I. Background

Top Previous Next References

Potassium channels are a diverse and ubiquitous family of membrane proteins present in both excitable and nonexcitable cells. Members of this channel family play critical roles in cellular signaling processes regulating neurotransmitter release, heart rate, insulin secretion, neuronal excitability, epithelial electrolyte transport, smooth muscle contraction, and cell volume regulation. Over 50 human genes encoding various K⁺ channels have been cloned during the past decade (Fig. 1), and precise biophysical properties, subunit stoichiometry, channel assembly and modulation by second messenger and ligands have been addressed to a large extent. More recently, the crystal structure of a K⁺ channel from Streptomyces lividans has become available (Doyle et al., 1998).

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Growth of genes encoding diverse K+ channels. The Shaker K+ channel gene was first cloned from Drosophila (Papazian et al., 1987). The gene products indicated along the y-axis include both K+ channel - and auxiliary subunits. The data were obtained from the Entrez database of the National Center for Biotechnology Information (NCBI).

Concurrent with this remarkable progress in our understanding of molecular diversity, structure, and function, a growing number of discoveries have linked K⁺ channel gene mutations with various diseases. Such diseases of the heart, kidney, pancreas, and central nervous system involve either mutation(s) in K⁺ channel gene(s) and/or altered regulation of K⁺ channel function. The enhanced understanding of these diseases, facilitated by a combination of genomic and biophysical approaches, has helped our understanding of how various mutations affect channel function, contributes to disease etiology, and rationalizes novel treatment strategies. In this review, we provide a comprehensive overview of our recent understanding of molecular defects of K⁺ channels in various diseases and its implications for the development of novel prophylactic or therapeutic approaches targeting distinct types of K⁺ channels.

A brief overview of the structural and functional diversity of K⁺ channels is initially provided to enable familiarity with the nomenclature and biophysical and pharmacological characteristics of diverse K⁺ channels. Several extensive reviews are already available on this subject that may be consulted for additional details (Doupnik et al., 1995; Coetzee et al., 1999). Diseases involving other voltage-gated ion channels have been reviewed elsewhere (Ackerman and Clapham, 1997; Lehmann-Horn and Rüdel, 1997; Cooper and Jan, 1999).

K⁺ channels are membrane-spanning proteins that selectively conduct K⁺ ions across the cell membrane along its electrochemical gradient at a rate of 10⁶ to 10⁸ ions/s. To accomplish this, K⁺ channels are endowed with a set of salient features: 1) a water-filled permeation pathway (pore) that allows K⁺ ions to flow across the cell membrane; 2) a selectivity filter that specifies K⁺ as permeant ion species; and 3) a gating mechanism that serves to switch between open and closed channel conformations (Hille, 1992). Since the first gene encoding a K⁺ channel was cloned from Drosophila Shaker mutant (Papazian et al., 1987), more than 200 genes encoding a variety of K⁺ channels have been identified (Fig. 1), all containing a homologous pore segment (S5-S6 linker) selective for K⁺ ions (Hartmann et al., 1991; Yellen et al., 1991). Accordingly, a general classification of K⁺ channels into families is based upon the primary amino acid sequence of the pore-containing subunit. Three groups with six, four, or two putative transmembrane segments are recognized. These include 1) voltage-gated K⁺ channels (Shaker-like) containing six transmembrane regions (S1-S6) with a single pore; 2) inward rectifier K⁺ channels containing only two transmembrane regions and a single pore; and 3) two-pore K⁺ channels containing four transmembranes with two pore regions (Fig. 2). Table 1 lists a generalized classification of various cloned K⁺ channel subunits.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Schematic representation of the structural classification of K+ channel subunits. A, 6-TM subunits. The voltage-gated K+ channels are composed of four subunits each containing six transmembrane segments (S1-S6) and a conducting pore (P) between S5 and S6 with a voltage sensor (positive charge of amino acid residues) located at S4. Some of the voltage-gated K+ channels include an auxiliary -subunit (Kv), which is a cytoplasmic protein with binding site located at the N terminus of the -subunit. The inset shows the general assembly of K+ channels. The homotetrameric K+ channel consists of four identical subunits while different -subunits form heterotetrameric K+ channels. B, 2-TM subunits. The inward rectifier K+ channel belongs to a superfamily of channels with four subunits each containing two transmembrane segments (M1 and M2) with a P-loop in between. C, 4-TM subunits. This represents a class of the K+ channels that has four transmembranes with two P-loops. IACh, muscarine-activated K+ current; IKDR, delayed rectifying K+ current; IKTO, transient outward delayed rectifier; IKUR, ultrarapid delayed rectifier; IKr, cardiac rapid delayed rectifier; IKs, cardiac slow delayed rectifier; IK1, inward rectifier; TWIK, two-pore weak inward rectifier; TASK, TWIK-related acid-sensitive K+ channel; TRAAK, TWIK-related arachidonic acid-stimulated K+ channel.

