting, at least in part, differences in the densities and properties of Ito. Markedly higher densities of Ito, together with the expression of the ultrarapid delayed rectifier, for example, accelerate the early phase of repolarization, leading to lower plateau potentials and shorter action potentials in atrial cells [14,19,66,67]. In addition, in human (as well as canine and feline) ventricles, Ito densities are much higher in epicardium and midmyocardium, than in endocardium [29,68,69]. The high density of Ito is reflected in the prominent phase 1 notch and the “spike and dome” morphology of epicardial and midmyocardial action potentials, that are largely absent in endocardium [13,28,69]. Ito densities are also reportedly higher in right than in left (midmyocardial and epicardial) myocytes, consistent with the more pronounced spike and dome morphology of right, compared with left, ventricular action potentials [70,71].
Regional heterogeneities in cardiac action potential waveforms also reflect the presence of two functionally distinct components of Ito. Although both components activate and inactivate quite rapidly, the rates of recovery from inactivation are distinct (Figure 1). The fast component, Ito,fast (Ito,f), for example, recovers very rapidly with time constants in the range of 60 to 100 ms [48,49]. In contrast, the slow component, Ito,slow (Ito,s), recovers slowly with time constants on the order of seconds [26,27]. Ito,f and Ito,s can also be distinguished based on differential sensitivities to several spider toxins, including the Heteropoda toxins [26,27,72,73] and phrixotoxins , that selectively block Ito,f, and have no measurable effects on other Kv channels, including Ito,s channels.
In human and ferret left ventricles, Ito in epicardial myocytes recovers rapidly, whereas a slowly recovering component of Ito dominates in endocardial myocytes, suggesting that Ito,f densities are higher in epicardial cells, whereas Ito,s predominates in endocardial myocytes [26,29,68]. In rat ventricles, although Ito,f and Ito,s are coexpressed, Ito,f densities are higher in epicardial, than in endocardial and septum, myocytes, and Ito,s densities are similar [75,76]. In mouse ventricles, Ito,f densities are significantly higher in right, than in left, ventricular myocytes and are low in the interventricular septum, whereas Ito,s is only detected in the septum [27,77,78] (Figure 1). In rabbit atrial and ventricular myocytes, Ito recovers slowly (time constants ≈600 ms to 8 s) [79,80], suggesting that Ito,s is the primary transient Kv current . In human, canine and mouse atrial myocytes, in contrast, only Ito,f appears to be expressed [7,80,82].
Functional Kv channels reflect the assembly of four (4) pore-forming Kv (α) subunits , each of which has six transmembrane segments (S1 to S6), including the S4 voltage sensor , a single pore (H) region between the S5 and S6, and cytoplasmic N- and C-termini (Figure 2). The H (pore) loop contains the G(Y/F)G signature motif and functions as the K+ selectivity filter . In the Kv1 to Kv4 subfamilies, N-terminal tetramerization (T1) domains immediately preceding S1 (Figure 2) play critical roles in Kv α subunit assembly [86,87]. In heterologous expression systems, Kv α subunits can assemble as homomultimers or as heteromultimers, containing different α subunits from the same subfamily [88–90].
Considerable evidence suggests that Kv4 α subunits underlie Ito,f channels. Expression of Kv4.3 predominates in human and canine ventricles [91–94], whereas both Kv4.2 and Kv4.3 are expressed in rodent and ferret ventricles [26,75,95,96]. In rodents, Kv4.2 expression is correlated with regional heterogeneities in Ito,f [75,95]. Heterologous expression of Kv4 α subunits gives rise to rapidly activating, inactivating and recovering (time constants ≈50 to 380 ms) Kv currents [75,91,97–99], and the currents are sensitive to 4-AP [97,98], as well as to the heteropodatoxins  and phirotoxins . Genetic manipulations in vitro and in vivo have provided direct evidence that Kv4 α subunits underlie Ito,f channels. Expression of mutant Kv4 α subunits that function as dominant-negatives, for example, attenuated Ito,f in rat ventricular myocytes in vitro , and eliminated Ito,f in mouse ventricular (Figure 1)  and atrial  myocytes in vivo. Selective gene silencing using antisense oligonucleotides (AsODNs) targeted against Kv4.2 and Kv4.3 reduced Ito,f in cultured rodent ventricular myocytes [102,103]. In rat atrial myocytes, AsODNs targeted against Kv4.2, but not Kv4.3, attenuated Ito,f , whereas in human atrial myocytes, Ito,f was significantly attenuated by Kv4.3, but not by Kv4.2, AsODNs . In addition, targeted gene deletion of Kcnd2 (Kv4.2) in mice eliminates ventricular Ito,f, further revealing the critical role of Kv4.2 in the generation of Ito,f channels in rodents .
