Channelopathies

vii

Foreword It was a simple question, that we tried to answer, when Bert Sakmann and myself made attempts to record single-channel currents in the early 1970s. What are the molecular mechanisms underlying the electrically and chemically induced permeability changes in excitable tissue: ion channels or other kinds of transporters? Our attention was focussed on the `traditional' excitable cells: nerve, muscle neuroendocrine. We did not anticipate that ion channels are found in basically any cell type and that they mediate an incredible variety of regulatory functions. Even less would we have anticipated that mutations in channels are the basis of a whole range of disorders. It is rewarding to see that the tools, we created to answer this simple question posed above, in combination with modern molecular genetics would allow us to dissect and to better understand as many diseases as discussed in this volume. Even more, it is fascinating to realize, that the precision with which an ion channels' function can be studied, may teach lessons on pathogenetic mechanisms in general. The development from single channel recording to the detailed knowledge on diseases, presented in this book, is evidence of the credo of `basic' researchers, that bene®ts for the public result in unpredictable ways from progress in basic knowledge and that any important advance in basic research sooner or later will contribute to solutions of very practical relevance. Dr Erwin Neher Professor of Biophysics GoÈttingen, June 14th, 2000

xi

Introduction

Ions channels ± an exciting topic! Editing a book on a topic indicates two things: 1. a considerable amount of knowledge has been accumulated; and 2. the nature of the material is well enough established to become teaching material. Remarkably, this book is possible even though the ®eld of ion channelopathies is only 10 years old ± and the term channelopathies itself was coined only 8 years ago at a European Neuromuscular Center Workshop held in Ulm in 1992 (Neuromusc. Dis. 3:161±168). But what predisposes ion channelopathies of so different medical specialties to be incorporated into one book? It is because the same basic pathomechanism underlies the clinical features, namely sustained membrane decharging (depolarization). Depending on the degree of this disturbance, affected tissue is rendered either in an hyperexcitable (i.e. myotonic muscle stiffness) or an unexcitable state (i.e. ¯accid weakness) and this is capable of explaining the symptoms patients present with! Of Becker and Bryant The work of two pioneer scientists in this ®eld is acknowledged: ² Peter E. Becker (born 1908), neurologist and geneticist; and ² Shirley H. Bryant (1924±1999), biophysicist Both were interested in muscle stiffness, but each with a different approach. While Becker clinically described hereditary myotonia in man, recessive Becker myotonia, for the ®rst time and on an incredibly large epidemiological scale, Bryant was the ®rst to propose that a change of ion conductance, Cl 2-ions, functionally causes the fear-evoked myotonic stiffness in the corresponding animal model, the fainting goats (see Chapter 2). That was 40 years ago, long before such thing as an ion channel was known to exist and when it was generally accepted that ions diffuse into the cells membrane through membrane `holes'. It took another 30 years for the ®rst causative gene encoding for an ion channel to be identi®ed, the adult skeletal

xii

Introduction

muscle sodium channel gene, mutations in which are associated with dyskalemic paralysis, a disease of both man and, ironically, American race horses (see Chapter 1). Patch clamp champs Knowing the gene responsible for an hereditary disorder alone is not of much help without an understanding of the functional defect brought about by the diseasecausing mutations therein. Unique among hereditary disorders is the availability of the technical possibilities to study exactly this by combining modern molecular biology with patch-clamp techniques (Erwin Neher and Bert Sakmann, Nobel Prize for medicine and physiology 1991). Current variants of this technique make the application of solution on the exterior and interior of whole cells and on membrane patches torn from the cell possible (outside-out or inside-out) ± every thinkable con®guration of solution and ion channel orientation that the heart of an ion channel researcher craves for. Channel panel Ion channels do not come alone, but rather in whole families of related proteins conducting each ion type with slightly modi®ed function and varying tissue expression patterns. Structures of importance like pore, selectivity ®lter, voltage sensors, ligand binding sites, and opening and closing gates show conservation apparently for more than 600 million years. This is an evolutionary trick to on the one hand mediate many functions with the aid of one basic mechanism, but on the other hand, to compensate for an eventually disturbed function by closely related channel siblings. Therefore, the ion channelopathies known to date do not lead to death, not even to continuous disability, but rather require an out-of-the-normal situation, a so-called trigger, to present with recognizable symptoms. Watch out for similarities in clinical symptom patterns, triggers, channel structure with localization of disease-causing mutations, functional consequences of mutants, and acute and prophylactic therapy when reading this book. You will be able to identify new channelopathies in your own research ®eld! Let us know!

ix

Preface In the last 10 years the combination of electrophysiological and molecular genetic investigations led to the exploration of the growing family of diseases caused by mutations in genes encoding voltage- and ligand-gated ion channels, the so-called channelopathies. Although the underlying mutations are rare and restricted to single genes expressed in a speci®c tissue, channelopathies may be important models for much more frequent disorders of non-monogenic etiology. Most channelopathies have a certain clinical pattern in common. Typically the symptoms occur as episodic attacks lasting from minutes to days that show spontaneous and complete remission, onset in the ®rst or second decade of life, and ± for some unknown reason ± show amelioration at the age of 40 or 50. Frequently the attacks can be provoked by rest following physical activity or exercise itself, hormones, stress and certain types of food. Surprisingly, many patients with channelopathies respond to acetazolamide, a carbonic anhydrase inhibitor. Most channelopathies show no chronic progression, however, there are a few exceptions which unfortunately do not respond to acetazolamide treatment. Examples are cerebellar degeneration in episodic ataxias or proliferation of the transverse tubular system in periodic paralysis both of which are probably associated with an altered gene expression in the affected cells triggered directly by the mutant proteins or the overall cell dysfunction. This book describes human hereditary ion channel diseases of voltage- and ligand-gated ion channels covering the diverse ®elds of medicine myology, neurology, cardiology, and nephrology requiring a wide and interdisciplinary readership. Interesting parallels in pathogenetic mechanisms of disease are especially emphasized to interest even highly specialized readers in entities outside of their ®elds. Each author has written an objective overview of his or her particular subject in a way that should allow the reader within a short period of time to obtain a comprehensive picture of the present state of the art. The authors and editor anticipate that many more channelopathies will be identi®ed after this book has gone to press. We believe that we know only the tip of the iceberg. Frank Lehmann-Horn Guest Editor

Karin Jurkat-Rott Co-editor

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

3

Chapter 1

Sodium and calcium channelopathies of sarcolemma: periodic paralyses, paramyotonia congenita and potassium-aggravated myotonia Nenad Mitrovic, Holger Lerche Departments of Applied Physiology and Neurology, University of Ulm, D-89069 Ulm, Germany

Abstract Periodic paralyses, paramyotonia congenita and potassium-aggravated myotonia are dominantly inherited muscle disorders caused by mutations in genes encoding voltage-gated sodium or calcium channels. Three so-called sodium channel diseases, hyperkalemic periodic paralysis (HyperPP), paramyotonia congenita (PC) and potassium-aggravated myotonia (PAM) are caused by mutations in the a-subunit of the human skeletal muscle sodium channel, whereas in hypokalemic periodic paralysis (HypoPP) mutations in the a-subunit of the L-type human skeletal muscle calcium channel are found. The clinical phenotype of HyperPP are transient episodes of muscle weakness or paralysis often provoked by rest after exercise or by ingestion of potassium-rich food. HypoPP is also characterized by episodes of generalized paralysis occurring usually in the morning with on average longer duration than in HyperPP. Attacks may be provoked by carbohydrate-rich food or heavy exercise the preceding day. Decisive for classi®cation is the level of serum potassium during a paralytic attack, which may fall below 2 mmol/l in HypoPP, whereas in the hyperkalemic form, it may rise above 5.5 mmol/l. PC is characterized by paradoxical myotonia, which is muscle stiffness increasing with repeated activity, and weakness triggered by exposure to cold. The clinical phenotype of PAM is muscle stiffness induced or aggravated by depolarizing agents such as potassium, typically without weakness or cold-sensitivity. This article provides an overview of clinical features, genetics, pathogenesis and therapy of the sodium and calcium channelopathies of sarcolemma. q 2000 Elsevier Science B.V. All rights reserved.

Introduction In the last 10±15 years, the combination of electrophysiological and molecular genetic investigations led to the identi®cation of the growing family of the so-called

4

N. Mitrovic, H. Lerche

`channelopathies', diseases caused by mutations in different voltage-gated or ligandgated ion channels. Since ion channels provide the basis for the regulation of excitability in nerve and muscle cells, it is not surprising that mutations in channel encoding genes, leading to a dysfunction of these highly speci®c, membrane spanning proteins, result in hyper- or hypoexcitability of the corresponding cells. Meanwhile, channelopathies involving tissues such as skeletal muscle, brain and heart are known. These are myotonias and periodic paralyses, inherited cardiac arrhythmias (the long-QT syndromes), episodic ataxias, familial hemiplegic migraine and some familial epileptic syndromes. The ®rst diseases in which the underlying pathophysiology has been identi®ed to be due to an ion channel defect were skeletal muscle diseases, the non-dystrophicmyotonias and periodic paralyses (Lehmann-Horn et al., 1994). Predominant symptoms of the myotonias and of the periodic paralyses are transiently occurring muscle stiffness and episodes of muscle weakness, respectively. Myotonia is caused by an increased excitability of the sarcolemma, the episodes of weakness by a reduced excitability. Both symptoms are caused by a depolarization of the muscle ®bre membrane. Whereas a slight depolarization will induce hyperexcitability, i.e. myotonia, a strong and sustained depolarization causes hypoexcitability and muscle weakness. The relation is so close that diseases exist that exhibit both symptoms. Non-dystrophic myotonias may be divided into chloride and sodium channel diseases. The chloride channel myotonias Thomsen and Becker are caused by mutations in the gene encoding the skeletal muscle chloride channel ClC-1 and will be described in the next chapter. Sodium channel myotonias are caused by mutations in the gene encoding the adult skeletal muscle sodium channel a -subunit (SCN4A). These are paramyotonia congenita and potassium-aggravated myotonia. Periodic paralysis is the common name for a number of rare diseases characterized by episodes of ¯accid weakness. There are primary and secondary forms. The primary forms show an autosomal dominant mode of inheritance and are commonly classi®ed upon the change in serum potassium concentration during attacks of weakness as hypokalemic, normokalemic and hyperkalemic. Secondary types of periodic paralyses occur in association with hyperthyroidism or with wastage or retention of body potassium. Before genetic studies were possible, an extensive electrophysiological survey, carried out with excised muscle specimens of all kinds of myotonia patients was performed (Lipicky et al., 1971; RuÈdel and Lehmann-Horn, 1985; Lehmann-Horn et al., 1987a,b). Normal excitability of muscle ®bres requires a high resting potential and short-lasting action potentials. Both requirements are not ful®lled in periodic paralyses and myotonias. Voltage-clamp studies on single ®bres have revealed that inactivation of Na 1 currents is incomplete for hyperkalemic periodic paralysis (HyperPP), paramyotonia congenita (PC) and potassium-aggravated myotonia (PAM) causing a depolarization of the sarcolemma (Lehmann-Horn et al.,

Sodium and calcium channelopathies of sarcolemma

5

1987a,b; Lerche et al., 1993). A depolarization is also the underlying pathomechnism for HypoPP (RuÈdel et al., 1984), however, up to now no abnormal membrane conductance has been observed. According to the electrophysiological abnormalities, subsequent genetic studies revealed indeed the gene encoding the a -subunit of the adult human skeletal muscle sodium channel as the site of the defect for HyperPP, PC and PAM. In contrast, HypoPP was linked to the gene encoding the a 1-subunit of the skeletal muscle Ltype calcium channel, the dihydropyridine (DHP) receptor. Several disease-causing point mutations in both channel genes were identi®ed up to now. In order to examine the functional consequences of these mutations on a molecular level, extensive studies using heterologous expression of the mutant genes and the patch clamp technique were carried out. These studies identi®ed different gating defects of the sodium channel affecting mainly the inactivation which can nicely explain the pathophysiology of HyperPP, PC and PAM. The link between the calcium channel mutations and the pathomechanism of HypoPP remains to be elucidated. Sodium channel diseases Voltage-gated sodium channels are responsible for the initiation and propagation of action potentials in excitable cells. There are three main conformational states of the channel protein, a closed or resting state at hyperpolarized potentials, an open state occurring upon depolarization and an inactivated state that follows the opening at maintained depolarization or can be directly reached from the closed state (Fig. 1). In order to regulate the action potential properly, the depolarizing sodium current needs to be quickly activated and inactivated. If inactivation is too slow or incomplete, the repolarizing phase of the action potential is delayed and a stable resting potential can not be maintained which is the common underlying pathomechanism of the sodium channel disorders. Mutations in SCN4A, the gene encoding the adult skeletal muscle sodium channel a -subunit, can cause three clinically distinct syndromes: paramyotonia congenita, potassium-aggravated myotonia and hyperkalemic periodic paralysis. Although many patients show the typical clinical phenotype of only one of these diseases, overlap syndromes do occur. The SCN4A gene contains 24 exons distributed over about 30 kb. The intron±exon boundaries are known and primer sets consisting of intron sequences for ampli®cation of all 24 exons by use of PCR are available (George et al., 1993). SCN4A is only expressed in skeletal muscle and its product is the only sodium channel detectable in the fully differentiated tissue. The adult human muscle sodium channel is a 260 kDa glycoprotein of about 2000 amino acids consisting of four highly homologous domains (I±IV) with six trans-

6

N. Mitrovic, H. Lerche

Fig. 1. A model of the sodium channel in resting, open and inactivated state. The sodium channel opens from a resting closed state upon depolarization and than closes spontaneously into the inactivated state. In order to open again the sodium channel has to recover from inactivation. Two channel gates are indicated, the activation gate (in red), coupled with the voltage sensor which opens the channel and the inactivation gate (in black) which closes the channel pore from the intracellular side.

membrane segments each (S1±S6, Fig. 2). The four interlinkers between segments S5 and S6 dip into the membrane and form the lining of the channel pore (Catterall, 1995). The S4 helices contain four to eight positively charged amino acids at every third position which serve as voltage sensors (StuÈhmer et al., 1989; Yang et al., 1996; Horn, 1998, Fig. 3). The third functional important structure is the interlinker connecting repeats III and IV, which is indispensable for fast inactivation of the channel. Most likely, this part of the protein acts as an inactivation gate or particle in a way that has been compared with a tethered ball or a hinged-lid (Armstrong and Bezanilla, 1977; West et al., 1992, Fig. 4). Yet unknown parts of the intracellular ori®ce of the pore or its surroundings may act as acceptor for the particle. The reader who is interested in naturally occurring animal models of sodium and calcium muscle channelopathies should refer to the overview table of known channelopathies and a recent review (Lehmann-Horn and Jurkat-Rott, 1999).

Sodium and calcium channelopathies of sarcolemma

7

Fig. 2. Mutations predicted in the skeletal muscle sodium channel a -subunit, hSkm-1. Conventional oneletter abbreviations are used for wild-type amino acids (aa). Each aa position is given by the respective number. In two positions (G1306, R1448) three different natural mutations have been detected. The different symbols used for the point mutations indicate the resulting diseases as explained at the bottom of the left-hand side. Included are the positions of mutations causing HyperPP, PC and PAM (modi®ed after Lehmann-Horn and RuÈdel, 1996).

Paramyotonia congenita Clinical features The hallmarks of this disease as ®rst described by Eulenburg (Eulenburg, 1886) and later con®rmed in many families by Becker (Becker, 1970) are: (i) paradoxical myotonia, de®ned as muscle stiffness increasing with continued activity; (ii) severe worsening of the myotonia by cold; (iii) weakness after longer exposure to cold in most cases. In some families patients have spontaneous attacks of weakness like those occurring in HyperPP. The condition is transmitted as a dominant with complete penetrance. Paramyotonic symptoms are present at birth and remain often unchanged for the entire lifetime. In the cold, the face may appear mask-like, and the eyes cannot be opened for several seconds. Working in the cold makes the ®ngers so stiff that the patient becomes unable to move them within minutes. The stiffness then gives way to weakness. After warming, the hands may not regain strength for several hours.

8

N. Mitrovic, H. Lerche

Fig. 3. Model for the movement of the voltage sensor. The voltage sensor S4 in domain IV contains eight positively charged residues, either arginines (R) or lysines (K). Two conformational states of IV/S4 are shown, the inward state at very negative potentials and the outward state at depolarized potentials. For details see Section 5 (taken from Yang et al., 1997 with kind permission).

Sodium and calcium channelopathies of sarcolemma

9

Fig. 4. Hinged-lid model for fast inactivation of the sodium channel. When inserted in the membrane, the four repeats of the sodium channel fold to generate a pore as schematically indicated. The outward movement of the voltage sensor that opens the channel pore is indicated. Three hydrophobic aa (isoleucine, fenylalanine, methionine) were proposed to form an inactivation particle closing the channel pore from the cytoplasmic side in a hinged-lid fashion (modi®ed after West et al., 1992).The hinge could be a pair of glycines, one of which is mutated in PAM (G1306A/V/E, Lerche et al., 1993; Mitrovic et al., 1995, see text).

Under warm conditions many patients have no complaints. Muscle pain or muscle atrophy are not typical for the disease. Diagnosis The diagnosis of PC is suggested by the clinical picture described above and a positive family history. The EMG shows generalized spontaneous activity in the form of ®brillation-like potentials and myotonic discharges, often also at a normal muscle temperature. The serum CK may be elevated, up to 10 times above normal. The diagnosis can be veri®ed by cooling hand and forearm in a water bath at about 158C for 15±30 min. Cooling induces muscle stiffness and later weakness which can be veri®ed by a reduction of the amplitude of the evoked compound muscle action potential (Subramony et al., 1983; Gutmann et al., 1986; Jackson et al., 1994). A more precise measure of myotonia and weakness can be obtained determining the isometric force and relaxation time of the long ®nger ¯exor muscles before and after cooling (Ricker et al., 1986). Relaxation time can be prolonged from 0.5 s up to 50 s and contraction force reduced by more than 50%. A muscle biopsy is NOT necessary to diagnose PC. The diagnosis may be con®rmed by a mutation within the SCN4A gene. Therapy Antiarrhythmic drugs, such as mexiletine, are effective in preventing muscle stiffness and weakness induced by physical activity or exposure to cold (Ricker et al., 1980; Streib, 1987). The majority of PC patients, however, require no treatment

10

N. Mitrovic, H. Lerche

and know best how to deal with their symptoms. In severe cases 360 mg mexiletine (Mexitil Depot w), once to twice a day is recommended. A cardial check-up should precede the antimyotonic mexiletine therapy and serum concentration should be carefully checked because of the small therapeutic breadth. Some patients wish to abolish their symptoms to be able to participate in special events like swimming or winter sports. In these cases 360 mg mexiletine should be taken 1 h before the event. In paramyotonic HyperPP, the combined use of mexiletine and hydrochlorothiazide can prevent stiffness and weakness induced by cold, and the spontaneous attacks of HyperPP (Ricker et al., 1986). Genetics After electrophysiological measurements had revealed the TTX-sensitive sodium channel to be a strong candidate for the site of the defect in PC and HyperPP (see below, Lehmann-Horn et al., 1987a,b), a candidate gene approach demonstrated linkage of HyperPP to SCN4A on chromosome 17q23 (Fontaine et al., 1990). This was the ®rst evidence for the existence of a human sodium channel disease. Three groups showed independently that also PC is linked to the SCN4A locus (Ebers et al., 1991; Koch et al., 1991; Ptacek et al., 1991a). The ®rst mutation in an ion channel was found in HyperPP (Rojas et al., 1991). Up to date, about 20 point mutations have been detected in different parts of the alpha subunit of the human skeletal muscle sodium channel (Fig. 2). Ten of them lead to PC and most of these are located within the voltage sensor IV/S4 (PtaÂcek et al., 1992a; Wang et al., 1995). One mutation situated in the III±IV interlinker is supposed to be a part of the inactivation gate of the channel (T1313M; McClatchey et al., 1992). Pathogenesis In contrast to myotonia congenita Becker and Thomsen, electrophysiology on excised muscle specimens of PC patients revealed a normal chloride conductance. Instead, the speci®c abnormality found was a non-inactivating component of the sodium current (Lehmann-Horn et al., 1987a,b). Studies on PC muscle ®bres have shown that upon cooling the resting membrane potential depolarizes to values around ±40 mV. This depolarization can be prevented by TTX and the weakness of muscle bundles in vitro can be antagonized by potassium channel openers which hyperpolarize the muscle membrane (Lehmann-Horn et al., 1987b; Lerche et al., 1996b). These results proved that weakness is caused by membrane depolarization due to an increased sodium in¯ux through voltage-gated sodium channels. Electrophysiological recordings on heterologously expressed mutant sodium channels causing PC typically show a pronounced slowing of the current decay and a small persistent current due to incomplete inactivation (Fig. 5), a shift of

Sodium and calcium channelopathies of sarcolemma

11

Fig. 5. Impaired inactivation of mutant sodium channels. Sodium currents conducted by WT and mutant channels expressed in a heterologous system (human embryonic kidney cells). Whole cell currents were elicited by a membrane depolarization from a holding potential of 100±0 mV. The recording in (A) shows V1589M channels (Mitrovic et al., 1994) and in (B) R1448P channels (Mitrovic et al., 1999), in comparison with the wild type (WT). (A) Shows a persistent current as a typical ®nding for mutations causing HyperPP or PAM. (B) Shows a strongly slowed inactivation as an example for PC. Both, the persistent current and the slowing of inactivation are caused by more frequent reopenings of sodium channels as shown in single channel recordings in (C). Recordings from native muscle specimens are shown for a normal control (left panel) and for a PC patient carrying the R1448P mutation (right panel, Lerche et al., 1996a).

the steady-state inactivation curve in the hyperpolarizing direction and a faster recovery from inactivation (Chahine et al., 1994; Yang et al., 1994; Lerche et al., 1996b; Mitrovic et al., 1996; Mitrovic et al., 1999; Hayward et al., 1996). Cooling enhanced the slowing of inactivation and the persistent sodium current, however, this effect was not speci®c for mutant channels (Lerche et al., 1996b; Fleischhauer et al., 1998; Mitrovic et al., 1999). The electrophysiological ®ndings may explain the

12

N. Mitrovic, H. Lerche

clinical symptoms as follows: the slowed current decay provides a plausible explanation for paradoxical myotonia, since the abnormal sodium in¯ux mainly occurs during an action potential. Hence, only continued exercise results in suf®cient accumulation of intracellular sodium to cause muscle depolarization and myotonia. Weakness upon cooling is most probably explained by a combination of an increased persistent current, exceeding a certain threshold in the cold, and the left-shift of steady-state inactivation decreasing the number of sodium channels available for an action potential (Lerche et al., 1996b; Wagner et al., 1997; Mitrovic et al., 1999). The faster recovery from inactivation also contributes to the pathophysiology of PC promoting the development of myotonia by shortening the refractory period after an action potential (Hayward et al., 1996). An important question in the pathophysiology of dominantly inherited diseases, such as channelopathies, is the level of expression of the mutant protein. Using a quantitative kinetic analysis of Na 1 currents measured in muscle specimens biopsied from two PC patients, we were recently able to estimate the percentage of functional mutant protein in the native muscle membrane, a so far unknown parameter in the pathophysiology of channelopathies and other dominantly inherited diseases (Mitrovic et al., 1999). In contrast to mRNA measurements, western blots, antibody staining or other molecular biological or biochemical approaches, which determine the level of mutant RNA or protein without regard to its function, the electrophysiogical evaluation allowed us to determine the ratio of the mutant protein that is pathophysiologically relevant. Our analysis suggests that no more than 38% of the sodium channels in the PC muscle specimen were of the mutant type, which is signi®cantly less than the 50% suggested by the autosomal dominant mode of inheritance. Thus, the clinical severity of the phenotype seems to depend not only on the extent of the electrophysiological dysfunction of the mutant channels but also on their fraction in the muscle membrane. For another Na 1 channelopathy, the equine periodic paralysis in Quarter horses, the severity of the clinical phenotype has been shown to correspond to mRNA levels in native muscle probes (Zhou et al., 1994).

Potassium-aggravated myotonia This disease has been newly de®ned when long-known clinical knowledge could be combined with recent genetic and molecular biologic information (Lerche et al., 1993; Heine et al., 1993; PtaÂcek et al., 1992b, 1994a). Becker (1977) investigated more than 100 families with non-dystrophic dominant myotonia and proposed several subtypes of what he thought was myotonia congenita. Molecular biology revealed that many of these conditions were in fact caused by mutations in the gene encoding the muscle sodium channel. A few forms could be classi®ed as special

Sodium and calcium channelopathies of sarcolemma

13

types of paramyotonia, as they did show cold- and exercise-induced stiffness, albeit no cold-induced weakness. Other conditions, however, were too inconsistent with the de®nition of PC. Clinical features As a characteristic ®nding, PAM patients never experience muscle weakness and are not substantially sensitive to cold. Four clinical phenotypes can be distinguished with regard to the severity of myotonia and response to therapy. One group of affected persons experiences muscle stiffness that tends to ¯uctuate from day to day, hence the name `myotonia ¯uctuans' (Ricker et al., 1990, 1994; Lennox et al., 1992). Their muscle stiffness is provoked by exercise, and often it occurs with some delay during rest after heavy exercise. The stiffness may then last for 0.5±2 h. On many days or even for weeks, af¯icted persons experience no symptoms at all. More severely affected are persons with a generalized moderate myotonia, that may also show a kind of delayed warm-up phenomenon. The most severe form of PAM and of myotonia in general was called `myotonia permanens' (Lerche et al., 1993). It is characterized by very severe and persisting myotonia. When the myotonia is aggravated, e.g. by intake of potassium-rich food, ventilation might be impaired by stiffness of the thoracic muscles. In particular, children can suffer from acute hypoventilation and this may lead to cyanosis and unconsciousness, so that such episodes were occasionally mistaken for epileptic seizures. In spite of the misdiagnosis, antiepileptic medication, e.g. administration of carbamazepine, was useful in these cases because of its antimyotonic effects. Such patients would probably not survive without continuous treatment. The fourth related subtype is associated with acetazolamide-responsiveness of myotonia (Trudell et al., 1987), also described as atypical myotonia congenita (PtaÂcek et al., 1992b). In addition to stiffness, patients also report of muscle pain. Both the stiffness and pain are alleviated by acetazolamide. In PAM, depolarizing agents such as potassium or suxamethonium may aggravate the myotonia, but do not induce weakness. It is well known for myotonic disorders that the risk of depolarizing relaxants inducing anaesthesia-related events is increased. The incidence of such events seems to be highest in myotonia ¯uctuans families (Ricker et al., 1994; Vita et al., 1995). There seems to be no other biological reason for this other than the frequent absence of clinical myotonia in these patients making the anaesthesiologists unaware of the condition. Diagnosis. The diagnosis of PAM is suggested by a generalized myotonia with dominant inheritance, by the absence of weakness and cold sensitivity and possibly with some of the special features described above. However, in many cases PAM cannot be

14

N. Mitrovic, H. Lerche

clinically differentiated from myotonia congenita (MC) Thomsen. In this case oral potassium loading which induces a myotonic attack in PAM but not in MC Thomsen might be useful. This test is contraindicated in patient with myotonia permanens since it may provoke severe attacks. The EMG shows generalized myotonic discharges, in myotonia ¯uctuans patients often in the absence of clinical myotonia. Therapy As in PC, mexiletine is the drug of choice in preventing muscle stiffness. However, treatment is usually only necessary in severe cases (360 mg mexiletine (Mexitil Depot w) once to twice daily). Genetics and pathogenesis Six point mutations at four different positions are responsible for PAM. Four of the substitutions are located in the supposed inactivation gate (Fig. 2). Three of them affect the same nucleotide, resulting in three different amino acid substitutes for one (G1306) of a pair of glycines (1306/7) supposed to be essential for proper inactivation. The more the substitutes differ from glycine by having side-chains of variable length and charge and/or rami®cation, the greater is the degree of the electrophysiological defect in vitro and the more severe are the clinical symptoms (McClatchey et al., 1992; Lerche et al., 1993; Mitrovic et al., 1995). Glutamic acid, having a long side-chain, causes myotonia permanens, the most severe form of myotonia known. Valine, an amino acid with a side-chain of intermediate size, is the substitute in patients with moderate myotonia and alanine, distinguished by a short side-chain, results in the benign myotonia ¯uctuans. Electrophysiological experiments with some of the PAM-mutants expressed in human embryonic kidney cells have shown slowed inactivation, increased persistent current (Fig. 5) and a shift of the steady-state inactivation curve in the depolarizing direction (Mitrovic et al., 1994, 1995; Hayward et al., 1996). As mentioned above (see PC), both a slowed inactivation and an increased persistent sodium current cause membrane depolarization and muscle membrane hyperexcitability. The right shift of the steady-state inactivation curve extends the availability of sodium channels at more positive potentials and may, in contrast to PC-causing mutants which show a left shift of the curve, prevent the development of paralysis. Mutations at the position G1306, additionally show slowed deactivation (Mitrovic et al., 1995; Hayward et al., 1996). This defect could increase the in¯ux of sodium ions during repolarization which might also contribute to myotonia. However, compared to the defects of fast inactivation, the contribution of impaired deactivation to clinical myotonia should be rather small. Deactivation defects were also described for two other mutations causing either PAM or PC (Richmond et al., 1997; Featherstone et al., 1998), however for the PC-causing

Sodium and calcium channelopathies of sarcolemma

15

mutation (R1448P), deactivation is not impaired in the voltage range where no inactivation occurs (Mitrovic et al., 1999). Hyperkalemic periodic paralysis The disease was ®rst described by Tyler et al. (1951); Helweg-Larsen et al. (1955) and was extensively investigated by Gamstorp (1956) who clearly differentiated it from `paroxysmal familial paralysis' and named it `adynamia episodica hereditaria'. Clinically, the most striking difference of the two diseases is that, during the paralytic episodes, serum potassium decreases in the former and increases in the latter. To stress this distinction, the names hypokalemic periodic paralysis and hyperkalemic periodic paralysis, respectively, are now preferred. Hyperkalemic periodic paralysis is transmitted as an autosomal dominant trait with complete penetrance, although incomplete penetrance was reported for families with rare mutations (McClatchey et al., 1992; Wagner et al., 1997). The disease has three clinically distinct variants. It can occur (i) without myotonia, (ii) with clinical or electromyographic myotonia, or seldom (iii) with paramyotonia. In some patients, a chronic progressive myopathy may develop which seems to be genetically determined (mutation T704M) (Bradley et al., 1989; PtaÂcek et al., 1991b; Lehmann-Horn et al., 1993). Clinical features The attacks usually begin in the ®rst decade of life. Initially they are rare but then increase in frequency and, in severe cases, may recur daily. The attack commonly starts in the morning before breakfast and lasts 15 min to an hour, and then spontaneously disappears. Often rest provokes the attack, and prior strenuous work usually aggravates it. Potassium loading, cold environment, emotional stress, glucocorticoids, and pregnancy provoke or worsen the attacks. After strenuous exercise, weakness can follow within a few minutes of rest. Sustained mild exercise after a period of strenuous exercise may postpone or prevent the weakness in the exercising muscle groups while the resting muscles become weak. The generalized weakness is usually accompanied by a signi®cant increase of serum potassium (more than 5.5 mM). The course of the paralytic attacks is the same in all three forms of HyperPP. Cooling can induce weakness, but not stiffness, and reheating restores contractile force quickly (except for the paramyotonic form). EMG studies are required to determine the presence or absence of myotonia which is usually very mild. In the non-myotonic form, clinical and electrical myotonia are both absent. Cooling may provoke weakness but does not cause substantial myotonia. Paramyotonic HyperPP is characterized by attacks of generalized muscle weakness associated with hyper-

16

N. Mitrovic, H. Lerche

kalemia and by paradoxical myotonia (for details see previous section on paramyotonia congenita). Normokalemic periodic paralysis, a variant of the hyperkalemic form This rare disorder resembles HyperPP in many respects but differs from it in that the serum potassium does not increase even during serious attacks. The existence of normokalemic periodic paralysis as a nosological entity has been questioned because some patients with this condition are sensitive to oral potassium salts (Poskanzer and Kerr, 1961). The disorder is transmitted as an autosomal dominant trait with high penetrance in both sexes. The attacks begin in the ®rst decade of life and are provoked or worsened by rest after exercise, exposure to cold and by potassium loading. A urinary potassium retention, the lack of a bene®cial effect of glucose, and failure of the serum potassium to increase in attacks are differences to primary HyperPP. However, in at least one such family, the condition is caused by the common T704M mutation in SCN4A normally associated with HyperPP (F. Lehmann-Horn, pers. commun.). Diagnosis The diagnosis of HyperPP is based on the presence of typical attacks of weakness or paralysis combined with an increased serum potassium during the attacks, the positive family history, and the myotonic or paramyotonic phenomena, if present. Except for some older patients with progressive myopathy, the muscles are well developed. The serum CK is sometimes elevated up to 200±300 U/l. When the diagnosis is unclear, a provocative test can be performed. An elegant test consists of exercise on a bicycle ergometer for 30 min so that the pulse increases to 120±160 beats/min followed by absolute rest in bed (Ricker et al., 1989). It should be preferably performed in the morning in the fasting state. The serum potassium rises during exercise and then declines to almost the pre-exercise level, as in healthy individuals. Ten to 20 min after the onset of rest, a second hyperkalemic period occurs in the patients in contrast to normal subjects, and during this period the patients become paralyzed. If a paralytic attack is not induced, the test can be combined with the administration of 40±80 mmol oral potassium chloride. This test should be performed in anesthesiologic stand-by and is contraindicated in subjects already hyperkalemic and in those who do not have adequate renal or adrenal reserve. Recordings of the evoked compound muscle action potential during rest and exercise are also helpful in con®rming the diagnosis of periodic paralysis (McManis et al., 1986). An abnormally high serum potassium level between attacks suggests secondary rather than primary HyperPP.

Sodium and calcium channelopathies of sarcolemma

17

Therapy Preventive therapy consists of frequent meals rich in carbohydrates, a low-potassium diet, and avoidance of fasting, strenuous work, and exposure to cold. Many patients are able to prevent or abort attacks by continuing slight exercise and/or by oral ingestion of carbohydrates at the onset of weakness (e.g. 2 g glucose per kg body weight). However, severe attacks may fail to respond to these measures (Gamstorp, 1956). Interestingly, attacks occur more frequently on holidays and weekends when patients rest in bed longer than usual. Thus, patients are advised to rise early and have a full breakfast. Some patients can abort or attenuate attacks by the prompt oral intake of a thiazide diuretic or acetazolamide, or by inhalation of a b-adrenergic agent. The bene®cial effect of the diuretics is probably due to their capacity to lower the serum potassium level. The effects of the b-adrenergic agents is probably mediated via stimulation of the sodium±potassium pump (Clausen, 1986). Calcium gluconate, 0.5±2 g given intravenously, has also terminated attacks in some patients. It is often advisable to prevent attacks by the continuous use of a thiazide diuretic (Gamstorp, 1956) or acetazolamide (McArdle, 1962; Riggs and Griggs, 1979). Diuretics in modest dosages at intervals from twice daily to twice weekly are very effective in mild cases. The drug should not lower the serum potassium below 3.3 mM or the serum sodium below 135 mM (McArdle, 1962). In severe cases, 50 or 75 mg of hydrochlorothiazide should be taken daily early in the morning. For the dosage of acetazolamide see HypoPP therapy. Genetics Four mutations in SCN4A were found to cause HyperPP (Fig. 2). T704M is the most frequent SCN4A mutation and, in addition to HyperPP with or without myotonia, often causes chronic progressive myopathy (PtaÂcek et al., 1991b). All other mutations cause myotonic HyperPP without permanent weakness. M1592V, the ®rst of all detected sodium channel mutations (Rojas et al., 1991), always causes HyperPP associated with myotonia. Two rare mutations (A1156T, M1360V) are of interest since they were discovered in families with overlap syndromes of HyperPP and PC showing incomplete penetrance in females (McClatchey et al., 1992; Wagner et al., 1997; F. Lehmann-Horn, pers. commun.). Pathogenesis In vitro electrophysiological studies on muscle specimens from patients with HyperPP revealed a large persistent sodium current due to incomplete inactivation of the sarcolemmal sodium channels depolarizing the muscle membrane to values of about 240 mV (Lehmann-Horn et al., 1987a,b, 1991; Cannon et al., 1991). At this

18

N. Mitrovic, H. Lerche

membrane potential all sodium channels are inactivated and no action potential can be elicited which causes paralysis. Clinically, attacks are provoked by potassium intake or by physical activity inducing a slight membrane depolarization, which in HyperPP muscle cannot be readily equilibrated. Interestingly, patients never become weak during activity despite an increase of the extracellular potassium level (Gamstorp, 1962). This could be connected to the work-related decrease in intracellular pH because lowering the pH of the highpotassium bathing solution normalizes the contractile force exerted by muscle bundles obtained from HyperPP patients (Lehmann-Horn et al., 1987a). Also, physical activity is associated with enhanced adrenalin release. Adrenalin stimulates the sodium±potassium pump (Clausen, 1986) which, in turn, helps to compensate for the abnormal sodium in¯ux into the muscle ®bres. Studies on heterologously expressed mutant channels revealed that an incomplete fast inactivation causes the persistent sodium current. It was more pronounced in HyperPP-causing mutations (~6% of the peak current,Cannon and Strittmatter, 1993) compared to mutations causing PAM (~1±3%, Mitrovic et al., 1994, 1995). Therefore it was proposed that a small persistent current observed with PAM mutants induces a slight membrane depolarization and hyperexcitability whereas HyperPP mutants cause a large persistent current with sustained depolarization and paralysis. Ruff (1994) proposed that a defect of fast sodium channel inactivation is not suf®cient to explain long periods of paralysis as observed in HyperPP patients. Up to Ruff, an intact slow inactivation should terminate the in¯ux of sodium ions into muscle ®bres through channels with defective fast inactivation within several minutes. Recent studies (Cummins and Sigworth, 1996; Hayward et al., 1997) showed that indeed HyperPP-causing mutations additionally disturb slow sodium channel inactivation to a signi®cant extent in contrast to many other sodium channel mutations causing PAM or PC. This should substantially increase the permanent sodium in¯ux into the muscle ®bres. Contrary to what one could have expected, in the heterologous expression system extracellular potassium had no direct effect on any of the mutant channels investigated (Cannon and Strittmatter, 1993; Cummins et al., 1993; Mitrovic et al., 1995; Wagner et al., 1997). Therefore, this triggering factor seem to exert its effect indirectly, i.e. via membrane depolarization.

Electrophysiological differences among the sodium channelopathies and further aspects The electrophysiological properties found for the mutations examined so far causing either PAM, PC or HyperPP are summarized in Table 1. Mutations causing PAM are characterized by an increased persistent current and/or slowing of fast

Sodium and calcium channelopathies of sarcolemma

19

Table 1 Summary of the electrophysiological properties of PAM, PC and HyperPP sodium channel mutants a

Disease

Slowing of fast inactivation

PC PAM HyperPP

1 1 1 2

Persistent current 1 1/1 1 1 1 1

Steady-state fast inactivation

Recovery from fast inactivation

à ! 2

1 1 2/1 2

Impaired slow inactivation 2 2 1

a 2 and 1 indicate the severity of the gating defect. The arrows show the direction of the steady-state fast inactivation curve shift. For details see text.

inactivation causing a moderate sodium inward current that explains slight depolarization and myotonia. A right shift of steady-state inactivation might contribute to prevent weakness. HyperPP mutants show a large persistent current and an incomplete slow inactivation inducing a large sodium inward current and a strong depolarization that cause paralysis. PC mutants are characterized by a strong slowing of fast inactivation providing an explanation for paradoxical myotonia, an acceleration of recovery from inactivation and a left shift of steady-state inactivation which in combination with an increased persistent current could explain the cold-induced weakness. The pathophysiological concept of sodium channel diseases was con®rmed by experiments introducing a sodium channel inactivation defect into healthy muscle ®bres by using anemone toxin II. The toxin caused myotonic runs in rat muscle ®bres by inducing a persistent sodium current of 2% of peak current (Cannon and Corey, 1993). Furthermore, a computer model simulating several of the experimentally found mechanisms of inactivation failure was able to reproduce both myotonia and paralysis, depending on the degree of the sodium channel inactivation defect (Cannon et al., 1993; Hayward et al., 1996, 1997). Patch-clamp studies on native sodium channel mutants expressed in human cell lines revealed not only detailed mechanisms of the pathophysiology of these diseases but also pointed to speci®c parts of the protein important for channel function. Studies on PC-causing mutants in segment S4 of domain IV (Chahine et al., 1994) initiated a series of experiments which led to the identi®cation of the voltage sensor. The mutation R1448C was used to study the accessibility of this channel region and proved that the voltage sensor moves outward relative to surrounding parts of the protein (Yang and Horn, 1995). Individual substitution of all arginines in IV/S4 by cysteines revealed that a part of the supposed S4-helix moves completely across the membrane through a yet unknown region called the `S4 channel' (Yang et al., 1996). Further experiments showed that only the voltage sensors of domain III and IV are important for channel inactivation, whereas those of

20

N. Mitrovic, H. Lerche

domain I and II for channel activation (Chen et al., 1996; Kontis et al., 1997; Mitrovic et al., 1998; Cha et al., 1999). Studies on the PAM-causing mutations G1306A/V/E extended the hinged-lid model for channel inactivation proposed by West et al. (1992), suggesting that the two glycines at position 1306/7 form a hinge of the lid (see Fig. 4; Lerche et al., 1993; Mitrovic et al., 1995; Hayward et al., 1996). Finally, all mutations at the cytoplasmic surface of the channel protein may be directly or indirectly involved in the formation of a receptor site for the IFM motif. Extended studies of mutations in the S4-S5 loop of domain IV were stimulated partly by disease-causing mutations in IV/S4 and IV/S4-S5. Although IV/S4S5 is probably not a receptor site for IFM, these studies revealed an important role of this channel region in fast inactivation and pointed to its secondary structure which is an alpha-helix (Mitrovic et al., 1996; Lerche et al., 1997; Filatov et al, 1998; McPhee et al., 1998; Tang et al., 1998).

Hypokalemic periodic paralysis There are two types of calcium channels expressed in skeletal muscle, the socalled dihydropyridine receptor (DHPR) and the ryanodine receptor (RYR1). Both are closely related in the triadic junctions of the transverse (T) tubular system and the sarcoplasmic reticulum (SR, see Fig. 6). The DHPR, located in the T-tubular membrane, is an L-type voltage-dependent calcium channel. RYR1, located in the SR membrane, is itself not voltage-dependent, but coupled to the DHPR. In contrast to the cardiac muscle, the voltage-gated L-type calcium channel of skeletal muscle appears to be physiologically unimportant as an ion-conducting channel. However, it functions as a voltage-sensor of the ryanodine receptor (RYR1) which releases calcium from the SR initiating contraction (Melzer et al., 1995). Disease-causing mutations are known in the genes for either channel. Certain mutations in CACNA1S, encoding the DHPR, cause HypoPP. Other mutations in the same gene, as well as mutations in the RYR1 gene cause malignant hyperthermia. Although the clinical phenotype of familial hypokalemic periodic paralysis (HypoPP) was described almost 200 years ago it was not recognized until 1934 that a decrease of serum potassium is associated with paralytic attacks (Biemond and Daniels, 1934). An early clinical summary about the disease came out in 1941 (Talbott, 1941). HypoPP is a disease that still baf¯es clinicians and basic scientists although the genetic cause of this dominantly inherited disease has been found. Although it is the most common form of the periodic paralyses in man, it is still a rare disease showing a prevalence of only about 1:100 000. The major symptoms of HypoPP, i.e. episodes of generalized paralysis are on average of longer duration than in HyperPP, however, the differential diagnosis between both syndromes is often dif®cult. The

Sodium and calcium channelopathies of sarcolemma

21

Fig. 6. The triadic junction between a transverse tubule and the sarcoplasmic reticulum. Two calcium channels of skeletal muscle, the dihydropyridine (DHP) receptor and the ryanodine receptor, are responsible for the so-called excitation±contraction coupling in skeletal muscle. The coupling mechanism between the two channels is not yet fully elucidated. Mutations in the respective genes may cause hypokalemic periodic paralysis, malignant hyperthermia or central core disease.

condition is transmitted as an autosomal dominant trait with reduced penetrance in women (the male to female ratio is 3±4:1,Cerny and Katzenstein-Sutro, 1952). Decisive for classi®cation is the level of serum potassium during a paralytic attack, which may fall below 2 mmol/l in HypoPP. Beside clear familial forms, also sporadic cases occur and are more frequent in men than in women. Occasionally a carrier is asymptomatic and the disease appears to skip a generation. Clinical features The main symptom of HypoPP is the episodic occurrence of attacks of ¯accid weakness. The attacks can be provoked by excessive intake of carbohydrates, stren-

22

N. Mitrovic, H. Lerche

uous exercise or by mental stress. Interestingly they do not occur during the physical stress but during a following rest, sometimes several hours later. Very common are paralytic attacks in the second half of the night or the early hours of the morning, so that the awakening patient is unable to move his arms, legs or trunk. Injection of antiphlogistics and/or local anaesthetics can trigger a severe attack after a few hours. Slight physical activity can sometimes prevent or delay mild attacks. The attacks vary in frequency, duration and severity. In severe attacks, the vital capacity may be reduced and death can occur from ventilatory failure or cardiac arrhythmia due to hypokalemia. However, usually patients show a normal life span. Independent of the occurrence of attacks, many patients develop a late onset, progressive myopathy (Links et al., 1990; Lehmann-Horn et al., 1994). Severe cases present in early childhood, mild cases as late as the third decade of life, and about 60% present before age 16 (Talbott, 1941). Initially the attacks are infrequent but after a few months or years they often increase in frequency and eventually recur daily. During major attacks, the serum potassium decreases, though not always below the normal range. Sinus bradycardia and ECG signs of hypokalemia may appear. Clinical or histopathologic signs of cardiomyopathy are absent (Links et al., 1990). Diagnosis The diagnosis of familial hypokalemic periodic paralysis is suggested by a decrease of the serum potassium level during a major attack and by a positive family history. Low serum potassium levels between attacks suggest secondary periodic paralysis. In these cases a search for renal or gastrointestinal potassium wastage is necessary. Thyrotoxic periodic paralysis resembles the familial form with respect to changes in serum and urinary electrolytes during attacks and its response to glucose, insulin, potassium, and rest after exertion. The attacks cease when the euthyroid state is restored. The serum CK may be slightly elevated but is usually normal. Electrical myotonia excludes the diagnosis of HypoPP. In the absence of myotonia, patients may still have non-myotonic HyperPP. Patients with permanent weakness often show myopathic changes and sometimes ®brillation potentials. During a severe attack the evoked compound muscle action potential is either abnormally small or absent. When the serum potassium of a patient cannot be investigated during a spontaneous attack, further tests are required to establish the diagnosis of periodic paralysis and to determine its type. The systemic provocative tests carry the risk of inducing a severe attack. Therefore they must be performed by an experienced physician, and the serum potassium and glucose levels and the ECG must be closely monitored. Provocative tests with glucose with or without the additional use of insulin must

Sodium and calcium channelopathies of sarcolemma

23

never be done in patients who are already hypokalemic and potassium chloride must not be given to patients unless they have adequate renal and adrenal function. The simplest systemic provocative test exploits the physiological potency of glucose, or of glucose plus insulin, to cause hypokalemia. The oral administration of glucose, 2 g/kg body weight, in the early morning combined with 10±20 international units of insulin, given subcutaneously, may provoke a paralytic attack within 2±3 h. Exercise and intake of carbohydrates the evening before increases the potency of the test. If the test is equivocal, intravenous administration of 1.5±3 g glucose per kg body weight over 60 min in combination with subcutaneous insulin may provoke an attack. In general, a serum potassium level of 3.0 mM or less should be achieved. A negative test does not exclude the diagnosis of primary hypokalemic periodic paralysis. Therapy and preventive measures Mild paralytic attacks need no treatment. Attacks of generalized paralysis should be treated by oral intake of 2±10 g potassium chloride. In most cases this causes muscle strength to recover considerably within 0.5±1 h, especially when the patient uses every opportunity for physical activity as strength returns. The dose may be repeated after 3±4 h. Some patients like to take potassium at the beginning of an attack. However, increasing potassium doses should be avoided since sometimes a potassium `dependency' can occur, which makes the disease more dif®cult to control. Basic recommendations for prevention are to avoid the ingestion of carbohydraterich meals and to avoid strenuous exertion. The preventive medication of choice is acetazolamide (Griggs et al., 1970). The dosage should be as low as possible, starting with 125 mg every other day. If the paralytic attacks continue, the dose can be increased up to a maximum of 250 mg twice daily. Adverse reactions to the drug include paresthesia, anorexia, transient myopia and an increased incidence of nephrolithiasis. Few patients have developed renal failure during protracted acetazolamide therapy. Interestingly, the medication is also effective in preventing attacks in HyperPP, but the mechanism of action for both syndromes is unclear. Patients refractory to acetazolamide may respond favourably to dichlorophenamide, another carbonic anhydrase inhibitor, at doses of 25 mg three times daily (Dalakas and Engel, 1983). Furthermore, the aldosterone antagonist spironolactone is effective by increasing the serum potassium level in dosages of 50±150 mg daily. For more therapeutic options in severe cases see Lehmann-Horn et al. (1994). Molecular genetics A systematic genome-wide search in three families (Fontaine et al., 1994) demonstrated that the disease is linked to chromosome 1q31-32 and co-segregates with CACLN1A3, the gene encoding the L-type calcium channel (DHP receptor) a 1-

24

N. Mitrovic, H. Lerche

subunit which is mapped to this region (Drouet et al., 1993; Gregg et al., 1993). Thus far, only for one family linkage of the disease to this locus was excluded (Plassart et al., 1994), suggesting genetic heterogeneity. The dihydropyridine (DHP) receptor/ calcium channel complex is located in the transverse tubular system and consists of ®ve subunits: a1, a2/d, b, and g (Catterall, 1995). The a1 subunit (Fig. 7) constitutes both the gating and permeation machinery of the channel and contains the receptor sites for dihydropyridines and phenylalkylamines. Three mutations within the CACLN1A3 gene have been identi®ed. Two of these are analogous predicting arginine to histidine substitutions within the highlyconserved S4 regions of repeats II and IV (R528H and R1239H, respectively), the third predicts an arginine to glycine substitution in IV-S4 (R1239G) (Jurkat-Rott et al., 1994; PtaÂcek et al., 1994b, Fig. 7). The majority of families carry either the R528H or the R1239H substitution (Elbaz et al., 1995; Grosson et al., 1996). Pathogenesis Electrophysiological investigations on biopsied muscle specimens from HypoPP patients showed that the paralysis induced by hypokalemia in vitro is due to a sustained depolarization of the sarcolemma to about 240 mV (RuÈdel et al., 1984). The crucial role of the depolarization for the pathogenesis of paralysis was

Fig. 7. Mutations in the a1-subunit of the skeletal muscle L-type calcium channel (dihydropyridine receptor, DHPR). Three mutations at the two positions indicated in the voltage sensors II/S4 and IV/S4 of the DHPR have been found so far to cause HypoPP (R528H, R1239H/G). The whole channel complex is built of ®ve subunits. The alpha1 subunit contains the gating machinery, channel pore and drug binding sites (for details see text, modi®ed after Lehmann-Horn and RuÈdel, 1996).

Sodium and calcium channelopathies of sarcolemma

25

con®rmed by experiments showing that the hyperpolarizing effect of ATP-sensitive potassium channel openers was able to prevent muscle weakness or restore normal force (Grafe et al., 1990). Although the genetic defect has now been elucidated, the pathophysiological mechanism of the membrane depolarization is still unclear. L-type calcium currents conducted by R528H and R1239H mutant channels were studied in various native and heterologous expression systems (Sipos et al., 1995; Lapie et al., 1996; Lerche et al., 1996a; Jurkat-Rott et al., 1998, Morrill et al., 1998). In summary, they did not reveal a signi®cant alteration of neither the gating of the mutant channels nor of the resulting calcium transients that could explain the occurrence of membrane depolarization following a decrease in serum potassium. An initially reported hyperpolarizing shift of steady-state inactivation for the R528H mutation (Sipos et al., 1995) could not be reproduced. Some groups found a reduction in current amplitude (Sipos et al., 1995; Lapie et al., 1996; Morrill et al., 1998), and one recent paper reported a slight slowing of the time course of activation (Morrill et al., 1998). Recently, a reduced potassium current conducted by ATP-dependent channels (KATP) was reported in biopsied muscle specimens from three HypoPP patients carrying the R528H mutation of the dihydropyridine receptor (Tricarico et al., 1999). The authors proposed a contribution of this ®nding to the pathogenesis of HypoPP.

Open questions The pathophysiology of the sodium channelopathies has been well elucidated by studying the effects of disease-causing mutations in molecular detail. Different mechanisms of channel inactivation failure cause either a small and short-lasting or a large and sustained sodium inward current into the muscle ®bres leading to the clinical symptoms of myotonia and paralysis. Not entirely clear is why patients with PC are temperature-sensitive and those with PAM and HyperPP are not, since a speci®c temperature dependence could not be found for any of the paramyotoniacausing mutations. On the other hand, the cold-induced weakness is clearly linked to a sodium current-induced membrane depolarization, so that a different mechanism ± other than via the sodium channel ± seems unlikely. Another interesting question is the aggravation of the clinical symptoms upon potassium intake with PAM and HyperPP patients. Initial electrophysiological experiments which showed a direct effect of potassium on mutant channels causing HypePP (Cannon et al., 1991) could not be reproduced (Cannon and Strittmatter, 1993; Wagner et al., 1997). Studies on PAM- or PC-causing mutations also showed no sensitivity to extracellular potassium (Chahine et al., 1994; Mitrovic 1994, 1995). Therefore, the effect of potassium is most likely explained by a membrane depolarization. An additional problem of

26

N. Mitrovic, H. Lerche

the sodium channelopathies which has not been elucidated is the occurrence of a myopathy with the HyperPP-causing mutation T704M. In contrast to the sodium channelopathies, the pathophysiological mechanism linking a calcium channel mutation to membrane depolarization and paralysis in HypoPP is entirely unresolved. Here, it seems likely that other structures are involved interacting somehow with the dihydropyridine receptor and/or reacting to hypokalemia. Note added in proof Very recently, a point mutation, Arg-669-His, in the voltage sensor of domain 2 of the sodium channel gene SCN4A was reported in a single HypoPP family with four affected individuals (Bulman et al., 1999). To test SCN4A as causative for the disease in several typical HypoPP families for which DHPR gene defects were excluded, genetic linkage studies and mutation screening were performed (JurkatRott et al., 2000). Tight linkage and two novel mutations, Arg-672-His and -Gly, were identi®ed. Heterologous expression revealed a 10 mV left-shift of the steadystate fast inactivation curve and a sodium current density that was reduced suggesting reduced function as the pathogenetic mechanism. This entity was named HypoPP-2. Acknowledgements The authors thank Drs Frank Lehmann-Horn and Reinhardt RuÈdel for reading the manuscript and for their support. References Armstrong, C.M., Bezanilla, F., 1977. Inactivation of the sodium channel: II. Gating current experiments. J. Gen. Physiol. 70, 567±590. Becker, P.E., 1977. Myotonia Congenita and Syndromes Associated with Myotonia. Georg Thieme, Stuttgart. Becker, P.E., 1970. Paramyotonia congenita (Eulenburg). Fortschritte der allgemeinen und klinischen Humangenetik. Georg Thieme, Stuttgart. Biemond, A., Daniels, A.P., 1934. Familial periodic paralysis and its transition into spinal muscular atrophy. Brain 57, 91±108. Bradley, W., Taylor, R., Rice, D., Hausmanowa-Petrusewicz, I., Adelman, L., Jenkison, M., Jedrzejowska, H., Drac, H., Pendlebury, W., 1989. Progressive myopathy in hyperkalemic periodic paralysis. Arch. Neurol. 47, 1013±1017. Bulman, D.E., Scoggan, K.A., van Oene, M.D., Nicolle, M.W., Hahn, A.F., Tollar, L.L., Ebers, G.C., 1999. A novel sodium channel mutation in a family with hypokalemic periodic paralysis. Neurology 53, 1932±1936.

Sodium and calcium channelopathies of sarcolemma

27

Cannon, S.C., Brown Jr., R.H., Corey, D.P., 1991. A sodium channel defect in hyperkalemic periodic paralysis: potassium-induced failure of inactivation. Neuron 6, 619±626. Cannon, S.C., Brown Jr., R.H., Corey, D.P., 1993. Theoretical reconstruction of myotonia and paralysis caused by incomplete inactivation of sodium channels. Biophys. J. 65, 270±288. Cannon, S.C., Corey, D.P., 1993. Loss of Na + channel inactivation by anemone toxin (ATX II) mimics the myotonic state in hyperkalaemic periodic paralysis. J. Physiol. (Lond.) 466, 501±20. Cannon, S.C., Strittmatter, S.M., 1993. Functional expression of sodium channel mutations identi®ed in families with periodic paralysis. Neuron 10, 317±326. Catterall, W.A., 1995. Structure and function of voltage-gated ion channels. Ann. Rev. Biochem. 64, 493± 531. Cerny, A., Katzenstein-Sutro, E., 1952. Die paroxysmale LaÈhmung. Schweizer Archiv fuÈr Neurologie und Psychiatrie 70, 259±338. Cha, A., Ruben, P.C., George Jr., A.L., Fujimoto, E., Bezanilla, F., 1999. Voltage sensors in domains III and IV, but not I and II, are immobilized by Na + channel fast inactivation. Neuron 22, 73±87. Chahine, M., George, A.L., Zhou, M., Ji, S., Sun, W., Barchi, R.L,. Horn, R., 1994. Sodium channel mutations in paramyotonia congenita uncouple inactivation from activation. Neuron 12, 281±294. Chen, L.Q., Santarelli, V., Horn, R., Kallen, R.G., 1996. A unique role for the S4 segment of domain 4 in the inactivation of sodium channels. J. Gen. Physiol. 108, 549±556. Clausen, T., 1986. Regulation of active Na +±K + transport in skeletal muscle. Physiol. Rev. 66, 542±580. Cummins, T.R., Zhou, J., Sigworth, F.J., Ukomadu, C., Stephan. M., Ptacek, L.J., Agnew, W.S., 1993. Functional consequences of a Na + channel mutation causing hyperkalemic periodic paralysis. Neuron 10, 667±678. Cummins, T.R., Sigworth, F.J., 1996. Impaired slow inactivation of mutant sodium channels. Biophys. J. 71, 227±236. Dalakas, M.C., Engel, W.K., 1983. Treatment of 'permanent' muscle weakness in familial hypokalemic periodic paralysis. Muscle Nerve 6, 82±186. Drouet, B., Garcia, L., Simon-Chazottes, D., Mattei, M.G., GueÂnet, J.-L., Schwartz, A., Varadi. G., PincËon-Raymond, M., 1993. The gene encoding for the a1 subunit of the skeletal dihydropyridine receptor (Cchlla3=mdg) maps to mouse chromosome 1 and human 1q32. Mamm. Genome 4, 499± 503. Ebers, G.C., George Jr., A.L., Barchi, R.L., Ting-Passador, M.S., Kallen, R.G., Lathrop, G.M., Beckmann, J.S., Hahn, A.F., Brown, W.F., Campbell, R.D., Hudson, A.J., 1991. Paramyotonia congenita and hyperkalemic periodic paralysis are linked to the adult muscle sodium channel gene. Ann. Neurol. 30, 810±816. Elbaz, A., Vale-Santos, J., Jurkat-Rott, K., Lapie, P., Ophoff, R.A., Bady, B., Links, T.P., Puissan, C., Villa, A., Monnier, N., Padberg, G.W., Abe, K., Feingold, N., Guimaraes, J., Wintzen, A.R., Van der Hoeven, J.H., Saudubray, J.M., Grundfeld, J.P., Lenoir, G., Nivet, H., Echenne, B., Frants, R.R., Fardeu, M., Lehmann-Horn, F., Fontaine, B., 1995. Hypokalemic periodic paralysis (hypoPP) and the dihydropyridine receptor (CACNL1A3): genotype/phenotype correlations for two predominant mutations and evidence for the absence of a founder effect in 16 Caucasian families. Am. J. Hum. Genet. 56, 374±380. È ber eine familiaÈre durch 6 Generationen verfolgbare Form congenitaler ParamyoEulenburg, A., 1886. U tonie. Neurologisches Zentralblatt 5, 265±272. Featherstone, D.E., Fujimoto, E., Ruben, P.C., 1998. A defect in skeletal muscle sodium channel deactivation exacerbates hyperexcitability in human paramyotonia congenita. J. Physiol. (Lond.) 506, 627±38. Fleischhauer, F., Mitrovic, N., Deymeer, F., Lehmann-Horn, F., Lerche, H., 1998. Effects of mexiletine and temperature on the F1473S Na + channel mutation causing paramyotonia congenita. P¯uÈgers Arch. Eur. J. Physiol. 436, 757±765.

28

N. Mitrovic, H. Lerche

Filatov, G.N., Nguyen, T.P., Kraner, S.D., Barchi, R.L., 1998. Inactivation and secondary structure in the D4/S4-5 region of the SkM1 sodium channel. J. Gen. Physiol. 111, 703±15. Fontaine, B., Khurana., T.S., Hoffman, E.P., Bruns, G.A.P., Haines, J.L., Trofatter, J., Hanson, M.P., Rich, J., McFarlane, H., McKenna-Yasek, D., Romano, D., Gusella, J.F., Brown Jr., R.H., 1990. Hyperkalemic periodic paralysis and the adult muscle sodium channel a-subunit gene. Science 250, 1000± 1003. Fontaine, B., Vale Santos, J.M., Jurkat-Rott, K., Reboul, J., Plassart, E., Rime, C.S., Elbaz, A., Heine, R., Guimaraes, J., Weissenbach, J., Baumann, N., Fardeau, M., Lehmann-Horn, F., 1994. Mapping of hypokalemic periodic paralysis (HypoPP) to chromosome 1q31-q32 by a genomewide search in three European families. Nature Genet. 6, 267±272. Gamstorp, I., 1956. Adynamia episodica hereditaria. Acta Paediatr. (Uppsala), Suppl. 108. Gamstorp, I., 1962. A study of transient muscular weakness. Acta Neurol. Scand. 38, 3±19. George Jr., A.L., Iyer, G.S., Kleinfeld, R., Kallen, R.G., Barchi, R.L., 1993. Genomic organization of the human skeletal muscle sodium channel gene. Genomics 15, 598±606. Grafe, P., Quasthoff, S., Strupp, M., Lehmann-Horn, F., 1990. Enhancement of K + conductance improves in vitro the contraction force of skeletal muscle in hypokalemic periodic paralysis. Muscle Nerve 13, 451±457. Gregg, R.G., Couch, F., Hogan, K., Powers, P.A., 1993. Assignment of the human gene for the a1-subunit of the skeletal muscle DHP-sensitive calcium channel (CACNL1A3) to chromosome 1q31-32. Genomics 15, 107±112. Griggs, R.C., Engel, W.K., Resnick, J.S., 1970. Acetazolamide treatment of hypokalemic periodic paralysis. Ann. Intern. Med. 73, 39±48. Grosson, C.L.S., Esteban, J., McKenna-Yasek, D., Gusella, J.F., Brown Jr., R.H., 1996. Hypokalemic periodic paralysis mutations: Con®rmation of mutation and analysis of founder effect. Neuromusc. Disord. 6, 27±31. Gutmann, L., Riggs, J., Brick, J., 1986. Exercise-induced membrane failure in paramyotonia congenita. Neurology 36, 130±132. Hayward, L.J., Brown Jr., R.H., Cannon, S.C., 1996. Inactivation defects caused by myotonia±associated mutations in the sodium channel III±IV linker. J. Gen. Physiol. 107:559±756. Hayward, L.J., Brown, R.H., Cannon, S.C., 1997. Slow inactivation differs among mutant Na channels associated with myotonia and periodic paralysis. Biophys. J. 72, 1204±1219. Heine, R., Pika, U., Lehmann-Horn, F., 1993. A novel SCN4A mutation causing myotonia aggravated by cold and potassium. Hum. Mol. Genet. 2, 1349±1353. Helweg-Larsen, H.F., Hauge, M., Sagild, U., 1955. Hereditary transient muscular paralysis in Denmark. Acta Genet. Stat. Med. 5, 263±281. Horn, R., 1998. Explorations of voltage-dependent conformational changes using cysteine scanning. Methods Enzymol. 293, 145±55. Jackson, C.E., Barohn, R.J., PtaÂcek, L.J., 1994. Paramyotonia congenita: abnormal short exercise test, and improvement after mexiletine therapy. Muscle Nerve 17, 763±768. Jurkat-Rott, K., Lehmann-Horn, F., Elbaz, A., Heine, R., Gregg, R.G., Hogan, K., Powers, P., Lapie, P., Vale-Santos, J.E., Weissenbach, J., Fontaine, B., 1994. A calcium channel mutation causing hypokalemic periodic paralysis. Hum. Mol. Genet. 3, 1415±1419. Jurkat-Rott, K., Uetz, U., Pika-Hartlaub, U., Powell, J., Fontaine, B., Melzer, W., Lehmann-Horn, F., 1998. Calcium currents and transients of native and heterologously expressed mutant skeletal muscle DHP receptor a1 subunits (R528H. FEBS Lett. 423, 198±204. Jurkat-Rott, K., Mitrovic, N., Hang, C., Iaizzo, P., Herzog, J., Lerche, H., Nicole, N., Vale-Santos, J., Chauveau, D., Fontaine, B., Lehmann-Horn, F., 2000. Novel voltage sensor sodium channel mutations cause hypokalemic periodic paralysis type 2 by enhanced inactivation and reduced current. Proc. Natl. Acad. Sci. U.S.A., in press.

Sodium and calcium channelopathies of sarcolemma

29

Koch, M.C., Ricker, K., Otto, M., Grimm, T., Bender, K., Zoll, B., Harper, P.S., Lehmann-Horn, F., RuÈdel, R., Hoffman, E., 1991. Linkage data suggesting allelic heterogeneity for paramyotonia congenita and hyperkalemic periodic paralysis on chromosome 17. Hum. Genet. 88, 71±74. Kontis, K.J., Rounaghi, A., Goldin, A.L., 1997. Sodium channel activation gating is affected by substitutions of voltage sensor positive charges in all four domains. J. Gen. Physiol. 110, 391±401. Lapie, P., Goudet, C., Nargeot, J., Fontaine, B., Lory, P., 1996. Electrophysiolocal properties of the hypokalemic periodic paralysis mutation (R528H) of the skeletal muscle alpha 1s subunit as expressed in mouse L cells. FEBS Lett. 382, 244±248. Lehmann-Horn, F., Jurkat-Rott, K., 1999. Voltage-gated ion channels and hereditary disease. Physiol. Rev. 79, 1317±1371. Lehmann-Horn, F., KuÈther, G., Ricker, K., Grafe, P., Ballanyi, K., RuÈdel, R., 1987a. Adynamia episodica hereditaria with myotonia: a non-inactivating sodium current and the effect of extracellular pH. Muscle Nerve 10, 363±374. Lehmann-Horn, F., RuÈdel, R., Ricker, K., 1987b. Membrane defects in paramyotonia congenita (Eulenburg). Muscle Nerve 10, 633±641. Lehmann-Horn, F., Iaizzo, P.A., Hatt, H., Franke, C., 1991. Altered gating and reduced conductance of single sodium channels in hyperkalemic periodic paralysis. P¯uÈgers Arch. 418, 297±299. Lehmann-Horn, F., Rudel, R., Ricker, K., 1993. Non-dystrophic myotonias and periodic paralyses. Neuromusc. Disord. 3, 161±168. Lehmann-Horn, F., Engel, A.G., Ricker, K., RuÈdel, R., 1994. The periodic paralyses and paramyotonia congenita. In: Engel, A.G., Franzini-Armstrong, C. (Eds.), Myology, 2nd edn. McGraw-Hill, New York, pp 1303±1334. Lennox, G., Purves, A., Marsden, D., 1992. Myotonia ¯uctuans. Arch. Neurol. 49, 1010±1011. Lerche, H., Heine, R., Pika, U., George, A.L., Mitrovic, N., Browatzki, M., Weiû, T., Rivet-Bastide, M., Franke, C., Lomonaco, M., Ricker, K., Lehmann-Horn, F., 1993. Human sodium channel myotonia: slowed channel inactivation due to substitutions for glycine within the III/IV linker. J. Physiol. (Lond.) 470, 13±22. Lerche, H., Klugbauer, N., Lehmann-Horn, F., Hofmann, F., Melzer, W., 1996a. Expression and functional characterization of the cardiac L-type calcium channel carrying a skeletal muscle DHP±receptor mutation causing hypokalaemic periodic. P¯uÈgers Arch. 431, 461±463. Lerche, H., Mitrovic, N., Dubowitz, V., Lehmann-Horn, F., 1996b. Pathophysiology of paramyotonia congenita: the R1448P sodium channel mutation in adult human skeletal muscle. Ann. Neurol. 39, 599±608. Lerche, H., Peter, W., Fleischhauer, R., Pika-Hartlaub, U., Malina, T., Mitrovic, N., Lehmann-Horn, F., 1997. Role in fast inactivation of the IV/S4-S5 loop of the human muscle Na + channel probed by cysteine mutagenesis. J. Physiol. (Lond.) 505, 345±352. Links, T.P., Zwarts, M.J., Wilmink, J.T., Molenaar, W.M., Oosterhuis, H.J.G.H., 1990. Permanent muscle weakness in familial hypokalemic periodic paralysis. Brain 113, 1873. Lipicky, R.J., Bryant, S.H., Salmon, J.H., 1971. Cable parameters, sodium, potassium, chloride, and water content and potassium ef¯ux in isolated external intercostal muscle of normal volunteers and patients with myotonia congenita. J. Clin. Invest. 50, 2091±2103. McArdle, B., 1962. Adynamia episodica hereditaria and its treatment. Brain 85, 121±148. McClatchey, A.I., McKenna-Yasek, D., Cros, D., Worthen, H.G., Kuncl, R.W., DeSilva, S.M., Cornblath, D.R., Gusella, J.F., Brown Jr., R.H., 1992. Novel mutations in families with unusual and variable disorders of the skeletal muscle sodium channel. Nature Genet. 2, 148±152. McManis, P.G., Lambert, L.H., Daube, J.R., 1986. The exercise test in periodic paralysis. Muscle Nerve 9, 704±710. McPhee, J.C., Ragsdale, D.S., Scheuer, T., Catterall, W.A., 1998. A critical role for the S4-S5 intracellular

30

N. Mitrovic, H. Lerche

loop in domain IV of the sodium channel alpha-subunit in fast inactivation. J. Biol. Chem. 273, 1121± 1129. 2+ Melzer, W., Herrmann-Frank, A., LuÈttgau, H.C., 1995. The role of Ca ions in excitation±contraction coupling of skeletal muscle ®bers. Biochim. Biophys. Acta 1241, 59±116. Mitrovic, N., George Jr., A.L., Heine, R., Wagner, S., Pika, U., Hartlaub, U., Zhou, M., Lerche, H., Fahlke, Ch., Lehmann-Horn, F., 1994. Potassium-aggravated myotonia: the V1589M mutation destabilizes the inactivated state of the human muscle sodium channel. J. Physiol. (Lond.) 478, 395±402. Mitrovic, N., George Jr., A.L., Lerche, H., Wagner, S., Fahlke, Ch., Lehmann-Horn, F., 1995. Different effects on gating of three myotonia-causing mutations in the inactivation gate of the human muscle sodium channel. J. Physiol. (Lond.) 487, 107±114. Mitrovic, N., Lerche, H., Heine, R., Fleischhauer, R., Pika-Hartlaub, U., Hartlaub, U., George Jr., A.L., Lehmann-Horn, F., 1996. Role in fast inactivation of conserved amino acids in the IV/S4-S5 loop of the human muscle Na + channel. Neurosci. Lett. 214, 9±12. Mitrovic, N., George, A.L., Horn, R., 1998. Independent versus coupled inactivation in sodium channels. Role of the domain 2 S4 segment. J. Gen. Physiol. 111, 451±462. Mitrovic, N., George, A.L., RuÈdel, R., Lehmann-Horn, F., Lerche, H., 1999. Mutant channels contribute less than 50% to Na + current in paramyotonia congenita muscle. Brain 122, 1085±1092. Morrill, J.A., Brown Jr., R.H., Cannon, S.C., 1998. Gating of the L-type Ca channel in human skeletal myotubes: an activation defect caused by the hypokalemic periodic paralysis mutation R528H. J. Neurosci. 18, 10320±10334. Plassart, E., Elbaz, A., Vale Santos, J., Reboul, J., Lapie, P., Chauveau, D., Jurkat-Rott, K., Guimaraes, J., Saudubray, J.M., Weissenbach, J., Lehmann-Horn, F., Fontaine, B., 1994. Genetic heterogeneity in hypokalemic periodic paralysis (HypoPP). Hum. Genet. 94, 551±556. Poskanzer, D.C., Kerr, D.N.S., 1961. A third type of periodic paralysis with normokalemia and favorable response to sodium chloride. Am. J. Med. 31, 328±342. PtaÂcek, L.J., Trimmer, J.S., Agnew, W.S., Roberts, J.W., Petajan, J.H., Leppert, M., 1991a. Paramyotonia congenita and hyperkalemic periodic paralysis map to the same sodium-channel gene locus. Am. J. Hum. Genet. 49, 851±854. PtaÂcek, L.J., George Jr., A.L., Griggs, R.C., Tawil, R., Kallen, R.G., Barchi, R.L., Robertson, M., Leppert, M.F., 1991b. Identi®cation of a mutation in the gene causing hyperkalemic periodic paralysis. Cell 7, 1021±1027. PtaÂcek, L.J., George Jr., A.L., Barchi, R.L., Griggs, R.C., Riggs, J.E., Robertson, M., Leppert, M.F., 1992a. Mutations in an S4 segment of the adult skeletal muscle sodium channel cause paramyotonia congenita. Neuron 8, 891±897. PtaÂcek, L.J., Tawil, R., Griggs, R.C., Storvick, D., Leppert, M., 1992b. Linkage of atypical myotonia congenita to a sodium channel locus. Neurology 42, 431±433. PtaÂcek, L.J., Tawil, R., Griggs, R.C., Meola, G., McManis, P., Barohn, R.J., Mendell, J.R., Harris, C., Spitzer, R., Santiago, F., Leppert, M.F., 1994a. Sodium channel mutations in acetazolamide±responsive myotonia congenita, paramyotonia congenita and hyperkalemic periodic paralysis. Neurology 44, 1500±1503. PtaÂcek, L., Tawil, R., Griggs, R.C., Engel, A., Layzer, R.B., Kwiecinski, H., McManis, P.G., Santiago, L., Moore, M., Fouad, G., Bradley, P., Leppert, M.F., 1994b. Dihydropyridine receptor mutations cause hypokalemic periodic paralysis. Cell 77, 863±868. Richmond, J.E., Van De Carr, D., Featherstone, D.E., George Jr., A.L., Ruben, P.C., 1997. Defective fast inactivation recovery and deactivation account for sodium channel myotonia in the I1160V mutant. Biophys. J. 73, 1896±1903. Ricker, K., Haass, A., RuÈdel, R., BoÈhlen, R., Mertens, H.G., 1980. Successful treatment of paramyotonia congenita (Eulenburg). Muscle stiffness and weakness prevented by tocainide. J. Neurol. Neurosurg. Psychiatry 43, 268±271.

Sodium and calcium channelopathies of sarcolemma

31

Ricker, K., Rohkamm, R., BoÈhlen, R., 1986. Adynamia episodica and paralysis periodica paramyotonica. Neurology 36, 682±686. Ricker, K., Camacho, L., Grafe, P., Lehmann-Horn, F., RuÈdel, R., 1989. Adynamia episodica hereditaria: what causes the weakness? Muscle Nerve 10, 883±891. Ricker, K., Lehmann-Horn, F., Moxley, R.T., 1990. Myotonia ¯uctuans. Arch. Neurol. 47, 268±272. Ricker, K., Moxley, R.T., Heine, R., Lehmann-Horn, F., 1994. Myotonia ¯uctuans, a third type of muscle sodium channel disease. Arch. Neurol., 51, 1095±1102. Riggs, J.E., Griggs, R.C., 1979. Diagnosis and treatment of the periodic paralyses. In: Klawans, H.L. (Ed.), Clinical Neuropharmacology, Vol. 4. Raven Press, New York, pp 123±138. Riggs, J.E., Moxley, R.T., Griggs, R.C., Horner, F.A., 1981. Hyperkalemic periodic paralysis: an apparent sporadic case. Neurology 31, 1157±1159. Rojas, C.V., Wang, J., Schwartz, L., Hoffman, E.P., Powell, B.R., Brown Jr., R.H., 1991. A MettoVal mutation in the skeletal muscle sodium channel a-subunit in hyperkalemic periodic paralysis. Nature 354, 387±389. RuÈdel, R., Lehmann-Horn, F., Ricker, K., KuÈther, G., 1984. Hypokalemic periodic paralysis: In vitro investigation of muscle ®ber membrane parameters. Muscle Nerve 7, 110±120. RuÈdel, R., Lehmann-Horn, F., 1985. Membrane changes in cells from myotonia patients. Physiol. Rev. 65, 310±356. Ruff, R.L., 1994. Slow sodium channel inactivation must be disrupted to evoke prolonked depolarizationinduced paralysis. Biophys. J. 66, 542. Sipos, I., Jurkat-Rott, K., Harasztosi, Cs., Fontaine, B., Kovacs, L., Melzer, W., Lehmann-Horn, F., 1995. Skeletal muscle DHP receptor mutations alter calcium currents in human hypokalaemic periodic paralysis myotubes. J. Physiol. (Lond.) 483, 299±306. Streib, E.W., 1987. Differential diagnosis of myotonic syndromes. Muscle Nerve 10, 603±615. StuÈhmer, W., Conti, F., Suzuki, H., Wang, X., Noda, M., Yahagi, N., Kubo, H., Numa, S., 1989. Structural parts involved in activation and inactivation of the sodium channel. Nature 339, 597± 603. Subramony, S.H., Malhotra, C.P., Mishra, S.K., 1983. Distinguishing paramyotonia congenita and myotonia congenita by electromyography. Muscle Nerve 6, 374±379. Talbott, J.H., 1941. Periodic paralysis. Medicine 20, 85±143. 1649 Tang, L., Chehab, N., Wieland, S.J., Kallen, R.G., 1998. Glutamine substitution at alanine in the S4-S5 cytoplasmic loop of domain 4 removes the voltage sensitivity of fast inactivation in the human heart sodium channel. J. Gen. Physiol. 111, 639±652. Tricarico, D., Servidei, S., Tonali, P., Jurkat-Rott, K., Camerino, D.C., 1999. Impairment of skeletal muscle adenosine triphosphate-sensitive K + channels in patients with hypokalemic periodic paralysis. J. Clin. Invest. 103, 675±682. Trudell, R.G., Kaiser, K.K., Griggs, R.C., 1987. Acetazolamide responsive myotonia congenita. Neurology 37, 488±491. Tyler, F.H., Stephens, F.E., Gunn, F.D., Perkoff, G.T., 1951. Studies in disorders of muscle. VII. Clinical manifestations and inheritance of a type of periodic paralysis without hypopotassemia. J. Clin. Invest. 30, 492±502. Vita, G.M., Olckers, A., Jedlicka, A.E., George Jr., A.L., Heiman-Patterson, T., Rosenberg, H., Fletcher, 1306 J.E., Levitt, R.C., 1995. Masseter muscle rigidity associated with glycine -to-alanine mutation in adult muscle sodium channel a-subunit gene. Anesthesiology 82, 1097±1103. Wagner, S., Lerche, H., Mitrovic, N., Heine, R., George, A.L., Lehmann-Horn, F., 1997. A novel sodium channel mutation causing a hyperkalemic paralytic and paramyotonic syndrome with reduced penetrance. Neurology 49, 1018±1025. Wang, P., Clausen, T., 1976. Treatment of attacks in hyperkalemic familial periodic paralysis by inhalation of salbutamol. Lancet ii, 221±223.

32

N. Mitrovic, H. Lerche

Wang, J., Dubowitz, V., Lehmann-Horn, F., Ricker, K., Michaels, J., PtaÂcek, L.J., Hoffman, E.P., 1995. In vivo structure/function studies: consecutive Arg1448 changes to Cys, His and Pro at the extracellular surface of IVS4. In: Dawson, D.C., Frizzell, R.A. (Eds.). The Rockefeller University Press, New York, pp 77±88. West, J.W., Patton, D.E., Scheuer, T., Wang, Y., Goldin, A.L., Catterall, W.A., 1992. A cluster of hydrophobic amino acid residues required for fast Na + channel inactivation. Proc. Natl. Acad. Sci. U.S.A. 89, 10910±10914. Yang, N., Ji, S., Zhou, M., PtaÂcek, L.J., Barchi, R.L., Horn, R., George Jr., A.L., 1994. Sodium channel mutations in paramyotonia congenita exhibit similar biophysical phenotypes in vitro. Proc. Natl. Acad. Sci. U.S.A. 91, 12785±12789. Yang, N., Horn, R., 1995. Evidence for voltage-dependent S4 movement in sodium channels. Neuron 15, 213±218. Yang, N., George Jr., A.L., Horn, R., 1996. Molecular basis of charge movements in voltage-gated sodium channels. Neuron 16, 113±122. Yang, N., George Jr., A.L., Horn, R., 1997. Probing the outer vestibule of a sodium channel voltage sensor. Biophys. J. 73, 2260±2268. Zhou, J., Spier, S.J., Beech, J., Hoffman, E.P., 1994. Pathophysiology of sodium channelopathies: Correlation of normal/mutant mRNA ratios with clinical phenotype in dominantly-inherited periodic paralysis. Hum. Mol. Genet. 3, 1599±1603.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

33

Chapter 2

Muscle chloride channelopathies: myotonia congenita Reinhardt RuÈdel Department of General Physiology, University of Ulm, 89069 Ulm, Germany

Abstract The muscle chloride channelopathies are caused by mutations in the gene CLCN1 (located on chromosome 7q35) that encodes the skeletal muscle-speci®c chloride channel CLC1. Altogether nine different human genes have been identi®ed as members of the CLC chloride channel gene family. They code for voltage-gated chloride channels that are structurally completely different from any other known class of ion channels. The membrane topology of the CLC channels is still a matter of debate and structure±function relations of the protein are not yet understood. The functional CLC1 is most likely to be a homodimer. Gain-offunction and loss-of-function mutations are known in CLCN1 causing two types of myotonia congenita, namely the dominantly transmitted Thomsen type and the recessively transmitted Becker type, respectively. The main symptom of either type is a slowed relaxation from voluntary contraction owing to a transient hyperexcitability of the muscle ®bers. Many myotonia congenita patients manage their disease without medication. Myotonic stiffness responds well to antiarrhythmic drugs and related agents. Thomsen-type myotonia congenita is very rare: less than ten families have been identi®ed on the molecular level. The frequency of the more common Becker type is between 1:23 000 and 1:50 000. More than 40 mutations in CLCN1 causing recessive myotonia have been identi®ed so far. Animal models of myotonia congenita exist for either type. q 2000 Elsevier Science B.V. All rights reserved.

Introduction The symptom of myotonia is characterized by an uncontrolled temporary muscle stiffness caused by a transient hyperexcitability of the muscle ®ber membrane. Diseases with this symptom are called `myotonias'. The group of hereditary myotonias can be divided into four major subgroups according to clinical differences and different pathogenesis.

34

R. RuÈdel

(1) Myotonia congenita. This group consists of two forms, Thomsen's disease, or dominant myotonia congenita, the ®rst myotonic disease to be described (Thomsen, 1876), and recessive myotonia congenita, the more common form separated from it (Becker, 1977b). Both forms are non-progressive and non-dystrophic. They have been shown to be caused by allelic mutations of the gene coding for the chloride channel of the skeletal muscle ®ber membrane, which is located on chromosome 7q35. Hence they are classi®ed as `muscular chloride channel diseases' or `muscle chloride channelopathies' (Lehmann-Horn and Jurkat-Rott, 1999). They are the main subject of this chapter. (2) Paramyotonia congenita, hyperkalemic periodic paralysis and potassiumaggravated myotonia. Paramyotonia congenita was already clinically separated from myotonia congenita when it was ®rst described by Eulenburg (1886), because in this disorder, the muscle stiffness is distinct from that in myotonia congenita by two characteristic features: (1) upon continued use of the muscles, the myotonia increases rather than decreases (`paradoxical myotonia'), and (2) the myotonia is induced and very much aggravated when the muscles are cooled. In hyperkalemic periodic paralysis, the major symptom is the occurrence of episodic attacks of muscular weakness or paralysis. The attacks may or may not begin with the symptom of myotonia. Potassium-aggravated myotonia is symptomatically very similar to myotonia congenita but its pathogenesis differs from it. The three diseases are all inherited as autosomal dominant traits. They are caused by allelic mutations of the gene encoding the adult-type sodium channel of the skeletal muscle ®ber membrane, located on chromosome 17q13.1-3. They are dealt with together in Chapter 1. (3) Myotonic dystrophy was ®rst described by Steinert (1909) and by Batten and Gibbs (1909). In this progressive disease with dominant inheritance, myotonia is only one of many symptoms, of which the most severe is usually muscle dystrophy. Myotonic dystrophy is clinically different from the above non-dystrophic myotonias, and also its pathogenesis, though not yet clearly understood, is different from all other myotonias, as it is not caused by a mutation in a channel gene. Myotonic dystrophy is the most common adult muscular dystrophy of most of the populations studied, and no racial group is exempt. Molecular biological techniques have shown that there are at least two genetically distinct forms of the disease. The most common, `classical' form 1, is linked to chromosome 19q13, the mutation being a marked increase of a tandemly arrayed trinucleotide (CTG) repeat sequence in the non-translated part of a gene coding for a protein kinase. The much rarer and only recently discovered form 2 was also called proximal myotonic myopathy (PROMM). In some families, the condition was shown to be linked to chromosome 3q. The gene product is still unknown. (4) Schwartz-Jampel syndrome is a very rare muscle disease ®rst described in 1962. Myotonia is very prominent in this disease, and is probably the reason for a number of secondary symptoms characteristic for the syndrome. The myotonia is

Muscle chloride channelopathies

35

non-progressive and non-dystrophic. The syndrome is inherited as an autosomal recessive trait based on a gene defect located to chromosome 1p34-36.1 The gene product is still unknown.