d. Subunit Interaction and Assembly Domains. As noted previously, K⁺ channels contain four -subunits, which surround a water-filled, K⁺-selective pore (Fig.

2. Two Transmembrane One-Pore Channels. The inward rectifier K⁺ channels (Kirs) belong to a distant superfamily of channels with four subunits each containing a two-transmembrane segment (M1 and M2) and a pore loop in between (Ho et al., 1993; Kubo et al., 1993). These channels conduct K⁺ currents more in the inward direction than outward, and they are important in setting the resting membrane potential. This inward rectification is attributed to gating mechanisms by internal Mg²⁺ and polyamines (spermine, spermidine, etc.) that occlude access of K⁺ to the internal vestibule of a conducting pore (Matsuda, 1991; Ficker et al., 1994; Lu and MacKinnon, 1994; Wible et al., 1994). Like the voltage-gated K⁺ channels, these channels are organized as tetramers (Yang et al., 1995), although a more complex octameric arrangement has been described, as in the case of the ATP-sensitive K⁺ channels involving four inward rectifiers contributing to ion conducting pore and four peripheral sulfonylurea receptors as regulatory subunits (Clement et al., 1997; Inagaki et al., 1997; Shyng and Nichols, 1997).

3. Four Transmembrane Two-Pore Channels. The more recently discovered tandem-pore domain family are weak inward rectifiers with four putative transmembrane domains and two pore domains (Ketchum et al., 1995; Lesage et al., 1996a). They represent perhaps the most abundant class of K⁺ channels (at least in C. elegans), with >50 distinct members (Wang et al., 1999). The G(Y/F)G residues of K⁺-selective motif is preserved in the first pore loop of the two-pore K⁺ channel, but it is replaced by GFG or GLG in the second pore loop. Although all the two-pore channels have a conserved core region between transmembrane segments M1 and M4, the amino- and carboxyl-terminal domains are quite diverse. With two-pore domain subunits, two such subunits would presumably form a channel to retain the tetrameric arrangement.

Auxiliary subunits that associate with many of the pore-forming subunits have also been described (reviewed in Isom et al., 1994). For example, the Kv1 channels associate with cytoplasmic -subunits to alter channel kinetics (reviewed in Xu and Li, 1998). More recently, chaperone proteins, such as KChAP, regulating the function and expression of some of the Kv channels, such as Kv2.1, Kv1.3, and Kv4.3, have been reported (Kuryshev et al., 2000b). Certain other Kv channels, such as Kv5, Kv6, Kv8, and Kv9, do not form functional channels themselves but associate with Kv2.1 channels to alter the biophysical properties (Salinas et al., 1997; Kramer et al., 1998). Other examples include distinct -subunits that associate with the calcium-activated K⁺ channels (Tseng-Crank et al., 1996; Wallner et al., 1999a; Behrens et al., 2000; Brenner et al., 2000; Meera et al., 2000), sulfonylurea receptors for the inward rectifiers Kir6.1 or Kir6.2 (Aguilar-Bryan et al., 1995; Inagaki et al., 1995a), and minK and minK-related peptides (MiRPs) for the cardiac delayed rectifier channels (Barhanin et al., 1996; Sanguinetti et al., 1996; Abbott et al., 1999). These subunits play roles as diverse as modulation of gating properties such as inactivation, cell surface expression, and/or trafficking of the ion channel complex, to serving as binding sites for both endogenous and exogenous ligands. Given the diversity of K⁺ channel subunits and the potential to vary the constituents to form diverse - or - heteromeric channel complexes to alter expression, cellular targeting, and biophysical and pharmacological properties in native cell types, understanding the precise composition of channel complexes in vivo remains a challenge.