Consistent with differences in biophysical properties, considerable evidence suggests that myocardial Ito,s channels are molecularly distinct from Ito channels and are encoded by the Kv1 α subunit, Kv1.4. Heterologous expression of Kv1.4 yields slowly recovering (time constants of seconds), 4-AP sensitive transient Kv currents [75,105–107] that closely resemble Ito,s [26,27,80,82]. In addition, expression of Kv1.4 has been demonstrated at the transcript level in rat [75,105,108], ferret , canine  and human  hearts, and at the protein level in mouse, rat and ferret ventricles [26,75,110] and rabbit atrium . In ferret left ventricles, the Kv1.4 protein is predominantly expressed in endocardial myocytes, but is essentially absent in epicardium, consistent with the observed gradient in Ito,s . A close relationship between Kv1.4 protein expression and Ito,s density has also been described in rat  and mouse  ventricles. Definitive evidence for a role of Kv1.4 was provided with the demonstration that Ito,s is eliminated in intraventricular septum myocytes from (Kv1.4−/−) animals in which the Kcna4 locus was disrupted by homologous recombination . In addition, the upregulated Ito,s in left ventricular apex cells expressing the mutant Kv4.2 subunit (Kv4.2DN), that functions as a dominant negative, is eliminated in the Kv1.4−/− background (Figure 1) .
Although heterologous expression of Kv4.3/4.3 or Kv1.4 α subunits generates Kv currents with biophysical properties similar to those of myocardial Ito,f and Ito,s, there are clear differences. Expressed Kv4 channels, for example, inactivate slower and recover from inactivation much more slowly than native Ito,f channels. These observations have long been interpreted as suggesting that additional subunits are required for the generation of cardiac Ito channels (Figure 3). Indeed, a number of molecularly diverse accessory subunits have been demonstrated to modulate the biophysical properties and the cell surface expression of heterologously expressed Kv4 channels [112,113]. Of these accessory subunits, only the cytoplasmic KChIP2 protein has been shown directly to be an essential component of myocardial Ito,f channels . The KChIPs, first identified in brain using the Kv4 N-terminus as bait in a yeast two-hybrid screen, belong to the neuronal calcium sensor (NCS) family of Ca2+ binding proteins with four EF-hand (Ca2+ binding) motifs [114,115]. Co-expression with KChIP subunits dramatically increases the cell surface expression of Kv4 channels, slows current inactivation and markedly accelerates recovery [114,116]. In addition, KChIP2 co-immunoprecipitates with Kv 4.2 (and Kv4.3) α subunits from adult mouse ventricles , and the targeted deletion of the KChIP2 locus (Kcnip2) eliminates (mouse) ventricular Ito,f . Interestingly, KChIP2 protein expression is dramatically reduced in Kv4.2−/− ventricles , suggesting that expression of the Kv4 and KChIP2 proteins is reciprocally regulated. In human and canine, there is a marked transmural gradient in KChIP2 expression across the ventricular wall [40–42] that parallels the gradient in Ito,f density. A modest epi/endocardial gradient in KChIP2 transcript expression has also been reported in mouse ventricles .
The molecular mechanisms underlying the enhancement of cell surface expression and the modulation of the gating properties of Kv4 channels by KChIP2, however, have not been elucidated. The proximal N-terminal residues preceding the Kv4.3 T1 domain (Figure 2) are essential for KChIPs binding [114,118]. It has been suggested that KChIP binding to the N-termini of Kv4 α subunits masks a retention signal, and releases channels from the endoplasmic reticulum (ER), allowing forward trafficking to the membrane [116,118]. To date, however, no functional ER retention signals in Kv4 N-termini have been identified. It has also been reported that KChIP1 stabilizes Kv4 protein expression  and drives Kv4 subunit tetramerization . Other studies, however, suggest critical roles for the Kv4 C-terminus  and the T1 domain  in mediating the interaction(s) with KChIPs. Recent structural analyses of KChIP1 with Kv4.3 N-termini revealed that each KChIP1 laterally binds to two neighboring Kv4.3 N-terminal domains, the proximal N-terminus in one Kv4.3 subunit and the T1 domain of the adjacent Kv4.3 subunit [122,123]. Another EF-hand motif-containing Ca2+ binding protein, neuronal calcium sensor-1 (NCS-1), also co-immunoprecipitates with Kv4 α subunits from adult mouse ventricles . Like the KChIPs, co-expression with NCS-1 increases Kv4 current densities and affects the rates of Kv4 current inactivation [124,125].