De®nition of myotonic signs The word myotonia was coined by StruÈmpell (1891) as a composition of the Greek words for muscle and tension. In myotonia congenita, and also in myotonic dystrophy, the myotonia decreases with continued muscle activity, a phenomenon termed warm-up. In paramyotonia congenita the myotonia usually increases with muscle activity, a phenomenon termed paradoxical myotonia. Myotonic muscles react to a blow with the percussion hammer by becoming indented for a few seconds. This was called percussion myotonia by Erb (1885), who also detected that myotonic muscles have a lowered electrical threshold and an increased tendency to react to direct current with a prolonged contraction. The combination of these mechanical and electrical abnormalities is called the myotonic reaction. This reaction was used to diagnose a myotonic disease in the early days of electrophysiology and has now been replaced by the electromyogram (EMG) where the muscles exhibit myotonic runs, i.e. repetitive activity. In very mild cases myotonic stiffness might not be clinically present, yet the EMG might reveal myotonic runs. This is termed latent myotonia. Myotonia is by de®nition myogenic as opposed to neurogenic. Proof of this feature can be achieved with curare (Brown and Harvey, 1939; Lanari, 1946). The myotonias are thus delineated from neuromyotonia, spontaneous motor unit activity originating from an unexplained hyperexcitability of the terminal motor nerve branches (Gamstorp and Wohlfart, 1959; Isaacs, 1961; Mertens and Zschocke, 1965; Greenhouse et al., 1967; Nakanishi et al., 1975; Negri et al., 1977; Nakanishi et al., 1978; Partanen et al., 1980). Animal species having myotonia congenita are the myotonic goat, a model for dominant myotonia congenita, and the ADR and MTO mice, models for recessive myotonia congenita.

Myotonia congenita The ®rst description of a myotonic disorder was by Thomsen (1876), who noticed what he called myotonia congenita in himself and his son, and thus concluded that the disease was transmitted as a dominant trait. In the 1950s, Becker (1957) recognized that in many families that had been diagnosed as having myotonia congenita the inheritance was recessive. In these families myotonia was more generalized than

36

R. RuÈdel

in Thomsen's disease. Therefore Becker named this type `recessive generalized myotonia'. It is now clear that both the dominant and the recessive forms are caused by different mutations in the same gene which codes for the major chloride channel of adult human skeletal muscle (Koch et al., 1992). The intensive search for mutations that followed this discovery showed that the dominant form is very rare, as less than 10 different families have been identi®ed at the molecular genetic level to date. The recessive form is much more common, and the estimation by Becker (1977a) of a frequency between 1:2000 and 1:5000 might still hold. Males seem to predominate at a rate of 3:1 when the Becker-type propositi are counted. However, family studies disclose that women are affected at the same frequency though to a much lesser degree (Deymeer et al., 1999).

Clinical signs of dominant myotonia congenita (Thomsen's disease) Usually the myotonia is recognized in early childhood but the milder cases may go unrecognized until late childhood. The myotonia is generalized with the legs being most affected causing the children to fall frequently. The cranial and upper limb musculature can be severely affected. Chewing is sometimes impaired. The myotonic stiffness is most pronounced when a forceful movement is abruptly initiated after the patient has rested for 5±10 min. For instance, after making a hard ®st, the patient may not be able to extend the ®ngers fully for several seconds (Fig. 1). The myotonia decreases or vanishes completely as the same movement is repeated several times (`warm-up phenomenon'), but it always recurs after a few minutes of rest. The patient may experience much dif®culty while getting up from a chair or stepping into a bus in a hurry. Occasionally a sudden noise may cause instantaneous generalized stiffness. The patient may then fall to the ground and remain rigid and helpless for some seconds or even minutes. Some patients have hypertrophied muscles and an athletic appearance (Fig. 2). Their muscle strength is normal or even greater than normal and they can be quite successful in sports where strength is more important than speed. A slight contracture of the calves may limit dorsi¯exion of the feet. Tapping a muscle produces an indentation that persists for a second or so (percussion myotonia, Fig. 3). Lid lag is usually present, and in some patients myotonia of the lid muscles causes blepharospasm after forceful eye closure. The tendon re¯exes are normal. In some families the degree of myotonia ¯uctuates at a very slow and irregular periodicity of up to several months. In these families af¯icted members may sometimes suffer from muscle pain due to muscle spasms.

Muscle chloride channelopathies

37

Fig. 1. Myotonic stiffness of the hand of a patient with generalized myotonia (Becker). The patient was asked to rest his ®ngers for 5 min, then to make a ®st as forcefully as possible for 3 s, and then to stretch the ®ngers. It took more than 10 s for the patient to open his ®st fully (from RuÈdel et al., 1994).

Clinical signs of Becker-type myotonia The clinical picture of recessive myotonia resembles that of the dominant form. A few special points are worth mentioning. In some patients the myotonia does not manifest until the age of 10 years or even later, although in a few it is present by the age of 2±3 years. The severity of the myotonia may slowly increase for a number of years but usually not after the age of 25±30 years. In general the myotonia is more severe in recessive than in dominant myotonia congenita. Patients with Becker myotonia are more handicapped in daily life. The disability stems mainly from myotonic stiffness affecting the leg muscles. In severely affected young patients the stiffness may lead to toe-walking. Even more disabling in Becker patients is a peculiar transient weakness commonly encountered. This is best demonstrated when the patient makes a tight ®st after a period of rest: the force exerted by the ®nger ¯exors vanishes almost completely within a few seconds. With repeated muscle contractions, the force returns within 20±60 s. This transient weakness is often generalized and troublesome, typically occurring when a patient attempts to rise from a recumbent position after rest or sleep. The leg and gluteal

38

R. RuÈdel

Fig. 2. Three siblings from a family with generalized myotonia (Becker). The girl on the left and the boy are affected, while the girl on the right is not. Note the different postures and the hypertrophy of the thighs and upper arms (from RuÈdel et al., 1994).

muscles are often markedly hypertrophied and a lordotic appearance is common. By contrast, the neck, shoulder and arm muscles appear poorly developed resulting in a characteristic disproportionate ®gure. Patients with severe recessive myotonia congenita are limited in their choice of occupation and they are unsuited for military service. Life expectancy is normal. In a few families the heterozygotes can be identi®ed by showing repetitive action potentials on EMG.

Muscle chloride channelopathies

39

Fig. 3. Percussion myotonia in the thigh of a patient with generalized myotonia (Becker) (from RuÈdel et al., 1994).

Pathogenesis The muscle stiffness is due to the continued generation of runs of action potentials in the muscle ®ber membrane for some seconds after a voluntary contraction. This continued activity prevents immediate muscle relaxation from occurring. Experiments with muscles of an animal model, the myotonic goat, showed that the overexcitability is caused by a permanent reduction of the resting chloride conductance of the muscle ®ber membranes (Bryant, 1969). The high chloride conductance is necessary for a fast repolarization of the transverse tubular membrane which becomes depolarized by potassium accumulation in the tubules during tetanic muscle excitation. The depolarization increases with every action potential and, in the range of small depolarizations, excitability increases with increasing depolarization. The myotonic run comes to an end when, due to increased depolarization, not enough sodium channels recover from inactivation to generate the next spike. The `warm-up phenomenon' familiar to all myotonia congenita patients who take advantage of it when time is available to prepare for a voluntary movement, is caused by the shortening of consecutive myotonic runs with repeated use of the muscle. It has been claimed, on the basis of in vitro experiments with a pharmacologically induced myotonia model, that the warm-up effect is due to the lowering of

40

R. RuÈdel

the intracellular pH that physiologically occurs during muscle activity (Birnberger and Klepzig, 1979). The basic pathology of the myotonic run, ®rst explained on the basis of pharmacological experiments with external intercostal muscle of the myotonic goat (Adrian and Bryant, 1974), was later also demonstrated to prevail in human dominant and recessive myotonia congenita (Lipicky, 1979; RuÈdel et al., 1988; Franke et al., 1991). The skeletal muscle chloride channel gene was therefore an excellent candidate for carrying disease-causing mutations.

Molecular genetics The starting point for an understanding of myotonia congenita at the molecular level was expression cloning of the chloride channel CLC0 in the electric organ of the ®sh Torpedo marmorata (Jentsch et al., 1990). Rat skeletal muscle chloride channel cDNA was then cloned by homology screening (Steinmeyer et al., 1991a). This was followed by demonstration of linkage of both dominant and recessive myotonia congenita to chromosome 7q35 (Koch et al., 1992). A large number of mutations have subsequently been identi®ed in this chloride channel gene in both forms of myotonia congenita. CLCN1, the gene encoding CLC1, the chloride channel responsible for the high resting membrane conductance of skeletal muscle cells, is a member of a newly identi®ed multigene family encoding chloride channels that are structurally not related to any other known class of ion channels. It spans at least 40 kb and contains 23 exons whose boundaries have been located (Lorenz et al., 1994). The complete coding sequence consists of 2964 base pairs. CLC1, the gene product, is a protein of 988 amino acids with a predicted molecular weight of 110 kDa. Fig. 4 is a revision (Kieferle et al., 1994; Middleton et al., 1994; Schmidt-Rose and Jentsch, 1997a,b) of the original membrane topology model that was based on hydropathy analysis performed for CLC0, the chloride channel of the electric organ of Torpedo (Jentsch et al., 1990).

Voltage-dependent chloride channels Meanwhile, altogether nine different human genes have been identi®ed as members of the CLC chloride channel gene family (CLCN1±7; CLCNA/B, Table 1). They encode voltage-gated chloride channels, termed CLC1±CLC7, CLCKA, and CLCKB (Jentsch et al., 1999). In general, they ful®l a variety of functions depending on their tissue distribution, i.e. stabilization of the membrane potential, regulation of the cell volume and transepithelial chloride transport

Muscle chloride channelopathies

41

Fig. 4. Revised version of the original membrane topology model that was based on hydropathy analysis of the skeletal muscle chloride channel monomer, CLC1. The different symbols designate known mutations leading to dominant Thomsen-type myotonia, recessive Becker-type myotonia, recessive mouse myotonia and dominant goat myotonia (see left-hand bottom). Conventional one-letter abbreviations used for replaced amino acids located at positions are given by the respective numbers of the human protein.

(Jentsch et al., 1995). The channels can be found both in the plasmalemma and in the lining of internal organelles. In axons, the voltage-dependent chloride conductance is so small that it is usually neglected when the contributions of ion component conductances to the generation of the action potential are considered. In resting skeletal muscle the chloride conductance is about four times larger than the potassium conductance (Bretag, 1987). Nevertheless, its electrophysiological identi®cation and characterization at the single-channel level was very dif®cult because the single-channel conductance is very low, i.e. near 1 pS as estimated from noise analysis (Pusch and Jentsch, 1994). The large macroscopic chloride conductance of the skeletal muscle ®ber membrane, therefore, must result from an extremely high channel density. The ®rst membrane topology model of CLC0 assumed 13 transmembrane helical segments D1±13 (Jentsch et al., 1990). Later, a combination of glycosylation and electrophysiological experiments on mutant proteins showed that both the N- and Cterminus must be located intracellularly whereas the S8±S9 interlinker must be on the extracellular side because it carries a glycosylation site. Two different possibilities of con®guration arose from these results, namely a model that places S4

Accession ID of GDB or GB

Chromosome location

Chloride channel type, synonyma

Tissue expression

CLCN1

GDB: 1346887, GB: Z25752-68, L25587

7q32qter

Skeletal muscle, placenta

CLCN2

GDB: 270664

3q2728

CLC1, 988-aa, major skeletal muscle channel, myotonia congenita CLC2

CLCN3

GDB: 270665, GB: AF029346, X78520

4q33, 4q32, 4pterqter

CLC3, protein kinase C± regulated channel (760-aa)

CLCN4

GDB: 270666, GB: X77197

Xp22.3

CLC4

CLCN5

GDB: 270667, GB: X91906

Xp11.22

CLC5 (746-aa), X-linked nephro-lithiasis

CLCN6

1p36

CLC6, KIAA0046 (97 kDa)

CLCN7

GDB: 3929143, GB: AF00924757, X83378 GDB: 3929156, GB: Z67743

16p13

CLC7 (89 kDa)

CLCNKA

GDB: 698471

1p36

CLCNKB

GDB: 698472

1p36

CLCKa, CLCK1, kidneyspeci®c channel CLCKb, CLCK2, kidneyspeci®c channel, Bartter III

Ubiquitous, kidney (S3 segment of the proximal tubule) Brain, skeletal muscle, lung, retina, kidney (type B intercalated cells of connecting tubule and collecting duct) Skeletal muscle, brain, heart, kidney, retina Kidney (type A intercalated cells of connecting tubule and collecting duct) Ubiquitous Brain, testes, skeletal muscle, kidney Kidney inner medulla: thin ascending limb of Henle Kidney: thin ascending limb of Henle and collecting ducts

a GDB, genome database (http://gdbwww.gdb.org); GB, GenBank (http://www.ncbi.nlm.nih.gov/Entrez); aa, amino acids (from Lehmann-Horn and Jurkat-Rott, 1999).

R. RuÈdel

Gene name

42

Table 1 Voltage-gated chloride channels. Listed are their gene names, accession numbers, protein names and tissue expression. When no information on tissue expression in humans was available, data for other mammals are given a

Muscle chloride channelopathies

43

extracellularly with a hydrophobic core of S9±S12 that crosses the membrane several times (Fig. 4, Pusch et al., 1994; Jentsch et al., 1995; Schmidt-Rose and Jentsch, 1997b) or, alternatively, a model that places S2 at the extracellular side (Adachi et al., 1994; Middleton et al., 1994). Co-expression of wild-type CLC0 with naturally occurring mutants that change the single-channel conductance resulted in chloride channels having two different conductance levels and largely independent pores (Ludewig et al., 1996; Middleton et al., 1996). This suggests that the CLC0 channel protein is a `double-barrelled' homodimer with two functional off-axis pores each with its own independent activation gate, and with a single slow inactivation gate in common. A similar homodimeric structure has also been postulated for the voltage-dependent skeletal muscle chloride channel CLC1 (Fahlke et al., 1997c). Closely related members of the same subfamily such as CLC1 and the voltage- and volume-sensitive, ubiquitously expressed CLC2 are also capable to assemble as heterodimers. CLC1 is functional when expressed in Xenopus oocytes (Steinmeyer et al., 1991b; Pusch and Jentsch, 1994; Pusch et al., 1994; Jentsch et al., 1995) or human embryonic kidney cells (Pusch and Jentsch, 1994; Fahlke and RuÈdel, 1995) without any other subunits. The channel conducts over the whole physiological voltage range, showing inward recti®cation in the negative potential range. It is activated upon depolarization, and with hyperpolarizing voltage steps it is deactivated to a non-zero steady-state level. As known from macroscopic experiments (Bryant and Morales-Aguilera, 1971; Palade and Barchi, 1977), the channel can be blocked by external iodide and monocarboxylic aromatic acids. The acid of choice, 9-anthracene carboxylic acid, is effective at low millimolar concentrations (Steinmeyer et al., 1991b; Pusch and Jentsch, 1994). The time course of chloride currents conducted by CLC1 expressed in Xenopus oocytes, human embryonic kidney (HEK-293) cells or the insect cell line Sf-9 (Astill et al., 1996) is similar to that found in native muscle ®bers (Fahlke and RuÈdel, 1995). Electrophysiological studies of wild-type and mutant channel proteins have provided the ®rst insights into the structure-function relationship of CLC1, and led to the identi®cation of regions involved in gating and ion permeation (Steinmeyer et al., 1994; Pusch et al., 1995; Fahlke et al., 1996; 1997a; KuÈrz et al., 1996; Rychkov et al., 1997). Fahlke and RuÈdel (1995) showed a negative charge in S1 to be involved in the voltage-sensing mechanism and postulated that the intracellular mouth of the pore changes its af®nity to a putative gating particle in three different grades, thus mediating the three known gating modes differentiated by their time course: fast, slow and time-independent. The grading could result from a set of two negatively charged voltage sensors combining to three different states. This model is able to explain all macroscopic gating properties so far described. The cytoplasmic face of the pore vestibule may be associated with a phosphorylation site for protein

44

R. RuÈdel

kinase C as suggested by an increase of non-deactivating channels and reduction of single-channel conductance without change in voltage-sensitivity of channel gating following protein kinase C activation (Rosenbohm et al., 1999). As mentioned above, inferences from experiments with CLC0 channel constructs (Ludewig et al., 1996; Middleton et al., 1996) and studies of CLC1 constructs (Fahlke et al., 1997b) strongly suggest that the functional channel is a homodimer. Molecular pathogenesis More than 30 point mutations and three deletions have been found in CLCN1 of affected families. They may cause either dominant or recessive myotonia congenita (Fig. 4, Table 2) by producing change or loss of function of the gene product, respectively. In general, gene dosage effects of loss-of-function mutations may lead to a recessive or dominant phenotype, depending on whether 50% of the gene product (supplied by the normal allele) is suf®cient for normal function. Experiments with myotonia-generating drugs on normal muscle showed that blockade of 50% of the physiological chloride current is not suf®cient to produce myotonia. This explains the existence of recessive transmission in the case of mutations that completely destroy the generation of a gene product. Dominant inheritance is explained by a mutant CLC1 that can bind to another CLC1 (i.e. form a channel dimer with a non-mutant CLC1) and, in doing so, changes its function in the sense of a dominant negative effect. The most common feature of the resulting chloride currents in dominant myotonia is a shift of the activation curve towards more positive membrane potentials. As a consequence of this shift the chloride conductance is reduced in the physiologic range of membrane potentials (Fig. 5). Surprisingly, the size of the shift and clinical severity sometimes disagree, e.g. Gln-552-Arg causes an unusually large shift but only a very mild clinical phenotype, called myotonia levior (Pusch and Jentsch, 1994; Lehmann-Horn et al., 1995). Animal models of myotonia Myotonic goats About 30 years after the ®rst description of myotonia in man, White and Plaskett (1904) described a breed of `fainting' goats raised in Tennessee (Fig. 6). The animals tended to have attacks of extreme muscle stiffness when attempting a quick forceful motion, so that they often fell to the ground for 5±20 s with extension of the limbs and neck. Clark et al. (1939) were the ®rst to refer to the disease as `a form of congenital myotonia in goats'. Excised external intercostal muscles from such goats were used by Bryant (1969) in his electrophysiological studies of the

Muscle chloride channelopathies

45

Table 2 CLCN1 mutations causing myotonia congenita. The gene encodes the major chloride channel of skeletal muscle a

Genotype

Segment

Region

Mutation

Trait

1 3/A ! T C202T C220T C313T A407G A449G T481G T494G C501G G598A G689A G706C A782G G854A G854A T857C C870G G871A C898T T920C G937A G950A G979A T986C G1013A 1095-96D T1238G C1244T 1262insC 1278-81D C1439T 1437-50D C1443A G1444A A1453G G1471A G1488T C1649T

Intron 1 Exon 2 Exon 2 Exon 3 Exon 3 Exon 4 Exon 4 Exon 4 Exon 4 Exon 5 Exon 5 Exon 6 Exon 7 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 8 Exon 9 Exon 9 Exon 10 Exon 11 Exon 11 Exon 12 Exon 12 Exon 13 Exon 13 Exon 13 Exon 13 Exon 13 Exon 13 Exon 14 Exon 15

N-term N-term N-term N-term S1 S1-S2 S1-S2 S2 S2 S2-S3 S3-S4 S4 S4-S5 S5 S5 S5 S5-S6 S5-S6 S5-S6 S5-S6 S6 S6 S6-S7 S6-S7 S6-S7 S7 S8 S8 S8-S9 S8-S9 S9-S10 S9-S10 S9-S10 S9-S10 S9-S10 S10 S10±S11 S11±S12

Splice site Gln-68-Stop Gln-74-Stop Arg-105-Cys Asp-136-Gly Tyr-150-Cys Phe-161-Val Val-165-Gly Phe-167-Leu Gly-200-Arg Gly-230-Glu Val-236-Leu Tyr-261-Cys Splice site Gly-285-Glu Val-286-Ala Iie-290-Met Glu-291-Lys Arg-300-Stop Phe-307-Ser Ala-313-Thr Arg-317-Gln Splice site Ile-329-Thr Arg-338-Gln fs 387-Stop Phe-413-Cys Ala-415-Val fs 429-Stop fs 433-Stop Pro-480-Leu fs 503-Stop Cys-481-Stop Gly-482-Arg Met-485-Val Splice site Arg-496-Ser Thr-550-Met

Recessive Recessive Recessive Recessive Recessive Recessive Recessive Recessive Recessive Dominant Dom./rec. Recessive Recessive Recessive Recessive Dominant Dominant Recessive Recessive Dom./rec. Dom./rec. Dominant Recessive Recessive Dom./rec. Recessive Recessive Recessive Recessive Recessive Dominant Recessive Recessive Recessive Recessive Recessive Recessive Recessive

R. RuÈdel

46

Table 2 (continued) Genotype

Segment

Region

Mutation

Trait

A1655G

Exon 15

S12

Gln-552-Arg

T1667A G1687A C2124G G2149D C2680T

Exon Exon Exon Exon Exon

S12 S12 C-term C-term C-term

Ile-556-Asn Val-563-Ile Phe-708-Leu Glu-717-Stop Arg-894-Stop

Dom. levior Dom./rec. Recessive Recessive Recessive Dom./rec.

15 15 17 17 23

a

fs, frame shift due to an insertion (ins) and/or deletion (K); both events cause altered amino acid sequence usually followed by premature termination, sometimes indicated by the given stop codon number (from Lehmann-Horn and Jurkat-Rott, 1999).

membrane conductance of myotonic muscle that led him to the fundamental ®nding of a reduced chloride conductance in resting myotonic muscle ®bers. The myotonic goat did not play a role in the ®nding of the gene defect that causes the reduced chloride conductance. Long after CLCN1 was localized and cloned for mouse (Steinmeyer et al., 1991a) and man (Koch et al., 1992), the mutation in the homologous goat gene was detected (Beck et al., 1996). It predicts an Ala-885-Pro substitution in the C terminus of the chloride channel protein (Fig. 4) that causes a right-shift of the activation curve of the chloride current, similarly to dominant mutations in man. Myotonic mice In the late 1970s, two spontaneous mouse mutations were detected (reviewed by RuÈdel, 1990), one in the A2G strain in London, the other in the SWR/J strain in Bar Harbor, Maine. The behavioral abnormalities of the affected animals were very similar, and in both mutants the symptom was transmitted as an autosomal recessive trait. The British scientists were struck by the observation that from days 10±12 onwards the affected animals had dif®culty in righting themselves when placed supine and therefore called the mutation adr for `arrested development of righting response' (Fig. 7). The Americans observed that shaking the cage provoked sustained extension of an animal's hind limbs, and since electrical myotonia was recorded in the EMG from the stiff muscles, this strain was called mto for `myotonic'. As far as the phenotype is concerned, the two models of myotonia are virtually indistinguishable and, as in the myotonic goat, the reason for the abnormal excitability is a reduced chloride conductance. The assumption of interspecies conservation of the genomic structure in the vicinity of the adr locus found for the mice led Jockusch (1990) to predict that the Becker myotonia gene is located

Muscle chloride channelopathies

47

Fig. 5. Altered channel properties of a naturally occurring CLC1 mutant (G200R) causing dominant (Thomsen-type) myotonia in man. Macroscopic currents recorded in the whole-cell mode from WT (top panel) and mutant (middle) channels are expressed in a mammalian cell line. Chloride currents were activated by voltage steps from a holding potential of 0 mV to potentials of 2145 to 195 mV, and deactivated after 400 ms by polarization to 2105 mV. Bottom panel: voltage dependence of relative open probability deduced from currents as illustrated above shows that the `activation curve' of the mutant channel is shifted to the right, so that in the physiological potential range the open probability is much reduced. All mutations that cause such a voltage shift cause dominant mode of inheritance (modi®ed after Wagner et al., 1998).

48

R. RuÈdel

Fig. 6. Severe muscle stiffness in myotonic goats, a naturally occurring animal model of dominant myotonia, provoked by the noise (courtesy of S. Bryant and A.L. George).

on human chromosome 7 before CLC1 was cloned. When this was ®nally accomplished (Steinmeyer et al., 1991a, see above), the spontaneous mutation of the ADR mouse was soon identi®ed which destroys the gene's coding potential for several membrane spanning domains (Fig. 4, Steinmeyer et al., 1991b). From this and the lack of recombination between the CLC1 gene and the adr locus, it was concluded that a lack of functional chloride channels is the primary cause of mouse myotonia.

Fig. 7. `Arrested development of righting' due to severe muscle stiffness in the myotonic ADR mouse (left) compared to the response of a control mouse (right). Both animals had been turned on their back at the same time (courtesy of Brinkmeier).

Muscle chloride channelopathies

49

Therapy Many myotonia congenita patients can manage their disease without medication. Should treatment be necessary myotonic stiffness responds well to drugs that reduce the increased excitability of the cell membrane by interfering with the sodium channels, i.e. local anesthetics, anti®brillar and antiarrhythmic drugs and related agents. These drugs suppress myotonic runs by decreasing the number of available sodium channels and have no known effect on chloride channels. Of the many drugs tested, mexiletine (RuÈdel et al., 1980; Leheup et al., 1986) is the drug of choice. A simple method for scoring the severity of the myotonia before starting therapy and for evaluating the effect of the treatment is provided by the stair test (Birnberger et al., 1975). The patient should rest for 10±15 min in a chair at the foot of the stairs, then get up and climb 10 steps as quickly as possible. A healthy person needs about 3 s, a patient with severe myotonia up to 30 s. Immediate repetition of the test demonstrates the warm-up phenomenon.

Open questions Many basic questions on myotonia congenita are now understood. We know the mutated gene and its product, and fairly well understand the physiologic mechanisms underlying the pathology of muscle stiffness. No modern experiments have been reported designed to test the explanation for the warm-up phenomenon that stems unchallenged from the pre-molecular biology days (Birnberger and Klepzig, 1979). An interesting, so far unanswered question arose with the increased search for myotonia families who might have a so far unreported mutation. Several mutations were found that lead to myotonia which under certain circumstances is transmitted as a dominant trait and under other circumstances as a recessive trait. What the decisive circumstance is has so far not been elucidated. Perhaps it is connected to one or the other polymorphism in CLCN1 that has no functional consequences within the wildtype. As to dominant inheritance, as mentioned above this is believed to be caused by mutations that lead to the production of an altered protein that is still able to form dimers though with compromised function. All dominant mutations have been found to shift the activation curve of CLC1 towards more positive membrane potentials. There is, however, no simple relation between the amount of shift that was measured with the various mutant CLC1 channels and the severity of symptoms usually seen with patients carrying the respective mutation. Apparently other, so far undetected factors act in an ancillary fashion. Such factors may also play a role in cases of recessive myotonia where the symptoms ¯uctuate (Wagner et al., 1998). Even more

50

R. RuÈdel

enigmatic is the ®nding that some `recessive' mutations do not lead to the loss of a gene product but to channels that, when expressed in one of the usual expression systems, conduct chloride currents with normal amplitude and normal gating behavior. Relatively little ± in comparison to our knowledge on cation channels ± is known about the structure/function relation of CLC1. This is of course due to the fact that the CLC family of chloride channels has been detected only recently and that the structure of its members is so totally different from that of other, well understood, channel family members. Molecular biological methods, such as site-directed mutagenesis, use of channel chimeras, etc. will hopefully help to overcome this unsatisfactory situation soon. References Adachi, S., Uchida, S., Ito, H., Hata, M., Hiroe, M., Marumo, F., Sasaki, S., 1994. Two isoforms of a chloride channel predominantly expressed in thick ascending limb of Henle's loop and collecting ducts of rat kidney. J. Biol. Chem. 269, 17677±17683. Adrian, R.H., Bryant, S.H., 1974. On the repetitive discharge in myotonic muscle ®bres. J. Physiol. Lond. 240, 505±515. Astill, D.S., Rychkov, G., Clarke, J.D., Hughes, B.P., Roberts, M.L., Bretag, A.H., 1996. Characteristics of skeletal muscle chloride channel ClC-1 and point mutant R304E expressed in Sf-9 insect cells. Biochim. Biophys. Acta 1280, 178±186. Batten, F.E., Gibbs, H.P., 1909. Myotonia atrophica. Brain 32, 187. Beck, C.L., Fahlke, C., George Jr., A.L., 1996. Molecular basis for decreased muscle chloride conductance in the myotonic goat. Proc. Natl. Acad. Sci. USA 93, 11248±11252. Becker, P.E., 1957. Zur Frage der Heterogenie der erblichen Myotonien. Nervenarzt 28, 455±460. Becker, P.E. 1977a. Myotonia Congenita and Syndromes Associated with Myotonia. Georg Thieme, Stuttgart. Becker, P.E. 1977b. Syndromes associated with myotonia: clinical-genetic classi®cation. In: Rowland, L.P. (Ed.). Pathogenesis of Human Muscular Dystrophies. Excerpta Medica, Amsterdam, pp. 699± 703. Birnberger, K.L., Klepzig, M., 1979. In¯uence of extracellular potassium and intracellular pH on myotonia. J. Neurol. 222, 23±35. Birnberger, K.L., RuÈdel, R., Struppler, A., 1975. Clinical and electrophysiological observations in patients with myotonic muscle disease and the therapeutic effect of N-propyl-ajmalin. J. Neurol. 210, 99±110. Bretag, A.H., 1987. Muscle chloride channels. Physiol. Rev. 67, 618±724. Brown, G.L., Harvey, A.M., 1939. Congenital myotonia in the goat. Brain 62, 341. Bryant, S.H., 1969. Cable properties of external intercostal muscle ®bres from myotonic and nonmyotonic goats. J. Physiol. Lond. 204, 539±550. Bryant, S.H., Morales-Aguilera, A., 1971. Chloride conductance in normal and myotonic muscle ®bres and the action of monocarboxylic aromatic acids. J. Physiol. Lond. 219, 367±383. Clark, S.L., Luton, F.H., Cutler, J.T., 1939. A form of congenital myotonia in goats. J. Nerv. Ment. Dis. 90, 297±309. È zdemir, C., 1999. Deymeer, F., Lehmann-Horn, F., Serdaroglu, P., Cakirkaya, S., Benz, S., RuÈdel, R., O Electrical myotonia in heterozygous carriers of recessive myotonia congenita. Muscle Nerve 22, 123± 125.

Muscle chloride channelopathies

51

Erb, W., 1885. Klinisches und Pathologisch-Anatomisches von der Thomsenschen Krankheit. Neurol. Zentralbl. 13, 289. È ber eine familiaÈre durch 6 Generationen verfolgbare Form congenitaler ParamyoEulenburg, A., 1886. U tonie. Neurol. Zentralbl. 5, 265±272. Fahlke, C., RuÈdel, R., 1995. Chloride currents across the membrane of mammalian skeletal muscle ®bres. J. Physiol. Lond. 484, 355±368. Fahlke, C., Rosenbohm, A., Mitrovic, N., George Jr., A.L., RuÈdel, R., 1996. Mechanism of voltagedependent gating in skeletal muscle chloride channels. Biophys. J. 71, 695±706. Fahlke, C., Beck, C.L., George Jr., A.L., 1997a. A mutation in autosomal dominant myotonia congenita affects pore properties of the muscle chloride channel. Proc. Natl. Acad. Sci. USA 94, 2729±2734. Fahlke, C., Knittle, T., Gurnett, C.A., Campbell, K.P., George Jr., A.L., 1997b. Subunit stoichiometry of human muscle chloride channels. J. Gen. Physiol. 109, 93±104. Fahlke, C., Yu, H.T., Beck, C.L., Rhodes, T.H., George Jr., A.L., 1997c. Pore-forming segments in voltage-gated chloride channels. Nature 390, 529±532. Franke, C., Iaizzo, P.A., Hatt, H., Spittelmeister, W., Ricker, K., Lehmann-Horn, F., 1991. Altered Na 1 channel activity and reduced Cl 2 conductance cause hyperexcitability in recessive generalized myotonia (Becker). Muscle Nerve 14, 762±770. Gamstorp, I., Wohlfart, G., 1959. A syndrome characterised by myokymia, myotonia, muscular wasting and increased perspiration. Acta Psychiatr. Scand. 34, 181. Greenhouse, A.H., Bicknell, J.M., Pesch, R.N., Seelinger, D.F., 1967. Myotonia, myokymia, hyperhydrois and wasting of muscle. Neurology 17, 263±268. Isaacs, H., 1961. Syndrome of continuous muscle-®ber activity. J. Neurol. Neurosurg. Psychiatry 24, 319. Jentsch, T.J., Steinmeyer, K., Schwarz, G., 1990. Primary structure of Torpedo marmorata chloride channel isolated by expression cloning in Xenopus oocytes. Nature 348, 510±514. Jentsch, T.J., GuÈnther, W., Pusch, M., Schwappach, B., 1995. Properties of voltage-gated chloride channels of the ClC gene family. J. Physiol. Lond. 482, 19S±25S. Jentsch, T.J., Friedrich, T., Schriever, A., Yamada, H., 1999. The CLC chloride channel family. P¯uÈgers Arch. 437, 783±795. Jockusch, H., 1990. Molecular aspects of myotonia: the ADR mouse as a model. J. Neurol. Sci. 88, 9. Kieferle, S., Fong, P., Bens, M., Vandewalle, A., Jentsch, T.J., 1994. Two highly homologous members of the ClC chloride channel family in both rat and human kidney. Proc. Natl. Acad. Sci. USA 91, 6943± 6947. Koch, M.C., Steinmeyer, K., Lorenz, C., Ricker, K., Wolf, F., Otto, M., Zoll, B., Lehmann-Horn, F., Grzeschik, K.H., Jentsch, T.J., 1992. The skeletal muscle chloride channel in dominant and recessive human myotonia. Science 257, 797±800. KuÈrz, L.L., Wagner, S., George Jr., A.L., RuÈdel, R., 1996. Probing the major skeletal muscle chloride channel with Zn 21 and other sulfhydryl-reactive compounds. P¯uÈgers Arch. 433, 357±363. Lanari, A., 1946. Mechanism of myotonic contraction. Science 104, 221. Leheup, B., Himon, F., Morali, A., Brichet, F., Vidailhet, M., 1986. Value of mexiletine in the treatment of Thomsen±Becker myotonia. Arch. Fr. Pediatr. 43, 49±50. Lehmann-Horn, F., Jurkat-Rott, K., 1999. Voltage-gated ion channels and hereditary disease. Physiol. Rev. 79, 1317±1372. Lehmann-Horn, F., MailaÈnder, V., Heine, R., George Jr., A.L., 1995. Myotonia levior is a chloride channel disorder. Hum. Mol. Genet. 4, 1397±1402. Lipicky, R.J. (1979). Myotonic syndromes other than myotonic dystrophy. In: Vinken, P.J., Bruyn, G.W. (Eds.). Handbook of Clinical Neurology, Vol. 40. Elsevier, Amsterdam, pp. 533±571. Lorenz, C., Meyer-Kleine, C., Steinmeyer, K., Koch, M.C., Jentsch, T.J., 1994. Genomic organization of the human muscle chloride channel CIC-1 and analysis of novel mutations leading to Becker-type myotonia. Hum. Mol. Genet. 3, 941±946.

52

R. RuÈdel

Ludewig, U., Pusch, M., Jentsch, T.J., 1996. Two physically distinct pores in the dimeric ClC-0 chloride channel. Nature 383, 340±343. Mertens, H.G., Zschocke, S., 1965. Neuromyotonie. Klin. Wochenschr. 43, 917. Middleton, R.E., Pheasant, D.J., Miller, C., 1994. Puri®cation, reconstitution, and subunit composition of a voltage-gated chloride channel from Torpedo electroplax. Biochemistry 33, 13189±13198. Middleton, R.E., Pheasant, D.J., Miller, C., 1996. Homodimeric architecture of a ClC-type chloride ion channel. Nature 383, 337±340. Nakanishi, T., Sugita, H., Shimada, Y., Toyokura, Y., 1975. Neuromyotonia. A mild case. J. Neurol. Sci. 26, 599±604. Nakanishi, T., Shimada, Y., Sakuta, M., Toyokura, Y., 1978. The initial positive component of the scalprecorded somatosensory evoked potential in normal subjects and in patients with neurological disorders. Electroenceph. Clin. Neurophysiol. 45, 26±34. Negri, S., Caraceni, T., Boiardi, A., 1977. Neuromyotonia. Report of a case. Eur. Neurol. 16, 35±41. Palade, P.T., Barchi, R.L., 1977. On the inhibition of muscle membrane chloride conductance by aromatic carboxylic acids. J. Gen. Physiol. 69, 879±896. Partanen, V.S., Soininen, H., Saksa, M., Riekkinen, P., 1980. Electromyographic and nerve conduction ®ndings in a patient with neuromyotonia, normocalcemic tetany and small-cell lung cancer. Acta Neurol. Scand. 61, 216±226. Pusch, M., Jentsch, T.J., 1994. Molecular physiology of voltage-gated chloride channels. Physiol. Rev. 74, 813±827. Pusch, M., Steinmeyer, K., Jentsch, T.J., 1994. Low single channel conductance of the major skeletal muscle chloride channel. ClC-1. Biophys. J. 66, 149±152. Pusch, M., Steinmeyer, K., Koch, M.C., Jentsch, T.J., 1995. Mutations in dominant human myotonia congenita drastically alter the voltage dependence of the CIC-1 chloride channel. Neuron 15, 1455± 1463. Rosenbohm, A., RuÈdel, R., Fahlke, C., 1999. Regulation of the human skeletal muscle chloride channel hClC-1 by protein kinase C. J. Physiol. Lond. 514, 677±685. RuÈdel, R., 1990. The myotonic mouse - a realistic model for the study of human recessive generalized myotonia. Trends Neurosci. 13, 1±3. RuÈdel, R., Dengler, R., Ricker, K., Haass, A., Emser, W., 1980. Improved therapy of myotonia with the lidocaine derivative tocainide. J. Neurol. 222, 275±278. RuÈdel, R., Ricker, K., Lehmann-Horn, F., 1988. Transient weakness and altered membrane characteristic in recessive generalized myotonia (Becker). Muscle Nerve 11, 202±211. RuÈdel, R., Lehmann-Horn, F., Ricker, K. (1994). Altered excitability of the muscle cell membrane. In: Engel, A.G., Franzini-Armstrong, C. (Eds.). Myology, 2nd ed., Vol. 2, pp. 1291±1302 Rychkov, G., Astill, D.S., Bennetts, B., Hughes, B.P., Bretag, A.H., Roberts, M.L., 1997. pH-dependent interactions of Cd 21 and a carboxylate blocker with the rat ClC-1 chloride channel and its R304E mutant in the Sf-9 insect cell line. J. Physiol. Lond. 501, 355±362. Schmidt-Rose, T., Jentsch, T.J., 1997a. Reconstitution of functional voltage-gated chloride channels from complementary fragments of ClC-1. J. Biol. Chem. 272, 20515±20521. Schmidt-Rose, T., Jentsch, T.J., 1997b. Transmembrane topology of a ClC chloride channel. Proc. Natl. Acad. Sci. USA 94, 7633±7638. È ber das klinische und anatomische Bild des Muskelschwundes der Myotoniker. Steinert, H., 1909. U Dtsch. Z. Nervenheilkd. 37, 38. Steinmeyer, K., Klocke, R., Ortland, C., Gronemeier, M., Jockusch, H., GruÈnder, S., Jentsch, T.J., 1991a. Inactivation of muscle chloride channel by transposon insertion in myotonic mice. Nature 354, 304± 308. Steinmeyer, K., Ortland, C., Jentsch, T.J., 1991b. Primary structure and functional expression of a developmentally regulated skeletal muscle chloride channel. Nature 354, 301±304.

Muscle chloride channelopathies

53

Steinmeyer, K., Lorenz, C., Pusch, M., Koch, M.C., Jentsch, T.J., 1994. Multimeric structure of ClC-1 chloride channel revealed by mutations in dominant myotonia congenita (Thomsen). EMBO J. 13, 737±743. StruÈmpell, A., 1891. Tonische KraÈmpfe in willkuÈrlich bewegten Muskeln (Myotonia congenita). Berl. Klin. Wochenschr. 9, 119. Thomsen, J., 1876. Tonische KraÈmpfe in willkuÈrlich beweglichen Muskeln in Folge von ererbter psychischer Disposition. Arch. Psychiatr. Nervenkrankh. 6, 702. Wagner, S., Deymeer, F., KuÈrz, L.L., Benz, S., Schleithoff, L., Lehmann-Horn, F., Serdaroglu, P., È zdemir, C., RuÈdel, R., 1998. The dominant chloride channel mutant G200R causing ¯uctuating O myotonia: clinical ®ndings, electrophysiology, and channel pathology. Muscle Nerve 21, 1122±1128. White, G.R., Plaskett, J., 1904. `Nervous', `stiff-legged', or `fainting' goats. Am. Vet. Rev. 28, 556±560.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

55

Chapter 3

Molecular aspects of malignant hyperthermia and central core disease Patrick J. Lynch, Tommie V. McCarthy Department of Biochemistry, University College Cork, Cork, Ireland

Abstract Malignant hyperthermia (MH) is a pharmacogenetic disorder of skeletal muscle that is manifest in response to anesthetic triggering agents. Central core disease (CCD) is a closely related muscle weakness disorder. Both MH and CCD are primarily disorders of calcium regulation abnormalities in skeletal muscle. MH susceptible (MHS) individuals are healthy and triggering of MH is avoidable if the susceptibility status of an individual is known. Consequently, the ultimate goal driving MH related research is the development of a simple non-invasive test for routine diagnosis of MHS. To date, this goal has not yet been realized and although discoveries at the genetic level are promising with respect to development of gene-based diagnosis for MHS, more extensive genetic analysis in humans is necessary to address the key questions raised by molecular genetic analysis in MH families. Molecular genetic linkage studies implicate the ryanodine receptor gene (RYR1) in more than 50% of MH cases and in all CCD cases tested to date. This suggests that MH is a very heterogeneous disorder. However, relatively extensive genetic analysis has only identi®ed two families where MH can be con®rmed to be caused by a gene other than RYR1. Screening of the RYR1 gene in MH and CCD has identi®ed many mutations to date for these disorders. Functional characterization of several MHS and CCD mutations in cultured cells has provided valuable insights into the molecular etiology of these disorders and suggests that the spectrum of phenotype outcomes seen in MHS and CCD re¯ects the effect of RYR1 mutations on the size of the releasable calcium stores in muscle, the resting calcium concentration and the level of compensation achieved with respect to maintenance of calcium homeostasis. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Malignant hyperthermia (MH) is an unusual and enigmatic pharmacogenetic disorder of skeletal muscle that is manifest in response to anesthetic triggering

56

P.J. Lynch, T.V. McCarthy

agents. It is one of the major causes of death due to anesthesia. The ®rst cases of MH were reported at the start of this century and were often described in the literature as `late ether convulsions'. The familial aspects of the disorder were ®rst described by Denborough in 1962 in a large family where several individuals died during anesthesia (Denborough et al., 1962). MH reactions can be triggered by the use of non-halogenated anesthetics such as ether, ethylene and cyclopropane and halogenated anesthetics such as halothane, iso¯urane, des¯urane, en¯urane and sevo¯urane (Ellis and Heffron, 1985). The chance of a crisis is increased by the administration of depolarizing muscle relaxants, such as succinylcholine and suxamethonium, after halothane inhalation. Anesthetic agents considered safe for individuals with MH include nitrous oxide, barbiturates, propofol and ketamine. Surprisingly, a phenotypically similar disorder exists in swine. In susceptible swine, porcine malignant hyperthermia or porcine stress syndrome (PSS) can be triggered by halogenated anesthetics as well as by stress (Gronert, 1996). Stress induced triggering of the syndrome in swine is prevalent in animals homozygous for the susceptible locus and has a greatly reduced penetrance (the proportion of subjects of a particular genotype that manifest its phenotype in a given environment) in heterozygous animals. Homozygous affected individuals are rare in the human population and the role of stress and heat in triggering MH in human heterozygotes is controversial. A MH frequency of one per 50 000 to one per 100 000 anesthetics in adults and one per 15 000 anesthetics in children has been estimated (Britt and Kalow, 1970). The increased incidence of MH in children may re¯ect the reduced penetrance of MH with increasing age.

Clinical presentation Both swine and people respond to certain anesthetic agents and drugs with a striking increase in aerobic and anaerobic metabolism that results in increased production of heat, carbon dioxide and lactate. MH can be triggered by any potent volatile agent but the onset is usually more abrupt when succinylcholine is used. Once initiated a vicious cycle is established and a fulminant syndrome evolves in which the body temperature may exceed 438C (1098F). The arterial blood carbon dioxide tension may rise above 100 mm Hg and the arterial blood pH can fall below 7.0. This picture is generally accompanied by tachycardia and other signs of circulatory and metabolic stress. Almost all swine and about 75% of people show muscle rigidity during acute MH, which is caused by a contracture rather than a contraction of muscles. The rigidity may progress directly to rigor and death. Active MH results in increased permeability of muscle that causes increased serum potassium, ionized calcium, creatine kinase, myoglobin and sodium; and their is marked edema. Excessive release of myoglobin from muscle can result in gross myoglobinuria. There is

Molecular aspects of MH and CCD

57

probably eventual metabolic exhaustion with a more generalized increase in muscle permeability. If the MH episode is not fatal, the muscles during recovery may be edematous and tender and the creatine kinase returns to normal in 10±15 days (Gronert, 1986). A MH reaction can vary from the severe `classical/fulminant' case to more mild forms of the condition. The severity of a reaction varies greatly and depends on a number of known and unknown factors including age of patient, the concentration of triggering agent(s) used and the time elapsed before treatment of the disorder begins (Britt and Kalow, 1970). Many susceptible patients present with some, but not all, of the classical signs with variable intensity when exposed to agents known to induce the condition. Some known susceptible patients have had previous triggering anesthetics without any complications. Once an individual is known to be susceptible to MH, triggering anesthetics can be avoided. Thus, the diagnosis of MH susceptibility is of prime importance as early clinical diagnosis of MH is often dif®cult. MH treatment The chances of surviving a MH reaction is approximately 90±95%. This progress in treatment from earlier times where mortality was as high as 80% is attributed to an increased awareness about the symptoms of MH as well as the introduction of a clinical antidote, dantrolene sodium. This lipid-soluble hydantoin derivative was ®rst used as a therapy in porcine MH. Dantrolene inhibits halothane and caffeine induced muscle contractures and normalizes myoplasmic calcium concentrations in MHS muscle cells treated with triggering agents (Mickelson and Louis, 1996). The ryanodine receptor (RyR1) protein resides in the sarcoplasmic reticulum (SR) and has been identi®ed as the key calcium release channel involved muscle excitation± contraction coupling. Diagnosis of susceptibility to malignant hyperthermia The identi®cation of MH susceptible individuals prior to the administration of anesthetic agents by a non-invasive, sensitive and speci®c test is a major goal of MH research. If an individual is known to be MHS, the triggering of a MH reaction can be circumvented by the use of alternative anesthetics and non-depolarizing muscle relaxants. Susceptibility to MH is diagnosed by observing the magnitude of contractures induced in strips of muscle tissue in vitro by caffeine and halothane. Two standardized versions of the IVCT have been established ± the European MH group (EMHG) protocol and the North American MH Registry (NAMHR) protocol (European Malignant Hyperthermia Group, 1984; Larach et al., 1987).

58

P.J. Lynch, T.V. McCarthy

A standardized European in vitro contracture test (IVCT) was established in 1984 (European Malignant Hyperthermia Group, 1984). It allows the following diagnoses to be made: MH susceptible (MHS), MH equivocal (MHE) and MH normal (MHN). If a muscle biopsy strip produces a sustained increase in muscle tension of 0.2 g at 2% halothane or less and independently at a caffeine concentration of 2 mM or less the patient is considered MHS. MHN is diagnosed if the 0.2 g threshold is not attained at these concentrations while MHE is diagnosed if the threshold tension is attained with caffeine (MHE(c)) or with halothane (MHE(h)) but not with both. The clinical and biochemical relationship between the MHS and MHE phenotype is unclear. The North American protocol is relatively similar to the European protocol in that the response of the muscle to caffeine and halothane is measured and allows diagnosis of MHS and MHN. However, the diagnostic criteria are not highly standardized and there is no equivocal result (Larach et al., 1987). Other contracture tests such as the ryanodine contracture test and a test using 4-chloro-m-cresol (a speci®c activator of the RyR1) show promising results but are not yet widely accepted.

Sensitivity and speci®city of the IVCT Sensitivity in the IVCT is de®ned as the percentage of positive test results in the diseased population and is calculated from the formula: 100£ (true positives/(true positives 1 false negatives)). Speci®city is de®ned as the percentage of negative test results in the absence of the disease and is calculated from the formula: 100£ (true negatives/(true negatives 1 false positives)). Because failure to detect MHS individuals can result in a serious or fatal outcome, sensitivity approaching 100% is more important for clinical diagnosis than speci®city. As a result, the diagnosis of MHS is considered to err on the side of false positive diagnosis which reduces speci®city. Analysis of individuals tested by the American protocol revealed a sensitivity of 92±97% and a speci®city of 53±78% (Allen et al., 1998). Analysis of the European protocol's test results, carried out on 105 individuals with previous fulminant MH and 202 normal subjects, produced a sensitivity value of 99% and a speci®city value of 93.6% (Ording et al., 1997). However, false negative IVCT results were reported for four patients despite clinical evidence of MH susceptibility (Isaacs and Badenhorst, 1993). The IVCT has high value as a clinical test, however, the lack of 100% speci®city in the IVCT has serious implications for genetic analysis as even one false diagnosis in a family will produce a recombinant individual for the RYR1 gene and wrongly implies genetic heterogeneity. By increasing the cut-off values in the IVCT, higher speci®city is achieved but at the risk of reducing sensitivity. Raising the cut-off value in the North American 3% halothane IVCT from 0.5 to 0.7 g of tension

Molecular aspects of MH and CCD

59

increased speci®city from 78 to 81% but sensitivity dropped from 97 to 88%. Altering the IVCT cut-off points permitted investigators to link MH families to the RYR1 gene which were previously unlinked when conventional cut-off points were applied (MacKenzie et al., 1991; Healy et al., 1996; Serfas et al., 1996). These studies emphasize that arbitrary cut-off points, which were calculated empirically from analysis of a relatively small population, may be inadequate for genetic investigations in every family. Inheritance and penetrance Segregation studies on Irish and European MH families indicate that MHS is inherited as an autosomal dominant trait with apparently complete penetrance. By contrast, the penetrance of clinical MH in susceptible individuals is dif®cult to calculate as not all susceptible individuals are exposed to MH trigger agents. However, many susceptible patients have undergone anesthesia with trigger agents without any complications suggesting that the penetrance of the MH trait is low. In one study of 18 families where two or more relatives suffered a clinical episode of MH, six parents were also affected giving an estimated apparent penetrance of 6=18 ˆ 0:33. In 32 solitary MH cases in different families, MH reactions were not reported in any other relatives. Thus, in this case, if MHS is an autosomal dominant trait, the penetrance would be less than 0.033 (Kalow, 1987). Factors affecting penetrance are unclear and further studies will be necessary to solve this issue and to determine if the actual incidence of MHS in the population is far greater than previously estimated. Association of MH with other human diseases The only major disorder which consistently exhibits a MHS phenotype is central core disease (CCD). Apart from CCD, the majority of patients identi®ed as MHS do not appear to have any other associated disorder. However, in a small minority of cases, anesthesia-related complications similar to MH have been reported for patients with King±Denborough syndrome, myotonias and periodic paralyses, Duchenne and Becker muscular dystrophies, malignant neuroleptic syndrome and sudden infant death syndrome. Central core disease CCD is an inherited myopathy that was ®rst described by Shy and Magee in 1956 and which is closely associated with MH (Shuaib et al., 1987). The disorder exhibits

60

P.J. Lynch, T.V. McCarthy

great variability both clinically and histologically and the severity of symptoms may vary from normal to signi®cant muscle weakness. The clinical course of the disorder can be slow or non-progressive and progressive weakness leading to immobility is uncommon. Muscle weakness of the lower extremities is frequently the leading complication and the clinical course of the disorder can be slow or non-progressive. The main dif®culties are in running, rising from sitting, or climbing stairs. Mild to moderate muscle atrophy has frequently been reported. Increases in serum creatine kinase levels are unusual but elevated levels have been noted in some adult CCD patients. CCD is unusual in that the affected patients bene®t substantially from muscle exercise. Diagnosis of CCD is by histology of skeletal muscle biopsy samples and typical CCD samples show amorphous central areas (cores) in type 1 ®bers. In CCD there is a predominance of type I ®bers expressing the slow twitch calcium-ATPase pump. The molecular nature of the progressive disappearance of type II ®bers expressing the fast twitch and slow twitch calcium-ATPase pump is currently unknown. Identi®cation of cores is facilitated by staining for oxidative enzyme activity. As the cores are depleted of mitochondria they appear as negative areas within the normal enzyme activity areas of the surrounding muscle ®ber. Fiber type predominance is usually present and type I ®bers are usually in the majority. Electron microscopic investigations of CCD muscle tissue have demonstrated: (a) amorphous central areas (cores) in type 1 ®bers with a relative lack, if not complete absence, of mitochondria in core regions, (b) less numerous glycogen granules, (c) less well de®ned myo®brils and loss of myo®bril alignment in adjacent sarcomeres, (d) contracted sarcomeres, (e) Z disc streaming, and (f) pathological changes in the SR and t-tubules in both core and non-core regions (Hayashi et al., 1989). The absence of mitochondria per se in central core regions may not contribute signi®cantly to muscle weakness since there does not appear to be a direct relationship between the extent of the central cores and the clinical severity of the disease. Patients with CCD are at risk for MH and in almost all cases and in one study Shuaib et al. (1987) reached the conclusion that `all patients with central core disease should be considered at risk for malignant hyperthermia unless in vitro contracture tests show that the particular patient is free of the trait'.

MH genetics The recognition of the close similarities between porcine and human MH reactions prompted the use of the swine as an animal model for human MH. The gene for the MH trait in swine is referred to as halothane sensitivity (Hal) and is tightly linked to muscle leanness in the animals. Muscle from the homozygous affected animals is signi®cantly more hypertrophic than the heterozygous

Molecular aspects of MH and CCD

61

animals which in turn is more hypertrophic than the normal animals. Thus, the presence of the gene apparently contributes directly to muscle hypertrophy. Earlier biochemical studies on MH indicate that both human MH, and porcine MH, are disorders of regulation of the intracellular free calcium concentration of skeletal muscle (Mickelson and Louis, 1996). In swine and to a lesser extent in human, these studies indicated that calcium release from the SR in skeletal muscle was abnormal. The key calcium release channel in SR was identi®ed in rabbit. The puri®cation and characterization of the channel was greatly aided by the observation that ryanodine, a plant alkaloid bound speci®cally to the channel. Consequently, the channel is commonly known as the ryanodine receptor (RyR1). Biochemical evidence that RyR1 is defective in MH has been provided by the demonstration that the binding of ryanodine and the gating properties of the channel are altered in skeletal muscle isolated from affected animals. The cDNA for the rabbit RYR1 gene was isolated in 1989 and encodes for a protein that has a subunit size of over 560 kDa (Takeshima et al., 1989). This protein forms an elaborate tetrameric structure that acts both as a calcium release channel and `foot' structure bridging the gap between the SR and the t-tubule in skeletal muscle. The RYR1 gene is one of the largest genes known and the RyR1 channel is the largest channel known to date (Mickelson and Louis, 1996). In 1977, genetic linkage between porcine MH and polymorphisms in the gene encoding GPI locus on pig chromosome 6 was established (Andresen and Jensen, 1977). In a genetic study using a series of markers for the syntenic region on human chromosome 19q12-13.2, linkage was demonstrated between markers in the GPI region and MHS in three large Irish pedigrees (McCarthy et al., 1990). In a separate genetic study, the RYR1 gene was mapped to chromosome19q13.1-13.2 and RYR1 polymorphic markers were shown to segregate with MHS in nine Canadian families (MacKenzie et al., 1990; MacLennan et al., 1990). The co-localization of the MHS and RYR1 loci to the q12-13.2 region of chromosome 19 with the known biochemical data identi®ed the RYR1 gene as the main candidate for MHS. This was con®rmed by sequence analysis of the porcine RYR1 gene on chromosome 6q12 which showed that a single point mutation is present in MH swine and results in an Arg to Cys substitution at position 615 (Fujii et al., 1991). Interestingly, the same mutation is present in all affected swine which indicates that the mutation arose through a founder effect, i.e. all swine with the mutation have a common ancestor, and the mutation was disseminated through the domestic swine population through selective breeding for muscle quality. A mechanism by which the Arg615Cys RYR1 mutation gives rise to lean, heavilymuscled swine has been proposed (MacLennan and Phillips, 1992). A hypersensitive skeletal muscle calcium release channel could stimulate spontaneous muscle contractions in an animal and the resulting repetitive toning of the muscle could lead to muscle hypertrophy. The greater energy utilization would limit fat deposition.

62

P.J. Lynch, T.V. McCarthy

Mutation screening of the RYR1 gene in MHS affected individuals has to date led to the identi®cation of more than 20 RYR1 mutations (Table 1). The most common mutation identi®ed to date is the Gly341Arg mutation which accounts for about 10% of Caucasian MHS cases (Quane et al., 1994b) although reports indicate that the Arg614Cys (equivalent of the porcine Arg615Cys mutation) may be as common in parts of Europe (Deufel et al., 1995). Mutation screening of the RYR1 gene in a small number of CCD patients con®rmed that the RYR1 gene was indeed also mutated in CCD (Table 1). Mutations appear to cluster in two main regions of the RYR1 gene: a N-terminal region ranging from amino acid residues 35 to 614 (exons 2±17) and a mid region ranging from residues 2163 to 2458 (exons 39±46). However, the discovery of a Ile4898Thr mutation in a CCD kindred and a T4826I mutation in a large Maori pedigree suggests that the C-terminal region of the RyR1 may represent a novel mutation hot spot (Lynch et al., 1999; Brown et al., 2000). Further analysis of these three apparent `hot spots' in remaining MHS families is likely to yield more novel mutations, thus avoiding the laborious task of sequencing the entire 15 kb RYR1 cDNA in each individual. All three mutation hotspots are highly conserved across 14 ryanodine receptor molecules. It should be noted that since mutation screening strategies designed for RYR1 analysis have primarily been restricted to two regions of the RYR1 gene, the apparent hotspots mentioned may re¯ect a bias in mutation screening strategy rather than true mutation hotspots.

Genetic heterogeneity in malignant hyperthermia Genetic heterogeneity is not expected in MH since the disorder is essentially a disorder of calcium regulation and mutations in other proteins involved in calcium regulation might be expected to produce a similar phenotype to that seen with RYR1 mutations. While there is considerable genetic, biochemical and electrophysiological evidence to support the role of RYR1 mutations in MH, a number of nonchromosome 19 linked families have been reported (MacLennan, 1995). It is estimated that less than 50% of MHS families are linked to the RYR1 locus. The ®rst alternative MHS locus (MHS2) was assigned tentatively to chromosome 17q in North American families (Levitt et al., 1992). The validity of the chromosome 17 linkage has been questioned and even though considerable effort was made in Europe (Iles et al., 1993; Sudbrak et al., 1993), the result could not be replicated. One possible reason for the lack of replication other than validity of the statistical approach used is that diagnosis of MHS in the North American families is essentially different than that performed in Europe. Thus, the chromosome 17 linkage may be identifying a MHS population that is phenotypically different from the European MHS phenotype. Comparison of European and North American malignant

Table 1 Mutations in the human RYR1 gene in MHS and CCD

DNA substitution

MH/CCD

Incidence

Reference

Cys35Arg Arg163Cys Gly248Arg Gly341Arg Ile403Met Tyr522Ser Arg552Trp Arg614Cys Arg614Leu Arg2163Cys Arg2163His Val2168Met Thr2206Met Thr2206Arg Gly2434Arg

T103C (TGC ! CGC) C487T (CGC ! TGC) G742A (GGG ! AGG) G1021A (GGG ! AGG) C1209G (ATC ! ATG) A1565C (TAT ! TCT) C1654T (CGG ! TGG) C1840T (CGC ! TGC) G1841T (CGC ! CTC) C6487T (CGC ! TGC) G6488A (CGC ! CAC) G6496A (GTG ! ATG) C6617T (ACG ! ATG) C6617G (ACG ! AGG) G7300A (GGA ! AGA)

MH MH/CCD MH MH MH/CCD MH/CCD MH MH MH MH MH/CCD MH MH MH MH

One family 2% One family 6±10% One family One family One family 4% 2% 4% One family 8% One family One family 4±10%

Arg2435His Arg2435Leu Arg2454His Arg2454Cys Arg2458Cys Arg2458His Thr4826Ile Ile4898Thr

G7304A (CGC ! CAC) G7304T (CGC ! CTC) G7361A (CGC ! CAC) C73607 (CGC ! TGC) C7372T (CGC ! TGC) G7373A (CGC ! CAC) C14477T (ACC ! ATC) T14693C (ATT ! ACT)

MH/CCD MH/CCD MH MH MH MH MH CCD

One family One individual One family One family 4% 4% One family One family

Lynch et al. (1997) Quane et al. (1993) Gillard et al. (1992) Quane et al. (1994b) Quane et al. (1993) Quane et al. (1994a) Keating et al. (1997) Gillard et al. (1991) Quane et al. (1997) Manning et al. (1998b) Manning et al. (1998b) Manning et al. (1998b) Manning et al. (1998b) Brandt et al. (1999) Keating et al. (1994); Phillips et al. (1994) Zhang et al. (1993) Barone et al. (1999) Barone et al. (1999) Brandt et al. (1999) Manning et al. (1998a) Manning et al. (1998a) Brown et al. (2000) Lynch et al. (1999)

Molecular aspects of MH and CCD

Mutation

63

64

P.J. Lynch, T.V. McCarthy

hyperthermia diagnostic protocol outcomes for use in genetic studies has been reported and shows that the European protocol is more accurate for use in genetic studies largely due to inclusion of a MHE category in IVCT diagnosis and the exclusion of the MHE phenotype in genetic studies (Fletcher et al., 1999). Several large, apparently non-chromosome 19 linked European MH families have been investigated for linkage to loci other than RYR1. Since RyR1 is closely associated with the multi-subunit skeletal muscle DHP receptor voltage gated calcium channel, the genes for the individual DHP receptor subunits have been considered as candidates for MHS. The CACNL2A gene on chromosome 7q, encodes the a 2 and d -subunits of the DHP receptor, has been tentatively linked to MHS in a single European family (Iles et al., 1994). A chromosome 7q polymorphic marker was found to co-segregate with the MHS trait through 11 meioses in the family generating a lod score of 2.91. However, despite extensive efforts, analysis of the CACNL2A gene has not led to the identi®cation of a causative mutation. A systematic linkage study using a set of polymorphic microsatellite markers covering the entire human genome has also been performed using several large, apparently non-chromosome 19 linked European MH families. The MHS trait was found to co-segregate with markers on chromosome 3q13.1 in a single large family generating a lod score of 3.22 (Sudbrak et al., 1995). Genome wide analysis in three European families identi®ed one new MHS loci on chromosome 1q and a tentative locus on chromosome 5p (Robinson et al., 1997). The CACNA1S gene that encodes the a 1-subunit of the DHP receptor maps to the chromosome 1q locus but no known candidate gene has been mapped to the chromosome 5p locus. Sequence analysis of the chromosome 1q linked family led to the identi®cation of a Gly3333Ala mutation in the CACNA1S gene, generating a lod score of 4.38 at a recombination fraction of 0.00 (Monnier et al., 1997). The mutation, which results in an Arg1086His substitution, is located in the loop between domains III and IV of the channel. Interestingly, mutations linked to hypokalemic periodic paralysis (HypokPP) have been reported in the S4 segments of domain II (Arg528His) and IV (Arg1239His and Arg1239Gly) (Jurkat-Rott et al., 1994; Ptacek et al., 1994). However, individuals with hypokPP usually do not test positive in the IVCT and no members of the chromosome 1q family displayed symptoms associated with hypokPP. The discovery of a MHS mutation in the CACNA1S gene suggests that there is a direct functional interaction between the III±IV loop of the DHP receptor a 1-subunit and RyR1 and, indeed, this interaction has been demonstrated (Leong and MacLennan, 1998). A summary of the MHS loci is presented in Table 2. One striking feature of this table is that apart from linkage of MHS to RYR1, linkage to other loci has only been con®rmed in two families in total. Thus, despite the early estimation of 50% heterogeneity, de®nite evidence of linkage to alternative MHS loci in apparently nonchromosome 19 linked families has proved dif®cult to obtain. One of the critical

Table 2 Summary of human MHS loci

Incidence

Possible candidate gene(s)

Mutations

MHS1 19q13.1 MHS2 17q11.2-q24

~50% 31% (not replicated in European families)

RYR1 SCN4A, CACNLB1, CACNLG

See Table 1

MHS3 7q MHS4 3q13.1 MHS5 1q

One family (not con®rmed) a One family One family

CACNL2A ? CACNA1S

MH56 5p

One family (not con®rmed) a

?

a

Arg1086His

References Levitt et al. (1992); Olckers et al. (1992); Moslehi et al. (1998) Iles et al. (1994) Sudbrak et al. (1995) Robinson et al. (1997); Monnier et al. (1997) Robinson et al. (1997)

Molecular aspects of MH and CCD

Locus name and location

A lod score of $3.0 was not generated.

65

66

P.J. Lynch, T.V. McCarthy

problems in MH genetics is that a single incorrect IVCT diagnosis in a family can destroy linkage to the RYR1 gene and, so, falsely suggest the presence of a second MHS locus. It is likely that the estimation of 50% heterogeneity is a considerable underestimate due to single (or more) recombinants arising from a false positive or false negative IVCT diagnosis. In a linkage analysis study of 20 large, well de®ned MHS families, nine families were linked to chromosome 19q13.1, eight families contained a single recombinant, suggesting a phenotype misdiagnosis, and three families were clearly unlinked to chromosome 19q13.1 (Robinson et al., 1998). Assuming that a single recombinant is not suf®cient to exclude linkage between RYR1 and MHS, this study is perhaps a more accurate re¯ection of the degree of heterogeneity in MH and suggests that more than 85% of MHS families may arise from RYR1 mutations. A further consideration is that given a speci®city of diagnosis of MHS of 93.6% in Europe (Ording et al., 1997), one would expect a low number of MHS pedigrees with two or more false diagnosis in European families implying that some families that have been considered to be de®nitely non-RYR1 linked on the basis of two recombination events are in fact actually RYR1 linked. Examples of these types of families have already been reported (MacKenzie et al., 1991; Healy et al., 1996; Serfas et al., 1996) Thus, rather than measuring heterogeneity in MHS by exclusion of linkage, it may be more appropriate to measure heterogeneity by the number of families demonstrating con®rmed linkage to alternate MHS loci. A separate and largely non-addressed issue impacting on MH heterogeneity concerns the possibility of families containing two RYR1 mutations on separate haplotypes. Such families will generally give the appearance of heterogeneity and screening for only one of the mutations will produce discordant individuals who appear to be false positives but actually harbor the second undetected mutation. Estimates of the incidence of MH in the population is one per 50 000 to one per 100 000 anesthetics in adults. Thus, the chances of ®nding a family with two separate RYR1 MHS alleles was considered to be low. However, as the penetrance of the MH trait in MHS individuals is likely to be low, the incidence of the MHS in the general population is probably much higher. The European Malignant Hyperthermia Group has acknowledged the reality of this possibility and recommend testing of both parents of a proband by IVCT. In further support of this prospect, there are now several families previously excluded from linkage to the RYR1 gene where a portion of MHS individuals in the pedigree carry a speci®c RYR1 mutation. There is no simple solution for solving heterogeneity questions in MHS. The invasive nature of the IVCT precludes retesting of recombinant individuals. The size and complexity of the RYR1 gene makes it refractory to large scale analysis using current mutation screening technology. Thus the answer to heterogeneity questions are likely to appear slowly as each avenue is explored in each RYR1 recombinant family in a systematic fashion.

Molecular aspects of MH and CCD

67

Guidelines for genetic testing in MH families Genetic testing in MH families is unusual in that a false negative diagnosis can potentially be fatal. A further complication is that the incidence of MHE diagnosis in the population is exceptionally high and the chances of an individual diagnosed as MHN by genetic means has a high chance of being diagnosed as MHE by the IVCT. This creates signi®cant medico-legal problems and as such MHN diagnosis by genetic means alone is problematic. The EMHG has produced guidelines for genetic testing in MH families and recommends the following procedures: ² the proband's DNA should be tested for known RYR1 and DHP receptor mutations; ² if the proband tests negative for known MHS mutations, further genetic analysis may be appropriate. Once enough members of the proband's family have been phenotyped by the IVCT, genetic segregation to known MHS loci is recommended using DNA markers compiled by the Genetics Section of the EMHG.

In some large families, it will be possible to generate a lod score of 13.0 at u ˆ 0:0, thus establishing linkage to a MHS gene locus. In these families, it is possible to assign a MH status to individuals not tested by IVCT; individuals harboring the high-risk haplotype should be regarded as MH susceptible. However, individuals not containing the high-risk haplotype should not be considered MHN. These individuals must be IVCT tested to determine MH status. In families where linkage to a MHS locus is suggested but tentative because a lod score of 13.0 could not be achieved, then predictive testing based on haplotype analysis is not recommended. A single recombination between the haplotype and the IVCT result of an individual can eliminate linkage of a family to a MHS locus. If this occurs, the results of the genetic analysis and the IVCT should be scrutinized closely in an attempt to resolve the discordant result. For predictive diagnosis, the more cautious result, either MHS/MHE or inheritance of the high risk haplotype, must be taken into account for the clinical diagnosis. (For further details see http://www.emhg.org).

The ryanodine receptor Extensive studies have been carried out on RyR proteins and are the subject of a recent reviews (Coronado et al., 1994; Mickelson and Louis, 1996; Mackrill, 1999). At present, 14 RYR molecules have been cloned and sequenced from nine species including RYR2 and RYR3 isoforms. The 158 kb human RYR1 gene contains 106 exons and produces a 15kb mRNA that encodes a 5038 amino acid polypeptide. One practical consequence of complexity of the RYR1 gene at the genomic level and the

68

P.J. Lynch, T.V. McCarthy

huge size of the RYR1 gene at the cDNA level is that mutation screening is a costly, arduous and time consuming task and explains why the number of different RYR1 mutations known in MH is relatively low by comparison with other disorders. The N-terminal domain RyR1 is believed to form the large myoplasmic `foot' structure, containing many of the ligand binding sites, while the C-terminal ®fth of the molecule is believed to form the transmembrane calcium channel. Numerous skeletal muscle proteins have been shown to bind directly to RyR1 or else modify channel behavior (Mackrill, 1999). These include the dihydropyridine (DHP) receptor, FKBP12, calsequestrin, triadin, calmodulin, glyceraldehyde 3phosphate dehydrogenase, aldolase, phosphoglucomutase, annexin VI, 170 kDa low-density lipoprotein binding protein, 30 kDa DIDS binding protein, sarcalumenin, histidine-rich calcium binding protein and S-100 protein. A 158-amino acid sequence consisting of Arg954-Asp1112 was identi®ed as a region of interaction between RyR1 and the II±III and III±IV loops of DHP receptor a 1-subunit (Leong and MacLennan, 1998). A second area for EC-coupling was identi®ed as the sequence encompassed by amino acid residues Thr1303 and Leu1406 (Yamazawa et al., 1997). Deletion of this sequence preserved the function of the channel but resulted in the loss of EC-coupling. This region, which is known as the D2 region, is highly divergent between RyR1 and RyR2, and the corresponding sequence is absent in the RyR3 isoform. In addition to `receiving' the EC-coupling signal from the DHP receptor, RyR1 also `transmits' a retrograde signal that enhances the calcium channel activity of the DHP receptor. By expressing chimeric RyR1/RyR2 molecules in dyspedic myotubes, the region responsible for the retrograde signal was localized to the RyR1 residues 2659±3720 (Nakai et al., 1998). The activating and inhibiting effects of calcium on RyR1 channel activity have been attributed to high and low af®nity calcium binding sites, respectively, and a number of these sites have been identi®ed (Mickelson and Louis, 1996). RyR1 is also inhibited by Mg 21. No consensus sequences have yet been de®ned for Mg 21 binding sites but the inhibitory effect of Mg 21 may be due to the cation acting as a competitive inhibitor of calcium for activating calcium binding sites. (Coronado et al., 1994; Mickelson and Louis, 1996; Mackrill, 1999).

Exogenous RyR1 modulator binding sites The activity of RyR1 is modulated by a variety of exogenous ligands including ryanodine, caffeine, halothane and dantrolene sodium (Coronado et al., 1994; Zucchi and Ronca-Testoni, 1997). The plant alkaloid, ryanodine, modulates RyR1 in a biphasic manner with low concentrations (nanomolar) activating the channel and high concentrations (.100 mM) inhibiting the channel. Both high and low af®nity ryanodine binding sites are located between Arg4475 and the C-terminus

Molecular aspects of MH and CCD

69

of the RyR1 (Callaway et al., 1994). Caffeine, a trimethylxanthine, activates RyR1 calcium release by increasing the apparent af®nity of the calcium activation site for calcium (Meissner et al., 1997). By expressing the full length and C-terminal portion of the RyR1 in CHO cells, the caffeine binding site was localized to the N-terminal foot region of the RyR1 protein (Bhat et al., 1997). The major effect of clinical concentrations of halothane on skeletal muscle appears to be a stimulation of SR calcium release but the exact molecular mechanism by which halothane stimulates calcium release from RyR1 is unknown. However, halothane induces intracellular calcium release in HEK-293 cells transfected with RYR1 cDNA but is unable to elicit a response in non-transfected cells (Tong et al., 1997). The skeletal muscle relaxant dantrolene sodium inhibits the release of calcium from the SR during ECcoupling and suppresses the uncontrolled calcium release that underlies MH. Dantrolene acts directly on RyR1, possibly by limiting channel activation by calmodulin and calcium (Fruen et al., 1997). One possibility of a mechanistic role for MH and CCD mutations is that such mutations might directly affect RyR1 binding sites. However, there is no evidence to date that MH and CCD RYR1 mutations have a direct consequence on known RyR1 binding sites (Fig. 1). A plausible mechanistic possibility suggested from work using a monoclonal antibody directed against the Gly341 region of the one protein (Zorzato et al., 1996) is that alterations of the RyR1 channel activity due to MH mutations may be due to perturbation of RyR1 intramolecular interactions.

Functional expression of the RYR1 cDNA Functional expression of the RYR1 cDNA offers the prospect of comprehensive evaluation of MHS and CCD mutations on the RyR1 channel calcium homeostasis and on the response of the protein to halothane and caffeine. The cloned rabbit RYR1 cDNA has been successfully expressed in Chinese hamster ovary (CHO) cells and yields a protein indistinguishable from puri®ed RyR1 receptor in immunoreactivity, molecular size and [ 3H]ryanodine binding (Takeshima et al., 1989). Both caffeine (1±50 mM) and ryanodine (100 mM) induced release of calcium from intracellular stores of transformed CHO cells but not from control (non-transfected) CHO cells. Single channel recordings of RyR1 channels expressed in COS-1 cells demonstrated that the recombinant protein responded to calcium, ATP, Mg 21, ruthenium red and ryanodine in a similar way as the native channel (Chen et al., 1993). However, some differences, such as the presence of subconductance states and an untypical slow gating mode were found in expressed channels. A cellular calcium photometry assay was described that allows caffeine and halothane-induced calcium release to be measured in cultured cells transfected with a mammalian expression vector containing a rabbit RYR1 cDNA (Otsu et al., 1994; Treves et al., 1994; Tong

70

P.J. Lynch, T.V. McCarthy

Fig. 1. Location of RYR1 MH and CCD mutations with respect to known RYR1 binding sites.

et al., 1997). The single channel properties of recombinant calcium release channels expressed in HEK-293 cells were found to be virtually identical to the native channel (Chen et al., 1997). This system has been exploited successfully for functional expression and analysis of several RyR1 mutant channels.

Expression and functional analysis of RYR1 mutants The porcine mutation, Arg615Cys, was the ®rst mutant RyR1 molecule to be expressed and analyzed functionally by cellular calcium photometry. C2C12 cells expressing the mutant Cys615 RyR1 showed higher sensitivity to caffeine when compared to cells expressing the wild-type channel (Otsu et al., 1994). Exposure of mutant transfected cells to clinical doses of halothane resulted in a rapid increase in the intracellular calcium concentration, whereas no calcium changes were observed in cells expressing the wild-type RyR1. Furthermore, COS-1 cells expressing the Cys615 mutant were more sensitive to the RyR1 agonist, 4-chloro-m-cresol, when compared to cells expressing the wild-type channel (Treves et al., 1994). In both C2C12 and COS-1 cells, there was no difference in the resting cellular calcium concentration between wild-type and mutant transfected cells (Otsu et al., 1994; Treves et al., 1994). Analysis of 15 MH/CCD RYR1 mutant channels expressed in HEK cells showed that 14 out of the 15 mutant channels had signi®cantly reduced ED50 values for

Molecular aspects of MH and CCD

71

caffeine activation and all had signi®cantly reduced ED50 values for halothane activation when compared to the wild-type channel value (Tong et al., 1997). Thus, all RYR1 mutations investigated produced functional calcium release channels that were hypersensitive to the actions of caffeine and halothane. This supports the view that the agonist hypersensitivity of mutant MHS RyR1 molecules is responsible for abnormal contractures of MHS skeletal muscle in the IVCT. The validity of the in vitro expression system and calcium photometry assay was con®rmed when comparison of the caffeine responses from the cellular calcium photometry assay and the caffeine IVCT results from MHS individuals heterozygous for the mutations was performed. Comparison of the `cell culture assay system' with IVCT data collected from patients across Europe showed a highly signi®cant correlation with the caffeine response (Tong et al., 1997). The high degree of signi®cance of the correlation was unexpected given that the IVCT data is based on a muscle response whereas the `cell culture assay system' uses cells of kidney origin. Furthermore the IVCT data was collected from different laboratories, performed over a long period of time and individuals tested were from a diverse background. The highly signi®cant correlation between the `cell culture assay system' and the IVCT for the caffeine response validates the exploration of HEK cells transfected with mutant RYR1 genes as a potential assay system for MHS. However, no signi®cant correlation was observed between the halothane photometry and halothane IVCT results. This is not surprising given the lack of correlation observed between halothane threshold and tension values in the IVCT (Manning et al., 1998b). Interestingly, a signi®cant correlation was observed between the caffeine and halothane ED50 for the mutant channels expressed in the HEK-293 cells whereas no correlation was observed for the equivalent analysis of the IVCT, i.e. caffeine threshold versus halothane threshold in the IVCT (Tong et al., 1997; Manning et al., 1998b). These data suggest that experimental deviations in the halothane responses in the IVCT is the most likely explanation for the lack of the correlation and indicate that halothane hypersensitivity may be determined more accurately in the cellular calcium photometry assay than in the IVCT. In a follow-up paper, relatively extensive analysis on RYR1 transfected HEK-293 cells was performed (Tong et al., 1999) and showed the following: ² The expression of wild-type RyR1 increases the size of the calcium store and increases both content and activity of the calcium ATPase (SERCA2b). These results can be explained on the basis of an increased permeability of the endoplasmic reticulum (ER) calcium store which is compensated for by increased synthesis of SERCA2b, with a consequent enlargement of the calcium store. ² In cells expressing homotetrameric wild-type or mutant RyR1 channels, the amplitude of 10 mM caffeine-induced calcium release was correlated signi®cantly with the amplitude of carbachol- or thapsigargin-induced calcium release,

72

²

²

² ²

²

P.J. Lynch, T.V. McCarthy

indicating that maximal drug-induced calcium release depends on the size of the ER calcium store. Expression of MH or CCD mutant RyR1 enhances synthesis of SERCA2b, leading to a higher potential for calcium storage. This potential is not realized, however, because calcium ¯ux out of the store is higher than calcium ¯ux into the store and a lower ER calcium store results. Average resting cytosolic calcium concentration was elevated signi®cantly over wild-type and over MH mutant RyR1 proteins for the ®ve CCD mutant RyR1 proteins expressed in HEK-293 cells. However, although resting levels were higher for all CCD mutants, resting calcium levels were signi®cantly higher only for Tyr523Ser and Arg2163His mutants when analyzed independently. Maximal caffeine-induced calcium release, measured by both calcium photometry and calcium imaging, was lower in cells transfected with individual MH/CCD mutants than with wild-type RYR1. Caffeine ED50 values for calcium release through MH/CCD mutant proteins were linearly correlated with maximal caffeine responses, and the maximal caffeineinduced calcium release was also linearly correlated with clinical caffeine IVCT thresholds, indicating that higher ER calcium stores inhibit caffeine responses. Cells transfected with a CCD mutant have the highest resting calcium concentration and the lowest ER calcium store.

Although the conclusions drawn were obtained from transfection of RYR1 into non-muscle cells, they are entirely consistent with the single report of over-expression of a RYR1 CCD (R163C) mutant in muscle cells obtained from MHN individuals (Censier et al., 1998) Very few families with clinically severe CCD have been analyzed extensively. One such kindred with an unusually severe and highly penetrant form of the disorder has been investigated and the identi®ed RYR1 mutation functionally expressed and analyzed in HEK-293 cells (Lynch et al., 1999). The I4898T mutation identi®ed in this kindred, unlike other CCD mutations is in the C-terminus transmembrane/ lumenal region of the RyR1 protein and results in an Ile4898Thr substitution in a highly conserved region of the 5038 amino acid protein. Functional expression and analysis of this mutation in HEK-293 cells has shown it to be an exceptionally severe mutation resulting in an exceptionally leaky channel with perturbed ryanodine binding sites. Comparison with other co-expressed mutant/normal channels showed that the Thr4898 mutation produces the most abnormal MH/CCD RyR1 channels yet investigated and this level of abnormality is re¯ected in the severe and penetrant phenotype of affected CCD individuals in the kindred. Taken together, the functional analysis of MH/CCD RyR1 mutants has yielded several valuable insights into the etiology of MH and CCD: overall, the ®ndings suggest that CCD and MH mutant channels are more leaky than wild-type RyR1

Molecular aspects of MH and CCD

73

channels; CCD mutants have the most enhanced permeability and the extra compensatory mechanisms brought into play to restore calcium homeostasis can explain the shift in expression of the calcium-ATPase seen in CCD cases; the results are consistent with the view that RyR1 mutants can result in a spectrum of phenotype outcomes ranging from muscle hypertrophy to muscle atrophy and that these outcomes are a re¯ection of the sensitivity of the RYR1 mutant proteins to agonists, the size of the releasable calcium store, the resting calcium concentration and the level of compensation achieved by the muscle with respect to maintaining calcium homeostasis.

Model for MH and CCD Consideration of the various ®ndings for MH and CCD allow construction of a tentative model for the two related disorders MH and CCD (Figs. 2 and 3). MH The MH RyR2 channel in the presence of agonist releases calcium at an enhanced rate, increasing the resting myoplasmic calcium concentration with a consequential decrease in lumenal calcium. This accounts for the increase in baseline tension observed in MH muscle in the IVCT. The released calcium further stimulates RyR1 channel ef¯ux by calcium-induced calcium release (CICR). The consequential reduced SR lumenal calcium concentration exacerbates the situation as decreased lumenal calcium levels increase the sensitivity of RyR1 to agonists (Tong et al., 1999). Calcium stimulates ATP synthesis by activating phosphorylase kinase, which is responsible for the breakdown of glycogen to glucose-1-phosphate. Glucose-1-phosphate is then converted to lactic acid and ATP by glycolysis. Aerobic metabolism is also stimulated by calcium during a MH episode as two enzymes in the mitochondria, isocitrate dehydrogenase and 2-oxoglutarate dehydrogenase, are activated resulting in the synthesis of ATP, CO2 and heat. The sustained muscle contraction, along with the actions of the SR calcium ATPases deplete the muscle cell of ATP resulting in the inability of the cell to reduce the myoplasmic calcium concentrations and rigor resulting from the inability of the myo®brils to relax. The high temperature and low pH cause cell membrane damage resulting in the leakage of K 1, Mg 21, myoglobin and creatine kinase into the blood stream. These initiate secondary MH responses in other tissues: the hyperkalaemia resulting from the K 1 release induces cardiac arrhythmias and kidney failure arises from precipitation of myoglobin protein in the basement membrane of the glomerular ®ltration barrier.

74

P.J. Lynch, T.V. McCarthy

Fig. 2. A tentative model for MH.