Initial studies of the structure and function of K⁺ channels by a combination of mutagenesis and biophysical approaches have revealed domains that are responsible for K⁺ selectivity, gating, channel assembly, subunit interaction, and drug binding sites. However, the three dimensional structural implications remained largely speculative. Recent discovery of the crystal structure of the KCsA channel established a blueprint of K⁺ channel structure with 3.2 Å resolution (Doyle et al., 1998). The KCsA channel is encoded by a bacterial gene cloned from S. lividans on the basis of sequence homology to K⁺-selective motif GYG in the P-loop (Schrempf et al., 1995). The KCsA channel contains only two transmembrane domains with an intervening pore loop, although at the amino acid sequence level, this channel is more similar to the voltage-gated K⁺ channels. Functionally, it lacks any hint of voltage gating because of the lack of S4 region. X-ray analysis revealed that four identical subunits form a tetramer creating an inverted cone, cradling the selectivity filter of the pore in its outer end. The overall length of the conducting pore is 45 Å, and its diameter is variable along its distance. The internal vestibule of the pore begins as a tunnel of 18 Å in length that widens into a cavity (~10 Å across) near the middle of the membrane, with the narrow selectivity filter only 12 Å long. The remainder of the pore is wider and lined with hydrophobic amino acids. The selectivity filter is lined by the carbonyl oxygen atoms of the GYG signature sequence, which is held open by structural constraints to coordinate K⁺ ions (~3 Å) but not smaller Na⁺ ions because the diameter is too wide to substitute for the hydration energy of the Na⁺ ions (Doyle et al., 1998). The crystal structure of KCsA channel provides the first three-dimensional structure of the conduction pore that fits consistently with current understanding of the core functionality of K⁺ channels. However, structural information of the remaining transmembrane segments (S1-S4), particularly the voltage sensor and the gate coupling to channel opening and closing, remains to be elucidated. Nevertheless, the understanding of structural information can be applied to design selective compounds targeting K⁺ channels. For example, a structure-based design strategy allowed several charybdotoxin analogs to be prepared with about 20-fold higher affinity to block Ca²⁺-activated K⁺ channels versus voltage-gated Kv1.3 channels (Rauer et al., 2000). It is to be anticipated that a detailed understanding of the structural aspects would revolutionize and refine approaches targeting K⁺ channels for therapeutic purposes.

	II. Pathophysiologic Regulation of K+ Channels: Genetically Linked Diseases

Top Previous Next References

Advances in genetic linkage analysis during the past decade have greatly facilitated the identification of many disease-producing loci. Both positional cloning and candidate gene approaches have been used. Using positional cloning techniques, it has become possible to identify the location of genetic locus responsible for a given hereditary syndrome without prior knowledge of the biochemical or physiological abnormalities underlying the disease. Alternatively, following identification of genes encoding proteins that may be logically altered in a particular disease, the candidate gene approach may be used to examine genetic linkage to the hereditary disease of interest and screened for mutations.

As K⁺ channels play fundamental roles in the regulation of membrane excitability, it is to be expected that both genetic and acquired diseases involving altered functioning of neurons, smooth muscle, and cardiac cells could arise subsequent to abnormalities in K⁺ channel proteins. Genetically linked diseases of the cardiac, neuronal, renal, and metabolic systems involving members of voltage-gated K⁺ channels, inward rectifiers, and channel-associated proteins are discussed in the following sections (Table 1).