The diaminopeptidyl transferase-like protein 6 (DPP6 or DPPX) (Figure 3), first identified as a protein co-immunoprecipitating with Kv4.2 from rat brain , has also been suggested to be an integral component of brain  and heart  Kv4 channels. Reportedly lacking enzymatic activity , DPP6 co-expression increases the cell surface expression of Kv4 α subunits , shifts the voltage dependences of activation and inactivation of heterologously-expressed Kv4 currents to more negative potentials, and accelerates the rates of current activation, inactivation and recovery [44,126,128]. Interestingly, heterologous co-expression of DPP6 with Kv4.3 and KChIP2, produces Kv currents that closely resemble native cardiac Ito,f . Gene silencing of DPP6 in neurons, however, produces only very subtle effects on the voltage- and time-dependent properties of Kv4-encoded currents . Another member of the DPP-like subfamily of proteins, DPP10, has also been demonstrated to associate with Kv4.2 and KChIP3 in rat brain  and to have regulatory effects similar to DPP6 on heterologously expressed Kv4 (with and without KChIP) channels [130–133], and DPP10 has also been shown to be expressed in normal and failing human ventricles .
The MinK and MinK-related peptide (MiRP) subfamily of transmembrane accessory subunits, modulate the properties of multiple types of Kv channels, including Kv4 channels . Co-expression with MiRP1, for example, affects the kinetics and the voltage-dependent properties of Kv4 and Kv4-KChIP2 channels [45,136–138], and induces an “overshoot” in peak current amplitude during current recovery [45,137,138], a phenomenon evident in human epicardial Ito,f . MiRP1 is expressed at the protein level in human, canine, rat and mouse ventricles [46,139,140] and co-immunoprecipitates with Kv4.2 from adult mouse hearts . Although MiRP1 does not measurably affect the densities of expressed Kv4.2-KChIP2 currents , targeted deletion of the gene (Kcne2) encoding MiRP1 resulted in reduced (~25%) ventricular Ito,f densities with no measurable changes in total or surface Kv4.2 expression . Another member of this family, MiRP2 (Figure 3), co-immunoprecipitates with Kv4.3 from human atria . Interestingly, a missense mutation (R99H) KCNE3 (which encode MiRP2) was recently identified in an individual with Brugada syndrome and in phenotype-positive (but not phenotype-negative) family members . Heterologous expression experiments revealed that MiRP2 decreases Kv4.3- (with and without KChIP) encoded current densities [47,139,141]. In addition, co-expression of the Brugada syndrome MiRP2 mutant reversed the inhibitory effects of wild type MiRP2, leading to markedly increased Kv4.3 currents . Taken together, these intriguing observations suggest that MiRP2 is required for the physiological functioning of human Ito,f channels and that gain of function mutations in MiRP2 predispose to Brugada syndrome through augmentation of Ito,f .
Three homologous Kv β accessory subunits, Kvβ1, Kvβ2 and Kvβ3, as well as alternatively spliced variants of each, have been identified [142,143]. The Kvβ subunits are low molecular weight, cytosolic proteins that co-localize with Kv α subunits in brain [142–151]. In heterologous cells, co-expression of Kvβ subunits affects the properties and densities [148,149] of Kv α subunit-encoded channels. Although initially isolated using a Kv1 channel toxin  and suggested to be specific for Kv1 channels [145,148,149], both Kvβ1 and Kvβ2 co-localize with Kv4 α subunits in (rat) brain [146,147] and co-expression of Kvβ subunits alters the properties of heterologously expressed Kv4.2 currents . In addition, Kvβ2 reportedly associates with Kv4.3 in (rat) brain  and heterologous co-expression with Kvβ1, Kvβ2 or Kvβ3 increases Kv4.3-encoded Kv current densities [136,153,154].
In spite of the postulated link between Kvβ subunits and the generation of Kv1-encoded channels [144–151], studies on Kvβ1.1 deficient, Kvβ2 deficient and combined Kvβ1.1/Kvβ2 deficient mice did not reveal an essential role for these subunits in regulating the expression of Kv1.1 or Kv1.2 subunits/channels in the cerebellum or in peripheral nerve [155,156]. It is possible that Kv1 α subunit expression is altered in other parts of the central nervous system and/or in other tissues (including the heart) and/or that Kvβ3 expression is upregulated in these animals to compensate for the loss of Kvβ1.1 and/or Kvβ2; these possibilities were not explored [155,156]. These mouse lines do, however, display neurological phenotypes, including seizures [155,156], suggesting that loss of Kvβ1.1 and/or Kvβ2 affects neuronal excitability. Consistent with this suggestion, it has been reported that K+ currents and action potential waveforms are altered in hippocampal CA1 pyramidal neurons in (an independently generated line of) Kvβ1.1-deficient mice [157,158]. The observed reductions in rapidly inactivating K+ currents were interpreted as reflecting the loss of Kvβ1.1-mediated inactivation of Kv1-encoded A-type (IA) K+ currents [157,158]. Given the documented role for Kv4 α subunits in the generation of IA, however, the observed effects may suggest a role for Kvβ subunits in the generation of Kv4-encoded IA channels. Consistent with this hypothesis, it has been reported that deletion of Kvβ1 markedly attenuated Kv4-encoded mouse ventricular Ito,f, whereas Kv1-encoded currents were unaffected . Interestingly, structural data suggests that Kvβ subunits can interact with Kv4 subunits in the presence of KChIP (Figure 3), resulting in the formation of multimeric Kv4-channel complexes.