CCD In MHS cases RyR1 channels are likely to function within the normal range in the absence of external agonists such as halothane with respect to maintenance of calcium homeostasis given that compensation for leaky channels up to a certain level is likely to occur. By contrast, RyR1 channels in CCD cases are likely to have a leakage level in the absence of external agonists that requires the induction of major compensatory mechanisms. The level of compensation required results in change of the ®ber type II to predominantly ®ber type I. The functional analysis of mutant RyR1 channels in single HEK-293 cells has shown us three consequences of this abnormal channel gating: reduced SR lumenal calcium stores, reduced myoplasmic

Molecular aspects of MH and CCD

75

Fig. 3. A tentative model for CCD.

calcium transients and increased myoplasmic calcium concentration (Lynch et al., 1999). The reduced calcium concentration in the lumen of the SR may make the RyR1 channel more sensitive to activation by agonists. The increased myoplasmic calcium concentration could lead to many of the characteristics that are typical of CCD muscle histology. A calcium gradient could be established within the cell with the highest concentrations occurring in the center of the cell and lowest concentrations occurring at the cell periphery due to the enhancement of the calcium removal system by plasma membrane pumps and exchangers (MacLennan, 1996). The centrally elevated calcium concentration could lead to signi®cant differential force being established between centrally and peripherally positioned myo®brils resulting in Z-disc streaming and targeting of the damaged central area for degradation leading to core formation. The elevated myoplasmic calcium concentration

76

P.J. Lynch, T.V. McCarthy

could also lead to the excessive activation of the calcium dependent neutral proteases (calpains). The increased calpain activity would lead to proteolysis of myo®bril proteins, which could enhance Z-disc streaming. Furthermore, RyR1 is a substrate for calpain. Cleavage of the RyR1 protein would lead to disruption of the triadic regions resulting in muscle weakness. Calcium and calpains are also involved in the activation of the skeletal muscle glycogen phosphorylase kinase causing increased glycogen degradation and would account for reduction of glycogen granules observed in CCD muscle. Disappearance of mitochondria from core regions may result from calcium overload in the organelle as it attempts to buffer myoplasmic calcium concentrations. Another explanation may be that the mitochondria are destroyed by the increased activity of calpains and calcium-dependent phospholipases. One key inconsistency of a calcium-stimulated degration model is that in such a model exercise might be expected to have a harmful effect. By contrast, CCD is unusual in that exercise is bene®cial to affected individuals. A further consequence of the abnormal gating of a CCD RyR1 channel mutation may be a reduction in the myoplasmic calcium transient during EC-coupling. These reduced calcium transients could have three effects on skeletal muscle function. First, muscle weakness is likely as the contractile force generated in EC-coupling will be compromised due to the reduced calcium transients (which presumably re¯ect a reduced SR lumenal calcium concentration. Second, the amplitude of intracellular calcium transients is likely to have important effects on muscle maintenance as it is known that changes in amplitude can modulate the activity of CaM kinase II (De Koninck and Schulman, 1998) and regulate the expression of calciumdependent transcription factors such as NF-AT and NF-kB. Therefore a reduction in the amplitude of the calcium transient in CCD could result in a reduced muscle maintenance signal that would manifest as atrophy. Thirdly, speci®c blockage of RyR1 calcium release disrupts myosin thick ®lament assembly (Ferrari et al., 1998); thus, reduced RyR1 calcium transients could result in myo®bril degradation and Zdisc streaming. These latter two possibilities are particularly attractive and would explain why exercise is of signi®cant bene®t to individuals who suffer from CCD as exercise would increase the frequency and amplitude of such a signal. One extrapolation of this model is that strenuous regular exercise should be of more bene®t to CCD patients than mild exercise. Perspectives One of the most unusual aspects of RYR1 mutations in MH and CCD is that the phenotype at one end of the MH scale may be expressed as muscle hypertrophy positively re¯ecting cell growth, cell adaptation and cell maintenance while at the other end of the scale, the phenotype is expressed as severe muscle atrophy re¯ecting cell degeneration. Despite the variation in phenotypes resulting from the muta-

Molecular aspects of MH and CCD

77

tions, the reports to date argue in favor of the same mechanism giving rise to the different phenotypes, namely that the RYR1 mutations result in `leaky' RyR1 channels and that different levels of `leakiness' are responsible for the different spectrum of phenotypes. The spectrum of phenotype outcomes is likely to re¯ect the size of the releasable calcium store, the resting calcium concentration and the level of compensation achieved with respect to maintenance of calcium homeostasis. The simplest explanation for these ®ndings is that variable levels of `leakiness' of the RyR1 channel alters the calcium signaling and the subsequent cellular interpretation of the signal. The consequence is a cellular response that ranges from hypertrophy to atrophy depending on the level of `leakiness'. Further exploration of the effect of alterations of the size of the releasable calcium stores and the resting calcium concentration on cell models with respect to cellular interpretation, cellular response and cellular integration of calcium signals may yield fruitful insights into the subtleties of MH and CCD. MHS individuals are healthy and triggering of MH is avoidable if the susceptibility status of an individual is known. Consequently, the ultimate goal driving MH related research is the development of a simple non-invasive test for routine diagnosis of susceptibility to MH. While early discoveries at the genetic level were promising with respect to development of RYR1 gene based diagnosis for MHS, more extensive genetic analysis in humans has shown gene based diagnosis is currently problematic apart from a relatively small number of large families where RYR1 mutations and/or linked markers clearly segregate with the susceptibility phenotype. In contrast to the human situation, RYR1 gene based diagnosis for MHS in pigs is one of the most commercially successful gene based tests for an inherited disorder to date. This success is due to the fact that MHS in pigs is due to the same single founder RYR1 mutation in all commercial pig breeding lines (Fujii et al., 1991). The key problems hindering the development of a genetic based test for MHS in humans is heterogeneity and discordance between described causal mutations and MHS/MHN phenotype. The key question concerns the extent to which apparent genetic heterogeneity and causal mutation/phenotype discordance in MHS re¯ect inaccurate IVCT phenotyping versus the extent to which true genetic heterogeneity and true discordance exists in MHS. Since the speci®city and sensitivity of IVCT based MHS diagnosis is less than ,100%, apparent heterogeneity and causal mutation/phenotype discordance will be detected. A further predicted consequence of IVCT phenotyping that is not 100% accurate is that one would expect a decreasing number of MHS pedigrees with one, two, three or more false diagnosis, respectively, and indeed this prediction is consistent with the MH genetics literature. Consequently, estimation of genetic heterogeneity in MHS may be more accurately measured by the number of families demonstrating con®rmed linkage rather than exclusion to alternate MHS loci. If this approach is applied, then RYR1 accounts for

78

P.J. Lynch, T.V. McCarthy

all cases of MHS to date apart from a single family demonstrating linkage between MHS and markers on chromosome 3q13.1 and a single family bearing a mutation in the CACNA1S on chromsome 1 (Table 2). A further consideration for the heterogeneity question concerns the MHS phenotype as diagnosed by the European protocol versus the North American protocol. Comparison of European and North American malignant hyperthermia diagnostic protocol outcomes for use in genetic studies has been reported and shows that the European protocol is more accurate for use in genetic studies largely due to inclusion of a MHE category in IVCT diagnosis and the exclusion of the MHE phenotype in genetic studies (Fletcher et al., 1999). However, it is quite possible that the differences in the protocols result in detection of different phenotypes and are differently weighted with respect to identi®cation of modulating gene effects. Therefore, even though the chromosome 17 linkage reported for families analyzed by the North American protocol (Levitt et al., 1992) has not been reproduced in Europe, it may be identifying a locus conferring a phenotype that is only detectable by the North American protocol. Given that the IVCT phenotype is assigned on the basis of increased contracture in response to caffeine and halothane, the underlying mechanism must involve alterations of proteins that directly or indirectly in¯uence calcium release from the SR. Thus, the scope for heterogeneity in MHS is sizeable. However, the real extent of true heterogeneity in MHS remains to be determined and will require signi®cant efforts at the molecular genetic level. Many cases of discordance between well substantiated causal RYR1 mutations and MHS phenotype have been reported. There are likely to be several explanations for these ®ndings. On one hand, the IVCT as the `gold standard' has been a successful phenotyping tool in that it has permitted identi®cation of three con®rmed MHS loci to date. On the other hand, the IVCT is a relatively crude test by nature that has been derived empirically based on the response of normal and affected individuals of diverse genetic background and as such resolution of the IVCT is limited. A further factor that limits the resolution of the IVCT is that diagnosis is based on exceeding a threshold value rather than interpretation of a dose response curve (European Malignant Hyperthermia Group, 1984). A further explanation that has not been fully explored concerns the frequency of MHS in the population. If the penetrance of clinical MH in MHS individuals is low, then the frequency of MHS in the population may be considerably higher than previously expected. Investigation of 202 low-risk subjects showed that three were MHS, ®ve were MHEh, ®ve MHEc, and 189 MHN (Ording et al., 1997). It is possible that these data re¯ect a high level of true MHS in the population. A direct consequence of a high population incidence of MHS would be the occurrence of some families where MHS is due to mutations on different halplotypes. Linkage analysis in families in this category would exhibit exclusion to known MHS loci.

Molecular aspects of MH and CCD

79

Mutation analysis would show discordance between known MHS mutations and the MHS phenotype. A number of families consistent with this category have been reported (Deufel et al., 1995; Adeokun et al., 1997). However, the RYR1 gene has not been analyzed extensively in families in this category. If MHS proves to have a high incidence in the population, then the prospect of screening the RYR1 gene in discordant families must be viewed seriously. While the guidelines for genetic testing in MH families are useful, additional serious questions have been raised by the studies to date that need to be addressed. The key questions include the following: ² If a presenting proband carries a known RYR1 MHS mutation, does the proband need to be tested using the IVCT and should relatives not carrying the mutations be tested using the IVCT to verify MHN status? ² Is it appropriate to perform the IVCT on individuals from a MH family that do not carry a segregating MHS mutation to verify MHN status when the a priori risk of MHS or MHE diagnosis is high (6.5%)? ² If it is accepted that the IVCT has a sensitivity of less than 100%, should false positive and false negative IVCT diagnosis be reassigned? ² The MHE phenotype occurs at a relatively high but variable frequency across Europe and is an important diagnostic category. Yet, very little is understood about the factors governing the MHE phenotype. Major insights into such factors will have a very valuable impact on the direction that genetic testing for MH takes in the future.

In conclusion, the ultimate goal of a simple non-invasive test for routine diagnosis of susceptibility to MH has still not been achieved. The prospect of this elusive goal becoming a reality will require signi®cant re®nement of our understanding of this enigmatic calcium regulation disorder at the genetic, biochemical, physiological, epidemiological and clinical level. Acknowledgements This work was funded by the EU BIOMED program. References Adeokun, A.M., West, S.P., Ellis, F.R., Halsall, P.J., Hopkins, P.M., Foroughmand, A.M., Iles, D.E., Robinson, R.L., Stewart, A.D., Curran, J.L., 1997. The G1021A substitution in the RYR1 gene does not cosegregate with malignant hyperthermia susceptibility in a British pedigree. Am. J. Hum. Genet. 60, 833±841. Allen, G.C., Larach, M.G., Kunselman, A.R., 1998. The sensitivity and speci®city of the caffeine-

80

P.J. Lynch, T.V. McCarthy

halothane contracture test: a report from the North American Malignant Hyperthermia Registry. The North American Malignant Hyperthermia Registry of MHAUS. Anesthesiology 88, 579±588. Andresen, E., Jensen, P., 1977. Close linkage established between the HAL locus for halothane sensitivity and the PHI (phosphohexose isomerase) locus in pigs of the Danish Landrace breed. Nord. Vet. Med. 29, 502±504. Barone, V., Massa, O., Intravaia, E., Bracco, A., Di Martino, A., Tegazzin, V., Cozzolino, S., Sorrentino, V., 1999. Mutation screening of the RYR1 gene and identi®cation of two novel mutations in Italian malignant hyperthermia families. J. Med. Genet. 36, 115±118. Bhat, M.B., Zhao, J., Zang, W., Balke, C.W., Takeshima, H., Wier, W.G., Ma, J., 1997. Caffeine-induced release of intracellular calcium from Chinese hamster ovary cells expressing skeletal muscle ryanodine receptor. Effects on full-length and carboxyl-terminal portion of calcium release channels. J. Gen. Physiol. 110, 749±762. Brandt. A., Schleithoff, L., Jurkat-Rott, K., Klingler, W., Baur, C. Lehmann-Horn, F., 1999. Screening of the ryanodine receptor gene in 105 malignant hyperthermia families: novel mutations and concorndance with the in vitro contracture test. Hum. Mol. Genet. 8, 2055±2062. Britt, B.A., Kalow, W., 1970. Malignant hyperthermia: a statistical review. Can. Anaesth. Soc. J. 17, 293± 315. Brown, R.L., Pollock, A.N., Couchman, K.G., Hodges, M., Waaka, R., Lynch, P., McCarthy, T.V., Stowell, K.M., 2000. Novel ryanodine receptor mutation and genotype±phenotype correlation in a large malignant hyperthermia New Zealand Maori pedigree. Hum. Mol. Genet. in press. Callaway, C., Seryshev, A., Wang, J.P., Slavik, K.J., Needleman, D.H., Cantu III, C., Wu, Y., Jayaraman, T., Marks, A.R., Hamilton, S.L., 1994. Localization of the high and low af®nity [ 3H]ryanodine binding sites on the skeletal muscle calcium release channel. J. Biol. Chem. 269, 15876±15884. Censier, K., Urwyler, A., Zorzato, F., Treves, S., 1998. Intracellular calcium homeostasis in human primary muscle cells from malignant hyperthermia-susceptible and normal individuals. Effect Of overexpression of recombinant wild-type and Arg163Cys mutated ryanodine receptors. J. Clin. Invest. 101, 1233±1242. Chen, S.R.W., Vaughan, D.M., Airey, J.A., Coronado, R., MacLennan, D.H., 1993. Functional expression of cDNA encoding the calcium release channel (ryanodine receptor) of rabbit skeletal muscle sarcoplasmic reticulum in COS-1 cells. Biochemistry 32, 3743±3753. Chen, S.R.W., Leong, P., Imredy, J.P., Barlett, C., Zhang, L., MacLennan, D.H., 1997. Single-channel properties of the recombinant skeletal muscle calcium release channel (ryanodine receptor). Biophys. J. 73, 1904±1912. Coronado, R., Morrissette, J., Sukhareva, M., Vaughan, D.M., 1994. Structure and function of ryanodine receptors. Am. J. Physiol. 266, C1485±C1504. De Koninck, P., Schulman, H., 1998. Sensitivity of CaM kinase II to the frequency of calcium oscillations. Science 279, 227±230. Denborough, M.A., Forster, J.F.A., Lovell, R.R.H., Maplestone, P.A., Villiers, J.D., 1962. Anaesthetic deaths in a family. Br. J. Anaesth. 34, 395. Deufel, T., Sudbrak, R., Feist, Y., Rubsam, B., Du Chesne, I., Schafer, K.L., Roewer, N., Grimm, T., Lehmann-Horn, F., Hartung, E.J., Muller, C.R., 1995. Discordance, in a malignant hyperthermia pedigree, between in vitro contracture-test phenotypes and haplotypes for the MHS1 region on chromosome 19q12-13.2, comprising the C1840T transition in the RYR1 gene. Am. J. Hum. Genet. 56, 1334±1342. Ellis, F.R., Heffron, J.J.A., 1985. Clinical and biochemical aspects of malignant hyperthermia. In: Akinson, R.S., Adams, A.P. (Eds.). Recent Advances in Anesthesia and Analgesia. Churchill-Livingstone, New York, pp. 173±207. European Malignant Hyperthermia Group (1984). A protocol for the investigation of malignant hyperpyrexia (MH) susceptibility. Br. J. Anaesth. 56, 1267-1269.

Molecular aspects of MH and CCD

81

Ferrari, M.B., Ribbeck, K., Hagler, D.J., Spitzer, N.C., 1998. A calcium signaling cascade essential for myosin thick ®lament assembly in xenopus myocytes. J. Cell Biol. 141, 1349±1356. Fletcher, J.E., Rosenberg, H., Aggarwal, M., 1999. Comparison of European and North American malignanthyperthermia diagnostic protocol outcomes for use in genetic studies. Anesthesiology 90, 654± 661. Fruen, B.R., Mickelson, J.R., Louis, C.F., 1997. Dantrolene inhibition of sarcoplasmic reticulum calcium release by direct and speci®c action at skeletal muscle ryanodine receptors. J. Biol. Chem. 272, 26965±26971. Fujii, J., Otsu, K., Zorzato, F., De Leon, S., Khanna, V.K., Weiler, J.E., O'Brien, P.J., MacLennan, D.H., 1991. Identi®cation of a mutation in porcine ryanodine receptor associated with malignant hyperthermia. Science 253, 448±451. Gillard, E.F., Otsu, K., Fujii, J., Kanna, V.K., DeLeon, S., Derdemezi, J., Britt, B.A., Duff, C.L., Worton, R.G., MacLennan, D.H., 1991. A substitution of cysteine for arginine 614 in the ryanodine receptor is potentially causative of human malignant hyperthermia. Genomics 11, 751±755. Gillard, E.F., Otsu, K., Fujii, J., Duff, C.L., DeLeon, S., Khanna, V.K., Britt, B.A., Worton, R.G., MacLennan, D.H., 1992. Polymorphisms and deduced amino acid substitutions in the coding sequence of the ryanodine receptor (RYR1) gene in individuals with malignant hyperthermia. Genomics 13, 1247±1254. Gronert, G.A., 1986. Malignant hyperthermia. In: Engel, A.G., Banker, B.Q. (Eds.). Myology. McGrawHill, New York, pp. 1763±1784. Gronert, G.A., 1996. Porcine malignant hyperthermia. In: Morio, M., Kikuchi, H., Yuge, O. (Eds.). Malignant Hyperthermia: Proceedings of the 3rd International Symposium on Malignant Hyperthermia. Springer-Verlag, Tokyo, pp. 159±161. Hayashi, K., Miller, R.G., Brownell, K.W., 1989. Central core disease: ultrastructure of the sarcoplasmic reticulum and the t-tubules. Muscle Nerve 12, 95±102. Healy, J.M.S., Quane, K.A., Keating, K.E., Lehane, M., Heffron, J.J.A., McCarthy, T.V., 1996. Diagnosis of malignant hyperthermia: a comparison of the in vitro contracture test with the molecular genetic diagnosis in a large pedigree. J. Med. Genet. 33, 18±24. Iles, D.E., Segers, B., Sengers, R.C.A., Monsieurs, K., Heytens, L., Halsall, P.J., Hopkins, P.M., Ellis, F.R., Hall-Curran, J.L., Stewart, A.D., Wieringa, B., 1993. Genetic mapping of the b1- and g-subunits of the human skeletal muscle L-type voltage-dependent calcium channel on chromosome 17q and exclusion as candidate genes for malignant hyperthermia susceptibility. Hum. Mol. Genet. 2, 863±868. Iles, D.E., Lehmann-Horn, F., Scherer, S.W., Tsui, L.C., Olde Weghuis, D., Suijkerbuijk, R.F., Heytens, L., Mikala, G., Schwartz, A., Ellis, F.R., Stewart, A.D., Deufel, T., Wieringa, B., 1994. Localisation of the gene encoding the a2/d subunits of the L-type voltage-dependent calcium channel to chromosome 7q and analysis of the segregation of ¯anking markers in malignant hyperthermia susceptible families. Hum. Mol. Genet. 3, 969±975. Isaacs, H., Badenhorst, M., 1993. False-negative results with the muscle caffeine halothane contracture testing for malignant hyperthermia. Anesthesiology 79, 5±9. Jurkat-Rott, K., Lehmann-Horn, F., Elbaz, A., Heine, R., Gregg, R.G., Hogan, K., Powers, P.A., Lapie, P., Vale-Santos, J.E., Weissenbach, J., Fontaine, B., 1994. A calcium channel mutation causing hypokalemic periodic paralysis. Hum. Mol. Genet. 3, 1415±1419. Kalow, W., 1987. In: inheritance of malignant hyperthermia ± a review of published data. Britt, B.A. (Ed.). Malignant Hyperthermia. Martinus-Nijhoff, Amsterdam, pp. 155±179. Keating, K.E., Quane, K.A., Manning, B.M., Lehane, M., Hartung, E., Censier, K., Urwyler, A., Klausnitzer, M., Muller, C.R., Heffron, J.J.A., McCarthy, T.V., 1994. Detection of a novel RYR1 mutation in four malignant hyperthermia pedigrees. Hum. Mol. Genet. 3, 1855±1858. Keating, K.E., Giblin, L., Lynch, P.J., Quane, K.A., Lehane, M., Heffron, J.J.A., McCarthy, T.V., 1997. Detection of a novel mutation in the ryanodine receptor gene in an Irish malignant hyperthermia

82

P.J. Lynch, T.V. McCarthy

pedigree: correlation of the IVCT response with the affected and unaffected haplotypes. J. Med. Genet. 34, 291±296. Larach, M., Rosenberg, H., Larach, D.R., Broennle, A.M., 1987. Prediction of malignant hyperthermia susceptibility by clinical signs. Anesthesiology 66, 547±550. Leong, P., MacLennan, D.H., 1998. The cytoplasmic loops between Domains II and III and Domains III and IV in the skeletal muscle dihydropyridine receptor bind to a contiguous site in the skeletal muscle ryanodine receptor. J. Biol. Chem. 273, 29958±29964. Levitt, R.C., Olckers, A., Meyers, S., Fletcher, J.E., Rosenberg, H., Isaacs, H., Meyers, D.A., 1992. Evidence for the localization of a malignant hyperthermia susceptibility locus (MHS2) to human chromosome 17q. Genomics 14, 562±566. Loke, J., MacLennan, D.H., 1998. Malignant hyperthermia and central core disease: disorders of calcium release channels. Am. J. Med. 104, 470±486. Lynch, P.J., Krivosic-Horber, R., Reyford, H., Monnier, N., Quane, K., Adnet, P., Haudecoeur, G., Krivosic, I., McCarthy, T., Lunardi, J., 1997. Identi®cation of heterozygous and homozygous individuals with the novel RYR1 mutation Cys35Arg in a large kindred. Anesthesiology 86, 620±626. Lynch, P.J., Tong, J., Lehane, M., Mallet, A., Giblin, L., Heffron, J.J.A., Vaughan, P., Zafra, G., MacLennan, D.H., McCarthy, T.V., 1999. A mutation in the transmembrane/lumenal domain of the ryanodine receptor is associated with abnormal calcium release channel function and severe central core disease. Proc. Natl. Acad. Sci. U.S.A. 96, 4164±4169. MacKenzie, A.E., Korneluk, R.G., Zorzato, F., Fujii, J., Phillips, M., Iles, D., Wieringa, B., Leblond, S., Bailly, J., Willard, H.F., Duff, C., Worton, R.G., MacLennan, D.H., 1990. The human ryanodine receptor gene: its mapping to 19q13.1, placement in a chromosome 19 linkage group, and exclusion as the gene causing myotonic dystrophy. Am. J. Hum. Genet. 46, 1082±1089. MacKenzie, A.E., Allen, G., Lahey, D., Crossan, M.L., Nolan, K., Mettler, G., Worton, R.G., MacLennan, D.H., Korneluk, R., 1991. A comparison of the caffeine halothane muscle contracture test with the molecular genetic diagnosis of malignant hyperthermia. Anesthesiology 75, 4±8. Mackrill, J.J., 1999. Protein-protein interactions in intracellular calcium-release channel function. Biochem. J. 337, 345±361. MacLennan, D.H., 1995. Discordance between phenotype and genotype in malignant hyperthermia. Curr. Opin. Neurol. 8, 397±401. MacLennan, D.H., 1996. Mutations in the skeletal muscle ryanodine receptor (RYR1) gene are linked to malignant hyperthermia and central-core disease. In: Morio, M., Kikuchi, H., Yuge, O. (Eds.). Malignant Hyperthermia: Proceedings of the 3rd International Symposium on Malignant Hyperthermia, Springer-Verlag., Tokyo, pp. 79±86. MacLennan, D.H., Phillips, M.S., 1992. Malignant hyperthermia. Science 256, 789±794. MacLennan, D.H., Duff, C., Zorzato, F., Fujii, J., Phillips, M., Korneluk, R.G., Frodis, W., Britt, B.A., Worton, R.G., 1990. Ryanodine receptor gene is a candidate for predisposition to malignant hyperthermia. Nature 343, 559±561. Manning, B., Quane, K., Lynch, P.J., Urwyler, A., Tegazzin, V., Krivosic-Horber, R., Censier, K., Comi, G., Adnet, P., Wolz, W., Muller, C., McCarthy, T., 1998. Novel mutations at a CpG dinucleotide in the ryanodine receptor in malignant hyperthermia. Hum. Mutat. 11, 45±50. Manning, B.M., Quane, K.A., Ording, H., Urwyler, A., Tegazzin, V., Lehane, M., O'Halloran, J., Hartung, E., Giblin, L.M., Lynch, P.J., Vaughan, P., Censier, K., Bendixen, D., Comi, G., Heytens, L., Monsieurs, K., Fagerlund, T., Wolz, W., Heffron, J.J., Muller, C.R., McCarthy, T.V., 1998. Identi®cation of novel mutations in the ryanodine-receptor gene (RYR1) in malignant hyperthermia: genotype± phenotype correlation. Am. J. Hum. Genet. 62, 599±609. McCarthy, T.V., Healy, J.M.S., Heffron, J.J.A., Lehane, M., Deufel, T., Lehmann-Horn, F., Farrall, M., Johnson, K., 1990. Localization of the malignant hyperthermia susceptibility locus to human chromosome 19q12-13.2. Nature 343, 562±564.

Molecular aspects of MH and CCD

83

Meissner, G., Rios, E., Tripathy, A., Pasek, D.A., 1997. Regulation of skeletal muscle calcium release channel (ryanodine receptor) by calcium and monovalent cations and anions. J. Biol. Chem. 272, 1628±1638. Mickelson, J.R., Louis, C.F., 1996. Malignant hyperthermia: excitation-contraction coupling, calcium release channel, and cell calcium regulation defects. Physiol. Rev. 76, 537±592. Monnier, N., Procaccio, V., Stieglitz, P., Lunardi, J., 1997. Malignant-hyperthermia susceptibility is associated with a mutation of the alpha 1-subunit of the human dihydropyridine-sensitive L-type voltage dependent calcium-channel receptor in skeletal muscle. Am. J. Hum. Genet. 60, 1316± 1325. Moslehi, R., Langlois, S., Yam, I., Friedman, J.M., 1998. Linkage of malignant hyperthermia and hyperkalemic periodic paralysis to the adult skeletal muscle sodium channel (SCN4A) gene in a large pedigree. Am. J. Med. Genet. 76, 21±27. Nakai, J., Sekiguchi, N., Rando, T.A., Allen, P.D., Beam, K.G., 1998. Two regions of the ryanodine receptor involved in coupling with L-type calcium channels. J. Biol. Chem. 273, 13403±13406. Olckers, A., Meyers, D.A., Meyers, S., Taylor, E.W., Fletcher, J.E., Rosenberg, H., Isaacs, H., Levitt, R.C., 1992. Adult muscle sodium channel alpha-subunit is a gene candidate for malignant hyperthermia susceptibility. Genomics 14, 829±831. Ording, H., Brancadoro, V., Cozzolino, S., Ellis, F.R., Glauber, V., Gonano, E.F., Halsall, P.J., Hartung, E., Heffron, J.J., Heytens, L., Kozak-Ribbens, G., Kress, H., Krivosic-Horber, R., Lehmann-Horn, F., Mortier, W., Nivoche, Y., Ranklev-Twetman, E., Sigurdsson, S., Snoeck, M., Stieglitz, P., Tegazzin, V., Urwyler, A., Wappler, F., 1997. In vitro contracture test for diagnosis of malignant hyperthermia following the protocol of the European MH Group: results of testing patients surviving fulminant MH and unrelated low-risk subjects. The European Malignant Hyperthermia Group. Acta. Anaesthesiol. Scand. 41, 955±966. Otsu, K., Nishida, K., Kimura, Y., Hori, M., Kamada, T., Tada, M., 1994. The point mutation Arg615 ! Cys in the calcium release channel of skeletal sarcoplasmic reticulum is responsible for hypersensitivity to caffeine and halothane in malignant hyperthermia. J. Biol. Chem. 269, 9413±9415. Phillips, M.S., Khanna, V.K., DeLeon, S., Frodis, W., Britt, B.A., MacLennan, D.H., 1994. The substitution of Arg for Gly2433 in the human skeletal muscle ryanodine receptor is associated with malignant hyperthermia. Hum. Mol. Genet. 3, 2181±2186. Ptacek, L.J., Tawil, R., Griggs, R.C., Engel, A.G., Layzer, R.B., Kwiecinski, H., McManis, P.G., Santiago, L., Moore, M., Fouad, G., Bradley, P., Leppert, M.F., 1994. Dihydropyridine receptor mutations cause hypokalemic periodic paralysis. Cell 77, 863±868. Quane, K.A., Healy, J.M.S., Keating, K.A., Manning, B.M., Couch, F.J., Palmucci, L.M., Doriguzzi, C., Fagerlund, T.H., Berg, K., Ording, H., Bendixen, D., Mortier, W., Linz, U., Muller, C.R., McCarthy, T.V., 1993. Mutations in the ryanodine receptor gene in central core disease and malignant hyperthermia. Nat. Genet. 5, 51±55. Quane, K.A., Keating, K.E., Healy, J.M.S., Manning, B.M., Krivosic-Horber, R., Krivosic, I., Monnier, N., Lunardi, J., McCarthy, T.V., 1994. Mutation screening of the RYR1 gene in malignant hyperthermia: detection of a novel Tyr to Ser mutation in a pedigree with associated central cores. Genomics 23, 236±239. Quane, K.A., Keating, K.E., Manning, B.M., Healy, J.M.S., Monsieurs, K., Heffron, J.J.A., Lehane, M., Heytens, L., Krivosic-Horber, R., Adnet, P., Ellis, F.R., Monnier, N., Lunardi, J., McCarthy, T.V., 1994. Detection of a novel common mutation in the ryanodine receptor gene in malignant hyperthermia: implications for diagnosis and heterogeneity studies. Hum. Mol. Genet. 3, 471±476. Quane, K.A., Ording, H., Keating, K.E., Manning, B.M., Heine, R., Bendixen, D., Berg, K., KrivosicHorber, R., Lehmann-Horn, F., Fagerlund, T., McCarthy, T.V., 1997. Detection of a novel mutation at amino acid position 614 in the ryanodine receptor in malignant hyperthermia. Br. J. Anaest. 79, 332± 337.

84

P.J. Lynch, T.V. McCarthy

Robinson, R.L., Monnier, N., Wolz, W., Jung, M., Reis, A., Nuernberg, G., Curran, J.L., Monsieurs, K., Stieglitz, P., Heytens, L., Fricker, R., Van Broeckhoven, C., Deufel, T., Hopkins, P.M., Lunardi, J., Mueller, C.R., 1997. A genome wide search for susceptibility loci in three European malignant hyperthermia pedigrees. Hum. Mol. Genet. 6, 953±961. Robinson, R., Curran, J.L., Hall, W.J., Halsall, P.J., Hopkins, P.M., Markham, A.F., Stewart, A.D., West, S.P., Ellis, F.R., 1998. Genetic heterogeneity and HOMOG analysis in British malignant hyperthermia families. J. Med. Genet. 35, 196±201. Serfas, K.D., Bose, D., Patel, L., Wrogemann, K., Phillips, M.S., MacLennan, D.H., Greenberg, C.R., 1996. Comparison of the segregation of the RYR1 C1840T mutation with segregation of the caffeine/ halothane contracture test results for malignant hyperthermia susceptibility in a large Manitoba Mennonite family. Anesthesiology 84, 322±329. Shuaib, A., Paasuke, R.T., Brownell, K.W., 1987. Central core disease. Clinical features in 13 patients. Medicine 66, 389±396. Sudbrak, R., Golla, A., Hogan, K., Powers, P., Gregg, R., Du Chesne, I., Lehmann-Horn, F., Deufel, T., 1993. Exclusion of malignant hyperthermia susceptibility (MHS) from a putative MHS2 locus on chromosome 17q and of the a1, b1, and g subunits of the dihydropyridine receptor calcium channel as candidates for the molecular defect. Hum. Mol. Genet. 2, 857±862. Sudbrak, R., Procaccio, V., Klausnitzer, M., Curran, J.L., Monsieurs, K., Van Broeckhoven, C., Ellis, R., Heyetens, L., Hartung, E.J., Kozak-Ribbens, G., Heilinger, D., Weissenbach, J., LehmannHorn, F., Mueller, C.R., Deufel, T., Stewart, A.D., Lunardi, J., 1995. Mapping of a further malignant hyperthermia susceptibility locus to chromosome 3q13.1. Am. J. Hum. Genet. 56, 684±691. Takeshima, H., Nishimura, S., Matsumoto, T., Ishida, H., Kangawa, K., Minamino, N., Matsuo, H., Ueda, M., Hanaoka, M., Hirose, T., Numa, S., 1989. Primary structure and expression from complementary DNA of skeletal muscle ryanodine receptor. Nature 339, 439±445. Tong, J., Oyamada, H., Demaurex, N., Grinstein, S., McCarthy, T.V., MacLennan, D.H., 1997. Caffeine and halothane sensitivity of intracellular calcium release is altered by 15 calcium release channel (Ryanodine receptor) mutations associated with malignant hyperthermia and/or central core disease. J. Biol. Chem. 272, 26332±26339. Tong, J., McCarthy, T.V., MacLennan, D.H., 1999. Measurement of resting cytosolic calcium concentrations and calcium store size in HEK-293 cells transfected with malignant hyperthermia or central core disease mutant calcium release channels. J. Biol. Chem. 274, 693±702. Treves, S., Larini, F., Menegazzi, P., Steinberg, T.H., Koval, M., Vilsen, B., Andersen, J.P., Zorzato, F., 1994. Alteration of intracellular calcium transients in COS-7 cells transfected with the cDNA encoding skeletal-muscle ryanodine receptor carrying a mutation associated with malignant hyperthermia. Biochem. J. 301, 661±665. Yamazawa, T., Takeshima, H., Shimuta, M., Iino, M., 1997. A region of the ryanodine receptor critical for excitation-contraction coupling in skeletal muscle. J. Biol. Chem. 272, 8161±8164. Zhang, Y., Chen, H.S., Khanna, V.K., De Leon, S., Phillips, M.S., Schappert, K., Britt, B.A., Brownell, K.W., MacLennan, D.H., 1993. A mutation in the human ryanodine receptor gene associated with central core disease. Nat. Genet. 5, 46±50. Zorzato, F., Menegazzi, P., Treves, S., Ronjat, M., 1996. Role of malignant hyperthermia domain in the regulation of calcium release channel (ryanodine receptor) of skeletal muscle sarcoplasmic reticulum. J. Biol. Chem. 271, 22759±22763. Zucchi, R., Ronca-Testoni, S., 1997. The sarcoplasmic reticulum calcium channel/ryanodine receptor: modulation by endogenous effectors, drugs and disease states. Pharmocol. Rev. 49, 1±51.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

85

Chapter 4

Mutations affecting muscle nicotinic acetylcholine receptors and their role in congenital myasthenic syndromes David Beeson, John Newsom-Davis Neurosciences Group, Institute for Molecular Medicine, John Radcliffe Hospital, Headington, Oxford OX3 9DS, UK

Abstract Congenital myasthenic syndromes are a heterogeneous group of inherited disorders which have in common the defective transmission of the signal from nerve to muscle and fatigable muscle weakness. The principal syndromes involve mutations of the muscle acetylcholine receptor (AChR), a pentameric ligand-gated ion channel (a2bgd in foetal and denervated muscle, a2bd1 in adult muscle). Acetylcholine receptor de®ciency (autosomal recessive) presents in the ®rst 2 years of life, is generally non-progressive, responds to anti-cholinesterase medication, and in the great majority of patients is due to homozygous or heteroallelic mutations in the 1 subunit including its promoter region. Where these mutations lead to loss of function of this subunit, neuromuscular transmission is presumed to occur either through expression of the foetal form of the receptor or possibly through a2,b,d2 pentamers. The low-af®nity, fast channel syndrome (autosomal recessive) has a similar phenotype to AChR de®ciency but is much rarer. Mutations at different sites lead to fewer and shorter channel activations. The slow channel syndrome (autosomal dominant or sporadic), in contrast to the other disorders, presents in childhood, adolescence or adult life, does not respond to anticholinesterase, and is progressive. Mutations can occur in different subunits and in different functional domains, and result in a gain of function (prolongation of channel activations) that leads to subsynaptic damage. Acetylcholinesterase de®ciency (autosomal recessive) presents in infancy, is non-progressive but, like the slow channel syndrome, fails to respond to acetylcholinesterase medication. Mutations are in the ColQ gene which encodes the ColQ protein that attaches the asymmetric form of acetylcholinesterase to the basal lamina. q 2000 Elsevier Science B.V. All rights reserved.

86

D. Beeson, J. Newsom-Davis

Introduction Congenital myasthenic syndromes are a heterogeneous group of inherited disorders which have in common the defective transmission of the signal from nerve to muscle and consequent fatigable muscle weakness. They are rare compared to the autoimmune forms of myasthenia (myasthenia gravis and the Lambert±Eaton myasthenic syndrome). Nerve to muscle transmission depends principally on acetylcholine (ACh) release from the nerve terminal and its reaction with the muscle nicotinic acetylcholine receptor (AChR), a pentameric ligand-gated ion channel. This chapter will focus primarily on the disorders that result from the many mutations that have now been detected in the AChR subunit genes, but because AChR function is strongly in¯uenced by acetylcholinesterase (AChE), the enzyme that breaks down ACh, we will also brie¯y describe congenital myasthenic syndromes that arise from mutations affecting AChE. The principal clinical syndromes involving AChR mutations are AChR de®ciency (AChRdef), the slow channel syndrome (SCS) and the low-af®nity, fast channel syndrome (FCS), as listed in Table 1. Each of these can arise from mutations at a number of different sites, and can show distinctive clinical and electromyographic features. Endplate acetylcholinesterase de®ciency (EAD) results from mutations at various sites in the ColQ gene that encodes the collagen-like tail that attaches AChE to the basal laminar at the neuromuscular junction. Case histories of AChRdef and SCS are given, that serve to remind the reader that these congenital disorders are not always evident at birth. They are followed by a brief description of neuromuscular transmission. Illustrative case histories Acetylcholine receptor de®ciency caused by an 1 -subunit promoter mutation (Nichols et al., 1999) A 28-year-old man ®rst became aware of weakness of his legs at the age of 13 years, although family members noticed drooping of his eyelids at about 6 years of age. Weakness would develop progressively during exercise, and he would lose the ability to run. He occasionally experienced double vision. Later his arms would fatigue similarly, but his symptoms were generally only mildly progressive. The patient was the product of consanguineous parents in that they had great-grandfathers who were brothers. He had seven siblings, of whom one died in infancy, and another had had very similar symptoms to his since the age of 10 years (Fig. 1). Examination at the age of 16 years showed limitation of eye movements, fatigable ptosis and fatigable limb weakness, that responded unequivocally to intravenous edrophonium (an AChE inhibitor). Serum anti-acetylcholine receptor antibodies

Table 1 Some clinical features and results of investigations into congenital myasthenic syndromes

Inheritance a

Electromyography Morphological and and electrophysiology biochemical studies on biopsied muscle on biopsied muscle

Clinical picture

Treatment

Synaptic Acetylcholinesterase de®ciency

AR

Repetitive response to single nerve stimulus. MEPP decay phase prolonged

AChE absent or markedly reduced at neuromuscular junction

Onset in neonatal period

No effect of anticholinesterases

Postsynaptic Acetylcholine receptor de®ciency

AR

Decrement on stimulation at low frequency. MEPP amplitudes markedly reduced Repetitive response to single nerve stimulus. MEPP decay phase prolonged

Elongation of motor endplate. Reduced number of a-BuTx binding sites

Onset neonatal or during infancy

Anticholinesterases and 3,4-DAP are helpful

Slow-channel syndrome AD or Sporadic

Low- af®nity fastchannel syndrome

a

AR

Onset in childhood or adult. Slowly progressive with severe involvement of scapular and forearm muscles Onset in neonatal Decrement on Normal endplate stimulation at low morphology. Normal period numbers of a-BuTx frequency. MEPP amplitudes very small binding sites

Anticholinesterases and 3,4-DAP are detrimental. Quinidine-type drugs may be helpful Anticholinesterases produce some improvement

87

AR, autosomal recessive; AD, autosomal dominant.

Degenerative changes in postsynaptic folds and endplate cytoplasm

Congenital myasthenic syndromes

Syndrome

88

D. Beeson, J. Newsom-Davis

Fig. 1. Pedigree of the AChR de®ciency family. Shaded boxes indicate affected individuals.

were not detected and there was no improvement following thymectomy suggesting the absence of an autoimmune pathology. Examination at the age of 28 years revealed full eye movements apart from slight limitation of abduction, and generalized mild weakness that seemed very similar to that found 12 years earlier. Electromyography showed a 12% decrement in the compound muscle action potential amplitude at 3 Hz stimulation, and a highly signi®cant increase in mean jitter during single ®bre studies. Anti-AChR antibodies were again undetectable. Microelectrode electrophysiological recordings from biopsied intercostal muscle showed miniature endplate potentials that were abnormally small in the presence of eserine (an anticholinesterase preparation), and which were undetectable in its absence. Endplate acetylcholine receptor numbers were substantially reduced, and the endplates were elongated. The clinical and laboratory ®ndings were thus consistent with AChR de®ciency of relatively late onset. Screening for mutations within the coding sequence of the AChR 1-subunit gene did not detect any abnormalities; however a single point mutation (1-156C ! T) was identi®ed within the promoter region of this gene. This mutation cosegregates with the recessive inheritance of the disorder and disrupts a critical 6-bp promoter element known as the N-box. The N-box has been shown to be important in the control of synapse-speci®c transcriptional regulation in other species. Whereas mRNA for the a, b, g and d subunits could be readily detected in a muscle biopsy from this patient, mRNA for the 1 subunit was absent. Thus, the inherited myasthenic syndrome in this patient is due to a promoter mutation that causes a de®ciency of the adult form of the AChR

Congenital myasthenic syndromes

89

Slow channel syndrome (Oosterhuis et al., 1987) A 29-year-old woman developed arm weakness towards the end of her ®rst pregnancy. Weakness increased post-partum and she had transient dif®culty in swallowing. Fatigable weakness in her legs was evident in the following year and she fell on several occasions. Intermittent double vision and transient eyelid drooping were also experienced. Her parents were unrelated and, like her only brother, they had no neurological symptoms. Myasthenia gravis was suspected, and electromyography showed a characteristic `myasthenic' decrement (25%) in the compound muscle action potential amplitude at 3 Hz stimulation. However, she failed to respond to intravenous edrophonium. Serum anti-acetylcholine receptor antibodies were not detected. Thymectomy was undertaken, but her thymus was normal and she continued to deteriorate. About 6 months later, high dose (80 mg daily) prednisone treatment was begun, but again she failed to improve over the subsequent year. Examination at this time revealed wasting of the hand muscle with weakness present generally but particularly for ®nger and wrist extension. Further electromyographic studies showed a repetitive response to a single shock and increased jitter and blocking on single ®bre analysis. Microelectrode intracellular recordings from her biopsied intercostal muscle showed a highly signi®cant increase in the rise time of the miniature endplate potential and in its decay time constant, and endplate AChE staining was present, consistent with the diagnosis of the slow channel syndrome. Molecular genetic studies later demonstrated the presence of a single heterozygous missense mutation aS269I in the AChR a-subunit, located between the M2 and M3 transmembrane domains. Recordings of the effect of this mutation on AChR function demonstrate channel activations that are approximately 7-fold longer than normal (Croxen et al., 1997). The aS269I mutation causes a gain of function for the AChR ion channel. The prolonged AChR activations are thought to cause the damage to the muscle in the endplate region which underlies many of the clinical symptoms observed in this patient. Neuromuscular transmission When a nerve impulse reaches the nerve terminal the depolarization of the membrane is sensed by P-type voltage-gated calcium channels (VGCC) (Fig. 2). The VGCC open brie¯y and the in¯ux of Ca 21 ions triggers the near-simultaneous fusion of ACh-containing synaptic vesicles with the membrane i.e. the `quantal' release of ACh. ACh rapidly diffuses across the synaptic cleft. On reaching the postsynaptic membrane it binds to AChR causing the central ion pore to open. The in¯ux of cations, mainly Na 1, results in an endplate potential (EPP) that, in turn, activates voltage sensitive sodium channels (VGNaC) through which further

90

D. Beeson, J. Newsom-Davis

Fig. 2. Diagrammatic representation of the neuromuscular junction.

Na 1 ions enter, generating an action potential that is propagated along the length of the muscle ®bre leading to contraction. The action of ACh is terminated by its dissociation from the AChR and removal through diffusion and enzymatic inactivation by AChE. The amplitude of the EPP must reach a critical threshold to activate the VGNaC, and conditions that prevent it reaching this threshold result in a myasthenic disorder. The difference between the actual EPP amplitude and the threshold EPP amplitude required to trigger the muscle ®bre action potential is known as the `safety margin of neuromuscular transmission'. In humans, the extent to which the EPP exceeds the safety margin is quite small (around 10 mV) and signal transmission is thus sensitive to pathological changes. As well as the ACh release in response to depolarization of the nerve terminal described above, ACh may be released from resting nerve terminals as a result of the spontaneous fusion of ACh-containing vesicles with the membrane. This release of single `packets' (quanta) of ACh into the synaptic cleft causes a small depolarization of the postsynaptic membrane, called a miniature endplate potential (MEPP). MEPPs are usually around 1 mV in amplitude and are insuf®cient to propagate an action potential. MEPP and EPP amplitudes depend upon a number of factors that include the number of ACh molecules released as a `packet', the number of AChRs available to bind ACh, the ion-channel properties of the AChR, the properties of AChE in the synaptic cleft, and the structural geometry of the neuromuscular junction (NMJ). In the clinical investigation of patients with muscle weakness, information about neuromuscular transmission may be obtained by electromyography (recording the compound muscle action potential (CMAP) in response to stimulation of the corre-

Congenital myasthenic syndromes

91

sponding nerve) and by electrophysiological analysis of muscle biopsies. Clues about the nature of a disorder may be gleaned from a change in the initial amplitude of the CMAP, changes in the CMAP during stimulation at different frequencies or after exercise, or the appearance of a repetitive response to a single nerve stimulus. Similarly, a change in the amplitude or duration of MEPPs and miniature endplate currents (MEPCs), measured intracellularly in biopsied muscle, may help clinical diagnosis. Clinical features AChR de®ciency syndrome (AChRdef) and low-af®nity fast-channel syndrome (FCS) These two syndromes have almost indistinguishable clinical manifestations and will therefore be considered together in this section. Weakness is usually evident at birth or within the ®rst year or two of life, and is characterized by feeding dif®culties, ptosis, impaired eye movements and delayed motor milestones. Strength sometimes improves during adolescence, and in contrast to the slow channel syndrome (see below), does not exhibit a progressive course. Re¯exes are usually brisk and muscle wasting does not occur. These patients respond to anti-AChE medication (e.g. intravenous edrophonium). Clinical electromyography (EMG) typically shows increased decrement of the compound muscle action potential at 3 Hz stimulation and single ®bre studies reveal an increase in jitter and blocking. Intracellular microelectrode recording from the endplate region of biopsied muscle will show reduced amplitudes of miniature endplate potentials (MEPPs) and currents (MEPCs) in both syndromes. However biochemical and morphological analyses of muscle biopsies may help to differentiate between the two syndromes. Muscle biopsies from FCS patients may show near normal endplate morphology and 125 I-a-BuTx binding, whereas biopsies from AChRdef patients frequently show reduced numbers of AChR that are often distributed abnormally along the muscle ®bre, coincident with elongated areas of AChE staining. The slow channel syndrome This syndrome was ®rst described by Engel et al. (1982). Although muscle weakness may be apparent in infancy more usually it is not evident until early adulthood and as a result may often be initially misdiagnosed as MG. Sometimes, as in the illustrative case history, it ®rst manifests itself during pregnancy. Weakness and associated wasting of cervical and scapular muscles and of the ®nger extensors may be an early feature. Unlike AChRdef and FCS, the disorder is slowly progressive, involving respiratory as well as limb and bulbar muscles.

92

D. Beeson, J. Newsom-Davis

Patients with SCS do not improve with intravenous edrophonium. Clinical electromyography often reveals the characteristic double response to a single nerve stimulus. Jitter is increased in single ®bre EMG. Microelectrode studies on biopsied intercostal muscle show prolonged decay of the MEPPS and MEPCs. Ultrastructural studies often show an endplate myopathy, particularly in affected muscles, which is thought to be caused by `calcium overload' of the muscle in the endplate region. Endplate acetylcholinesterase de®ciency (EAD) Patients with EAD have generalized weakness that is exacerbated by exercise. Symptoms are apparent at birth or in early childhood and may progress during early life. Staining of endplates shows loss or a marked reduction of AChE and patients, therefore, do not response to AChE inhibitors such as edrophonium. Owing to the absence of AChE, synaptic currents are prolonged. On electromyography there is a repetitive response to a single nerve stimulus, and the MEPP has an extended decay phase. Patients show endplate myopathy and morphological changes that bear similarities to the myopathy of the SCS. Typically, there may be a reduced nerveterminal area, and degeneration of junctional folds causing reduced levels of AChR binding, both of which are likely to be a consequence of the prolonged synaptic currents. Differential diagnosis Myasthenia gravis (MG) is distinguished by the presence of anti-AChR antibodies in the great majority. Moreover, onset at less than 1 year is very rare. Fluctuation in the clinical course is also characteristic of MG, and 10% of patients have a thymoma. Both seronegative and seropositive MG respond to plasma exchange and immunosuppressive medication. Electromyographic studies should distinguish congenital myasthenic syndromes (CMS) from congenital myopathies. Genetics In myasthenic disorders events that affect the generation of the EPP can occur at the presynaptic nerve terminal, within the synaptic cleft, or at the postsynaptic membrane. As pointed out in the introduction, clinical, ultrastructural and in vitro microelectrode studies of the neuromuscular junctions of patients with CMS have de®ned a heterogeneous group of disorders. Although a few patients appear to have presynaptic defects which are not yet fully characterized, the majority involve postsynaptic defects associated with de®ciency of AChR (Vincent et al., 1993) or prolonged opening of the AChR (Engel et al., 1982) and a de®ciency of AChE within the synaptic cleft (Hutchinson et al., 1993). The information gained from

Congenital myasthenic syndromes

93

the detailed analysis of neuromuscular transmission in CMS patients enables a candidate gene approach to be used in the search to establish the underlying genetics of the disorders. In particular, a kinetic abnormality of the AChR implicates a mutation within one of the AChR subunits, whereas de®ciency of AChE or AChR could involve mutations within the genes encoding the respective proteins, within proteins that control their expression or structural proteins involved in anchoring and localization. Acetylcholine receptors Muscle AChRs are glycosylated transmembrane molecules that mediate synaptic transmission. They are allosteric (existing in several different conformations) and the binding of two ACh to each receptor is thought to favour a conformational change that results in a brief opening or activation of the channel and the in¯ux of cations. Each AChR molecule has a molecular mass of around 250 kDa, and is made up of ®ve subunits, arranged in a pentameric structure around a central ion pore. In mammalian muscle there are two types of AChR. A form found in foetal and denervated muscle (foetal AChR) that consists of a2bgd subunits and an adult form, a2bd1, in which the g subunit is replaced by the 1 (Fig. 3). The subunits are homologous, and are thought to have a transmembrane topology consisting of a long N-terminal extracellular domain, followed by three transmembrane domains, termed M1, M2 and M3, a long cytoplasmic loop, and a fourth transmembrane domain, with the C-terminus extracellular to the membrane. Evidence from photo-af®nity labelling with non-competitive channel blockers,

Fig. 3. Diagrammatic representation of the AChR showing the subunit composition of the adult and foetal subtypes.

94

D. Beeson, J. Newsom-Davis

site-directed mutagenesis, and from the accessibility of substituted cysteine residues to charged sulphydryl-speci®c reagents strongly supports the notion that the M2 transmembrane domains line the channel pore (Changeux et al., 1992). All subunits have highly conserved disulphide-linked cysteine residues at positions analogous to amino acids 128 and 142 (human a subunit numbering) that are essential for AChR assembly. Residues in the N-terminal regions of the a, g , d and 1 subunits are thought to contribute to the ACh binding sites, which are located on the interface of ag, ad and a1 subunits. By the early 1990s cDNA and genomic clones encoding the ®ve human AChR subunits had been cloned and sequenced (Noda et al., 1983; Shibahara et al., 1985; Beeson et al., 1989; Luther et al., 1989; Beeson et al., 1990b, 1993). Each of the subunits is encoded by a separate gene of between 10 and 12 exons. The genes encoding the human a, g and d subunits are located on chromosome 2 (Beeson et al., 1990a), and the human b and 1 subunit genes on chromosome 17 (Beeson et al., 1990a, 1993) (Table 2).

Table 2 Polymorphisms identi®ed in the AChR subunit genes

Polymorphism

Allele frequency

Reference

aIVS1 1 59G/T aIVS2-13insT a900C/T b26A/G bIVS6 1 6T/C bIVS10 1 17T/C d-52A/G d57G/A d799C/G d1543A/G 1-130C/T 1-8G/T 1IVS4-7C/T 1IVS8 1 15C/G 1IVS9-6C/T 1IVS11 1 33G/C 1IVS11 1 10del20 11233C/T 11254A/G 11483C/T 11566T/C

11/28 29/70 13/278 40/124 13/54 13/60 32/64 32/64 ND 10/38 7/98 6/252 5/98 34/186 3/58 ND ND 40/196 6/108 24/82 39/82

Milone et al. (1997) Milone et al. (1997) Engel et al. (1996b) Milone et al. (1997) Engel et al. (1996b) Engel et al. (1996b) Milone et al. (1997) Ohno et al. (1997) Engel et al. (1996b) Milone et al. (1997) Milone et al. (1997) Engel et al. (1996b) Milone et al. (1997) Ohno et al. (1998a) Engel et al. (1996b) Sine et al. (1995) Sine et al. (1995) Ohno et al. (1998a) Milone et al. (1997) Ohno et al. (1998a) Ohno et al. (1998a)

Congenital myasthenic syndromes

95

Fetal to adult (g ±1) AChR switch In embryonic muscle, prior to innervation, AChR are distributed along the length of the muscle ®bre and are of the foetal form (a2bgd). During innervation the AChR are lost from the extrasynaptic membrane and become both clustered and localized on the postsynaptic membrane of the NMJ. At the same time expression of the g subunit mRNA is repressed, and is replaced by 1-subunit mRNA transcribed from subsynaptic nuclei. Newly synthesized 1 subunits are incorporated into the receptor pentamers to generate the adult AChR (a2bd1) found at endplates in mature muscle. In humans the g subunit is readily detectable at the endplate until around 31 weeks of gestation (Hesselmans et al., 1993). The g /1 switch results in a change of AChR ion channel properties. Adult AChR has a shorter open time and a higher conductance than foetal AChR (Sakmann et al., 1985). For human AChR expressed in Xenopus oocytes the mean burst duration and single channel conductance of adult AChR were 4.1 ms and 62.2 pS, respectively, as opposed to 7.9 ms and 37.9 pS for foetal AChR (Newland et al., 1995). Mutations underlying AChR de®ciency Like the fast channel syndrome, AChRdef is a recessive disorder in which the actions of ACh are diminished. The great majority of the mutations affect the 1 subunit gene. Both homozygous and recessive heteroallelic mutations occur. Homozygous mutations are frequently the consequence of consanguinity within the pedigree. The mutations are located along the length of the gene and also include changes within transcription control elements of the promoter (Table 3). Promoter mutations Two homozygous mutations, 1-156C ! T and 1-155G ! A, have been identi®ed in the promoter region of the 1-subunit gene (Brengman et al., 1998a; Nichols et al., 1998). Both mutations are within the human N-box sequence (±CCGGAA±), located immediately 5 0 to the transcription initiation start site (Fig. 4). The N-box has been proposed as the site for regulation of synapse speci®c expression of the AChR subunits. ARIA (neuregulin) secreted from the motor nerve terminal is thought to act via postsynaptic receptors (erbB) to increase ras/mitogen activated protein (MAP) kinase activity. It is proposed that MAP-kinase responsive transcription factors, such as the Ets family member GABP, then act at the level of transcription via the N-box. In in vitro studies of the N-box sequence in mice the transition equivalent to 1-156C ! T produced a 70% reduction in reporter gene expression and a 90% reduction in GABP binding. However, in vivo, the situation appears more severe. RT-PCR analysis of RNA obtained from one of two siblings with 1-156C ! T failed to detect mRNA encoding the 1 subunit (Nichols et al., 1999),indicating that this mutation impairs transcription, leading to loss of adult

96

D. Beeson, J. Newsom-Davis

Table 3 Mutations that cause reduced AChR expression

Mutation

Domain

Reference

1-156C ! T 1-155G ! A 1G-8R 159ins5 170insG 1127ins5 1R64X 1C128S 1S143L 1R147L 1553del7 1P245L 1IVS7 1 2 ! C 1R311W 1346delN 1IVS9-1G ! C 11033delG 11101insT 11206ins19 11254ins18 11260del23 11267delG 11276delG 11293insG

Promoter (N-box) Promoter (N-box) Signal peptide Extracellular Extracellular Extracellular Extracellular Disulphide loop Glycosylation site Extracellular Extracellular M1 Intracellular M1- M2 M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop M3-M4 intracellular loop

Nichols et al. (1998) Brengman et al. (1998b) Ohno et al. (1996) Ohno et al. (1998a) Ohno et al. (1998a) Ohno et al. (1997) Ohno et al. (1997) Milone et al. (1998a) Ohno et al. (1996) Ohno et al. (1997) Ohno et al. (1997) Ohno et al. (1997) Ohno et al. (1998a) Ohno et al. (1997) Shen et al. (1998) Shen et al. (1998) Brengman et al. (1998a) Engel et al. (1996a) Ohno et al. (1998a) Milone et al. (1998a) Brengman et al. (1998a) Croxen et al. (1998) Ohno et al. (1998a) Engel et al. (1996a)

Fig. 4. Diagrammatic representation of the AChR 1-subunit gene promoter region, illustrating elements thought to be involved in the control of gene transcription and the two mutations identi®ed within the Nbox that may cause AChR de®ciency.

Congenital myasthenic syndromes

97

AChR expression. The identi®cation of further CMS mutations upstream of the AChR coding regions may highlight additional elements essential for the control of AChR synthesis. Missense mutations of essential residues Missense mutations have highlighted a number of key residues within domains essential for generating the correctly folded 1 subunit within an AChR pentamer. 1G-8R is within the signal peptide sequence and presumably causes inef®cient signal sequence cleavage and the retention of the 1 subunit chain within the endoplasmic reticulum (Ohno et al., 1996). 1C128S changes one of the cysteines that forms the extracellular disulphide bridge, which is present in all members of the `cys-loop' gene superfamily, and is essential for correct subunit folding (Milone et al., 1998a). 1S143L disrupts the consensus site for N-glycosylation (Asn-X-Ser, Asn-X-Thr) that is conserved in each of the ®ve AChR subunits (Ohno et al., 1996) (Fig. 5).

Fig. 5. Topological model of the AChR 1 subunit indicating the positions of mutations underlying AChR de®ciency.

98

D. Beeson, J. Newsom-Davis

Mutations causing truncation of the 1 -subunit Insertions, deletions, splice site and nonsense mutations that lead to truncation of the 1 subunit have been identi®ed. They occur throughout the length of the 1-subunit gene, and each of them causes a translation frameshift that results in varying numbers of missense amino acids prior to peptide chain termination. Expression in HEK293 cells of truncated 1 subunits, in combination with wild-type a, b and d subunits, results in reduced 125I-a-BuTx surface binding and a biphasic binding curve characteristic of a2bd2 pentamers, indicating that the truncated 1 subunits are not incorporated into the surface receptors. How each of these mutations affect assembly of the AChR pentamer has not been studied in detail. It is unlikely that short peptides generated by truncation close to the N-terminus (i.e. 159ins5, 170insG, 1127ins5, Ohno et al., 1996; Ohno et al., 1997) will interact with other subunits and thus, in expression studies with these mutants, a2bd2 pentamers would be expected. Mutations that truncate the 1 subunit close to its normal C-terminus, however, should allow interaction with other AChR subunits. Newland et al. (1998) report that 1 subunits generated by the mutation 11267delG are incorporated into functional AChR pentamers expressed in Xenopus oocytes although surface expression is reduced. However. experiments in HEK293 cells using 11276delG, a mutation that generates a similar truncated subunit, show that only a2bd2 pentamers are formed, and that levels of surface 125I-a-BuTx binding are below those recorded when abd without 1 was expressed (Ohno et al., 1998a). These results suggest similarities to the DF508 mutation of CFTR, where abnormally folded protein remains bound to chaperone proteins in the endoplasmic reticulum and undergoes accelerated degradation in transfected cells at 378C. At temperatures lower than 378C CFTRD508 transfected cells produce increased amounts of fully glycosylated functional protein. The 11267delG mutation is one of the more common CMS mutations and has been detected in at least six unrelated families (Croxen et al., 1998). In this mutation the ®rst nucleotide of exon 12 is deleted. Inef®cient excision of intron 11 from the 1subunit mRNA transcript has been reported (Newland et al., 1998). Retention of intron 11 in combination with the 11267delG mutation returns exon 12 to the correct reading frame and thus conserves the M4 domain. This may account for the relatively mild clinical phenotype in patients with this mutation, although it should be noted that polymorphisms within intron 11 may affect the reading frame. Regulating the assembly of the AChR subunits provides a mechanism for controlling the properties of the expressed receptor. In studies of AChR de®ciency it is often dif®cult to differentiate between mutations that directly affect assembly per se, and those that lead to misfolding and rapid intracellular degradation. Subunit polypeptides contain N-terminal signal sequences and are co-translationally inserted into the membrane of the endoplasmic reticulum where, following initial subunit folding, oligomerization occurs. Only correctly folded and oligomerized subunits are

Congenital myasthenic syndromes

99

ef®ciently transported to the cell surface. Incorrectly folded unassembled subunits and assembly intermediates are retained within the endoplasmic reticulum by association with chaperone proteins. Retention within the endoplasmic reticulum will lead to degradation of the subunits. More detailed studies of the de®ciency mutations may reveal novel domains within the 1 subunit crucial for cell surface expression. Mutations affecting the function of the AChR ion channel Abnormalities of AChR function can potentially increase or decrease the response to ACh. Mutations that increase the response to ACh are dominantly inherited. Theoretically, mutations in the a subunit could result in the dominant inheritance of a decreased response to ACh (dominant-negative effect), but to date all the mutations identi®ed that lead to a loss of response to ACh show recessive inheritance (Fig. 6). Gain of function kinetic mutations The slow channel congenital myasthenic syndrome The slow channel congenital myasthenic syndrome is dominantly inherited, though a number of sporadic cases have also been reported. To date, 14 different mutations underlying the SCS have been reported. The mutations occur in different subunits and in different functional domains within the subunits (Fig. 7). Eight are located within the M2 domain, three within the M1 domain, two are

Fig. 6. Illustration of the effect of kinetic mutations on AChR channel function.

100 D. Beeson, J. Newsom-Davis

Fig. 7. Topological model of the AChR subunits indicating the position of the mutations that give rise to the slow channel syndrome.

Congenital myasthenic syndromes

101

close to residue aW149 that has been shown to contribute to the ACh binding site, and one is extracellular to the membrane, midway between M2 and M3. Each is a single amino acid change, that causes an increase in channel burst length, and thus a pathologic gain of function for the AChR. The mutations are heterozygously expressed in the patients. Therefore the number of different AChR subunit types in the patient's muscle will depend upon whether the mutation is in a non-a subunit, in which case it will be singly represented within a pentamer, or in the doubly represented a subunit. Genotype±phenotype correlation The phenotype of SCS is variable and, as pointed out earlier, onset of weakness can occur neonatally or may not arise until adolescence, adulthood or during pregnancy. Indeed, some family members may carry the mutation, show characteristic electrophysiological changes on biopsy, and yet show no clinical symptoms. Although no clear genotype±phenotype correlation has yet emerged there is some evidence that the mutations that give rise to the longest channel activations, particularly some within the M2 domain, may have the most severe phenotype. However, it is necessary to consider each mutation in the context of its overall effect on AChR function. Loss of function kinetic mutations The low-af®nity fast-channel syndrome Contrasting with the SCS, other CMS cases have been reported in which miniature endplate potentials were very small and single channel activation episodes much shorter than normal (Fig. 8). Genetic analysis shows that this is due to recessive inheritance of mutations that reduce AChR ion channel function. A missense mutation, 1P121L, was identi®ed in combination with a signal peptide mutation 1G-8R in one patient and glycosylation consensus site mutation 1S143L in a second. The combination of the null (1143L) or low expression (1G-8R) allele of the 1 subunit with the 1P121L mutation unmasks its phenotypic effects which are generated by the AChR activations being fewer and shorter than normal (Ohno et al., 1996). More recently two additional fast-channel mutations have been identi®ed, aV285I located within the M3 domain and 11254ins18 located within the long cytoplasmic loop between M3 and M4 (Milone et al., 1998a,b). In a similar fashion to the slow channel syndrome, it is probable that further fast channel mutations will be identi®ed in a series of different functional domains.

102

D. Beeson, J. Newsom-Davis

Fig. 8. Positions of the slow channel mutations within the M2 channel pore region.

Pathogenesis Functional disturbances at the molecular level Residues that contribute to the ACh binding sites and the ion channel have been identi®ed, but the mechanism through which the ACh binding results in the channel opening is not yet clear. The physically distinct nature of the binding sites and Ê below the putative ACh binding channel pore (the ion channel lies about 50 A sites, Unwin, 1993) suggests an allosteric mechanism, i.e. that the AChR can exist in distinct conformations and that the binding of ACh promotes a particular conformation of the molecule that affects the topologically distinct ion channel gate. The AChR is thought to adopt multiple conformations, and at least three general interconvertible functional states have been recognized: a resting state in the absence of agonist in which the probability of channel opening is small; an active state in the presence of agonist in which the probability of opening is high; and a closed state that results from prolonged exposure to high concentrations of agonist in which ACh binds with high af®nity but the AChR is inactive or `desensitised' (Fig. 9). A simplistic but illustrative mechanism (derived from analysis of single channel currents) for activation of the AChR is shown in Fig. 10.

Congenital myasthenic syndromes

103

Fig. 9. Topological model of the AChR a and 1 subunits showing the positions of fast channel mutations.

104

D. Beeson, J. Newsom-Davis

Fig. 10. Basic kinetic framework for analysis of the activation of muscle AChR. A, ACh; R, AChR; A2Rw indicates the ion channel is in the open state.

This scheme does not take into account channel opening of unliganded or monoliganded AChRs, or the existence of desensitized states and assumes the two ACh binding sites are equivalent. During neuromuscular transmission at the normal neuromuscular junction there is a transient saturating concentration of ACh in the synaptic cleft. The high ACh concentration ensures rapid binding (A2R, where the two binding sites are occupied but the channel is closed) and the fast opening rate b ensures rapid channel opening (A2R*). It takes around 100 ms for miniature endplate currents to reach their peak. The hydrolysis of ACh by AChE rapidly clears ACh from the synaptic cleft so that its concentration falls to near zero and there is no rebinding once ACh dissociates from the receptor. Because the rates for channel opening (b ) and ACh dissociation (k22) are roughly similar the bi-liganded receptor oscillates between the open and closed states (A2R* and A2R) before ACh dissociates. The entire oscillatory period between the ®rst opening of the channel and the ®nal closure that precedes ACh dissociation is termed an activation or burst (Edmonds et al., 1995). At the synapse, because of the high initial ACh concentration and then its very rapid removal, the duration of activation of wild-type receptors is dictated by the channel opening rate (b ) the closure rate (a ) and the agonist dissociation rate (k22). Mutations that give rise to altered channel kinetics are likely to affect one or more of these rates, and in addition could affect desensitization, and unliganded and monoliganded open states that probably occur rarely during normal endplate currents. Kinetic analysis of slow channel syndrome mutations aG153S has been identi®ed in three unrelated families (Sine et al., 1995; Croxen et al., 1997). Single channel recordings from mutant receptors engineered in HEK293 cells corresponded to single channel recordings made directly from biopsied muscle. The recordings demonstrated prolonged channel activations with the receptor oscillating many more times than wild-type AChR between the open and closed states. Kinetic analysis shows that the prolonged activation arises primarily through a reduction in the rate of dissociation of ACh from the AChR (k22), thereby increasing the number of reopenings during ACh occupancy. aG153S is located close to aW149, which has been shown by photoaf®nity labelling to contribute to

Congenital myasthenic syndromes

105

the ACh binding site. Thus, the primary effect of aG153S is on the af®nity of the AChR for ACh, possibly by a direct effect on the binding pocket. In addition it causes an increased propensity to desensitization. Interestingly, aV156M (Croxen et al., 1997), although located in the same region, appears to act through a different mechanism, but detailed kinetic analysis of this mutation has yet to be reported. The majority of SCS mutations are located in the M2 domain. M2 mutations have been identi®ed in each of the four adult AChR subunits. The ®rst to be describes was 1T264P (Ohno et al., 1995). This mutation slows the rate of channel closure (a ), stabilizing the AChR in the open state, so that within an activation, the duration of individual openings are increased. Although not all M2 mutations have been analyzed in detail it is likely they act through a similar primary mechanism. A second feature of M2 mutations, which has been observed for 1T264P, 1V265A (Ohno et al., 1998c), 1L269F, bV266M and aV249F, is increased channel openings in the absence of ACh. It was not seen for aT254I, although this may be due to expression in Xenopus oocytes as opposed to HEK293 cells. Other properties of the AChR may be affected depending on the particular mutation. For instance, aV249F, is thought to enhance af®nity for binding ACh in the resting state, and increase desensitization. aN217K, and two recently identi®ed mutations, bV229F (Gomez et al., 1998) and 1L221F are located within the amino-terminal third of the M1 transmembrane domain. Single channel recordings of aN217K give a qualitatively similar pro®le to that seen for aG153S and kinetic analysis shows that its primary effect is to slow the rate of ACh dissociation from the binding site (k22) (Wang et al., 1997), leading to repeated reopening of the channel during ACh occupancy. This slightly surprising Ê above the plane of the result (the ACh binding site is thought to be 25±30 A membrane) illustrates how a mutation distant from the binding site may cause a local conformational change that is propagated to the binding site and enhances the ®t of ACh for AChR in the resting state. It is not yet known if bV229F and 1L221F similarly affect ACh dissociation, or stabilize the AChR in the open state. aS269I is located midway between the M2 and M3 domains and is thus extracellular. It could have a distant effect on ACh binding sites, in a similar fashion to aN217K, or it may directly affect the transduction of the conformational change involved in channel gating. Kinetic analysis of the low-af®nity fast-channel syndrome By contrast to SCS the FCS shows a decreased response to ACh. Patients have normal neuromuscular ultrastructure and AChR density, but very small MEPPs. Single channel recordings at the endplate from a FCS patient muscle biopsy resemble those of the 1P121L mutant expressed in HEK293 cells (Ohno et al., 1996). 1P121L slows the rate of channel opening (b ), but has little effect on the rate of ACh dissociation (k22). Since the probability that a given occupancy will result in a

106

D. Beeson, J. Newsom-Davis

channel opening depends on the ratio of b /k22, 1P121L results in reduced channel reopenings during ACh occupancy, and consequently shorter than normal activations. It also results in a reduced frequency of apparent activations since some ACh occupancies will not result in any openings. In addition, the binding af®nity of the mutant AChR in the open and desensitized states for ACh is reduced. At the endplate the small MEPPS result from reduced numbers of AChR being activated and the shorter than normal burst length. The mutation 11254ins18, in which a six amino acid sequence (STRDQE) within the long cytoplasmic loop is duplicated, causes an FCS in which channel opening (b ) is slowed and closure (a ) is faster. This mutation also has the interesting property of altering the kinetic mode of activation. As well as its fast channel characteristics, 11254ins18 also reduces the number of surface AChRs to less than 50% of wild-type AChRs, thus compounding the kinetic effects. Missense mutations affecting AChR expression and kinetics Mutations that combine reduced surface expression with kinetic effects have been identi®ed. 1P245L increases burst duration and 1R311W reduces burst duration (Ohno et al., 1997). However, in both these cases the kinetic effect is likely to be inconsequential because of the low surface expression. Phenotypic consequences of AChR mutations AChR de®ciency and fast-channel syndrome The phenotypes of AChR de®ciency and fast channel syndromes are similar, as already described. In both syndromes there is a reduced response to ACh, although in AChR de®ciency this results from loss of AChR number whereas in the fast channel it is the result of fewer and shorter AChR activations. Muscle biopsies from fast channel patients do not usually show ultrastructural changes of the neuromuscular junction whereas AChR de®ciency is often associated with a lower density of AChR distributed abnormally along the muscle ®bre and coincident with elongated areas of AChE staining. In addition, de®ciency patients frequently show the reduced utrophin abundance which is associated with loss of AChRs (Slater et al., 1997; Sieb et al., 1998). Genetic analysis indicates that the focus for AChR de®ciency mutations is the AChR 1 subunit. Moreover, many mutations result in non-functional 1 subunits that fail to be incorporated into AChR pentamers. In these cases neuromuscular transmission must occur either through a2bd2 pentamers or through expression of the foetal form of the receptor (a2bgd). Studies of 1 subunit `knock out' mice (see below) suggest that neuromuscular transmission does not readily occur through

Congenital myasthenic syndromes

107

a2bd2 pentamers. By contrast the AChR g subunit has been detected by immunohistochemical and (probably) by patch clamp techniques at the neuromuscular junctions of AChR de®ciency patients (Engel et al., 1996a; Ohno et al., 1997), fast channel syndrome patients (Ohno et al., 1996) and also in some SCS patients (Ohno et al., 1995). It is thought likely, therefore, that in patients who have null alleles for the 1 subunit neuromuscular transmission occurs through foetal AChRs and the presence of the g subunit rescues the phenotype from mutations that otherwise would be fatal. Animal models of AChR de®ciency Mice with targeted disruption of the 1 subunit have been generated by two groups (Witzmann et al., 1996; Missias et al., 1997). In both studies the mice were weak and showed reduced number and density of AChR at the post-synaptic membrane, dying between 10±12 weeks of age. AChRs were gradually lost from the endplates and at the age of 8 weeks, when the mutant mice began to die, their endplates typically had only 5% of wild-type AChR numbers. The size of the synaptic area was near normal, although some thin axonal processes extended beyond the endplate region. In mice, the g subunit is usually undetectable at the endplate from 2 weeks after birth; however, both immunohistochemical and electrophysiological recordings showed that foetal AChR was present at the endplates of the adult mutant mice. The half-life of foetal AChR is normally around 24 h whereas for adult AChR at the endplate it is around 10 days. RNase protection assays failed to detect g subunit mRNA expression in the mutant mice, suggesting that survival is principally due to stabilization at the endplate of the g containing AChRs rather than compensatory upregulation of transcription. In addition to the loss of AChR the postsynaptic junctional folds failed to develop normally in the mutant mice. The changes at the neuromuscular junctions of 1 subunit `knock out' mice resemble those seen in AChR de®ciency patients in several respects. However, there is one clear and distinctive clinical difference between the patients and this animal model, namely that the patients are relatively mildly affected whereas the knock out mice die. A possible explanation for this species difference lies in the transcriptional control of the respective g subunits. In mice, following innervation and induction of 1 subunit expression, the repression of g subunit mRNA synthesis is tightly controlled, and even using RT-PCR it is dif®cult to detect g subunit mRNA. By contrast, in both normal and 1 subunit de®cient human muscle biopsies g subunit mRNA can be readily detected both by RT-PCR and by RNase protection assays (MacLennan et al., 1997). Thus, in human muscle, a low level of the g subunit, suf®cient for survival, is likely to be available for recruitment into AChR pentamers.

108

D. Beeson, J. Newsom-Davis

The slow channel syndrome Mutations underlying SCS prolong channel activations The phenotype is extremely variable. It is likely that a number of factors arising as a consequence of the mutations affect the safety margin for neuromuscular transmission. Ultrastructural changes are evident at the endplates of patients and include evidence of calcium accumulation, Ca 21 ions make up approximately 6% of the normal current through adult AChR (Villarroel and Sakmann, 1996). It is proposed that delayed closure of the AChR allows excess Ca 21 to ¯ow through the channel, that in turn leads to localized Ca 21 overload and the activation of Ca 21-sensitive degradative enzymes (Engel et al., 1982). Leakage due to channel openings in the absence of ACh, which has been reported for many of the M2 mutations, may enhance this excitotoxic effect. The resultant `endplate myopathy' is characterized by breakdown and loss of junctional folds, degeneration of organelles such as mitochondria, apoptosis of some subsynaptic nuclei, autophagic vacuoles, and widening of the synaptic space. The breakdown of the postsynaptic membrane may result in loss of AChR, although in one study a 6±45-fold increase in perijunctional 125I-a-BuTx binding sites was reported (Gomez et al., 1996). In some SCS patients a minor population of foetal AChR has been recorded at their endplates (Ohno et al., 1995; Engel et al., 1996b), that, possibly, are induced because the loss of postsynaptic electrical stimulation has a `denervation-like' effect on the muscle. Features of the mutant channel other than the endplate myopathy may also contribute to dysfunction of neuromuscular transmission. MEPPS are sometimes reduced in amplitude as well as exhibiting a prolonged decay phase, and a decrement during repetitive stimulation is not uncommon. An explanation for this may be that in mutations such as aG153S, aN217K, 1L269F, and aV249F the propensity for the mutant channel to enter the desensitized state is increased, thus leading to fewer AChRs available for signal transmission. Another cause of dysfunction may be that, at physiological rates of stimulation, the prolonged endplate potentials summate, leading to persistent depolarization at the endplate and inactivation of the voltagegated sodium channels that generate the muscle action potentials. Animal models of the slow channel syndrome Transgenic mice containing mutations dS262T, aC418W,aL251T and 1L269F have been established and two of these dS262 and 1L269F have been analyzed in detail (Gomez et al., 1996, 1997). Both mutations prolong AChR channel activations. The effect of dS262T was identi®ed from in vitro mutagenesis experiments, whereas that of 1L269F was recognized from studies of two unrelated families with SCS (Gomez and Gammack 1995; Engel et al., 1996b). The major component of channel opening is prolonged approximately 3-fold by dS262T and 6±8-fold by

Congenital myasthenic syndromes

109

1L269F. Transgenic lines for both mutations demonstrate repetitive compound muscle action potentials after single nerve stimulation and prolonged MEPPS. Transgenic mice bearing the dS262T mutation showed neither weakness nor ultrastructural changes at the endplate up to 1 year of age, and endplate 125I-aBuTx binding was normal. However, the amplitudes of both EPC and MEPC during repetitive stimulation were signi®cantly reduced compared to controls, suggesting that for this mutation increased desensitization may reduce the safety margin for neuromuscular transmission. By contrast, transgenic lines expressing the mutant 1 subunit 1L269F developed progressive weakness, fatigability and endplate myopathy. Signs of neuromuscular dysfunction were evident at 3 weeks. The ®rst ultrastructural change at the synapse was the accumulation of vacuolar structures in the sarcoplasm and a slight reduction in AChR number. By 3±4 months synaptic degeneration was clearly evident, with degeneration of subsynaptic nuclei and mitochondria, and reduced postsynaptic folds and numbers of AChR. By 15 months the number of AChR was reduced to around 34% of controls and additional ¯attening of the postsynaptic fold and widening of the synaptic cleft had occurred. While the ultrastructural changes correlate with the progressive weakness in older mice, they are unlikely to account for the weakness observed at an early age, suggesting that additional factors, such as increased desensitization of mutant receptors and summation of the prolonged endplate potentials leading to depolarization block, may contribute to neuromuscular dysfunction. The 1L269F transgenic lines provide a useful disease model of an excitotoxic channelopathy. The endplate myopathy of the mice bear striking similarities to the endplate myopathy observed in patients harbouring the 1L269F SCS mutation. Moreover, the transgenic mice lines demonstrate the association of progressive degenerative endplate changes with the excitotoxic kinetic effect generated by a single point mutation.

Subsequent research aims Recent therapeutic approaches An understanding of the pathogenic mechanisms of AChR mutations provides a rational basis for therapy. For instance, until recently there has been no effective therapy for SCS. Studies using quinidine sulphate, a long-lived blocker of the AChR channel, demonstrate that it may be used to reduce channel open duration. At a 5 mM concentration it can be used to bring burst durations of mutant AChR back to near normal; moreover, this concentration can be achieved at patient endplates. All six SCS patients enrolled in a trial of quinidine sulphate showed improved muscle

110

D. Beeson, J. Newsom-Davis

strength, although adverse side effects occurred in two (Fukudome et al., 1998; Harper and Engel 1998). Patients in whom there is reduced response to ACh bene®t from AChE inhibitors and 3,4-diaminopyridine (3,4-DAP) (Palace et al., 1991). AChE inhibitors prolong the lifetime of ACh within the synaptic cleft, thus enabling activation of a greater number of AChRs. 3,4-DAP acts principally by blocking presynaptic voltage-gated potassium channels, thereby prolonging the depolarization of the nerve terminal which results in increased quantal release of ACh. Thus, 3,4-DAP increases the effective concentration of ACh within the synaptic cleft. 3,4-DAP has been found to be particularly effective for patients with fast channel syndrome, although some may experience side effects. Mutations in other proteins located at the neuromuscular junction As pointed out in this chapter, CMS comprise a heterogeneous group of disorders. While it is now established that mutations within the AChR genes are a cause in many of them, it is highly likely that mutations in neuromuscular junction proteins that associate with AChRs either through function or physical interaction will affect signal transmission (Vincent et al., 1997). Dysfunction of proteins that regulate synthesis, clustering, or attachment to the cytoskeleton could all compromise AChR function, as could mutations that affect the concentration of ACh in the synaptic cleft. Recent reports con®rm that mutations outside the transcribed regions of the AChR genes may cause CMS. First, the mutations in the promoter region of the AChR 1 subunit gene (Brengman et al., 1998b; Nichols et al., 1998) demonstrate how the control of AChR synthesis underlies a CMS. Secondly, reports have now identi®ed mutations within the ColQ gene, which encodes the collagen-like tail that attaches the asymmetric form of AChE to the basal lamina at the neuromuscular junction (Donger et al., 1998; Ohno et al., 1998b; Ohno et al., 1998d). The reports describe 11 mutations that underlie recessive inheritance of acetylcholinesterase de®ciency (EAD) syndrome. ColQ is located on chromosome 3p24.2. The ColQ protein contains an N-terminal proline rich domain (PRAD) that binds AChE tetramers, a collagen domain with 63 Gxy repeats, and a C-terminal domain responsible for anchoring in the basal lamina and for initiating assembly of the collagen triple helix structure. The mutations (P59Q, 107del215, W148X, S169X, E214X, 788insC, 806insC, R282X, D342E, 1082delC, Y431S) which may be homozygous or heteroallelic, are present in each of the function domains, and illustrate genetic heterogeneity for this disorder. The result of these mutations is the loss of the asymmetric form of AChE from the synaptic cleft, and consequently an increase in the time that ACh is available to bind AChR. The physiological consequences show similarities to SCS. Prolonged exposure to ACh will cause desensitization of

Congenital myasthenic syndromes

111

the AChR and at physiological rates of stimulation, persistent depolarization of the endplate will inactivate the voltage-gated sodium channels in the post-synaptic membrane and thus block signal transmission. In addition, persistent stimulation will cause an endplate myopathy. EAD illustrates a pathology generated by AChR although caused by an associated functional molecule. Mutations in postsynaptic proteins associated with localizing AChR on the postsynaptic membrane may well give rise to phenotypes similar to AChR de®ciency. Perspectives The neuromuscular junction is relatively well understood. Moreover, what we know about muscle AChR provides the basis for many studies of the cysteine loop ion channel gene superfamily and their role within the CNS. Most of the CMS are channelopathies for which functional effects may be directly recorded from muscle biopsies, where pathogenic channel function can be modelled in vitro in mammalian cells and in vivo in transgenic mice, and where channel dysfunction can be correlated directly with the clinical phenotype. The diversity of mutations and clinical phenotypes is providing novel insights both into the detail of ion channel function and the pathogenic consequences of dysfunction. AChRs were the ®rst ligand-gated ion channels to be cloned and remain the best understood of the neurotransmitter receptors. The methodologies and experimental models developed for their study have acted as a paradigm that has underpinned research into the ion channels of the CNS. Similarly, the CMS should provide a model system for understanding synaptic dysfunction within the CNS, and investigating new therapeutic strategies. For instance SCS mutations provide the key elements required of a model system to test the potential of ribozymes to annul excitotoxic damage in a neurological disorder. Finally, understanding the underlying cause of CMS provides a basis for genetic counselling and rational treatment for a clinically heterogeneous group of disorders.