K⁺ channels are critical to cardiac excitability because they play a fundamental role in repolarization of the action potential. Unlike the action potentials of nerves that last only a few milliseconds, the action potentials of ventricular myocytes can last several hundred milliseconds. This prolonged depolarization phase is essential for normal excitation-contraction coupling process and renders the myocytes relatively refractory to premature excitation. Various classes of K⁺ channels with different time and voltage dependencies and pharmacological properties function in concert to regulate the heart rate by setting the resting membrane potential, amplitude, and duration of action potential and its refractoriness (Barry and Nerbonne, 1996; Roden and Kupershmidt, 1999; Snyders, 1999). The Kir2.1 current sets the resting membrane potential and contributes to the terminal phase of repolarization. The transient outward K⁺ current (Kv4.3 or Kv1.4), which is Ca²⁺-independent and expressed in a species- and cell type-specific fashion, is important for the early phase of repolarization. The long ventricular action potentials that result from the slow onset of repolarization are controlled mainly by two types of delayed rectifier K⁺ currents, i.e., IKs (derived from KCNQ1/minK) and IKr (derived from hERG/MiRP1). Both genetic linkage analysis and the candidate gene approach revealed that mutations in these delayed rectifier K⁺ channel subunits form the molecular basis of LQT syndromes (Curran et al., 1995; Sanguinetti et al., 1995; Schott et al., 1995; Wang et al., 1996; Neyroud et al., 1997; Splawski et al., 1997b; Abbott et al., 1999).

The LQT syndromes are inherited genetic disorders characterized by prolonged or delayed ventricular repolarization, manifested on the electrocardiogram (ECG) as a prolongation of the QT interval. Table 2 lists K⁺ and other ion channel genes involved in various forms of inherited LQT syndromes, LQT1 through LQT6. The inherited LQT causes syncopal attacks and high risk of sudden death as result of torsade de pointes polymorphic ventricular tachycardia, typically triggered by adrenergic arousal (Ackerman and Clapham, 1997; Sanguinetti and Spector, 1997; Vincent et al., 1999). Based on genetic origins, two allelic diseases are recognized: 1) the Romano-Ward syndrome inherited as a dominant trait and 2) the autosomal recessive Jervell and Lange-Nielsen syndrome. In the case of the latter, the patient suffers from a severe congenital bilateral deafness in addition to the cardiac disorder (Vincent et al., 1999). Note that in addition to genetically linked LQT syndromes, many drugs are also known to cause QT prolongation leading to torsade de pointes (see Section III.).

1. Long-QT1 and Long-QT5 Syndromes: KCNQ1 (KvLQT1) and minK. KvLQT1, encoded by the KCNQ1 gene, in association with the minK subunit, a short peptide of 130 residues, constitutes the IKs responsible for phase 3 repolarization in the heart (Barhanin et al., 1996; Sanguinetti et al., 1996b). Several mutations in the KCNQ1 gene, including missense mutations, intragenic deletion, and insertions, are involved in chromosome 11-linked LQT1 syndrome, the most common form of inherited LQT in families with Jervell and Lange-Nielsen and Romano-Ward syndromes (Russell et al., 1996; Wang et al., 1996; Donger et al., 1997; Tanaka et al., 1997; van den Berg et al., 1997; Saarinen et al., 1998; Li et al., 1998; Neyroud et al., 1999;). Functional analysis of mutant channels in COS cells cotransfected with the minK subunit revealed that these mutations either alter gating properties or fail to produce functional homomeric channels and reduced K⁺ current when coexpressed with the wild-type subunit (Chouabe et al., 1997; Shalaby et al., 1997; Wollnik et al., 1997; Franqueza et al., 1999).

2. Long-QT2 Syndrome and Human ether-a-go-go-Related K⁺ Channel. The hERG gene encoding a rapidly activating IKr is a major subunit responsible for repolarization during cardiac action potential (Sanguinetti et al., 1995). Interaction with hERG channels has been shown to be a primary mechanism involved in the therapeutic actions of the class III antiarrhythmic agents and the potential cardiotoxicity of second generation H₁ receptor antagonists, such as terfenadine and astemizole, as well as certain antidepressants and neuroleptics (Vincent et al., 1999).