Biochemical experiments on extracts from neonatal rat ventricles suggest that the voltage-gated Na+ (Nav) channel accessory subunit, Navβ1, associates with Kv4.2/Kv4.3 α subunits . It has also been reported that co-expression of Navβ1 increased Kv4.3-encoded current densities in heterologous cells , and gene silencing of Navβ1 dramatically reduced Ito,f densities and decreased KChIP2 transcript expression in neonatal rat ventricular myocytes . These data suggest the intriguing possibility that cardiac Ito,f and Nav channels are functionally coupled through accessory (Navβ) subunits. The K+ channel-associated protein, KChAP, also increases the cell surface expression, without influencing the time- and/or voltage-dependent properties, of Kv4-encoded channels .
Similar to other ion channels, accumulating evidence suggests that Kv channel α and accessory (β) subunit assemblies associate with the actin cytoskeleton. These interactions likely are important in determining channel stability, trafficking, localization and biophysical properties, but also for providing microdomains for regulatory molecules. Consistent with this suggestion, the properties of rat ventricular Ito,f channels are regulated by the actin cytoskeleton, particularly under pathological conditions. Pharmacological disruption of actin microfilaments by cytochalasin D, for example, reportedly decreased Ito,f densities in hypertrophied myocytes, whereas stabilization of actin microfilaments by phalloidin increased the currents . Neither treatment had effects on Ito,f densities in non-hypertrophied cells . Heterologously expressed Kv4.2 currents, however, are augmented on exposure to cytochalasin D . These seemingly disparate observations might suggest important roles for Kv4 channel accessory/regulatory proteins in mediating interactions between Kv4 channels and the actin cytoskeleton and that these accessory/regulatory proteins are differentially expressed in different cell types . It has also been shown that the insulin-induced recovery of Ito,f densities (to normal) in diabetic rat ventricular myocytes requires an intact cytoskeleton, actin microfilaments and microtubules, presumably to traffic newly synthesized channels to the membrane .
Several actin-binding proteins and other scaffolding proteins have been suggested to mediate the interactions between Kv channels and the actin cytoskeleton. Filamin, for example, interacts directly with the Kv4.2 C-terminus through a C-terminal filamin binding domain (that is also present in Kv4.3 (Figure 2)) and increases the cell surface expression of Kv4.2 channels . The actin binding protein, α-actinin-2, in contrast, binds to the Kv1.4 N-terminus . A number of scaffolding proteins, including the PDZ-domain-containing membrane-associated guanylate kinases (MAGUKs), have been implicated in the clustering and localization of ion channels (and neurotransmitter receptors) in brain [166–168], and more recently, in the heart . Of these, the 95 kDa post-synaptic density protein (PSD-95) and the synapse-associated protein 97 (SAP97) have been shown to interact directly with Kv α subunits. PSD-95, for example, binds directly to the PDZ-binding domains in the C-termini of Kv1α subunits and co-immunoprecipitates with Kv1.4 from brain . Interactions with PSD-95 increase the clustering and cell surface expression of Kv1 channels in heterologous cells [167,170,171]. Co-expression with PSD-95 localizes Kv1.4 channels to specific membrane microdomains known as “lipid rafts” [172–174] and interaction with PSD-95 appears to stabilize Kv1.4-encoded channels in the membrane , presumably by suppressing internalization . PSD-95 also reportedly binds to C-terminal PDZ-binding motifs in Kv4 α subunits , although the functional impact of this binding on Kv4 channel surface expression and clustering is not clear [167,177]. Although not detected in rat heart [178,179], PSD-95 is reportedly expressed in human atria .
The PDZ-binding motifs in the C-termini of Kv1α subunits also interact directly with another MAGUK protein, SAP97  and, in contrast with PSD-95, this interaction results in retention of Kv1 channels in the endoplasmic reticulum [170,171,181]. It has also been reported that SAP97 regulates the functional expression of Kv4 channels in neurons  and in the heart . In rat ventricles and human atria, for example, Kv4.2/Kv4.3 α subunits interact with SAP97 via C-terminal PDZ-binding motifs, and the Kv4-SAP97 proteins are co-localized at the plasma membrane in rat atrial myocytes . Heterologous co-expression with SAP97 enhanced Kv4.3 channel clustering at the membrane surface, increased current densities and slowed inactivation in heterologous cells, and similar effects were observed on rat atrial Ito,f by the overexpression of SAP97 . Conversely, SAP97 RNA interference significantly decreased Ito,f densities in rat atrial myocytes, consistent with modulatory effects of SAP97 on endogenous Ito,f channels . In addition, recent work suggests that SAP97 functions as a molecular adaptor of the Ca2+/calmodulin-dependent protein kinase II (CaMKII), mediating phosphorylation-dependent regulation of cardiac (and neuronal) Kv4 channels [168,169]. Consistent with this hypothesis, it was recently demonstrated that SAP97 and CaMKII co-immunoprecipitate and that the expression level of SAP97 affects the interaction between CaMKII and Kv4.3 in rat heart . Additional studies in heterologous cells revealed that reduced expression of SAP97 led to loss of CaMKII-induced modulation of Kv4.3-encoded currents and eliminated the effects of CaMKII inhibitors .