References Beeson, D., Brydson, M., Newsom-Davis, J., 1989. Nucleotide sequence of human muscle acetylcholine receptor beta subunit. Nucl. Acids Res. 17, 4391. Beeson, D., Jeremiah, S., West, L., Povey, S., Newsom-Davis, J., 1990. Assignment of the human nicotinic acetylcholine receptor genes: a- and d-subunit genes to chromosome 2 and b-subunit gene to chromosome 17. Ann. Hum. Genet. 54, 199±208. Beeson, D., Morris, A., Vincent, A., Newsom-Davis, J., 1990. The human muscle acetylcholine receptor a subunit exists as two isoforms: a novel exon. EMBO J. 9, 2101±2106. Beeson, D., Brydson, M., Betty, M., Jeremiah, S., Povey, S., Vincent, A., Newsom-Davis, J., 1993. Primary structure of the human muscle acetylcholine receptor: cDNA cloning of the gamma and epsilon subunits. Eur. J. Biochem. 215, 229±238. Brengman, J., Ohno, K., Shen, X-M., Engel, A.G., 1998. Congenital myasthenic syndrome (CMS) due to

112

D. Beeson, J. Newsom-Davis

two novel mutations in the acetylcholine receptor (AChR) 1 subunit gene. Muscle Nerve (Suppl. 7, S120. Brengman, J., Ohno, K., Anlar, B., Engel, A.G., 1998. A mutation in the promoter region of the acetylcholine receptor (AChR) ( subunit gene causes congenital myasthenic syndrome. Am. J. Hum. Genet. 63, A3532044. Changeux, J.-P., Galzi, J.-L., Devilliers-Thiery, A., Bertrand, D., 1992. The functional architecture of the acertylcholine nicotinic receptor explored by af®nity labelling and site-directed mutagenesis. Qu. Rev. Biophys. 25, 395±432. Croxen, R., Newland, C., Beeson, D., Oosterhuis, H., Chauplannez, G., Vincent, A., Newsom-Davis, J., 1997. Mutations in different functional domains of the human muscle acetylcholine receptor a subunit in patients with slow-channel congenital myasthenic syndrome. Hum. Mol. Genet. 6, 767±773. Croxen, R., Beeson, D., Newland, C., Betty, M., Vincent, A., Newsom-Davis, J., 1998. A single nucleotide deletion in the 1 subunit of the acetylcholine receptor (AChR) in ®ve congenital myasthenic syndrome patients with AChR de®ciency. Ann. N. Y. Acad. Sci. 841, 195±198. Donger, C., Krejci, E., Serradell, A., Eymard, B., Bon, S., Nicole, S., Chateau, D., Gary, F., Fardeau, M., Massoulie, J., Guicheeney, P., 1998. Mutation in the human acetylcholinesterase-associated collagen gene. ColQ, is responsible for congenital myasthenic syndrome with end-plate acetylcholine esterase de®ciency (type 1c). Am. J. Hum. Genet. 63, 967±975. Edmonds, B., Gibb, A., Colquhoun, D., 1995. Mechanisms of activation of muscle nicotinic acetylcholine receptors and the time course of endplate currents. Ann. Rev. Physiol. 57, 469±493. Engel, A.G., Lambert, E.H., Mulder, D., Torres, C.F., Sahashi, K., Bertorini, T., Whitaker, J.N., 1982. A newly recognised congenital myasthenic syndrome attributed to a prolonged open time of the acetylcholine-induced ion channel. Ann. Neurol. 11, 553±569. Engel, A.G., Ohno, K., Bouzat, C., Sine, S.M., Griggs, R.G., 1996. Endplate acetylcholine receptor de®ciency due to nonsense mutations in the 1 subunit. Ann. Neurol. 40, 810±817. Engel, A.G., Ohno, K., Milone, M., Wang, H-L., Nakano, S., Bouzat, C., Pruitt, J., Huchinson, D., Brengman, J.M., Bren, N., Sieb, J.P., Sine, S.M., 1996. New mutations in acetylcholine receptor subunit genes reveal heterogeneity in the slow-channel congenital myasthenic syndrome. Hum. Mol. Genet. 5, 1217±1227. Fukudome, T., Ohno, K., Brengman, J., Engel, A.G., 1998. AChR channel blockade by quinidine sulphate reduces channel open duration in the slow-channel congenital myasthenic syndrome. Ann. N. Y. Acad. Sci. 841, 199±202. Gomez, C., Gammack, J.T., 1995. A leucine to phenylalanine substitution in the acetylcholine receptor ion channel in a family with the slow-channel syndrome. Neurology 45, 982±985. Gomez, C., Bhattacharyya, B., Charnet, P., Day, J., Labarca, C., Wollmann, R., Lambert, E.H., 1996. A transgenic mouse model of the slow-channel syndrome. Muscle Nerve 19, 79±87. Gomez, C., Maselli, R., Gundeck, J., Chao, M., Day, J., Tamamizu, S., Lasalde, J., McNamee, M., Wollmann, R., 1997. Slow-channel transgenic mice: a model of postsynaptic organellar degeneration at the neuromuscular junction. . J. Neurosci. 17, 4170±4179. Gomez, C.M., Maselli, R., Staub, J., Day, J., Cens, T., Wollmann, R., Charnet, P., 1998. Novel d and b subunit acetylcholine receptor mutations in the slow-channel syndrome demonstrate phenotypic variability. J. Neurosci. 24, 1876. Harper, R., Engel, A.G., 1998. Safety and ef®cacy of quinidine sulphate in slow-channel congenital myasthenic syndrome. Ann. N. Y. Acad. Sci. 841, 203±206. Hesselmans, L., Jennekens, F., Van Den Oord, C., Veldman, H., Vincent, A., 1993. Development of innervation of skeletal muscle ®bers in man: relation to acetylcholine receptors. Anat. Rec. 236, 553± 562. Hutchinson, D., Walls, T., Nakano, S., Camp, S., Taylor, P., Harper, C.M., Groover, R., Peterson, H.,

Congenital myasthenic syndromes

113

Jamieson, D., Engel, A.G., 1993. Congenital endplate acetylcholinesterase de®ciency. Brain 116, 633±653. Luther, M., Schoepfer, R., Whiting, P., Casey, B., Blatt, Y., Montal, M., Lindstrom, J., 1989. A muscle AChR is expressed in the human cerebella medulloblastoma line TE671. . J. Neurosci. 9, 1082±1096. MacLennan, C., Beeson, D., Biujs, A-M., Vincent, A., Newsom-Davis, J., 1997. Acetylcholine expression in human extraocular muscles and their susceptibility to myasthenia gravis. Ann. Neurol. 41, 423±431. Milone, M., Wang, H-L., Ohno, K., Fukadome, T., Pruitt, J., Bren, N., Sine, S.M., Engel, A.G., 1997. Slow-channel myasthenic syndrome caused by enhanced activation, desensitisation and agonist binding af®nity attributable to mutation in the M2 domain of the acetylcholine receptor a subunit. . J. Neurosci. 17, 5651±5665. Milone, M., Wang, H-L., Ohno, K., Prince, R., Fukadome, T., Shen, X-M., Brengman, J., Griggs, R., Sine, S., Engel, A.G., 1998. Mode switching kinetics produced by a naturally occurring mutation in the cytoplasmic loop of the human acetylcholine receptor a subunit. Neuron 20, 575±580. Milone, M., Ohno, K., Brengman, J.M., Batocchi, A.P., Tonali, P.A., Engel, A.G., 1998. Low-af®nity fastchannel congenital myasthenic syndrome caused by new missense mutations in the acetylcholine receptor a subunit. Neurology 50 (Suppl. 4, A432±A433. Missias, A., Mudd, J., Cunningham, J., Steinbach, J., Merlie, J., Sanes, J., 1997. De®cient development and maintenance of postsynaptic specialisations in mutant mice lacking an adult acetylcholine receptor subunit. Development 124, 5075±5086. Newland, C., Beeson, D., Vincent, A., Newsom-Davis, J., 1995. Functional and non-functional isoforms of the human muscle acetylcholine receptor. J. Physiol. 489, 767±778. Newland, C., Croxen, R., Vincent, A., Newsom-Davis, J., Beeson, D., 1998. Functional properties of novel acetylcholine receptor 1 subunits identi®ed in normal and congenital myasthenic muscle. J. Neurosci. 24, 33222. Nichols, P., Croxen, R., Vincent, A., Newsom-Davis, J., Beeson, D., 1998. Congenital myasthenia associated with muscle AChR subunit gene promoter mutations. J. Neurol. 245, 381 (abstract). Nichols, P.N., Croxen, R., Vincent, A., Rutter, R., Hutchinson, M., Newsom-Davis, J., Beeson, D., 1999. Mutation of the acetylcholine receptor 1-subunit promoter in congenital myasthenic syndrome. Ann. Neurol. 45, 439±443. Noda, M., Furutani, Y., Takahashi, H., Toyosato, M., Tanabe, T., Schimitzu, S., Kikyotani, S., Kayano, T., Hirose, T., Inayama, S., Numa, S., 1983. Cloning and sequence analysis of calf cDNA and human genomic DNA encoding a subunit precursor of muscle acetylcholine receptor. Nature 305, 818±823. Ohno, K., Hutchinson, D., Milone, M., Brengman, J.M., Bouzat, C., Sine, S.M., Engel, A.G., 1995. Congenital myasthenic syndrome caused by a prolonged acetylcholine receptor channel openings due to a mutation in the M2 domain of the 1 subunit. Proc. Natl. Acad. Sci. USA 92, 758±762. Ohno, K., Wang, H-L., Milone, M., Bren, N., Brengman, J.M., Nakano, S., Quiram, P., Pruitt, J., Sine, S.M., Engel, A.G., 1996. Congenital myasthenic syndrome caused by decreased agonist binding af®nity due to a mutation in the acetylcholine receptor 1 subunit. Neuron 17, 157±170. Ohno, K., Quiram, P., Milone, M., Wang, H.-L., Harper, C.M., Pruitt, J., Brengman, J., Pao, L., Fishbeck, K., Crawford, T., Sine, S.M., Engel, A.G., 1997. Congenital myasthenic syndromes due to heteroallelic nonsense/missense mutations in the acetylcholine receptor 1 subunit gene: identi®cation and functional characterisation of six new mutations. Hum. Mol. Genet. 6, 753±766. Ohno, K., Anlar, B., Ozdirim, E., Brengman, J., DeBlecker, J., Engel, A.G., 1998. Myasthenic syndromes in Turkish kinships due to mutations in the acetylcholine receptor. Ann. Neurol. 44, 234±241. Ohno, K., Brengman, J., Tsujino, A., Engel, A.G., 1998. Human endplate acetycholinesterase de®ciency caused by mutations in the collagen-like tail subunit (ColQ) of the asymmetric enzyme. Proc. Natl. Acad. Sci. USA 95, 9654±9659. Ohno, K., Milone, M., Brengman, J.M., LoMonaco, M., Evoli, A., Tonali, P.A., Engel, A.G., 1998. Slow-

114

D. Beeson, J. Newsom-Davis

channel congenital myasthenic syndrome caused by a novel mutation in the acetylcholine receptor 1 subunit. Neurology, 50 (Suppl. 4) A432. Ohno, K., Brengman, J., Milone, M., Shen, X-M., Tsujino, A., Anlar, B., Engel, A.G., 1998. Congenital endplate acetylcholinesterase de®ciency: novel missense and null mutations in the collagen-like tail subunit of the asymmetric form. Am. J. Hum. Genet. 63, (Suppl. A377) 2189. Oosterhuis, H., Newsom-Davis, J., Wokke, J., Molenaar, P., Weerden, T., Oen, B., Jennekens, F., Veldman, H., Vincent, A., Wray, D., Prior, C., Murray, N., 1987. The slow channel syndrome: two new cases. Brain 110, 1061±1079. Palace, J., Wiles, C., Newsom-Davis, J.3., 1991. 4-Diaminopyridine in the treatment of congenital (hereditary) myasthenia. J. Neurol. Neurosurg. Psychiatry 54, 1069±1072. Sakmann, B., Methfessel, C., Mishina, M., Takahashi, T., Takai, T., Kurasaki, M., Fukuda, K., Numa, S., 1985. Role of acetylcholine receptor subunits in gating of the channel. Nature 318, 538±543. Shen, X-M., Ohno, K., Fukadome, T., Brengman, J., Engel, A.G., 1998. Congenital myasthenic syndrome (CMS) due to a splice site and a novel missense mutation in the acetylcholine receptor (AChR) 1 subunit gene. Muscle Nerve (Suppl. 7, S120. Shibahara, S., Kubo, T., Perski, H., Takahashi, H., Noda, M., Numa, S., 1985. Cloning and sequence analysis of human genomic DNA encoding the gamma subunit precursor of muscle acetylcholine receptor. Eur. J. Biochem. 146, 15±22. Sieb, J-P., Dor¯er, P., Tzartos, S., Wewer, U., Ruegg, M., Meyer, D., Baumann, I., Lindemuth, R., Jakschik, J., Ries, F., 1998. Congenital myasthenic syndromes in two kinships with end-plate acetylcholine receptor and utrophin de®ciency. Neurology 50, 54±61. Sine, S., Ohno, K., Bouzat, C., Auerback, A., Milone, M., Pruitt, J., Engel, A.G., 1995. Mutation of the acetylcholine receptor a subunit causes a slow channel myasthenic syndrome by enhancing agonist binding af®nity. Neuron 15, 229±239. Slater, C.R., Young, C., Wood, S.J., Bewick, G.S., Anderson, L., Baxter, P., Fawcett, P., Roberts, M., Jacobson, L., Kuks, J., Vincent, A., Newsom-Davis, J., 1997. Utrophin abundance is reduced at neuromuscular junctions of patients with both inherited and acquired acetylcholine receptor de®ciencies. Brain 120, 1513±1531. Ê resolution. J. Mol. Biol. 229, 1101±1124. Unwin, N., 1993. Nicotinic acetylcholine receptor at 9 A Villarroel, A., Sakmann, B., 1996. Calcium channel permeability increase of endplate channels in rat muscle during postnatal development. J. Physiol. 496, 331±338. Vincent, A., Newsom-Davis, J., Wray, D., Shillito, P., Harrison, J., Betty, M., Beeson, D., Mills, K., Palace, J., Molenaar, P., Murray, N., 1993. Clinical and experimental observations in patients with congenital myasthenic syndromes. Ann. N.Y. Acad. Sci. 681, 451±460. Vincent, A., Newland, C., Croxen, R., Beeson, D., 1997. Genes at the junction ± candidates for congenital myasthenic syndromes. Trends. Neurosci. 20, 15±22. Wang, H-L., Auerbach, A., Bren, N., Ohno, K., Engel, A.G., Sine, S.M., 1997. Mutation in the M1 domain of the acetylcholine receptor a subunit decreases the rate of agonist dissociation. J. Gen. Physiol. 109, 757±766. Witzmann, V., Schwartz, M., Koenen, C., Berberich, H., Villarroel, A., Wernig, H., Brenner, R., Sakmann, B., 1996. Acetylcholine receptor epsilon subunit deletion causes muscle weakness and atrophy in juvenile and adult mice. Proc. Natl. Acad. Sci. USA. 93, 13286±13291.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

115

Chapter 5

Ion channel defects in primary electrical diseases of the heart Glenn E. Kirsch Case Western Reserve University, MetroHealth Medical Center, Rammelkamp Center for Research, 2500 MetroHealth Drive, Cleveland, OH 44109-1998, USA

Abstract Abnormal function or expression of cardiac ion channels has the potential for triggering arrhythmia that can result in sudden death. Although most arrhythmias are associated with structural heart disease (e.g. coronary artery disease, heart valve disease or cardiomyopathy), some have been attributed to primary electrical diseases in which malfunctions associated with ion channels play an important role. In recent years remarkable progress toward understanding the molecular basis of this problem has been achieved through a combined approach of genetics, molecular biology and cellular electrophysiology. In particular, a number of inherited defects in cardiac K 1 and Na 1 channels have been identi®ed as the underlying causes of long QT syndrome, a disease in which malignant ventricular arrhythmia is associated with delayed repolarization phase of the cardiac action potential. A similar approach is underway in the study of electrical disturbances that do not involve repolarization abnormalities, such as those associated with idiopathic ventricular ®brillation. These results hold promise for the development of new diagnostic and therapeutic strategies as well as for understanding the ionic basis of cardiac arrhythmia. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Alterations in cardiac excitability and conduction that have the potential for triggering arrhythmia and sudden death even in the absence of structural heart disease (e.g. coronary artery disease, heart valve disease or cardiomyopathy) have been attributed to primary electrical diseases in which ion channels play an important role (Viskin et al., 1998). These diseases are particularly insidious because they often have a familial component. Their initial appearance is sometimes a fatal arrhythmic episode that strikes young, otherwise healthy individuals. Without

116

G.E. Kirsch

preventive measures, those that survive the initial occurrence are at strong risk of reoccurrence. Considerable progress in understanding the molecular basis of this problem has been made through a combined approach of genetics, molecular biology and cellular electrophysiology. The results hold the promise of new diagnostic and therapeutic strategies. Correlations between ion channel dysfunction and pathogenesis are made simpler in heart than in the nervous system because of well-established relationships between various ionic current components and the cardiac action potential, and between the action potential waveform and rhythmicity. Moreover, the identi®cation of speci®c ion channel genes with observed ionic currents has facilitated the elucidation of the genetic component of these disorders. Fig. 1 summarizes the essential relationships. The initial upstroke of the action potential (phase 0, Fig. 1B) is driven by a brief surge of inward current through Na 1 channels. Under normal conditions these channels activate rapidly in response to pacemaker depolarization (not illustrated) and then close into an inactivated state while the cell membrane remains partially depolarized during phases 1 and 2 due to other depolarizing effects. The major structural subunit is encoded by the SCN5A gene located on chromosome 3 in humans (Fig. 1C). Mutations in SCN5A have been linked to one form (LQT3) of long-QT syndrome (LQTS) and to a form of idiopathic ventricular ®brillation (Brugada syndrome). The return of the membrane potential to the diastolic level (phase 4) progresses through several steps, each controlled by different channels. A transient outward current (Ito1) carried primarily by K 1 channels of the Kv4 subfamily controls the early repolarization phase 1. Although not genetically linked to disease, down-regulation of Kv4.3 may contribute to action potential prolongation and arrhythmia in congestive heart failure (Kaab et al., 1998). The phase 2 plateau is controlled mainly by inward currents through L-type Ca 21 channels with a minor contribution from Na 1 channels. The involvement of Ca 21 channel defects in primary electrical disease has not been established but a reduction in the amplitude of ICa-L may be responsible for action potential shortening in atrial ®brillation (Le Grand et al., 1994; Ouadid et al., 1995). Late repolarization (phase 3) is controlled by an assortment of K 1 channels whose different functional properties produce characteristic ultra-rapid (IKur), rapid (IKr) and slow (IKs) components of the sustained outward current (IK, delayed rectifying). Mutations of the structural genes that encode the channels associated with both IKr and IKs have been linked to several forms of LQTS. Also, it is noteworthy that ion channel genes are not expressed equally throughout the myocardium. Non-uniformity of ion channel composition is thought to be a source of regional heterogeneity in action potential conduction and repolarization. Moreover, altered patterns of expression may be important factors in arrhythmogenesis associated with increased QTc dispersion and triggered activity (Antzelevitch and Sicouri, 1994).

Cardiac conduction defects

117

Fig. 1. Ionic basis of the cardiac action potential. Schematic diagrams of the surface electrocardiogram (A) and intracellular action potential (B) show the temporal relationships between the two waveforms. In (A) the letters correspond to the standard ECG nomenclature. In (B) the phases are numbered according to standard classi®cation described in (C). The ion channels responsible for the different phases of the action potential (C) are related to the known genes that encode their major structural components.

Long-QT syndrome Clinical aspects Long-QT syndrome (LQTS) is a serious cardiac disorder that causes loss of

118

G.E. Kirsch

consciousness (syncope) and sudden death in otherwise healthy individuals. LQTS is typically detected in childhood as syncope or cardiac arrest often triggered by emotional stress or physical exertion. The underlying problem is episodic ventricular arrhythmia characterized by Torsade de Pointes (TdP, an unusual disruption of the ECG in which the QRS complex twists around the isoelectric axis) and ventricular ®brillation. Between episodes, LQTS patients show delays in repolarization such that the QTc interval (corrected for heart rate) of the surface ECG exceeds 0.45 s (or .0.47 s in asymptomatic individuals, Keating et al., 1991a,b), or other ECG disturbances (such as T-wave alternans) associated with abnormal repolarization (Schwartz et al., 1995a). Often a family history of syncopal episodes or sudden death exists, but sporadic cases have been reported possibly involving de novo mutations (Koh et al., 1994; Curran et al., 1995). Two forms of inherited LQTS are recognized: Jervell±Lange±Nielsen syndrome, which is associated with congenital deafness is transmitted in an autosomal recessive pattern, and Romano±Ward syndrome, the more common form, which is not associated with deafness and follows an autosomal dominant pattern of inheritance. In addition, acquired LQTS syndrome that most often occurs as a side effect of antiarrhythmic drugs that prolong the action potential or in association with hypokalemia, has similar characteristics but differs in response to b-adrenergic interventions (Schwartz et al., 1995a,b). In both the acquired and inherited forms, LQTS is characterized by TdP that can lead to ventricular ®brillation and cardiac arrest. Traditionally the acquired form has been characterized as pause-dependent (TdP preceded by a long pause or by a short±long±short interval), whereas the inherited form is adrenergic-dependent (TdP follows a sudden increase in sympathetic tone; Jackman et al., 1988). Whether the two forms can be distinguished by these criteria has become less certain in view of clinical data that pause dependent TdP is often observed in congenital LQTS (Viskin et al., 1996) and some subsets of congenital LQTS are not adrenergicdependent (Schwartz et al., 1995a). Moreover, the same ion channels appear to be involved in both types of LQTS. Recent genetic analysis has pinpointed mutations in several different genes which Table 1 Congenital forms of long QT syndrome a

Disease

Chromosome

Gene

Ionic current

Clinical classi®cation

LQT1 LQT2 LQT3 LQT4 LQT5

11p15.5 7q35-36 3p21-24 4q25-27 21q22

KCNQ1 (KvLQT1) HERG SCN5A (hH1) ? KCNE1 (minK, IsK)

IKs IKr INa ? IKs

RWS and JLNS RWS RWS RWS RWS and JLNS

a

JLNS and RWS are the Jervell±Lange±Nielsen and Romano±Ward variants of the long QT syndrome.

Cardiac conduction defects

119

lead to a similar clinical presentation. Conversely, in some instances mutations of the same genes can produce clinically distinct phenotypes depending on homo- or heterozygosity. The molecularly distinct inherited forms are denoted LQT1±LQT5 (Table 1), but as we shall see the ionic basis and clinical manifestations can sometimes overlap. Both RWS and JLNS are caused by repolarization defects usually associated with different mutations in either of two K 1 channel genes (KCNQ1 and KCNE1). RWS is generally associated with heterozygous and JLNS with homozygous mutations in these genes. RWS can also be caused by heterozygous mutations in an unrelated K 1 channel gene (HERG), a Na 1 channel gene (SCN5A) and a presently unidenti®ed gene in chromosome 4q25±27. A recent study by Zareba et al. (1998) has provided initial correlations between genotype and clinical outcome in congenital LQTS. They found that LQT1 and LQT2 patients (K 1 channel defects) showed more frequent occurrence of cardiac events than LQT3 (Na 1 channel) patients. But a fatal outcome was more likely for the less frequent events in LQT3 than in LQT1 or LQT2 patients. Therefore, the cumulative mortality from K 1 channel and Na 1 channel defects was nearly equal. Basic mechanisms of LQTS Both the acquired and inherited forms appear to involve abnormal action potential repolarization and hence the possible involvement of cardiac ion channels, particularly K 1 and Na 1 channels. The underlying mechanism is that delayed repolarization, caused by suppression of outward, repolarizing currents conducted by K 1 channels (Roden, 1993; Curran et al., 1995) or by potentiation of inward, depolarizing currents conducted by Na + channels (Boutjdir et al., 1984), allows a reactivation of Ca 21 channels (January and Riddle, 1988), and secondary triggered activity due to early afterdepolarizations (EAD; Wit and Rosen, 1991; Zhou et al., 1992; Tan et al., 1995). The inherited form is caused by defects in ion channel structural genes that lead to abnormalities in channel expression or function. The acquired form appears to involve drug or environment-induced abnormalities of the same ion channels. Whether susceptibility to acquired LQTS has a genetic component has not been clari®ed. The notion that an ion channel defect underlies inherited LQTS is supported on theoretical grounds. Na 1 channel involvement derives from the fact that although rapidly inactivating sodium currents are primarily responsible for the phase 0 upstroke (Fig. 1), a late, slowly inactivating phase contributes a depolarizing current that helps maintain the phase 2 plateau (Gintant et al., 1984; Carmeliet, 1987; Kiyosue et al., 1989). Pharmacological agents that slow down Na 1 inactivation (e.g. aconitine, veratridine and anthopleurin A) mimic some aspects of LQTS (e.g. EADs and delayed repolarization) in ventricular muscle and Purkinje ®bers (Imanaga, 1967; Shimizu et al., 1979; Boutjdir et al., 1984). The potential involve-

120

G.E. Kirsch

ment of Ca 21 channels is supported by pharmacological results. Ca 21 channel agonists (e.g. BAY K8644) have been shown to generate EADs in isolated cell systems (January et al., 1988) and Ca 21 channel blockers have been used successfully to prevent EADs in LQTS patients (De Ferrari et al., 1994). K 1 channel involvement is clearly present in the acquired form of LQTS where the most common mechanism is drug-induced blockade of K 1 channels (Roden, 1990). The observations that quinidine at therapeutic concentrations, but not at higher doses is known to induce TdP (Roden et al. 1986), and that amiodarone rarely causes TdP can be rationalized from the pharmacological pro®les of these drugs. Quinidine, although included in the class I category, blocks K 1 channels at therapeutic concentrations (Roden and Hoffman 1985; Salata and Wasserstrom, 1988) and at higher levels blocks both K 1 and Na 1 channels (Snyders and Hondeghem, 1990; Balser et al., 1991). Thus, at low concentrations it would tend to delay repolarization, but at high concentrations the additional block of Na 1 channels would tend to restore repolarization. Similarly, amiodarone is a broad spectrum channel blocker with relatively little speci®city between Na 1, Ca 21 and K 1 channels (Kodama et al., 1997), and therefore its blockade of K 1 channels is compensated by Na 1 and Ca 21 channel block. Hypokalemia may act as a contributing factor in acquired LQTS on the basis of in vitro results from guinea pig ventricular myocytes where reduced extracellular [K 1] suppressed inward recti®er K 1 currents, prolonged the plateau phase and generated EADs (Hiraoka et al., 1992). Although inherited LQTS cannot be cured, once diagnosed, the life-threatening arrhythmias are usually controllable by therapeutic strategies based on antiadrenergic interventions (b-adrenergic blockade or left cardiac sympathetic denervation), pacemaker implant, or cardioverter de®brillator implant (Moss, 1998). However, these treatments do not correct the underlying abnormality. Recent results have opened the possibility of new gene-speci®c strategies based on ion channel-selective drugs that can correct the repolarization defect (Priori et al., 1997). The effects of mutations on channel function are broadly classi®ed into loss-offunction and altered-function categories. Most of the Na 1 channel mutations are of the altered-function type, as would be predicted, since increased amounts of persistent inward current rather than an overall reduction in Na 1 conductance would be required to prolong repolarization. By contrast, decreased outward current through K 1 channels would be required to produce a similar effect on repolarization. In principle a decrease could result from loss-of-function mutations in which a reduction in conductance is proportional to the amount of non-functional protein subunits produced (haplo-insuf®ciency) or through an alteration in gating that prevents mutant channels from opening normally in response to stimulation. The tetrameric structure of K 1 channels opens the possibility of more severe effects through dominant negative inhibition. In the heterozygous condition mutant subunits could be non-functional but retain the ability to assemble with normal a-subunits to produce

Cardiac conduction defects

121

non-functional, heteromeric channels. Thus, interfering with the functionality of the normal subunits (e.g. Fig. 2) and extending the inhibitory effect beyond the proportionality predicted by simple haplo-insuf®ciency. Dominant negative effects are often observed in LQTS mutations associated with K 1 channels but not for Na 1 channels that consist of a single a-subunit. LQT1 and LQT5 Genetic linkage analysis in several families localized LQTS to a site on the short arm of chromosome 11 (11p15.5; Keating et al., 1991a,b). The Harvey ras-1 (H-ras1) proto-oncogene was originally considered to be a candidate since its product, a small GTP-binding protein, had previously been shown to modulate a cardiac K 1 channel (Yatani et al., 1990). But it was eliminated as a causative factor when no mutations of the gene were found in affected individuals (Schulze-Bahr et al., 1995). Instead, a previously unknown K 1 channel gene (KvLQT1, also denoted KCNQ1) was found to be the actual LQT1 target (i.e. the ®rst gene known to be associated with LQTS) on the basis that intragenic deletions or missense mutations were found in affected members of 16 different LQT1 families (Q. Wang et al., 1996). These mutations have been hypothesized to reduce cardiac K 1 current and thereby prolong action potential repolarization. When ®rst isolated, the partial clone KvLQT1 (lacking a complete N-terminus) was suggested to encode the a-subunit of a voltage-gated K 1 channel based on hydropathy analysis and sequence comparison to other known K 1 channels of the voltage-gated, six-transmembrane segment type (Kv channels). RNA transcripts of the gene were detected in heart tissue, but whether they encoded a functional channel was not known (Q. Wang et al. 1996). Isolation of full length KvLQT1 (Sanguinetti et al., 1996b; Yang et al., 1997) and heterologous expression revealed a voltage-gated K 1 channel with gating properties unlike those of any known cardiac K 1 currents. Although it

Fig. 2. Dominant negative effects of electrically silent Kv5.1 channel a subunits on Kv2.1 expression. Typical current±voltage families recorded in voltage-clamped Xenopus oocytes injected with Kv2.1 cRNA at 0.06 ng (A), Kv5.1 alone at 0.6 ng (B), and a mixture of 0.06 ng of Kv2.1 and 0.6 ng of Kv5.1. Currents were evoked by test pulses of 230 to 50 mV (20 mV increments) from a holding potential of 280 mV. The cells were bathed in normal Ringers solution.

122

G.E. Kirsch

showed a general similarity to IKs (the slow component of the cardiac delayed recti®er current), including potentiation by cAMP and blockade by clo®lium but not dofetilide (Yang et al., 1997), the ionic current produced by heterologous expression of KvLQT1 activated too rapidly and showed a greater degree of inactivation compared to IKs in cardiac myocytes. These results suggested that the channel responsible for IKs might be a heteromer assembled from a-subunits and modulatory b-subunits. Coexpression of KvLQT1 with an auxiliary b-subunit, minK (also called IsK, encoded by KCNE1) recapitulated the functional characteristics of IKs including slow activation gating and a lack of detectable inactivation (Barhanin et al., 1996; Sanguinetti et al., 1996b). Moreover, coexpression of KvLQT1 and minK resulted in much larger macroscopic K 1 current than KvLQT1 alone (Romey et al. 1997). Thus the molecular basis of IKs that controls phase 3 repolarization, is a channel formed by coassembly of polypeptides encoded by both KvLQT1 and KCNE1 genes. Furthermore, mutations in either KCNE1 or KvLQT1 might cause suppression of IKs. In fact, several mutations in KCNE1 (located on the long arm of chromosome 21), have been shown to be associated with LQTS in some families (Tyson et al., 1997). Therefore these mutations have been assigned to the LQT5 form of the disease (Duggal et al., 1998) although both LQT1 and LQT5 affect the same ionic current and have a similar clinical manifestation. Functional IKs channels are thought to be assembled as oligomers of four asubunits (Fig. 3) by analogy with other Kv channels. Addition of one or more bsubunits, although not required for ion conduction, modify the functional properties. Thus, suppression of IKs could result from a-subunit mutations that produce truncated, non-functional polypeptides and might be expected to have a dominant negative effect. Such is likely to be the case for the frame shift and splice donor mutations that cause premature termination (Table 2). Alternatively, mutations in either a- or b-subunits that produce full length protein (particularly missense or deletions) might be expected to modify or even inhibit channel function without suppressing channel assembly. Thus, heteromers might be non-functional through the inclusion of one or more mutant subunits or a failure to reach the plasmalemma. The vast majority of the missense and deletion LQT1 mutations appear to have this loss-of-function mechanism. Partial reduction of ionic currents might also result from alterations in gating or conduction in heteromeric channels (i.e. reduced single channel conductance or decreased probability of opening). Relatively few LQT1 mutations fall into this altered-function category. Nonetheless in all cases the net effect would be an inhibition of IKs in heterozygous individuals expressing both normal and mutated alleles, and would account for the autosomal dominant pattern of inheritance characteristic of the Romano±Ward form of the disease. Several missense mutations (Table 2) in the KvLQT1 gene have been characterized by engineering the nucleotide changes into cDNA clones and expressing the mutated channels in heterologous systems. Low-level constitutive expression of an

Cardiac conduction defects

123

Fig. 3. Membrane spanning topology of the a subunit KvLQT1 (LQT1) and the b subunit minK (KCNE1, LQT5). For KvLQT1, the N- and C-termini are located on the intracellular side of the membrane. Putative transmembrane segments are denoted by membrane-embedded cylinders. The linker between segments 5 and 6 (P) forms part of the ion-conducting pore. minK is represented as a separate polypeptide with a single transmembrane (TM) segment. Functional channels are assembled from four of a subunits and at least two b subunits (Sesti and Goldstein, 1998). The approximate location of LQT1 mutations in KvLQT1 and LQT5 mutations in minK are identi®ed using notation described in the legends of Tables 2 and 3. Standard single letter amino acid abbreviations are used.

endogenous KvLQT1-like K 1 channel in Xenopus oocytes (Sanguinetti et al., 1996b) makes this system less than ideal for detecting poorly expressing exogenous channels, but has the advantage of allowing consistent delivery of cRNA by microinjection. Alternatively, mammalian cell lines such as Chinese hamster ovary cells and human embryonic kidney cells have been employed using vector-mediated transfection. In nearly all cases the mutant a-subunit (homomeric expression) by itself or together with minK b-subunits, did not express K 1 currents. The addition of wild-type a-subunits (heteromeric expression) resulted in markedly diminished levels of current when compared with equivalent amounts of wild-type by itself. Moreover, the gating kinetics of the residual currents observed in heteromeric expression usually were indistinguishable from wild-type, indicating a total lossof-function in channels that incorporated mutant subunits.

124

G.E. Kirsch

Table 2 KvLQT1 (LQT1) mutations a

Mutation

F167W/DG168 R174C A178P/T G189R Fs/R190 R190Q V254M G269D L273F A300T W305S G306R Sp/V307 T312I I313M G314S Y315S D317N G325R DF339 A341E A341V L342F sp/A344 G345E R366P fs/Q484

Location

S2 S2-S3 link S2-S3 link S2-S3 link S2-S3 link S2-S3 link S2-S3 link S5 S4-S5 link P P P P P P P P P S6 S6 S6 S6 S6 S6 S6 S6 C-term

Disease

RWS RWS RWS RWS RWS, JLNS RWS RWS RWS RWS RWS (homozygous) JLNS (homozygous) RWS RWS RWS RWS RWS RWS RWS RWS RWS RWS RWS RW RWS RWS RWS JLNS, RWS

Phenotype Homomeric

Heteromeric Gating

n.x. n.x.

d.n. d.n.

n.e. n.e.

n.x. x x n.x. n.x.

d.n. d.n. n.d. n.e. d.n.

n.e. n.e.

n.x.

d.n.

n.e.

n.x. n.x. n.x.

d.n. d.n. d.n.

n.e. n.e. n.e.

n.x.

d.n.

n.e.

n.e. n.e.

Relatively few mutations fall into the altered-function category. In fact only one mutation, L273F in the S5 segment (Fig. 3), produced measurable K 1 current when expressed as a homotetramer. Its net effect in heteromeric expression with wild-type subunits was to reduce current levels, but the mechanism is unclear since gating and single channel conductance were not fully examined (Shalaby et al., 1997). Another mutation, R555C in the C-terminus, expressed current as a homotetramer but only when minK was present. Moreover, its activation required depolarizations 50 mV greater than in wild-type and its closing rate upon repolarization was doubled. Thus,

Cardiac conduction defects

125

Table 2 (continued) Mutation

Location

Disease

Phenotype Homomeric

fs/Q544 R555C

C-term C-term.

JLNS RWS

Heteromeric Gating

n.x. n.e. x (with minK) d.n.

n.e. shift

a The codon numbering system is for the full-length human cardiac KvLQT1 clone (Yang et al., 1997). Some original reports were based on partial clones with truncated N-termini. For instance, A341V was originally reported as A212V (Wang et al., 1996) and A246V (Li et al., 1998). Missense mutations resulting in single substitutions are abbreviated using standard notation (wild-type amino acid residue followed by codon number and substituted residue). F168W/DG169 is a three base-pair deletion that results in both a substitution (F168W) and a deletion (DG169). Fs/R190 is an insertion at codon 190 that results in a frame shift and premature termination. Sp/V307 is a three base-pair deletion (DGGT) in the pore region that disrupts a splice-donor sequence in the pore region and causes premature termination. Sp/ A344 is a single nucleotide substitution (g ! a) that disrupts a splice-donor sequence in S6 and causes premature termination. DF339 is a three base-pair deletion that results in a deletion of Phe 339. Fs/Q484 is an insertion/deletion at codon 484 that results in a frameshift and premature termination. Fs/Q544 is an insertion/deletion at codon 544 that results in a frameshift and premature termination. Abbreviations: x, expression observed; n.x., no expression observed; d.n., dominant negative suppression of wild-type expression; n.d., not determined, n.e., no effect; shift, positive shift of the voltage-dependence of steady-state activation. Unless otherwise indicated RWS and JLNS are found in individuals who are heterozygous and homozygous, respectively for the mutation. References: Russell et al. (1996), Wang et al. (1996); Chouabe et al. (1997), Donger et al. (1997), Neyroud et al. (1997), Shalaby et al. (1997), Tanaka et al. (1997), Wollnik et al. (1997), Ackerm an et al. (1998), Kanters et al. (1998), Li et al. (1998), Priori et al. (1998), Saarinen et al. (1998).

its net effect appears to be a stabilization of the closed state of the channel, resulting in a positive shift in the steady-state voltage-dependence of activation (Chouabe et al., 1997). In homozygous individuals these channels would be unlikely to open in the physiological range of potentials. The coexpression of wild-type and R555C produced channels with a 30 mV shift of the activation-voltage relationship that may be responsible for the observed dominant negative effect on heterologously expressed currents (Chouabe et al., 1997). Thus, R555C is one of the few LQT1 mutations that have been reported to in¯uence the gating properties of the KvLQT1 channel. The location of this mutation in the intracellular C-terminal tail gives little insight into its mechanism since this region has not previously been implicated as an important determinant of gating in voltage-gated K 1 channels. Moreover, the Cterminus of regulatory minK subunits that markedly alter KvLQT1 gating (Barhanin et al., 1996; Sanguinetti et al., 1996b), does not interact strongly with this region of KvLQT1 (Romey et al., 1997). Therefore, the C-terminus of KvLQT1 may play a novel role in regulating IKs gating. Another KvLQT1 mutation that results in altered function, A300T, is unusual

126

G.E. Kirsch

Table 3 KCNE1, (LQT5) mutations a

Mutation

Location

Disease

Phenotype

T59P 1 L60P S74L D76N

TM C-term C-term

JLNS RWS JLNS, RWS

? Gating d.n./gating

a KCNE1 encodes minK (also called IsK), an auxiliary K 1 channel subunit. TM is the transmembrane segment of minK. JLNS and RWS are the Jervell±Lange±Nielsen and Romano±Ward variants of the long QT syndrome. d.n., dominant negative. References: Takumi et al. (1991), Schulze-Bahr et al. (1997); Splawski et al. (1997); Tyson et al. (1997); Duggal et al. (1998).

because it is associated with LQTS only in homozygous individuals (Priori et al., 1998). In Xenopus oocytes A300T when co-expressed with minK expresses onefourth the amplitude of currents produced by wild-type. The mechanism is unclear but cannot be attributed to altered voltage dependence of activation since a negative shift was observed. Reduced single channel is a possibility since Ala 300 occupies a position in the P region (Fig. 3). Compared to the loss of function produced by other mutations, the relatively mild effects of A300T are consistent with a recessive pattern of inheritance. In addition to mutations of the KvLQT1 a-subunit, mutations in its auxiliary bsubunit, minK (Table 3), have been linked to both RWS and JLNS (LQT5; Duggal et al., 1998). The minK protein (Fig. 3), unlike its KvLQT1 partner, is a small polypeptide of 130 amino acid residues with only a single putative transmembrane segment (Takumi et al., 1988). It is encoded by the KCNE1 gene located on chromosome 21q22 (Chevillard et al., 1993). The regulatory in¯uence of minK on KvLQT1 is multi-faceted. In heterologous systems expression of KvLQT1 alone yields a K 1 current that activates rapidly and deactivates slowly with a characteristic `hooked' tail current waveform that arises from a relatively rapid recovery from inactivation (Sanguinetti et al., 1996b). Coexpression with minK has two major effects: it increases the amplitude of the K 1 currents by at least 2-fold and slows the time course of activation (Barhanin et al., 1996). The underlying mechanism appears to involve protein±protein interactions in which the intracellular C-terminus of minK interacts with the P-region of KvLQT1 (Romey et al. 1997) such that minK forms an integral part of the KvLQT1 pore (K. Wang et al., 1996; Tai and Goldstein, 1998). Alteration of the KvLQT1 pore structure by minK residues would be expected to have profound effects on both ion conduction and gating by analogy with other Kv channels (De Biasi et al., 1993; Heginbotham et al., 1994). Coexpression of minK with KvLQT1 has been reported to increase single channel conductance (Pusch, 1998; Sesti and Goldstein, 1998) in agreement with its effect on macroscopic current (Barhanin et al., 1996). In addition, minK in¯uences gating

Cardiac conduction defects

127

by causing a nearly complete abolition of KvLQT1 inactivation (Pusch et al., 1998; Tristani-Firouzi and Sanguinetti, 1998), and would therefore be expected to increase the amount of activatable current. It is noteworthy that minK can also interact with HERG K 1 channel subunits to increase the magnitude of macroscopic IKr (Yang et al., 1995; McDonald et al., 1997), but the physiological signi®cance is uncertain (Drici et al., 1998). Although the regulatory mechanism is complex, mutations in minK would be expected to alter the level of expression of IKs in the heart. A double mutation in KCNE1 resulting in the substitutions T59P and L60P has been identi®ed in a JLNS family (Tyson et al., 1997). The functional effects of these mutations are not known but they would be predicted to render minK non-functional and unable to coassemble with KvLQT1 on the basis of their location in the putative helical transmembrane domain and the helix-disruptive nature of proline side chains. A D76N mutation associated with both JLNS and RWS (Schulze-Bahr et al., 1997; Splawski et al., 1997; Duggal et al., 1998) and a S74L mutation associated with RWS (Splawski et al., 1997), both located in the cytoplasmic C-terminus, have been characterized in the Xenopus oocyte system (Splawski et al., 1997; Sesti and Goldstein, 1998). Co-injection of cRNA encoding the D76N minK mutant with KvLQT1 has been reported to produce either no expression of exogenous ionic current (Splawski et al., 1997) or channels with markedly reduced single channel conductance (Sesti and Goldstein, 1998). Both results are consistent with a strong dominant negative effect. To mimic the heterozygous condition, mixtures of D76N mutant, wild-type minK and KvLQT1 were co-injected (Splawski et al., 1997; Sesti and Goldstein, 1998). This combination expressed currents with reduced single channel amplitude and with gating kinetic parameters shifted in the depolarizing direction. Both effects would tend to reduce macroscopic currents. Co-injection of the S74L minK mutant with KvLQT1 had similar effects on both single channel conductance and gating. These mutations suppress macroscopic IKs by changing both the gating and conductance properties of the channels without eliminating the ability of mutant minK subunits to coassemble with KvLQT1. The effects of IKs suppression on action potential repolarization are thought to be more pronounced at fast heart rates. Because of its slow deactivation rate, IKs channels cannot close completely in the brief diastolic intervals at short cycle lengths. This accumulation of open K 1 channels allows adaptive shortening of the action potential by offsetting inward Ca 21 and Na 1 currents during phase 2. A decrease in the number of functional IKs channels would have its most pronounced effect on action potential duration at the fast heart rates associated with physical or emotional stress, as observed in LQTS patients.

128

G.E. Kirsch

LQT2 LQTS was known to be genetically heterogeneous before additional loci were identi®ed (Curran et al., 1993; Towbin et al., 1994). Curran et al. (1995) showed linkage to chromosome 7q35-36 in six LQT families and identi®ed several mutations in the HERG (human ether-a-go-go related) gene that encodes another cardiac K 1 channel subunit. This second form of congenital LQTS is denoted as LQT2. HERG was originally cloned from a human hippocampal cDNA library (Warmke and Ganetzky, 1994) and identi®ed as a K + channel of the six-transmembrane type (Fig. 4), by homology with the Drosophila gene (eag) that has been shown to encode a K + channel when expressed in Xenopus ooyctes (Bruggemann et al., 1993). HERG mRNA is expressed at high levels in heart (Curran et al., 1995) and heterologous expression in oocytes yields a K 1 current whose biophysical and pharmacological characteristics are very similar to IKr, the rapidly activating component of the delayed recti®er K 1 current in heart (Sanguinetti et al., 1995; Trudeau et al., 1995). Like IKr, HERG-expressed currents are blocked by the class III antiarrhythmic agent, dofetilide (Spector et al., 1996a) and show an unusual current±voltage

Fig. 4. Membrane spanning topology of the a subunit of HERG (LQT2). N- and C-termini are located on the intracellular side of the membrane. Putative transmembrane segments are denoted by membraneembedded cylinders. The linker between segments 5 and 6 (P) forms part of the ion-conducting pore. Note that the functional channel is assembled from four of these subunits. The approximate location of LQT2 mutations are identi®ed using notation described in the legend of Table 4.

Cardiac conduction defects

129

relationship in which inward recti®cation (i.e. more ef®cient conduction of inward than of outward currents) results from rapid inactivation concurrent with the activation process upon depolarization and rapid recovery from inactivation concurrent with a slow deactivation process upon repolarization (Smith et al., 1996; Spector et al., 1996b). The structural basis of this form of rapid inactivation, however, is unlike that of the well-studied N-type inactivation of Drosophila Shaker K 1 channels in which the peptide residues in the amino terminus act as blocking particles that occlude the intracellular mouth of the pore (Hoshi et al., 1990). In fact, truncation of the amino terminus of HERG does not eliminate inactivation (Schonherr et al., 1996; Spector et al., 1996a) instead, inactivation has been found to be highly sensitive to mutation of residues in the pore region (Schonherr and Heinemann, 1996; Smith et al., 1996; Herzberg et al., 1998). Thus, LQT2 mutations especially those in the pore region, that affect both gating kinetics and ion conduction of the open channel have the potential to drastically alter the fundamental current±voltage relationship of IKr. Despite the qualitative similarities with IKr, currents expressed by HERG activate and deactivate at slower rates than IKr in native myocardial cells (Sanguinetti and Jurkiewicz, 1990; Yang et al., 1994). This has raised questions about the molecular identity of IKr and the possibility of alternate isoforms or auxiliary subunits. Since the original HERG clone was isolated from brain, a search for cardiac-speci®c isoforms has been mounted. Two alternatively processed, cardiac-speci®c isoforms have been identi®ed: a splice variant that has a much shorter intracellular amino terminus than HERG (Lees-Miller et al., 1997; London et al., 1997) and a splice variant with truncated intracellular carboxyl-terminus (Kupershmidt et al., 1998). When co-expressed with HERG or by itself, the truncated amino-terminal isoform displays faster deactivation gating kinetics (Lees-Miller et al., 1997; London et al., 1997). The truncated carboxyl terminus variant did not express current by itself but upon coexpression with full-length HERG produced currents with accelerated activation kinetics. Thus, in heterologous systems coexpression produced currents that more closely approximated IKr and, therefore heteromers of full-length and truncated subunits may form the channels responsible for IKr in vivo. Thus far functional analysis of LQT2 mutations has only been performed only in homomeric, full-length HERG channels. Nonetheless, these results together with analysis of arti®cial terminal deletion constructs of HERG (Schonherr and Heinemann, 1996; Spector et al., 1996a) indicate important roles for both amino and carboxyl termini in the regulation of HERG gating. In principle, HERG mutations that suppress the rise in potassium conductance mediated by IKr would be expected to slow phase 3 repolarization and prolong the QT interval. This could be achieved either by decreasing the number of functional channels or by impairing the gating mechanism that opens them. Decreased numbers of functional channels could result either from loss of function mutations or a more

130

G.E. Kirsch

severe dominant negative effect. Similarly, gating defects could reduce functional expression by multiple mechanisms. For instance, a depolarizing shift in the voltage dependence of activation could prevent channels from opening in the physiological range or a hyperpolarizing shift in the voltage dependence of inactivation that induces inactivation in diastole could reduce the pool of available channels upon depolarization. Phenotypic analysis of mutated channels expressed in heterologous systems has been used to characterize about half of the known LQT2 mutations. In nearly all cases (Table 4) expression of the mutant subunit by itself produced no detectable ionic current. But when coexpressed with wild-type subunits either suppression (dominant negative effect) or no effect on wild-type function was observed. Underlying mechanisms include defects in protein traf®cking resulting in failure of the channels to reach the surface membrane (e.g. Y611H and V822M; Zhou et al., 1998), failure of mutant channels to coassemble with wild-type (e.g. fs/421, DI500-F508, Sanguinetti et al., 1996a), or increased levels of inactivation in heteromeric channels (e.g. A614V and V630L, Nakajima et al., 1998). In some cases, however, the biophysical defect is unclear. For instance, the N470D mutant subunit by itself expresses ionic current but when coinjected with wild-type causes a modest reduction in wild-type expression without gross alteration of gating (Sanguinetti et al., 1996a). This mutation is unlikely to reduce single channel conductance given its location far from the pore region, but a subtle effect on gating that reduces the probability of channel opening is possible. A nearby mutation T474I is more problematic because of disagreements in published data. Nakajima et al. (1998) reported that T474I injected into Xenopus oocytes did not express current whereas transfection of the same construct in HEK 293 cells was reported to express current with altered gating properties (Zhou et al., 1998). The observed alteration, a large hyperpolarizing shift of the voltage dependence of activation however, would be expected to increase rather than suppress K 1 currents. Disagreements also exist in reports on the fs/421 mutation. This mutation produces a non-functional truncated polypeptide that has no effect when coinjected with wild-type in Xenopus oocytes (Sanguinetti et al., 1996a), but produces a modest dominant negative effect when cotranfected with wild-type in COS cells (Li et al., 1997). Whether these discrepancies are due to different expression systems remains to be determined. LQT3 A decrease in outward K 1 currents through voltage-gated KvLQT1 and HERG channels would be expected to increase action potential duration by prolonging phase 3 repolarization. An alternative mechanism would be to increase the duration of the phase 2 plateau through augmentation of inward currents carried by either voltage-gated Ca 21 or Na 1 channels. To date no involvement of Ca 21 channel genes

Cardiac conduction defects

131

Table 4 HERG (LQT2) mutations a

Mutation

Dbp1261 N470D T474I DI500-F508 A561V A561I G572C N588D I593R G601S Y611H V612L A614V G628S N629D N629S V630A V630L N633S sp/intron 3 V822M

Location

S1 S2 S2-S3 link S3 S5 S5 S5 S5 P P P P P P P P P P P C-term C-term

Phenotype Homomeric

Heteromeric

Gating

n.x. x. n.x. n.x. n.x.

d.n. d.n. d.n. n.e. d.n.

n.e. n.e. n.e. n.e. n.e.

n.x. n.x.

d.n. d.n.

Inactivation shift

n.x.

d.n.

Inactivation shift

n.x.

n.x.

n.x. n.x.

a Hetero- and homomeric, respectively refer to expression of mutant subunits alone and to their coexpression with wild-type subunits. d.n., dominant negative, i.e. no expression of ionic currents when cRNA is injected into oocytes but co-injection with wild-type cRNA results in decreased levels of current and, in some cases altered gating; n.x., no expression of ionic currents after cRNA is injected into oocytes; n.e., no effect. Dbp1261 is a single base pair deletion that causes a frame shift and premature termination in the S1 segment. DI500-F508 is a deletion of nine amino acid residues in the S3 segment. Sp/intron 3 is a single base substitution that disrupts a splice-donor sequence in intron 3 affecting the cyclic nucleotide binding domain in the C-terminus. References: Curran et al. (1995); Benson et al. (1996); Dausse et al. (1996), Sanguinetti et al. (1996), Li et al. (1997); Nakajima et al. (1998), Satler et al. (1998), Splawski et al. (1998); Zhou et al. (1998).

in LQTS has been found. However, in a small number of families LQTS has been linked to SCN5A, the gene that encodes the cardiac Na 1 channel. This form of the disease is referred to as LQT3. In animal models, Na 1 channel blockade by tetrodotoxin at submaximal concentrations shortens the duration of the action potential (Attwell et al., 1979; Coraboeuf et al., 1979). The biophysical basis appears to be an inhibition of persistent inward current conducted by Na 1 channels that do not

132

G.E. Kirsch

completely inactivate (Table 5). Although small in magnitude, this persistent depolarizing current could have a pronounced effect on the action potential plateau since membrane input impedance is elevated during this phase. Several mutations in SCN5A have been linked to LQT3 (Fig. 5) and in all cases have been shown to affect the inactivation gating mechanism. The most common biophysical defect is a marked increase in persistent late current. At the single channel level, current appears in two forms (Fig. 6), either brief, dispersed reopenings that probably arise from decreased stability of the inactivated state (e.g. N1325S and R1644H mutants, Dumaine et al., 1996), or prolonged bursts of openings that probably result from a modal switch in gating (DK1505-Q1507 mutant, Bennett et al., 1995). Wild-type channels can switch from a normal mode of gating in which brief openings are terminated by entry into an absorbing inactivated state, to an inactivation-de®cient mode in which long openings and repetitive reopenings cluster to form long bursts. Mode switching is extremely infrequent in wild-type channels but increases dramatically in the DK1505-Q1507 mutant (Bennett et al., 1995). The mechanism of mode switching and its regulation are unclear. A similar phenomenon has been observed in skeletal muscle (SCN4a) and brain Na 1 channel (SCN2a and SCN3a) isoforms, where, in heterologous systems, the absence of a b1 subunit promotes bursting (Bennett et al., 1993; Patton et al., 1994). This is unlikely to be the mechanism for cardiac Na 1 channels where the b1 subunit, although present in native tissue, has no effect on modal switching in heterologous expression (Makita et al., 1994). Another brain isoform, SCN8a, generates persistent current in the presence of b1 subunits (Smith et al., 1998) but in its primary structure has none of the amino acid substitutions corresponding to LQT3 mutations. Two additional LQT3 mutations have been uncovered recently that have different biophysical actions. R1623Q was identi®ed as a de novo mutation located in the S4 transmembrane segment of domain 4 (Fig. 5), that resulted in an especially severe form of LQTS (Yamagishi et al., 1997). Electrophysiological analysis revealed no persistent current, but rather a marked slowing of inactivation (Kambouris et al., 1998; Makita et al., 1998). The corresponding position, R1448, in the SCN4a skeletal muscle isoform, when mutated to Cys or His in some forms of paramyotonia, produces a similar phenotype (Chahine et al., 1994). Based on these and other results the D4/S4 region is believed to be a critical determinant of Na 1 channel inactivation. Another mutation has been found in an intracellular C-terminal region not previously correlated with channel function. D1790G has been identi®ed by genetic linkage to LQTS in a single family (Benhorin et al., 1997). Although this residue is highly conserved in Na 1 channels of different species, the mutation does not produce a distinctive phenotype when expressed by itself in human embryonic kidney cells. However, coexpression with b1 subunits produces a hyperpolarizing shift of the voltage-dependence of steady-state inactivation compared to coexpression of wild-type and b1 subunits (An et al., 1998). The result is surprising in view of

Cardiac conduction defects

133

Table 5 SCN5A mutations (LQT3 and IVF) a

Mutation

Disease

Location

sp/intron 7 R1232W 1 T1620M D4/S3-S4 link N1325S Fs/1398 DK1505-Q1507 R1623Q R1644H D1790G

IVF IVF inactivation shift LQT3 IVF LQT3 LQT3 LQT3 LQT3

D1 truncation D3/S1-S2 link D3/S4-S5 link D3/P truncation D3-D4 link D4/S4 D4/S4 C-term

Phenotype

Persistent current Persistent current Slow inactivation Persistent current Inactivation shift

a

sp/intron 7 is a two base-pair insertion that disrupts a splice-donor site in intron 7 and predicts truncation of the polypeptide in the S2±S3 linker of domain 1. R1232W 1 T1620M is a double missense mutation. Fs/1398 is a single base-pair deletion in codon 1398 that causes a frameshift and premature termination in the pore region of domain 3. DK1505-Q1507 is a deletion of three residues in the intracellular linker between domains 3 and 4. LQT3 is chromosome 3 linked long QT-syndrome and IVF is the Brugada variant of idiopathic ventricular ®brillation. References: Bennett et al. (1995); Wang et al. (1995a); Wang et al. (1995b)), Dumaine et al. (1996); Benhorin et al. (1997), Yamagishi et al. (1997); An et al. (1998); Chen et al. (1998); Kambouris et al. (1998); Makita et al. (1998).

Fig. 5. Membrane spanning topology of the a subunit of hH1 (SCN5A). The protein consists of a single polypeptide containing four homologous repeats (domains 1±4) each of which has the 6 transmembrane segment (1±6) structure reminiscent of K 1 channels. The linker between segments 5 and 6 (P) forms part of the ion-conducting pore and the interdomain linker between domains 3 and 4 forms an essential component of the inactivation gate. The approximate locations of LQT3 and IVF mutations are identi®ed using notation described in the legend of Table 5.

134

G.E. Kirsch

Fig. 6. Unitary Na 1 channel currents in normal and LQT3 mutant channels. Each column shows representative recordings from cell-attached patches in Xenopus oocytes that were injected with cRNA encoding normal (wild-type) or mutant (N1325S, R1644H and DK1505-Q1507) channels. Test pulses to 210 mV from a holding potential of 2120 mV. Openings are downward de¯ections of the current tracings. External (pipette) solution was normal Ringers.

previous structure-function studies indicating that the a/b subunit interaction surface is located on extracellular domains (Makita et al. 1996; McCormick et al., 1998). Also, the relationship of this particular alteration in gating to prolongation of the action potential is unclear; a negative shift of inactivation would tend to suppress slow mode bursting (Dumaine and Kirsch, 1998) of Na 1 channels. The in¯uence of this mutation on cardiac excitation is therefore likely to be indirect. Experimental models of LQTS Although many of the mutations associated with LQTS have been characterized in heterologous expression systems, direct veri®cation that mutated channels alter the excitability of cardiac myocytes or change the ECG in whole animals has been

Cardiac conduction defects

135

much more dif®cult to obtain. Moreover, development of suitable models is a necessary step toward testing new therapies. At the cellular level mutant phenotypes have been mimicked by application of channel-speci®c gating modi®ers or blockers that increase action potential duration. This approach has been used in the analysis of both LQT2 and LQT3 since selective blockers of HERG channels and gating modi®ers of SCN5A Na 1 channels are well established tools in channel biophysics. Priori et al. (1996) used isolated guinea pig ventricular myocytes treated with either anthopleurin-A (AP-A, a toxin that selectively slows inactivation in Na 1 channels) or dofetilide (a class III drug that selectively blocks HERG K 1 channels) as models of LQT3 and LQT2, respectively. They found that while both treatments prolonged the action potential by ,20%, the addition of mexiletine (a class Ib drug that selectively inhibits late current in LQT3 mutant Na 1 channels, Dumaine et al., 1996) restored nearly normal action potential duration in the LQT3 but not the LQT2 model. The two models could also be distinguished by their response to increased pacing. LQT3 model cells showed 3-fold greater action potential shortening upon increasing the stimulation rate than control cells and 2-fold greater shortening than the LQT2 model cells. These results, together with suggestive data that LQT3 patients bene®t from mexiletine therapy more than LQT2 patients, and show greater QT-interval adaptation to increased heart rate compared to normal (Schwartz et al., 1995b), support the possibility of gene-speci®c treatment of LQTS. However, several questions remain unanswered. For instance, the accuracy of the dofetilide model of LQT2 can be questioned on the basis that in mouse heart dominant negative suppression of IKr by transgenic methods produces a much greater prolongation of the intracellular action potential than treatment with high concentrations of dofetilide (Babij et al., 1998). Also, the notion that Na 1 channel blockers shorten action potential duration only in the presence of mutation- or toxinmodi®ed gating con¯icts with evidence that tetrodotoxin (Attwell et al., 1979; Coraboeuf et al., 1979) and lidocaine (Carmeliet and Saikawa, 1982) can shorten action potential duration in normal animal hearts and more recent evidence for regional differences in the response of guinea pig ventricular action potential to mexiletine (Shimizu and Antzelevitch, 1997). The latter experiments show that in canine ventricular myocytes isolated from the mid-wall (M cells) action potential prolongation induced by HERG blockade can be restored to normal by mexiletine treatment. This result suggests that mexiletine treatment may be valuable in patients with K 1 channel as well as those with Na 1 channel defects, and raise questions about its genetic-speci®city. Another problem is to understand the mechanism responsible for the increase in rate adaptation in the AP-A LQT3 model compared with either the LQT2 model or normal cells. One possibility is that the effect is related to the voltage-dependent binding of this class of Na 1 channel toxins. Upon depolarization, dissociation of toxin molecules restores normal gating. Repolarization of the membrane allows

136

G.E. Kirsch

toxin re-association and re-establishment of the modi®cation (Strichartz and Wang, 1986). Therefore, during the action potential as toxin dissociates, a fraction of the channels would revert to normal gating thereby reducing the amount of depolarizing inward current. Acceleration of the heart rate would decrease the repolarization interval available for toxin re-association, reduce the population of gating-modi®ed channels and shorten the action potential. But this mechanism for rate-dependent adaptation requires state-dependent gating modulation. It is unclear that either wild type or LQT3-mutated Na 1 channels would show similar voltage-dependent characteristics. Moreover, a careful comparison of the mutation-induced and anthopleurin-induced modi®cations in gating reveals other potentially important differences. For instance, AP-A-modi®ed cardiac Na 1 channels show slower than normal inactivation and recovery, but do not display persistent non-inactivating currents (Wasserstrom et al., 1993; Hanck and Sheets, 1995). By contrast persistent current is the hallmark of LQT3 mutant channels (Bennett et al., 1995; Dumaine et al. 1996) and, at least for the DKPQ mutation, both onset and recovery from inactivation are markedly accelerated compared to wild-type (D. Wang et al., 1996). Another sea anemone toxin (ATX-II) that has been used in skeletal muscle Na 1 channels to model periodic paralysis (Cannon and Corey, 1993) and to model LQT3 in cardiac Na 1 channels (Sicouri et al., 1997) may be more accurate since it has been demonstrated to produce persistent currents, but like other site 3 toxins, would suffer from voltage dependent binding (Strichartz and Wang, 1986). Thus, these drug and toxin models of LQTS models provide useful insights but are unlikely to be correct in detail. Genetic manipulation of whole animals although technically dif®cult, provides a conceptually more appealing approach since phenomena can be studied at all levels from whole animal to molecular. Transgenic methods have been used in mice to over-express dominant negative cardiac K 1 channel subunit fragments that selectively suppress expression of normal K 1 currents and therefore inhibit repolarization in a manner similar to that of naturally occurring LQT1 and LQT2 mutations in humans. Several K 1 channel subunits have been targeted using this approach including minK (KCNE1, associated with LQT5), HERG (associated with LQT2), Kv1.5 (a component of the ultra-rapid delayed recti®er, IKur, thus far not implicated in LQTS), and Kv4.3 (a component of Ito1, also not implicated in LQTS). The results have been both intriguing, by providing new insights, and perplexing because of limitations of the mouse model. Direct attempts to introduce LQT2 and LQT5 mutations in mouse heart have been successful in con®rming dominant negative and loss-of-function mechanisms suspected from heterologous expression systems, but have not produced an accurate electrophysiological model of human LQTS. In part this is due to the markedly different characteristics of K 1 currents in mouse and human hearts. Because the basal heart rate in mice is at least ®ve times faster than in humans, K 1 channels that

Cardiac conduction defects

137

display very fast activation kinetics (IKur and Ito1) dominate the repolarization phase of the mouse action potential. Therefore genetic manipulation of the slower IKr and IKs channel components that play an important role in human disease, may have little effect on rhythmicity and ECG waveform in the mouse. This expectation has been ful®lled in several transgenic studies. For instance, Babij et al. (1998) overexpressed the HERG G638S mutation that is associated with LQT2 in humans and is thought to exert a strong dominant negative in¯uence based on Xenopus oocyte expression (Sanguinetti et al., 1996). Over-expression of G628S subunits in mouse heart eliminated IKr and prolonged the action potential in isolated myocytes at room temperature, but had little effect on action potential duration in intact ventricular tissue at physiological temperature, and did not cause QT interval prolongation in vivo (Babij et al., 1998). Similarly, ventricular repolarization was not affected in a minK (KCNE1) knockout mouse model (Charpentier et al., 1998) although inner ear defects similar to those associated with human Jervell±Lange± Nielsen syndrome were detected (Vetter et al., 1996). By contrast, dominant negative suppression in transgenic mice of genes that encode the a-subunits of K 1 channels associated with either ITo (Kv4.2/4.3, Barry et al., 1998) or IKur (Kv1.5, London et al., 1998), produced a marked QT interval prolongation closely correlated with reductions in fast K 1 currents and protein subunit expression. These results illustrate the power of transgenic models for analyzing the functional heterogeneity and structural complexities that underlie adaptation of cardiac K 1 currents to different physiological situations. Thus far no transgenic Na 1 channel manipulations that might provide more accurate models of LQT3 have been reported. Unresolved issues in LQTS Gender differences Although LQTS is inherited in an autosomal dominant fashion, cardiac events occur at least twice as frequently in female as in male carriers of the mutations (Locati et al., 1998). Moreover, administration of class III antiarrhythmic drugs that prolong the QT interval causes TdP three times more often in women than in men (Lehmann et al., 1996). This greater female susceptibility may be related to the fact that the QTc interval in normal adult women is slightly longer than in adult men (Merri et al., 1989). Thus, intrinsic gender-related differences in cardiac repolarization, including differences in K 1 or Na 1 channel function or expression, may be responsible. For instance a lower level of K 1 channel expression in female hearts might increase susceptibility to QT prolongation induced by mutations or drugs that further reduce K 1 conductance. In fact, Lehmann et al. (1997) have shown that in LQT1 and LQT2 patients, adult male carriers showed signi®cantly shorter QTc intervals than adult females. At present, insuf®cient data exist to determine gender-speci®city of the Na 1 channel-linked LQT3 form.

138

G.E. Kirsch

A K 1 channel mechanism is supported by observations in rabbits where a prolonged QT interval in female hearts (at slow heart rates) is correlated with reduced IKr and IK1 current densities (Liu et al. 1998). The mechanism could involve hormonal modulation on ion channel expression or function. Indeed, Kv1.5 and minK mRNA in rabbit heart has been shown to be down-regulated by estrogen (Drici et al., 1996). But the situation is complicated because testosterone gave the same result. Moreover, the expression of HERG mRNA (encoding the a-subunit of IKr) is insensitive to either estrogen or testosterone (Drici et al., 1996), and the hormonal regulation of IK1 expression has not been determined. The minK result may be related to the observation that minK can up-regulate HERG expression in a heterologous system (McDonald et al., 1997). Gender-speci®c or androgen-sensitive regulation of cardiac Na 1 channels has not been examined directly. Androgen treatment has been shown to reduce Na 1 current density in mouse skeletal muscle precursor cells (Tabb et al., 1994) that produce tetrodotoxin-resistant currents through expression of the cardiac Na channel isoform. Sympathetic nervous system involvement Before genetic linkage analyses, LQTS was considered to be primarily of neurogenic origin. A congenital imbalance between left and right sympathetic in¯uence resulting in excessive left nerve activation was suggested as the underlying cause (Schwartz et al. 1975). This idea was supported by experimental evidence of the arrhythmogenic potential of left stellate ganglion stimulation or right sympathetic denervation in animal models (Schwartz et al., 1979; Yanowitz et al., 1966). In fact b-blocking drugs or left cardiac sympathetic denervation remain effective treatments in many cases (Schwartz et al., 1991). The sympathetic imbalance mechanism explained several clinical ®ndings including the observation that cardiac events often were triggered by increased adrenergic activity (e.g. physical or emotional stress) in LQTS patients. Although the sympathetic imbalance hypothesis has fallen into disfavor as a primary cause of LQTS, its role as a contributing factor in at least some forms of the disease needs to be re-evaluated in light of the primary ion channel defects. For instance, the correlation of arrhythmia with increased stress appears to hold for carriers of HERG (LQT2) K 1 channel defects, but not for carriers of SCN5A (LQT3) Na 1 channel defects (Schwartz et al., 1995a,b). The observation that the rat homologue of HERG (Shi et al., 1997), is expressed abundantly in sympathetic neurons and the adrenal gland opens the possibility that hyperexcitability in adrenergic as well as cardiac systems may result from HERG mutations that reduce K 1 conductance. In contrast, expression of SCN5A appears to be restricted to heart and denervated skeletal muscle (Rogart et al., 1989; Gellens et al., 1992) and this might explain the lack of a sympathetic component in LQT3. The importance of sympathetic activation in LQT1/LQT5 is less well-established.

Cardiac conduction defects

139

However, the KvLQT1 gene, like HERG, is expressed strongly in the adrenal gland as well as the heart (Yang et al., 1997). Thus, defects in K 1 channels that make up the majority of congenital LQTS cases may act not only to increase the intrinsic excitability of cardiac membranes, but also to increase sympathetic nerve activity and raise the levels of circulating catecholamines, all of which may contribute to arrhythmogenesis. In addition, under normal conditions IKs is potentiated by adrenergic activation (Walsh and Kass, 1988). Therefore, its suppression by LQT1 or LQT5 mutations would be felt to a greater extent during periods of heightened sympathetic activation. Acquired LQT In contrast to congenital LQTS, much less is known about the ionic basis of the more common, acquired LQTS. Many agents that are capable of prolonging the QT interval can cause TdP.Most notably, quinidine a non-selective antiarrhythmic drug that blocks both Na 1 and many types of K 1 channels, can cause polymorphic ventricular tachycardia, the so-called `quinidine syndrome' at therapeutic dosage in some individuals. Because the drug is used for treatment of atrial ®brillation, it is probably the most common cause of acquired LQTS, accounting for 6% of the general population (Jackman et al., 1984). Other antiarrhythmic agents that can cause acquired LQTS include sotalol, procainamide and disopyramide, have in common with quinidine, the ability to block K 1 channels and prolong the action potential. Acquired LQTS also is associated with factors such as bradycardia, cardiac ischemia, metabolic abnormalities (including hypokalemia and hypomagnesemia), starvation (anorexia nervosa), and various non-cardiac medications including general anesthetics, antibiotics and antihistamines. A mechanistic link between inherited and acquired LQTS has been suggested on the basis that TdP can be caused by mutations in IKr channels encoded by HERG (Curran et al., 1995) as well as by drugs that block IKr (Roden 1990). For instance, the antihistamine terfenadine (Seldane), has been reported to induce TdP in both LQTS patients (Koh et al., 1994) and healthy individuals (Monahan et al., 1990). Moreover, in heterologous expression IKr is sensitive to block by terfenadine (Roy et al., 1996; Suessbrich et al., 1996). In another instance of clinical/experimental correlation, acquired LQTS often occurs on a background of hypokalemia (Schwartz et al., 1995a,b), and in cellular models the ef®cacy of IKr blockade by quinidine increased 10-fold upon reduction of extracellular [K 1] from 8 to 1 mM (Yang and Roden, 1996). Moreover, elevation of extracellular [K 1] by itself promotes IKr (Sanguinetti and Jurkiewicz, 1992) and elevated serum K 1 has been shown to improve repolarization in LQT2 patients (Compton et al., 1996). Therefore HERG channel blockade has been proposed as an explanation for acquired LQTS. However, most of the treatments that lead to acquired LQTS are not HERG-selective, but block other K 1 channels as well. Moreover, additional predisposing conditions, including genetic factors, such

140

G.E. Kirsch

as Na 1 or K 1 ion channel gene mutations, are likely since only a small fraction of patients treated with K 1 channel blockers develop acquired LQTS.

Idiopathic ventricular ®brillation (IVF) Spontaneous ventricular ®brillation in the presence of normal baseline QT intervals and in the absence of structural heart disease may also be associated with ion channel abnormalities. One form of IVF, Brugada syndrome (Brugada and Brugada, 1992), is characterized by an ECG pattern resembling right bundle branch block with elevated ST segment (sometimes referred to as a prominent `J-wave'; Bjerregaard et al. 1994) and a familial history of syncope or sudden cardiac death. In a canine model the J-wave ECG morphology has been associated with heterogeneity in the expression of ITo, the K 1 current that regulates phase 1 action potential repolarization, in different cell layers of the ventricular myocardium. Notably blockade of Ito1 by 4-aminopyridine reduces the heterogeneity and the amplitude of the Jwave (Yan and Antzelevitch, 1996). Therefore, a K 1 channel abnormality was suspected. However, genetic linkage analysis of six Brugada syndrome families has identi®ed three different Na 1 channel (SCN5A) mutations (Chen et al. 1998). Two of the mutations (Table 5) cause premature termination of the a-subunit either in domain 1 or domain 4 (Fig. 5) and would be expected to cause loss of function (StuÈhmer et al., 1989). A third case was linked to a double missense mutation that resulted in substitution of Trp and Met for Arg 1232 and Thr 1620, respectively. Thr 1620 is a highly conserved residue located in a region previously identi®ed with regulation of Na 1 channel inactivation (Chahine et al., 1994) and, indeed the T1620M mutation caused a depolarizing shift in the voltage-dependence of inactivation and an acceleration of recovery from inactivation (Chen et al., 1998). The arrhythmogenic potential of these changes may be related to the theory that antiarrhythmic class Ib drugs such as lidocaine act by producing the opposite effect, i.e. a negative shift in the availability versus membrane potential curve and a slowing of the time course of recovery. However, the correlation of this biophysical defect in the Na 1 channel with the ECG characteristics of IVF is obscure (Yan and Antzelevitch, 1996). Moreover, the relationship of the very different loss-of-function and altered-function mutations to the clinical pattern of Brugada syndrome remains to be established. The involvement of other ion channel genes would seem likely in view of the families that were not linked to SCN5A. Unlike LQTS that is much more prevalent in females, IVF shows a male prevalence, the basis for which remains to be determined (Viskin and Belhassen, 1998).

Cardiac conduction defects

141

Conclusions Genetic linkage analysis together with advances in ion channel structure-function analysis have opened new avenues for understanding the complex relationships between cardiac ionic currents and life-threatening arrhythmia. The advances have been particularly well developed in regard to repolarization abnormalities such as those associated with long QT syndrome and have highlighted the importance of HERG, KvLQT1, minK and SCN5A genes in this form of arrhythmia. However, a substantial gulf still exists between the molecular and clinical realms. A particularly dif®cult problem is the correlation of biophysical defects, that can be characterized in exquisite detail at the molecular level, with complex characteristics of arrhythmias observed at higher tissue and organ levels. Development of accurate animal models as well as more extensive correlations between genotype, molecular phenotype and clinical observations will be required to exploit these new avenues.

Note added in proof Since submission of this manuscript several notable observations have been published. For LQT1 several new mutations in KCNQ1 have been identi®ed and their phenotypes characterized in heterologous expression. These include a two base pair deletion in the S2 segment leading to a truncated, non-functional product (Q. Chen et al., 1999); point mutations R160Q (S2-S3 linker), R243H (S4), R533W and R539W (C-terminus; Chouabe et al., 2000); and point mutations R243C, W248R and E261K (S4-S5 linker, Franqueza et al., 1999). All of these mutations result in either reduced K + channel activity due to altered gating when co-expressed with KCNE1, or no functional expression. Several new mutations associated with LQT2 have been identi®ed and in both the N- (J. Chen et al., 1999) and C-termini (Berthet et al., 1999) of HERG. The Nterminal mutations have been characterized as reducing K + current through altered activation gating (J. Chen et al., 1999). Also, the possible involvement of auxiliary subunits in regulating IKr currents has received signi®cant support form the observation (Abbott et al., 1999) that MiRP1, a minK-related protein, coassembles with HERG to generate K + currents that more closely resemble native IKr than HERG alone. Moreover three missense mutations in the KCNE2 (the gene that encodes MiRP1) were found to be associated with either sporadic or acquired (drug-induced) arrhythmias. KCNE2, therefore, is the ®rst gene to be linked to increased risk of the acquired form of LQTS. New mutations of SCN5A associated with LQT3 include a missense mutation E1784K (cytoplasmic C-terminus) that results in persistent Na + current (Wei et al.,

142

G.E. Kirsch

1999), and two uncharacterized missense mutations (Wattanasirichaigoon et al., 1999), T1645M (D4/S4) and T1304M (D3/S4). Several additional mutations of SCN5A have been identi®ed with Brugada syndrome including: 1795insD (insertion of aspartate at position 1795 in the cytoplasmic C-terminus, Bezzina et al., 1999), R1512W and R1432G (Deschenes et al., 2000). Heterologous expression of all of these mutations resulted in either altered gating or loss of function, but no persistent Na + current. The ECG characteristics associated with the 1795insD mutation however, were unique in showing both long QT interval with elevated ST segment. This observation suggests that although LQT3 and IVF disorders are distinct at the molecular level, the clinical patterns may sometimes overlap. It is also noteworthy that SCN5A has recently been found to be expressed in discrete regions of the brain (Hartmann et al., 1999) raising the possibility that overlapping neurological and cardiac disorders may arise from SCN5A mutations. Finally, although the gene at chromosome 4q25±27 associated with LQT4 has not been identi®ed, a functional interaction of ankyrin B (also localized to 4q25±27) with cardiac Na + channels has been established in mice (Chauhan et al., 2000); myocytes from neonatal mice lacking ankyrin B show abnormally delayed single channel openings, consistent with prolonged action potential duration. This result raises the intriguing possibility that both LQT3 and LQT4 are caused by primary defects in Na + channel function.

Acknowledgements The author's work was supported by the American Heart Association (AHA) and with funds contributed in part by the AHA, Ohio Valley Af®liate.

References Abbott, G.W., Sesti, F., Splawski, I., Buck, M.E., Lehmann, M.H., Timothy, K.W., Keating, M.T., Goldstein, S.A., 1999. MiRP1 forms IKr potassium channels with HERG and is associated with cardiac arrhythmia. Cell 97, 175±187. An, R.H., Wang, X.L., Kerem, B., Benhorin, J., Medina, A., Goldmit, M., Kass, R.S., 1998. Novel LQT-3 mutation affects Na + channel activity through interactions between a- and b1-subunits. Circ. Res. 83, 141±146. Antzelevitch, C., Sicouri, S., 1994. Clinical relevance of cardiac arrhythmias generated by afterdepolarizations. Role of M cells in the generation of U waves, triggered activity and torsade de pointes. J. Am. Coll. Cardiol. 23, 259±277. Attwell, D., Cohen, I., Eisner, D., Ohba, M., Ojeda, C., 1979. The steady state TTX-sensitive (`window') sodium current in cardiac Purkinje ®bres. P¯uÈgers Arch. 379, 137±142. Babij, P., Askew, G.R., Nieuwenhuijsen, B., Su, C.M., Bridal, T.R., Jow, B., Argentieri, T.M., Kulik, J., DeGennaro, L.J., Spinelli, W., Colatsky, T.J., 1998. Inhibition of cardiac delayed recti®er K + current

Cardiac conduction defects

143

by overexpression of the long-QT syndrome HERG G628S mutation in transgenic mice. Circ. Res. 83, 668±678. Balser, J.R., Bennett, P.B., Hondeghem, L.M., Roden, D.M., 1991. Suppression of time-dependent outward current in guinea pig ventricular myocytes. Actions of quinidine and amiodarone. Circ. Res. 69, 519±529. Barhanin, J., Lesage, F., Guillemare, E., Fink, M., Lazdunski, M., Romey, G., 1996. KvLQT1 and IsK (minK) proteins associate to form the I(Ks) cardiac potassium current. Nature 384, 78±80. Barry, D.M., Xu, H., Schuessler, R.B., Nerbonne, J.M., 1998. Functional knockout of the transient outward current, long-QT syndrome, and cardiac remodeling in mice expressing a dominant-negative Kv4 a-subunit. Circ. Res. 83, 560±567. Benhorin, J., Goldmit, M., MacCluer, J., Blangero, J., Goffen, R., Leibovitch, A., Rahat, A., Wang, Q., Medina, A., Towbin, J., Kerem, B., 1997. Identi®cation of a new SCN5A mutation associated with the long QT syndrome. Hum. Mutat., 153. Bennett Jr., P.B., Makita, N., George Jr., A.L., 1993. A molecular basis for gating mode transitions in human skeletal muscle Na 1 channels. FEBS Lett. 326, 21±24. Bennett, P.B., Yazawa, K., Makita, N., George Jr., A.L., 1995. Molecular mechanism for an inherited cardiac arrhythmia. Nature 376, 683±685. Benson, D.W., MacRae, C.A., Vesely, M.R., Walsh, E.P., Seidman, J.G., Seidman, C.E., Satler, C.A., 1996. Missense mutation in the pore region of HERG causes familial long QT syndrome. Circulation 93, 1791±1795. Berthet, M., Denjoy, I., Donger, C., Demay, L., Hammoude, H., Klug, D., Schulze-Bahr, E., Richard, P., Funke, H., Schwartz, K., Coumel, P., Hainque, B., Guicheney, P., 1999. C-terminal HERG mutations: the role of hypokalemia and a KCNQ1-associated mutation in cardiac event occurrence. Circulation 99, 1464±1470. Bezzina, C., Veldkamp, M.W., van Den Berg, M.P., Postma, A.V., Rook, M.B., Viersma, J.W., van Langen, I.M., Tan-Sindhunata, G., Bink-Boelkens, M.T., van Der Hout, A.H., Mannens, M.M., Wilde, A.A., 1999. A single Na + channel mutation causing both long-QT and Brugada syndromes. Circ. Res. 85, 1206±1213. Bjerregaard, P., Gussak, I., Kotar, S.L., Gessler, J.E., Janosik, D., 1994. Recurrent syncope in a patient with prominent J wave. Am. Heart J. 127, 1426±1430. Boutjdir, M., Restivo, M., Wei, Y., Stergiopoulos, K., El-Sherif, N., 1994. Early afterdepolarization formation in cardiac myocytes: analysis of phase plane patterns, action potential, and membrane currents. J. Cardiovasc. Electrophysiol. 5, 609±620. Brugada, P., Brugada, J., 1992. Right bundle branch block, persistent ST segment elevation and sudden cardiac death: a distinct clinical and electrocardiographic syndrome. A multicenter report. J. Am. Coll. Cardiol. 20, 1391±1396. Bruggemann, A., Pardo, L.A., Stuhmer, W., Pongs, O., 1993. Ether-a-go-go encodes a voltage-gated channel permeable to K 1 and Ca 21 and modulated by cAMP. Nature 365, 445±448. Cannon, S.C., Corey, D.P., 1993. Loss of Na 1 channel inactivation by anemone toxin (ATX II) mimics the myotonic state in hyperkalaemic periodic paralysis. J. Physiol. (Lond.) 466, 501±520. Carmeliet, E., 1987. Slow inactivation of sodium current in rabbit cardiac Purkinje ®bres. P¯uÈgers Arch. 408, 18±26. Carmeliet, E., Saikawa, T., 1982. Shortening of the action potential and reduction of pacemaker activity by lidocaine, quinidine, and procainamide in sheep cardiac Purkinje ®bers. An effect on Na 1 or K 1 currents? Circ. Res. 50, 257±272. Chahine, M., George Jr., A.L., Zhou, M., Ji, S., Sun, W., Barchi, R.L., Horn, R., 1994. Sodium channel mutations in paramyotonia congenita uncouple inactivation from activation. Neuron 12, 281±294.