2. Benign Familial Neonatal Convulsions and KCNQ2/KCNQ3. Recent application of genetic analysis to hereditary epilepsy has provided the impetus for the identification of mutations in genes encoding various ion channels, including K⁺ channels (Biervert et al., 1998; Charlier et al., 1998; Singh et al., 1998). BFNC is an idiopathic form of epilepsy beginning within the first six months after birth. Seizures are generalized and mixed, starting with tonic posture, ocular symptoms, and apnea, and often progress to clonic movements and motor automatisms. Seizures last 1 to 2 s and occur three to six times per day. Two forms of benign familial neonatal convulsions, BFNC1 and BFNC2, are typically observed in families as an autosomal dominant inheritance and have been previously mapped into chromosomes 20q and 8q, respectively (Leppert and Singh, 1999). By positional cloning techniques, the voltage-gated K⁺ channel KCNQ2, spanning the deletion region of chromosome 20q13.3 that cosegregates with seizures in a BFNC family, was identified (Biervert et al., 1998; Singh et al., 1998). Missense mutation, frameshifts, and splice-site mutations in KCNQ2 were also found in other BFNC families. By a homology search of expressed sequence tag database and genotyping approaches, a missense mutation in the pore region of another voltage-gated K⁺ channel, KCNQ3, was also identified from families with BFNC2 previously linked to chromosome 8q24 (Biervert et al., 1998; Charlier et al., 1998; Schroeder et al., 1998).

3. Neurodegeneration and Kir3.2. The progressive loss of dopaminergic neurons in the weaver mouse is similar to the pathological symptom of Parkinson's disease where cell death of dopaminergic neurons in the substantia nigra is observed, leading to striatal dopaminergic deficit and a clinical syndrome dominated by disorders of movement (Yamada et al., 1990; Gaspar et al., 1994). The weaver phenotype in mice is an autosomal recessive neurological and reproductive disorder characterized behaviorally by severe ataxia, hyperactivity, and tremors that are manifested within 2 weeks after birth. These behavioral changes are attributable to the degeneration of cerebellar granule cells and dopaminergic neurons in the substantia nigra (Rakic and Sidman, 1973a,b). In addition, wv/wv genotype causes death or impaired function of dopaminergic neurons in the substantia nigra, male infertility, and sporadic tonic-clonic seizures (Hess, 1996; Harrison and Roffler-Tarlov, 1998). While heterozygous mice are not ataxic, they have seizures and a significant reduction in the number of granule cells.

4. Schizophrenia and SK3 (hKCa3). Although initially differentiated on the basis of biophysical and differential toxin sensitivity, distinct genes are now known to encode various calcium-activated K⁺ channels (Vergara et al., 1998; Castle, 1999; Wallner et al., 1999b). Abnormal function of calcium-activated K⁺ channels has been noted in platelets of patients with Alzheimer's disease, although its relevance to the pathology is not clear (de Silva et al., 1998). The CAG triplet repeat in KCNN3 gene encoding a small conductance calcium-activated K⁺ channel, hKCa3, mapped to chromosome 1q21 has been reported to be associated with schizophrenia (Chandy et al., 1998), although subsequent investigations to confirm these findings have been met with mixed results (Austin et al., 1999; Dror et al., 1999).

Much progress has been made in the area of identifying genes defective in hearing and balance disorders, with over 40 such genes described (Holt and Corey, 1999). One of the genes reported to be the locus for hereditary hearing impairment is another K⁺ channel belonging to the KCNQ channel superfamily, i.e., KCNQ4. The KCNQ4 gene, isolated from a human retina library using KCNQ3 partial cDNA as a probe, exhibits 38, 44, and 37% identity to KCNQ1, KCNQ2, and KCNQ3, respectively (Kubisch et al., 1999). Reverse transcriptase-polymerase chain reaction analysis revealed high expression of KCNQ4 in the vestibular system and brain. In cochlea sections from mice at postnatal day P12, sensory outer hair cells were strongly labeled with a KCNQ4 antisense probe but not in the inner hair cells and stria vascularis where KCNQ1 expression was detected. Expression of KCNQ4 in Xenopus oocytes generated a voltage-dependent K⁺ current, similar to KCNQ1, KCNQ2, and KCNQ3, except with slower activation. Unlike KCNQ1, KCNQ4 did not interact with minK. However, coexpression of KCNQ3 with KCNQ4 yielded currents resembling an M-channel, but with only weak inhibition (75% inhibition at 200 µM) by linopirdine, unlike those observed with the KCNQ2/KCNQ3 combination. The similarity of currents from KCNQ3/KCNQ4 to M-channel indicated that KCNQ3/KCNQ4 might potentially form another M-channel variant in the nervous system (Kubisch et al., 1999).