Syntaxin 1A, a membrane associated protein involved in neurotransmitter exocytosis, has also been shown to interact with Kv4.2 α subunits. In addition, syntaxin 1A co-immunoprecipitates with Kv4.2 from rat heart and co-expression of syntaxin 1A with Kv4.2 decreases cell surface channel expression and modulates gating kinetics . It appears that syntaxin 1A shares the KChIP2 binding site (Figure 2) in the Kv4.2 N-terminus and affects KChIP2-mediated modulation of Kv4.2 channels . Accumulating evidence also suggests that ion channels are functionally modulated by cell-cell and cell-extracellular matrix interactions. In neonatal rat ventricular myocytes in culture, for example, cell-cell contact results in down-regulation of Kv4.2 expression . Although a relatively unexplored area in cardiac physiology, it seems certain that there will be increased emphasis on defining the physiological role(s) of extracellular matrix components, as well as components of the intracellular actin cytoskeleton, in the regulation of myocardial Ito (and other Kv) channels and on delineating the underlying molecular mechanisms controlling these pathways under physiological and pathophysiological conditions.
Also first described in Purkinje fibers , it is now well recognized that ventricular and atrial Ito,f densities are modulated physiologically by adrenergic receptor activation . It has also been suggested that adrenergiic stimulation is an important determinant of Ito,f densities in the diseased myocardium . The effects of myocardial α-adrenergic and β-adrenergic receptor stimulation are mediated through the activation of intracellular second messender cascades, culminating in the activation of protein kinases and resulting in the regulated phosphorylation and dephosphorylation of downstream targets, including ion channels. Indeed, several protein kinases and phosphatases have been linked to the regulation of myocardial Kv channels, particularly Ito channels , and multiple potential phosphorylation sites for cAMP-dependent protein kinase (PKA), protein kinase C (PKC), extracellular signal regulated kinase (ERK) and CaMKII have been identified in Kv4 (Figure 2) and Kv1.4 α subunits. Of these, CaMKII has emerged as a pivotal regulator of both cardiac Ito,f (Kv4.2/4.3) and Ito,s (Kv1.4) channels. A serine-threonine kinase, CaMKII, is activated by binding of Ca2+-bound calmodulin (Ca2+/CaM), resulting in CaMKII autophosphorylation . In heterologous cells, CaMKII activation slows the inactivation and accelerates the recovery from inactivation of Kv1.4 channels via phosphorylation of an N-terminal consensus CaMKII phosphorylation site, suggesting a physiological role for CaMKII in the regulation of Ito,s .
Considerable evidence also suggests a role for CaMKII in the regulation of Ito,f channels. CaMKII, for example, co-immunoprecipitates with Kv4.2/Kv4.3 subunits from rat heart [169,188]. In addition, inhibition of CaMKII accelerates Ito,f inactivation in human atrial and rat ventricular myocytes [188,189]. Modulatory effects of CaMKII on heterologously expressed Kv4 channels have also been described and, interestingly, the effects on Kv4.2- and Kv4.3-encoded channels are distinct [188,190]. In HEK-293 and CHO cells, inhibition of CaMKII accelerates the inactivation of Kv4.3 currents [169,188,190]. The opposite effect is produced by autophosphorylated (activated) CaMKII  and is mediated by phosphorylation of a CaMKII consensus motif in the C-terminus (Figure 2) of Kv4.3 . The association of CaMKII with Kv4.3 occurs in the absence of increased intracellular free Ca2+ concentration ([Ca2+]i), and the Kv4.3 protein is phosphorylated at low [Ca2+]i . In contrast, co-transfection of activated CaMKII with Kv4.2 and KChIP3 in COS-7 cells increased total and cell surface Kv4.2 expression via phosphorylation of two CaMKII consensus motifs in the C-terminus of Kv4.2, resulting in increased current amplitudes without measurable changes in gating . In HEK-293 cells, increasing [Ca2+]i reportedly resulted in Kv4.2 phosphorylation and a slowing of inactivation . The latter effect is prevented by inhibition of CaMKII, suggesting that increased [Ca2+]i results in the modulation of Kv4.2 channels by endogenous CaMKII .