144

G.E. Kirsch

Charpentier, F., Merot, J., Riochet, D., Le Marec, H., Escande, D., 1998. Adult KCNE1-knockout mice exhibit a mild cardiac cellular phenotype. Biochem. Biophys. Res. Commun. 251, 806±810. Chauhan, V.S., Tuvia, S., Buhusi, M., Bennett, V., Grant, A.O., 2000. Abnormal cardiac Na + channel properties and QT heart rate adaptation in neonatal ankyrin B knockout mice. Circ. Res. 86, 441±447. Chen, J., Zou, A., Splawski, I., Keating, M.T., Sanguinetti, M.C., 1999. Long QT syndrome-associated mutations in the Per-Arnt-Sim (PAS) domain of HERG potassium channels accelerate channel deactivation. J. Biol. Chem. 274, 10113±10118. Chen, Q., Kirsch, G.E., Zhang, D., Brugada, R., Brugada, J., Brugada, P., Potenza, D., Moya, A., Borggrefe, M., Breithardt, G., Ortiz-Lopez, R., Wang, Z., Antzelevitch, C., O'Brien, R.E., SchulzeBahr, E., Keating, M.T., Towbin, J.A., Wang, Q., 1998. Genetic basis and molecular mechanism for idiopathic ventricular ®brillation. Nature 392, 293±296. Chen, Q., Zhang, D., Gingell, R.L., Moss, A.J., Napolitano, C., Priori, S.G., Schwartz, P.J., Kehoe, E., Robinson, J.L., Schulze-Bahr, E., Wang, Q., Towbin, J.A., 1999. Homozygous deletion in KVLQT1 associated with Jervell and Lange±Nielsen syndrome. Circulation 99, 1344±1347. Chevillard, C., Attali, B., Lesage, F., Fontes, M., Barhanin, J., Lazdunski, M., Mattei, M.G., 1993. Localization of a potassium channel gene (KCNE1) to 21q22.1-q22.2 by in situ hybridization and somatic cell hybridization. Genomics 15, 243±245. Chouabe, C., Neyroud, N., Guicheney, P., Lazdunski, M., Romey, G., Barhanin, J., 1997. Properties of KvLQT1 K 1 channel mutations in Romano±Ward and Jervell and Lange±Nielsen inherited cardiac arrhythmias. EMBO J. 16, 5472±5479. Chouabe, C., Neyroud, N., Richard, P., Denjoy, I., Hainque, B., Romey, G., Drici, M.D., Guicheney, P., Barhanin, J., 2000. Novel mutations in KvLQT1 that affect IKs activation through interactions with IsK. Cardiovasc. Res. 45, 971±980. Compton, S.J., Lux, R.L., Ramsey, M.R., Strelich, K.R., Sanguinetti, M.C., Green, L.S., Keating, M.T., Mason, J.W., 1996. Genetically de®ned therapy of inherited long-QT syndrome. Correction of abnormal repolarization by potassium. Circulation 94, 1018±1022. Coraboeuf, E., Deroubaix, E., Coulombe, A., 1979. Effect of tetrodotoxin on action potentials of the conducting system in the dog heart. Am. J. Physiol. 236, H561±H567. Curran, M., Atkinson, D., Timothy, K., Vincent, G.M., Moss, A.J., Leppert, M., Keating, M., 1993. Locus heterogeneity of autosomal dominant long QT syndrome. J. Clin. Invest. 92, 799±803. Curran, M.E., Splawski, I., Timothy, K.W., Green, E.D., Vincent, G.M., Keating, M.T., 1995. A molecular basis for cardiac arrhythmia: HERG mutations cause long QT syndrome. Cell 80, 795± 803. Dausse, E., Berthet, M., Denjoy, I., Andre-Fouet, X., Cruaud, C., Bennaceur, M., Faure, S., Coumel, P., Schwartz, K., Guicheney, P., 1996. A mutation in HERG associated with notched T waves in long QT syndrome. J. Mol. Cell. Cardiol. 28, 1609±1615. De Biasi, M., Hartmann, H.A., Drewe, J.A., Taglialatela, M., Brown, A.M., Kirsch, G.E., 1993. Inactivation determined by a single site in K 1 pores. P¯uÈgers Arch. 422, 354±363. De Ferrari, G.M., Nador, F., Beria, G., Sala, S., Lotto, A., Schwartz, P.J., 1994. Effect of calcium channel block on the wall motion abnormality of the idiopathic long QT syndrome. Circulaton 89, 2126±2132. Deschenes, I., Baroudi, G., Berthet, M., Barde, I., Chalvidan, T., Denjoy, I., Guicheney, P., Chahine, M., 2000. Electrophysiological characterization of SCN5A mutations causing long QT (E1784K) and Brugada (R1512W and R1432G) syndromes. Cardiovasc. Res. 46, 55±65. Donger, C., Denjoy, I., Berthet, M., Neyroud, N., Cruaud, C., Bennaceur, M., Chivoret, G., Schwartz, K., Coumel, P., Guicheney, P., 1997. KVLQT1 C-terminal missense mutation causes a forme fruste longQT syndrome. Circulation 96, 2778±2781. Drici, M.D., Burklow, T.R., Haridasse, V., Glazer, R.I., Woosley, R.L., 1996. Sex hormones prolong the

Cardiac conduction defects

145

QT interval and downregulate potassium channel expression in the rabbit heart. Circulation 94, 1471± 1474. Drici, M.D., Arrighi, I., Chouabe, C., Mann, J.R., Lazdunski, M., Romey, G., Barhanin, J., 1998. Involvement of IsK-associated K 1 channel in heart rate control of repolarization in a murine engineered model of Jervell and Lange±Nielsen syndrome. Circ. Res. 83, 95±102. Duggal, P., Vesely, M.R., Wattanasirichaigoon, D., Villafane, J., Kaushik, V., Beggs, A.H., 1998. Mutation of the gene for IsK associated with both Jervell and Lange±Nielsen and Romano±Ward forms of Long-QT syndrome. Circulation 97, 142±146. Dumaine, R., Kirsch, G.E., 1998. Mechanism of lidocaine block of late current in long Q-T mutant Na 1 channels. Am. J. Physiol. 274, H477±H487. Dumaine, R., Wang, Q., Keating, M.T., Hartmann, H.A., Schwartz, P.J., Brown, A.M., Kirsch, G.E., 1996. Multiple mechanisms of Na 1 channel-linked long-QT syndrome. Circ. Res. 78, 916±924. Franqueza, L., Lin, M., Shen, J., Splawski, I., Keating, M.T., Sanguinetti, M.C., 1999. Long QT syndrome-associated mutations in the S4±S5 linker of KvLQT1 potassium channels modify gating and interaction with minK subunits. J. Biol. Chem. 274, 21063±21070. Gellens, M.E., George Jr., A.L., Chen, L.Q., Chahine, M., Horn, R., Barchi, R.L., Kallen, R.G., 1992. Primary structure and functional expression of the human cardiac tetrodotoxin-insensitive voltagedependent sodium channel. Proc. Natl. Acad. Sci. USA 89, 554±558. Gintant, G.A., Datyner, N.B., Cohen, I.S., 1984. Slow inactivation of a tetrodotoxin-sensitive current in canine cardiac Purkinje ®bers. Biophys. J. 45, 509±512. Hanck, D.A., Sheets, M.F., 1995. Modi®cation of inactivation in cardiac sodium channels: ionic current studies with Anthopleurin-A toxin. J. Gen. Physiol. 106, 601±616. Hartmann, H.A., Colom, L.V., Sutherland, M.L., Noebels, J.L., 1999. Selective localization of cardiac SCN5A sodium channels in limbic regions of rat brain. Nat. Neurosci. 2, 593±595. Heginbotham, L., Lu, Z., Abramson, T., MacKinnon, R., 1994. Mutations in the K 1 channel signature sequence. Biophys. J. 66, 1061±1067. Herzberg, I.M., Trudeau, M.C., Robertson, G.A., 1998. Transfer of rapid inactivation and sensitivity to the class III antiarrhythmic drug E-4031 from HERG to M-eag channels. J. Physiol. (Lond.) 511, 3±14. Hiraoka, M., Sunami, A., Fan, Z., Sawanobori, T., 1992. Multiple ionic mechanisms of early afterdepolarizations in isolated ventricular myocytes from guinea-pig hearts. Ann. N. Y. Acad. Sci. 644, 33±47. Hoshi, T., Zagotta, W.N., Aldrich, R.W., 1990. Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science 250, 533±538. Imanaga, I., 1967. Effects of some ions and drugs on the aconitine-induced ®brillation of the Purkinje ®bers. Jpn. Circ. J. 31, 1819±1831. Jackman, W.M., Clark, M., Friday, K.J., Aliot, E.M., Anderson, J., Lazzara, R., 1984. Ventricular tachyarrhythmias in the long QT syndromes. Med. Clin. North Am. 68, 1079±1109. Jackman, W.M., Friday, K.J., Anderson, J.L., Aliot, E.M., Clark, M., Lazzara, R., 1988. The long QT syndromes: a critical review, new clinical observations and a unifying hypothesis. Prog. Cardiovasc. Dis. 31, 115±172. January, C.T., Riddle, J.M., 1988. Early afterdepolarizations: mechanism of induction and block, a role for L-type Ca 21 current. Circ. Res. 64, 977±990. January, C.T., Riddle, J.M., Salata, J.J., 1988. A model for early afterdepolarizations: induction with the Ca 21 channel agonist Bay K 8644. Circ. Res. 62, 563±571. Kaab, S., Dixon, J., Duc, J., Ashen, D., Nabauer, M., Beuckelmann, D.J., Steinbeck, G., McKinnon, D., Tomaselli, G.F., 1998. Molecular basis of transient outward potassium current downregulation in human heart failure: a decrease in Kv4.3 mRNA correlates with a reduction in current density. Circulation 98, 1383±1393. Kambouris, N.G., Nuss, H.B., Johns, D.C., Tomaselli, G.F., Marban, E., Balser, J.R., 1998. Phenotypic

146

G.E. Kirsch

characterization of a novel long-QT syndrome mutation (R1623Q) in the cardiac sodium channel. Circulation 97, 640±644. Kanters, J.K., Larsen, L.A., Orholm, M., Agner, E., Andersen, P.S., Vuust, J., Christiansen, M., 1998. Novel donor splice site mutation in the KVLQT1 gene is associated with long QT syndrome. J. Cardiovasc. Electrophysiol. 9, 620±624. Keating, M., Atkinson, D., Dunn, C., Timothy, K., Vincent, G.M., Leppert, M., 1991. Linkage of a cardiac arrhythmia, the long QT syndrome, and the Harvey ras-1 gene. Science 252, 704±706. Keating, M., Dunn, C., Atkinson, D., Timothy, K., Vincent, G.M., Leppert, M., 1991. Consistent linkage of the long-QT syndrome to the Harvey ras-1 locus on chromosome 11. Am. J. Hum. Genet. 49, 1335± 1339. Kiyosue, T., Arita, M., 1989. Late sodium current and its contribution to action potential con®guration in guinea pig ventricular myocytes. Circ. Res. 64, 389±397. Kodama, I., Kamiya, K., Toyama, J., 1997. Cellular electropharmacology of amiodarone. Cardiovasc. Res. 35, 13±29. Koh, K.K., Rim, M.S., Yoon, J., Kim, S.S., 1994. Torsades de pointes induced by terfenadine in a patient with long QT syndrome. J. Electrocardiol. 27, 343±346. Kupershmidt, S., Snyders, D.J., Raes, A., Roden, D.M., 1998. A K 1 channel splice variant common in human heart lacks a C-terminal domain required for expression of rapidly activating delayed recti®er current. J. Biol. Chem. 273, 22735±27231. Le Grand, B.L., Hatem, S., Deroubaix, E., Couetil, J.P., Coraboeuf, E., 1994. Depressed transient outward and calcium currents in dilated human atria. Cardiovasc. Res. 28, 548±556. Lees-Miller, J.P., Kondo, C., Wang, L., Duff, H.J., 1997. Electrophysiological characterization of an alternatively processed ERG K 1 channel in mouse and human hearts. Circ. Res. 81, 719±726. Lehmann, M.H., Hardy, S., Archibald, D., Quart, B., MacNeil, D.J., 1996. Sex difference in risk of torsade de pointes with d,l-sotalol. Circulation 94, 2535±2541. Lehmann, M.H., Timothy, K.W., Frankovich, D., Fromm, B.S., Keating, M., Locati, E.H., Taggart, R.T., Towbin, J.A., Moss, A.J., Schwartz, P.J., Vincent, G.M., 1997. Age-gender in¯uence on the ratecorrected QT interval and the QT-heart rate relation in families with genotypically characterized long QT syndrome. J. Am. Coll. Cardiol. 29, 93±99. Li, H., Chen, Q., Moss, A.J., Robinson, J., Goytia, V., Perry, J.C., Vincent, G.M., Priori, S.G., Lehmann, M.H., Den®eld, S.W., Duff, D., Kaine, S., Shimizu, W., Schwartz, P.J., Wang, Q., Towbin, J.A., 1998. New mutations in the KVLQT1 potassium channel that cause long-QT syndrome. Circulation 97, 1264±1269. Li, X., Xu, J., Li, M., 1997. The human delta1261 mutation of the HERG potassium channel results in a truncated protein that contains a subunit interaction domain and decreases the channel expression. J. Biol. Chem. 272, 705±708. Liu, X.K., Katchman, A., Drici, M.D., Ebert, S.N., Ducic, I., Morad, M., Woosley, R.L., 1998. Gender difference in the cycle length-dependent QT and potassium currents in rabbits. J. Pharmacol. Exp. Ther. 285, 672±679. Locati, E.H., Zareba, W., Moss, A.J., Schwartz, P.J., Vincent, G.M., Lehmann, M.H., Towbin, J.A., Priori, S.G., Napolitano, C., Robinson, J.L., Andrews, M., Timothy, K., Hall, W.J., 1998. Age- and sexrelated differences in clinical manifestations in patients with congenital long-QT syndrome: ®ndings from the International LQTS Registry. Circulation 97, 2237±2244. London, B., Trudeau, M.C., Newton, K.P., Beyer, A.K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Satler, C.A., Robertson, G.A., 1997. Two isoforms of the mouse ether-a-go-go-related gene coassemble to form channels with properties similar to the rapidly activating component of the cardiac delayed recti®er K 1 current. Circ. Res. 81, 870±878. London, B., Wang, D.W., Hill, J.A., Bennett, P.B., 1998. The transient outward current in mice lacking the potassium channel gene Kv1.4. J. Physiol. (Lond.) 509, 171±182.

Cardiac conduction defects

147

Makita, N., Bennett Jr., P.B., George Jr., A.L., 1994. Voltage-gated Na 1 channel b1 subunit mRNA expressed in adult human skeletal muscle, heart, and brain is encoded by a single gene. J. Biol. Chem. 269, 7571±7578. Makita, N., Bennett, P.B., George, J.r., A, L., 1996. Molecular determinants of b1 subunit-induced gating modulation in voltage-dependent Na 1 channels. J. Neurosci. 16, 7117±7127. Makita, N., Shirai, N., Nagashima, M., Matsuoka, R., Yamada, Y., Tohse, N., Kitabatake, A., 1998. A de novo missense mutation of human cardiac Na 1 channel exhibiting novel molecular mechanisms of long QT syndrome. FEBS Lett 423, 5±9. McCormick, K.A., Isom, L.L., Ragsdale, D., Smith, D., Scheuer, T., Catterall, W.A., 1998. Molecular determinants of Na 1 channel function in the extracellular domain of the b1 subunit. J. Biol. Chem. 273, 3954±3962. McDonald, T.V., Yu, Z., Ming, Z., Palma, E., Meyers, M.B., Wang, K.W., Goldstein, S.A., Fishman, G.I., 1997. A minK-HERG complex regulates the cardiac potassium current I(Kr). Nature 388, 289±292. Merri, M., Benhorin, J., Alberti, M., Locati, E., Moss, A.J., 1989. Electrocardiographic quantitation of ventricular repolarization. Circulation 80, 1301±1308. Monahan, B.P., Ferguson, C.L., Killeavy, E.S., Lloyd, B.K., Troy, J., Cantilena Jr., L.R., 1990. Torsades de pointes occurring in association with terfenadine use. JAMA 264, 2788±2790. Moss, A.J., 1998. Management of patients with the hereditary long QT syndrome. J. Cardiovasc. Electrophysiol. 9, 668±674. Nakajima, T., Furukawa, T., Tanaka, T., Katayama, Y., Nagai, R., Nakamura, Y., Hiraoka, M., 1998. Novel mechanism of HERG current suppression in LQT2: shift in voltage dependence of HERG inactivation. Circ. Res. 83, 415±422. Neyroud, N., Tesson, F., Denjoy, I., Leibovici, M., Donger, C., Barhanin, J., Faure, S., Gary, F., Coumel, P., Petit, C., Schwartz, K., Guicheney, P., 1997. A novel mutation in the potassium channel gene KVLQT1 causes the Jervell and Lange±Nielsen cardioauditory syndrome. Nat. Genet. 15, 186±189. Ouadid, H., Albat, B., Nargeot, J.,1995. Calcium currents in diseased human cardiac cells. J. Cardiovasc. Pharmacol. 25, 282±291. Patton, D.E., Isom, L.L., Catterall, W.A., Goldin, A.L., 1994. The adult rat brain b1 subunit modi®es activation and inactivation gating of multiple sodium channel a subunits. J. Biol. Chem. 269, 11755± 17649. Priori, S.G., Napolitano, C., Cantu, F., Brown, A.M., Schwartz, P.J., 1996. Differential response to Na 1 channel blockade, b-adrenergic stimulation, and rapid pacing in a cellular model mimicking the SCN5A and HERG defects present in the long-QT syndrome. Circ. Res. 78, 1009±1015. Priori, S.G., Napolitano, C., Paganini, V., Cantu, F., Schwartz, P.J., 1997. Molecular biology of the long QT syndrome: impact on management. Pacing Clin. Electrophysiol. 20, 2052±2057. Priori, S.G., Schwartz, P.J., Napolitano, C., Bianchi, L., Dennis, A., De Fusco, M., Brown, A.M., Casari, G., 1998. A recessive variant of the Romano±Ward long-QT syndrome? Circulation 97, 2420±2425. Pusch, M., 1998. Increase of the single-channel conductance of KvLQT1 potassium channels induced by the association with minK. P¯uÈgers Arch. 437, 172±174. Pusch, M., Magrassi, R., Wollnik, B., Conti, F., 1998. Activation and inactivation of homomeric KvLQT1 potassium channels. Biophys J. 75, 785±792. Roden, D.M., 1990. Clinical features of arrhythmia aggravation by antiarrhythmic drugs and their implications for basic mechanisms. Drug Dev. Res. 19, 153±172. Roden, D.M., 1993. Early after-depolarizations and torsade de pointes: implications for the control of cardiac arrhythmias by prolonging repolarization. Eur. Heart J. 14, 56±61. Roden, D.M., Hoffman, B.F., 1985. Action potential prolongation and induction of abnormal automaticity by low quinidine concentrations in canine Purkinje ®bers. Relationship to potassium and cycle length. Circ. Res. 56, 857±867.

148

G.E. Kirsch

Roden, D.M., Thompson, K.A., Hoffman, B.F., Woosley, R.L., 1986. Clinical features and basic mechanisms of quinidine-induced arrhythmias. J. Am. Coll. Cardiol. 8, 73A±78A. Rogart, R.B., Cribbs, L.L., Muglia, L.K., Kephart, D.D., Kaiser, M.W., 1989. Molecular cloning of a putative tetrodotoxin-resistant rat heart Na 1 channel isoform. Proc. Natl. Acad. Sci. USA 86, 8170± 8174. Romey, G., Attali, B., Chouabe, C., Abitbol, I., Guillemare, E., Barhanin, J., Lazdunski, M., 1997. Molecular mechanism and functional signi®cance of the minK control of the KvLQT1 channel activity. J. Biol. Chem. 272, 16713±16716. Roy, M., Dumaine, R., Brown, A.M., 1996. HERG, a primary human ventricular target of the nonsedating antihistamine terfenadine. Circulation 94, 817±823. Russell, M.W., Dick II, M., Collins, F.S., Brody, L.C., 1996. KVLQT1 mutations in three families with familial or sporadic long QT syndrome. Hum. Mol. Genet. 5, 1319±1324. Saarinen, K., Swan, H., Kainulainen, K., Toivonen, L., Viitasalo, M., Kontula, K., 1998. Molecular genetics of the long QT syndrome: two novel mutations of the KVLQT1 gene and phenotypic expression of the mutant gene in a large kindred. Hum. Mutat. 11, 158±165. Salata, J.J., Wasserstrom, J.A., 1988. Effects of quinidine on action potentials and ionic currents in isolated canine ventricular myocytes. Circ. Res. 62, 324±337. Sanguinetti, M.C., Jurkiewicz, N.K., 1990. Two components of cardiac delayed recti®er K 1 current. Differential sensitivity to block by class III antiarrhythmic agents. J. Gen. Physiol. 96, 195±215. Sanguinetti, M.C., Jurkiewicz, N.K., 1992. Role of external Ca 21 and K 1 in gating of cardiac delayed recti®er K 1 currents. P¯uÈgers Arch. 420, 180±186. Sanguinetti, M.C., Jiang, C., Curran, M.E., Keating, M.T., 1995. A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the IKr potassium channel. Cell 81, 299± 307. Sanguinetti, M.C., Curran, M.E., Spector, P.S., Keating, M.T., 1996a. Spectrum of HERG K 1-channel dysfunction in an inherited cardiac arrhythmia. Proc. Natl. Acad. Sci. USA 93, 2208±2212. Sanguinetti, M.C., Curran, M.E., Zou, A., Shen, J., Spector, P.S., Atkinson, D.L., Keating, M.T., 1996b. Coassembly of KvLQT1 and minK (IsK) proteins to form cardiac IKs potassium channel. Nature 384, Satler, C.A., Vesely, M.R., Duggal, P., Ginsburg, G.S., Beggs, A.H., 1998. Multiple different missense mutations in the pore region of HERG in patients with long QT syndrome. Hum. Genet. 102, 265±272. Satler, C.A., Vesely, M.R., Duggal, P., Ginsburg, G.S., Beggs, A.H., 1998. Multiple different missense mutations in the pore region of HERG in patients with long QT syndrome. Hum. Genet. 102, 265±272. Schonherr, R., Heinemann, S.H., 1996. Molecular determinants for activation and inactivation of HERG, a human inward recti®er potassium channel. J. Physiol. (Lond.) 493, 635±642. Schulze-Bahr, E., Haverkamp, W., Wiebusch, H., Schulte, H., Hordt, M., Borggrefe, M., Breithardt, G., Assmann, G., Funke, H., 1995. Molecular analysis at the Harvey Ras-1 gene in patients with long QT syndrome. Mol. Med. 73, 565±569. Schulze-Bahr, E., Wang, Q., Wedekind, H., Haverkamp, W., Chen, Q., Sun, Y., 1997. KCNE1 mutations cause jervell and Lange±Nielsen syndrome. Nat. Genet. 17, 267±268. Schwartz, P.J., Stone, H.L., 1979. Effects of unilateral stellectomy upon cardiac performance during exercise in dogs. Circ. Res. 44, 637±645. Schwartz, P.J., Periti, M., Malliani, A., 1975. The long Q-T syndrome. Am. Heart J. 89, 378±390. Schwartz, P.J., Locati, E.H., Moss, A.J., Crampton, R.S., Trazzi, R., Ruberti, U., 1991. Left cardiac sympathetic denervation in the therapy of congenital long QT syndrome. A worldwide report. Circulation 84, 503±511. Schwartz, P.J., Locati, E.H., Napolitano, C., Priori, S.G., 1995a. The long QT syndrome. In: Zipes, D.P., Jalife, J. (Eds.). Cardiac Electrophysiology. From Cell to Bedside, W.B. Saunders, Philadelphia, PA, pp. 788±811. Schwartz, P.J., Priori, S.G., Locati, E.H., Napolitano, C., Cantu, F., Towbin, J.A., Keating, M.T.,

Cardiac conduction defects

149

Hammoude, H., Brown, A.M., Chen, L.S., 1995b. Long QT syndrome patients with mutations of the SCN5A and HERG genes have differential responses to Na 1 channel blockade and to increases in heart rate. Implications for gene-speci®c therapy. Circulation 92, 3381±3386. Sesti, F., Goldstein, S.A.N., 1998. Single-channel characteristics of wild-type IKs channels and channels formed with two minK mutants that cause long QT syndrome. J. Gen. Physiol. 112, 651±663. Shalaby, F.Y., Levesque, P.C., Yang, W.P., Little, W.A., Conder, M.L., Jenkins-West, T., Blanar, M.A., 1997. Dominant-negative KvLQT1 mutations underlie the LQT1 form of long QT syndrome. Circulation 96, 1733±1736. Shi, W., Wymore, R.S., Wang, H.S., Pan, Z., Cohen, I.S., McKinnon, D., Dixon, J.E., 1997. Identi®cation of two nervous system-speci®c members of the erg potassium channel gene family. J. Neurosci. 17, 9423±9432. Shimizu, W., Antzelevitch, C., 1997. Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade de pointes in LQT2 and LQT3 models of the long-QT syndrome. Circulation 96, 2038±2047. Shimizu, T., Iwamura, N., Toyama, J., Yamada, K., Shibata, S., 1979. Effect of cardiotonic polypeptide anthopleurin-A on canine Purkinje and ventricular muscle ®bers. Eur. J. Pharmacol. 56, 7±13. Sicouri, S., Antzelevitch, D., Heilmann, C., Antzelevitch, C., 1997. Effects of sodium channel block with mexiletine to reverse action potential prolongation in in vitro models of the long term QT syndrome. J. Cardiovasc. Electrophysiol. 8, 1280±1290. Smith, P.L., Baukrowitz, T., Yellen, G., 1996. The inward recti®cation mechanism of the HERG cardiac potassium channel. Nature 379, 833±836. Smith, M.R., Smith, R.D., Plummer, N.W., Meisler, M.H., Goldin, A.L., 1998. Functional analysis of the mouse Scn8a sodium channel. J. Neurosci. 18, 6093±6102. Snyders, D.J., Hondeghem, L.M., 1990. Effects of quinidine on the sodium current of guinea pig ventricular myocytes: evidence for a drug-associated rested state with altered kinetics. Circ. Res. 66, 565±579. Spector, P.S., Curran, M.E., Keating, M.T., Sanguinetti, M.C., 1996. Class III antiarrhythmic drugs block HERG, a human cardiac delayed recti®er K 1 channel. Open-channel block by methanesulfonanilides. Circ. Res. 78, 499±503. Spector, P.S., Curran, M.E., Zou, A., Keating, M.T., Sanguinetti, M.C., 1996. Fast inactivation causes recti®cation of the IKr channel. J. Gen. Physiol. 107, 611±619. Splawski, I., Tristani-Firouzi, M., Lehmann, M.H., Sanguinetti, M.C., Keating, M.T., 1997. Mutations in the hminK gene cause long QT syndrome and suppress IKs function. Nat. Genet. 17, 338±340. Splawski, I., Shen, J., Timothy, K.W., Vincent, G.M., Lehmann, M.H., Keating, M.T., 1998. Genomic structure of three long QT syndrome genes: KVLQT1. HERG, and KCNE1. Genomics 51, 86±97. Strichartz, G.R., Wang, G.K., 1986. Rapid voltage-dependent dissociation of scorpion a-toxins coupled to Na 1 channel inactivation in amphibian myelinated nerves. J. Gen. Physiol. 88, 413±435. StuÈhmer, W., Conti, F., Suzuki, H., Wang, X.D., Noda, M., Yahagi, N., Kubo, H., Numa, S., 1989. Structural parts involved in activation and inactivation of the sodium channel. Nature 339, 597±603. Suessbrich, H., Waldegger, S., Lang, F., Busch, A.E., 1996. Blockade of HERG channels expressed in Xenopus oocytes by the histamine receptor antagonists terfenadine and astemizole. FEBS Lett. 385, 77±80. Tabb, J.S., Fanger, G.R., Wilson, E.M., Maue, R.A., Henderson, L.P., 1994. Suppression of sodium channel function in differentiating C2 muscle cells stably overexpressing rat androgen receptors. J. Neurosci. 14, 763±773. Takumi, T., Ohkubo, H., Nakanishi, S., 1988. Cloning of a membrane protein that induces a slow voltagegated potassium current. Science 242, 1042±1045. Takumi, T., Moriyoshi, K., Aramori, I., Ishii, T., Oiki, S., Okada, Y., Ohkubo, H., Nakanishi, S., 1991. Alteration of channel activities and gating by mutations of slow ISK potassium channel. J. Biol. Chem. 266, 22192±22198.

150

G.E. Kirsch

Tai, K.K., Goldstein, S.A., 1998. The conduction pore of a cardiac potassium channel. Nature 391, 605± 608. Tan, H.L., Hou, C.J.Y., Lauer, M.R., Sung, R.J., 1995. Electrophysiologic mechanisms of the long QT interval syndromes and Torsade de Pointes. Arch. Int. Med. 122, 701±714. Tanaka, T., Nagai, R., Tomoike, H., Takata, S., Yano, K., Yabuta, K., Haneda, N., Nakano, O., Shibata, A., Sawayama, T., Kasai, H., Yazaki, Y., Nakamura, Y., 1997. Four novel KVLQT1 and four novel HERG mutations in familial long-QT syndrome. Circulation 95, 565±567. Towbin, J.A., Li, H., Taggart, R.T., Lehmann, M.H., Schwartz, P.J., Satler, C.A., Ayyagari, R., Robinson, J.L., Moss, A., Hejtmancik, J.F., 1994. Evidence of genetic heterogeneity in Romano±Ward long QT syndrome. Analysis of 23 families. Circulation 90, 2635±2644. Tristani-Firouzi, M., Sanguinetti, M.C., 1998. Voltage-dependent inactivation of the human K 1 channel KvLQT1 is eliminated by association with minimal K 1 channel (minK) subunits. J. Physiol. (Lond. 510, 37±45. Trudeau, M.C., Warmke, J.W., Ganetzky, B., Robertson, G.A., 1995. HERG, a human inward recti®er in the voltage-gated potassium channel family. Science 269, 92±95. Tyson, J., Tranebjaerg, L., Bellman, S., Wren, C., Taylor, J.F., Bathen, J., Aslaksen, B., Sorland, S.J., Lund, O., Malcolm, S., Pembrey, M., Bhattacharya, S., Bitner-Glindzicz, M., 1997. IsK and KvLQT1: mutation in either of the two subunits of the slow component of the delayed recti®er potassium channel can cause Jervell and Lange±Nielsen syndrome. Hum. Mol. Genet. 6, 2179±2185. Vetter, D.E., Mann, J.R., Wangemann, P., Liu, J., McLaughlin, K.J., Lesage, F., Marcus, D.C., Lazdunski, M., Heinemann, S.F., Barhanin, J., 1996. Inner ear defects induced by null mutation of the isk gene. Neuron 17, 1251±1264. Viskin, S., Belhassen, B., 1998. Polymorphic ventricular tachyarrhythmias in the absence of organic heart disease: classi®cation, differential diagnosis, and implications for therapy. Prog. Cardiovasc. Dis. 41, 17±34. Viskin, S., Alla, S.R., Barron, H.V., Heller, K., Saxon, L., Kitzis, I., Hare, G.F., Wong, M.J., Lesh, M.D., Scheinman, M.M., 1996. Mode of onset of torsade de pointes in congenital long QT syndrome. J. Am. Coll. Cardiol. 28, 1262±1268 . Walsh, K.B., Kass, R.S., 1988. Regulation of a heart potassium channel by protein kinase A and C. Science 242, 67±69. Wang, Q., Shen, J., Li, Z., Timothy, K., Vincent, G.M., Priori, S.G., Schwartz, P.J., Keating, M.T., 1995. Cardiac sodium channel mutations in patients with long QT syndrome, an inherited cardiac arrhythmia. Hum. Mol. Genet. 4, 1603±1607. Wang, Q., Shen, J., Splawski, I., Atkinson, D., Li, Z., Robinson, J.L., Moss, A.J., Towbin, J.A., Keating, M.T., 1995. SCN5A mutations associated with an inherited cardiac arrhythmia, long QT syndrome. Cell 80, 805±811. Wang, D.W., Yazawa, K., George Jr., A.L., Bennett, P.B., 1996. Characterization of human cardiac Na 1 channel mutations in the congenital long QT syndrome. Proc. Natl. Acad. Sci. USA 93, 13200±13205. Wang, K.W., Tai, K.K., Goldstein, S.A., 1996. MinK residues line a potassium channel pore. Neuron 16, 571±577. Wang, Q., Curran, M.E., Splawski, I., Burn, T.C., Millholland, J.M., VanRaay, T.J., Shen, J., Timothy, K.W., Vincent, G.M., de Jager, T., Schwartz, P.J., Toubin, J.A., Moss, A.J., Atkinson, D.L., Landes, G.M., Connors, T.D., Keating, M.T., 1996. Positional cloning of a novel potassium channel gene: KvLQT1 mutations cause cardiac arrhythmias. Nat. Genet. 12, 17±23. Warmke, J.W., Ganetzky, B., 1994. A family of potassium channel genes related to eag in Drosophila and mammals. Proc. Natl. Acad. Sci. USA 91, 3438±3442. Wasserstrom, J.A., Kelly, J.E., Liberty, K.N., 1993. Modi®cation of cardiac Na 1 channels by anthopleurin-A: effects on gating and kinetics. P¯uÈgers Arch. 424, 15±24. Wattanasirichaigoon, D., Vesely, M.R., Duggal, P., Levine, J.C., Blume, E.D., Wolff, G.S., Edwards,

Cardiac conduction defects

151

S.B., Beggs, A.H., 1999. Sodium channel abnormalities are infrequent in patients with long QT syndrome: identi®cation of two novel SCN5A mutations. Am. J. Med. Genet. 86, 470-476. Wei, J., Wang, D.W., Alings, M., Fish, F., Wathen, M., Roden, D.M., George Jr., A.L., 1999. Congenital long-QT syndrome caused by a novel mutation in a conserved acidic domain of the cardiac Na + channel. Circulation 99, 3165±3171. Wit, A.L., Rosen, M.R., 1991. Afterdepolarizations and triggered activity: distinction from automaticity as an arrhythmogenic mechanism. In: Fozzard, H.A., Haber, E., Jennings, R.B., Katz, A.M., Morgan, H.E. (Eds.). The Heart and Cardiovascular System, Raven Press, New York, pp. 2113±2163. Wollnik, B., Schroeder, B.C., Kubisch, C., Esperer, H.D., Wieacker, P., Jentsch, T.J., 1997. Pathophysiological mechanisms of dominant and recessive KVLQT1 K + channel mutations found in inherited cardiac arrhythmias. Hum. Mol. Genet. 6, 1943±1949. Yamagishi, H., Furutani, M., Kamisago, M., Morikawa, Y., Kojima, Y., Hino, Y., Furutani, Y., Kimura, M., Imamura, S., Takao, A., Momma, K., Matsuoka, R., 1997. A de novo missense mutation (R1623Q) of the SCN5A gene in a Japanese girl with sporadic long QT Syndrome. Hum. Mutat., 140. Yan, G.X., Antzelevitch, C., 1996. Cellular basis for the electrocardiographic J wave. Circulation 93, 372±379. Yang, T., Wathen, M.S., Felipe, A., Tamkun, M.M., Snyders, D.J., Roden, D.M., 1994. K 1 currents and K 1 channel mRNA in cultured atrial cardiac myocytes (AT-1 cells). Circ. Res. 75, 870±878. Yang, T., Kupershmidt, S., Roden, D.M., 1995. Anti-minK antisense decreases the amplitude of the rapidly activating cardiac delayed recti®er K 1 current. Circ. Res. 77, 1246±1253. Yang, T., Roden, D.M., 1996. Extracellular potassium modulation of drug block of I Kr. Implications for torsade de pointes and reverse use-dependence. Circulation 93, 407±411. Yang, W.P., Levesque, P.C., Little, W.A., Conder, M.L., Shalaby, F.Y., Blanar, M.A., 1997. KvLQT1, a voltage-gated potassium channel responsible for human cardiac arrhythmias. Proc. Natl. Acad. Sci. USA 94, 4017±4021. Yanowitz, F., Preston, J.B., Abildskov, J.A., 1966. Functional distribution of right and left stellate innervation to the ventricles. Production of neurogenic electrocardiographic changes by unilateral alteration of sympathetic tone. Circ. Res. 18, 416±428. Yatani, A., Okabe, K., Polakis, P., Halenbeck, R., McCormick, F., Brown, A.M., 1990. ras p21 and GAP inhibit coupling of muscarinic receptors to atrial K 1 channels. Cell 61, 769±776. Zareba, W., Moss, A.J., Schwartz, P.J., Vincent, G.M., Robinson, J.L., Priori, S.G., Benhorin, J., Locati, E.H., Towbin, J.A., Keating, M.T., Lehmann, M.H., Hall, W.J., 1998. In¯uence of genotype on the clinical course of the long-QT syndrome. International Long-QT Syndrome Registry Research Group. N. Engl. J. Med. 339, 960±965. Zhou, Z., Gong, Q., Epstein, M.L., January, C.T., 1998. HERG channel dysfunction in human long QT syndrome. Intracellular transport and functional defects. J. Biol. Chem. 273, 21061±21066. Zhou, J.T., Zheng, L.R., Liu, W.Y., 1992. Role of early afterdepolarization in familial long QTU syndrome and torsade de pointes. Pacing Clin. Electrophysiol. 15, 2164±2168.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

155

Chapter 6

Migraine and ataxias Anne Ducros a, b, Christian Denier a, Elisabeth TournierLasserve a, c a

INSERM EPI 9921, Faculte de MeÂdecine Saint-Louis LariboisieÁre, 10 avenue de Verdun, 75010 Paris, France b Service de Neurologie, HoÃpital LariboisieÁre, 2 rue Ambroise PareÂ, 75475 Paris Cedex 10, France c Laboratoire de CytogeÂneÂtique, HoÃpital LariboisieÁre, 2 rue Ambroise PareÂ, 75475 Paris Cedex 10, France

Abstract Three additional hereditary neurological conditions, familial hemiplegic migraine (episodic ataxia type 2 (EA2) and spino cerebellar ataxia type 6 (SCA6) have been shown recently to be caused by mutations within a calcium channel gene, CACNA1A. These clinically distinct conditions are associated with distinct types of mutations, missense mutation in familial hemiplegic migraine (FHM), truncating ones in EA2 and expansion of a CAG triplet in SCA6. The main objective of researchers involved in this ®eld is now to understand the mechanisms leading from these mutations to the observed phenotypes. Various approaches are currently used, including in vitro electrophysiological studies comparing calcium currents in cells expressing wild type and mutant CACNA1A as well as detailed analysis of CACNA1A tottering and leaner mutant mice. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Hemiplegic migraine, episodic ataxia and progressive cerebellar ataxia represent, together with some speci®c epileptic syndromes, the most recent members of the growing list of neurological channelopathies. Recently, distinct types of mutation in the neuronal P/Q type calcium channel alpha 1A subunit gene (CACNA1A) on chromosome 19p13 were shown to cause three human autosomal dominant disorders: familial hemiplegic migraine (HM), episodic ataxia type 2 (EA2), and chronic spinocerebellar ataxia type 6 (SCA6) (Ophoff et al., 1996; Zhuchenko et al., 1997). In the mutant mice tottering and leaner, which are characterized by epilepsy and

156

A. Ducros et al.

ataxia, mutations were identi®ed in the mouse ortholog of CACNA1A (Fletcher et al., 1996). The following chapter presents the various clinical patterns and molecular alterations observed in these calcium channel disorders. Episodic ataxia type 1, another autosomal dominant disorder clinically related to EA2, which is due to mutations in the potassium channel gene KCNA1, is detailed in the note added in proof.

Clinical features Diagnostic criteria Hemiplegic migraine Familial hemiplegic migraine (HM) is an hereditary form of migraine with aura characterized by the presence of a motor weakness during the aura (IHS 1988). HM is the only migraine subtype for which a monogenic, autosomal dominant, mode of inheritance has been clearly established. Besides the familial forms, some sporadic cases have been reported. Typical attacks include a unilateral motor de®cit associated with paresthesias, speech disturbances or visual signs (Blau et al., 1955; Whitty, 1953). The degree of motor de®cit is highly variable, ranging from mild clumsiness to total hemiplegia. Most patients have a moderate hemiparesis. Bilateral sensory-motor symptoms occur in about 25% of patients, one side after the other or both sides simultaneously (Haan et al., 1994, 1995; Terwindt et al., 1996; Ducros et al., 2000). These aura symptoms last ten minutes to a few hours and are followed by a migrainous headache. Those hemiplegic migraine features are not stereotyped. Onset's order, progression, topography, intensity and duration of the various aura symptoms may vary from one attack to another within a given patient, as may vary the various headache features. Moreover, those features may be highly variable among patients from a given family (Ducros et al., 2000). In addition to those usual attacks, severe episodes with prolonged aura (up to several days or weeks), consciousness impairment ranging from confusion to profound coma, agitation, fever and meningismus occur in 40% of the patients (Fitzsimons et al., 1985; MuÈnte et al., 1990; Gardner et al., 1997; Spranger et al. 1999; Ducros et al., 2000). A few patients may have seizures during a severe episode (tonic-clonic generalized seizures or partial motor seizures). Other forms of migraine may occur: about 15% of the patients have migraine with `non hemiplegic' aura alternating with HM attacks and 34% have migraine without aura (Ohta et al., 1967; Young et al., 1970; Fitzsimons et al., 1985; MuÈnte et al., 1990; Haan et al. 1995; Marchioni et al., 1995; Ducros et al., 2000). Triggering factors are reported by about two-thirds of the patients, the most frequent being stress and minor head trauma (Terwindt et al., 1996; Ducros et al.,

Migraine and ataxias

157

2000). In several cases, a severe HM episode was precipitated by injection of contrast enhancement products during cerebral or extracerebral angiography (Blau et al., 1955). Age of onset is usually between 10 and 15 years old, but ranges from 1± 75 years old (Rajput et al., 1995). The frequency of attacks varies from several per week to only a few in the whole life, with an average of 3±4 per year. In some patients, this frequency may range from daily attacks, usually in the ®rst years of evolution, to years-long free intervals (Ducros et al., 2000). In general, the attackfrequency decreases after 20±25 years old. In 20% of unselected families, some HM patients have permanent cerebellar symptoms such as nystagmus and slowly progressive mild to moderate stato-kinetic ataxia. Cerebellar ataxia may be diagnosed prior to the ®rst HM attack and progresses independently of the frequency and/or severity of these attacks. Autonomous gait remains generally possible even after years of evolution (Ohta et al., 1967; Young et al., 1970; Zifkin et al., 1980; Fitzsimons et al., 1985; Joutel et al., 1993; Joutel et al., 1994; Ophoff et al., 1994; Terwindt et al., 1996; Elliott et al., 1996; Ducros et al., 1999a,b). Other associated neurological symptoms have been reported in a few HM families: essential tremor (Zifkin et al. 1980; Ducros et al., 1998), Usher's syndrome and cataract (Young et al., 1970), cognitive impairment (Marchioni et al., 1995) and mental retardation (Zifkin et al., 1980; Fitzsimons et al., 1985). Episodic ataxia type 2 The autosomal dominant episodic ataxias belong to a clinically and genetically heterogeneous group of conditions characterized by recurrent attacks of generalized ataxia usually starting during childhood or adolescence (Gancher et al., 1986). In episodic ataxia type 1 (EA1), short ataxic spells (less than 5±10 min) are typically precipitated by sudden movement or startle, and are increased by exercise, emotional stress, fatigue, fever and hunger (Brandt et al., 1997). Unsteadiness, limb incoordination, dysarthria and generalized myokimia are characteristically present during attacks. Other symptoms may occur: minimal vertigo, nausea, diplopia, dystonic posture or postural tremor of hands and head. Application of cold water to face and hands is used by some patients to shorten attacks. Sleep has been reported to abort some longer attacks. EA1 usually starts in early childhood. The frequency of attacks is highly variable among patients, from several per day to one in several weeks. Frequency of attacks often decrease after 20 years of age. Neurological examination during and between attacks shows continuous muscle movements. No nystagmus is observed. EA1 is caused by mutations in the voltage-gated potassium channel KCNA1, mapped on chromosome 12p, and is thus clinically and genetically different from EA2 (Browne et al., 1994; Adelman et al., 1995; Comu et al., 1996; Sanguinetti et al., 1997; D'Adamo et al., 1998). Episodic ataxia type 2 (EA2), also called acetazolamide-responsive hereditary paroxysmal cerebellar ataxia is also an autosomal dominant disorder (Parker,

158

A. Ducros et al.

1946). Sporadic cases have been reported. In EA2, attacks are more prolonged (15 min to several hours or days), often provoked by emotional or physical stress, alcohol or coffee, but not by startle, and associated with interictal nystagmus (Hill et al., 1968; Griggs et al., 1978; Donat et al., 1979; Feeney et al., 1989; Van Bogaert et al., 1993; Baloh et al., 1997). Acetazolamide responsiveness is a common feature (Griggs et al., 1978; Zasorin et al., 1983). Attacks usually include a major unsteadiness of rapid onset associated with limb incoordination, dysarthria and gaze-evoked nystagmus. Other symptoms are occasionally reported, such as vertigo, nausea, diplopia, headache, confusion, generalized sweating and rarely oscillopsia. Prolonged asthenia may follow the attacks. A brief sleep may often relieve all attack features. Age of onset is usually during childhood or adolescence, but varies from 1 to 30 years. The frequency of attacks is highly variable and tends to decrease with age in some patients. Between attacks, nystagmus on lateral or vertical gaze is the most prominent sign. In addition, some patients develop a mild permanent and progressive cerebellar ataxia, with gait unsteadiness, limb uncoordination and dysarthria (Baloh et al., 1997). Impaired smooth pursuit eye movements are also frequently noted. Gait usually remains autonomous, but in some patients ataxia and dysarthria have an early onset or progress to a severe disability. Isolated severe progressive ataxia without any paroxysmal symptoms has also been reported in some families. Some degree of mental retardation is occasionally reported (Donat et al., 1979; Feeney et al., 1989; von Brederlow et al., 1995; Denier et al., 2000). Spinocerebellar ataxia type 6 On the contrary of HM and EA, spinocerebellar ataxia type 6 (SCA6) is a lateonset progressive neurological condition (Zhuchenko et al., 1997). Autosomal dominant cerebellar ataxias (SCA) are a clinically and genetically heterogeneous group of disorders in which ataxia may be associated to ophtalmoplegia, pyramidal and extrapyramidal signs, dysarthria, amyotrophy, and pigmentary retinopathy. SCA6 is responsible for gait and limb ataxia, with dysarthria, nystagmus, hypotonia, sometimes hypopallesthesia and loss of lower limbs positional sense (Geschwind et al., 1997; Matsumura et al., 1997; Stevanin et al., 1997; Nagai et al., 1998). Progression of ataxia is often so slowly that patients are not aware of the disease onset. They complain from brief spells of unsteadiness precipitated by rapid movements during years before the onset of a permanent unsteadiness, usually in their ®fties. Most patients remain ambulatory at 60 years of age, and become wheel-chair bound after 15±25 years of evolution. Differential diagnosis The diagnosis of HM and EA is entirely dependent upon obtaining a precise description of the transient neurological episodes and a family history of similar

Migraine and ataxias

159

attacks. In HM, the major diagnosis problem concerns the severe attacks with consciousness impairment and prolonged aura. When inaugural, those severe episodes are often diagnosed as being meningoencephalitis. Some children with recurrent severe attacks are considered as having alternating hemiplegia. In HM and EA, other differential diagnosis include complex partial seizures, transient ischemic attacks, metabolic disorders (such as pyruvate decarboxylase de®ciencies, pyruvate deshydrogenase de®ciencies or aminaciduria) and other forms of hereditary paroxysmal neurological disorders such as paroxysmal kinesigenic or dystonic choreoathetosis. Patients with EA2 are unfortunately very often considered as being hysterical or simulating. In SCA6, or in HM and EA2 patients in whom cerebellar ataxia is the major clinical feature, the differential diagnosis include all other hereditary and nonhereditary causes of progressive cerebellar ataxia. Histological ®ndings To our knowledge, no anatomopathological study has been performed in HM and EA2. In SCA6, autopsy dicloses a major cerebellar atrophy, with very mild atrophy of the brain stem (Subramony et al., 1996; Sasaki et al., 1998; Takahashi et al., 1998). In patients deceased from SCA6, microscopic examination showed a severe loss of cerebellar Purkinje cells, predominanting in the superior vermis and the hemispheres, with moderate loss of granular cells and of neurons of dentate nucleus and inferior oliva. In one case, who died after only seven years of evolution of SCA6 from another cause, neuronal loss affected only Purkinje and granular cells (Sasaki et al., 1998). Neurological investigations In HM, cerebrospinal ¯uid is often abnormal during severe attacks with fever and unconsciousness, showing an elevated white cell count (12±290 cells/mm 3) (Fitzsimons et al., 1985). Protein and glucose levels were always normal. During usual and severe attacks, electroencephalography (EEG) shows a diffuse slow wave activity predominating on the hemisphere controlateral to the weak limbs (Gastaut et al., 1981). Periodic sharp waves (16), or dysrythmia (MuÈnte et al., 1990) have been rarely reported. EEG abnormalities may persist several days or weeks after the attack. Cranial CT scan or magnetic resonance imaging (MRI) performed during one severe attack may show aspects of hemispheric oedema. Interictal imaging is normal, except in some HM patients having permanent nystagmus or ataxia, in whom a cerebellar atrophy predominating on the anterior vermis may be observed (Fitzsimons et al., 1985; Joutel et al., 1993; Elliott et al., 1996). In EA2, CT scan or MRI may be normal or show cerebellar atrophy, either

160

A. Ducros et al.

affecting the anterior vermis, or more diffuse (Vighetto et al., 1988). Electromyogram is normal, the characteristic pattern of myokymia, i.e. spontaneous and periodic or rythmic activity of motor units at rest, always present in EA1, is not observed in EA2. In SCA6, brain imaging shows an isolated cerebellar atrophy, predominating on the anterior vermis. The aspects of cerebellar atrophy observed in SCA6 are indistinguishable from those observed in some cases of HM and EA2 (Satoh et al., 1998). Therapy In EA2, acetazolamide is remarkably effective in preventing attacks, whereas its withdrawal leads to the recurrence of attacks within a few days (Griggs et al., 1978; Zasorin et al., 1983). The effective dose ranges from 125 to 250 mg, two or three times daily. Rarely, some patients may not respond or require increasing doses of acetazolamide. The most frequent side effect is nephrolithiasis. Sodium valproate and ¯unarizine are also effective in preventing EA2 attacks. It is not clear if acetazolamide or the treatments which prevent attacks have any bene®t in preventing the progression of permanent cerebellar ataxia. No symptomatic treatment has been reported to abort attacks. In HM, treatment can be preventive or symptomatic (Ducros et al., 2000). In most patients, HM attacks are infrequent and prophylactic therapy is considered only to reduce the frequency of associated other forms of migraine attacks. In a few patients, HM attacks are frequent enough to justify a prophylactic therapy. However, due to the relative rarity of this condition, management of HM is mostly based on what is known about treatment other forms of migraine with aura. For preventive treatment, use of beta-blocking agents, speci®cally propanolol, calcium-blocking agents (¯unarizil, verapamil and nimodipine), phenytoin, papaverine and phenobarbital has been reported. Since the demonstration that HM and EA2 were allelic conditions due to mutations within the same gene, acetazolamide has been used to prevent HM attacks. But the number of reported cases is too low to conclude on its ef®ciency. Symptomatic treatment of HM attacks will try to relieve pain, nausea and vomiting. As full recovery is the rule even in severe attacks, the use of speci®c antimigraine drugs which effects have not been evaluated in HM is debatable. All vasoconstrictor agents should be avoided, such as ergotamine, dihydroergotamine, sumatriptan and other triptans. Genetics Incidence, mode of transmission HM, EA2 and SCA6 are autosomal dominant disorders. Precise incidence of HM

Migraine and ataxias

161

and EA2 is unknown. If the number of published families is considered to be related to the disease frequency, then HM is more frequent than EA2. In France, more than 80 HM and 14 EA2 families have been collected. In addition to familial forms, sporadic cases of HM, EA2 and SCA6 have been described. Chromosomal linkage Hemiplegic migraine HM has been shown to be genetically heterogeneous (Joutel et al., 1993, Joutel et al., 1994; Ophoff et al., 1994, Ophoff et al., 1996, Ducros et al., 1997). The ®rst responsible gene, located on chromosome 19p13.1 (Joutel et al., 1993), was identi®ed in 1996 as being CACNA1A, which encodes the main pore-forming a1A subunit of P/Q-type voltage-gated calcium channels (Ophoff et al., 1996). CACNA1A has been involved, on the basis of linkage or mutation screening data, in all 18 families with HM and progressive cerebellar ataxia (HM/PCA) reported so far (Joutel et al., 1994; Ophoff et al., 1994; Ducros et al., 1997, 1998). On the contrary, pure HM has been linked to at least three different genes: CACNA1A (Joutel et al., 1993, Joutel et al., 1994; Ophoff et al., 1996), a second gene mapped on the long arm of chromosome 1 (Ducros et al., 1997; Gardner et al., 1997), and at least a third gene still to be localized (Ducros et al., 1997). The HM gene located on 1q is yet unidenti®ed. Moreover, an American group found linkage to 1q31 in a large family, whereas linkage to 1q21±23 has been demonstrated in three French families. Further analysis have to be done to disclose whether chromosome 1q is the site of one or two HM genes. Except for cerebellar ataxia which appears to be present only in chromosome 19 linked families, very few differences have been found between families linked to different loci. In a study comparing clinical features between three chromosome 19linked families and two unlinked families, patients belonging to the former group were more likely to have attacks triggered by minor head trauma and to have severe attacks with unconsciousness (Terwindt et al., 1996). In another study comparing clinical and genetic data between three HM family groups (ten chromosome 19linked families, three chromosome 1-linked families and four families unlinked to both loci), two major genotype-phenotype correlations were observed (Ducros et al., 1998). First, penetrance seemed to be much lower in chromosome 1-linked families. Second, associated permanent cerebellar symptoms were observed in 50% of chromosome 19-linked families and in those families only. No signi®cant difference was observed between the three family-groups with regards to the characteristics of HM attacks, the occurrence of severe attacks, the existence of other migraine subtypes and the disease course. The incomplete penetrance of HM has several implications. First, an affected subject may have no ®rst or even second degree affected relative, making it dif®cult to diagnose familial hemiplegic migraine. Some of the apparent

162

A. Ducros et al.

sporadic cases of HM may thus result from the incomplete penetrance of the condition. Second, an asymptomatic family member may have affected children. Third, in the mapping of HM genes, only affected recombinants should be considered. Finally, the clinical variability within a given family and the incomplete penetrance suggests that modifying genetic and/or environmental factors play a role in the expression of the HM phenotype. Episodic ataxia type 2 EA2 was linked to the same interval than HM on chromosome 19p13 (Kramer et al., 1995; Tean Teh et al., 1995; Vahedi et al., 1995; Von Bredelow et al., 1995). Both disorders are characterized by transient neurological episodes starting during childhood or adolescence, associated in some patients to permanent nystagmus and ataxia with cerebellar atrophy. Those genetic and clinical similarities had led to the hypothesis that both conditions could be due to mutations within the same gene. Moreover, the identi®cation of ion channel genes' mutations in several peripheral neurological paroxysmal disorders responsive to acetazolamide suggested that the gene for HM and EA2 could encode an ion channel (Vahedi et al., 1995). These hypothesis were con®rmed in 1996 by the identi®cation of mutations in the CACNA1A gene encoding the a1A subunit of P/Q type voltage gated calcium channels in both disorders (Ophoff et al., 1996). SCA6 SCA6 is due to a CAG repeat expansion in the CACNA1A gene located on chromosome 19p13 (Zhuchenko et al., 1997). CACNA1A and the a 1A subunit of P/Q type voltage-gated calcium channels Therefore, three distinct autosomal dominant neurological conditions HM, EA2 and SCA6 are caused by mutations within CACNA1A. This gene encodes the main subunit of a voltage-dependant calcium channel. Those channels are responsible for the speci®c in¯ux of calcium into the cell in response to membrane depolarization (Catterall, 1995; Ackerman et al., 1997). They are involved in a great diversity of physiological processes including control of membrane excitability, enzyme activity regulation, gene expression, muscle contraction and hormone secretion. In neurons, they play a major role in synaptogenesis and axonal growth during development, and in synaptic transmission. Voltage-gated calcium channels Six functional subclasses of voltage-dependent calcium channels are de®ned based on electrophysiological and pharmacological criteria (Zhang et al., 1993; De Waard et al., 1996). Two major classes are distinguished: low-voltage activated

Migraine and ataxias

163

(T type) and high-voltage activated (L, N, P, Q and R) channels. Calcium channels are multimeric complexes containing a major transmembrane pore-forming a1 subunit associated with smaller auxillary subunits (b, a2/d and g) (Perez-Reyes et al., 1995; De Waard et al. 1996).

a 1 subunits The a1 subunit is a large (190±260 kDa) hydrophobic protein which forms the ionic pore, is responsible for the calcium selectivity, the voltage sensitivity and the channel gating properties. At least nine different genes encoding a1 subunits have been identi®ed (CACNA1A, B, C, D, E, F, G and H). Those different subunits are responsible for the electrophysiological and pharmacological diversity of calcium currents. All a1 subunits have the same structure, which is also found in voltage-gated sodium a1 subunits. They contain four homologous domains (I±IV) each containing six putative a-helix membrane spanning segments (S1±S6). The four domains are linked by intracytoplasmic loops and fold within the cellular membrane to build the ionic pore. Small hydrophobic loops between each S5 and S6 segments are called Ploops (P standing for pore) because they line the inner part of the pore. Each P-loop harbors a conserved glutamate residue, and the combination of the four glutamates is though to form a speci®c and dynamic ®lter for calcium ions. The four S4 segments are responsible for voltage sensitivity. They contain positive charged arginines regularly disposed every three or four amino acids (Varadi et al., 1995). During membrane depolarization, these charges are moving within the electrical ®eld, inducing conformational changes and thus pore opening. The structures responsible for channel inactivation seem to be the IS6 segment, the linker between domains I and II, and probably some regions of the S5 and S6 segments. The molecular diversity of a1 subunits is increased by alternative splicing. For all known a1 subunits genes, the location of alternatively spliced regions is similar, affecting the N-terminal region, the I±II loop, IIS6, the II±III loop, IIIS2, IVS3 and the C-terminal region (Perez-Reyes et al., 1995). Auxillary subunits The b subunit is an intracytoplasmic protein interacting with the a1 subunit through a conserved region of the I±II cytoplasmic linker, called alpha interaction domain (De Waard et al., 1994). These subunits play an important regulatory role (Varadi et al., 1995; Yamaguchi et al., 1998). Their coexpression with a1 subunits inXenopus oocytes immediately increases the activation and inactivation kinetics of calcium currents, and secondary dramatically increases calcium currents amplitudes and expression of a1 subunits in the membrane. Four genes encoding b subunits have been identi®ed, and all are submitted to alternative splicing. All b subunits can interact with the various a1 subunits, producing different calcium currents. The regulatory effects of b subunits is probably mediated by to different mechanisms. First, interaction with b subunits could

164

A. Ducros et al.

produce an allosteric modulation of a1 subunits, inducing a change from inactivated to activated forms. Second, b subunits could act are chaperons, increasing the expression of a1 subunits in the membrane (Yamaguchi et al., 1998). The a2/d subunit is encoded by a single gene and consists of a glycosylated extracellular a2 protein linked by disul®de bonds to a transmembrane d protein. Coexpression of a2/d with a1 and b subunits moderately increases calcium currents and modulates inactivation kinetics role (Varadi et al., 1995). The g subunit is a transmembrane protein which stabilizes the channel inactivation. This subunit was, until recently, though to be part only from L-type calcium channels expressed in skeletal muscle. A neuronal g subunit (g2) was recently cloned, and is probably playing a role in the molecular structure of P/Q channels (Letts et al., 1998). CACNA1A, a 1A subunits and P/Q type channels The CACNA1A gene is localized in humans on chromosome 19p13. The 47 known exons span about 350 kb. Intron 7 contains a polymorphic CA repeat (D19S1150) (Ophoff et al., 1996). Exon 47 contains a polymorphic CAG repeat which is predicted to code for a polyglutamine tract in three of the six known human isoforms (Zhuchenko et al., 1997). The gene encodes a 9.8-kb RNA (including 7800 bp of coding sequence) which is almost exclusively expressed in central and peripheral neurons (motoneurons). The RNA, detected in rat brain by in situ hybridization with riboprobes, is heavily expressed in the cerebellum, mainly in Purkinje cells, but also in granular neurons (Stea et al., 1994; Ludwig et al., 1997). High level of expression are also found in the hippocampus, less in cortex, olfactory bulb, thalamus, hypothalamus and brainstem. Expression of the a1A subunit, analyzed with anti-a1A antibodies, is concordant with in situ hybridization (Westenbroek et al., 1995). At the subcellular level, a1A subunits are particularly dense at presynaptic terminals (Sakurai et al., 1995). They are also present on the dendrites and the neuronal bodies. In association with a b, a2/d and probably the g 2subunits, a1A subunits form the ionic pore of P- and Q-type channels. In addition, P/Q channels contain a 95 kDa protein, which was recently identi®ed as being a truncated form of a1A, resulting from alternative splicing or proteolysis, which role is still unknown (Scott et al., 1998). P/Q-type channels are expressed in a large variety of neurons where they play an important role in the control of membrane excitability, neurotransmitter release and gene expression (Catterall, 1995). They are the predominant calcium channel in Purkinje cells (P-type currents) and also highly expressed in cerebellar granule cells (P- and Q-type currents) (Llinas et al., 1992; Stea et al., 1994; Westenbroek et al., 1995). The a1A subunit is able to generate, at least, the two different P- and Q-type calcium currents. Both are high voltage activated calcium channels, insensitive to dihydropyridines. But P- and Q-types can be distinguished based on different sensi-

Migraine and ataxias

165

tivity to other toxins and by their inactivation kinetics. P-type currents are slowly inactivating whereas Q-type currents have a fast inactivation. Finally, P-type currents form 90% of all calcium currents observed in Purkinje cells, whereas Qtype currents are important in cerebellar granular cells (35% of all calcium currents), but also in hippocampal neurons where their proeminant of the role in glutamatergic neuronal transmission has been established. Differences between P- and Q-types could result from alternative splicing of a1A subunits or from the association with different b subunits. Indeed, it has been shown that alternative splicing generates a1A subunits with distinct subcellular localization, biochemical and electrophysiological properties (Sakurai et al., 1995). In addition, the same a1A isoform coexpressed with different ancillary b subunits in a mammalian cell line gives raise to distinct types of currents (Moreno et al., 1997). Other hypothesis could be the association with other regulatory subunits or the effects of postraductional changes. CACNA1A mutations and genetic screening Hemiplegic migraine Thirteen CACNA1A mutations have been identi®ed so far in 25 families and two sporadic cases (Ophoff et al., 1996; Terwindt et al. 1998; Ducros et al., 1999a,Ducros et al., 1999b; Gardner et al. 1999). All these mutations are missense mutations (see Table 1). All are located within exons coding for S4±S6 segments. T666M, a predominant substitution was detected in ten out of those 25 families and in one sporadic case. This genetic defect was demonstrated to arose through recurrent mutation events and to be speci®cally associated with the presence of a permanent cerebellar ataxia (Ducros et al., 1999a). Three other recurrent mutations (R583Q, R1668W and I1811L) have been also identi®ed in hemiplegic migraine with ataxia. A de novo mutation was shown to cause a severe form of hemiplegic migraine with ataxia in one sporadic case (Vahedi et al., 1999). Only ®ve of all these identi®ed mutations are associated with pure FHM (Ophoff et al., 1996; Ducros et al., 1999a). Episodic ataxia type 2 Eleven different CACNA1A mutations have been identi®ed in seven families and four sporadic cases (see Table 2) (Ophoff et al., 1996; Yue et al., 1998; Jen et al., 1999; Denier et al., 1999,Denier et al., 2000). In one of the sporadic cases, a de novo nonsense mutation of CACNA1A was demonstrated (Yue et al., 1998). In a second non-familial case, it was established that incomplete penetrance was the cause for the occurrence of sporadic EA2 (Denier et al., 1999). Ten of these mutations are predicted to lead to truncated or aberrant a1A subunits. Interestingly, one mutation (a 3-bp deletion) leads neither to a truncated protein nor to a frame shift.

166

Table 1 CACNA1A mutations identi®ed in HM and EA2

Phenotype

Number of families

Location

Domain

Nucleotide change

Consequence

Ducros et al. (1999a,b) Ophoff et al. (1996); Ducros et al. (1999a,b) Ducros et al. (1999a,b) Ducros et al. (1999a,b) Ducros et al. (1999a,b) Gardner et al. (1999) Ducros et al. (1999a) Ophoff et al. (1996) Ophoff et al. (1996) Ducros et al. (1999a,b) Ophoff et al. (1996) Ducros et al. (1999a,b) Ducros et al. (1999a,b)

HM 1 ataxia HM 1 ataxia

3 11

Exon 13 Exon 16

IIS4 IIP loop

CGa ! cAa aCg ! aTg

Arg583Glu Thr666Met

HM 1 ataxia HM 1 ataxia HM 1 ataxia HM 1 ataxia HM 1 ataxia HM 1 ataxia Pure HM Pure HM Pure HM Pure HM Pure HM

1 De novo 2 1 1 2 1 1 1 1 1

Exon 17 Exon 26 Exon 32 Exon 32 Exon 32 Exon 36 Exon 4 Exon 4 Exon 17 Exon 25 Exon 33

IIS6 IIIS5 IVS4 IVS4-S5 IVS4-S5 IVS6 IS4 IS4 IIS6 IIIS3-S4 IVS5

gaC ! gaG TAc ! tGc Cgg ! Tgg CTn ! cCn Tgg ! Cgg Atc ! Ctc CGa ! cAa aGg ! aAg gTg ! gCg Aaa ! Gaa Gtc ! Atc

Asp715Glu Tyr1385Cys Arg1668Trp Leu1682Pro Trp1664Arg Ile1811Leu Arg192Gln Arg195Lys Val714Ala Lys1336Glu Val1696Ile

A. Ducros et al.

References

Table 2 CACNA1A mutations in EA2

Numberof families

Location

Domain

Nucleotide change

Consequence

Denier et al. (1999) Denier et al. (1999) Denier et al. (1999) Ophoff et al. (1996) Yue et al. (1998) Ophoff et al. (1996) Denier et al. (1999) Denier et al. (1999) Denier et al. (1999) Jen et al. (1999) Denier et al. (1999)

1 1 1 1 De novo 1 1 1 1 1 1

Intron 11 Exon 16 Exon 22 Exon 22 Exon 23 Intron 24 Intron 26 Exon 27 Exon 29 Exon 29 Exon 30

± IIS6 IIIS1 IIIS1 IIIS2 ± ± IIIPloop IVS1 IVS1 IVS2

TTTgt ! TTTat Del AG2259±60 Del C3797 Del C4073 Cga ! Tga g4270 1 1 ! a TCGgt ! TCGat taC ! taG Cga ! Tga C4914T Del CTT4778±80

Aberrant splicing Frameshift and stop780 Frameshift and stop1293 Frameshift and stop1293 Arg1279Stop Aberrant splicing Aberrant splicing Tyr1443Stop Arg1546Stop Premature stop Del Tyr1594 and Ala1593Asp

Migraine and ataxias

References

167

168

A. Ducros et al.

SCA6 SCA6 is caused by small expansions of a CAG repeat, located within the 3 0 end of CACNA1A and predicted to code for a polyglutamine tract in three of the six known human splice variants (Zhuchenko et al., 1997). As in other neurological conditions due to a CAG expansion, the number of repeats is inversely correlated to the age of onset of clinical symptoms. However, the CAG expansion in SCA6 seems to be stable during transmission from parents to children: only two intergenerational changes have been reported in more than 178 families (Jodice et al., 1997; Matsuyama et al., 1997). The difference between normal and pathological alleles is very small: in normal subjects, the number of CAG repeats ranges from 4 to 20, whereas SCA6 patients have between 21 and 33 repeats. The polyglutamine expansion is shorter than in other CAG expansion related disorders such as Huntington disease, SCA1, SCA2, SCA3, SCA7, DRPLA and Kenedy's syndrome (CAG ˆ 36± 121). Surprisingly, CAG expansions were identi®ed in three families associating paroxysmal and permanent progressive ataxia (Geschwind et al., 1997; Jodice et al., 1997). Finally, a missense mutation (G293A) located within the IS5-S6 loop, was detected in a single family in which patients suffered from both paroxysmal episodes indistinguishable from those observed in EA2 and a rapidly progressive and very severe permanent ataxia (Yue et al., 1997). Genetic screening Diagnostic genetic testing is now theoretically possible in HM, EA2 and SCA6. In subjects with suspected SCA6, the search for a CAG repeat ampli®cation in CACNA1A by the polymerase chain reaction (PCR) provides an easy and reliable diagnosis tool. This is not the case for HM and EA2 in which causative mutations are located all over the coding sequence of the gene. In HM, it could be proposed to directly search for the T666M substitution which is present in more than 50% of families affected by HM and cerebellar ataxia. New technologies providing time and cost-effective mutation detection method are needed to routinely search for CACNA1A mutations as a diagnosis tool in HM. In EA2, same technical problems have to be solved. In addition to screening of the whole coding sequence, since some families with EA2 have CAG expansions, one should analyze the CAG length. Genotype-phenotype correlation in CACNA1A related conditions Three distinct types of CACNA1A mutations seem to be related to three different phenotypes: missense mutations to HM, truncating mutations to EA 2 and CAG expansions to SCA6. However, the genotype-phenotype correlations are not so clear cut: indeed, CAG expansions may cause a clinical picture indistinguishable from

Migraine and ataxias

169

EA2 (Geschwind et al., 1997; Jodice et al., 1997) and a missense mutation can produce a very severe condition with both episodic and progressive cerebellar ataxia (Yue et al., 1997). In hemiplegic migraine, both the pure form of the condition and the form with associated cerebellar ataxia are caused by missense mutations located in important functional domains of the channel (segments S4±S6 and P-loops). However, the mechanisms leading from a given mutation to the presence or absence or cerebellar symptoms in addition to HM are unknown. One important issue is the high clinical variability observed within given HM and EA2 families having the same mutation. This variability concerns both episodic symptoms (hemiplegic migraine attacks or ataxic spells) and permanent symptoms. This profound variability suggests that other factors than the sole CACNA1A mutation are important to produce the clinical phenotype. Modifying genetic or environmental factors may play a role. With regards to allelic modifying factors, two studies found no role of the length of the intragenic CAG repeat on the severity of episodic as well as permanent symptoms in HM with ataxia (Ducros et al., 1999a) and in EA2 (Denier et al., 2000).

Pathogenesis of CACNA1A mutation related conditions Functional disturbances on molecular level Only four CACNA1A mutations causing HM (R192Q, V714A, T666M and I1811L) have been investigated so far in order to detect their putative effects on a1A calcium currents (Kraus et al., 1998; Hans et al., 1999). In the ®rst study, mutant rabbit cDNAs were expressed inXenopus oocytes. Three of the mutations were demonstrated to change the a1A subunit gating properties (Kraus et al., 1998). Substitutions T666M and I1811L (causing HM with ataxia) changed the inactivation kinetics of a1A channels. But, T666M decreased the rate of recovery from inactivation whereas I1811L increased the rate of this recovery. Moreover, and on the contrary of expected, mutations causing pure HM and mutations causing HM with ataxia do not lead to clearly distinct electrophysiological patterns. For R192Q (pure HM), which removes one positive charge in IS4, no modi®cation could be detected in this ®rst study. Finally, mutation V714A (causing pure HM), located in IIS6, accelerates recovery from inactivation in a very similar way than I1811L which causes HM with ataxia. In a second study, the same four mutations were introduced in human a1A cDNAs and expressed in human embryonic kidney 293 cells (Hans et al., 1999). All four mutations, including R192Q, were showed to affect both the biophysical properties and the density of functional channels. Mutation R192Q increased the

170

A. Ducros et al.

density of functional P/Q-type channels and their open probability. Mutation T666M decreased both the density of functional channels and their unitary conductance. Mutations V714A and I1811L were again shown to have similar consequences despite the fact that V714A causes pure HM and I1811L HM with ataxia: they both shifted the voltage range of activation toward more negative voltages, increased both the open probability and the rate of recovery from inactivation, and decreased the density of functional channels. Interestingly, the reduction in single-channel conductance induced by mutations T666M and V714A was not observed in some patches or periods of activity, suggesting that the abnormal channel may switch on and off, perhaps depending on some unknown factor. These data suggest that the HM mutations can lead to both gain- and loss-of-function of human P/Q-type calcium channels. Functional consequences of truncating mutations causing EA2 and CAG expansions causing SCA6 have not been analyzed. Animal models for CACNA1A disorders: the tottering and leaner mutant mice Two recessive mutations in the murine homolog of CACNA1A have been identi®ed as responsible for the tottering and leaner phenotypes (Fletcher et al., 1996). Tottering These mutant mice have a neurological phenotype associating paroxysmal and permanent symptoms. They have absence-epilepsy, with a critical electroencephalogram showing generalized regular spike-wave discharges similar to those observed during an absence in humans. they also have rare motor seizures, characterized by clonic movements of both lower limbs, slowly progressing to the upper limbs and trunk before ending abruptly. These attacks last 20±30 min, during which the EEG remains normal. It has recently been shown that during those motor attacks, the initial expression of c-fos is cerebellar with a secondary activation of the cerebral cortex. Tottering mice develop after the third week of life a moderate cerebellar ataxia responsible for a hopping gait. No cerebellar atrophy is found. Electronic microscopy disclose some shrunken Purkinje cells. Tottering mice have a clear histological abnormality of the locus coeruleus, with an abnormal synaptogenesis of noradrenergic ®bers. Moreover, an abnormal persistent expression of tyrosine hydroxylase (TH) in Purkinje cells has been demonstrated (TH is normally expressed from the third to the ®fth week) (Hess et al., 1991). The tottering mutation (C1802T; Pro601Leu) is located in the IIS5-S6 loop, and produces the phenotype in homozygous mice (Fletcher et al., 1996). Interestingly, this mutation is located at a few bases from the T666M mutation causing HM with ataxia. The consequences of the tottering mutation on calcium currents is unknown.

Migraine and ataxias

171

Leaner The leaner phenotype is also transmitted as an autosomal recessive trait. Like tottering, leaner mice have absence epileptic seizures and persistent TH expression in Purkinje cells. But they do not have motor seizures, they develop a severe cerebellar ataxia and their life-span is reduced. Histological examination shows a dramatic loss of Purkinje cells, granular and Golgi neurons, which increases in a few months. Despite the uniform expression of a1A subunits within wild type mice cerebellum, the loss of Purkinje cells in leaner is not homogeneous but follows a parasagittal disposition, with bands of severe neuronal loss separated by bands containing surviving cells. The 15% of surviving Purkinje cells show an axonal swelling and an abnormal dendritic development. The leaner mutation alters a CACNA1A splice site and induces the production of two aberrant transcripts after the fourth domain: one two long resulting from the absence of splicing out an intron and the other truncated (Fletcher et al., 1996). It was recently showed by whole-cell recording of Purkinje neurons from leaner mice, that v-Aga-IVA sensitive currents, normally representing 85% of the whole calcium currents, were reduced by 65% (Lorenzon et al. 1998). This work suggests that a1A subunits generates v-Aga-IVA sensitive P-type currents and that these currents are necessary to the normal cerebellar function. Tottering and leaner as animal models for HM, EA2 or SCA6? Tottering and leaner are recessive mutations. HM, EA2 and SCA6 are autosomal dominant human conditions. Mechanisms leading from genotypes to phenotypes are certainly different. However, the analysis of these mutants provides some important information. First, the decrease of calcium in¯ux mediated by mutated a1A channels may cause cell death. Second, this neuronal loss is not homogeneous, but spars some cells according to a parasagittal banding pattern. The factors favoring the survival of these neurons are unclear. Third, the abnormalities of the surviving Purkinje cells in leaner, and of the locus coeruleus in tottering suggest an implication of a1A subunits in neuronal development. Fourth, the abnormal TH expression in Purkinje cells is consistent with the role of calcium in the regulation of gene expression. This abnormality suggests that P-type channels, which are the major calcium channels located on Purkinje cells neuronal bodies, are implicated in this regulation. CACNA1A disorders: from the genotype to the phenotype To date, our knowledge about these disorders is not suf®cient to propose realistic hypothesies. However, all these data suggest that channels formed by a1A subunits have speci®c functions in speci®c neurons, of which only some are modi®ed by a given mutation. These functions could be the development of speci®c neuronal populations, the control of speci®c genes' expression or the neurotransmission in

172

A. Ducros et al.

some peculiar neuronal pathways. This hypothesis is underlined by the remarkable neuronal phenotype restriction in the tottering and leaner mice. This restriction, which could appear in contradiction with the very large expression of a1A subunits throughout the central and peripheral neurons, favors the hypothesis of a specialization of the various a1A channels. This specialization could be mediated by the association of with different auxillary subunits according to the neurons, the existence of functionally different a1A isoforms, the rate of expression of a1A currents related to other calcium currents in a given neuron and the balance between activating and inactivating interaction. Perspectives New therapeutic approaches Despite the identi®cation of the causative genes, few therapeutic progresses have been made in HM, EA2 and SCA6. Based on the ef®ciency of acetazolamide in preventing EA2 attacks, acetazolamide should be tried in severe cases of HM with frequent attacks. In EA2, it is still not known how acetazolamide acts to prevent the attacks. Subsequent research aims In CACNA1A related conditions, the perspectives are the following: ² Characterization of the neuronal and subcellular expression of the various a1A isoforms in normal human brain. ² Electrophysiological analysis of the various isoforms coexpressed with different regulatory subunits. ² Consequences of CACNA1A mutations on these various currents. ² Animal models. The major problem of animal models in paroxysmal disorders such as HM and EA2 is related to the phenotype itself. How to diagnose that a mouse is having a hemiplegic migraine attack ? The answer supposes to develop an objective method of diagnosis.

Note added in proof Episodic ataxia type 1 (EA1) The ®rst human channelopathy detected involving a voltage-gated potassium channel is linked to the Shaker homologue, KCNA1 on chromosome 12p13 (Litt et al., 1994). EA1 is an autosomal dominant disease characterized by episodic failure

Migraine and ataxias

173

of excitation of cerebellar neurons and sustained ®ring of the motoneurons (Van Dyke et al., 1975). The resulting interictal myokymia in facial muscles and distal extremities is associated with rhythmic activity of motor units in the EMG. Onset of motion and exercise may provoke attacks of atactic gait and jerking extremity movements that last for seconds to minutes. Adequate treament, as expected for neuronal hyperexcitability, consists of anticonvulsants such as carbamazepine. The kinesigenic attacks also respond to some extent to acetazolamide (Gancher and Nutt, 1986). Interestingly EA1 families show an over-representation of epilepsy (Zuberi et al., 1999). Several missense mutations have been detected in the resulting gene product Kv1.1 (Table 3), most of which have been expressed in Xenopus oocytes. While four of the mutants did not yield signi®cant currents, others showed change of function leading to enhanced deactivation and C-type inactivation (Val-408-Ala, Gly-311-Ser), marked reduction of the channel open duration (Val-408-Ala) or right-shift of voltage dependence of activation and slowing of time course of activation (Phe-184-Cys, Val-1174-Phe, Glu-325-Asp, Thr-226-Ala/Met) (Adelman et al., 1995; Boland et al., 1997; D'Adamo et al., 1998, 1999; Zerr et al., 1998a,b). Similar effects have been observed when these mutations were expressed in mammalian cells (Bretschneider et al., 1999). Coexpression of the mutants with wild-type-mimicking in vivo conditions revealed current reduction between 26 and 100% indicating a dominant negative interaction (Adelman et al., 1995). Heteromeric channels containing human Kv1.1 and Kv1.2 subunits revealed biophysical and pharmacological properties intermediate between the respective homomers (D'Adamo et al., 1999). Table 3 KCNA1 mutations. The intronless gene encodes kv1.1, the human analog of the Shaker gene. The mutations cause episodic ataxia type 1 (EA1)

Genotype

Region

Mutation

Reference

C520A T527A T551G A676G C677T C677G G715T T745A

S1 S1 S1 S2 S2 S2 S2 S2/S3 S4/S5 S4/S5 S6 S6

Val-174-Phe Ile-176-Arg Phe-184-Cys Thr-226-Ala Thr-226-Met Thr-226-Arg Arg-239-Ser Phe-249-Ile Gly-311-Ser Glu-325-Asp Val-404-Ile Val-408-Ala

Browne et al. (1994) Scheffer et al. (1998) Browne et al. (1994) Scheffer et al. (1998) Comu et al. (1996) Zuberi et al. (1999) Browne et al. (1994) Browne et al. (1994) Zerr et al. (1998b) Browne et al. (1994) Scheffer et al. (1998) Browne et al. (1994)

G975C G1210A A1223G

174

A. Ducros et al.

Disease pathogenesis might therefore be explained by a reduced repolarizing effect of the delayed recti®er leading to broadening of the action potentials and prolongation of transmitter release. Due to the strong expression of KCNA1 in basket cells of the cerebellum, an imbalance between inhibitory and excitatory input could well destabilize motor control under stress or exercise leading to kinesiogenic ataxia. The restriction of clinical symptoms to cerebellum and perhaps the peripheral motoneuron despite almost ubiquitious expression in nervous tissue (Veh et al., 1995), may be due to the synergistic function of KCNA2 generating the so-called dendrotoxin-sensitive delayed recti®er (Rhodes et al., 1997).