Using fluorescence in situ hybridization to human chromosomes, KCNQ4 was mapped to chromosome 1p34, a region also hosting DFNA2, a locus for autosomal dominant progressive hearing loss (Kubisch et al., 1999). One 13-bp deletion mutation and four missense mutations (G285S, G285C, W276S, and G321S) were identified from families with autosomal dominant progressive hearing loss linked to the DFNA2 locus (Coucke et al., 1999; Kubisch et al., 1999). The G285S and G285C mutations alter the first glycine residue in the GYG signature sequence of K⁺ channel pore. Mutations in these amino acids disrupt the selectivity filter and, in most cases, abolish channel function. An identical change in amino acids at the equivalent position has also been reported in the KCNQ1 gene of a patient with the dominant LQT1 (Russell et al., 1996). Functional analysis reveals that the mutant channel did not produce current when the cRNA was injected into oocytes, whereas the mutation exerted a dominant-negative effect when coexpressed with wild-type KCNQ1. Whereas mutations in KCNQ1 affect endolymph secretion, the mechanism leading to KCNQ4-related hearing loss appears to be in outer hair cells (Kubisch et al., 1999), inner ear, and the central auditory pathway (Kharkovets et al., 2000).

It must be pointed out that in addition to mutations in KCNQ4, mutations in GJB3, which encodes the connexin 31 component of gap junctions and was mapped to human chromosome 1p33-p35, were identified from the DFNA2 family with nonsyndromic autosomal dominant hearing loss (Xia et al., 1998). Although at least two or three genes responsible for hearing impairment are located close together on chromosome 1p34, KCNQ4 mutations may be a relatively frequent cause of autosomal dominant hearing loss.

Several transporters and ion channels in the renal epithelium play important roles in urine production, fluid balance, and electrolyte metabolism. Genetic analysis reveals that dysfunction of an inward rectifier K⁺ channel Kir1.1 is linked to Bartter's syndrome, an autosomal recessive inherited renal tubular disorder characterized by hypokalemia, metabolic alkalosis, hyper-reninism and hyperaldosteronism. Patients have normal or low blood pressure and renal salt loss despite increased plasma renin activity and high serum aldosterone levels (Karolyi et al., 1998; Simon and Lifton, 1998; Scheinman et al., 1999). At least three phenotypically different renal tubulopathies have been identified: antenatal Bartter's syndrome (hyperprostaglandin E syndrome), classic Bartter's syndrome, and Gitelman's syndrome. Of these, polyhydramnios, premature delivery, hypokalemic alkalosis, hypercalciuria, and dehydration at birth characterize the antenatal Bartter's syndrome (hypokalemic alkalosis with hypercalciuria). Children with the antenatal Bartter's syndrome present the typical pattern of impaired salt reabsorption in the thick ascending limb of Henle's loop resulting in the marked ante- and postnatal salt wasting.

Genetic heterogeneity of antenatal Bartter's syndrome has been demonstrated initially by identification of mutations in the SLC12A1 gene, encoding for the bumetanide-sensitive sodium potassium 2 chloride cotransporter (NKCC2) leading to defective reabsorption of sodium chloride in the thick ascending limb of Henle's loop (Simon et al., 1996a; Vargas-Poussou et al., 1998). Subsequently, several mutations in KCNJ1, encoding the apical renal outer medullary inward rectifying K⁺ channel (Kir1.1), were identified in patients with antenatal Bartter's syndrome by single strand conformation polymorphism analysis (Simon et al., 1996b; Derst et al., 1997; Feldmann et al., 1998; Vollmer et al., 1998). Functional studies revealed that mutant channels expressed none or significantly reduced currents compared with the wild-type channel. This impaired K⁺ flux and loss of tubular K⁺ channel function probably prevents apical membrane potassium recycling with secondary inhibition of Na-K-2Cl cotransport in the thick ascending limb of Henle's loop (Derst et al., 1997). The mechanisms underlying impaired Kir1.1 function involve abnormalities in phosphorylation, proteolytic processing, and/or protein trafficking (Schwalbe et al., 1998).