It has also been reported that increased CaMKII activity down-regulates KCND3 (Kv4.3) transcript expression, resulting in decreased Ito,f densities in isolated canine ventricular myocytes . Interestingly, rapid pacing increases CaMKII activity in Langendorff-perfused rabbit hearts  and in isolated canine ventricular cells . Accumulating evidence suggests that calcineurin and its downstream target, the nuclear factor of activated T-cells (NFAT), transcription factor, are important in the regulation of Ito,f channels. In isolated adult rat ventricular myocytes, for example, increased Ca2+ influx results in decreased Ito,f densities as a result of down-regulation of Kv4.2 transcript expression mediated by calcineurin, and independent of CaMKII . In mouse left ventricles, transmural differences in [Ca2+]i appear to underlie the differential activation of calcineurin and NFAT across the ventricular wall . Higher activity of calcineurin/NFAT in the endocardium down-regulates Kv4.2 expression, resulting in reduced Ito,f densities . In a mouse model of acute myocardial infarction, Ito,f density is decreased as a result of activation of the calcineurin/NFAT pathway and the transcriptional down-regulation of Kv4.2 . In isolated neonatal rat ventricular myocytes, however, it has been reported that expression of a constitutively active form of calcineurin increases Ito,f densities through the up-regulation of Kv4.2 transcript expression .
It has been reported PKA activation by cAMP inhibits Ito in rat ventricular myocytes , although another study revealed no effects of cAMP/PKA on Ito . Nevertheless, there are two consensus PKA phosphorylation sites in the N- and C-terminal cytoplasmic domains of Kv4.2 (Figure 2), and both are phosphorylated in heterologous cells by PKA . Interestingly, PKA-mediated phosphorylation of serine 552 in Kv4.2 modulated the properties of Kv4.2-encoded currents only in the presence of KChIP3 . It has also been reported that (serine 552) phosphorylation of the Kv4.2 protein is increased with KChIP co-expression . In embryonic rat neurons, PKA activation (and serine 552 phosphorylation) leads to Kv4.2 internalization . Although there are also PKA phosphorylation sites in KChIP3, phosphorylation of these sites is not necessary for PKA modulation of Kv4.2-KChIP3 channels . Similarly, Kv1.4 is a substrate for PKA, and an N-terminal PKA phosphorylation site in Kv1.4 has been identified . In diabetes, angiotensin II (AT-II) activation suppresses PKA activity, resulting in inhibition of Ito,f in (rat) ventricular myocytes .
Inhibitory effects of PKC have been observed on heterologously expressed Kv4.2/4.3- and Kv1.4-encoded  currents and activation of PKC reduces Ito amplitudes in rat and canine ventricular myocytes [206,208–211]. Two alternatively spliced variants of Kv4.3 have been identified in rat [212,213] and human [93,214], and the full-length variant contains consensus PKC phosphorylation sites (Figure 2) [93,214] that have been suggested to underlie PKC-dependent down-regulation of Ito,f mediated by α-adrenergic receptor activation . It has also been suggested that PKC-mediated phosphorylation underlies endothelin-induced inhibition of Kv1.4- and Kv4.3-encoded currents . Two PKC phosphorylation sites have also been identified in the C-terminus of Kv4.2 , and inhibition of PKC phosphorylation increases Kv4.2 surface expression and the amplitudes of Kv4.2-encoded currents .
In rat ventricles, activation of PKC negatively regulates Kv4.2 protein  and KChIP2 transcript  expression. In neonatal rat ventricles, PKC-dependent transcriptional down-regulation of KChIP2 appears to be mediated by the activation of the mitogen-activated protein kinase/ERK kinase (MEK) pathway . Inhibition of MEK increased KChIP2 transcript expression . In brain, basal ERK phosphorylation of the Kv4.2 C-terminus is observed [219,220], and inhibition of MEK increases neuronal Kv4.2-encoded current (IA) amplitudes and shifts the voltage-dependences of current activation and inactivation [221,222]. There are three ERK phosphorylation sites in the Kv4.2 C-terminus . Of these, phosphorylation at threonine 607 mimics the effects of ERK activation on Kv4.2-KChIP3 encoded currents . Interestingly, co-expression with KChIP3 appears necessary for the modulatory effects of ERK-mediated phosphorylation of Kv4.2 . It has also been suggested that the MEK/ERK pathway acts downstream of PKA and PKC [222,223].
Other protein kinases also reportedly modulate Ito,f channels. Cardiac-specific overexpression of p90 ribosomal S6 kinase (p90RSK), for example, attenuates Ito,f densities in mouse ventricular myocytes , and activation of p90RSK appears to mediate hydrogen perioxide-induced modulation of heterologously expressed Kv4.3- (with and without KChIP2) encoded currents . In addition, inhibition of protein tyrosine kinase (PTK) activity suppresses Ito,f in human atrial  and rat ventricular  myocytes. It has also been reported, however, that inhibition of PTK attenuates Ito,f only in ventricular myocytes from diabetic animals . In heterologous cells, inhibition of PTK reduced, and inhibition of tyrosine phosphatase augmented, Kv4.3-encoded currents . In addition, Kv4.3 co-immunoprecipitates with a non-receptor PTK, c-Src, and is phosphorylated on tyrosine . It was also reported that the phosphorylation of KChIP3 by G-protein-coupled kinase (GRK)2 negatively regulates the membrane trafficking of Kv4.2-KChIP3 channels in heterologous cells .