References Ackerman, M.J., Clapham, D.E., 1997. Ion channels. Basic science and clinical disease. N. Engl. J. Med. 336, 1575±1586. Adelman, J.P., Bond, C.T., Pessia, M., Maylie, J., 1995. Episodic ataxia results from voltage-dependent potassium channels with altered functions. Neuron 15 (6), 1449±1454. Baloh, R.W., Yue, Q., Furman, J.M., Nelson, S.F., 1997. Familial episodic ataxia: clinical heterogeneity in four families linked to chromosome 19p. Ann. Neurol. 41 (1), 8±16. Blau, J.N., Whitty, C.W.M., 1955. Familial hemiplegic migraine. Lancet 2, 1115±1116. Boland, L.M., Price, D.L., Jackson, K.A., 1997. Functional consequences of potassium channel mutations identi®ed in families with inherited episodic ataxia. Biophys. J. 72, A140. Brandt, T., Strupp, M., 1997. Episodic ataxia type 1 and 2 (familial periodic ataxia/vertigo). Audiol. Neurootol. 2 (6), 373±383. Bretschneider, F., Wrisch, A., Lehmann-Horn, F., Grissmer, S., 1999. Expression in mammalian cells and electrophysiological characterization of two mutant kv1.1 channels causing episodic ataxia type 1 (EA-1). Eur. J. Neurosci. 11, 2403±2412. Browne, D.L., Gancher, S.T., Nutt, J.G., Brunt, E.R.P., Smith, E.A., Kramer, P., Litt, M., 1994. Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA1. Nat. Genet. 8, 136±140. Catterall, W.A., 1995. Structure and function of voltage-gated ion channels. Annu. Rev. Biochem. 64, 493±531. Comu, S., Giuliani, M., Narayanan, V., 1996. Episodic ataxia and myokymia syndrome: a new mutation of potassium channel gene Kv1.1. Ann. Neurol. 40 (4), 684±687. D'Adamo, M.C., Liu, Z., Adelman, J.P., Maylie, J., Pessia, M., 1998. Episodic ataxia type-1 mutations in the hKv1.1 cytoplasmic pore region alter the gating properties of the channel. EMBO J. 17 (5), 1200± 1207. D'Adamo, M.C., Imbrici, P., Sponcichetti, F., Pessia, M., 1999. Mutations in the KCNA1 gene associated with episodic ataxia type-1 syndrome impair heteromeric voltage-gated K(1) channel function. FASEB J. 13, 1335±1345. De Waard, M., Pragnell, M., Campbell, K.P., 1994. Ca 21 channel regulation by a conserved beta subunit domain. Neuron 13 (2), 495±503. De Waard, M., Gurnett, C.A., Campbell, K.P., 1996. Structural and functional diversity of voltageactivated calcium channels. In: Narahashi, T. (Ed.). Ion Channels, Plenum Press, New York, pp. 41±87. Denier, C., Ducros, A., Joutel, A., Vahedi, K., Tournier-Lasserve, E., 1999. High prevalence of

Migraine and ataxias

175

CACNA1A truncating mutations and broader clinical spectrum in episodic ataxia type 2. Neurology 52(Suppl 2), A116. Denier, C., Ducros, A., Vahedi, K., Joutel, A., Thierry, P., Ritz, A., Castelnovo, G., Deonna, T., GeÂrard, P., Devoize, J.L., Gayou, A., Perrouty, B., Soisson, T., Autret, A., Warter, J.M., Vighetto, A., Van Bogaert, P., Alamowitch, S., Roulet, E., Tournier-Lasserve, E., 2000. High prevalence of CACNA1A truncations and broader clinical spectrum in episodic ataxia type 2. Neurology, 52, 1816±1821. Donat, J.R., Auger, R., 1979. Familial periodic ataxia. Arch. Neurol. 36, 568±569. Ducros, A., Joutel, A., Vahedi, K., Cecillon, M., Ferreira, A., Bernard, E., Verier, A., Echenne, B., 1997. Lopez de Munain A. Bousser M.G., Tournier-Lasserve E. Mapping of a second locus for familial hemiplegic migraine to 1q21±q23 and evidence of further heterogeneity. Ann. Neurol. 42 (6), 885±890. Ducros, A., Joutel, A., Vahedi, K., Bousser, M.G., Tournier-Lasserve, E., 1998. Genotype-phenotype correlations in familial hemiplegic migraine. Neurology 50 (Suppl 4), A352. Ducros, A., Denier, C., Joutel, A., Vahedi, K., Bousser, M.G., Tournier-Lasserve, E., 1999. Characterization of CACNA1A mutations in familial and sporadic hemiplegic migraine. Neurology 52 (Suppl 2), A273. Ducros, A., Denier, C., Joutel, A., Vahedi, K., Michel, A., Darcel, F., Madigand, M., Guerouaou, D., Tison, F., Julien, J., Hirsch, E., Chedru, F., Bisgard, C., Lucotte, G., Despres, P., Billard, C., Barthez, M.A., Ponsot, G., Bousser, M.G., Tournier-Lasserve, E., 1999. Recurrence of the T666M calcium channel CACNA1A gene mutation in familial hemiplegic migraine with progressive cerebellar ataxia. Am. J. Hum. Genet. 64 (1), 89±98. Ducros, A., Campbell, K., 2000. Familial hemiplegic migraine. In: Olesen, J., Telft-Hansen, P., Welch, K.M.A. (Eds.). The Headaches, 2nd edition, Lippincott Williams & Wilkins, New York, Chapter 62. Elliott, M.A., Peroutka, S.J., Welch, S., May, E.F., 1996. Familial hemiplegic migraine, nystagmus, and cerebellar atrophy. Ann. Neurol. 39 (1), 100±106. Feeney, G.F.X., Boyle, R.S., 1989. Paroxysmal cerebellar ataxia. Aust. NZ J. Med. 19, 113±117. Fitzsimons, R.B., Wolfenden, W.H., 1985. Migraine coma. Meningitic migraine with cerebral oedema associated with a new form of autosomal dominant cerebellar ataxia. Brain 108 (3), 555±577. Fletcher, C.F., Lutz, C.M., O'Sullivan, T.N., Shaughnessy Jr, J.D., Hawkes, R., Frankel, W.N., Copeland, N.G., Jenkins, N.A., 1996. Absence epilepsy in tottering mutant mice is associated with calcium channel defects. Cell 87 (4), 607±617. Gancher, S.T., Nutt, J.G., 1986. Autosomal dominant episodic ataxia: a heterogeneous syndrome. Mov. Disord. 1 (4), 239±253. Gardner, K., Barmada, M.M., Ptacek, L.J., Hoffman, E.P., 1997. A new locus for hemiplegic migraine maps to chromosome 1q31. Neurology 49 (5), 1231±1238. Gardner, K., Bernal, O., Keegan, M., Badger, J., Gerber, O., Lowry, N., Hoffman, E.P., 1999. A new mutation in the Chr19p calcium channel gene CACNL1A4 causing hemiplegic migraine with ataxia. Neurology 52 (Suppl 2), A115. Gastaut, J.L., Yermenos, E., Bonnefoy, M., Cros, D., 1981. Familial hemiplegic migraine: EEG and CT scan study of two cases. Ann. Neurol. 10 (4), 392±395. Geschwind, D.H., Perlman, S., Figueroa, K.P., Karrim, J., Baloh, R.W., Pulst, S.M., 1997. Spinocerebellar ataxia type 6. Frequency of the mutation and genotype- phenotype correlations. Neurology 49 (5), 1247±1251. Griggs, R.C., Moxley, R.T., Lafrance, R.A., 1978. Mc Quilen. J. Hereditary paroxysmal ataxia: response to acetazolamide. Neurology 28, 1259±1264. Haan, J., Terwindt, G.M., Bos, P.L., Ophoff, R.A., Frants, R.R., Ferrari, M.D., 1994. Familial hemiplegic migraine in The Netherlands. Dutch Migraine Genetics Research Group. Clin. Neurol. Neurosurg. 96 (3), 244±249. Haan, J., Terwindt, G.M., Ophoff, R.A., Bos, P.L., Frants, R.R., Ferrari, M.D., Krommenhoek, T.,

176

A. Ducros et al.

Lindhout, D.L., 1995. Sandkuyl L.A.,Van Eyk R., Is familial hemiplegic migraine a hereditary form of basilar migraine? Cephalalgia 15 (6), 477±481. Hans, M., Luvisetto, S., Williams, M.E., Spagnolo, M., Urrutia, A., Tottene, A., Brust, P.F., Johnson, E.C., Harpold, M.M., Stauderman, K.A., Pietrobon, D., 1999. Functional consequences of mutations in the human alpha1A calcium channel subunit linked to familial hemiplegic migraine.. J. Neurosci. 19 (5), 1610±1619. Hess, E.J., Wilson, M.C., 1991. Tottering and leaner mutations perturb transient developmental expression of tyrosine hydroxylase in embryonicaly distinct Purkinje cells. Neuron 6, 123±132. Hill, W., Sherman, H., 1968. Acute intermittent familial cerebellar ataxia. Arch. Neurol. 18, 350±357. IHS, 1988. Classi®cation and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Headache Classi®cation Committee of the International Headache Society. Cephalalgia 8 (Suppl 7), 1±96, Jen, J.C., Yue, Q., Yu, F.H., Litt, M., Nutt, J.G., Nelson, S.F., Baloh, R.W., 1999. Incomplete penetrance and phenotypic variability of a novel nonsense mutation in CACNA1A causing episodic ataxia and hemiplegic migraine. Neurology 52 (Suppl 2), A553. Jodice, C., Mantuano, E., Veneziano, L., Trettel, F., Sabbadini, G., Calandriello, L., Francia, A., Spadaro, M., Pierelli, F., Salvi, F., Ophoff, R.A., Frants, R.R., Frontali, M., 1997. Episodic ataxia type 2 (EA2) and spinocerebellar ataxia type 6 (SCA6) due to CAG repeat expansion in the CACNA1A gene on chromosome 19p. Hum. Mol. Genet. 6 (11), 1973±1978. Joutel, A., Bousser, M.G., Biousse, V., Labauge, P., Chabriat, H., Nibbio, A., Maciazek, J., Meyer, B., Bach, M.A., Weissenbach, J., et al., 1993. A gene for familial hemiplegic migraine maps to chromosome 19. Nat. Genet. 5 (1), 40±45. Joutel, A., Ducros, A., Vahedi, K., Labauge, P., Delrieu, O., Pinsard, N., Mancini, J., Ponsot, G., Gouttiere, F., Gastaut, J.L., et al., 1994. Genetic heterogeneity of familial hemiplegic migraine. Am. J. Hum. Genet. 55 (6), 1166±1172. Kramer, P.L., Yue, Q., Gancher, S.T., Nutt, J.G., Baloh, R., Smith, E., Browne, D., Bussey, K., Lovrien, E., Nelson, S., et al., 1995. A locus for the nystagmus-associated form of episodic ataxia maps to an 11-cM region on chromosome 19p. Am. J. Hum. Genet. 57 (1), 182±185. Kraus, R.L., Sinnegger, M.J., Glossmann, H., Hering, S., Striessnig, J., 1998. Familial hemiplegic migraine mutations change alpha1A Ca 21 channel kinetics. J. Biol. Chem. 273 (10), 5586±5590. Letts, V.A., Felix, R., Biddlecome, G.H., Arikkath, J., Mahaffey, C.L., Valenzuela, A., Bartlett, F.S., Mori, Y., Campbell, K.P., Frankel, W.N., 1998. The mouse stargazer gene encodes a neuronal Ca 21channel gamma subunit. Nat. Genet. 19 (4), 340±347. Litt, M., Kramer, P., Browne, D., Gancher, S., Brunt, E.R.P., Root, D., Phromchotikul, T., Dubay, C.J., Nutt, J.G., 1994. A gene for episodic ataxia/myokymia maps to chromosome 12p13. Am. J. Hum. Genet. 55, 702±709. Llinas, R., Sugimori, M., Hillman, D.E., Cherksey, B., 1992. Distribution and functional signi®cance of the P-type, voltage- dependent Ca 21 channels in the mammalian central nervous system. Trends Neurosci. 15 (9), 351±355. Lorenzon, N.M., Lutz, C.M., Frankel, W.N., Beam, K.G., 1998. Altered calcium channel currents in Purkinje cells of the neurological mutant mouse leaner.. J. Neurosci. 18 (12), 4482±4489. Ludwig, A., Flockerzi, V., Hofmann, F., 1997. Regional expression and cellular localization of the alpha1 and beta subunit of high voltage-activated calcium channels in rat brain.. J. Neurosci. 17 (4), 1339± 1349. Marchioni, E., Galimberti, C.A., Soragna, D., Ferrandi, D., Maurelli, M., Ratti, M.T., Bo, P., Montalbetti, L., Albergati, A., Savoldi, F., 1995. Familial hemiplegic migraine versus migraine with prolonged aura: an uncertain diagnosis in a family report. Neurology 45 (1), 33±37. Matsumura, R., Futamura, N., Fujimoto, Y., Yanagimoto, S., Horikawa, H., Suzumura, A., Takayanagi,

Migraine and ataxias

177

T., 1997. Spinocerebellar ataxia type 6. Molecular and clinical features of 35 Japanese patients including one homozygous for the CAG repeat expansion. Neurology 49 (5), 1238±1243. Matsuyama, Z., Kawakami, H., Maruyama, H., Izumi, Y., Komure, O., Udaka, F., Kameyama, M., Nishio, T., Kuroda, Y., Nishimura, M., Nakamura, S., 1997. Molecular features of the CAG repeats of spinocerebellar ataxia 6 (SCA6). Hum. Mol. Genet. 6 (8), 1283±1287. Moreno, H., Rudy, B., Llinas, R., 1997. beta subunits in¯uence the biophysical and pharmacological differences between P- and Q-type calcium currents expressed in a mammalian cell line. Proc. Natl. Acad. Sci. USA 94 (25), 14042±14047. MuÈnte, T.F., MuÈller-Vahl, H., 1990. Familial migraine coma: a case study.. J. Neurol. 237 (1), 59±61. Nagai, Y., Azuma, T., Funauchi, M., Fujita, M., Umi, M., Hirano, M., Matsubara, T., Ueno, S., 1998. Clinical and molecular genetic study in seven Japanese families with spinocerebellar ataxia type 6 [In Process Citation]. J. Neurol. Sci. 157 (1), 52±59. Ohta, M., Araki, S., Kuroiwa, Y., 1967. Familial occurence of migraine with an hemiplegic syndrome and cerebellar manifestations. Neurology 17, 813±817. Ophoff, R.A., van Eijk, R., Sandkuijl, L.A., Terwindt, G.M., Grubben, C.P., Haan, J., Lindhout, D., Ferrari, M.D., Frants, R.R., 1994. Genetic heterogeneity of familial hemiplegic migraine. Genomics 22 (1), 21±26. Ophoff, R.A., Terwindt, G.M., Vergouwe, M.N., van Eijk, R., Oefner, P.J., Hoffman, S.M., Lamerdin, J.E., Mohrenweiser, H.W., Bulman, D.E., Ferrari, M., Haan, J., Lindhout, D., van Ommen, G.J., Hofker, M.H., Ferrari, M.D., Frants, R.R., 1996. Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca 21 channel gene CACNL1A4. Cell 87 (3), 543±552. Parker, H.L., 1946. Periodic ataxia. Mayo Clin Proc 38, 642±645. Perez-Reyes, E., Schneider, T., 1995. Molecular biology of calcium channels. Kidney Int 48 (4), 1111± 1124. Rajput, V., Kramer, E.D., 1995. Adult onset familial hemiplegic migraine. Headache 35 (7), 423±427. Rhodes, K.J., Strassle, B.W., Monaghan, M.M., Bekele Arcuri, Z., Matos, M.F., Trimmer, J.S., 1997. Association and colocalization of the Kva1 and Kva2-subunits with Kv1 a-subunits in mammalian brain K 1 channel complexes. J. Neurosci. 17, 8246±8258. Sakurai, T., Hell, J.W., Woppmann, A., Miljanich, G.P., Catterall, W.A., 1995. Immunochemical identi®cation and differential phosphorylation of alternatively spliced forms of the alpha 1A subunit of brain calcium channels. J. Biol. Chem. 270 (36), 21234±21242. Sanguinetti, M.C., Spector, P.S., 1997. Potassium channelopathies. Neuropharmacology 36 (6), 755±762. Sasaki, H., Kojima, H., Yabe, I., Tashiro, K., Hamada, T., Sawa, H., Hiraga, H., Nagashima, K., 1998. Neuropathological and molecular studies of spinocerebellar ataxia type 6 (SCA6). Acta Neuropathol. (Berl.) 95 (2), 199±204. Satoh, J.I., Tokumoto, H., Yukitake, M., Matsui, M., Matsuyama, Z., Kawakami, H., Nakamura, S., Kuroda, Y., 1998. Spinocerebellar ataxia type 6: MRI of three Japanese patients. Neuroradiology 40 (4), 222±227. Scheffer, H., Brunt, E.R.P., Mol, G.J.J., van der Vlies, P., Stulp, R.P., Verlind, E., Mantel, G., Averyanov, Y.N., Hofstra, R.M.W., Buys, C.H.C.M., 1998. Three novel KCNA1 mutations in episodic ataxia type I families. Hum. Genet. 102, 464±466. Scott, V.E., Felix, R., Arikkath, J., Campbell, K.P., 1998. Evidence for a 95 kDa short form of the alpha1A subunit associated with the omega-conotoxin MVIIC receptor of the P/Q-type Ca 21 channels.. J. Neurosci. 18 (2), 641±647. Spranger, M., Spranger, S., Schwab, S., Benninger, C., Dichgans, M., 1999. Familial hemiplegic migraine with cerebellar ataxia and paroxysmal psychosis. Eur. Neurol. 41 (3), 150±152. Stea, A., Tomlinson, W.J., Soong, T.W., Bourinet, E., Dubel, S.J., Vincent, S.R., Snutch, T.P., 1994. Localization and functional properties of a rat brain alpha 1A calcium channel re¯ect similarities to neuronal Q- and P-type channels. Proc. Natl. Acad. Sci. USA 91 (22), 10576±10580.

178

A. Ducros et al.

Stevanin, G., Durr, A., David, G., Didierjean, O., Cancel, G., Rivaud, S., Tourbah, A., Warter, J.M., Agid, Y., Brice, A., 1997. Clinical and molecular features of spinocerebellar ataxia type 6 [see comments]. Neurology 49 (5), 1243±1246. Subramony, S.H., Fratkin, J.D., Manyam, B.V., Currier, R.D., 1996. Dominantly inherited cerebelloolivary atrophy is not due to a mutation at the spinocerebellar ataxia-I. Machado-Joseph disease, or dentato-rubro-pallido-luysian atrophy locus. Mov. Disord. 11, 174±180. Takahashi, H., Ikeuchi, T., Honma, Y., Hayashi, S., Tsuji, S., 1998. Autosomal dominant cerebellar ataxia (SCA6): clinical, genetic and neuropathological study in a family. Acta Neuropathol. (Berl) 95 (4), 333±337. Tean Teh, B.T., Silburn, P., Lindblad, K., Betz, R., Boyle, R., Schalling, M., Larsson, C., 1995. Familial periodic cerebellar ataxia without myokymia maps to a 19-cM region on 19p13. Am. J. Hum. Genet. 56 (6), 1443±1449. Terwindt, G.M., Ophoff, R.A., Haan, J., Frants, R.R., Ferrari, M.D., 1996. Familial hemiplegic migraine: a clinical comparison of families linked and unlinked to chromosome 19.DMG RG. Cephalalgia 16 (3), 153±155. Terwindt, G.M., Ophoff, R.A., Haan, J., Vergouwe, M.N., van Eijk, R., Frants, R.R., Ferrari, M.D., 1998. Variable clinical expression of mutations in the P/Q-type calcium channel gene in familial hemiplegic migraine. Dutch Migraine Genetics Research Group. Neurology 50 (4), 1105±1110. Vahedi, K., Joutel, A., Van Bogaert, P., Ducros, A., Maciazeck, J., Bach, J.F., Bousser, M.G., TournierLasserve, E., 1995. A gene for hereditary paroxysmal cerebellar ataxia maps to chromosome 19p. Ann. Neurol. 37 (3), 289±293. Vahedi, K., Denier, C., Ducros, A., Bousson, V., Levy, C., Haguenau, M., Tournier-Lasserve, E., Bousser, M.G., 1999. Sporadic hemiplegic migraine with a de novo CACNA1A missense mutation. Neurology 52 (Suppl2), A274. Van Bogaert, P., Van Nechel, C., Goldman, S., Szliwowski, H.B., 1993. Acetazolamide-responsive hereditary paroxysmal ataxia: report of a new family. Acta Neurol. Belg. 93 (5), 268±275. Van Dyke, D.H., Griggs, R.C., Murphy, M.J., Goldstein, M.N., 1975. Hereditary myokymia and periodic ataxia. J. Neurol. Sci. 25, 109±118. Varadi, G., Mori, Y., Mikala, G., Schwartz, A., 1995. Molecular determinants of Ca 21 channel function and drug action. Trends Pharmacol. Sci. 16, 43±49. Veh, R.W., Lichtinghagen, R., Sewing, G.S., Wunder, F., Grumbach, I.M., Pongs, O., 1995. Immunohistochemical localization of ®ve members of the Kv1 channel subunits: contrasting subcellular locations and neuron-speci®c co-localizations in rat brain. Eur. J. Neurosci. 7, 2189±2205. Vighetto, A., Froment, J., Trillet, M., Aimard, G., 1988. Magnetic resonance imaging in familial paroxysmal ataxia. Arch Neurol. 45, 547±549. Von Bredelow, B., Hahn, A.F., Koopman, W.J., Ebers, G.C., Bulman, D.E., 1995. Mapping the gene for acetozolamide responsive hereditary paroxysmal cerebellar ataxia to chromosome 19p. Hum. Mol. Genet. 4, 279±284. von Brederlow, B., Hahn, A.F., Koopman, W.J., Ebers, G.C., Bulman, D.E., 1995. Mapping the gene for acetazolamide responsive hereditary paryoxysmal cerebellar ataxia to chromosome 19p. Hum. Mol. Genet. 4 (2), 279±284. Westenbroek, R.E., Sakurai, T., Elliott, E.M., Hell, J.W., Starr, T.V., Snutch, T.P., Catterall, W.A., 1995. Immunochemical identi®cation and subcellular distribution of the alpha 1A subunits of brain calcium channels. J. Neurosci. 15 (10), 6403±6418. Whitty, C.W.M., 1953. Familial hemiplegic migraine. J. Neurol. Neurosurg. Psychiatr. 16, 172±177. Yamaguchi, H., Hara, M., Strobeck, M., Fukasawa, K., Schwartz, A., Varadi, G., 1998. Multiple modulation pathways of calcium channel activity by a beta subunit. Direct evidence of beta subunit participation in membrane traf®cking of the alpha1c subunit. J. Biol. Chem. 273 (30), 8±19356.

Migraine and ataxias

179

Young, G.F., Leon-Barth, C.A., Green, J., 1970. Familial hemiplegic migraine, retinal degeneration, deafness, and nystagmus. Arch. Neurol. 23 (3), 201±209. Yue, Q., Jen, J.C., Nelson, S.F., Baloh, R.W., 1997. Progressive ataxia due to a missense mutation in a calcium-channel gene. Am. J. Hum. Genet. 61 (5), 1078±1087. Yue, Q., Jen, J.C., Thwe, M.M., Nelson, S.F., Baloh, R.W., 1998. De novo mutation in CACNA1A caused acetazolamide-responsive episodic ataxia. Am. J. Med. Genet. 77 (4), 298±301. Zasorin, N.L., Baloh, R.W., Myers, L.B., 1983. Acetazolamide-responsive episodic ataxia syndrome. Neurology 33 (9), 1212±1214. Zerr, P., Adelman, J.P., Maylie, J., 1998a. Characterization of three episodic ataxia mutations in the human Kv1.1 potassium channel. FEBS Lett. 431, 461±464. Zerr, P., Adelman, J.P., Maylie, J., 1998b. Episodic ataxia mutations in Kv1.1 alter potassium channel function by dominant negative effects or haploinsuf®ciency. J. Neurosci. 18, 2842-2848. Zhang, J.F., Randall, A.D., Ellinor, P.T., Horne, W.A., Sather, W.A., Tanabe, T., Schwarz, T.L., Tsien, R.W., 1993. Distinctive pharmacology and kinetics of cloned neuronal Ca 21 channels and their possible counterparts in mammalian CNS neurons. Neuropharmacology 32 (11), 1075±1088. Zhuchenko, O., Bailey, J., Bonnen, P., Ashizawa, T., Stockton, D.W., Amos, C., Dobyns, W.B., Subramony, S.H., Zoghbi, H.Y., Lee, C.C., 1997. Autosomal dominant cerebellar ataxia (SCA6) associated with small polyglutamine expansions in the alpha 1A-voltage-dependent calcium channel. Nat. Genet. 15 (1), 62±69. Zifkin, B., Andermann, E., Andermann, F., Kirkham, T., 1980. An autosomal dominant syndrome of hemiplegic migraine, nystagmus, and tremor. Ann. Neurol. 8, 329±332. Zuberi, S.M., Eunson, L.H., Spauschus, A., De Silva, R., Tolmie, J., Wood, N.W., McWilliam, R.C., Stephenson, J.P., Kullmann, D.M., Hanna, M.G., 1999. A novel mutation in the human voltage-gated potassium channel gene (Kv1.1) associates with episodic ataxia type 1 and sometimes with partial epilepsy. Brain 122, 817±825.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

181

Chapter 7

Genetic analysis of idiopathic epilepsies: the role of ion channel mutations Ortrud K. Steinlein Institute for Human Genetics, University of Bonn, Wilhelmstrasse 31, 53111 Bonn, Germany

Abstract The ®rst gene defect underlying an idiopathic epilepsy was described in 1995: mutations in the a 4 subunit gene of the neuronal nicotinic acetylcholine receptor (CHRNA4) were found to lead to the syndrome of autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE). Recently, the gene defects for two more idiopathic epilepsies, benign familial neonatal convulsions (BFNC) and generalized epilepsy with febrile seizures plus (GEFS 1) have been found. They are also caused by mutations of ion channels: the voltage gated potassium channels KCNQ2 and KCNQ3 and the voltage gated sodium channel SCN1B, respectively. Although all three syndromes are rare monogenic disorders it seem to be plausible to assume that ion channel mutations play an important role in the aetiology of the common idiopathic epilepsies, too. q 2000 Elsevier Science B.V. All rights reserved.

Introduction More than 2000 years ago Hippokrates already assumed that idiopathic epilepsies had a strong genetic background. However, even today, our knowledge of the underlying genetic factors is still fragmentary. This is mainly due to the fact that a genetic aetiology in idiopathic epilepsies rarely means monogenetic inheritance. Twin studies demonstrate a high concordance rate in monozygotic pairs, while the rates for dizygotic twins are only slightly raised if compared with the co-occurrence of epilepsy in siblings (Conrad, 1935; Berkovic et al., 1990, 1998). These results show that the aetiology of idiopathic epilepsy is mainly genetic, but that in most syndromes the mode of inheritance is complex rather than monogenetic. Furthermore, recurrence rates are dropping markedly when the degree of consanguinity decreases. Thus, it is likely that the contribution of the underlying genes is multiplicative rather than additive (Zara et al., 1995). Most syndromes, especially the

182

O.K. Steinlein

common forms, like juvenile myoclonic epilepsy or childhood/juvenile absence epilepsy probably have an oligogenic or polygenic background, even if some rare families suggest a major gene effect. Furthermore, genetic studies are complicated by heterogeneity within a syndrome, as well as by overlapping genetic aetiology between different syndromes. Heterogeneity and aetiological overlap could be one explanation for the con¯icting results of recent linkage studies in common idiopathic epilepsies. For example, one of the ®rst chromosomal assignments was the localization of juvenile myoclonic epilepsy at chromosome 6p (Greenberg et al., 1988, Table 1). This assignment has been controversially discussed over the years (Durner et al., 1991; Weissbecker et al., 1991; Whitehouse et al., 1993), but so far no gene has been identi®ed. Only a few rare syndromes have been described as single gene disorders, and, unsurprisingly, genetic approaches were most successful here. To date gene defects underlying three different idiopathic epilepsies have been discovered: autosomal dominant nocturnal frontal lobe epilepsy (Steinlein et al., 1995a; 1997), benign familial neonatal convulsions (Biervert et al., 1998; Charlier et al., 1998; Singh et al., 1998) and generalized epilepsy with febrile seizures plus (Wallace et al., 1998, Table 2). In each of these syndromes the seizures are caused by ion channel mutations. Clinical features Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) ADNFLE is characterized by clusters of brief nocturnal motor seizures, which occur mostly during light sleep, either shortly after falling asleep or in the early Table 1 Gene localization in idiopathic epilepsy

Generalized epilepsy with febrile seizures plus Benign familial infantile convulsions Juvenile myoclonus epilepsy Febrile convulsions Febrile convulsions Autosomal dominant nocturnal frontal lobe epilepsy Idiopathic generalized epilepsy Juvenile myoclonic epilepsy Complex partial epilepsy Benign familial neonatal convulsions Benign familial neonatal convulsions Juvenile myoclonic epilepsy

19q 19q 15q 19p 8q 20q

Wallace et al. (1998) Guipponi et al. (1997) Elmslie et al. (1997) Dubovsky et al. 1996) Wallace et al. (1996) Phillips et al. (1995)

8q 6p 10q 8q 20q 6p

Zara et al. (1995) Liu et al. (1995) Ottman et al. (1995) Lewis et al. (1993) Leppert et al. (1989) Greenberg et al. (1988)

Ion channel mutations in epilepsy

183

Table 2 Gene identi®cation in idiopathic epilepsy

Syndrome

Chromosome Gene

Reference

Generalized epilepsy with febrile seizures plus (GEFS 1) Benign familial neonatal convulsions (BFNC) Benign familial neonatal convulsions Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE)

19q13.1

SCN1B

Wallace et al. (1998)

8q24

KCNQ3

Charlier et al. (1998)

20q13.3

KCNQ2

20q13.3

CHRNA4

Biervert et al. (1998); Singh et al. (1998) Steinlein et al. (1995a)

morning hours. The epilepsy shows considerable intrafamilial variation in age of onset as well as in severity. In about 50% of the patients the seizures start in the ®rst or second decade of life, but an age of onset later in life has also been observed. Seizures might be preceded by an aura, and can start with a gasp or grunt, or a vocalization. The motor features are described as thrashing hyperkinetic activity or tonic stiffening with superimposed clonic jerking. Secondary generalization with loss of consciousness can occur, but the majority of patients remain conscious through most of their seizures. Interictal EEG abnormalities are rare, and, for anatomical and technical reasons, ictal recordings are often showing non-speci®c or inconclusive patterns. Therefore nocturnal video-polysomnography is most helpful for differential diagnosis. ADNFLE is often misdiagnosed as benign nocturnal parasomnia, night terror, hyperactivity, or hysteria (Scheffer et al., 1994; Oldani et al., 1996; Hayman et al., 1997) Benign familial neonatal convulsions (BFNC) BFNC is a rare autosomal dominant idiopathic epilepsy. Multifocal or generalized tonic-clonic convulsions start around day 3 after birth and in most cases disappear spontaneously after a few weeks or months of life. The seizures last between seconds and several minutes, but status epilepticus has rarely been reported (Ronen et al., 1993; Wakai et al., 1994). Initially observed symptoms are usually tonic posture of the trunk and limbs, apnea, a shrill cry, ocular symptoms, or skin color change. Interictal EEG activity is usually normal, while ictal activity often starts with a general suppression of amplitude. Neurocognitive development is normal in most affected individuals. Learning disorders or mild cognitive impairment have only been reported in a minority of those affected (Ronen et al., 1993). However, 11± 15% of the BFNC patients will have epileptic seizures again later in life, which are mostly described as generalized tonic or tonic±clonic seizures with a variable age of

184

O.K. Steinlein

onset and duration. They often occur in response to provocative stimuli like auditory or emotional stress. Generalized epilepsy with febrile seizures plus (GEFS ^) GEFS 1 has only recently been described as a new epilepsy syndrome, and so far clinical descriptions are rare. The syndrome seems to be characterized by a wide range of different seizure types (Scheffer and Berkovic, 1997; Wallace et al., 1998). These include typical febrile convulsions, febrile seizures persisting beyond the age of 6 years (named seizures and not convulsions to differentiate them from typical febrile convulsions which normally do not occur after the age of 6 years), as well as different types of afebrile seizures (i.e. tonic±clonic, myoclonic, myoclonic±astatic, absence or atonic seizures). In most cases the course of the epilepsy seems to be benign and self-limiting, but at least one individual with myoclonic±astatic epilepsy leading to moderate intellectual disability has been described (Scheffer and Berkovic, 1997). However, it remains unclear if this was due to coincidence or if myoclonic±astatic epilepsy indeed belongs to the spectrum of GEFS 1. Genetics and pathogenesis Both ADNFLE and BFNC are inherited as autosomal dominant traits with incomplete penetrance (proportion of gene defect-carriers not showing the phenotype) (Rett and Teubel, 1964; Ryan et al., 1991; Scheffer et al., 1994; Steinlein et al., 1995a; Biervert et al., 1998). Thus, the mutation of a single gene is responsible for the disease in a given family. Each affected individual, male or female, transmits this defect gene to (statistically) 50% of his or her children. However, not every individual who inherits the defect gene develops the disease. The penetrance for both diseases, ADNFLE and BFNC, is approximately 70±80%. ADNFLE The ®rst gene locus for ADNFLE has been assigned to the genomic region 20q13.3 by performing linkage analysis in a large Australian pedigree with 27 affected individuals spanning six generations (Phillips et al., 1995). Previously, the a 4 subunit gene of the neuronal nicotinic acetylcholine receptor (CHRNA4) has been mapped to the same region (Steinlein et al., 1994, Fig. 1). By screening the Australian ADNFLE pedigree the ®rst mutation ever described for an idiopathic epilepsy was detected. A C ! T nucleotide transition led to the replacement of a neutral serine by a complex aromatic phenylalanine in the sixth amino acid position (Ser248Phe) of the second transmembrane domain of the CHRNA4 gene. All 21 affected members of the Australian family available for DNA analysis carried the

Ion channel mutations in epilepsy

185

Fig. 1. Localisation and transcriptional orientation of KCNQ2 and CHRNA4 on chromosome 20q13.3. D20S20 and D20S24, anonymous polymorphic markers.

mutation, as well as four obligate carriers and three individuals which were either pre-symptomatic or non-penetrant carriers (Steinlein et al., 1995a). CHRNA4 is one of the most abundant subunits of the neuronal nicotinic acetylcholine receptor (nAChR) found in mammalian brain (Whiting et al., 1991). nAChRs consist of heterologous pentamers comprising various combinations of approximately 11 subunits (a2±a9; b2±b4), which are differentially expressed throughout the brain and form physiologically and pharmacologically distinct receptors (Table 3). The a subunits have two adjacent cysteines (at positions 192 and 193 Table 3 Neuronal nicotinic acetylcholine receptor subunits

Alpha-type polypeptides

Beta-type polypeptides

Subunit

Gene symbol

Localisation

Subunit

Gene symbol

Localisation

a2 a3 a4 a5 a6 a7 a8 a9

CHRNA2 CHRNA3 CHRNA4 CHRNA5 CHRNA6 CHRNA7 CHRNA8 CHRNA9

8p21 15q24 20q13.3 15q24 Cloned, but not located 15q14 Not cloned in human Not cloned in human

b2 b3 b4

CHRNB2 CHRNB3 CHRNB4

1q21-q22 8 15q24

186

O.K. Steinlein

of the a1 muscle type subunit) and participate in the formation of agonist binding sites (Galzi and Changeux, 1995). The b subunits, which lack the cysteine pair are considered structural subunits, although they contribute towards the pharmacological speci®city of AChR subtypes (Luetje and Patrick, 1991). The second transmembrane domains of the ®ve subunits contribute to the walls of the ion channel (Fig. 2). According to the model proposed by Unwin (Unwin, 1993) Ser248 would be located directly beneath a tight hydrophobic ring at the most constricted part of the closed pore. Electrophysiological experiments in Xenopus oocytes revealed a profound effect of the mutation on the function of the receptor (Weiland et al., 1996). The mutant receptor exhibited an accelerated desensitization rate as well as a prolonged resensitization time, possibly indicating a destabilization of the ion channel open con®guration. Therefore it seems plausible that the Ser248Phe mutation causes the seizures in the ADNFLE patients from the Australian family by diminishing the activity of nAChRs containing the a4 subunit. The aetiological role of CHRNA4 in ADNFLE had been further supported by the observation of a second mutation in an Norwegian ADNFLE-family (Steinlein et al., 1997). Here ADNFLE is associated with the insertion of three additional nucleotides

Fig. 2. Model of a heteromeric nAChR with position of CHRNA4 mutations identi®ed in two families with autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE, Steinlein et al., 1995a, 1997).

Ion channel mutations in epilepsy

187

into the CHRNA4 gene. The insertion does not alter the reading frame, but leads to an additional leucine residue at the extracellular end of the second transmembrane domain (776ins3). Again, incomplete penetrance was demonstrated by the detection of the mutation in an obligate carrier in the Norwegian family. The mutation was not present in an control sample made up from 254 independent blood donors. Studies performed with this mutant showed that although it reconstitutes an functional channel if co-expressed with the b2 subunit, the receptor channel permeability for calcium is signi®cantly reduced (Steinlein et al., 1997). Thus, although acting by different mechanisms, the Ser248Phe as well as the 776ins3 mutation lead to hypoactivity of a4-containing neuronal nicotinic acetylcholine receptors by reducing the permeability for calcium and enhancing the desensitization sensitivity (Bertrand et al., 1998). However, several questions remain unanswered. For example, it is unclear how a CHRNA4 mutation could lead to a localization-related form of epilepsy. Mapping studies have shown that the CHRNA4 gene is expressed in nearly all brain tissues, making it one of the most abundant nAChR subunits. Thus, other unknown factors have to account for the frontal lobe origin of the epilepsy. One possibility might be a localisation-restricted co-assemblance with another subunit preferentially present in frontal cortex, or, contrary, the absence of a functionsubstituting subunit in this part of the brain. Furthermore, the subcellular localization of the nAChR is still unclear. There is evidence for both a postsynaptic as well as a presynaptic localization (De la Garza et al., 1987; SchroÈder et al., 1989; Wessler, 1992; Wonnacott et al., 1995; Wonnacott, 1997). A presynaptic localization would allow nAChR an important role in controlling and modulating the release of various neurotransmitters, which would explain that mutations likes those found in ADNFLE could lead to uncontrolled neuronal excitability and, subsequently, to epileptic seizures. Other ADNFLE families have failed to show linkage to chromosome 20q13.3, implicating genetic heterogeneity of this trait. With only two families found so far, CHRNA4 probably presents a minor locus for ADNFLE. Recent results of linkage studies found another family showing evidence for linkage to the CHRNA3/ CHRNA5/CHRNB4 cluster of nAChR subunits on chromosome 15q24. So far, no mutation has been identi®ed (Phillips et al., 1998). In the same report, linkage to 15q24 was excluded in other ADNFLE families, demonstrating genetic heterogeneity with at least three different genes for ADNFLE. BFNC In more than 90% of families BFNC is linked to chromosome 20q13.3 (Leppert et al., 1989; Malafosse et al., 1992). Only a few families have been reported to show linkage to chromosome 8q24 (Lewis et al., 1993; Steinlein et al., 1995b). It has been supposed that the genetic heterogeneity might be related to clinical heterogeneity

188

O.K. Steinlein

with chromosome 8q-linked families having a lesser risk of developing epilepsy later in life (Lewis et al., 1993). However, the numbers of chromosome 8q-linked families are far to low to allow such a prediction. The genes responsible for BFNC have recently been identi®ed: The major locus is a voltage-gated potassium channel gene (KCNQ2) on chromosome 20q (Biervert et al., 1998; Singh et al., 1998, Fig. 1). The highly homologous gene KCNQ3, showing 64% nucleotide homology in the open reading frame, codes for the minor locus on chromosome 8q (Charlier et al., 1998). Both voltage-gated potassium channels are highly brain-speci®c and expressed in nearly all regions of the central nervous system. Hybridization on Northern blots showed that at least three alternatively spliced transcripts (9.5, 3.8, 1.5 kb) of unknown importance are present in the brain (Biervert et al., 1998). Both KCNQ2 and KCNQ3 are showing signi®cant homologies (70% nucleotide homology in the coding region) to a previously cloned gene, KVLQT1, a potassium channel which is mutated in certain cases of the long QT syndrome (Wang et al., 1996). KVLQT1, KCNQ2 and KCNQ3 de®ne a new subfamily of the multiple member class of potassium channel genes, the ®rst one therefore is renamed as KCNQ1 (Table 4). KCNQ2 and KCNQ3 show the characteristic features of voltage-gated potassium channel genes (Fig. 3). The open reading frame of KCNQ2 encodes a potassium channel protein of 844 amino acids, with six transmembrane domains as well as a pore domain, followed by a long cytoplasmatic COOH-terminus. The S4 domain assembles a putative voltage sensor. Homologies between KCNQ2 and KVLQT1 are highest in the transmembrane domain regions as well as in short stretches of the COOH-terminus. Analysis of the genomic structure showed that KCNQ2 is coded by at least 18 exons ranging in size from 30 (exon 8 and 10) to 4.6 kb (exon 17). Exons 1±17 are distributed over approximately 50 kb of genomic DNA. Additional exons are probably located further upstream from exon 1 (Biervert and Steinlein, unpublished results). The ®rst fully characterized KCNQ2 mutation has been detected in an Australian BFNC family. A 5 bp insertion was found in a conserved 3 0 stretch of unknown function, downstream from the transmembrane domain regions (Biervert et al., 1998). Due to a frame shift, which led to a premature stop codon, KCNQ2 is shortened by more than 300 amino acids. Expression of the truncated protein in Table 4 Potassium channels: KCNQ subfamily

Name

Alias

Localisation

Disease

KCNQ1 KCNQ2 KCNQ2

KCNA9,KVLQT1,LQT1 KvEBN1 KvEBN2

11p15.5 20q13.3 8q24

LQT syndrome BFNC BFNC

Ion channel mutations in epilepsy

189

Fig. 3. Schematic representation of KCNQ2 mutations identi®ed in patients with benign familial neonatal convulsions (BFNC, Biervert et al., 1998; Singh et al., 1998; Steinlein et al., unpublished results). P, missense mutation; I/D, insertion/deletion; S, splice site mutation; 11, voltage sensor.

Xenopus oocytes did not lead to detectable currents, indicating that the mutation abolished channel function. If both the wild-type and the mutated protein were coexpressed at a 1:1 ratio currents were reduced when compared to those recorded from wild-type injections only. No dominant negative effect was observed, assuming that the seizures in BFNC patients are caused by haploinsuf®ciency of the KCNQ2 channel (Biervert et al., 1998). However, little is known about the stoichiometry and properties of this new potassium channel in vivo. For example, the KVLQT1 channel associates with a small b subunit (known as minK or IsK), which has a signi®cant effect on the electrophysiological properties of the channel (Barhanin et al., 1996). It is possible that a similar b subunit exists for KCNQ2, and that the binding and function of it would be affected by the mutation found in the Australian BFNC family. Furthermore, KCNQ2 might, together with other homologous subunits, build heteromeric receptors with in vivo properties different to that of the Xenopusmodel tested so far. It has already been shown that co-expression of KCNQ2 and KCNQ3 in Xenopus oocytes results in channels with properties distinct from either KCNQ2 or KCNQ3 channels alone (Yang et al., 1998). The heteromeric channels revealed a marked increase in current amplitude. If this model matches the situation in brain it could explain why mutations of both genes, KCNQ2 or KCNQ3, give rise to the same epileptic phenotype. Interestingly, the heteromeric KCNQ2/KCNQ3 channel shows biophysical properties and pharmacological sensitivities which are identical to what was formerly known to characterize the so-called M-current (Wang et al., 1996). The M-current is a well known potassium current showing a slow activating and deactivating conductance. Until recently nothing was known about the molecular identity of this current. With the identi®cation of the KCNQ2/KCNQ3 channel the proteins that make up the M-channel have probably been identi®ed. The M-current is expressed in peripheral

190

O.K. Steinlein

sympathetic neurons as well as in many different parts of the central nervous system. It reduces the neurons excitability by letting positively charged potassium channels ¯ow out while an action potential is built up. Thus the haploinsuf®ciency found to be caused by KCNQ2 mutations (Biervert et al., 1998; Schroeder et al., 1998) might cause BFNC by weakening the excitability-controlling function of the M-current. However, it remains unclear why the convulsions are restricted to the ®rst few weeks of life. One explanation might be that only the developing, immature brain is vulnerable to the mutation-caused disturbances of brain excitability. Another possibility, so far not investigated in human brain tissues, would be either a time-limited expression of KCNQ2 or KCNQ3, or an upregulation of another potassium channel some time after birth, which could substitute the loss of function caused by a BFNCmutation. Several other KCNQ2 mutations and one KCNQ3 mutation have been found in BFNC families. Most of them are either clustered in the sixth transmembrane domain, the pore domain, or in the COOH-terminal region (Charlier et al., 1998; Singh et al., 1998; Biervert and Steinlein, 1999). The disease can be caused by different types of mutations, point mutations are observed as well as small, frame shift causing deletions/insertions, splice site mutations or large deletions (Table 5). Some of the latter have been detected by observation of non-mendelian inheritance of the polymorphic marker D20S24. So far no correlation between the type of mutation and the severity of the phenotype has been found. None of these mutations, as far as they have been expressed in Xenopus oocytes, show a dominant-negative effect, regardless of whether they are studied in homomeric or heteromeric receptor models. With respect to the important function of the M-channel it is possible that a mutation causing such an effect would lead to a lethal phenotype. GEFS ^ The mode of inheritance of GEFS 1 seems to be less clear. This is probably due to the heterogeneous clinical picture. With several different types of seizures occurring in GEFS 1 families phenocopies are more likely to occur. For example, febrile convulsions affect about 3% of all children under the age of 6 years. Therefore in larger pedigrees a certain number of FC-individuals may present phenocopies. Furthermore, bilinear inheritance of epilepsy has been described in some families. Thus, a major gene effect on a background of complex inheritance seem to be responsible for GEFS 1. So far, linkage to chromosome 19q13.1 has been described only for one GEFS 1 family. A point mutation (C121W) in the b1 subunit (SCN1B) of the voltage-gated sodium channel was found to change a conserved cysteine residue in this family, probably disrupting a disul®de bridge between two conserved cysteine residues in the extracellular part of the protein (Wallace et al., 1998, Fig. 4). The mutation was found in all eight individuals matching the stringent criteria for

Ion channel mutations in epilepsy

191

Table 5 KCNQ2 mutations a

Mutation

Region

Function

Reference

1600ins5 1846delT 2513delG 1218 21G/C 1187 12T/G Large deletion Large deletion 283insGT Y284C A306T 494del13 516 21G/A

Exon 16 Exon 16 Exon 16 Intron 11 Intron 9 Exon 12±14? ? Exon 5 (pore) Exon 5 (pore) Exon 5 (S6) Exon 13 Intron 13

Reading frame shift Reading frame shift Reading frame shift Splice site mutation Splice site mutation ? ? Reading frame shift Missense mutation Missense mutation Reading frame shift Splice site mutation

Biervert et al. (1998) Biervert and Steinlein (1999) Lerche et al. (1999) Steinlein et al. (1999) Steinlein et al. unpublished Steinlein et al. unpublished Singh et al. (1998) Singh et al. (1998) Singh et al. (1998) Singh et al. (1998) Singh et al. (1998) Singh et al. (1998)

a

Exon and amino acid numbering according to Biervert and Steinlein (1999).

GEFS 1, but only in six of the twelve individuals with febrile seizures alone. No SCN1B mutations were found in 50 additional unrelated probands originating from

Fig. 4. Schematic representation of the SCN1B mutation identi®ed in a family with a history of generalized epilepsy with febrile seizures plus (GEFS 1, Wallace et al., 1998).

192

O.K. Steinlein

25 GEFS 1 families and isolated cases, demonstrating the heterogeneity of this syndrome. Voltage-gated sodium channels are activated by membrane depolarization, producing action potentials by a chain reaction of additional sodium channel activation by further depolarization. They consist of a large a subunit and one or two small b subunits. The a subunits show four repetitive regions, each containing six hydrophobic transmembrane segments. The a subunits can generate a functional channel when expressed alone in Xenopus oocytes, but association with b subunits modi®es channel function. b subunits are small proteins containing only a single transmembrane segment. So far, only two different b subunits have been characterized, b-1 and b-2 (Makita et al., 1994; Isom et al. 1995). The b-1 subunit SCN1B, which is mutated in GEFS 1, is probably stabilizing and thereby hastens the inactivation process as well as accelerating the recovery from inactivation of the channel by binding non-covalently to an a subunit (Krafte et al., 1988, 1990; Isom et al., 1992). The C121W mutation leads to a slowing of the inactivation process without showing any measurable effect on the recovery time (Wallace et al., 1998). This might result in an enhanced Na 1 in¯ux, causing a more depolarized membrane potential which could lead to hyperexcitability of the neuron.

Perspectives So far, all gene defects detected in idiopathic epilepsies are affecting ion channels: the a -nAChR subunit in ADNFLE, the voltage-gated potassium channels KCNQ2 and KCNQ3 in BFNC, and the voltage-gated sodium channel SCN1B in GEFS 1. Thus, ion channel mutations obviously play an important role in the pathogenesis of idiopathic epilepsies, con®rming the concept of ion channel diseases as paroxysmal disorders (Ptacek, 1997; Lehmann-Horn and Jurkat-Rott, 1999). Most of the so far identi®ed ion channel diseases, like hypereplexia, episodic ataxia/myokymiasyndrome, hyperkalemic periodic paralysis, hemiplegic migraine, and the long QT-syndrome, (Ptacek et al., 1991; Shiang et al., 1993; Browne et al., 1994; Curran et al., 1995; Ophoff et al., 1996; Wang et al., 1996), are characterized by episodic attacks in otherwise healthy individuals. Epilepsy, de®ned by chronically recurrent spontaneous or provoked seizures, ®ts well into this concept. Other monogenic or oligogenic idiopathic epilepsies, like the syndromes of benign familial infantile convulsions (BFIC) or benign epilepsy of childhood with centrotemporal spikes (rolandic epilepsy), for which loci have been assigned to chromosomes 19 and 15, respectively (Guipponi et al., 1997; Neubauer et al., 1998), are likely candidate diseases for ion channel mutations. However, different mechanisms seem to underlie symptomatic epilepsies, like progressive myoclonic epilepsies of the Unverricht±Lundborgh type or of the Lavora type (Pennacchio et

Ion channel mutations in epilepsy

193

al., 1996; Minassian et al., 1998). No ion channels were found to be involved in this subclass of epilepsies so far. For example, Unverricht±Lundborgh disease is caused by mutations of a cysteine protease inhibitor gene (cystatin B), while a protein tyrosine phosphatase (laforin) is responsible for Lafora disease. It will be interesting to learn about the pathomechanisms underlying borderline diseases like the newly described syndrome of BFIC with choreoathetosis (Szepetowski et al., 1997). Even if the success of the molecular approach in human idiopathic epilepsies so far is restricted to some rare monogenic syndromes, it is highly likely that comparable ®ndings will be made by analyzing the genetic mechanisms underlying the epilepsies with a non-monogenic mode of inheritance, like juvenile myoclonic epilepsy or the absence epilepsies of childhood and juvenile onset. Innumerable subtypes of ion channels are present in the mammalian brain, trying to maintain a sensitive balance in the electrical activity of the neuronal network. Theoretically, each of these ion channels can be a candidate target for epilepsy-causing mutations. Thus, the identi®cation of genes in diseases with a complex genetic background, including oligo- or polygenic inheritance and genetic heterogeneity, provides a great challenge at least for the next decade of research. Note added in proof A second gene coding for GEFS+ has been recently identi®ed on chromosome 2 (Escayg et al., 2000). Acknowledgements This work was supported by grants from the Deutsche Forschungsgemeinschaft (Ste769/1-1/1-2; Sonderforschungsbereich 400/B5). References Barhanin, J., Lesage, F., Guillemare, E., Fink, M., Lazdunski, M., Romey, G., 1996. KVLQT1 and IsK (minK) proteins associate to form the IKS cardiac potassium current. Nature 384, 78±80. Berkovic, S.F., Howell, R.A., Hopper, J.L., Hay, D.A., Andermann, E., 1990. A twin study of epilepsy. Epilepsia 31, 813. Berkovic, S.F., Howell, R.A., Hay, D.A., Hopper, J.L., 1998. Epilepsies in twins: genetics of the major epilepsy syndromes. Ann. Neurol. 43, 435±445. Bertrand, S., Weiland, S., Berkovic, S.F., Steinlein, O.K., Bertrand, D., 1998. Properties of neuronal nicotinic acetylcholine receptor mutants from humans suffering from autosomal dominant nocturnal frontal lobe epilepsy. Br. J. Pharmacol. 125, 751±760. Biervert, C., Steinlein, O.K., 1999. Structural and mutational analysis of KCNQ2, the major gene locus for benign familial neonatal convulsions. Hum. Genet. 104, 234±240.

194

O.K. Steinlein

Biervert, C., Schroeder, B.C., Kubisch, C., Berkovic, S.F., Propping, P., Jentsch, T.J., Steinlein, O.K., 1998. A potassium channel mutation in neonatal human epilepsy. Science 279, 403±406. Browne, D.L., Gancher, S.T., Nutt, J.G., Brunt, E.R., Smith, E.A., Kramer, P., Litt, M., 1994. Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA1. Nature Genet. 8, 136±140. Charlier, C., Singh, N.A., Ryan, S.G., Lewis, T.B., Reus, B.E., Leach, R.J., Leppert, M., 1998. A pore mutation in a novel KQT-like potassium channel gene in an idiopathic epilepsy family. Nature Genet. 18, 53±55. Conrad, K., 1935. Erbanlage und Epilepsie: Untersuchungen an einer Serie von 253 Zwillingspaaren. Z. Ges. Neurol. Psychiatry 153, 271±326. Curran, M.E., Splawski, I., Timothy, K.W., Vincent, G.M., Green, E.D., Keating, M.T., 1995. A molecular basis for cardiac arrhythmia: HERG mutations cause long QT syndrome. Cell 80, 795±803. de la Garza, R., Bickford-Wimer, P.C., Hoffer, B.J., Freedman, R., 1987. Heterogeneity of nicotine actions in the rat cerebellum: an in vivo electrophysiologic study. J. Pharmacol. Exp. Ther. 240, 689±695. Dubovsky, J., Weber, J.L., Orr, H.T., Rich, S.S., Gil-Nagel, A., Anderson, V.E., Leppik, I.E., Johnson, E.W., 1996. A second locus for febrile convulsions maps on chromosome 19p. Am. J. Med. Genet. 59, A223. Durner, M., Sander, T., Greenberg, D.A., Johnson, K., Beck-Mannagetta, G., Janz, D., 1991. Localization of idiopathic generalized epilepsies on chromosome 6p in families ascertained through juvenile myoclonic epilepsy patients. Neurology 41, 1651±1655. Elmslie, F.V., Rees, M., Williamson, M.P., Kerr, M., Kjeldsen, M.J., Pang, K.A., Sundqvist, A., Friis, M.L., Chadwick, D., Richens, A., Covanis, A., Santos, M., Arzimanoglou, A., Panayiotopoulos, C.P., Curtis, D., Whitehouse, W.P., Gardiner, R.M., 1997. Genetic mapping of a major susceptibility locus for juvenile myoclonic epilepsy on chromosome 15q. Hum. Mol. Genet. 6, 1329±1334. Escayg, A., MacDonald, B.T., Meisler, M.H., Baulac, S., Huberfeld, G., An-Gour®nkel, I., Brice, A., LeGuern, E., Moulard, B., Chaigne, D., Buresi, C., Malafosse, A., 2000. Mutations of SCN1A, encoding a neuronal sodium channel, in two families with GEFS+2. Nat. Genet. 24(4), 345. Galzi, J.L., Changeux, J.P., 1995. Neuronal nicotinic receptors: molecular organization and regulations. Neuropharmacology 34, 563±582. Greenberg, D.A., Delgado-Escueta, A.V., Widelitz, H., Sparkes, R.S., Treiman, L., Maldonado, H.M., Parks, M.S., Terasaki, P.I., 1988. Juvenile myoclonic epilepsy (JME) may be linked to the Bf and HLA loci on chromosome 6. Am. J. Med. Genet. 31, 185±192. Guipponi, M., Rivier, F., Vigevano, F., Beck, C., Crespel, A., Echenne, B., Lucchini, P., Sebastianelli, R., Baldy-Moulinier, M., Malafosse, A., 1997. Linkage mapping of benign familial infantile convulsions (BFIC) to chromosome 19q. Hum. Mol. Genet. 6, 473±477. Hayman, M., Scheffer, I.E., Chinvarun, Y., Berlangieri, S.U., Berkovic, S.F., 1997. Autosomal dominant nocturnal frontal lobe epilepsy: demonstration of focal frontal onset and intrafamilial variation. Neurology 49, 969±975. Isom, L.L., De Jongh, K.S., Patton, D.E., Reber, B.F., Offord, J., Charbonneau, H., Walsh, K., Goldin, A.L., Catterall, W.A., 1992. Primary structure and functional expression of the beta 1 subunit of the rat brain sodium channel. Science 256, 839±842. Isom, L.L., Ragsdale, D.S., De Jongh, K.S., Westenbroek, R.E., Reber, B.F.X., Scheuer, T., Catterall, W.A., 1995. Structure and function of the beta-2 subunit of brain sodium channels, a transmembrane glycoprotein with a CAM motif. Cell 83, 433±442. Krafte, D.S., Snutch, T.P., Leonard, J.P., Davidson, N., Lester, H.A., 1988. Evidence for the involvement of more than one mRNA species in controlling the inactivation process of rat and rabbit brain Na channels expressed in Xenopus oocytes. J. Neurosci. 8, 2859±2868.

Ion channel mutations in epilepsy

195

Krafte, D.S., Goldin, A.L., Auld, V.J., Dunn, R.J., Davidson, N., Lester, H.A., 1990. Inactivation of cloned Na channels expressed in Xenopus oocytes. J. Gen. Physiol. 96, 689±706. Lehmann-Horn, F., Jurkat-Rott, K., 1999. Voltage-gated ion channels and hereditary disease. Physiol. Rev. 79, 1317±1371. Leppert, M., Anderson, V.E., Quattlebaum, T., Stauffer, D., O'Conell, P., Lathrop, M., Lalouel, J.M., White, R., 1989. Benign neonatal convulsions linked to genetic markers on chromosome 20. Nature 337, 647±648. Lerche, H., Bievert, C., Alekov, A.K., Schleithoff, L., Linder, M., Klingler, W., Bretschneider, F., Mitrovic N., Jurkat-Rott, K., Lehmann-Horn, F., Steinlein, O.K., 1999. A reduced K + current due to a novel mutation in KCNQ2 causes neonatal convulsions. Ann. Neurol. 46, 305±312. Lewis, T.B., Leach, R.J., Ward, K., O'Connell, P., Ryan, S.G., 1993. Genetic heterogeneity in benign familial neonatal convulsions: identi®cation of a new locus on chromosome 8q. Am. J. Hum. Genet. 53, 670±675. Liu, A.W., Delgado-Escueta, A.V., Serratosa, J.M., Alonso, M.E., Medina, M.T., Gee, M.N., Cordova, S., Zhao, H.Z., Spellman, J.M., Peek, J.R., 1995. Juvenile myoclonic epilepsy locus in chromosome 6p21.2-p11: linkage to convulsions and electroencephalography trait. Am. J. Hum. Genet. 57, 368± 381. Luetje, C.W., Patrick, J., 1991. Both alpha and beta subunits contribute to the agonist sensitivity of neuronal nicotinic acetylcholine receptors. . J. Neurosci. 11, 837±845. Makita, N., Sloan-Brown, K., Weghuis, D.O., Ropers, H.H., George, A.L., 1994. Genomic organization and chromosomal assignment of the human voltage-gated Na(1) channel beta-1 subunit gene (SCN1B). Genomics 23, 628±634. Malafosse, A., Leboyer, M., Dulac, O., Navelet, Y., Plouin, P., Beck, C., Laklou, H., Mouchnino, G., Grandscene, P., Vallee, L., Guilloud-Bataille, M., Samolyk, D., Baldy-Moulinier, M., Feingold, J., Mallet, J., 1992. Con®rmation of linkage of benign familial neonatal convulsions to D20S19 and D20S20. Hum. Genet. 89, 54±58. Minassian, B.A., Lee, J.R., Herbrick, J.A., Huizenga, J., Soder, S., Mungall, A.J., Dunham, I., Gardner, R., Fong, C.Y., Carpenter, S., Jardim, L., Satishchandra, P., Andermann, E., Snead, O.C., Lopes-Cendes, I., Tsui, L.C., Delgado-Escueta, A.V., Rouleau, G.A., Scherer, S.W., 1998. Mutations in a gene encoding a novel protein tyrosine phosphatase cause progressive myoclonus epilepsy. Nat. Genet. 20, 171±174. Oldani, A., Zucconi, M., Ferini-Strambi, L., Bizzozero, D., Smirne, S., 1996. Autosomal dominant nocturnal frontal lobe epilepsy: electroclinical picture. Epilepsia 37, 964±976. Ophoff, R.A., Terwindt, G.M., Vergouwe, M.N., van Eijk, R., Oefner, P.J., Hoffman, S.M., Lamerdin, J.E., Mohrenweiser, H.W., Bulman, D.E., Ferrari, M., Haan, J., Lindhout, D., van Ommen, G.J., Hofker, M.H., Ferrari, M.D., Frants, R.R., 1996. Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca 21 channel gene CACNL1A4. Cell 87, 543±552. Ottman, R., Risch, N., Hauser, W.A., Pedley, T.A., Lee, J.H., Barker-Cummings, C., Lustenberger, A., Nagle, K.J., Lee, K.S., Scheuer, M.L., 1995. Localization of a gene for partial epilepsy to chromosome 10q. Nat. Genet. 10, 56±60. Pennacchio, L.A., Lehesjoki, A.E., Stone, N.E., Willour, V.L., Virtaneva, K., Miao, J., D'Amato, E., Ramirez, L., Faham, M., Koskiniemi, M., Warrington, J.A., Norio, R., de la Chapelle, A., Cox, D.R., Myers, R.M., 1996. Mutations in the gene encoding cystatin B in progressive myoclonus epilepsy. Science 271, 1731±1734. Ptacek, L.J., George, A.L., Griggs, R.C., Tawil, R., Kallen, R.G., Barchi, R.L., Robertson, M., Leppert, M.F., 1991. Identi®cation of a mutation in the gene causing hyperkalemic periodic paralysis. Cell 67, 1021±1027. Ptacek, L.J., 1997. Channelopathies: ion channel disorders of muscle as a paradigm for paroxysmal disorders of the nervous system. Neuromusc. Disord. 7, 250±255.

196

O.K. Steinlein

Phillips, H.A., Scheffer, I.E., Berkovic, S.F., Hollway, G.E., Sutherland, G.R., Mulley, J.C., 1995. Localization of a gene for autosomal dominant nocturnal frontal lobe epilepsy to chromosome 20q13.2. Nat. Genet. 10, 117±118. Phillips, H.A., Scheffer, I.E., Crossland, K.M., Bhatia, K.P., Fish, D.R., Marsden, C.D., Howell, S.J., Stephenson, J.B., Tolmie, J., Plazzi, G., Eeg-Olofsson, O., Singh, R., Lopes-Cendes, I., Andermann, E., Andermann, F., Berkovic, S.F., Mulley, J.C., 1998. Autosomal dominant nocturnal frontal-lobe epilepsy: genetic heterogeneity and evidence for a second locus at 15q24. Am. J. Hum. Genet. 63, 1108±1116. Rett, A., Teubel., R. (1964). NeugeborenenkraÈmpfe im Rahmen einer epileptische belasteten Familie. Wiener Klin. Wochenschrift 76, 609-613. Neubauer, B.A., Fiedler, B., Himmelein, B., Kampfer, F., Lassker, U., Schwabe, G., Spanier, I., Tams, D., Bretscher, C., Moldenhauer, K., Kurlemann, G., Weise, S., Tedroff, K., Eeg-Olofsson, O., Wadelius, C., Stephani, U., 1998. Centrotemporal spikes in families with rolandic epilepsy: linkage to chromosome 15q14. Neurology 51 (6), 1608±1612. Ronen, G.M., Rosales, T.O., Connolly, M., Anderson, V.E., Leppert, M., 1993. Seizure characteristics in chromosome 20 benign familial neonatal convulsions. Neurology 43, 1355±1360. Ryan, S.G., Wiznitzer, M., Hollman, C., Torres, M.C., Szekeresova, M., Schneider, S., 1991. Benign familial neonatal convulsions: evidence for clinical and genetic heterogeneity. Ann. Neurol. 29, 469± 473. Scheffer, I.E., Berkovic, S.F., 1997. Generalized epilepsy with febrile seizures plus. A genetic disorder with heterogeneous clinical phenotypes. Brain 120, 479±490. Scheffer, I.E., Bhatia, K.P., Lopes-Cendes, I., Fish, D.R., Marsden, C.D., Andermann, F., Andermann, E., Desbiens, R., Cendes, F., Manson, J.I., 1994. Autosomal dominant frontal epilepsy misdiagnosed as sleep disorder. Lancet 26, 515±517. SchroÈder, H., Zilles, K., Maelicke, A., Hajos, F., 1989. Immunohisto- and cytochemical localization of cortical nicotinic cholinoceptors in rat and man. Brain Res. 502, 287±295. Schroeder, B.C., Kubisch, C., Stein, V., Jentsch, T.J., 1998. Moderate loss of function of cyclic-AMPmodulated KCNQ2/KCNQ3 K 1 channels causes epilepsy. Nature 396, 687±690. Singh, N.A., Charlier, C., Stauffer, D., DuPont, B.R., Leach, R.J., Melis, R., Ronen, G.M., Bjerre, I., Quattlebaum, T., Murphy, J.V., McHarg, M.L., Gagnon, D., Rosales, T.O., Peiffer, A., Anderson, V.E., Leppert, M., 1998. A novel potassium channel gene. KCNQ2, is mutated in an inherited epilepsy of newborns. Nat. Genet. 18, 25±29. Shiang, R., Ryan, S.G., Zhu, Y.Z., Hahn, A.F., O'Connell, P., Wasmuth, J.J., 1993. Mutations in the alpha 1 subunit of the inhibitory glycine receptor cause the dominant neurologic disorder, hyperekplexia. Nat. Genet. 5, 351±358. Steinlein, O., Smigrodzki, R., Lindstrom, J., Anand, R., KoÈhler, M., Tocharoentanaphol, C., Vogel, F., 1994. Re®nement of the localization of the gene for neuronal nicotinic acetylcholine receptor a4 subunit (CHRNA4) to human chromosome 20q13.2-1-3.3. Genomics 22, 493±495. Steinlein, O., Mulley, J.C., Propping, P., Wallace, R.H., Phillips, H.A., Sutherland, G.R., Scheffer, I.E., Berkovic, S.F., 1995a. A missense mutation in the neuronal nicotinic acetylcholine receptor a4 subunit is associated with autosomal dominant nocturnal frontal lobe epilepsy. Nat. Genet. 11, 201±203. Steinlein, O., Schuster, V., Fischer, C., HaÈussler, M., 1995b. Benign familial neonatal convulsions: con®rmation of genetic heterogeneity and further evidence for a second locus on chromosome 8q. Hum. Genet. 95, 411±415. Steinlein, O., Magnusson, A., Stoodt, J., Bertrand, S., Weiland, S., Berkovic, S.F., Nakken, K.O., Propping, P., Bertrand, D., 1997. An insertion mutation of the CHRNA4 gene in a family with autosomal dominant nocturnal frontal lobe epilepsy. Hum. Mol. Genet. 6, 943±947. Szepetowski, P., Rochette, J., Berquin, P., Piussan, C., Lathrop, G.M., 1997. Monaco AP Familial

Ion channel mutations in epilepsy

197

infantile convulsions and paroxysmal choreoathetosis: a new neurological syndrome linked to the pericentromeric region of human chromosome 16. Am. J. Hum. Genet. 61, 889±898. Unwin, N., 1993. Nicotinic acetylcholine receptor at 9 A resolution. J. Mol. Biol. 229, 1101±1124. Wakai, S., Kamasaki, H., Itoh, N., Sueoka, H., Kawamoto, Y., Hayasaka, H., Tsutsumi, H., Chiba, S., 1994. Classi®cation of familial neonatal convulsions. Lancet 344, 1376. Wallace, R.H., Berkovic, S.F., Howell, R.A., Sutherland, G.R., Mulley, J.C., 1996. Suggestion of a major gene for familial febrile convulsions mapping to 8q13-21. J. Med. Genet. 33, 308±312. Wallace, R.H., Wang, D.W., Singh, R., Scheffer, I.E., George, A.L., Phillips, H.A., Saar, K., Reis, A., Johnson, E.W., Sutherland, G.R., Berkovic, S.F., Mulley, J.C., 1998. Febrile seizures and generalized epilepsy associated with a mutation in the Na 1-channel beta1 subunit gene SCN1B. Nat. Genet. 19, 366±370. Wang, Q., Curran, M.E., Splawski, I., Burn, T.C., Millholland, J.M., VanRaay, T.J., Shen, J., Timothy, K.W., Vincent, G.M., de Jager, T., Schwartz, P.J., Toubin, J.A., Moss, A.J., Atkinson, D.L., Landes, G.M., Connors, T.D., Keating, M.T., 1996. Positional cloning of a novel potassium channel gene: KVLQT1 mutations cause cardiac arrhythmias. Nat. Genet. 12, 17±23. Wessler, I., 1992. Acetylcholine at motor nerves: storage, release, and presynaptic modulation by autoreceptors and adrenoceptors. Int. Rev. Neurobiol. 34, 283±384. Weiland, S., Witzemann, V., Villarroel, A., Propping, P., Steinlein, O., 1996. An amino acid exchange in the second transmembrane segment of a neuronal nicotinic receptor causes partial epilepsy by altering its desensitization kinetics. FEBS Lett. 398, 91±96. Weissbecker, K.A., Durner, M., Janz, D., Scaramelli, A., Sparkes, M.A., Spence, M.A., 1991. Con®rmation of linkage between juvenile myoclonic epilepsy and the HLA region on chromosome 6. Am. J. Med. Genet. 38, 32±36. Whitehouse, W.P., Rees, M., Curtis, D., Sunquist, A., Parker, K., Chung, E., Baralle, D., Gardiner, R.M., 1993. Linkage analysis of idiopathic generalized epilepsy (IGE) and marker loci on chromosome 6p in families of patients with juvenile myoclonic epilepsy: No evidence for an epilepsy locus in the HLA region. Am. J. Hum. Genet. 53, 652±662. Whiting, P., Schoepfer, R., Lindstrom, J., Priestly, T., 1991. Structural and pharmacological characterization of the major brain nicotinc acetylcholine receptor subtype stably expressed in mouse ®broblasts. Mol. Pharmacol. 40, 463±472. Wonnacott, S., 1997. Presynaptic nicotinic ACh receptors. Trends Neurosci. 20, 92±98. Yang, W.P., Levesque, P.C., Little, W.A., Conder, M.L., Ramakrishnan, P., Neubauer, M.G., Blanar, M.A., 1998. Functional expression of two KvLQT1-related potassium channels responsible for an inherited idiopathic epilepsy. J. Biol. Chem. 273, 9±19423. Zara, F., Bianchi, A., Avanzini, G., Di Donato, S., Castellotti, B., Patel, P.I., Pandolfo, M., 1995. Mapping of genes predisposing to idiopathic generalized epilepsy. Hum. Mol. Genet. 4, 1201±1207.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

199

Chapter 8

The inhibitory glycine receptor as a model of hereditary channelopathies Kristina Becker, Cord-Michael Becker, Hans-Georg Breitinger Institut fuÈr Biochemie, UniversitaÈt Erlangen-NuÈrnberg, Fahrstrasse 17, D-91054 Erlangen, Germany

Abstract Inhibitory glycine receptors mediate synaptic transmission in the mammalian spinal cord and brain stem. They constitute a family of ligand gated chloride channels encoded by homologous subunit genes. Gene structure and localization, as well as tissue distribution and developmental regulation of the receptor have been studied extensively. Likewise, there is a large and increasing amount of information on glycine receptor protein structure and ion channel function, and a picture of the functional architecture of this receptor and its associated proteins begins to form. Lack or functional impairment of spinal glycine receptors have been shown to underlie the hereditary human motor disorder hyperekplexia (startle disease, stiff baby syndrome, STHE, OMIM # 138491). In many cases, the disease could be traced to single point mutations of the GLRA1 gene. Yet despite a wealth of detailed knowledge, a speci®c and selective modulation of glycinergic function is not available. This review summarizes new insights into the structure function relationship of glycine receptors with focus on mutations causing hypertonic motor disorders in humans and animals. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Motor control requires a highly structured and well-organized system of both excitation and inhibition. In the spinal cord, glycine receptors predominate as mediators of fast inhibitory signal transmission. On a physiological level, glycinergic function in the spinal cord is best studied in recurrent and reciprocal feedback loops which modulate the ®ring rate of motoneurons (Eccles, 1964). The presence of glycine receptors in this system has been known for a long time (Curtis et al., 1968a,b). Due to the advances in molecular cloning techniques, the molecular components within these spinal inhibitory circuits have been subject of intense

200

K. Becker et al.

investigation, resulting in characterization of receptor function in normal and pathological states. The glycine receptor belongs among the ligand-gated ion channel superfamily and shares considerable sequence homology as well as common structural elements with the other members of this receptor family, such as the nicotinic acetylcholine receptor, the 5-HT3 receptor, and g-aminobutyric acid (GABAA) and GABAC receptors. The discovery of strychnine as a selective high-af®nity ligand for the inhibitory glycine receptor has greatly facilitated receptor investigation. In fact, the presence of high-af®nity strychnine binding is often equated with the presence of glycine receptors. Accordingly, the symptoms of defective glycinergic inhibition, whether congenital or as a result of trauma, are similar to those of mild strychnine poisoning. Based on this clinical resemblance, glycine receptor dysfunction has long been considered a candidate mechanism of hypertonic motor disorders in humans and animals (Becker, 1990, 1995). The human disease hyperekplexia (startle disease, stiff baby syndrome) is a congenital motor disorder that displays dominant and recessive modes of inheritance. Affected patients exhibit an exaggerated startle response and increased muscle tone ± initial description in: Suhren et al. (1966). It should be noted that the disease most critically affects children during their ®rst years of life due to continuous muscle stiffness. Following extensive genomic mapping, Shiang et al. (1993) were the ®rst to identify a missense mutation in the GLRA1 gene encoding the a1-subunit of the receptor in patients with startle disease. To date, seven different mutations associated with startle disease have been described in patients and their families ± reviewed by Becker (1995). In this article, we summarize ®ndings regarding glycine receptor genetics, protein structure and function, and its role in human and murine motor disorders.

Glycine receptor genetics Analysis of protein and gene structure of individual ligand-gated ion channels has revealed a remarkable degree of homology between neurotransmitter receptors of entirely different pharmacology on one hand and also for the `same' receptor in different species on the other hand. These ®ndings have led to the hypothesis that the nicotinic acetylcholine receptor-type superfamily of ligand-gated receptors evolved from common ancestral genes (Le Novere and Changeux, 1995). The best understood of these receptors is the acetylcholine receptor (AChR), due to the fact that it is present in large amounts in the electric organs of ®sh such as electric eel (Electrophorus electricus) or ray (Torpedo californica). In fact, most of the structural and biochemical knowledge to date comes from studies using these tissues, thus promoting the nicotinic AChR to model status for the rest of this receptor family. The common features (Fig. 1) of these ligand-gated ion channels include:

The inhibitory glycine receptor as a model of hereditary channelopathies

201

Fig. 1. Ligand-gated ion channels: (A) subunit topology, (B) scheme of pentameric receptor.

² pentameric receptors, made up of ®ve individual, non-covalently linked subunits; ² size of each subunit is ca. 50 kDa; ² common topology of subunits: large N-terminus, four transmembrane (TM) segments; ² TM2 of each subunit forms the inner lining of the channel pore; ² a ring of charged amino acids at the extracellular `end' of the pore conveys ion selectivity onto the receptor.

The inhibitory glycine receptor is a typical member of this ligand-gated receptor superfamily with respect to protein size, topology and elementary steps of function. To date, four different a-subunits (a1±4) and one b-subunit have been identi®ed (Table 1). The a-subunits are assumed to contain the ligand binding pocket, while the bsubunit is considered to be responsible for attaching to the tubulin-binding protein gephyrin, thereby anchoring the receptor at the synapse and control clustering of receptors (Feng et al., 1998). Expression of a1-subunits in Xenopus oocytes (Schmieden et al., 1989) or in human embryonic kidney cells (Sontheimer et al., 1989) results in formation of functional homo-oligomeric channels, while b-subunits alone do not form functional channels in these systems (Grenningloh et al., 1990a). Co-expression of a- and b-subunits yields functional glycine receptors with a dramatically diminished sensitivity to the plant alkaloid picrotoxinin (Pribilla et al., 1992). Eight individual amino acid residues within the N-terminal domain of the glycine receptor a1 subunit were shown to be crucial for homo-oligomeric receptor formation. Oligomerization and glycosylation were critical for intracellular trans-

202

K. Becker et al.

Table 1 Glycine receptor subunits ± gene loci

Subunit

Chromosomal locus

Reference

Human

Mouse

a1

5q31.2

11 (29 cM)

a2 a3 a4 b

Xp21.2±p21.3 4q33±q34 (?Xq21±q22?) 4q31.3

X (70 cM) 8 (25 cM) X (56 cM) 3 (36 cM)

Shiang et al. (1993); Matzenbach et al. (1994) Grenningloh et al. (1990b) Nicolic et al. (1998) Matzenbach et al. (1994) Handford et al. (1996); Kingsmore et al. (1994); Milani et al. (1996); MuÈlhardt et al. (1994)

port while subunit surface contacts and protein conformation determine ion channel assembly (Griffon et al., 1999). All glycine receptor subunits exhibit a high degree of homology for the same subunit across species. Finally, b-subunits differ markedly from their a counterparts, as shown in Table 2. Homology data were compiled from sequences published in the SwissProt database. Gene structures are highly conserved between individual glycine receptor subunits. Glycine receptor genes consist of nine exons and complexity is added by the Table 2 Glycine receptor subunits ± protein sequence homologies a

Subunit

Human a1

Human a2

Human a3

Human b

Human a1 Human a2 Human a3 Human b Mouse a1 Mouse a4 Mouse b Rat a1 Rat a2 Rat a3 Rat b

(100)* 85 87 64 98* 98 64 98* 85 84 64

78 (100)* 83 46 78 91 47 78 98* 77 47

82 85 (100)* 49 82 92 49 82 85 99* 51

57 59 59 (100)* 47 60 97* 47 49 49 97*

a Homology data were compiled from sequences published in the SwissProt database. Corresponding subunits in each column, as well as b-subunits, are indicated by an asterisk (*).

The inhibitory glycine receptor as a model of hereditary channelopathies

203

presence of alternative splice sites. These were identi®ed in glycine receptor a1-, a2and a3-subunits. The number of receptor subunits identi®ed on the nucleic acid level exceeds the functional diversity that has been observed in native receptors. Notwithstanding this lack of information ± which essentially is due to the considerable dif®culty of identifying and studying different receptor subtypes in tissue ± it can be assumed that the variety of different subunit compositions, plus the use of alternative splicing, allows an organism to ®ne-tune its glycinergic function (Fig. 2). It is of interest to note that identi®ed alternative splice sites in the glycine receptor are located in exon 8 (a3) (Nicolic et al., 1998) and exon 9 (a1) (Malosio et al., 1991; Matzenbach et al., 1994), leading to proteins that differ in sequence in the cytoplasmic loop after TM 3. In the case of the a2-subunit, an alternative splice acceptor site is located in exon 3, leading to transcripts 3a and 3b (Kuhse et al., 1991). In rodents, the subunit composition of glycine receptors in the spinal cord is under strict developmental control (Becker et al., 1988; Malosio et al., 1991; Takahashi et al., 1992). The neonatal isoform GlyRN, which is considered an a2-homomer, prevails in juvenile animals. Within 2±3 weeks after birth, it is replaced by the adult isoform GlyRA, consisting of a1- and b-subunits (Becker et al., 1988). A change in the kinetics of spinal inhibitory postsynaptic currents (IPSCs) is temporally correlated with this subunit switch (Takahashi et al., 1992). IPSCs recorded in slices of spinal cord become shorter during postnatal development and glycine receptors in excised patches from spinal motor neurons exhibit a signi®cantly shorter mean open time in more mature animals. Thus, the developmental regulation of glycine receptors bears similarity to the changes in nicotinic acetylcholine receptors at the endplate, caused by switching from the g- to the 1-subunit (Mishina et al., 1986). Whereas factors governing developmental regulation of muscle AChRs are well characterized to date ± reviewed in Fischbach and Rosen (1997) and Sanes and Lichtman (1999) ± very little is known about regulation of glycine receptors. This is in part due to the dif®culty in establishing a suitable system for assaying candidate factors, since in primary neuronal cultures from spinal cord, the switch from GlyRN

Fig. 2. Glycine receptor subunit genes: structure and alternative splicing.

204

K. Becker et al.

to GlyRA does not occur using standard culturing protocols (Hoch et al., 1989). That activity-dependent mechanisms may be essential for the regulation of receptor expression is further suggested by the ®nding that blockade of NMDA receptors by the non-competitive antagonist MK-801 results in appearance of GlyRA in spinal cord cultures (Hoch et al., 1992). There are functional observations indicating that in some cases gephyrin may be needed for synaptic clustering (Kirsch et al., 1993). Synaptic activity may also be essential for synaptic clustering of glycine receptors (Kirsch and Betz, 1998). The presence of functional glycine receptors was also shown to be crucial for synaptic clustering (Levi et al., 1998). Repression of receptor expression in non-neuronal tissues is an important aspect of receptor regulation. The neuronal gene repressor RE1 (repressor element 1) acts by binding to speci®c DNA motifs, thereby effectively suppressing transcription (Thiel et al., 1998). A neuron restrictive silencer element (NRSE) was found in the upstream UTRs of neuronal genes, including synapsin I, SCG10, nicotinic acetylcholine receptor b2-, and muscarinic acetylcholine receptor m4- subunits, and also for the glycine receptor a1-subunit. It remains to be seen, if expression of individual glycine receptor subunits is under transcriptional control, or if it is mainly orchestrated during receptor translation, assembly and transport.

The glycine receptor protein The observation about the location of the alternative splice sites is of importance, since it supports the general idea of subdividing glycine receptor subunits into structural domains. It is dif®cult to subdivide a protein into several quasi-independent units, since it must be expected to behave as one unit; several examples, most prominently sickle cell anemia, show that in fact the exchange of one amino acid residue can lead to dramatically altered function of the entire protein. In contrast, many examples in the literature demonstrate, that in other cases amino acid exchanges are tolerated and do not seem to affect the properties or function of the protein in question. Most notably alanine (Lynch et al., 1997), and cysteine substitution scanning ± SCAM, Karlin and Akabas (1998) have given very useful information concerning ion channel function. Numerous studies regarding TM2 of ligand gated ion channels, where SCAM-investigation of the channel pore was carried out, suggested that this TM segment is helical, these ®ndings being in good agreement with other structural (Unwin, 1995, 1996) and functional (Bormann et al., 1987,1993; Lynch et al., 1997) observations. In close analogy to the nicotinic AChR, the generally accepted topology of a glycine receptor subunit shows structurally distinct regions (Fig. 1). Independently, studies of receptor function focussing on ligand binding, channel opening and clos-

The inhibitory glycine receptor as a model of hereditary channelopathies

205

Fig. 3. Glycine receptor subunit ± functional domains.

ing etc., revealed that the receptor can also be subdivided into functionally different elements, as will be discussed below. The functions of the glycine receptor domains and their physical realization are described in the following sections and Fig. 3 and Table 3. Table 3 Functional and structural domains of the glycine receptor a-subunit

Function

Structural representation

Comments

Ligand binding

N-terminal region 1 TM2± TM3-loop N-terminal region 1 TM2± TM3-loop Loop TM1±2, Loop TM2±3 TM2 Loop TM3±4

N-terminal region immediately preceding TM1 Three-loop-model of channel opening `Hinges' for TM2 movement Inner lining of ion channel Location of alternative splice sites, consensus sites for phosphorylation

Signal transduction Gating Pore characteristics Functional modi®cation

206

K. Becker et al.

Ligand binding The large N-terminal domain is the primary site of interaction with exogenous ligands. Photoincorporation of [ 3H]strychnine into rat spinal cord membrane preparations, followed by fractionation, showed that strychnine and glycine bind speci®cally to the a1-subunit of the glycine receptor, and in this case, binding was competitive (Graham et al., 1983; Betz et al., 1989). From limited proteolysis, where the ®rst 100 amino acids of the glycine receptor were left, the ligand binding site was suggested to be located near the ®rst transmembrane segment (Graham et al., 1981; Grenningloh et al., 1987; Betz, 1990). Similarly, [ 3H]strychnine was incorporated into peptide fragments from this region of the glycine receptor (Ruiz-Gomez et al., 1990). In further re®nements, individual amino acid residues that are in direct contact with the ligands or at least are involved in shaping the binding pocket, were identi®ed by chemical modi®cation of amino acid residues with chemical modifying agents (Young and Snyder, 1974a,b; Marvizon et al., 1986; RuizGomez et al., 1989; O'Connor et al., 1996; Breitinger and Becker, 1998) or by site-directed mutagenesis (Betz, 1990; Kuhse et al., 1990b). For instance, an individual glutamic acid residue (position 167) was found to be critical for strychnine sensitivity of glycine receptors (Kuhse et al., 1990b). Based upon the considerable number of binding data, attempts were made to model the ligand binding site (Aprison et al., 1995; Galvez-Ruano et al., 1995), and to determine thermodynamic parameters of ligand binding (Ruiz-Gomez et al., 1989). While some studies report a competitive displacement of strychnine by glycine (Gundlach and Beart, 1981; Fry and Phelan, 1985; White, 1985; Becker et al., 1986; O'Connor, 1989) as well as competitive inhibition of glycine gated currents by strychnine (Sontheimer et al., 1989), it was also reported that modi®cation of glycine binding by introducing a point mutation at the glycine binding site also altered glycine-induced displacement of strychnine (Marvizon et al., 1986). Inhibition of glycine-gated currents in embryonic rat medullary neurons by strychnine showed hallmarks of both competitive and allosteric inhibition (Lewis et al., 1989). The data available so far suggest adjacent or even overlapping, but not completely identical binding sites for glycine and strychnine (Aprison et al., 1987; Grenningloh et al., 1987; O'Connor, 1989). For pharmacological information the reader is also directed to reviews treating this subject in somewhat greater depth (Becker, 1992; Becker and Langosch, 1997; Breitinger and Becker, 1998). Signal transduction Studies using photoaf®nity labeling (White and Cohen, 1992), crosslinking and mutagenesis (Czajkowski et al., 1993; Fu and Sine, 1994; Karlin and Akabas, 1998) suggested for the nicotinic acetylcholine receptor that the binding site is located at

The inhibitory glycine receptor as a model of hereditary channelopathies

207

the interface between adjacent subunits (Smith and Olsen, 1995). GABAA receptors also show sequence similarities and functional properties similar to the other ligandgated ion channels. Smith and Olsen (1995) showed, that the 4-loop model suggested for these channels (Betz, 1990), also applies to the GABAA receptor. A common element is a Cys-Cys-loop of 13 amino acids in length, another loop from the N-terminal region of the same subunit, and a third loop which comes from the adjacent subunit (Betz, 1990). Recent mutagenesis studies identi®ed the amino acid residues I93, A101, and N102 in the N-terminal domain to be crucial for ligandbinding, but not for signal transduction (Vafa et al., 1999). In the same region of the protein, residues K104, F108, and T112 were shown to be involved in agonist binding (Schmieden et al., 1999). This part of the protein also holds a site that regulates receptor modulation by Zn 2+ (Harvey et al., 1999). Ligand-binding and the transduction of this activating event to channel-opening are consecutive processes that should be seen in connection. For the glycine receptor, it is generally found that more than one glycine molecule is required for receptor activation. Mostly deduced from Hill coef®cients, which are generally found to range between 1.7 and 3.0, or from mechanistic studies, results for native receptors from acutely dissociated or cultured neurons (Barker et al., 1983; Werman et al., 1986; Lewis et al., 1989; Tokutomi et al., 1989; Walstrom and Hess, 1994), as well as for recombinant receptors in Xenopus oocytes (Schmieden et al., 1989; Grenningloh et al., 1990b; Kuhse et al., 1990a,b), or in HEK 293 cells (Sontheimer et al., 1989; Grewer, 1999) show that binding of two or even three molecules of glycine is needed for receptor activation. At the same time, the partial agonists b-alanine and taurine lack cooperativity (Schmieden et al., 1989), and similarly, glycine displacement of the antagonist [ 3H]strychnine also had a Hill coef®cient of 1 (Gundlach and Beart, 1981; Fry and Phelan, 1985; White, 1985; Becker et al., 1986; O'Connor, 1989), as was found for strychnine inhibition (Sontheimer et al., 1989). On the other hand, [ 3H]strychnine was reported to have a Hill coef®cient .1 (Young and Snyder, 1974b). It should be borne in mind, that such data often do not allow unambiguous assignment of one mechanism and should, therefore, be treated with care. Channel gating Except for ligand binding (see above) and the structure of the ion pore of the channel (below), recent studies suggest the short loops ¯anking TM2 to be a critical determinant of channel gating (Lynch et al., 1997; Saul et al., 1999). Several hyperekplexia mutations were found in this region of the protein, leading to altered channel function, while the ligand-binding steps were much less affected (see below). For the intracellular TM1-2-loop, as in the mutant P250T (Saul et al., 1999), it was shown that hydrophobicity and bulk appears to be the crucial property governing channel function and conductance without dramatically affecting gating

208

K. Becker et al.

behavior (Breitinger et al., in preparation). Thus, the observed functional differences between mutant and wild type channels could, again, be attributed to modi®cations in well-de®ned structural elements, showing that the processes of ligand-binding and channel gating occur in spatially and functionally separate domains of the receptor protein. Analyzing glycine receptor function, one must always be aware of the interplay of several elementary reaction steps (ligand-binding, channel-opening, channel closing, receptor desensitization), that occur on different, yet often overlapping time scales. A kinetic analysis of glycine receptor currents in rat ventral cochlear nucleus revealed the existence of hitherto unknown rapidly desensitizing current phases (Harty and Manis, 1998). These experiments were carried out using a fast ¯ow system, and are in agreement with ®ndings of fast desensitizing chloride currents in embryonic spinal cord neurons in an earlier study (Walstrom and Hess, 1994). The laser-pulse photolysis method was used to analyze the functional mechanism of recombinant glycine receptor a1 homomers. Here, agonist binding was found to be slower than previously assumed, i.e. not controlled by diffusion. The individual steps of channel-opening and closing were determined in this extensive study to be 2500 s 21 and 300 s 21, respectively (Grewer, 1999). Characteristics of the glycine receptor ion pore For the ion pore of the acetylcholine receptor, mutation studies (Hille, 1992) had shown that the TM2-segment is responsible for the properties of the ion channel, and that in fact TM2 forms the lining of the ion pore (Stroud et al., 1990; Smith and Olsen, 1995). Structural studies, again on the muscle-type acetylcholine receptor (Unwin, 1995; 1996), suggest a helical arrangement of this segment. Scanning cysteine modi®cation of the TM2 segment (Karlin and Akabas, 1998) also suggested a helical structure for TM2. Modelling the electrostatic environment and ionization state of the ring of charged amino acids at the mouth of the ion pore led to the suggestion that not only the TM2 helix, but also extramembrane domains of the receptor may be crucial determinants for ionic selectivity (Adcock et al., 1998). On murine spinal cord neurons, single-channel conductance states of 10, 18, and 29 pS were found (Bormann et al., 1987; Hamill et al., 1983). Even in homooligomeric receptors, several conductance states were reported (see Table 4). Kinetic analysis of desensitization suggested the presence of multiple, independent receptor subtypes present in embryonic mouse spinal cord neurons (Walstrom and Hess, 1994). This diversity can also be expected to exist in other neuronal tissues, particularly in view of the pronounced developmental regulation of glycine receptor subunits and the possibility of post-translational receptor modi®cation mentioned above. As is evident from conductance states, incorporation of the b-subunit does not necessarily generate entirely different pore properties, but rather shifts the equili-

The inhibitory glycine receptor as a model of hereditary channelopathies

209

Table 4 Conductance states of recombinant glycine receptors (after Bormann et al., 1993)

Subunit(s)

a1 a2 a3 a1/b a2/b a3/b

Conductance states (single channel conductances in pS) I

II

III

IV

V

VI

± 111 105 ± 112 ±

86 91 85 ± 80 ±

64 66 62 ± ± ±

46 48 42 44 54 48

30 36 30 29 36 34

18 23 20 20 ± 23

brium of population of the conductance states that are available. A similar redistribution of the population of conductance states was found for the 5-HT3 receptor (Van Hooft and Vijverberg, 1995), where receptor phosphorylation and cellular environment were the determining factors. Channel closing rate constants of glycine receptors in zebra®sh were shown to be voltage-dependent (Legendre, 1999), although mammalian glycine receptors are generally found to have linear current± voltage relationships over the physiological range of transmembrane potentials. Not surprisingly, properties of the ion pore are signi®cantly affected by the ionic environment of the receptor. Physiologically, the important permeant ions are chloride and bicarbonate (Backus et al., 1998). Permeant ions include the halogenides and pseudohalogenides. A detailed patch-clamp study identi®ed two anion-binding sites within the channel pore (Bormann et al., 1987). Di- and polyvalent cations such as Zn 21, Pb 21, and La 31 do not traverse the channel, but modulate glycine receptor responses (Laube et al., 1995; Kumamoto and Murata, 1996). Zinc potentiation of the glycine receptor was shown to be following an allosteric mechanism (Lynch et al., 1998), while the biphasic modulation of glycine receptors by Zn 21 (,2 mM Zn 21 ˆ potentiation, .10 mM Zn 21 ˆ inhibition) was demonstrated in retinal ganglion cells (Han and Wu, 1999); the amino acid residues responsible for Zn 21 interaction, namey His 107 and His 108 were recently identi®ed (Harvey et al., 1999). When ionic effects are studied, the counterions must also be considered. Ammonium as counterion of permeant ions suppresses strychnine binding to the receptor, while in the presence of sodium strychnine does bind (Young and Snyder, 1974a; MuÈller and Snyder, 1978; Marvizon et al., 1986). Strychnine binding to the receptor was also affected by anions, here a concentration-dependent biphasic regulation was observed (Marvizon et al., 1986; Marvizon and Skolnick, 1988). For a more detailed review see Breitinger and Becker (1998). Finally, Cs 1 itself can activate the glycine receptor, and this activation can be antagonized by strychnine (Lewis et al., 1989). Chemical modi®cation suggested different sites of modulatory

210

K. Becker et al.

action of Na 1 and NH41 on the receptor (Marvizon et al., 1986), yet it seems that potential activation sites may also exist. Functional modi®cation Other important elements in the regulation of receptor function are sites for posttranslational modi®cation. The identi®cation of alternative splice sites in glycine receptor a1 ( Malosio et al., 1991; Matzenbach et al., 1994) and a3-subunits (Nicolic et al., 1998) showed, that stretches within the TM3-4-loop that contained consensus phosphorylation sites were affected by alternative splicing. The splice variants were found to have different desensitization properties. Altered desensitization (and gating) due to receptor phosphorylation was already observed for nicotinic acetylcholine (Huganir et al., 1986; Huganir and Greengard, 1990). In case of 5-HT3 receptors (Van Hooft and Vijverberg, 1995), conductance states were differently populated as a result of phosphorylation and dephosphorylation. Similarly, glycine receptors showed altered functionality upon phosphorylation by PKA and PKC (Ruiz-Gomez et al., 1991; Vaello et al., 1994). The availability for modi®cation also depended on receptor activity: liganded receptors were preferentially phosphorylated (Vaello et al., 1992). It was recently shown that the cAMPdependent protein kinase, PKA, but not PKC, did modulate glycine-activated chloride currents in rat VTA neurons. Thus it was suggested that the crucial parameter affected by phosphorylation is glycine af®nity (Ren et al., 1998). Phosphorylation was reported to modulate ethanol sensitivity of recombinant a1 glycine receptors expressed in Xenopus oocytes (Mascia et al., 1998). Modulation of glycine receptor function by ethanol and other aliphatic alcohols was shown to depend on residue 267, whereby potentiation or inhibition of glycinergic responses correlate with molecular volume at this position (Mascia et al., 1996a,b; Wick et al., 1998; Ye et al., 1998). Other, non-speci®c environmental effects that were reported to affect glycine receptor function include pressure ± studied in recombinant receptors expressed in Xenopus oocytes (Roberts et al., 1996), as well as in animal models of pressure-induced pain (Koltchine et al., 1996). Density of receptors expressed in Xenopus oocytes also appear to have an effect on receptor function (Maammar et al., 1997). Finally, subunit composition is an important regulatory element for the function of any ion channel. Presence of b-subunits confers picrotoxinin-insensitivity on recombinant glycine receptors (Pribilla et al., 1992), and the presence of the b-subunit also induces moderate inward recti®cation of the voltage-dependence of single-channel conductances. However, maximum currents and dose-response behavior of glycine-mediated currents are not signi®cantly affected by the absence or presence of b-subunits.