The signs and symptoms of Bartter's syndrome are usually a consequence of hypokalemia. Maintaining normal serum K⁺ levels and limiting the degree of metabolic alkalosis are some of the treatment approaches, and potassium supplements and potassium-sparing diuretics are frequently used (Gordon and Stokes, 1994).

PHHI is an autosomal recessive disorder characterized by increased irregularity in insulin secretion leading to hypoglycemia, coma, and severe brain damage in children. Both sporadic and familial variants of PHHI are recognized; familial forms are common in communities with high rates of consanguinity where the incidence may be as high as 1 in 2500 live births and is the most common cause of hypoglycemia in newborns (Aynsley-Green and Hawdon, 1997). Recent genetic linkage analysis has identified mutations in the K_ATP channel complex that regulates insulin secretion from pancreatic -cells. The K_ATP channels predominantly determine the resting potential of -cell and couple cellular metabolism to electrical activity (Ashcroft and Rorsman, 1989; Dukes and Philipson, 1996). When plasma glucose is elevated, increases in intracellular ATP/ADP ratio lead to closure of K_ATP channels and membrane depolarization that, in turn, lead to the activation of voltage-dependent Ca²⁺ channel, rise in intracellular Ca²⁺, and insulin secretion.

The -cell K_ATP channel, like other K_ATP channels described in neurons, cardiac, smooth, and skeletal muscle, are inhibited by intracellular ATP, and recent molecular cloning has shown that the channel is an octamer composed of four subunits of the sulfonylurea receptor SUR1 coupled to four subunits of the inward rectifier Kir6.2 (Inagaki et al., 1995a, 1997; Clement et al., 1997; Shyng and Nichols, 1997). Over 28 naturally occurring mutations in SUR1 (Thomas et al., 1995b; Dunne et al., 1997; Verkarre et al., 1998) and two different mutations in Kir6.2 subunits have been identified in families with PHHI (Thomas et al., 1996a; Nestorowicz et al., 1997; Meissner et al., 1999). No K_ATP channel activity was observed in -cells isolated from a homozygous patient or after coexpression of recombinant Kir6.2 and mutant SUR1 (V187D) (Otonkoski et al., 1999). Detailed functional analysis in COS cells by cotransfection of Kir6.2 with various single mutations of SUR1 identified in the PHHI family suggested this lack of K_ATP channel activity or reduction of K_ATP channel sensitivity to MgADP (Shyng et al., 1998). In fact, patients with mutations in SUR1 either failed to respond to diazoxide or showed diminished sensitivity to treatment (Thornton et al., 1998).

The role of K_ATP channels in -cell function has been evaluated in transgenic mice carrying a dominant-negative form of Kir6.2 (G132S) generated by substituting the glycine lining the pore with serine (Miki et al., 1997). These mice develop hypoglycemia with hyperinsulinemia in neonates and hyperglycemia with hypoinsulinemia and decreased -cell population in adults. K_ATP channel function was found to be impaired in the -cell of transgenic mice with hyperglycemia. These results imply that the K_ATP channel complex might play a significant role in -cell survival and regulation in insulin secretion, suggesting that modulation of Kir6.2 may offer additional opportunities in treatment of diabetes and related conditions of abnormal glucose regulation. More recently, it has been shown that the SUR1 knockout mice, unlike the Kir6.2 counterpart, are not insulin-hypersensitive, although their -cells lacks K_ATP channels and show spontaneous Ca²⁺ transients similar to those seen in PHHI patients. SUR1 knockout mice were normoglycemic until stressed, unlike in PHHI patients whose glucose levels are persistently low suggestive of a role for K_ATP-independent pathways that regulate insulin secretion, at least in mice (Seghers et al., 2000).