Other post-translational modifications of Kv4- and Kv1.4-channels, although suggested, have been much less well studied. In heterologous cells, for example, glycosylation of Kv1.4 increases channel stability, trafficking and cell surface expression . Palmitoylation, in contrast, appears to regulate Kv channel membrane expression through effects on accessory subunits. Palmitoylation of PSD-95, for example, has been shown to be required for the trafficking and clustering of Kv1.4 channels [230,231]. In addition, palmitoylation of KChIP2 appears to play a role in regulating the membrane expression of Kv4 channels .
Hormonal regulation of Ito channels has also been described. In pregnancy-induced cardiac hypertrophy (in mice), for example, increased estrogen was associated with reduced Ito,f densities . A subsequent study reported that MiRP1 expression is up-regulated in response to elevated estrogen , further suggesting a role for MiRP1 in the regulation of Ito,f. In diabetes, however, estrogen appears to prevent the reduction of Ito,f and Kv4.2 expression in female rat ventricular myocytes through inhibitory effects on rennin-angiotensin system by suppression of oxidative stress, PKA inhibition and receptor PTK activation [200,235,236]. In contrast, testosterone significantly increases Kv4.3 expression, leading to marked shortening of ventricular repolarization in canine heart . Importantly, sex hormone-dependent regulation of Ito,f channels has been suggested to underlie gender differences in arrhythmia susceptibility, such as the observed male predominance in the Brugada syndrome [238,239]. Thyroid hormone treatment also reportedly differentially regulates Ito channels and Kv4.2/Kv4.3 expression in rodent ventricular myocytes by a transcriptional mechanism [240–243], whereas Kv1.4 expression is negatively regulated [240–242,244]. Indeed, it has been suggested that thyroid hormone plays a role in postnatal up-regulation of Ito,f channels and the developmental change (Kv1.4 to Kv4.2/Kv4.3) in relative Kv α subunit expression levels [240,241,245].
Autocrine/paracrine factors also regulate Ito channels. AT-II, for example, decreased Ito,f densities in ventricular myocytes isolated from neonatal rats  and in adult canine epicardial myocytes  via the activation of the angiotensin II type-1 receptor (AT1R) . In addition, the inhibitory effect of AT-II on Ito,f is implicated in the initiation of short-term cardiac memory . Incubation with AT II induced transient reductions of Kv4.3 transcript levels via AT1R activation , an effect mediated by superoxide and the downstream signaling, apoptosis signal-regulating kinase 1 (ASK1)-p38 kinase pathway [249,250]. In diabetic rat, AT-II levels are elevated  and ventricular Ito,f densities are reduced, reflecting AT-II-dependent inhibition of PKC and PKA, the activation of receptor PTK and reduced Kv4.2 expression [200,210]. It has also been shown that AT-II and AT1R control the surface expression of Ito,f channels . In canine ventricles, AT1R forms a complex with Kv4.3, and in heterologous cells, application of AT-II leads to the internalization of Kv4.3-AT1R complexes, resulting in marked reductions in Ito,f densities . It has also been suggested that AT-II-mediated (inhibitory) regulation of Ito,f contributes to the transmural gradient of Ito,f in canine ventricles . Consistent with this hypothesis, AT-II transcript expression is substantially higher in the endocardium, than in the epicardium, in human ventricles . It has also been demonstrated that AT-II and AT1R activation modulate Ito,f gating in canine and mouse ventricular myocytes [247,253] and Kv4.3-KChIP2-encoded currents in heterologous cells .