The inhibitory glycine receptor as a model of hereditary channelopathies

211

Positioning of transmembrane segments A ®nal structural feature common to all glycine a and some GABAA receptor subunits is the termination of each TM segment with an arginine or lysine residue. Apparently, the charge on the amino acid ensures proper insertion and orientation of the TM regions. If these arginines are missing, the correct translocation of the receptor into the plasma membrane is affected, thus leading to dramatic loss of the number of receptors (Langosch et al., 1993; Kling et al., in preparation). In synthetic peptides which only contained the TM2-part of the glycine receptor, it was found necessary to add a series of four lysines to the C-terminal end of the peptide to ensure proper insertion into membranes. Upon incorporation into membrane bilayers, these peptides did form functional channels for chloride and ¯uids; chloride ¯uxes through these `minimal' channels were even blocked by strychnine (Langosch et al., 1993; Wallace et al., 1997). Glycine receptor defects Hyperekplexia Mutant alleles of the glycine receptor a- and b-subunit genes have been found to underlie hypertonic motor disorders in humans and in mice. Symptoms of the hereditary human disorder hyperekplexia (startle disease, stiff baby syndrome, STHE, OMIM # 138491) include an exaggerated startle response to unexpected acoustic and tactile stimuli as well as episodic muscle stiffness. Affected neonates usually exhibit more severe symptoms with profound startle responses and muscular hypertonia, sometimes resulting in fatal apneic attacks. Muscle tone normalizes during childhood, whereas excessive startling persists throughout life. Even in adulthood potentially life-threatening situations can still be encountered by hyperekplexia patients, since the startle response can trigger a loss of postural control with immediate and uncontrolled falling as a result (Andermann et al., 1980). The clinical picture, however, often varies considerably even within families with more than one af¯icted member, thereby imposing dif®culties in establishing a correct clinical diagnosis. Despite the problems regarding diagnostic criteria, the genomic locus for the disease was assigned by linkage analysis using microsatellite markers to chromosome 5q by Ryan et al. (1992) in a large, ®ve generation family with startle disease. Since this region contains a number of interesting neurological genes (Ryan et al., 1992), the demonstration that the GLRA1 gene also maps to this region (Baker et al., 1994; Warrington, 1994), in combination with the close clinical resemblance of hyperekplexia and subconvulsive strychnine poisoning, made the glycine receptor a1- subunit gene a perfect candidate for hyperekplexia. Shiang et al. (1993) were the ®rst to identify missense mutations in the GLRA1 gene in four different families with

212

K. Becker et al.

hereditary hyperekplexia, all occurring in the same base pair of exon 6 and resulting in the substitution of an uncharged amino acid (leucine or glutamine) for the highly conserved Arg 271, which delineates the extracellular border of TM2 in the mature protein. Thus, GLRA1 was the ®rst gene for a neurotransmitter receptor in the central nervous system to be identi®ed as the site of mutation in a human disorder. In the course of characterizing this mutation, it was found that this arginine residue is a critical determinant for the occurrence of higher conductance states (see above), which are absent in heterologously expressed channels containing either one of the two hyperekplexia-substitutions (Langosch et al., 1994; Rajendra et al., 1995). The af®nity of mutant receptors for glycine is severely reduced compared to wild-type channels (230- and 410-fold for the R271L and R271Q substitution, respectively), while desensitization kinetics are not altered. A number of additional missense mutations have been found in recessive and dominant forms of hyperekplexia, all of which result in single amino acid substitutions and are localized in a region of the receptor ranging from the intracellular loop immediately after TM1 to the extracellular loop following TM2 (Table 5, Fig. 4). Similar to the ®rst mutation to be identi®ed in hyperekplexia patients, R271L/Q, four other mutations are localized to the loops connecting either TM1 and TM2 (I244D, P250T) or TM2 and TM 3 (K276E, Y279C). In the Unwin model of the nAChR (Unwin, 1995; 1996), these loops have been postulated to act as `hinges', enabling TM2 to be rotated sideways upon activation of the channel (see above). Thus, these mutations are expected to affect channel gating. In a detailed study of these residues, Lynch et al. (1997) ± all, except P250T, see above ± demonstrated that, indeed, these mutations disrupt the long range allosteric transduction mechanTable 5 Mutant alleles of the human glycine receptor a1-subunit gene

Allelic variant

Phenotype

Mode of inheritance

References

Ser 231 Arg Ile 244 Asn Pro 250 Thr Gln 266 His Arg 271 Leu Arg 271 Gln Lys 276 Glu

Hyperekplexia Hyperekplexia Hyperekplexia Hyperekplexia Hyperekplexia Hyperekplexia Hyperekplexia spastic paraparesis Hyperekplexia Hyperekplexia

Recessive Recessive Dominant Dominant Dominant Dominant Dominant

Reuter et al. (unpublished) Rees et al. (1994) Saul et al. (1999) Milani et al. (1996) Shiang et al. (1993) Shiang et al. (1993) Elmslie et al. (1996)

Dominant Recessive

Shiang et al. (1993) Brune et al. (1996) Becker et al. (unpublished)

Tyr 279 Cys GLRA1 null

The inhibitory glycine receptor as a model of hereditary channelopathies

213

Fig. 4. Locations of hereditary mutations of the glycine receptor a1-subunit.

ism coupling ligand binding to ion channel gating. In all cases, glycine sensitivity and glycine evoked whole-cell currents were profoundly decreased. In addition, I244D mutant channels differed, such that expression ef®cacy was found to be severely reduced. Whether reduced expression is a contributing factor in vivo remains to be determined. The Q266H mutation also is characterized by a reduced af®nity of glycine for the receptor as indicated by a 6-fold shift in the concentrationresponse curve (Moorhouse et al., 1999). A defect in channel gating is causal for the functional effect of this mutation as well, as there is a large decrease in single channel open time without accompanying changes in ligand-binding parameters. Even though the substitution is located in close proximity to the channel pore, ion selectivity and permeation are not affected. This apparent discrepancy to the domain model could be reconciled by assuming that the mutated amino acid does not extend into the lumen of the ion pore (Moorhouse et al., 1999). Interestingly, mutations clustered around TM2 exhibit a dominant mode of inheritance, whereas mutations outside of this segment are recessive. In two cases of recessive hyperekplexia, functional null alleles of the GLRA1 gene were identi®ed,

214

K. Becker et al.

characterized by a deletion of exon 1-6 in one (Brune et al., 1996) and by a deletion of exons 1±7 in the other case (K. Becker, unpublished observation). Both children are born to consanguineous parents and are homozygous for the GLRA1 null allele, consistent with a complete loss of gene function. Despite the complete lack of mature glycine receptors, both children exhibit relatively mild symptoms including a disinhibition of vestigial brain stem re¯exes, i.e. exaggerated startle responses and head retraction jerks (Brune et al., 1996); as one of the children already has reached early adolescence, it appears that the loss of the glycine receptor a1-subunit is tolerated in humans much better compared to mice harboring an analogous genetic defect. (Mutant mice: oscillator, spasmodic, spastic). Whereas in patients with startle disease only mutations in GLRA1 have been linked to the disease, the situation is more complex in phenotypic mice, where both subunits constituting the mature spinal glycine receptor are affected. All identi®ed murine glycine receptor mutations show a recessive mode of inheritance. A summary of the different mutations encountered in murine glycine receptor subunit genes is given in Table 6. In the mutant mouse oscillator (spd ot), a microdeletion of 7 bp within exon 8 of the glycine receptor a1-subunit gene causes a translational frameshift (Buckwalter et al., 1994). Depending on alternative exon usage, the Glra1 spd-ot allele encodes two mutant transcripts: in the predominant a1 mRNA species, the site of the microdeletion corresponding to TM3 is followed by 127 missense amino acids, or in the alternate transcript, the polypeptide chain is terminated after 19 residues due to an premature stop codon (Buckwalter et al., 1994). In either case, the mutation results in loss of the second cytoplasmic loop and the TM4 domain of the polypeptide. In spinal cord of homozygous animals, GlyRA protein is completely absent, indicating that these domains are of crucial importance for receptor assembly and function (Kling et al., 1997). In contrast to the human null-allele of GLRA1, the phenotype in mice is lethal at 3 weeks of age, and even heterozygous animals exhibit exaggerated acoustic startle responses in behavioral testing paradigms (Kling et al., 1997). The glycine receptor a1-subunit gene is also the site of mutation in spasmodic (spd) mice, where an amino acid substitution in the N-terminal domain at position 52 (Ala 52 Ser) leads to reduced af®nity of the receptors to glycine, taurine and b-alanine, while strychnine binding af®nity is unchanged (Ryan et al., 1994; Saul et al., 1994). As glycine receptor content in spinal cord and native molecular weight are not altered, this suggests that the spasmodic phenotype results from an reduced neurotransmitter sensitivity of the mutant a1A52S subunit. The phenotype of homozygous mice is less severe compared to spd ot/spd ot mice, since the defect is not lethal. But similar to the ®ndings in oscillator mice, the response to acoustic stimuli is markedly enhanced in homozygous as well as in heterozygous animals (Plappert et al., in preparation). Why do different mutations within the same gene do not give rise to a similar phenotype? As the mutation in spasmodic mice does not exert its effects

Subunit

Gene

Gyndrome

Mode of inheritance

Mutation

Reference

a1

Glra1

Spasmodic (spd)

Recessive

Missense mutation (A52S)

a1

Glra1

Oscillator (spd ot)

Recessive

b1

Glrb

Spastic (spa)

Recessive

Functional null allele with microdeletion, causing translational frameshift Intronic LINE-1 insertion causing splicing defect

Ryan et al. (1994); Saul et al. (1994) Buckwalter et al. (1994); Kling et al. (1997) Kingsmore et al. (1994); MuÈlhardt et al. (1994)

The inhibitory glycine receptor as a model of hereditary channelopathies

Table 6 Glycine receptor mutations of the mouse

215

216

K. Becker et al.

on the level of GlyRA expression, but leads to a change in pharmacological function, it has to be assumed, that although glycinergic inhibition is reduced, the mutant receptors still contribute to inhibition in the spinal cord. In contrast, when GlyRAs are completely absent, as it is the case in homozygous oscillator mice, other inhibitory systems, such as GABAA/GABAB-mediated inhibition, are apparently not suf®cient to compensate for this defect. This hypothesis is challenged, however, by the ®nding of a functional null allele in two patients with hyperekplexia (Brune et al., 1996, K. Becker et al., unpublished; see above). Recently, compound heterozygosity for two point mutations in the GlyR a1 gene was shown to be responsible for a hyperekplexia phenotype. Two different mutant alleles encoding for the amino acid exchanges R252H and R392H were identi®ed in non-consanguineous parents. Hyperekplexia was only diagnosed in two of four children who were heterozygous for both alleles. Individuals carrying only one mutation showed no clinical symptoms (Vergouwe et al., 1999). Idiopathic hyperekplexia was observed in a patient who showed no glycine receptor defects but suffered from a subtle brainstem damage (Gambardella et al., 1999), indicating that systemic factors may also contribute to individual occurence or severity of hyperekplexia symptoms. Spastic mice (spa) carry a mutation in the glycine receptor b-subunit, which in homozygous animals causes a dramatic reduction of GlyRA expression within the spinal (Becker et al., 1986, 1992; Becker, 1990, 1995) and in other CNS areas, as the glycine receptor b-subunit is abundantly expressed ( Malosio et al., 1991; MuÈlhardt et al., 1994). In two independent studies, Kingsmore et al. (1994); MuÈlhardt et al. (1994) demonstrated a chromosomal re-arrangement of a region on chromosome 3, caused by an insertion of a full-length LINE-1 (or L1) element into intron 5 of the Glrb-gene. L1 elements are mobile elements which comprise about 15±20% of the genome. In contrast to mobile elements which retrotranspose autonomously using retroviral-like mechanisms, these `master mobile elements' retrotranspose via targetprimed reverse transcription ± reviewed in Kazazian (1998). In the case of the spastic mouse, L1 insertion interferes with the correct splicing of b-subunit pre-mRNA, resulting in exon skipping (Kingsmore et al., 1994; MuÈlhardt et al., 1994). The predominant Glyrb mRNA variants in homozygous mice are truncated either due to skipping of exon 5 or of both exon 4 and exon 5, leading to non-functional proteins. As the skipping of exons is incomplete, however, some full-length mRNA is still produced, resulting in low level expression of functional GlyRA in mutants (MuÈlhardt et al., 1994). The subunit composition of the remaining receptors is not affected, comprising both a1- and b-subunits in addition to the postsynaptic anchoring protein, gephyrin (Becker et al., 1986). Thus, the crucial role of the b-subunit for glycine receptor assembly and function in vivo is highlighted by the fact that a1-subunits alone ef®ciently assemble and function in vitro (Schmieden et al., 1989; Sontheimer et al., 1989). For regular synaptic function, the clustering of receptors to the postsynaptic membrane via cytoskeletal anchors might be necessary, which would be

The inhibitory glycine receptor as a model of hereditary channelopathies

217

accomplished by the glycine receptor b-subunit via its gephyrin binding domain. Consistent with the a numerical receptor defect in spastic, the phenotype can be rescued by a transgene expressing b-subunit mRNA (Hartenstein et al., 1996). Phenotypically, adult spa/spa mice are indistinguishable from homozygous spasmodic mice, but, as in hyperekplexia patients carrying the same mutation, substantial phenotypic variability can be observed. This is most likely due to variations in the genetic background, as it becomes apparent that mutations can result in very different phenotypes depending on the genetic backgrounds (Abeliovich et al., 1993; Smithies and Maeda, 1995; Rozmahel et al., 1996; Silva et al., 1997). In the case of the mutant mouse spastic, the phenotype varies from being lethal at three weeks of age (resembling spd ot/spd ot mice) in spa/spa C57Bl/6J mice to normal life span exhibiting only mild symptoms in hybrid spa/spa B6C3Fe mice (Becker et al., 1992; Becker and Becker, in preparation). Conclusions and perspectives The analysis of the inhibitory glycine receptor has provided valuable tools for the study of a class of hypertonic motor disorders in humans and mice. The discovery of novel mutations in patients with startle disease permitted new insights into structure and function of glycine receptor channels, and will continue to do so as new mutations are being discovered. Patients with startle disease, who do not carry mutations in the a1-subunit gene, are obvious candidates for a broader screening procedure, including other glycine receptor subunit genes in addition to candidates such as gephyrin. To date, a human hereditary disorder syntenic to the spastic locus of mice has not been identi®ed. Since the murine models offer the possibility of detailed analysis of the effects of glycinergic dysfunction in vivo, they contribute to our knowledge of glycine receptors in CNS areas other than spinal cord. With growing insight into the function of glycine receptors in cortex or subcortical structures such as the cerebellum, it will become increasingly possible to identify the molecular basis of other human neurological disorders. References Abeliovich, A., Paylor, R., Chen, C., Kim, J.J., Wehner, J.M., Tonegawa, S., 1993. PKC gamma mutant mice exhibit mild de®cits in spatial and contextual learning. Cell 75, 1263±1271. Adcock, C., Smith, G.R., Sansom, M.S., 1998. Electrostatics and the ion selectivity of ligand-gated channels. Biophys. J. 75, 1211±1222. Andermann, F., Keene, D.L., Andermann, E., Quesney, L.F., 1980. Startle disease or hyperekplexia: Further delineation of the syndrome. Brain 103, 985±997. Aprison, M.H., Lipkowitz, K.B., Simon, J.R., 1987. Identi®cation of a glycine-like fragment on the strychnine molecule. J. Neurosci. Res. 17, 209±213.

218

K. Becker et al.

Aprison, M.H., Galvez, R.E., Lipkowitz, K.B., 1995. On a molecular comparison of strong and weak antagonists at the glycinergic receptor. J. Neurosci. Res. 41, 259±269. Backus, K.H., Deitmer, J.W., Friauf, E., 1998. Glycine-activated currents are changed by coincident membrane depolarization in developing rat auditory brainstem neurones. J. Physiol. 507, 783±794. Baker, E., Sutherland, G.R., Scho®eld, P.R., 1994. Localization of the glycine receptor alpha 1 subunit gene (GLRA1) to chromosome 5q32 by FISH. Genomics 22, 491±493. Barker, J.L., McBurney, R.N., Mathers, D.A., 1983. Convulsant-induced depression of amino acid responses in cultured mouse spinal neirone studies under voltage clamp. Br. J. Pharmacol. 80, 619± 629. Becker, C.M., 1990. Disorders of the inhibitory glycine receptor: the spastic mouse. Faseb J 4, 2767± 2774. Becker, C.-M., 1992. Selective neurotoxicity: convulsants acting at the inhibitory glycine receptor. In: Herken, H., Hucho, F.F. (Eds.), Handbook of Experimental Pharmacology. Springer, Heidelberg, pp. 539±575. Becker, C.-M., 1995. Glycine receptors: molecular heterogeneity and implications for disease. Neuroscientist 1, 130±141. Becker, C.-M., Langosch, D., 1997. The inhibitory glycine receptor. In: Stephenson, F.A., Turner, A.J. (Eds.), Amino Acid Neurotransmission. Portland, London, pp. 93±112. Becker, C.-M., Hermans-Borgmeyer, I., Schmitt, B., Betz, H., 1986. The glycine receptor de®ciency of the mutant mouse spastic: evidence for normal glycine receptor structure and localization. J. Neurosci. 6, 1358±1364. Becker, C.M., Hoch, W., Betz, H., 1988. Glycine receptor heterogeneity in rat spinal cord during postnatal development. EMBO J. 7, 3717±3726. Becker, C.M., Schmieden, V., Tarroni, P., Strasser, U., Betz, H., 1992. Isoform-selective de®cit of glycine receptors in the mouse mutant spastic. Neuron 8, 283±289. Betz, H., 1990. Ligand-gated ion channels in the brain: the amino acid receptor family. Neuron 5, 383± 392. Betz, H., Becker, C.M., Grenningloh, G., Hoch, W., Knaus, P., Langosch, D., Malosio, M.L., Schmitt, B., Thomas, L., 1989. Homology and analogy in transmembrane channel design: lessons from synaptic membrane proteins. J. Protein Chem. 8, 325. Bormann, J., Hamill, O.P., Sakmann, B., 1987. Mechanism of anion permeation through channels gated by glycine and g-aminobutyric acid in mouse cultured spinal cord neurons. J. Physiol. (Lond) 385, 243±286. Bormann, J., Rundstrom, N., Betz, H., Langosch, D., 1993. Residues within transmembrane segment M2 determine chloride conductance of glycine receptor homo- and hetero-oligomers [published erratum appears in EMBO J. 1994 Mar 15;13(6):1493]. EMBO J. 12, 3729±3737. Breitinger, H.G., Becker, C.M., 1998. The inhibitory glycine receptor: prospects for a therapeutic orphan?. Curr. Pharm. Des. 4, 315±334. Brune, W., Weber, R.G., Saul, B., von Knebel Doeberitz, M., Grond Ginsbach, C., Kellerman, K., Meinck, H.M., Becker, C.M., 1996. A GLRA1 null mutation in recessive hyperekplexia challenges the functional role of glycine receptors. Am. J. Hum. Genet. 58, 989±997. Buckwalter, M.S., Cook, S.A., Davisson, M.T., White, W.F., Camper, S.A., 1994. A frameshift mutation in the mouse alpha 1 glycine receptor gene (Glra1) results in progressive neurological symptoms and juvenile death. Hum. Mol. Genet. 3, 2025±2030. Curtis, D.R., HoÈsli, L., Johnston, G.A.R., 1968a. A pharmacological study of the depression of spinal neurones by glycine and related amino acids. Exp. Brain Res. 6, 1±18. Curtis, D.R., HoÈsli, L., Johnston, G.A.R., Johnston, I.H., 1968b. The hyperpolarization of spinal motoneurones by glycine and related amino acids. Exp. Brain Res. 5, 235±258. Czajkowski, C., Kaufmann, C., Karlin, A., 1993. Negatively charged amino acid residues in the nicotinic

The inhibitory glycine receptor as a model of hereditary channelopathies

219

receptor delta subunit that contribute to the binding of acetylcholine. Proc. Natl. Acad. Sci. USA 90, 6285±6289. Eccles, J.C., 1964. The Ionic Mechanism of Synaptic Inhibition. The Nobel Foundation, Stockholm. Elmslie, F.V., Hutchings, S.M., Spencer, V., Curtis, A., Covanis, T., Gardiner, R.M. and Rees, M., 1996. Analysis of GLRA1 in hereditary and sporadic hyperekplexia: a novel mutation in a family cosegregating for hyperekplexia and spastic paraparesis. J. Med. Genet. 33, 435±436. Feng, G., Tintrup, H., Kirsch, J., Nichol, M.C., Kuhse, J., Betz, H., Sanes, J.R., 1998. Dual requirement for gephyrin in glycine receptor clustering and molybdoenzyme activity [see comments]. Science 282, 1321±1324. Fischbach, G.D., Rosen, K.M., 1997. ARIA: a neuromuscular junction neuregulin. [Review]. Annu. Rev. Neurosci. 20, 429±458. Fry, J.P., Phelan, P.P., 1985. Interaction of glycine receptor ligands in the normal mouse spinal cord. J. Physiol. 373, 21±43. Fu, D.X., Sine, S.M., 1994. Competitive antagonists bridge the alpha-gamma subunit interface of the acetylcholine receptor through quaternary ammonium-aromatic interactions. J. Biol. Chem. 269, 26152±26157. Galvez-Ruano, E., Lipkowitz, K.B., Aprison, M.H., 1995. On identifying a second molecular antagonistic mechanism operative at the glycine receptor. J. Neurosci. Res. 41, 775±781. Gambardella, A., Valentino, P., Annesi, G., Oliveri, R.L., Bono, F., Mazzei, R.L., Conforti, F.L., Aguglia, U., Zappia, M., Pardatscher, K., Quattrone, A., 1999. Hyperekplexia in a patient with a brainstem vascular anomaly. Acta Neurol. Scand. 99, 255±259. Graham, D., Pfeiffer, F., Betz, H., 1981. UV light-induced cross-linking of strychnine to the glycine receptor of rat spinal cord membranes. Biochem. Biophys. Res. Commun. 102, 1330±1335. Graham, D., Pfeiffer, F., Betz, H., 1983. Photoaf®nity-labelling of the glycine receptor of rat spinal cord. Eur. J. Biochem. 131, 519±525. Grenningloh, G., Rienitz, A., Schmitt, B., Methfessel, C., Zensen, M., Beyreuther, K., Gundel®nger, E.D., Betz, H., 1987. The strychnine-binding subunit of the glycine receptor shows homology with nicotinic acetylcholine receptors. Nature 328, 215±220. Grenningloh, G., Pribilla, I., Prior, P., Multhaup, G., Beyreuther, K., Taleb, O., Betz, H., 1990a. Cloning and expression of the 58 kd beta subunit of the inhibitory glycine receptor. Neuron 4, 963±970. Grenningloh, G., Schmieden, V., Scho®eld, P.R., Seeburg, P.H., Siddique, T., Mohandas, T.K., Becker, C.M., Betz, H., 1990b. Alpha subunit variants of the human glycine receptor: primary structures, functional expression and chromosomal localization of the corresponding genes. EMBO J. 9, 771± 776. Grewer, C., 1999. Investigation of the alpha(1)-glycine receptor channel-opening kinetics in the submillisecond time domain. Biophys. J. 77, 727±738. Griffon, N., Buttner, C., Nicke, A., Kuhse, J., Schmalzing, G., Betz, H., 1999. Molecular determinants of glycine receptor subunit assembly. EMBO J. 18, 4711±4721. Gundlach, A.L., Beart, P.M., 1981. [3]H-Strychnne binding suggests glycine receptors in the ventral tegmental area of rat brain. Neurosci. Lett. 22, 289±294. Hamill, O.P., Bormann, J., Sakmann, B., 1983. Activation of multiple-conductance state chloride channels in spinal neurones by glycine and GABA. Nature 305, 805±808. Han, Y., Wu, S.M., 1999. Modulation of glycine receptors in retinal ganglion cells by zinc. Proc. Natl. Acad. Sci. USA 96, 3234±3238. Handford, C.A., Lynch, J.W., Baker, E., Webb, G.C., Ford, J.H., Sutherland, G.R., Scho®eld, P.R., 1996. The human glycine receptor beta subunit: primary structure, functional characterisation and chromosomal localisation of the human and murine genes. Brain Res. Mol. Brain Res. 35, 211±219. Hartenstein, B., Schenkel, J., Kuhse, J., Besenbeck, B., Kling, C., Becker, C.M., Betz, H., Weiher, H.,

220

K. Becker et al.

1996. Low level expression of glycine receptor beta subunit transgene is suf®cient for phenotype correction in spastic mice. EMBO J. 15, 1275±1282. Harty, T.P., Manis, P.B., 1998. Kinetic analysis of glycine receptor currents in ventral cochlear nucleus. J. Neurophysiol. 79, 1891±1901. Harvey, R.J., Thomas, P., James, C.H. Wilderspin, A., Smart, T.G., 1999. Identi®cation of an inhibitory Zn 2+ binding site on the human glycine receptor alpha1 subunit. J. Physiol. (Lond.) 520 (1), 53±64. Hille, B., 1992. Ionic Channels of Excitable Membranes, 2nd ed. Sinauer, Sunderland, MA. Hoch, W., Betz, H., Becker, C.M., 1989. Primary cultures of mouse spinal cord express the neonatal isoform of the inhibitory glycine receptor. Neuron 3, 339±348. Hoch, W., Betz, H., Schramm, M., Wolters, I., Becker, C.-M., 1992. Modulation by NMDA-receptor antagonists of glycine receptor isoform expression in cultured spinal cord neurons. Eur. J. Neurosci. 4, 389±395. Huganir, R.L., Delcour, A.H., Greengard, P., Hess, G.P., 1986. Phosphorylation of the nicotinic acetylcholine receptor regulates its rate of desensitization. Nature 321, 774±776. Huganir, R.L., Greengard, P., 1990. Regulation of neurotransmitter receptor desensitization by protein phosphorylation. Neuron 5, 555±567. Karlin, A., Akabas, M.H., 1998. Substituted-cysteine accessibility method. Methods Enzymol. 293, 123± 145. Kazazian, H.H., 1998. Mobile elements and disease. Curr. Opin. Genet. Dev. 8, 343±350. Kingsmore, S.F., Giros, B., Suh, D., Bieniarz, M., Caron, M.G., Seldin, M.F., 1994. Glycine receptor betasubunit gene mutation in spastic mouse associated with LINE-1 element insertion. Nat. Genet. 7, 136± 141. Kirsch, J., Betz, H., 1998. Glycine-receptor activation is required for receptor clustering in spinal neurons. Nature 392, 717±720. Kirsch, J., Wolters, I., Triller, A., Betz, H., 1993. Gephyrin antisense oligonucleotides prevent glycine receptor clustering in spinal neurons [see comments]. Nature 366, 745±748. Kling, C., Koch, M., Saul, B., Becker, C.M., 1997. The frameshift mutation oscillator (Glra1(spd-ot)) produces a complete loss of glycine receptor alpha1-polypeptide in mouse central nervous system. Neuroscience 78, 411±417. Koltchine, V.V., Ye, Q., Finn, S.E., Harrison, N.L., 1996. Chimeric GABAA/glycine receptors: expression and barbiturate pharmacology. Neuropharmacology 35, 1445±1456. Kuhse, J., Schmieden, V., Betz, H., 1990a. Identi®cation and functional expression of a novel ligand binding subunit of the inhibitory glycine receptor. J. Biol. Chem. 265, 22317±22320. Kuhse, J., Schmieden, V., Betz, H., 1990b. A single amino acid exchange alters the pharmacology of neonatal rat glycine receptor subunit. Neuron 5, 867±873. Kuhse, J., Kuryatov, A., Maulet, Y., Malosio, M.L., Schmieden, V., Betz, H., 1991. Alternative splicing generates two isoforms of the alpha 2 subunit of the inhibitory glycine receptor. FEBS Lett 283, 73±77. Kumamoto, E., Murata, Y., 1996. Glycine current in rat septal cholinergic neuron in culture: monophasic positive modulation by Zn 21. J. Neurophysiol. 76, 227±241. Langosch, D., Herbold, A., Schmieden, V., Borman, J., Kirsch, J., 1993. Importance of Arg-219 for correct biogenesis of alpha 1 homooligomeric glycine receptors. FEBS Lett 336, 540±544. Langosch, D., Laube, B., Rundstrom, N., Schmieden, V., Bormann, J., Betz, H., 1994. Decreased agonist af®nity and chloride conductance of mutant glycine receptors associated with human hereditary hyperekplexia. EMBO J. 13, 4223±4228. Laube, B., Kuhse, J., Rundstrom, N., Kirsch, J., Schmieden, V., Betz, H., 1995. Modulation by zinc ions of native rat and recombinant human inhibitory glycine receptors. J. Physiol. (Lond) 483, 613±619. Legendre, P., 1999. Voltage dependence of the glycine receptor channel kinetics in the zebra®sh hindbrain. J. Neurophysiol. 82, 2120±2129.

The inhibitory glycine receptor as a model of hereditary channelopathies

221

Le Novere, N., Changeux, J.P., 1995. Molecular evolution of the nicotinic acetylcholine receptor: an example of. J. Mol. Evol. 40, 155±172. Levi, S., Vannier, C., Triller, A., 1998. Strychnine-sensitive stabilization of postsynaptic glycine receptor clusters. J. Cell Sci. 111, 335±345. Lewis, C.A., Ahmed, Z., Faber, D.S., 1989. Characteristics of glycine-activated conductances in cultured medullary neurons from embryonic rat. Neurosci. Lett. 96, 185±190. Lynch, J.W., Rajendra, S., Pierce, K.D., Handford, C.A., Barry, P.H., Scho®eld, P.R., 1997. Identi®cation of intracellular and extracellular domains mediating signal transduction in the inhibitory glycine receptor chloride channel. EMBO J. 16, 110±120. Lynch, J.W., Jacques, P., Pierce, K.D., Scho®eld, P.R., 1998. Zinc potentiation of the glycine receptor chloride channel is mediated by allosteric pathways. J. Neurochem. 71, 2159±2168. Maammar, M., Rodeau, J.L., Taleb, O., 1997. Permeation and gating of alpha1 glycine-gated channels expressed at low and high density in Xenopus oocyte. FEBS Lett. 414, 99±104. Malosio, M.L., Grenningloh, G., Kuhse, J., Schmieden, V., Schmitt, B., Prior, P., Betz, H., 1991a. Alternative splicing generates two variants of the alpha 1 subunit of the inhibitory glycine receptor. J. Biol. Chem. 266, 2048±2053. Malosio, M.L., Marqueze, P.B., Kuhse, J., Betz, H., 1991b. Widespread expression of glycine receptor subunit mRNAs in the adult and developing rat brain. EMBO J. 10, 2401±2409. Marvizon, J.C., Garcia-Calvo, M., Vazquez, J., Mayor, F.J., Ruiz-Gomez, A., Valdiviesco, F., Benavides, J., 1986. Activation and inhibition of 3H-strychnine binding to the glycine receptor by Eccles' anions: modulatory effects of cations. Mol. Pharmacol. 30, 598±602. Marvizon, J.C., Skolnick, P., 1988. [3H]glycine binding is modulated by Mg 21 and other ligands of the NMDA receptor-cation channel complex. Eur. J. Pharmacol., 151±158. Marvizon, J.C., Vazquez, J., Garcia-Calvo, M., Mayor, F.J., Ruiz-Gomez, A., Valdivieso, F., Benavides, J., 1986. The glycine receptor: pharmacological studies and mathematical modeling of the allosteric interaction between the glycine- and strychnine binding sites. Mol. Pharmacol. 30, 590± 597. Mascia, M.P., Machu, T.K., Harris, R.A., 1996a. Enhancement of homomeric glycine receptor function by long-chain alcohols and anaesthetics. Br. J. Pharmacol. 119, 1331±1336. Mascia, M.P., Mihic, S.J., Valenzuela, C.F., Scho®eld, P.R., Harris, R.A., 1996b. A single amino acid determines differences in ethanol actions on strychnine-sensitive glycine receptors. Mol. Pharmacol. 50, 402±406. Mascia, M.P., Wick, M.J., Martinez, L.D., Harris, R.A., 1998. Enhancement of glycine receptor function by ethanol: role of phosphorylation. Br. J. Pharmacol. 125, 263±270. Matzenbach, B., Maulet, Y., Sefton, L., Courtier, B., Avner, P., Guenet, J.L., Betz, H., 1994. Structural analysis of mouse glycine receptor alpha subunit genes. Identi®cation and chromosomal localization of a novel variant. J. Biol. Chem. 269, 2607±2612. Milani, N., Dalpra, L., del Prete, A., Zanini, R., Larizza, L., 1996. A novel mutation (Gln266 ! His) in the alpha 1 subunit of the inhibitory glycine-receptor gene (GLRA1) in hereditary hyperekplexia [letter]. Am. J. Hum. Genet. 58, 420±422. Mishina, M., Takai, T., Imoto, K., Noda, M., Takahashi, T., Numa, S., Methfessel, C., Sakmann, B., 1986. Molecular distinction between fetal and adult isoforms of muscle acetylcholine receptor. Nature 321, 406±411. Moorhouse, A.J., Jacques, P., Barry, P.H., Scho®eld, P.R., 1999. The startle disease mutation Q266H, in the second transmembrane domain of the human glycine receptor, impairs channel gating. Mol. Pharmacol. 55, 386±395. MuÈlhardt, C., Fischer, M., Gass, P., Simon Chazottes, D., Guenet, J.L., Kuhse, J., Betz, H., Becker, C.M., 1994. The spastic mouse: aberrant splicing of glycine receptor beta subunit mRNA caused by intronic insertion of L1 element. Neuron 13, 1003±1015.

222

K. Becker et al.

MuÈller, W.E., Snyder, S.H., 1978. Strychnine binding associated with synaptic glycine receptors in rat spinal cord membranes: ionic in¯uences. Brain Res. 147, 107±116. Nicolic, Z., Laube, B., Weber, R.G., Lichter, P., Kioschis, P., Poustka, A., Mulhardt, C., Becker, C.-M., 1998. The human glycine receptor subunit a3: GLRA3 gene structure, chromosomal location, and functional characterization of alternative transcripts. J. Biol. Chem. 273, 8±19714. O'Connor, V.M., 1989. Chemical modi®cation of overlapping but conformationally distinct recognition sites for glycine and strychnine in isolated spinal cord membranes. J. Physiol. (Lond) 415, 49P. O'Connor, V., Phelan, P.P., Fry, J.P., 1996. Interactions of glycine and strychnine with their receptor recognition sites in mouse spinal cord. Neurochem. Int. 29, 423±434. Pribilla, I., Takagi, T., Langosch, D., Bormann, J., Betz, H., 1992. The atypical M2 segment of the beta subunit confers picrotoxinin resistance to inhibitory glycine receptor channels [published erratum appears in EMBO J 1994 Mar 15;13(6):1493]. EMBO J. 11, 4305±4311. Rajendra, S., Lynch, J.W., Pierce, K.D., French, C.R., Barry, P.H., Scho®eld, P.R., 1995. Mutation of an arginine residue in the human glycine receptor transforms beta-alanine and taurine from agonists into competitive antagonists. Neuron 14, 169±175. Rees, M.I., Andrew, M., Jawad, S., Owen, M.J., 1994. Evidence for recessive as well as dominant forms of startle disease (hyperekplexia) caused by mutations in the alpha 1 subunit of the inhibitory glycine receptor. Hum. Mol. Genet. 3, 2175±2179. Ren, J., Ye, J.H., McArdle, J.J., 1998. cAMP-dependent protein kinase modulation of glycine-activated chloride current in neurons freshly isolated from rat ventral tegmental area. Brain Res. 811, 71±78. Roberts, R.J., Shelton, C.J., Daniels, S., Smith, E.B., 1996. Glycine activation of human homomeric alpha 1 glycine receptors is sensitive to pressure in the range of the high pressure nervous syndrome. Neurosci. Lett. 208, 125±128. Rozmahel, R., Wilschanski, M., Matin, A., Plyte, S., Oliver, M., Auerbach, W., Moore, A., Forstner, J., Durie, P., Nadeau, J., Bear, C., Tsui, L.C., 1996. Modulation of disease severity in cystic ®brosis transmembrane conductance regulator de®cient mice by a secondary genetic factor [published erratum appears in Nat. Genet. 1996;13(1):129]. Nat. Genet. 12, 280±287. Ruiz-Gomez, A., Garcia-Calvo, M., Vazquez, J., Marvizon, J.C., Valdivieso, F., Mayor, F.J., 1989. Thermodynamics of agonist and antagonist interaction with the strychnine-sensitive glycine receptor. J. Neurochem. 52, 1775±1780. Ruiz-Gomez, A., Morato, E., Garcia, C.M., Valdivieso, F., Mayor, F.J., 1990. Localization of the strychnine binding site on the 48-kilodalton subunit of the glycine receptor. Biochemistry 29, 7033±7040. Ruiz-Gomez, A., Vaello, M.L., Valdivieso, F., Mayor Jr, F., 1991. Phosphorylation of the 48-kDa subunit of the glycine receptor by protein kinase C. J. Biol. Chem. 266, 559±566. Ryan, S.G., Dixon, M.J., Nigro, M.A., Kelts, K.A., Markand, O.N., Terry, J.C., Shiang, R., Wasmuth, J.J., O'Connell, P., 1992. Genetic and radiation hybrid mapping of the hyperekplexia region on chromosome 5q. Am. J. Hum. Genet. 51, 1334. Ryan, S.G., Buckwalter, M.S., Lynch, J.W., Handford, C.A., Segura, L., Shiang, R., Wasmuth, J.J., Camper, S.A., Scho®eld, P., O'Connell, P., 1994. A missense mutation in the gene encoding the alpha 1 subunit of the inhibitory glycine receptor in the spasmodic mouse. Nat. Genet. 7, 131±135. Sanes, J.R., Lichtman, J.W., 1999. Development of the vertebrate neuromuscular junction. Annu. Rev. Neurosci. 22, 389±442. Saul, B., Schmieden, V., Kling, C., Mulhardt, C., Gass, P., Kuhse, J., Becker, C.M., 1994. Point mutation of glycine receptor alpha 1 subunit in the spasmodic mouse affects agonist responses. FEBS Lett 350, 71±76. Saul, B., Kuner, T., Sobetzko, D., Brune, W., Hanefeld, F., Meinck, H.M., Becker, C.M., 1999. Novel GLRA1 missense mutation (P250T) in dominant hyperekplexia de®nes an intracellular determinant of glycine receptor channel gating. J. Neurosci. 19, 869±877.

The inhibitory glycine receptor as a model of hereditary channelopathies

223

Schmieden, V., Grenningloh, G., Scho®eld, P.R., Betz, H., 1989. Functional expression in Xenopus oocytes of the strychnine binding 48 kd subunit of the glycine receptor. EMBO J. 8, 695±700. Schnieden, V., Kuhse, J., Betz, H., 1999. A novel domain of the inhibitory glycine receptor determining antagonist ef®cacies: further evidence for partial agonism resulting from self-inhibition. Mol. Pharmacol. 56, 464±472. Shiang, R., Ryan, S.G., Zhu, Y.Z., Hahn, A.F., O'Connell, P., Wasmuth, J.J., 1993. Mutations in the alpha 1 subunit of the inhibitory glycine receptor cause the dominant neurologic disorder, hyperekplexia. Nat. Genet. 5, 351±358. Silva, A.J., Simpson, E.M., Takahashi, J.S., Lipp, H., Nakanishi, S., Wehner, J.M., Giese, K.P., Tully, T., Abel, T., Chapman, P.F., Fox, K., Grant, S., Itohara, S., Lathe, R., Mayford, M., McNamara, J.O., Morris, R.J., Picciotto, M., Roder, J., Shin, H.-S., Slesinger, P.A., Storm, D.R., Stryker, M.P., Tonegawa, S., Wang, Y., Wolfer, D.P. 1997. Mutant mice and neuroscience: recommendations concerning genetic background. Neuron 19, 755±759. Smith, G.B., Olsen, R.W., 1995. Functional domains of GABAA receptors. Trends Pharmacol. Sci. 16, 162±168. Smithies, O., Maeda, N., 1995. Gene targeting approaches to complex genetic diseases: atherosclerosis and essential hypertension. Proc. Natl. Acad. Sci. USA 92, 5266±5272. Sontheimer, H., Becker, C.M., Pritchett, D.B., Scho®eld, P.R., Grenningloh, G., Kettenmann, H., Betz, H., Seeburg, P.H., 1989. Functional chloride channels by mammalian cell expression of rat glycine receptor subunit. Neuron 2, 1491±1497. Stroud, R.M., McCarthy, M.P., Shuster, M., 1990. Nicotinic acetylcholine receptor superfamily of ligandgated ion channels. Biochemistry 29, 11009±11023. Suhren, O., Bruyn, G.W., Tuyman, J.A., 1966. Hyperekplexia: a hereditary startle syndome. J. Neurol. Sci. 3, 577±605. Takahashi, T., Momiyama, A., Hirai, K., Hishinuma, F., Akagi, H., 1992. Functional correlation of fetal and adult forms of glycine receptors with developmental changes in inhibitory synaptic receptor channels. Neuron 9, 1155±1161. Thiel, G., Lietz, M., Cramer, M., 1998. Biological activity and modular structure of RE-1-silencing transcription factor (REST), a repressor of neuronal genes. J. Biol. Chem. 273, 26891±26899. Tokutomi, N., Kaneda, M., Akaike, N., 1989. What confers speci®city on glycine for its receptor site?. Br. J. Pharmacol. 97, 353±360. Unwin, N., 1995. Acetylcholine receptor channel imaged in the open state. Nature 373, 37±43. Unwin, N., 1996. Projection structure of the nicotinic acetylcholine receptor: distinct conformations of the alpha subunits. J. Mol. Biol. 257, 586±596. Vaello, M.L., Ruiz-Gomez, A., Mayor Jr, F., 1992. Glycinergic ligands modulate the rate of phosphorylation of the glycine receptor by protein kinase C. Biochem. Biophys. Res. Commun. 188, 813±819. Vaello, M.L., Ruiz-Gomez, A., Lerma, J., Mayor Jr, F., 1994. Modulation of inhibitory glycine receptors by phosphorylation by protein kinase C and cAMP-dependent protein kinase. J. Biol. Chem. 269, 2002±2008. Vafa, B., Lewis, T.M., Cunningham, A.M., Jacques, P., Lynch, J.W., Scho®eld, P.R., 1999. Identi®cation of a new ligand binding domain in the alpha1 subunit of the inhibitory glycine receptor. J. Neurochem. 73, 2158±2166. Van Hooft, J.A., Vijverberg, H.P., 1995. Phosphorylation controls conductance of 5-HT3 receptor ligandgated ion channels. Rec. Chan. 3, 7±12. Vergouwe, M.N., Tijssen, M.A., Peters, A.C., Wielaard, R., Frants, R.R., 1999. Hyperekplexia phenotype due to compound heterozygosity for the GLRA1 gene mutations. Ann. Neurol. 46, 634±638. Wallace, D.P., Tomich, J.M., Iwamoto, T., Henderson, K., Grantham, J.J., Sullivan, L.P., 1997. A synthetic peptide derived from glycine-gated Cl 2 channel induces transepithelial Cl- and ¯uid secretion. Am. J. Physiol. 272, C1672±C1679.

224

K. Becker et al.

Walstrom, K.M., Hess, G.P., 1994. Mechanism for the channel-opening reaction of strychnine-sensitive glycine receptors on cultured embryonic mouse spinal cord cells. Biochemistry 33, 7718±7730. Warrington, J.A., 1994. High resolution physical mapping of human 5q31-q33 using three methods: radiation hybrid mapping, interphase ¯uorescence in situ hybridization, and pulsed-®eld gel electrophoresis. Genomics 24, 395±398. Werman, R., Davidoff, R.A., Aprison, M.H., 1986. Inhibitory action of glycine on spinal neurons in the cat. Nature 214, 681±683. White, W.F., 1985. The glycine receptor in the mutant mouse spastic (spa): strychnine binding characteristics and pharmacology. Brain Res. 329, 1±6. White, B.H., Cohen, J.B., 1992. Agonist-induced changes in the structure of the acetylcholine receptor M2 regions revealed by photoincorporation of an uncharged nicotinic non-competitive antagonist. J. Biol. Chem. 267, 15770±15783. Wick, M.J., Mihic, S.J., Ueno, S., Mascia, M.P., Trudell, J.N., Brozowski, S.J., Ye, Q., Harrison, N.L., Harris, R.A., 1998. Mutations of g-aminobutyric acid and glycine receptors change alcohol cutoff: evidence for an alcohol receptor?. Proc. Natl. Acad. Sci. USA 95, 6504±6509. Ye, Q., Koltchine, V.V., Mihic, S.J., Mascia, M.P., Wick, M.J., Finn, S.E., Harrison, N.L., Harris, R.A., 1998. Enhancement of glycine receptor function by ethanol is inversely correlated with molecular volume at position alpha267. J. Biol. Chem. 273, 3314±3319. Young, A.B., Snyder, S.H., 1974a. The glycine synaptic receptor: evidence strychnine binding is associated with the ionic conductance mechanism. Proc. Natl. Acad. Sci. USA 71, 4002±4005. Young, A.B., Snyder, S.H., 1974b. Strychnine binding in rat spinal cord membranes associated with the synaptic glycine receptor: cooperativity of glycine interactions. Molec. Pharmac. 10, 790±809.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

227

Chapter 9

Cystic ®brosis R. Greger Physiologisches Institut, Albert-Ludwigs-UniversitaÈt, Hermann-Herder-Straûe 7, D-79104 Freiburg, Germany

Abstract Cystic ®brosis (CF) is one of the most frequent genetic diseases with one case in 2000±4000 life births. It is a very serious disease with a mean life expectancy of 20±40 years. The patients suffer mostly from chronic pulmonal infections which lead to lung destruction, right heart insuf®ciency and heart failure. The disease has been de®ned as a clinical entity only in this century and the underlying pathophysiology, namely a genetic defect in epithelial Cl 2 transport, has only been recognized 30 years ago. It took another 15 years to generate general awareness of this defect and to increase the research intensity in the basic mechanisms of this disease. Since then CF has made it into headlines in biomedical research. The genetic defect is known. The defective gene product, the cystic ®brosis transmembrane conductance regulator (CFTR), was identi®ed exactly 10 years ago. In this decade CFTR has been studied intensively with all technologies available and CF research has become one of the major topics in epithelial physiology. As a result we know today that this complex molecule functions as a Cl channel and as a regulator of other ion channels and transporters. A defect in CFTR leads to reduced NaCl and water secretion in the airways and in other epithelia. In addition NaCl absorption is enhanced. As a result the clearing of the airways is impeded and chronic colonization by pathogenic bacteria such as Pseudomonas aeruginosa leads to airway destruction. CF is one of the ®rst genetic diseases in which genetic therapy is being attempted. The initial enthusiasm has meanwhile been spoiled by the very limited success in animal models and no convincing bene®t for the patient. Currently new approaches for gene and even more so for `classical' therapy are under study. In addition, the CFTR molecule with all its complex functions is the target of basic research. It is entirely feasible that closer understanding of this molecule, its synthesis and maturation and its interaction with other transporters will lead to new and maybe unexpected therapeutic strategies. q 2000 Elsevier Science B.V. All rights reserved.

228

R. Greger

Introduction A typical case A girl aged 18 months is presented to the pediatrician with the symptoms of recurrent airway infection and with the failure to thrive. After weaning 1 year ago the baby has fed quite normally. The mother has noted bulky stools. During the past year the girl has suffered from recurrent infections which have been treated by the family doctor. The physical examination is not very instructive apart from the fact that the weight and height are clearly below average, the rectal temperature is 388C. The chest X-ray shows signs of pulmonal infection, leukocytes are elevated, the plasma electrophoresis indicates a chronic infection. Amongst other tests the pediatrician also performs a so-called sweat test in which pilocarpin is applied locally to the skin and sweat is collected. The sweat Cl 2 concentration is pathologically high: 100 mmol/l. The suspected diagnosis cystic ®brosis (CF) is made very likely by a genetic analysis in as much as the father and the daughter share the DF508 mutation of the CF gene on one allele of chromosome 7. The other allele is unknown, although this center searches routinely for the 20 most frequent mutations in their population. The parents are informed about the disease and the child, from now on, is treated as an outpatient of this CF center. This case history reads much less mysteriously than the same story some 500 years ago, when midwives used to say: `the child whose forehead when kissed tastes salty is bewitched and will die soon'. This chapter will attempt to summarize the current understanding of the pathophysiology of this disease. Why does this channelopathy lead to airway infection? What is the function of this gene? Which other defects can be understood today? What could be new therapeutic strategies? Clinical features Diagnostic criteria A clear diagnostic criterion for cystic ®brosis (CF) was established only some 45 years ago when di Sant'Agnese et al. (1953) noted that the sweat NaCl concentration was pathologically high in these patients. Then the sweat test was introduced as a routine procedure. Nowadays it is performed as the pilocarpin iontophoresis stimulation test. Normal sweat NaCl concentrations should be below 40 mmol/l. Higher values are regarded as positive. Some children suffering from CF (10±20%) present at birth with meconium ileus. Even in the children who do not suffer from meconium ileus the albumin content of the meconium may be markedly elevated, indicating a pancreatic insuf®ciency. In

Cystic ®brosis

229

the adult a meconium ileus equivalent is found frequently, which is caused by solid mucus containing masses in the terminal ileum and colon ascendens, where usually the contents are liquid. Also circulating trypsin may be elevated in CF newborns and can be diagnosed by the immuno reactive trypsin test. Today, with the knowledge of the genetic defect, DNA analysis has become a routine procedure. In 60±70% of our CF population one allele contains the mutation DF508 (cf. Fig.6 for the putative structure of CFTR). Other frequent mutations in our population (with decreasing frequency) are: G542X, G551D, N1303K, W1282X, R553X, 621 1 1G ! T, 1717-1G ! A, R117H, R1162X. Centers will nowadays be able to test for all these and more mutations, but the search for any of the .800 known mutations is not routinely feasible. Therefore, in several instances the genetic analysis will be negative, unless the entire gene is sequenced for the given patient (cf. below). Due to the fact that the sweat test can be borderline falsely positive or falsely negative and that the genetic information is not conclusive, additional functional tests have been developed over the years. One test is the measurement of the nasal transepithelial voltage. The voltage in the nose is measured against a skin or i.v. reference. CF patients usually have a larger voltage and amiloride, a speci®c inhibitor of the epithelial Na 1 channel, has a larger effect (Knowles et al., 1981). The understanding of the altered sweat test and the nasal transepithelial voltage requires an explanatory pathophysiological concept. This concept has been developed in pioneering studies by Schulz (1969) 30 years ago, but has been neglected until the same basic observation was reported by Quinton (1983) in 1983. The basic result is depicted in Fig. 1A. Here the transepithelial voltage in sweat ducts is measured before and after stimulation of secretion. It is evident that the lumen negative voltage is larger in CF patients under control conditions and even more so after stimulation. Schulz concluded that this larger voltage was caused by the fact that the luminal membrane was impermeable to Cl 2 but permeable to Na 1 (Fig. 1B). This gave a ready explanation for the failure of the sweat gland duct to absorb NaCl and hence for the elevated NaCl concentration in sweat. Obviously the same pathophysiology applies to the respiratory epithelium and causes the augmented voltages across the nasal epithelium. Knowles et al. (1981) have noticed that amiloride has a stronger effect in CF children. This was expected on the basis of the scheme shown in Fig. 1B. However, they postulated that above and beyond the Cl 2 channel defect the amiloride-inhibitable Na 1 conductance is increased in CF patients. This was shown directly by the same group a few years ago (Stutts et al., 1995, also cf. below). More recently the examination of the colonic mucosa has been introduced as a very sensitive additional test. In the colonic mucosa, depending on the functional status, NaCl absorption and NaCl secretion can occur (Greger et al., 1997). Cl 2 secretion is under the control of the second messengers cAMP, cGMP and Ca 21. Na 1 absorption is controlled by the mineralocorticoid aldosterone. A small speci-

230

R. Greger

Fig. 1. (A) Measurements of transepithelial voltage in sweat ducts from normal individuals (solid circles) and CF patients (open circles). Con, control; Stim, cholinergic stimulation of secretion (Schulz, 1969). (B) Cell model explaining the increased voltage in CF patients. Arrow, ion channel; circle with ATP ˆ …Na1 1 K1 †-ATPase. In CF patients the voltage is increased because the Cl 2 channels are defective.

men of the colonic/rectal mucosa can be obtained by biopsy. A very small piece is suf®cient for an Using chamber analysis in our new chamber (aperture: 0.95 mm 2). A typical experiment of a healthy individual and a CF patient is depicted in Fig. 2. In the healthy individual as well as in the CF patient the spontaneous transepithelial voltage can be lumen negative. Three ®ndings differentiate the normal individual from the CF patient: (1) cAMP producing agonists, forskolin, membrane permeable

Cystic ®brosis

231

Fig. 2. Using chamber measurements of rectal biopsies of a normal individual (A) and a CF patient (B). The spontaneous transepithelial voltage (Vte) is plotted as a function of time. Every few seconds a short current pulse is injected producing a voltage de¯ection (DVte) which is used to calculate transepithelial resistance (Rte). The equivalent short circuit current …I sc † ˆ V te =Rte corresponds to Cl 2 secretion. Carbachol (CCH, 0.1 mmol/l) applied to the serosal side induces a lumen negative voltage de¯ection which is increased in the presence of forskolin (adenylyl cyclase activator, 1mmol/l, serosal side) and serosal isobutylmethylxanthine (IBMX, 0.01 mmol/l, inhibitor of phosphodiesterase). Therefore, CCH is a strong secretagogue in collaboration with cAMP (cf. Fig. 3). In the presence of indomethacin (cyclooxygenase inhibitor, 0.1 mmol/l, serosal side) the spontaneous secretion is abolished and now CCH causes a secretion of K 1 (lumen positive voltage de¯ection, cf. Fig. 3). In the CF patient no spontaneous Cl 2 secretion is present. Therefore, CCH causes only K 1 secretion. Also forskolin and IBMX are without effect (Mall et al., 1999).

232

R. Greger

cAMP or isobutylmethylxanthine (IBMX) cause an increase in lumen negative voltage in the healthy individual but not in CF. (2) Amiloride, the blocker of epithelial Na 1 channels (ENaC), reduces the lumen negative voltage in normal tissue but has an augmented effect in CF tissue (not shown in Fig. 2). (3) In the presence of cAMP stimulation by carbachol (CCH) augments the lumen negative voltage but has the opposite effect in CF tissue (Mall et al., 1999). It is important that CF patients may possess residual function with respect to all three criteria and this partial function may be of high predictive value for the course of the disease. Genotype±phenotype correlations for this essay system are currently established. The basic coordination of NaCl secretion by the colonocyte is summarized in Fig. 3. Differential diagnosis It took quite some time to differentiate CF from celiac disease. Today the differential diagnosis should not be too dif®cult. Celiac disease does not affect the airways. The sweat test and the other tests mentioned above should guide to the correct diagnosis. Other diseases with malabsorption and a failure to thrive such as malabsorption syndromes, meconium plugs, inborn pancreas hypoplasia, gastrointestinal allergy and other rare causes can all be distinguished by the above tests. Pathological ®ndings CF is a generalized channelopathy. Many of the tissues which express CFTR are affected. The mostly affected organs are: (1) sweat glands; (2) airway epithelia; (3) gut; (4) pancreas and other exocrine glands; (5) reproductive organs; (6) bile ducts and gallbladder. CFTR plays a crucial role in exocrine secretion. Thus exocrine glands are affected. In the pancreas, like in salivary glands, the Cl 2 channels present in the luminal membrane are different from CFTR. In pancreatic acini and in tear gland acini Ca 21 activated Cl 2 channels of very small conductance have now been identi®ed (Zdebik et al., 1997). Nevertheless the pancreas is heavily affected in many but not all CF patients. In fact, the name cystic ®brosis alludes to the cystic degeneration of the pancreas in the affected children. Moreover, malnutrition is largely due to pancreatic insuf®ciency. This pancreatic insuf®ciency is caused by the CF defect in the pancreatic ducts rather than in the acini. In pancreatic ducts CFTR Cl 2 channels in the luminal membrane are of importance for Cl 2 recycling which is necessary for Cl 2/HCO32 mediated secretion of HCO32 (Novak and Greger, 1988; Gray et al., 1994). Defects in CFTR reduce the ability to secrete HCO32 and cause and increased MUC6 mucin secretion, which obstructs the ducts (Reid et al., 1997; Lee et al., 1999). The enzymes produced by the acini then destroy the exocrine and later even the endocrine pancreas. Maldigestion and even diabetes mellitus are the consequences.

Cystic ®brosis

233

Fig. 3. Scheme of the mechanisms of Cl 2 and K 1 secretion in a colonocyte. Arrow, ion channel; circle with ATP ˆ …Na1 1 K1 †-ATPase; circle, Na 1 2Cl 2 K 1 cotransporter; HTX, heat stable toxin from E. coli; 293B, chromanol inhibiting KVLQT1 K 1 channels (Greger et al., 1997); Clot, clotrimazol, an inhibitor of small Ca 21 activated K 1 channels; Azose, azosemide, an inhibitor of colonic Na 1 2Cl 2 K 1 cotransporter; PGE2, prostaglandin E2; PKA, proteinkinase A; PKC, proteinkinase C; GKII, cGMP activated proteinkinase II; Ach, acetylcholine. Two major activating pathways are shown: (a) the cAMP pathway activating CFTR Cl 2 channels directly and KVLQT1 K 1 channels indirectly. (b) The Ca 21 pathway which is activated by IP3 mediated Ca 21 release from cytosolic stores and by Ca 21 in¯ux. An increase in cytosolic Ca 21 has probably four effects: activation of basolateral K 1 channels, activation of a luminal K 1 channel, activation of Cl 2 channels (?) and activation of PKC. In addition the cGMP pathway also converges onto the CFTR Cl 2 channel. The cAMP and the Ca 21 pathways interact in positive and negative manner: (a) on the one hand, cAMP and Ca 21 assist each other as secretagogues, because activated K 1 channels provide additional driving force for Cl 2 secretion; on the other hand, the cAMP induced depolarization (opening of Cl 2 channels) reduces Ca 21 in¯ux and hence inhibits the Ca 21 pathway. In CF patients the primary defect is on the luminal CFTR type Cl 2 channel. Therefore, cAMP producing agonists are without effect (Fig. 2B). CCH only activates K 1 conductances. The increase in luminal membrane K 1 conductance leads to a lumen positive voltage de¯ection (Fig. 2B). The same is seen in normals if the cAMP pathway is inactivated by indomethacin (Fig. 2A).

234

R. Greger

The defect in the respiratory tract, despite intense research efforts during the past decade, is not completely understood at this stage. The distal airways possess the ability to absorb and to secrete ¯uid and NaCl. Absorption probably occurs in the surface cells, whilst secretion is characteristic of the bronchial glands. The ®ne balance of the two processes determines the liquid ¯uid layer covering the surface of the ciliated cells. CF apparently disturbs this equilibrium in two major ways: (1) Cl 2 secretion especially by glandular cells is impaired because of the CFTR Cl 2 channel defect in the luminal membrane of these glandular cells; (2) the absorption of Na 1 is probably enhanced in the surface cells, and this process too is caused by the CFTR effect (cf. below). Both processes contribute to the reduced mucociliary clearance of CF airways, which predisposes for chronic airway infection (Pilewski and Frizzell, 1999). Additional factors have also been incriminated such as reduced glycosylation and sialylation and increased sulfation and fucosylation of surface glycoproteins, the NaCl concentration in surface ¯uid, the absence of defensins, etc. (Imundo et al., 1995; Smith et al., 1996; Hill et al., 1997; Pilewski and Frizzell, 1999). It is generally accepted now that the defect in CFTR is causally related to lung pathophysiology. The defect in sweat glands has already been discussed above because it is of primordial importance for diagnosis. Pathophysiologically the CFTR defect in CF sweat glands is less relevant. On the one hand, the sweat excretion of NaCl is increased and NaCl losses may become a general problem especially when sweating becomes the major effector mechanism in thermoregulation (Greger, 1995). On the other hand, the production of primary sweat in the gland coil is reduced, just as it is decreased in other glands, and this may reduce sweat rates. It has been claimed that this defect, when partial in heterozygotes, has provided a genetic advantage for these individuals and has contributed to the large incidence of CFTR mutations (also cf. below). The defect in oviduct function, in the composition of cervical mucus (female), and the defect in the ductus deferens or its complete absence (male) may cause infertility. The function of the epididymis has been studied recently in detail and it was found that CFTR is involved in NaCl secretion and that the cAMP pathway is defective in CF mice (Leung et al., 1996). The gut is also affected in many patients. This is entirely expected in as much as CFTR is involved in Cl 2 secretion especially in the crypts. A defect in secretion would predispose to obstructive disease and to meconium ileus in the newborn or meconium ileus equivalent in the adult. It has been claimed that CF protects against bacteria toxin-induced diarrhea (cholera toxin, E. coli heat stable toxin) and that this may represent a genetic advantage for the heterozygous patient. This may be the reason why CF mutations have been preserved so long and why there are so many (probably old) mutations. A recent examination of this issue in CF knock-out mice is not entirely convincing because the design would have to be that of a large ®eld

Cystic ®brosis

235

study going on for many years (Cuthbert et al., 1995). In the colon CFTR does not only determine the cAMP- and cGMP-mediated Cl 2 secretion it also controls the function of ENaC channels which are present especially in the apical membrane (Ecke et al., 1996). CFTR also seems to play a major role in mucus secretion. In fact, in the many knockout and transgenic CF mice produced over the past 7 years the dominant pathology is found in the intestine, and most of these mice die early after birth from intestinal obstructive disease (Grubb and Boucher, 1999). CFTR is also present in the bile ducts and in the gallbladder, and CF, in some patients goes along with bilary disease and even liver cirrhosis. The detailed pathophysiological mechanisms have not been studied thus far. Recent experiments suggest that CFTR is involved in the secretion in the gallbladder, but there is little evidence for CFTR defects in hepatocyte function (Peters et al., 1997; Colombo et al., 1998; Chinet et al., 1999). CFTR is present in various tubule segments of the kidney, including the proximal tubule and collecting duct (Montrose-Ra®zadeh and Guggino, 1990). A clear-cut function for CFTR in NaCl absorption has never been veri®ed. Moreover, there is no clear primary clinical defect in kidney function in CF patients. It has been speculated that CFTR is the relevant Cl 2 channel for NaCl absorption in the thick ascending limb of the loop of Henle, but this could never be shown convincingly and it is clear now that the relevant basolateral Cl 2 channel belongs to the CLC family (CLCNKB) because a genetic defect of this channel can generate Bartter's syndrome (Simon et al., 1997). The function of CFTR in the kidney awaits further exploration. Recent data in collecting duct cell cultures may suggest that CFTR functions as a regulator of ENaC (Letz and Korbmacher, 1997). Even more mysterious is the role of CFTR in heart muscle function. There is no doubt that CFTR is expressed in cardiomyocytes, but again, no clear pathology is seen in CF patients, and the role of CFTR Cl 2 channels for the cardiomyocyte action potential is not clear (Gadsby et al., 1995). Functional clinical tests The relevant tests to establish the diagnosis CF have been described above. The severity of the disease can vary largely. Some children are diagnosed early and suffer from serious airway infection as well as from maldigestion (pancreatic insuf®ciency) already in their ®rst years of life, others go undiagnosed for years and even decades. Since the identi®cation of the genetic defect in 1989 by Tsui and co-workers in 1989 (Kerem et al., 1989; Riordan et al., 1989) it has been speculated that there is some correlation between genotype and phenotype in the sense that DF508 predisposes to severe courses, whilst other mutations are regarded mild. Such correlations between genotype, pancreatic function and general course of the disease are

236

R. Greger

obviously multifactorial and must not be used for the prediction of the disease course in the individual patient (Rozmahel et al., 1996). The residual pancreatic function of the patient will determine the need for enzyme medication. The sputum bacteriology will provide information with respect to colonization of the airways and tests of lung function will be used to quantify the respiratory function. Later additional tests of cardiorespiratory function will provide evidence for the degree of lung ®brosis and right heart failure. Therapy The therapy of this disease has been improved over the past decades and resulted in a substantial increase in life expectancy. The general strategies which have proven successful can be summarized as follows: (1) early diagnosis; (2) treatment and continuous monitoring by local and national centers, including psychological guidance and group work; (3) avoidance of hospitalization; (4) physical therapy; (5) antibiotics; (6) appropriate nutrition and enzyme substitution; (7) other speci®c medication; (8) gene therapy. Gene therapy, which has been the great hope (Collins, 1992) after the identi®cation of the gene, has not yet reached the level of general clinical application. Much more work and patience is needed before the possible role of gene therapy can be evaluated. Until then the above and maybe new `conservative' strategies will be required. In the following the above strategies will be discussed brie¯y especially with respect to their pathophysiological basis. (1) Early diagnosis is of very high relevance. Only after the diagnosis has been established at a very early age can all the other therapeutic approaches be utilized in an ef®cient manner. In fact, the correct diagnosis is nowadays made much earlier than was previously done. This is possible for all the cases in which the genetic status is clear-cut. Unfortunately, even with the inclusion of all meanwhile known .800 mutations, only some ,80% of the disease can be explained in the Northern European population (cf.http://www.genet.sickkids.on.ca). Therefore, the symptomatic diagnosis by sweat tests, nasal and colonic mucosa voltage measurements have a very high relevance and the recent improvement of the measurements in the colonic mucosa will probably be extremely helpful (Mall et al., 1998b). (2) Treatment and continuous monitoring by local and national centers, including psychological guidance and group work. The affected children must be monitored by specialized pediatricians. It is of speci®c importance to keep a close eye on the nutritional status of the children. Respiratory tract infections must be treated according to microbiological testing. Enzyme substitution must be adjusted to avoid maldigestion as well as overdosing. Physical therapy has to be supervised. With all these efforts the pediatrician is just one member of a larger team. The co-operation of the patient and of the parents play an enormous role as well as the assistance of patient and parent associations. These associations have not only proven to be of importance

Cystic ®brosis

237

in the coordination of treatment, they distribute information, channel and catalyze contacts. These institutions have also proven to be ¯exible and ef®cient agencies for fundraising and have had an enormous impact on the intensity of research in the CF ®eld. Psychological guidance is required not only for the patient but also for the relatives. (3) Avoidance of hospitalization. Colonization by very pathogenic and dif®cult to treat bacteria must be avoided for as long as possible. Therefore, hospitalization should be avoided and the course of bacterial colonization should be monitored. (4) Physical therapy is extremely relevant to improve and sustain respiratory function. The patient must be trained to clear his/her airways in order to compensate for the reduced mucociliary clearance. These exercises must be by trained and supervised physiotherapists and by the parents. Inhalation is of speci®c importance. The role of mucolytic drugs is less clear. It has been suggested that these drugs also act by stimulation of Cl 2 secretion (KoÈttgen et al., 1996). (5) Antibiotics must be used with great care but fullheartedly. The speci®c drugs are continuously improved. Their usage goes beyond the scope of this short review. (6) Enzyme substitution is required to compensate for the defect in exocrine pancreatic function. Usually the children do have a good appetite, but, because of indigestion, they produce bulky and fatty stools. To the extent that the pancreas undergoes ®brosis enzymes have to be substituted together with food intake. Care is necessary to avoid overdosing which might produce ®brotic gut strictures (®brosing colonopathy). (7) Other speci®c medication will be required depending on the individual case of this multifaceted disease (Schultz et al., 1999). Above and beyond, new approaches are currently under investigation which all are aimed at improving the mucociliary clearance. The scheme depicted in Fig. 4 summarizes the possible strategies to increase NaCl secretion and minimize NaCl absorption in the airways. NaCl undersecretion and overabsorption contribute to the pathophysiology of CF (Pilewski and Frizzell, 1999). Several general basic approaches have been tested, some may lead to new drugs or new applications. ² Phosphodiesterase inhibitors and related drugs. CFTR is activated by PKAdependent protein phosphorylation. Therefore, inhibition of cAMP degradation will enhance any residual function (step 9 in Fig. 4). It has been claimed that some of these drugs have a more direct function on CFTR (step 5 in Fig. 4) but this conclusion does not appear to be justi®ed (Cohen et al., 1997; Kunzelmann et al., 1998). ² Improve CFTR processing (Jiang et al., 1998; Kopito, 1999). This approach will be very promising if successful, because in the case of DF508 CFTR very little of the synthesized mutated CFTR reaches the plasma membrane, but the mutated

238

R. Greger

Fig. 4. Possible therapeutic approaches in a CF airway cell. Arrow, ion channel; circle with ATP ˆ …Na1 1 K1 †-ATPase; circle, Na 1 2Cl 2 K 1 cotransporter; a luminal ATP receptor is shown. The different possible targets for therapy (1±9) are explained in the legend.

protein, once in the plasma membrane may have a residual function (steps 2±4 in Fig. 4). ² Activation of basolateral K 1 channels. This issue is of high relevance. Inspection of Fig. 4 reveals that Cl 2 secretion across the luminal membrane does not only require activation of these CFTR type channels but also a net driving force. This driving force is provided by the K 1 channels of the cell. An increase in driving force may be the largest and most important secretagogue function of carbachol (CCH), ATP, neurotensin and other agonists activating phospholipase C and mobilizing cytosolic Ca 21 (steps 6 and 7 in Fig. 4). Therefore, any drug which will mimic this effect will act as a secretagogue and any inhibitor of K 1 channels will block Cl 2 secretion (Greger et al., 1997). Several imidazole compounds such

Cystic ®brosis

239

as EBIO have been shown to act as activators of Ca 21-regulated K 1 channels in the basolateral membrane and others are inhibitors of these channels such as NS004 (Schultz et al., 1999). ² Ca 21 mobilizing agonists may in addition open other Ca 21 regulated Cl 2 channels in the luminal membrane. The evidence for this from patch clamp data in intact preparations is sparse. Such an effect has been postulated for the respiratory tract and for colonocytes (T84, mHT29, Caco±2) (Stutts et al., 1992; Greger, 1996), and the secretagogue response to inhaled ATP or UTP is believed to be caused by these channels (Knowles et al., 1991) (step 6 in Fig. 4). It is entirely feasible that the limited bene®cial effect from these agonists is also due to the activation of residual CFTR via Ca 21 regulated PKC (Gadsby and Nairn, 1999) or by the above effect on K 1 channels. There is no direct evidence for Ca 21 activated Cl 2 channels in the colon and this may also explain the severe intestinal pathology of CF knockout and transgenic mice (Grubb and Boucher, 1999). ² Direct activation of residual CFTR. A direct and stimulatory effect for the imidazole NS004 has been reported in some but not all studies (Gribkoff et al., 1994; Schultz et al., 1999). The tyrosine kinase inhibitor genistein has been postulated as another CFTR-activating compound, but the data is not uniform and side effects at the used high concentrations are extremely likely (Illek et al., 1995). The claimed effects of phosphodiesterase inhibitors have been discussed above (step 5 in Fig. 4). ² CFTR acts as a modulator of other ion channels. In terms of the pathophysiology of CF, inhibition of epithelial Na 1 channels (ENaC) (Rossier, 1997) appears to be of high relevance (also cf. below). When intact CFTR is activated by PKAdependent phosphorylation. It inhibits ENaC, but this effect is absent in the case of mutated CFTR (Stutts et al., 1995; Mall et al., 1996; Grubb and Boucher, 1999). This disinhibition of ENaC in CF patients may cause overabsorption of ¯uid and hence reduce the mucociliary clearance further (Boucher et al., 1983). From this perspective amiloride inhalation trials are justi®ed (step 8 in Fig. 4). (8) Gene therapy for CF is still at an experimental level. In man, in knockout and transgenic mice adenoviruses, cationic liposomes and even yeast arti®cial chromosomes (YAC) have been tried thus far. The success in general was very limited. One major problem is the effective transfection of distal airways, which would have to be reached for effective repair ( Rich et al., 1990; Manson et al., 1997; Porteous et al., 1997; Grubb and Boucher, 1999). Much more work and patience is required before this general approach can be evaluated and maybe become practically relevant (step 1 in Fig. 4).

240

R. Greger

Genetics The inherited basis of CF as a recessive disorder linked to the long arm of chromosome 7 has long been appreciated. With incredible speed and in an enormous joint effort the site of mutation was narrowed down to in the mid 1980s by Tsui and many collaborators. Eventually, chromosome walking and jumping techniques were used and led to the identi®cation of the gene located at 7q31.2 and coding for a protein with 1440 amino acids (Riordan et al., 1989). It was also clear from their early studies that one deletion defect, leading to the loss of phenylalanine in position 508 (DF508) accounts for approximately 66% of all mutations in our population (Kerem et al., 1989). Incidence and mode of transmission Since the isolation of the CF gene in 1989 (Riordan et al., 1989) an international CF consortium has accumulated more than 800 mutations which still only account for approximately 80% of the diseases (http://www.genet.sickkids.on.ca). The 26 most frequent mutations are listed here in the sequence of their reported incidence (in parenthesis): DF508 (28,948), G542X (1,062), G551D (717), N1303K (589), W1282X (536), R553X (322), 621 1 1G ! T (315), 1717-1G ! A (284), R117H (133), R1162X (125), R347P (106), 3849 1 10kbC ! T (104), DI507 (93), 394delTT (78), G85E (67), R560T (67), A455E (62), 1078delT (57), 2789 1 G ! A (54), R334W (53), 1898 1 1G ! T (53), 711 1 1G ! T (49), 2183AA ! G (40), 3905insT (38), S549N (30), 2184delA (29) (for orientation within CFTR also cf. Fig.6). The incidence of mutations on one allele, i.e. the incidence of heterozygotes is extremely frequent with approximately 50 individuals per 1000. Simple Mendelian considerations leads to a predicted disease frequency of: 0:05 £ 0:05 £ 0:25 ˆ 0:0006, i.e. 1/1600. Indeed the reported incidence is 1/2000± 1/4000. As stated above the incidence of DF508 in heterozygotes and in one allele of homozygotes is approximately 66%. Given the ®ndings that there are so many mutations, that many of these are rare, and that even these 800 mutations explain only some 80% of all cases, makes population screening very problematic, because a negative test in a couple does, on the one hand, not guarantee that the child will be healthy. On the other hand, prenatal diagnosis is possible and available for heterozygote parents. Geneticists have discussed at length why heterozygotes for CF are so frequent and have suggested that they possess one or several heterozygote advantages. It has been proposed that the sweat production by b-agonists is reduced in heterozygotes. It has also been suggested that heterozygotes are more resistant to enteral infections by E. coli or Vibrio cholerae. It even has been suggested that CFTR mediates cellular

Cystic ®brosis

241

uptake of Shigellaand that CFTR defects hence protect against typhoid fever (Pier et al., 1998). Basically these considerations are supported by in vitro studies in CF knockout and transgenic mice (Grubb and Boucher, 1999), although short term studies in these mice have been negative (Cuthbert et al., 1995). The genetic defects in CFTR have been classi®ed according to the site of defect in the protein synthesis (Pilewski and Frizzell, 1999). This is schematically depicted in Fig. 5(1)). Stop mutations lead to truncated mRNA. Therefore, no CFTR is translated and processed. One example is G442X. (2) Mutations like DF508 lead to a folding defect, which is monitored by the respective chaperones and the protein is degraded by proteases. It is remarkable that this step is temperature dependent. At 378C little DF508 CFTR reaches the plasma membrane, but much more at room temperature. Even if it reaches the plasma membrane the channel function of DF508 is disturbed. (3) Other mutations such as G551D are normally transcribed, translated

Fig. 5. Classes of mutations in CFTR. The different classes (1±4) are explained in the legend and text (Pilewski and Frizzell, 1999).

242

R. Greger

and even processed. The protein resides in the plasma membrane but it cannot be properly activated by cAMP. (4) Some mutations such as R347P reside in the plasma membrane and can even be activated by cAMP, but the channel conductance is too low. (5) Recently several splice mutations have been reported. They lead to a partial CFTR defect. One example is the mutation 3849 1 10 kb C ! T. This classi®cation is very helpful to dissect several categories of mutations. It should, however, be clear that the most distal steps in CFTR activation and its incorporation into the plasma membrane are far from being understood and, hence, it is clear that mutations could also affect these distal steps of activation. Most of our understanding of the function of mutations stems from expression systems where it is a salient assumption that channel regulation represents that of the intact and diseased tissue. Chromosomal location and CF gene The chromosomal location has been discussed in the introduction to this section. Even now, after having knowledge of the gene and of the very many mutations, it cannot be ruled out completely that other genes are involved in this disease. Evidence comes from unexpected deviations from expected genotype±phenotype correlations and from the fact that the pathophysiology of CFTR does not completely explain the generally altered mucin secretion (Fanen et al., 1997; Quinton, 1999). Genotype±phenotype correlations The clinical picture of CF can vary widely. It spans from very mild forms with borderline sweat test and little other pathology to very severe cases with pancreatic involvement and early colonization of the airways by Pseudomonas or other dif®cult to treat bacteria such as B. Cepacia. It has been claimed that DF508 goes along with severe disease and that other mutations such as G551D, A455E, R117H; splice mutations or defects in channel function like R347P are mild mutations (Pilewski and Frizzell, 1999). These classi®cations, although predicted on a functional basis, may be not precise enough for several reasons: (1) the individual patient may have more or less pathology depending on other individual factors; (2) the functional parameters used for these classi®cations may not be accurate enough (tests of lung function; nasal transepithelial voltage; tests of pancreatic function). The recently improved methods to examine colonic biopsies may provide more accuracy; (3) polymorphisms in the intron 8 splice acceptor site may lead to altered exon 9 splicing and hence to non-functional CFTR; (4) other factors may control CFTR function and even modulation by other genes cannot be ruled out. It has been shown recently that pancreatic insuf®ciency caused by C225R can go along with a mild lung phenotype (Fanen et al., 1997). It does not appear prudent to draw such correlations prematurely on an individual basis, and even to inform parents and patients,

Cystic ®brosis

243

before more is known on such correlations with more sensitive tests and before the genetic control of CFTR expression, its maturation and processing is better understood. Pathogenesis It has already been discussed above extensively that CF is caused by mutations in CFTR. The knowledge that these mutations should lead to defects in Cl 2 transport is based on the above quoted studies by Schulz and later by Quinton (Quinton, 1983; Schulz, 1969). When patch clamp techniques became available and were applied to airway and colonic epithelia it was initially postulated by some investigators that the Cl 2 channel responsible for this disease was of intermediate conductance (named outwardly rectifying Cl 2 channel (ORCC) or intermediate conductance outwardly rectifying Cl 2 channel (ICOR)) and this view is still proposed by some investigators (Frizzell et al., 1986; Li et al., 1988; Sheppard and Welsh, 1999). Our data have argued against this view because the incidence of ICOR channels was equally frequent in CF and normal respiratory cells and these channels were only activated after excision (Kunzelmann et al., 1989). After the cloning of CFTR it became clear that the Cl 2 channels corresponding to CFTR had properties distinct from ICOR or ORCC. In several expression systems CFTR produced a cAMP activated linear Cl 2 current with a single channel conductance of 4±12 pS (Greger, 1996; Sheppard and Welsh, 1999). CFTR has since been studied intensively and much is known about the functional relevance of the ®ve domains of this molecule (Fig. 6). In the following the function of CFTR will be discussed in some detail. Disturbances on molecular level: CFTR CFTR, general properties Fig. 6 depicts the putative structure of CFTR as it has been proposed in the early studies by Riordan et al. (1989). This general scheme has held up extremely well for 10 years, but, of course, structural studies are not available for the entire molecule but only for small components thereof. The molecule consists of 1440 amino acids. Both, the N- and the C-termini are on the cytosolic side. The molecule has internal symmetry with two sets of 6 a-helical membrane spanning domains. Between these two domains a nucleotide binding (NBF1) and a regulatory domain (R) are interposed. The C-terminus carries another nucleotide binding domain (NBF2). This general structure is not at all related to what we know from other ion channels and it resembles the general structure of P-glycoproteins such as the multi drug resistant protein (pump) or ATP binding cassette (ABC) proteins (Gadsby and Nairn, 1999; Sheppard and Welsh, 1999). It was, as it turns out, very wise to give this protein the name cystic ®brosis transmembrane conductance regulator (CFTR)

244

R. Greger

and not simply CF Cl 2 channel, because the function of this molecule is more complex than that of other ion channels. It functions as a phosphorylation controlled ATP hydrolyzing ion transporter which in addition regulates other membrane transporters such as other Cl 2 channels, K 1 channels, water channels, Na 1 channels and probably even more (Dawson et al., 1999; Gadsby and Nairn, 1999; Kunzelmann, 1999; Schwiebert et al., 1999; Sheppard and Welsh, 1999). CFTR when expressed e.g. in Xenopus oocytes, Sf9 insect cells, Hela cells, CHO cells, ®broblasts, etc. or when examined in normally CFTR expressing cells produces an almost linear anion current with a conductance selectivity sequence: Br 2 $ Cl 2 . I 2 . F 2 . gluconate but a permeability sequence of PI . PBr . PCl . PF. Gluconate and I 2 block Cl 2 currents. The single channel conductance has been a matter of some controversy. With symmetrical Cl 2 concentrations of 150 mmol/l most studies report values between 6 and 12 pS, but smaller values have also been postulated. In cell attached patches there is some outward recti®cation, but in excised patches the current±voltage curves are linear. There are no speci®c inhibitors for CFTR. Diphenylaminocarboxylate (DPC) and glibenclamide have been used in very high concentrations. The same holds for the stilbenes DIDS and DNDS. None of these compounds are able to ®ngerprint these channels and to distinguish them from others (Cabantchik and Greger, 1992; Schultz et al., 1999). The activity of these channels can be enhanced by protein kinase A (PKA). The respective sites of phosphorylation are not known, but many putative sites are present in the R domain (Fig. 6). Phosphorylation by protein kinase C can also increase channel activity. Also a regulatory role of protein kinase GII (cGMP controlled), calmodulin dependent kinase and tyrosine kinase has been postulated (Gadsby and Nairn, 1999; Sheppard and Welsh, 1999). The NBF1 and NBF2 of CFTR apparently work as ATPases and control the gating. As a working model it has been proposed that sequential ATP hydrolysis at NBF1 and NBF2 controls the opening and the closure of this channel, respectively (Gadsby and Nairn, 1999). Many site directed mutagenesis studies have been performed to identify the pore region. A plethora of data points to an important role of the 6th membrane spanning the a helix (Fig. 6). With some mutations in this transmembrane domain the ion selectivity could be changed. In other studies this could not be veri®ed (Anderson et al., 1991; Hipper et al., 1995; Dawson et al., 1999). It is important to keep in mind that all these studies make the assumption that, all other things staying constant, the charge pattern of the pore is changed. This is, however, not conclusive because any speci®c change in one amino acid may have indirect effects by rearranging the secondary and tertiary structure of this protein. Spatial analysis of CFTR awaits studies with other technologies.