	III. Disease- and Drug-Induced Regulation of K+ Channels

Top Previous Next References

K⁺ channels are targets for the actions of transmitters, hormones, or drugs that modulate cardiac functions. Changes in the densities and/or properties of these K⁺ channels that occur during the normal development or as a result of damage or disease can have profound physiological consequences (Matsubara et al., 1993; Xu et al., 1996; Yao et al., 1999). Cardiac failure, a pathophysiologic condition with numerous etiologies including myocardial infarction, hypertension, and myocarditis (Wilson, 1997) is characterized by action potential prolongation and, accordingly, altered expression of a variety of depolarizing and hyperpolarizing membrane currents. In an attempt to compensate for the reduction in cardiac function in cardiac failure, the sympathetic nervous system, the renin-angiotensin-aldosterone systems, and other neurohumoral mechanisms are activated. Adaptive changes at the level of the cardiac myocyte include cellular hypertrophy and altered gene expression. Electrical remodeling in cardiac myocytes leading to action potential prolongation is a common finding in human heart failure and in animal models of cardiac hypertrophy. Changes in a wide range of plasma membrane receptors and intracellular signals such as increased intracellular calcium, cAMP, inositol phosphates, and diacylglycerol concentrations are associated with cardiac hypertrophy and failure (Morgan and Baker, 1991; Gopalakrishnan and Triggle, 1990; Wickenden et al., 1998).

A reduction in the current density of the transient outward current (I_TO) is the most consistent ionic current change in cardiac hypertrophy and failure (Nabauer and Kaab, 1998; Wickenden et al., 1998; Pinto and Boyden, 1999; Tomaselli and Marban, 1999). This outward repolarizing K⁺ current activates and inactivates rapidly with an inactivation constant of ~60 ms (Dixon et al., 1996; Kong et al., 1998). The down-regulation of this current has profound effects on phase 1 and the level of plateau of the action potential, and it also alters currents that are subsequently active along the cardiac action potential. The Kv4.3-containing channel is thought to underlie the bulk of I_TO found in the mammalian heart, although Kv1.4 or Kv4.2 channels might represent another fraction of I_TO with distinct kinetics in different regions of the heart (Dixon et al., 1996; Kong et al., 1998). By ribonuclease protection assays and whole-cell electrophysiological recording, Kaab et al. (1998) found that the level of Kv4.3 mRNA decreased by 30% in human failing hearts compared with nonfailing controls. This observation correlated with the reduction in peak I_TO density measured in ventricular myocytes isolated from adjacent regions of the heart.

It has been known that action potential durations vary across the myocardial wall and in different regions of the mammalian heart (Litovsky and Antzelevitch, 1989; Fedida and Giles, 1991; Lukas and Antzelevitch, 1993; Di Diego et al., 1996). The density of I_TO also varies regionally and transmurally in the heart (Wettwer et al., 1994; Nabauer et al., 1996). Electrophysiological recording from myocytes isolated from patients with aortic stenosis and compensated left ventricular hypertrophy indicates that macroscopic I_TO was absent in superficial subendocardial cells, whereas I_TO current density was not significantly altered in the deeper layers (Bailly et al., 1997). A region-dependent alteration in the density of I_TO current was also observed in the catecholamine-induced hypertrophy in animals (Bryant et al., 1999). It is possible that this region-dependent suppression of I_TO current might, in part, underlie the regional heterogeneity in action potential prolongation in cardiac hypertrophy and may predispose to ventricular arrhythmias, a cause of sudden death in patients with cardiac failure.

As discussed later, an approach to the treatment of heart failure would be to normalize K⁺ channel gene expression by gene transfer or pharmacologic modulation. Recent studies have shown that thyroid hormone treatment can increase Kv4.2 or Kv4.3 expression at the transcriptional level and enhance the recovery rate from the inactivation of I_TO in rat ventricular myocytes (Shimoni et al., 1997