Considerable evidence now suggests that alterations in cellular metabolism also have marked effects on the expression/functioning of cardiac Ito channels, although the underlying molecular mechanisms are poorly understood. In the streptozotocin-induced model of type 1 diabetes, for example, ventricular Ito,f amplitudes are reduced [254–257] and the expression Kv4.2/Kv4.3 transcripts and proteins are decreased, whereas Kv1.4 expression is increased [257–259]. Electrophysiological studies, however, did not reveal measurable changes in Ito recovery kinetics [254–256]. Reduced ventricular Ito density was also documented in (male) db/db mice, a genetic model of type 2 diabetes [235,260], although no changes in Ito,f densities/properties were reported in another (rat) model of type 2 diabetes produced by high fructose feeding . In the diabetic heart, >90% of the ATP is generated from oxidation of fatty acids [262,263], and, incubation with insulin [163,261,264], as well as agents that enhance glucose oxidation , normalize Ito,f densities in ventricular myocytes isolated from diabetic rats. In addition, in vivo insulin administration partially prevents diabetes-induced down-regulation of Ito,f and Kv4.3 expression in canine ventricles . Interestingly, ventricular Ito,f densities are also decreased in a mouse model of diabetic cardiomyopathy in which fatty acid uptake and utilization are increased as a result of overexpression of peroxisome proliferator-activated receptor α (PPARα) , although Ito,f is not affected in another mouse model of lipotoxic diabetic cardiomyopathy produced by overexpression of the fatty acid transport protein 1 . In addition, increased intracellular fatty acid metabolites reportedly suppress Ito,f densities in rat ventricular myocytes . It has also been suggested that insulin signaling via the MEK/ERK pathway plays a role in the regulation of Ito,f expression in ventricular myocytes from diabetic rats .
Metabolic changes associated with diabetes, including hyperglycemia, impaired glucose metabolism and increased AT-II, are all associated with an oxidized cellular environment [236,268], and oxidative stress has been implicated in the remodeling Ito,f. It has, for example, been shown that thioredoxin and glutathione reductase activities are decreased in the diabetic (rat) heart and that inhibition of thioredoxin and glutathione reductase actually prevents the up- regulation of Ito,f in diabetic rat ventricular myocytes exposed to insulin [269,270]. Interestingly, atrial Ito,f is not suppressed in experimentally-induced diabetes, presumably reflecting an inhibitory action of atrial natriuretic peptide on the renin-angiotensin system (and oxidative stress) . In experiments on left ventricular myocytes from non-diabetic rats, experimentally-induced oxidative stress decreased Ito,f densities, and down-regulated Kv4.2 expression , an effect mediated by receptor tyrosine kinase signaling  and the activation of phosphatidylinositol 3 (PI3)-kinase, MEK and p38 MAP kinases .
Considerable progress has been made in efforts focused on understanding the properties and the roles of cardiac Ito channels, as well on defining the molecular mechanisms that regulate the functioning of these channels. In atrial and ventricular myocardium, Ito channels play a prominent role in the early phase of repolarization. The two functionally distinct cardiac Ito channels, Ito,f and Ito,s, are encoded by Kv4.3/Kv4.2 and Kv1.4 a subunits, respectively. A number of Kv channel accessory (β) subunits have been suggested to function in the generation of cardiac Ito channels. Of these, a crucial role has been definitively demonstrated only in the case of KChIP2 . Recently, however, a potentially important role for MiRP2 has been suggested with the identification of a Brugada syndrome mutation in KCNE3 (MiRP2) and the demonstration that co-expression of the mutant MiRP2 protein augments Kv4.3 current densities . Studies in rodents also suggest functional roles for the accessory Kvβ1  and MiRP1  subunits in the regulation of ventricular Ito,f channel densities. In addition, several recent studies suggest that cardiac Ito channels interact with components of the actin cytoskeleton and that these interactions contribute to the regulation of Ito. The SAP97 scaffolding protein, for example, appears to regulate the surface expression of Kv4 channels and to mediate interactions between Kv4.3 and CaMKII . A number of other protein kinases (as well as phosphatases) have been postulated to function in the physiological and/or pathophysiological regulation of Ito channel expression and/or properties [199,205,215,224–226].
In spite of the progress made in defining the molecular determinants of cardiac Ito, the functional roles of the many putative accessory subunits and other regulatory proteins in the regulation of native Ito channels remain poorly understood. In addition, although recent structural information on Kv4 N-terminus-KChIP1 complexes has provided new insights into the sites of interaction of these two proteins [122,123], the molecular mechanisms underlying KChIP2-mediated regulation of Kv4 channel cell surface expression and functioning remain to be defined. Similarly, the mechanisms involved in the regulation of Ito channels by actin cytoskeletal/membrane-associated proteins or the extracellular matrix (and cell-cell contact) are poorly understood. In addition, a number of regulatory proteins and intracellular second messenger signaling cascades have been implicated in the regulation of Ito channels under physiological, as well as pathological, conditions, contributing to the “remodeling” of Ito channels. These signaling mechanisms, however, are quite complex reflecting the interplay between a variety of regulatory proteins, any one of which may be altered in the diseased myocardium. Clearly, an important emphasis of future research will be on deciphering these mechanisms and detailing the molecular determinants of Ito channel functioning in the (normal and diseased) myocardium.
The authors thank past and present members of the laboratory for their contributions to our understanding of the molecular determinants of cardiac Ito channel expression, function and regulation. We are also indebted to Rick Wilson for his expert assistance with the generation of tables and figures. In addition, we gratefully acknowledge the long-standing and continued financial support for our research endeavors provided by the American Heart Association and the National Heart, Lung and Blood Institute of the National Institute of Health.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.