Cystic ®brosis

245

Activation of CFTR As discussed above the gating of CFTR requires the phosphorylation of the R domain by PKA. cAMP and PKA have been proposed to have additional activating effects by enhancing the density of CFTR molecules in the plasma membrane. This could be caused by the exocytosis of membrane vesicles containing CFTR. In fact, some data point to such an effect (Bradbury, 1999). However, a recent reexamination with a modi®ed and very sensitive patch clamp method to monitor membrane capacitance has revealed that the changes in membrane capacitance in CFTR expressing cells produced by cAMP were very small in contrast to the large conductance changes observed in the same cells (Hug et al., 1997; Greger et al., 1998). The fact that no net changes in membrane capacitance, i.e. membrane area, were seen in

Fig. 6. Putative structure of CFTR. The molecule consists of ®ve domains: two sets of transmembrane (TM) domains with a helices; two nucleotide binding folds (NBF1, NBF2); and one regulatory domain (R). Both, the amino and the carboxy terminus are on the cytosolic side. The numbers refer to the amino acids. The negative and positive charges of the TM domains are shown. Some of these charges have been shown to be of importance for pore function (especially in the sixth a helix). A and C are phosphorylation sites in the R domain. NBF1 and NBF2 function as ATPases (Riordan et al., 1989; Dawson et al., 1999).

246

R. Greger

these studies does not exclude that CFTR is still rearranged in the plasma membrane and made accessible for antibodies by cAMP mediated regulation. CFTR a regulator of membrane function This issue has come up in the previous sections. CFTR has been shown (a) to control membrane traf®cking (Bradbury, 1999); (b) to regulate other Cl 2 channels; (c) to alter the properties of ROMK K 1 channels; (d) to activate KVLQT-K 1 channels; (e) to activate aquaporin 3 water channels; (f) to increase ATP export from the cell; and (g) to inhibit epithelial (amiloride inhibited) Na 1 channels (ENaC) (McNicholas et al., 1994; Stutts et al., 1995; Kunzelmann et al., 1997; Mall et al., 1998a; Bradbury, 1999; Kunzelmann, 1999; Schreiber et al., 1999). All these regulatory effects of CFTR have been shown in heterologous expression systems. It is therefore pertinent to question whether all these functions are of physiological or pathophysiological relevance or whether they are caused, at least in part, by the arti®cial conditions of the respective expression systems. The role of CFTR as a cAMP regulated ATP pump has been discussed controversially. Today it appears likely that CFTR itself does not have such an effect (Grygorczyk and Hanrahan, 1997). For some of these regulatory effects, however, it has been shown beyond doubt that they are also present in intact cells and that they are of functional relevance. In this context the effect of CFTR on ENaC and water channels will be discussed in more detail. CFTR as a regulator of water channels Originally, it has been shown that Xenopus oocytes expressing CFTR have an increased water permeability, which was ascribed to CFTR itself (Hasegawa et al., 1992). We have then provided evidence that this is not the case and that these oocytes possess an endogenous aquaporin 3 (AQP3) which is upregulated by cAMP activated CFTR but not by mutated CFTR. In respiratory epithelial cells AQP3 and CFTR coexist, and hence it is likely that the upregulation of water permeability by activation of CFTR is of functional relevance for water transport and that this function is disturbed in CF (Schreiber et al., 1999). CFTR as a regulator of ENaC Na 1 channels Boucher and his group have postulated for decades that Na 1 channels of the luminal membrane of the respiratory tract are upregulated in CF patients (Boucher et al., 1983). Several years ago they showed in heterologous expression systems that ENaC channels are downregulated by activated CFTR (Grubb and Boucher, 1999). Subsequently it has been shown that this mechanism of CFTR dependent downregulation of ENaC channels is of importance for the determination of the direction of net transport in mid crypt cells of the colon and that this channel interaction can be shown in the respiratory tract and is defective in CF (Greger et al., 1997; Grubb

Cystic ®brosis

247

and Boucher, 1999; Kunzelmann, 1999). In other words, in CF patients the ENaC channels are permanently active and not downregulated by cAMP producing agonists, which otherwise prime the respiratory tract for NaCl and water secretion. Therefore, in terms of net balance, the respiratory epithelium of CF patients fails to respond to secretagogues andhas an increased rate of absorption, resulting in a marked attenuation of the mucociliary clearance. These results are of importance from a pathophysiological point of view and they are also intriguing with respect to the underlying mechanism. It has been speculated (1) that CFTR and ENaC interact directly, but it has also been proposed (2) that this effect is caused by another yet unknown protein binding to both channel proteins or (3) that the cytosolic Cl 2 or Na 1 concentrations mediate this crosstalk via ion sensor proteins (Kunzelmann, 1999). CFTR and respiratory tract function There is no doubt that altered NaCl and water transport in airway epithelia is the key pathophysiological component causing airway infections in CF patients. The detailed mechanisms of increased susceptibility, however, are still not clear. The key issue is the colonization of airways by pathogens. It has also been proposed that defective endosomal CFTR does not support luminal acidi®cation and that, therefore, mucins and glycoproteins destined for the luminal membrane are not correctly glycosylated and sialylated but oversulfated and fucosylated instead. This is supposed to increase the binding of pathogens such as Pseudomonas aeruginosa to the airway surfaces, thus facilitating colonization (Bradbury, 1999). The binding of pathogens then induces local in¯ammatory responses which lead to the destruction of respiratory epithelial cells. Moreover it has been proposed that the NaCl concentration in the surface ¯uid plays an important role in colonization, with lower concentrations preventing and high concentrations facilitating colonization (Smith et al., 1996). Whether this hypothesis is of relevance depends on the accurate measurement of NaCl concentration in the airway surface ¯uid of normal individuals and CF patients. The reported results are controversial at this stage. Correlation to previous knowledge of pathomechanism The detailed knowledge of the gene defect in CFTR and the studies on CFTR function have changed our view of the disease CF considerably. The disturbed function of sweat glands, the intestine, exocrine glands, bile ducts and gallbladder, the reproductive organs and of the airways can now be explained. Complex disturbances, however, such as the altered mucin secretion and the colonization of airways are still far from being understood. Other aspects, of which we are only aware of since the molecular identi®cation and organ localization of CFTR, are also still

248

R. Greger

entirely unclear. For instance, why is kidney function not altered directly and what role does CFTR play in the heart? The diagnostic strategies have been improved substantially due to the molecular de®nition of the disease. In addition, the functional tests have been improved. As a result the correct diagnosis is made much earlier today. This has a large impact on the therapeutic outcome. On the one hand, the identi®cation of CFTR by itself thus far has had no direct therapeutic impact. Initial attempts of gene therapy were not very promising and many methodological problems will have to be solved before this approach will become of any relevance. On the other hand, established strategies have been modi®ed successfully, and this, together with the early diagnosis, has increased life expectancy considerably. Additional strategies have been developed, e.g. heart± lung and lung transplantation. Further `traditional' therapeutic strategies are under development. The key issues in this area are the enhancement of airway mucociliary clearance, the improvement of CFTR maturation and traf®cking to the plasma membrane, the activation of residual CFTR, the activation of other Cl 2 channels, the increase in driving force for Cl 2 exit across the luminal membrane, the inhibition of ENaC channels in the luminal membrane, etc. Comparison with animal models Knockout mice were generated only 3 years after the molecular identi®cation of CFTR (Grubb and Boucher, 1999). Meanwhile ten models are available. Six CFTR are knocked out and four are de®ned mutations (DF508 and G551D). These mouse models have been extremely helpful in examining the pathophysiology in detail. In turned out that most of these mice have a reduced life expectancy. They usually die from gastrointestinal obstruction by mucin plugs. The intestine shows all the expected changes. In the respiratory tract no obvious pathology was found, even though the basic alteration in transepithelial nasal voltage was easily demonstrable. One might argue that intestinal pathology is so marked that the animals do not live long enough to develop airway pathology. A transgenic approach introducing human CFTR under an intestine speci®c promoter has been used in a CF mouse to improve intestinal function and to be able to study airway pathology in detail. Likewise, for the respiratory tract, pancreatic function was not grossly altered in most mouse models. The mouse models have also been and will be relevant for the examination of strategies of gene therapy. Perspectives Given the speed with which CF research has progressed during the past 15 years I have little doubt that many questions which await clari®cation today will be

Cystic ®brosis

249

answered within the next few years. CFTR will also serve as a paradigm of successful modern research in many areas: at the molecular level, on the basic mechanisms of the regulation of this protein, on new therapeutic strategies including gene therapy and at the level of the patient. New therapeutic approaches It has been discussed above that recent developments in the CF ®eld may permit the design of new strategies. Given the predicted speed of success, drug therapy at several levels (cf. above) will probably be available before substantial progress in the area of gene therapy will be made. Future research aims It has become clear in this brief review that several areas of research deserve closer attention. (1) At the genetic level additional mutations will have to be identi®ed and, even more importantly, the substantial gap between the incidence of the disease and the explanation on the basis of mutations, which still amounts to 20%, will have to be closed. (2) From a clinical point of view the correlation between genotype and phenotype is still not understood. Why is this the case? Is our de®nition of the genotype insuf®cient, or is our functional characterization inaccurate? (3) The time point of diagnosis and the conservative strategies of treatment have to be further improved. (4) New strategies for gene therapy will have to be developed using appropriate animal models (CF sheep?). (5) From a pharmacological point of view new compounds will have to be developed which interact more speci®cally with CFTR. The known compounds or their route of application which apparently have some effect on airway Cl 2 secretion or Na 1 absorption in CF patients have to be improved to become more effective. The mechanisms of CFTR maturation have to be characterized in more detail and possible therapeutic strategies have to be developed at this level, because the most frequent mutation, DF508, gets stuck and is degraded during maturation. (6) The CFTR molecule has to be understood at a real structural level. This will require biophysical methodologies to de®ne the accurate tertiary structure and the effect which certain mutations have on this spatial structure. (7) The detailed mechanisms of how this funny channel-ATPase functions will have to be further elaborated. (8) Finally the mechanisms of regulation of CFTR and that of CFTR on other membrane proteins will have to be de®ned in much more detail.

250

R. Greger

Acknowledgements The work in the author's laboratory has been continuously supported by Deutsche Forschungsgemeinschaft: GR 480. References Anderson, M.P., Gregory, R.J., Thompson, S., Souza, D.W., Paul, S., Mulligan, R.C., Smith, A.E., Welsh, M.J., 1991. Demonstration that CFTR is a chloride channel by alteration of its anion selectivity. Science 253, 202±205. Boucher, R.C., Knowles, M.R., Stutts, M.J., Gatzy, J.T., 1983. Epithelial dysfunction in cystic ®brosis lung disease. Lung 161, 1±17. Bradbury, N.A., 1999. Intracellular CFTR: localization and function. Physiol. Rev. 79, S175±S191. Cabantchik, Z.I., Greger, R., 1992. Chemical probes for anion transporters of mammalian cell membranes. Am. J. Physiol. 262, C803±C827. Chinet, T., Fouassier, L., Dray-Charier, N., Imam-Ghali, M., Morel, H., Mergey, M., Dousser, B., Parc, R., Paul, A., Housset, C., 1999. Regulation of electrogenic anion secretion in normal and cystic ®brosis gallbladder mucosa. Hepatol. 29, 5±13. Cohen, B.E., Lee, G., Jacobson, K.A., Kim, Y.C., Huang, Z., Sorscher, E.J., Pollard, H.B., 1997. 8Cyclopentyl-1,3-dipropylxanthine and other xanthines differentially bind to the wild-type and DF508 ®rst nucleotide binding fold (NBF-1). domains of the cystic ®brosis transmembrane conductance regulator. Biochemistry 36, 6455±6461. Collins, F.S., 1992. Cystic ®brosis: molecular biology and therapeutic implications. Science 256, 774±779. Colombo, C., Battezzati, P.M., Strazzabosco, M., Podda, M., 1998. Liver and bilary problems in cystic ®brosis. Semin. Liver Dis. 18, 227±235. Cuthbert, A.W., Halstead, J., Ratcliff, R., Colledge, W.H., Evans, M.J., 1995. The genetic advantage hypothesis in cystic ®brosis heterocygotes: a murine study. J. Physiol. (Lond.) 482 (2), 449±454. Dawson, D.C., Smith, S.S., Mansoura, M.K., 1999. CFTR: mechanism of anion conduction. Physiol. Rev. 79, S47±S75. di Sant'Agnese, P.A., Darling, R., Perera, G., 1953. Abnormal electrolyte composition of sweat in cystic ®brosis of the pancreas. Pediatrics 12, 549±563. Ecke, D., Bleich, M., Greger, R., 1996. The amiloride inhibitabe Na 1 conductance of rat colonic cells is suppressed by forskolin. P¯ug. Arch. Eur. J. Physiol. 431, 984±986. Fanen, P., Labarthe, R., Granier, F., Benharouga, M., Goossens, M., Edelman, A., 1997. Cystic ®brosis phenotype associated with pancreatic insuf®ciency does not always re¯ect the cAMP-dependent chloride conductive pathway defect. Analysis of C225R-CFTR and R1066C-CFTR. J. Biol. Chem. 272, 30563±30566. Frizzell, R.A., Rechkemmer, G., Shoemaker, R.L., 1986. Altered regulation of airway epithelial cell chloride channels in cystic ®brosis. Science 233, 558±560. Gadsby, D.C., Nairn, A.C., 1999. Control of CFTR channel gating by phosphorylation and nucleotide hydrolysis. Physiol. Rev. 79, S77±S107. Gadsby, D.C., Nagel, G., Hwang, T.C., 1995. The CFTR chloride channel of mammalian heart. Annu. Rev. Physiol. 57, 387±416. Gray, M.A., Winpenny, J.P., Porteous, D.J., Dorin, J.R., Argent, B.E., 1994. CFTR and calcium-activated chloride currents in pancreatic duct cells of a transgenic CF mouse. Am. J. Physiol. 266, C213±C221. Greger, R., 1995. The formation of sweat. In: Greger, R., Windhorst, U. (Eds.). Human Physiology, from Cellular Mechanisms to Integration, Springer-Verlag, New York, pp. 2219±2228.

Cystic ®brosis

251

Greger, R., 1996. The membrane transporters regulating epithelial NaCl secretion. P¯ug. Arch. Eur. J. Physiol. 432, 579±588. Greger, R., Bleich, M., Leipziger, J., Mall, M., Ecke, D., Kunzelmann, K., 1997. Regulation of ion transport in colonic crypts. News Physiol. Sci. 12, 62±66. Greger, R., Thiele, I., Warth, R., Bleich, M., 1998. Does stimulation of NaCl secretion in in vitro perfused rectal gland tubules of Squalus acanthias increase membrane capacitance?. P¯ug. Arch. Eur. J. Physiol. 436, 538±544. Gribkoff, V.K., Champigny, G., Barbry, P., Dworetzky, S.I., Meanwell, N.A., Lazdunski, M., 1994. The substituted benzimidazolone NS004 is an opener of the cystic ®brosis chloride channel. J. Biol. Chem. 269, 10983±10986. Grubb, B.R., Boucher, R.C., 1999. Pathophysiology of gene-targeted mouse models for cystic ®brosis. Physiol. Rev. 79, S193±S214. Grygorczyk, R., Hanrahan, J.W., 1997. CFTR-independent ATP release fromepithelial cells triggered by mechanical stimuli. Am. J. Physiol. 272, C1058±C1066. Hasegawa, H., Skach, W., Baker, O., Calayag, M.C., Lingappa, V., Verkman, A.S., 1992. A multifunctional aqueous channel formed by CFTR. Science 258, 1477±1479. Hill, W.G., Harper, G.S., Rozaklis, T., Boucher, R.C., Hopwood, J.J., 1997. Organ-speci®c over-sulfation of glycosaminoglycans and altered extracellular matrix in a mouse model of cystic ®brosis. Biochem. Mol. Med. 62, 113±122. Hipper, A., Mall, M., Greger, R., Kunzelmann, K., 1995. Mutations in the putative pore-forming domain of CFTR do not change anion selectivity of the cAMP activated Cl 2 conductance. FEBS Lett. 374, 312±316. Hug, M.J., Thiele, I., Greger, R., 1997. The role of exocytosis in the activation of the chloride conductance in Chinese hamster ovary cells (CHO). Stably expressing CFTR. P¯ug. Arch. Eur. J. Physiol. 434, 779±784. Illek, B., Fischer, H., Santos, G.F., Widdicombe, J.H., Machen, T.E., Reenstra, W.W., 1995. cAMPindependent activation of CFTR Cl channels by the tyrosine kinase inhibitor genistein. Am. J. Physiol. 268, C886±C893. Imundo, L., Barasch, J., Prince, A., Al-Awqati, Q., 1995. Cystic ®brosis epithelial cells have a receptor for pathogenic bacteria on their apical surface. Proc. Natl. Acad. Sci. USA 92, 3019±3023. Jiang, C., Fang, S.L., Xiao, Y.F., O'Connor, S.P., Nadler, S.G., Lee, D.W., Jefferson, D.M., Kaplan, J.M., Smith, A.E., Cheng, S.H., 1998. Partial restoration of cAMP-stimulated CFTR chloride channel activity in DF508 cells by deoxyspergualin. Am. J. Physiol. 275, C171±C178. Kerem, B.-S., Rommens, J.M., Buchanan, J.A., Markiewicz, D., Cox, T.K., Chakravarti, A., Buchwald, M., Tsui, L.-C., 1989. Identi®cation of the cystic ®brosis gene: genetic analysis. Science 245, 1073± 1080. Knowles, M.R., Gatzy, J.T., Boucher, R.C., 1981. Increased biolelectric potential difference across respiratory epithelia in cystic ®brosis. N. Engl. J. Med. 305, 1489±1495. Knowles, M.R., Clarke, L.L., Boucher, R.C., 1991. Activation by extracellular nucleotides of chloride secretion in the airway epithelia of patients with cysic ®brosis. N. Engl. J. Med. 325, 533±538. Kopito, R.R., 1999. Biosynthesis and degradation of CFTR. Physiol. Rev. 79, S167±S173. KoÈttgen, M., Busch, A.E., Hug, M.J., Greger, R., Kunzelmann, K., 1996. N-Acetyl-L-cysteine and its derivatives activate a Cl 2 conductance in epithelial cells. P¯ug. Arch. Eur. J. Physiol. 431, 549±555. Kunzelmann, K., 1999. CFTR. Rev. Physiol. Biochem. Pharmacol. 137, 1±70. Kunzelmann, K., PavenstaÈdt, H., Greger, R., 1989. Properties and regulation of chloride channels in cystic ®brosis and normal airway cells. P¯ug. Arch. Eur. J. Physiol. 415, 172±182. Kunzelmann, K., Mall, M., Briel, M., Hipper, A., Nitschke, R., Ricken, S., Greger, R., 1997. The cystic ®brosis transmembrane conductance regulator attenuates the endogenous Ca 21 activated Cl 2 conductance of Xenopus oocytes. P¯ug. Arch. Eur. J. Physiol. 435, 178±181.

252

R. Greger

Kunzelmann, K., Briel, M., Schreiber, R., Ricken, S., Nitschke, R., Greger, R., 1998. No evidence for direct activation of the cystic ®brosis transmembrane regulator by 8-cyclopentyl-1,3-dipropylxanthine. Cell Physiol. Biochem. 8, 185±193. Lee, M.G., Wigley, W.C., Zeng, W., Noel, L.E., Marino, C.R., Thomas, P.J., Muallem, S., 1999. Regulation of Cl 2/HCO32 exchange by cystic ®brosis transmembrane conductance regulator expressed in NIH 3T3 and HEK 293 cells. J. Biol. Chem. 274, 3414±3421. Letz, B., Korbmacher, C., 1997. cAMP stimulates CFTR-like Cl 2 channels and inhibits amiloride-sensitive Na 1 channels in mouse CCD cells. Am. J. Physiol. 272, C657±C666. Leung, A.Y.H., Wong, P.Y.D., Yankaskas, J.R., Boucher, R.C., 1996. cAMP- but not Ca 21-regulated Cl 2 conductance is lacking in cystic ®brosis mice epididymides and seminal vesicles. Am. J. Physiol. 271, C188±C193. Li, M., McCann, J.D., Liedtke, C.M., Nairn, A.C., Greengard, P., Welsh, M.J., 1988. Cyclic AMPdependent protein kinase opens chloride channels in normal but not cystic ®brosis airway epithelium. Nature 331, 358±360. Mall, M., Hipper, A., Greger, R., Kunzelmann, K., 1996. Wild-type CFTR but not DF508 inhibits Na 1 channels in Xenopus oocytes. FEBS Lett. 381, 47±52. Mall, M., Bleich, M., Greger, R., Schreiber, R., Kunzelmann, K., 1998a. The amiloride-inhibitable Na 1 conductance is reduced by the cystic ®brosis transmembrane conductance regulator in normal but not in cystic ®brosis airways. J. Clin. Invest. 102, 15±21. Mall, M., Bleich, M., Schuerien, M., Kuehr, J., Seydewitz, H.H., Brandis, M., Greger, R., Kunzelmann, K., 1998b. Cholinergic ion secretion in human colon requires coactivation by cAMP. Am. J. Physiol. 275 (6 Pt 1), G1274±G1281. Mall, M., Wissner, A., Seydewitz, H.H, KuÈhr, J., Brandis, M., Greger, R., Kunzelmann, K., 2000. Defective cholinergic Cl 2 secretion and detection of K(+) secretion in rectal biopsies from cystic ®brosis patients. Am. J. Physiol. Gastrointest. Liver Physiol. 278(4), G617±G624. Manson, A.L., Trezise, A.E., Mac Vinish, L.J., Kasschau, K.D., Birchall, N., Episkopou, V., Vassaux, G., Evans, M.J., Colledge, W.H., Cuthbert, A.W., Huxley, C., 1997. Complementation of null CF mice with a human CFTR YAC transgene. EMBO J. 16, 4238±4249. McNicholas, C.M., Wang, W., Ho, K., Hebert, S.C., Giebisch, G., 1994. Regulation of ROMK1 K 1 channel activity involves phosphorylation processes. Proc. Natl. Acad. Sci. USA 91, 8077±8081. Montrose-Ra®zadeh, C., Guggino, W.B., 1990. Cell volume regulation in the nephron. Annu. Rev. Physiol. 52, 761±772. Novak, I., Greger, R., 1988. Properties of the luminal membrane of isolated perfused rat pancreatic ducts. P¯ug. Arch. Eur. J. Physiol. 411, 546±553. Peters, R.H., van Doorninck, J.H., French, P.J., Ratcliff, R., Evans, M.J., Colledge, W.H., Bijman, J., Scholte, B.J., 1997. Cystic ®brosis transmembrane conductance regulator mediates the cyclic adenosine monophosphate-induced ¯uid secretion but not the inhibition of resorption in mouse gallbladder epithelium. Hepatology 25, 270±277. Pier, G.B., Grout, M., Zaidi, T., Meluleni, G., Mueschenborn, S.S., Banting, G., Ratcliff, R., Evans, M.J., Colledge, W.H., 1998. Salmonella typhi uses CFTR to enter intestinal epithelial cells. Nature 393, 79± 82. Pilewski, J.M., Frizzell, R.A., 1999. Role of CFTR in airway disease. Physiol. Rev. 79, S215±S255. Porteous, D.J., Dorin, J.R., McLachlan, G., Davidson-Smith, H., Davidson, H., Stevenson, B.J., Carothers, A.D., Wallace, W.A., Moralee, S., Hoenes, C., Kallmeyer, G., Michaelis, U., Naujoks, K., Ho, L.P., Samways, J.M., Imrie, M., Greening, A.P., Innes, J.A., 1997. Evidence for safety and ef®cacy of DOTAP cationic liposome mediated CFTR gene transfer to the nasal epithelium of patients with cystic ®brosis. Gene Ther. 4, 210±218. Quinton, P.M., 1983. Chloride impermeability in cystic ®brosis. Nature 301, 421±422.

Cystic ®brosis

253

Quinton, P.M., 1999. Physiological basis of cystic ®brosis: a historical perspective. Physiol. Rev. 79, S3± S22. Reid, C.J., Hyde, K., Ho, S.B., Harris, A., 1997. Cystic ®brosis of the pancreas; involvement of MUC6 mucin in obstruction of pancreatic ducts. Mol. Med. 3, 403±411. Rich, D.P., Anderson, M.P., Gregory, R.J., Cheng, S.H., Paul, S., Jefferson, D.M., McCann, J.D., Klinger, K.W., Smith, A.E., Welsh, M.J., 1990. Expression of cystic ®brosis transmembrane conductance regulator corrects defective chloride channel regulation in cystic ®brosis airway epithelial cells. Nature 347, 358±363. Riordan, J.R., Rommens, J.M., Kerem, B.-S., Alon, N., Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.-L., Drumm, M.L., Iannuzzi, M.C., Collins, F.S., Tsui, L.-C., 1989. Identi®cation of the cystic ®brosis gene: cloning and characterization of complementary DNA. Science 245, 1066± 1073. Rossier, B., 1997. Homer Smith award lecture. Cum grano salis: the epithelial sodium channel and the control of blood pressure. J. Am. Soc. Nephrol. 8, 980±992. Rozmahel, R., Wilschanski, M., Martin, A., Plyte, S., Oliver, M., Auerbach, W., Forstner, J., Durie, P., Nadeau, J., Bear, C., Tsui, L.-C., 1996. Modulation of disease severity in cystic ®brosis transmembrane conductance regulator de®cient mice by a secondary genetic factor. Nat. Genet. 12, 280±287. Schreiber, R., Nitschke, R., Greger, R., Kunzelmann, K., 1999. CFTR activates aquaporin 3 in airway epithelial cells. J. Biol. Chem. 274(17), 11811±11816. Schulz, I., 1969. Micropuncture studies of sweat formation in cystic ®brosis patients. J. Clin. Invest. 48, 1470±1477. Schultz, B.D., Singh, A.K., Devor, D.C., Bridges, R.J., 1999. Pharmacology of CFTR chloride channel activity. Physiol. Rev. 79, S109±S144. Schwiebert, E.M., Benos, D.J., Egan, M.E., Stutts, M.J., Guggino, W.B., 1999. CFTR is a conductance regulator as well as a chloride channel. Physiol. Rev. 79, S145±S166. Sheppard, D.N., Welsh, M.J., 1999. Structure and function of the CFTR chloride channel. Physiol. Rev. 79, S109±S144. Simon, D.B., Bindra, R.S., Mans®eld, T.A., Nelson-Williams, C., Mendonca, E., Stone, R., Schurman, S., Nayir, A., Alpay, H., Bakkaloglu, A., Rodriguez-Soriano, J., Morales, J.M., Sanjad, S.A., Taylor, C.M., Pilz, D., Brem, A., Trachtman, H., Griswold, W., Richard, G.A., John, E., Lifton, R.P., 1997. Mutations in the chloride channel gene, CLCNKB, cause Bartter's syndrome type III. Nat. Genet. 17, 171±178. Smith, J.J., Travis, S.M., Greenberg, E.P., Welsh, M.J., 1996. Cystic ®brosis airway epithelia fail to kill bacteria because of abnormal airway surface ¯uid. Cell 85, 229±236. Stutts, M.J., Chinet, T.C., Mason, S.J., Fullton, J.M., Clarke, L.L., Boucher, R.C., 1992. Regulation of Clchannels in normal and cystic ®brosis airway epithelial cells by extracellular ATP. Proc. Natl. Acad. Sci. USA 89, 1621±1625. Stutts, M.J., Canessa, C.M., Olsen, J.C., Hamrick, M., Cohn, J.A., Rossier, B., Boucher, R.C., 1995. CFTR as a CAMP-dependent regulator of sodium channels. Science 269, 847±850. Zdebik, A., Hug, M.J., Greger, R., 1997. Chloride channels in the luminal membrane of rat pancreatic acini. P¯ug. Arch. Eur. J. Physiol. 434, 188±194.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

255

Chapter 10

Dent's disease: an hereditary nephrolithiasis caused by dysfunction of a voltage-gated chloride channel Christoph Fahlke Institut fuÈr Physiologie, RWTH Aachen, Pauwelsstrasse 30, D-52057 Aachen, Germany Centro de Estudios Cienti®cos (CECS), Avenida Prat 514, Valdivia, Chile

Abstract Dent's disease is an inherited human kidney disease associated with low-weight proteinuria, hypercalciuria and nephrolithiasis. The Dent's disease gene encodes a kidney-speci®c chloride channel (ClC-5) that is a member of the ClC-family. The exact role of chloride channel dysfunction in the pathophysiology of Dent's disease is not completely understood, but several lines of evidence support the following hypothetical pathomechanism. ClC-5 is believed to be an intracellular chloride channel in endosomes of the proximal tubule that is necessary for the pH adjustment in these cell organelles. Its dysfunction causes defective endosomal acidi®cation and interferes with absorptive endocytosis and degradation of lowmolecular weight proteins thus causing proteinuria. Hypercalciuria is presently unexplained. Understanding the mechanisms underlying Dent's disease promises novel insights into the pathophysiology of nephrolithiasis, as well as into the physiology of renal tubular cells and the role of chloride channels in these processes. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Ion channels are critical for numerous cellular functions, such as excitability, synaptic transmission, muscle contraction, transepithelial transport, and cell migration. Alterations of ion channel function can affect organ functions and cause various human diseases. In recent years, an increasing number of `ion channelopathies', genetic diseases that are caused by mutations in genes coding for ion channels (Hoffman et al. (1995)), has been identi®ed. Channelopathies are interesting for several reasons. (1) Ion channels have been studied for decades and represent one of the best understood classes of proteins. It is likely that this knowledge can be used

256

Ch. Fahlke

to pharmacologically target certain genetically altered ion channels and thus establish rational pharmacological treatments of these diseases. (2) The ®nding that mutations in a certain channel protein result in organ dysfunction establishes the importance of this particular channel in vivo. The elucidation of the pathomechanism of this disease provides detailed insights into the physiological role of this particular channel. (3) The positional cloning of disease-causing genes has allowed de®nition of the primary sequences of important ion channel proteins, most notably CFTR (the cystic ®brosis transmembrane regulator), an epithelial anion channel involved in several cell functions and the KCNQ potassium channel subfamily, involved in inherited arrythmias, epilepsies and deafness. (4) Finally, naturally occurring mutations represent a powerful tool in the investigation of structure± function relationships of ion channels. Since these mutations cause a signi®cant defect in channel function, they are likely to be located in segments that are functionally important and their characterization will therefore provide important hints to the de®nition of the structural determinants of particular channel functions. Dent's disease is an X-linked recessively inherited kidney disease that was de®ned only recently. Between 1991 and 1993, four distinct clinical syndromes were described, in different countries, with low-weight proteinuria, hypercalciuria and nephrolithiasis. Although there are certain variations in the clinical symptoms, the ®nding that all these syndromes have the same gene locus established that they all represent a single disease, now referred to as Dent's disease (Akuta et al., 1997; Lloyd et al., 1997a,b). Positional cloning of the disease locus, by Dr. Rajesh Thakker's group in London (Fisher et al., 1994, 1995) identi®ed a gene (CLCN5) encoding an anion channel belonging to the ClC family of voltage-gated chloride channels (ClC-5). Shortly thereafter, several point mutations were identi®ed and functional characterization in heterologous expression systems demonstrated that these mutations alter ClC-5 channel function (Lloyd et al., 1996). These ®ndings established CLCN5 as the Dent's disease gene locus. Dent's disease is a fascinating example of an ion channelopathy as it demonstrates the variety of insights we can learn investigating this kind of disease. Understanding the pathomechanism of Dent's disease will illuminate previously unknown functional roles of ClC channels. It will provide novel insights into the molecular and cellular mechanisms of calcium and protein reabsorption in the kidney. Finally, it appears likely that disease-causing point mutations will help to identify structural determinants of novel ClC channel functions.

Clinical features In 1991, the evaluation of a large kindred with a familial nephrolithiasis associated with renal failure in the US allowed the de®nition of a novel kidney

Dent's disease

257

disease that was initially termed X-linked recessive nephrolithiasis (Frymoyer et al., 1991; Scheinman, 1998). The clinical ®ndings included low molecular weight proteinuria, proximal tubular dysfunction, calcium nephrolithiasis, nephrocalcinosis and progressive renal failure. Proteinuria and hematuria had a very early onset, in certain cases in the ®rst year of life and nephrolithiasis occurred as early as in the ®rst decade. Additional abnormalities of renal tubular function, such as polyuria, glucosuria, aminoaciduria, phosphaturia, kaliuresis and uricosuria were present to variable extent. Kidney dysfunction progressed in many, but not in all patients to renal failure and the severity of symptoms was quite variable among patients. All affected patients were male and the inheritance pattern was in agreement with an X-linked recessive inheritance mode. Female carriers had mild low-molecular weight proteinuria and sometimes hypercalciuria, but were never symptomatic. One year later, several families with similar symptoms were evaluated in Great Britain revealing an X-linked recessive genetic disorder with low molecular weight proteinuria, calciuria, nephrolithiasis, renal failure and childhood rickets (Wrong et al., 1992, 1994). The appearance of rickets or osteomalacia initially appeared to distinguish this disease from X-linked recessive nephrolithiasis. Among the studied families were the kindreds of two patients that were described in 1964 (Dent and Friedman, 1964) in the ®rst known clinical report about this disease; and the name Dent's disease was chosen for this apparently novel disease to honor the initial describer. Shortly thereafter, another disorder with similar symptoms but a more severe phenotype was de®ned in Italy and France and named X-linked recessive hypophosphatamic rickets (Bolino et al., 1993). In Japan, an annual urine-screening program of school and preschool children has been performed since 1974. This program allowed identi®cation of a number of patients with low molecular weight proteinuria accompanied with hematuria, glucosuria, aminoaciduria and hypophosphatamia and hypercalciuria. Based on differences in the clinical ®ndings, this syndrome was also believed to be a distinct disease and was called low-molecular weight proteinuria with hypercaliuria and nephrocalcinosis (Suzuki et al., 1985). Although these four syndromes initially appeared to be distinct, later genetic ®ndings demonstrated that all are caused by mutations in the same gene (CLCN5, see below) indicating that all the cases are the same disease entity and the name Dent's disease was proposed (Akuta et al., 1997; Lloyd et al., 1997b). The clinical symptoms as well as the pathological ®ndings indicate that these syndromes are disorders of tubular function. Dent's disease clearly is distinct from other forms of Fanconi's syndrome and is described by the following diagnostic criteria (Scheinman, 1998). (1) Proteinuria: Proteinuria is one of the primary abnormalities in Dent's disease. It is independent of infection or obstruction secondary to nephrolithiasis and is the

258

Ch. Fahlke

earliest and most consistent ®nding in patients with this disorder. The total excretion of proteins commonly equals to 0.5±2 g/day, 50±70% of which are low molecular weight proteins with a mass of less than 40 kDa such as b2 microglobulin, a1 microglobulin and lysozyme. In many patients, albuminuria can be observed, but this is believed to be caused by secondary glomerular damage. Nephrotic syndrome has not been observed in affected individuals. (2) Hypercalciuria, nephrocalcinosis and nephrolithiaisis: Prior to renal failure, nephrolithiasis is subjectively the most important clinical ®nding. Medullary nephrocalcinosis, identi®ed radiologically or histologically, can be observed in almost every affected individual. It is most signi®cantly caused by hypercalciuria and excretion of oxalate and citrate is either normal or slightly elevated. Renal excretion of calcium is very pronounced in children, up to 10 mg/kg body weight and tends to decrease in adulthood (4±6 mg/kg body weight). Hypercalciuria is present under dietary deprivation in most affected individuals and all patients show an increase of hypercalciuria after oral calcium loading. Hypercalciuria is associated with decreased levels of parathyriod hormone (PTH) and increased levels of 1,25 dihydrovitamin D. (3) Proximal tubule dysfunction: In addition to low-molecular weight proteinuria, other defects of proximal tubular function such as aminoaciduria, glucosuria, hypokalemia, hypophosphatemia and hyperuricosuria are observed in patients, although with less consistency. (4) Decreased urinary concentration: Many patients show moderate polyuria and relative resistance to vasopression. This abnormality is not caused by proximal tubular failure, as osmolar clearance is unaffected, but is most probably due to renal insuf®ciency or nephrocalcinosis. (5) Renal failure: The progression of renal failure is quite variable; some patients reach end-stage renal disease in their early adulthood, while others experience only a modest impairment for their whole life. Histological ®ndings are non-speci®c. Tubular atrophy and interstitial ®brosis, as well as glomerular hypertrophy or sclerosis, can be observed. (6) Osteomalacia and rickets: Bone disease occurs in only a minority of patients, mostly in Europe (Great Britain, Italy and France).

Therapy options At present, there are no established treatment procedures for Dent's disease. Clinical trials have not yet been performed due to the limited number of patients and the very recent identi®cation of this novel disease. The current understanding of the pathophysiology (see below) provides no logical strategies to reverse the cellular

Dent's disease

259

alterations caused by the genetic defect. Treatment of Dent's disease is therefore currently largely symptomatic (Scheinman, 1998). Hypercalciuria is the cause of nephrolithiasis and likely plays an important role in the development of end-stage renal disease in patients with Dent's disease. A current treatment concept is therefore to pharmacologically affect hypercalciuria with the therapeutic principles used for idiopathic hypercalciuria. Currently, thiazide diuretics are administered to decrease hypercalciuria by stimulating calcium reabsorption by the nephron. Although in certain patients thiazide diuretics cause signi®cant potassium wasting and one report describes hypokaliaemia and acute volume-depletion (Wrong et al., 1994), some patients tolerate diuretic treatment well and respond with a reduction of calcium excretion. Additional therapeutic strategies may be the use of amiloride, to increase calcium reuptake, dietary restrictions of sodium and meat and dietary protein restriction to slow the progression of renal failure (Wrong et al., 1994). Vitamin D has been suggested to treat osteomalacia (Wrong et al., 1994). The established therapies for renal failure, such as dialysis and kidney transplantation, can be successfully used for patients with Dent's disease after reaching end-stage renal disease.

Genetic ®ndings The clinical symptoms observed in Dent's disease are complex and it could not initially be explained by a single gene defect. It was therefore very dif®cult to de®ne candidate genes for this disorder. A breakthrough in our understanding of Dent's disease was the genetic evaluation of a large family of 102 members with X-linked nephrolithiasis by Scheinman and colleagues in 1993 (Scheinman et al., 1993). This work allowed linkage of this syndrome to the DXS255 locus with a LOD score of 5.91% at 3.6% recombination (Scheinman et al., 1993) and localized the disease gene to the pericentromeric region of the short arm of the X chromosome (Xp11.22). Next, Pook et al. (1993) mapped Dent's disease in ®ve unrelated British families to the same locus. In one family with Dent's disease, the genetic defect was subsequently identi®ed as a microdeletion in DXS225. The isolation of coding sequences falling within the deleted region and the subsequent sequencing revealed a gene that apparently coded for a novel ClC-type chloride channel (Fisher et al., 1994) and was termed CLCN5. The expression pattern and deletion mapping made this chloride channel gene a strong candidate for the Dent's disease gene (Fisher et al., 1994, 1995). The ®nal proof that this gene is the Dent's disease locus was the identi®cation of mutations in the CLCN5 gene. In 1995, Lloyd and colleagues reported mutations in 11 kindreds with three of the four related syndromes (Dent's disease, Xlinked recessive nephrolithiasis and X-linked recessive hypophosphatamic rickets)

260

Ch. Fahlke

and demonstrated, by heterologous expression in Xenopus oocytes, that these mutations affect chloride channel function (Lloyd et al., 1996). Shortly thereafter, a study conducted on Japanese kindreds with low-molecular weight proteinuria with hypercalciuria and nephrocalcinosis lead to similar results (Lloyd et al., 1997a).

Pathophysiology Although the ®nding that chloride channel dysfunction causes a renal syndrome characterized by hypercalciuria, nephrolithiasis and proteinuria was surprising at ®rst glance, a possible mechanism accounting for this ®nding was proposed very early. In the initial publication reporting the cloning of the Dent's disease gene, Fisher et al. (1994) reasoned that the affected chloride channel might play a role in endosomal acidi®cation and might thus be important for absorptive endocytosis in the proximal tubule. Hebert (1996) suggested that ClC-5 could play a role in neutralizing the charges resulting from active proton pumping by an H 1-ATPase into subcellular vesicles where low-molecular weight proteins are reabsorbed in the proximal tubules and degraded by an acid-dependent mechanism. In normal kidney function, the glomerular barrier prevents the ®ltration of large proteins such as albumin, but smaller proteins may readily pass. Low-molecular weight proteins are ®ltrated into the primary urine, then reabsorbed by proximal tubule cells by endocytosis (Wall and Maack, 1985; Christensen and Nielsen, 1991) and later enzymatically degraded in lysosomes. During reabsorption, proteins initially bind to sites on the luminal plasma membrane and subsequent invagination produces endocytotic vesicles. Endocytotic vacuoles subsequently fuse to secondary lysosomes where absorbed proteins are degraded (Christensen and Nielsen, 1991). All processes during protein internalization and degradation critically depend on an acidic environment in subcellular organelles (Gekle et al., 1995). The acidic pH inside endosomes is established by a proton pump, a H 1ATPase, that actively transports protons into these organelles. The transport is electrogenic and therefore capable of generating a positive membrane potential inside the organelle. In the absence of additional conductances, this potential would reach very high values and the transport of protons against this electrostatic ®eld would become energetically prohibitive (Al-Awqati et al., 1992). Chloride channels profoundly increase the conductance of this membrane and can therefore clamp the potential difference between vesicle inside and outside to zero. This mechanism is critical for the creation of the substantial proton gradients necessary for endosomal acidi®cation (Bae and Verkman, 1990; AlAwqati et al., 1992; Marshansky and Vinay, 1996). Chloride channel dysfunction

Dent's disease

261

will therefore likely result in hampered endosomal acidi®cation causing impaired reabsoptive protein endocytosis. This proposed pathomechanism can account for one of the most consistent and earliest ®ndings in patients with Dent's disease, the low-weight molecular proteinuria. In order to support this hypothesis, it was necessary to demonstrate that ClC-5 is indeed located in endosomes and that it colocalizes with a H 1-ATPase. A series of three subsequent papers demonstrated that the localization of ClC-5 is in agreement with this proposal. Using polyclonal ClC-5-speci®c antibodies, GuÈnther et al. (1998) investigated the tissue speci®city and subcellular localization of ClC-5 in rat. ClC-5 was shown to be highly kidney-speci®c and immunohistochemistry revealed that it is expressed throughout the entire proximal tubule (S1±S3 segment) and the intercalated cells of the collecting duct. Subcellularly, in the proximal tubule as well as in the intercalated cells, ClC-5 colocalizes with the H-ATPase and with internalized protein, thus strongly supporting a role of ClC-5 in acidi®cation and endocytosis. In another series of experiments, also with rat tissue, Luyckx et al. (1998) demonstrated a slightly different intrarenal localization. Again using immunohistochemistry, the authors showed that ClC-5 is only present in the S3 segment of the proximal tubules and the thick ascending limb (MTAL). An additional ¯ow cytometry analysis of subcellular membrane fractions again supported the notion that ClC-5 is present in the endosomal compartment. As rat and human kidneys differ in several aspects, it was important to reevaluate the localization of ClC-5 in human kidney. Devuyst et al. (1999) demonstrated expression in proximal renal tubule, the thick ascending limb of Henle and intercalated cells of the collecting duct and showed that ClC-5 associated with Rab4, a marker of recycling early endosomes. These studies consistently demonstrated an intracellular localization of ClC-5 and supported a role in absorptive endocytosis. (Fig. 1) An alteration of membrane recycling caused by the chloride channel dysfunction has been also suggested to account for other symptoms of Dent's disease. A defect in the endocytotic capability of the proximal tubule would affect endosomal recycling of various membrane proteins, such as transport proteins for calcium, glucose, amino acids and phosphate (Luyckx et al., 1998) thereby causing hypercalciuria, glucosuria, amino aciduria and phosphaturia. Devuyst et al. (1999) suggested a possible endocytotic pathway that reabsorbs calcium together with low molecular weight proteins. The receptor to which low molecular weight proteins bind prior to internalization by endocytosis is a glycoprotein, called megalin (Cui et al., 1996; Christensen et al., 1998). As it is known that megalin also binds calcium ions, the authors reasoned that calcium endocytosis in the proximale tubule could be impaired in patients with ClC-5 dysfunction due to the same mechanisms by which protein reabsorption is affected. Scheinman (1998) pointed out that certain features of the calcium metabolism in Dent's disease resemble idiopathic hypercalciuria. Patients exhibit low parathyroid hormone

262

Ch. Fahlke

Fig. 1. Current concept of the role of ClC-5 channels in endosomal vesicles of proximal tubule cells. A proton pump actively transports protons into the vesicle and chloride channels prevent depolarization of the intravesicular potential by dissipating the accumulated charge.

(PTH) and high 1,25 dihydroxyvitamin D levels which argues against a pathological tubular leak where a high PTH level would be expected. For these reasons, he speculated that hypercalciuria may be the consequence of an abnormal regulation of 1-hydroxylation of vitamin D1, likely due to a general dysfunction of the tubular cells (Scheinman, 1998). In summary, the current concept of the pathophysiology of Dent's disease suggests that the dysfunction of a vesicular chloride channel, ClC-5, causes defective reabsorptive protein endocytosis in the proximal tubule (Fig. 1). Due either to alteration in endocytosis or membrane recycling or because of a general dysfunction of the proximal tubule cells, increased levels of calcium are excreted in the urine and

Dent's disease

263

hypercalciuria causes nephrolithiasis. It is currently not clear whether renal failure is a consequence of renal damage by kidney stones or whether other factors contribute to end-stage renal disease.

The ClC family of voltage-gated chloride channels The ClC family was founded in 1990, as Dr. Thomas Jentsch and coworkers successfully isolated cDNAs encoding a voltage-gated chloride channel (ClC-0) from the electric organ of Torpedo marmorata (ClC-0) using a novel expression cloning technique (Jentsch et al., 1990). The subsequent description of homologous sequences in various organisms de®ned a novel gene family, the ClC-family of voltage-gated chloride channels. ClC channels are now known to be present in almost every living cell, from prokaryotes to mammalian tissues (Jentsch and GuÈnther, 1997), thus representing the largest known gene family coding for anion channels. At present, nine human isoforms are known (ClC-1±ClC-7 and ClC-Ka and ClC-Kb) (Fig. 2). A physiological role has been unambiguously established for only two human ClC channels. ClC-1 is the major muscle chloride channel (Steinmeyer et al., 1991) and is responsible for the high resting chloride conductance of the sarcolemma of adult muscle ®bers necessary for normal muscle excitation. Mutations in ClCN1

Fig. 2. Dendrogram of the nine known human members of the ClC family of voltage-gated chloride channels (Jentsch and GuÈnther, 1997).

264

Ch. Fahlke

cause myotonia congenita (Koch et al., 1992; George et al., 1993), a congenital human muscle disease characterized by muscle stiffness due to repetitive sarcolemmal action potentials. Mutations in hClCKb cause type III Bartter's syndrome, a genetic disease characterized by the inability of the kidney to concentrate urine (Simon et al., 1997). Based on its expression pattern along the nephron, hClCKb is believed to be the major basolateral chloride channel in the thick ascending limb of Henle. For all other human ClC-isoforms, the physiological role is unde®ned or remains a matter of scienti®c controversy. Among the mammalian ClC channels, only ClC-1, ClC-2, ClC-4 and ClC-5 can be reproducibly expressed in heterologous expression systems and have been shown to form anion-selective ion channels. For both ClC-5 in native tissue (see above) and ClC-6 in heterologous expression systems (Buyse et al., 1998) an intracellular location has been demonstrated. The investigation of mechanisms and structural determinants of speci®c ClC channel functions is still in its infancy and our knowledge of the basic properties of this type of ion channel is very limited especially when compared with voltagegated cation or ligand-gated channels. Experiments with ClC-0 (Ludewig et al., 1996, Middleton et al., 1996) and ClC-1 (Fahlke et al., 1997a) convincingly showed that functional ClC channels are formed by two subunits. In Xenopus oocytes, ClC-1 and ClC-2 have been demonstrated to form heterodimeric channels (Lorenz et al, 1996) and it is therefore generally assumed that distinct ClC isoforms can heterodimerize, though this has not yet been thoroughly tested. Several models for the transmembrane topology have been suggested (Middleton et al., 1994; Fahlke et al., 1997b; Schmidt-Rose and Jentsch, 1997), but currently no model can explain all of the experimental data. For ClC-1, a region between transmembrane domains D3 and D5 that comprises major determinants of ion selectivity was shown to form part of the ion conduction pathway (Fahlke et al., 1997b). Since this region is conserved in all known ClC channels, it is believed to form a major portion of the pore for all ClC channels. At present, there is controversy about the pore stoichiometry of ClC channels. Single-channel recordings of the prototypic ClC channel ClC-0 show two independently gated and equally spaced subconductance states. Experiments with heterodimeric ClC-0 channel constructs demonstrated that the amplitude and selectivity of each subconductance state only depends on one of the two subunits and is independent of the other conductance level (Ludewig et al., 1996; Middleton et al., 1996). These results were initially interpreted by postulating a unique `doublebarreled' architecture with two identical and independent ion conduction pathways. In contrast to this suggestion, for hClC-1, several lines of experimental evidence demonstrated that pore-forming regions from both subunits face the same ionic pore (Fahlke et al., 1998) indicating that each hClC-1 channel exhibits only a single ionic pore. As it is dif®cult to imagine that various ClC isoforms radically differ in their quaternary structure, we believe that all ClC channels exhibit a single ion conduction pathway that is jointly formed by both subunits as is the case for other ion

Dent's disease

265

channels. For ClC-2, an intracellular gate mediating opening and closing transitions much like the inactivation gate in sodium and potassium channels and its presumed receptor has been identi®ed (Jordt and Jentsch, 1997). Gating properties of different ClC-isoforms are quite variable and neither the mechanisms nor the structural determinants of these processes are suf®ciently understood.

Functional characterization of the CLCN5 gene product CLCN5 encodes a 746 amino acid protein with a molecular weight of 83 kDa (Fisher et al. 1995). Among the ClC channel family, ClC-5 is most similar to ClC-3 and ClC-4 (Kawasaki et al., 1994; Jentsch et al., 1995, Fig. 2). Heterologous expression of rClC-5 in Xenopus oocytes and functional characterization with the two-microelectrode voltage clamp technique (Steinmeyer et al., 1995) demonstrated a novel chloride current with unique properties (Fig. 3). These currents are profoundly outwardly rectifying, with only very small inward current at negative potentials. While it is dif®cult in the Xenopus oocyte expression system to distinguish such small currents from leak currents, a later study using whole-cell patch-clamp recordings in mammalian cells clearly indicated that ClC-5 currents have a non-zero open probability and unitary conductance at negative voltages (Friedrich et al., 1999). At positive potentials (.40 mV) a relatively fast, time-dependent increase of the current amplitude can be observed, but surprisingly, in each of the two expression systems, deactivation occurs too fast to be recorded with either whole-cell patch-clamping or two microelectrode voltage clamping. Anion substitution experiments revealed a NO32 . Cl 2 . Br 2 . HCO32 . I 2 conductivity sequence (Steinmeyer et al., 1995; Friedrich et al., 1999; Mo et al., 1999). ClC-5 has an unusual pharmacological pro®le. It is not blocked by any known inhibitors of chloride currents: anthracene-9-carboxylic acids (9-AC) (1 mM), DIDS, dephenylamine-2-carboxylic acid (DPC) (1 mM), 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) (1 mM), nor by ni¯umic acid (1 mM) (Steinmeyer et al., 1995; Mo et al., 1999). Although rClC-5 exhibits a consensus site for cAMP dependent phosphorylation at a presumed intracellular position, raising intracellular cAMP levels does not affect ClC-5 mediated chloride currents (Steinmeyer et al., 1995). The experimental ®nding that chloride currents through the surface membrane can be recorded in heterologous expression systems demonstrates that, although the majority of channels are located in intracellular compartments, a certain percentage reaches the surface membrane in overexpressing cells. It is currently unclear whether there are functional differences between surface and intracellular channels.

266

Ch. Fahlke

Fig. 3. Voltage-dependence of peak current amplitudes from oocytes injected with rClC-5 cRNA. Current recordings were performed with a two-microelectrode voltage-clamp and the external solution contained (in mM) 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, 5 Hepes, pH 7.4. Amplitudes were measured at the end of 350 ms-test steps applied from a holding potential of 230 mV and normalized for each oocyte to the amplitude obtained at 1 100 mV. Current amplitudes represent mean ^ SEM from four different oocytes. The insert shows one of the original traces. This ®gure is reprinted from (Steinmeyer et al., 1995) with permission.

Spectrum of observed mutations Following the initial report by Lloyd et al. (1996), a large number of mutations in CLCN5 causing Dent's disease have been identi®ed in several affected kindred in Asia, the US and Europe (Akuta et al., 1997; Lloyd et al., 1997a,b; Morimoto et al., 1998; Igarashi et al., 1998; Fig. 4, Table 1). Mutations in CLCN5 represent the full spectrum of genetic defects: missense, nonsense, splice junction, deletion, frameshifts and insertions have all been observed. There are two obvious clusters of missense mutations in transmembrane domains D6 and in D11 (Fig. 4). So far,

Dent's disease

267

Fig. 4. Spectrum of known CLCN5 mutations found in patients with Dent's disease. The exact transmembrane topology of ClC channels is currently unknown. The topology shown is based on a proposal by Schmidt-Rose and Jentsch (Schmidt-Rose and Jentsch, 1997). It has been modi®ed to account for results of substituted cysteine accessibility studies in a putative pore region (between D3 and D5) (Fahlke et al., 1997b). To account for the experimentally demonstrated extracellular location of the D6-D7 linker (Kuchenbecker et al., 2000), D5 and D6 have been moved together as a broad hydrophobic domain.

the functional role of these segments has not been investigated. D6 is close to the presumed pore region (Fahlke et al., 1997b). It therefore appears possible that these mutations interfere with ion conduction. Furthermore, there are several diseasecausing deletions in the C-terminus, and, based on previous work with ClC-1, this region was suggested to play a role in intracellular processing of ClC-proteins (Hryciw et al., 1998). The functional effects of many of these mutations have been investigated in Xenopus oocytes. The majority of Dent's disease causing mutations abolishes chloride currents in injected oocytes, though some only decrease current levels (Hins30; G57V (Lloyd et al., 1997b), S244L (Lloyd et al., 1996), R280P (Lloyd et al., 1997a), L278F (Igarashi et al., 1998). The functional effects of naturally occurring mutations in ClC-5 therefore clearly differ from those of myotonia±congenita mutations in ClC-1, where a large percentage cause speci®c alterations of gating or permeation and only a small percentage completely abolishes function. An obvious complication in the characterization of Dent's disease mutations is that, due to methodological reasons, all information stems solely from recording of ClC-5 channels at the membrane surface, which is not the channel population responsible for the presumed cellular function. It would be therefore interesting to reconstitute intracellular WT and mutant ClC-5 channels in planar bilayers and to subject them to detailed single-channel analysis in order to learn more about the exact functional consequences of the disease-causing mutations.

268

Ch. Fahlke

Table 1 Published mutations in CLCN5 in patients with Dent's disease

Missense W22G G57V L200R S244L S270R L278F R280P G506E G512R G513E R516W S520P E527D

Trp to Gly Gly to Val Leu to Arg Ser to Leu Ser to Arg Leu to Phe Arg to Pro Gly to Glu Gly to Arg Gly to Glu Arg to Trp Ser to Pro Glu to Asp

Morimoto et al. (1998) Lloyd et al. (1997a) Lloyd et al. (1996) Lloyd et al. (1996) Igarashi et al. (1998) Igarashi et al. (1998) Lloyd et al. (1997a) Lloyd et al. (1996) Lloyd et al. (1997b) Akuta et al. (1997) Akuta et al. (1997) Lloyd et al. (1996) Lloyd et al. (1997b)

Nonsense R34X E118X W279X W343X R347X R648X R704X

Loss Loss Loss Loss Loss Loss Loss

Cox et al. (1999) Morimoto et al. (1998) Lloyd et al. (1997) Lloyd et al. (1997) Morimoto et al. (1998) Lloyd et al. (1996) Lloyd et al. (1996)

Insertion 30 ACC in-frame insertion

of 696 AA of 628 AA of 467 AA of 404 AA of 399 AA of 278 AA of 42 AA

Insertion of histidine at position 30

Lloyd et al. (1997a)

Splice-site Del 132-172

Loss of D2

Lloyd et al. (1996)

Deletions Del 132-241 Deletion of entire gene Del C 695

Loss of D2 to D4 Absence of protein Loss of C-terminal 47 AA

Lloyd et al. (1996) Lloyd et al. (1996) Lloyd et al. (1997a)

Open questions and an hypothesis Remarkable progress in our understanding of Dent's disease has been made in the last few years. Little time has passed from the initial characterization of this disease to the identi®cation of the gene locus and the establishment of a likely pathome-

Dent's disease

269

chanism that is able to explain many of the clinical ®ndings in affected patients. There are nevertheless a number of open questions and some inconsistencies in the current concept that need to be resolved in future studies. ClC-5 chloride currents in heterologous expression systems are strongly outwardly rectifying. When we assume that channels insert into the surface membrane by vesicle fusion, this ®nding indicates that ClC-5 exhibits a higher conductance pathway for anion movement out of the vesicle than in the opposite direction. Because of this recti®cation, ClC-5 channels will provide only a small conductance at 0 mV, the assumed voltage drop over the endosomal membrane. It is apparent that the functional properties of ClC-5 differ profoundly from reports about native chloride channels in endosomal vesicles. In endocytotic vesicles from rabbit proximal tubules (Bae and Verkman, 1990; Marshansky and Vinay, 1996), activation of a stilbene-sensitive chloride conductance by protein kinase A was shown to cause acidi®cation of the vesicle lumen. In a subsequent study, Reenstra et al. (1992) demonstrated that the molecular weight of the phosphorylated endosomal channel (67 kDa) is distinct from the 83 kDa band that ClC-5 would produce. Using patch-clamp recordings on fused endocytotic vesicles from rat kidney cortex, Schmid et al. (1989) described a chloride channel with a linear I±V relationship and a slope conductance of 73 pS under symmetrical chloride concentrations, as well as a Cl 2 ˆ Br 2 ˆ I 2 . SO42 . F 2 permeability sequence that is sensitive to DIDS and NPPB. All the descriptions of native endosomal chloride channels are consistent, but differ profoundly from ClC-5. ClC-5 in heterologous expression systems exhibits a pronounced outward recti®cation, a NO32 . Cl 2 . Br 2 . HCO32 . I 2 conductivity sequence and is neither regulated by protein kinase A nor sensitive to stilbenes (Steinmeyer et al., 1995; Mo et al., 1999). Several possible explanations may resolve these inconsistencies. Because of the specialized localization of ClC-5 (see Pathophysiology), it is imaginable that the native channels are not from cells that express ClC-5, but from contamination of other cells expressing other chloride channel types. Alternatively, native endosomal chloride channels may be formed by ClC-5 and another pore-forming subunit. As both subunits appear to jointly form an ion conduction pathway in ClC channels (Fahlke et al., 1998), the permeability and pharmacology of these heteromultimeric channels could be very distinct from homomultimeric channels. Although coexpression experiments with ClC-3, ClC-4 and ClC-5 in Xenopus oocytes did not reveal formation of channel with novel functional properties (Steinmeyer et al., 1995), this scenario appears possible and must be addressed experimentally in the future. Another possibility is that ClC-5 indeed forms homodimeric channels with properties very similar to those observed in heterologous expression systems and that these ful®ll another functional role. In this hypothetical scenario, additional chloride channels would be present in the same membrane compartment as the ones described in experiments with native tissue. These chloride channels that are not

270

Ch. Fahlke

yet de®ned at the molecular level are responsible for providing a high anionic conductance of the endosomal vesicle. In the following I will propose a possible ClC-5 function that contributes to the acidi®cation of endosomes and is in complete agreement with the functional properties of homodimeric ClC-5 channels. It appears likely that the unique recti®cation properties permit ClC-5 to mediate bicarbonate transport out of the endosome against an electrochemical gradient, by coupling to membrane potential oscillations. Secondary active bicarbonate ef¯ux will cause a decrease of the intra-endosomal bicarbonate concentration and thus contribute to the acidi®cation of the endosome. (Fig. 5)

Fig. 5. Hypothetical role of ClC-5 channels in endosomal vesicles of proximal tubule cells. ClC-5 is suggested to serve as a bicarbonate channel solely permitting the ef¯ux out of the vesicle and supporting endosomal acidi®cation by a proton pump actively transporting protons into the vesicle. Additional chloride channels contribute to the charge dissipation preventing depolarization of the intravesicular potential.

Dent's disease

271

ClC-5 channels in heterologous expression systems exhibit a very high relative bicarbonate permeability with a relative permeability ratio of PHCO3 =PCl ˆ 0:83 (Mo et al., 1999). Due to its unique rectifying properties, ClC-5 channels in the endosomal membrane would provide a pathway that almost selectively permits the ef¯ux of anions, bicarbonate and chloride, out of the vesicle. The proximal tubule of the nephron reabsorbs a signi®cant amount of the ®ltered bicarbonate via proton secretion by a Na 1/H 1 exchanger in the apical membrane and a Cl 2/HCO32 exchanger in the basolateral membrane side. The cytoplasm of tubular cells is rich in carbonic anhydrase, an enzyme that catalyzes the reaction of protons and bicarbonate to form water and carbon dioxide. The pK of this reaction is 6.1 and proximal tubule cells exhibit a cytoplasmic pH between 7.15 and 7.30, resulting in a cytoplasmic bicarbonate concentration of 24 mM. As the majority of chloride channels are bicarbonate permeant (Poulsen et al, 1994; Nilius et al., 1999), it appears likely that certain chloride channels in the endosomal membrane provide also a bicarbonate permeability that allows an in¯ux of HCO32 into the vesicle. The CO2/HCO32 system buffers solutions to alkaline pH and would antagonize the action of the endosomal proton pump depending on the intravesicular bicarbonate concentration. Homodimeric ClC-5 channels mediating the ef¯ux of bicarbonate would therefore be necessary to acidify the vesicle. Due to its recti®cation, ClC-5 could act as a secondary active bicarbonate transporter by simply coupling to oscillations of the transmembrane potential (Fig. 6). The vesicle membrane is electrically active, as it exhibits an electrogenic H 1 pump and additional passive channel-mediated charge carriers. Under these conditions, the potential difference between vesicle interior and the cytoplasm would not be perfectly stable, but would oscillate around a mean membrane potential value of 0 mV (Fig. 6). Although the integral of the potential differences over time will equal zero, these oscillations will give rise to a net charge movement as the integral of the voltage times a voltage-dependence conductance will not result in zero but in net anion movement out of the vesicle (Fig. 6). The rectifying properties therefore may cause a decrease in the internal anion concentration. As ClC-5 is permeant to HCO32, this feature will decrease the intravesicular bicarbonate concentration and thus contribute to the acidi®cation of the endosome. Such a mechanism would require high membrane densities of ClC-5 channels and it is easily imaginable that a slight impairment of channel function, as observed in some of the known Dent's disease mutations, might affect the suggested secondary active transport. A loss of ClC-5 function would prevent this bicarbonate transport out of the vesicle, increase the intravesicular bicarbonate concentration and thus increase the vesicular pH. Membrane recycling occurs in a variety of tissues and the question therefore arises why an additional ion channel is involved in the acidi®cation of proximal tubule cells. It may be argued that due to excessive protein reabsorption and degradation

272

Ch. Fahlke

Fig. 6. Secondary active bicarbonate transport by a rectifying channel. (A) Membrane potential oscillation around a mean value of 0 mV. (B) Current responses of a rectifying channel to the potential oscillations shown in (A). (C) Net charge movement by a rectifying channel undergoing membrane oscillations shown in (A). (C) Represents the time integral of the current values shown in (B) demonstrating that there is an increasing net anion transport over the membrane. All values are given in arbitrary units.

vesicle acidi®cation is particularly important in these cells and therefore additional mechanisms of proton accumulation are necessary. The current concepts of Dent's disease pathophysiology are still speculative at present and are in need of more direct experimental proof. It will be important to directly show an endosomal acidi®cation defect in proximal tubule cells lacking ClC-5 or expressing ClC-5 channels carrying Dent's disease causing mutations. Techniques to measure the pH of intracellular compartments are available and the availability of knock-out or transgenic mice will allow investigation of this important question in the near future. Similar experiments under conditions that change the cytoplasmic bicarbonate concentration will allow testing of whether the suggested role of ClC-5 as a secondary active bicarbonate carrier plays a role in the pathogenesis of Dent's disease. We are anxiously awaiting the electrophysiological and biochemical characterization of native ClC-5 channels in proximal tubule cells to learn about their subunit composition and the possibility of additional accessory subunits modifying functional properties. Ideally, these experiments will allow study of the functional effects of Dent's disease-causing mutations under more realistic conditions. Finally, the pathogenesis of osteomalacia is not yet understood and remains an important challenge for the future. Perhaps, Dent's

Dent's disease

273

disease will help us to understand another surprising role of this most interesting ion channel family. Dent's disease has serious consequences for the well-being of affected patients and can, when resulting in end-stage renal disease, be a life-threatening condition. Novel therapeutic strategies need to be developed and experimentally and clinically tested. We all hope that the collaborative efforts of basic and clinical researchers that have been so successful in the past will complete this important task in the near future. Acknowledgements The author would like to thank Dr Al George for support, Drs Al George, Steve Hebert, JP Johnson and Chris Williams for helpful discussions and critical reading of the manuscript and Dr Kevin Strange for a stimulating discussion about possible roles of ClC-5 in proximal tubule cells. Since the number of references per chapter was limited, I could not cite all the relevant publications and I apologize to the authors of these studies that were omitted. References Akuta, N., Lloyd, S.E., Igarashi, T., Shiraga, H., Matsuyama, T., Yokoro, S., Cox, J.P.D., Thakker, R.V., 1997. Mutations of CLCN5 in Japanese children with idiopathic low molecular weight proteinuria, hypercalciuria and nephrocalcinosis. Kidney Int. 52, 911±916. Al-Awqati, Q., Barasch, J., Landry, D., 1992. Chloride channels of intracellular organelles and their potential role in cystic ®brosis. J. Exp. Biol. 172, 245±266. Bae, H.R., Verkman, A.S., 1990. Protein kinase A regulates chloride conductance in endocytic vesicles from proximal tubule. Nature. 348, 637±639. Bolino, A., Devoto, M., Enia, G., Zoccali, C., Weissenbach, J., Romeo, G., 1993. Genetic mapping in the Xp11.2 region of a new form of X-linked hypophosphatemic rickets. Eur. J. Hum. Genet. 1, 269±279. Buyse, G., Trouet, D., Voets, T., Missiaen, L., Droogmans, G., Nilius, B., Eggermont, J., 1998. Evidence for the intracellular location of chloride channel (CIC)-type proteins: co-localization of CIC-6a and CIC-6c with the sarco/endoplasmic-reticulum Ca 21 pump SERCA2b. Biochem. J. 330, 1015±1021. Christensen, E.I., Nielsen, S., 1991. Structural and functional features of protein handling in the kidney proximal tubule. Semin. Nephrol. 11, 414±439. Christensen, E.I., Birn, H., Verroust, P., Moestrup, S.K., 1998. Megalin-mediated endocytosis in renal proximal tubule. Renal Fail. 20, 191±199. Cox, J.P., Yamamoto, K., Christie, P.T., Wooding, C., Feest, T., Flinter, F.A., Goodyer, P.R., Leumann, E., Neuhaus, T., Reid, C., Williams, P.F., Wrong, O., Thakker, R.V., 1999. Renal chloride channel, CLCN5, mutations in Dent's disease. J. Bone Min. Res. 14, 1536±1542. Cui, S., Verroust, P.J., Moestrup, S.K., Christensen, E.I., 1996. Megalin/gp330 mediates uptake of albumin in renal proximal tubule. Am. J. Physiol. 271, F900±F907. Dent, C.E., Friedman, M., 1964. Hypercalciuric rickets associated with renal tubular damage. Arch. Disord. Child. 39, 240±249. Devuyst, O., Christie, P.T., Courtoy, P.J., Beauwens, R., Thakker, R.V., 1999. Intra-renal and subcellular

274

Ch. Fahlke

distribution of the human chloride channel, CLC-5, reveals a pathophysiological basis for Dent's disease. Hum. Mol. Genet. 8, 247±257. Fahlke, Ch., Knittle, T.J., Gurnett, C.A., Campbell, K.P., George Jr., A.L., 1997a. Subunit stoichiometry of human muscle chloride channels. J. Gen. Physiol 109, 93±104. Fahlke, Ch., Yu, H.T., Beck, C.L., Rhodes, T.H., George Jr, A.L., 1997b. Pore-forming segments in voltage-gated chloride channels. Nature 390, 529±532. Fahlke, Ch., Rhodes, T.H., Desai, R.R., George Jr., A.L., 1998. Pore stoichiometry of a voltage-gated chloride channel. Nature 394, 687±690. Fisher, S.E., Black, G.C.M, Lloyd, S.E., Hatchwell, E., Wrong, O., Thakker, R.V., Craig, I.W., 1994. Isolation and partial characterization of a chloride channel gene which is expressed in kidney and is a candidate for Dent's disease (an X-linked hereditary nephrolithiasis). Hum. Mol. Genet. 3, 2053± 2059. Fisher, S.E., Van Bakel, I., Lloyd, S.E., Pearce, S.H.S, Thakker, R.V., Craig, I.W., 1995. Cloning and characterization of CLCN5, the human kidney chloride channel gene implicated in Dent's disease (an X-linked hereditary nephrolithiasis). Genomics 29, 598±606. Friedrich, T., Breiderhoff, T., Jentsch, T.J., 1999. Mutational analysis demonstrates that ClC-4 and ClC-5 directly mediate plasma membrane currents. J. Biol. Chem. 274, 896±902. Frymoyer, P.A., Scheinman, S.J., Dunham, P.B., Jones, D.B., Hueber, X., Schroeder, E.T., 1991. X-linked recessive nephrolithiasis with renal failure. N. Eng. J. Med. 325, 681±686. Gekle, M., Mildenberger, S., Freudinger, R., Silbernagl, S., 1995. Endosomal alkalinization reduces Jmax and Km of albumin receptor-mediated endocytosis in OK cells. Am. J. Physiol. 268, F899±F906. George, A.L., Crackower, M.A., Abdalla, J.A., Hudson, A.J., Ebers, G.C., 1993. Molecular basis of Thomsen's disease (autosomal dominant myotonia congenita). Nat. Genet. 3, 305±310. GuÈnther, W., Luchow, A., Cluzeaud, F., Vandewalle, A., Jentsch, T.J., 1998. ClC-5, the chloride channel mutated in Dent's disease, colocalizes with the proton pump in endocytotically active kidney cells. Proc. Natl. Acad. Sci. USA 95, 8075±8080. Hebert, S.C., 1996. Crystal-clear chloride channels. Nature 379, 398±399. Hoffman, E.P., Lehmann-Horn, F., RuÈdel, R., 1995. Overexcited or inactive: ion channels in muscle disease. Cell 80, 681±686. Hryciw, D.H., Rychkov, G.Y., Hughes, B.P., Bretag, A.H., 1998. Relevance of the D13 region to the function of the skeletal muscle chloride channel, ClC-1. J. Biol. Chem. 273, 4304±4307. Igarashi, T., GuÈnther, W., Sekine, T., Inatomi, J., Shiraga, H., Takahashi, S., Suzuki, J., Tsuru, N., Yanagihara, T., Shimazu, M., Jentsch, T.J., Thakker, R.V., 1998. Functional characterization of renal chloride channel, CLCN5, mutations associated with Dent's Japan disease. Kidney Int. 54, 1850±1856. Jentsch, T.J., Steinmeyer, K., Schwarz, G., 1990. Primary structure of Torpedo marmorata chloride channel isolated by expression cloning in Xenopus oocytes. Nature 348, 510±514. Jentsch, T.J., GuÈnther, W., Pusch, M., Schwappach, B., 1995. Properties of voltage-gated chloride channels of the ClC gene family. J. Physiol. (Lond.) 482, 195±255. Jentsch, T.J., GuÈnther, W., 1997. Chloride channels: an emerging molecular picture. Bioessays 19, 117± 126. Jordt, S.E., Jentsch, T.J., 1997. Molecular dissection of gating in the CIC-2 chloride channel. EMBO J. 16, 1582±1592. Kawasaki, M., Uchida, S., Monkawa, T., Miyawaki, A., Mikoshiba, K., Maruma, F., Sasaki, S., 1994. Cloning and expression of a protein kinase C-regulated chloride channel abundantly expressed in rat brain neuronal cells. Neuron 12, 597±604. Koch, M.C., Steinmeyer, K., Lorenz, C., Ricker, K., Wolf, F., Otto, M., Zoll, B., Lehmann-Horn, F., Grzeschik, K.H., Jentsch, T.J., 1992. The skeletal muscle chloride channel in dominant and recessive human myotonia. Science 257, 797±800.

Dent's disease

275

Kuchenbecker, M., Schu, B., KuÈrz, L.L., RuÈdel, R., 2000. The structure/function relationship of the hCIC1 C1 2 channel studied using hydrophilic epitope insertions. P¯uÈgers Arch. 439, R36. Lloyd, S.E., Pearce, S.H.S., Fisher, S.E., Steinmeyer, K., Schwappach, B., Scheinman, S.J., Harding, B., Alessandra, B., Devota, M., Goodyear, P., Rigden, S.P.A., Wrong, O., Jentsch, T.J., Craig, I.W., Thakker, R.V., 1996. A common molecular basis for three inherited kidney stone diseases. Nature 379, 445±449. Lloyd, S.E., Pearce, S.H.S., GuÈnther, W., Kawaguchi, H., Igarashi, T., Jentsch, T.J., Thakker, R.V., 1997a. Idiopathic low molecular weight proteinuria associated with hypercalciuric nephrocalcinosis in Japanese children is due to mutations of the renal chloride channel (CLCN5). J. Clin. Invest. 99, 967±974. Lloyd, S.E., GuÈnther, W., Pearce, S.H.S., Thomson, A., Bianchi, M.L., Bosio, M., Craig, I.W., Fisher, S.E., Scheinman, S.J., Wrong, O., Jentsch, T.J., Thakker, R.V., 1997b. Characterisation of renal chloride channel, CLCN5, mutations in hypercalciuric nephrolithiasis (kidney stones) disorders. Hum. Mol. Genet. 6, 1233±1239. Lorenz, C., Pusch, M., Jentsch, T.J., 1996. Heteromultimeric CLC chloride channels with novel properties. Proc. Natl. Acad. Sci. USA 93, 13362±13366. Ludewig, U., Pusch, M., Jentsch, T.J., 1996. Two physically distinct pores in the dimeric CIC-0 chloride channel. Nature 383, 340±343. Luyckx, V.A., Goda, F.O., Mount, D.B., Nishio, T., Hall, A., Hebert, S.C., Hammond, T.G., Yu, A.S., 1998. Intrarenal and subcellular localization of rat CLC5. Am. J. Physiol. 275, F761±F769. Marshansky, V., Vinay, P., 1996. Proton gradient formation in early endosomes from proximal tubules. Biochim. Biophys. Acta. 1284, 171±180. Middleton, R.E., Pheasant, D.J., Miller, C., 1994. Puri®cation, reconstitution and subunit composition of a voltage-gated chloride channel from Torpedo electroplax. Biochemistry. 33, 13189±13198. Middleton, R.E., Pheasant, D.J., Miller, C., 1996. Homodimeric architecture of a ClC-type chloride ion channel. Nature 383, 337±340. Mo, L., Hellmich, H.L., Fong, P., Wood, T., Embesi, J., Wills, N.K., 1999. Comparison of amphibian and human ClC-5: similarity of functional properties and inhibition by external pH. J. Membrane Biol. 168, 253±264. Morimoto, T., Uchida, S., Sakamoto, H., Kondo, Y., Hanamizu, H., Fukui, M., Tomino, Y., Nagano, N., Sasaki, S., Marumo, F., 1998. Mutations in CLCN5 chloride channel in Japanese patients with low molecular weight proteinuria. J. Am. Soc. Nephrol. 9, 811±818. Nilius, B., Voets, T., Eggermont, J., Droogmans, G., 1999. VRAC: a multifunctional volume-regulated anion channel in vascular endothelium. In: Koslowski, R.Z. (Ed.). Cl 2 Channels. Isis Medical, Oxford, pp. 47-63. Pook, M.A., Wrong, O., Woodling, C., Norden, A.G.W., Feest, T.G., Thakker, R.V., 1993. Dent's disease, a renal Fanconi syndrome with nephrocalcinosis and kidney stones, is associated with a microdeletion involving DXS255 and maps to Xp11.22. Hum. Mol. Genet. 2, 2129±2134. Poulsen, J.H., H. Fischer, H., Illek, B., Machen, T.E., 1994. Bicarbonate conductance and pH regulatory capability of cystic ®brosis transmembrane conductance regulator. Proc. Natl. Acad. Sci. USA 91, 5340±5344. Reenstra, W.W., Sabolic, I., Bae, H.R., Verkman, A.S., 1992. Protein kinase A dependent membrane protein phosphorylation and chloride conductance in endosomal vesicles from kidney cortex. Biochemistry 31, 175±181. Scheinman, S.J., 1998. X-linked hypercalciuric nephrolithiasis: clinical syndromes and chloride channel mutations. Kidney Int. 53, 3±17. Scheinman, S.J., Pook, M.A, Wooding, C., Pang, J.T., Frymoyer, P.A., Thakker, R.V., 1993. Mapping the gene causing X-linked recessive nephrolithiasis to Xp11.22 by linkage studies. J. Clin. Invest. 91, 2351±2357.

276

Ch. Fahlke

Schmid, A., Burckhardt, G., GoÈgelein, H., 1989. Single chloride channels in endosomal vesicle preparations from rat kidney cortex. J. Membrane Biol. 111, 265±275. Schmidt-Rose, T., Jentsch, T.J., 1997. Transmembrane topology of a CLC chloride channel. Proc. Natl. Acad. Sci. USA 94, 7633±7638. Simon, D.B., Bindra, R.S., Mans®eld, T.A., Nilson-Williams, C., Mendonca, E., Stone, R., Schurman, S., Nayir, A., Alpay, H., Bakkaloglu, A., Rodriguez-Soriano, J., Morales, J.M., Sanjad, S.A., Taylor, C.M., Pilz, D., Brem, A., Trachtman, H., Griswold, W., Richard, G.A., John, E., Lifton, R.P., 1997. Mutations in the chloride channel gene, CLCNKB, cause Bartter's syndrome type III. Nat. Genet. 17, 171±178. Steinmeyer, K., Klocke, R., Ortland, C., Gronemeier, M., Jockusch, H., GruÈnder, S., Jentsch, T.J., 1991. Inactivation of muscle chloride channel by transposon insertion in myotonic mice. Nature 354, 304± 308. Steinmeyer, K., Schwappach, B., Bens, M., Vandewalle, A., Jentsch, T.J., 1995. Cloning and functional expression of rat CLC-5, a chloride channel related to kidney disease. J. Biol. Chem. 270, 31172± 31177. Suzuki, Y., Okada, T., Higuchi, A., Mase, D., Kobayashi, O., 1985. Asymptomatic low molecular weight proteinuria: a report on ®ve cases. Clin. Nephrol. 23, 249±254. Wall, D.A., Maack, T., 1985. Endocytic uptake, transport and catabolism of proteins by epithelial cells. Am. J. Physiol. 248, C12±C20. Wrong, O., Norden, G.A., Feest, T.G., 1992. X-linked recessive nephrolithiasis with renal failure. N. Eng. J. Med. 326, 1029±1030. Wrong, O.M., Norden, A.G., Feest, T.G., 1994. Dent's disease; a familial proximal renal tubular syndrome with low-molecular-weight proteinuria, hypercalciuria, nephrocalcinosis, metabolic bone disease, progressive renal failure and a marked male predominance. Q. J. Med. 87, 473±493.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

277

Chapter 11

Liddle's syndrome and pseudohypoaldosteronism type I Stefan GruÈnder ENT University Hospital, Department of Otolaryngology, Sensory Biophysics, RoÈntgenweg 11, 72076 TuÈbingen, Germany

Abstract In recent years kidney physiology has made a big step forward through the molecular identi®cation and characterization of the major transporters and ion channels responsible for ion and solute transport. This has made it possible to test these molecules as candidate genes for hereditary kidney diseases demonstrating in only a few years that mutations in transporters and channels in different parts of the kidney are responsible for some major monogenic forms of hereditary defects in salt homeostasis. One of the most illustrative examples is the case of the epithelial sodium channel ENaC. ENaC provides the apical entry mechanism for sodium in the distal nephron. Its activity is under the tight control of aldosterone. Intriguingly, two genetic diseases with almost opposite symptoms are caused by mutations in the same molecule. Liddle's syndrome, a form of severe hypertension, is caused by gain-of-function mutations and pseudohypoaldosteronism type I, a form of hypovolaemia in infancy, is caused by loss-of-function mutations in ENaC. This not only highlights the importance of tight control of ENaC activity for salt homeostasis and blood pressure but provides also new insights into the molecular mechanisms controlling sodium reabsorption in the distal nephron. q 2000 Elsevier Science B.V. All rights reserved.

278

S. GruÈnder

Liddle's syndrome Mutations in the epithelial sodium channel cause Liddle's syndrome The epithelial sodium channel is a candidate gene for Liddle's syndrome Clinical phenotype In 1963, Grant Liddle and co-workers described a family in which many siblings had early onset of severe hypertension (Liddle et al., 1963). Moreover, in some of the patients hypertension was associated with hypokalaemia, indicating that the hypertension might be due to increased sodium reabsorption in the distal nephron where potassium secretion is indirectly coupled to mineralocorticoid-regulated sodium reabsorption. The clinical symptoms are typical of hyperaldosteronism but actual rates of aldosterone excretion in the patients were suppressed, explaining the classi®cation as `pseudoaldosteronism`. Accordingly, spironolactone, a mineralocorticoid antagonist, did not ameliorate the symptoms, while triamterene, a K 1-sparing diuretic, normalized the blood pressure and corrected the hypokalaemia. These therapeutic features are still indicative for Liddle's syndrome (OMIM #177200) and clearly pointed to a renal mechanism as the cause of the syndrome. As predicted by Liddle, plasma renin activity in Liddle's patients proved to be low (Gordon, 1995) and kidney transplantation could rescue the hypertensive phenotype (Botero-Velez et al., 1994). It was also established that Liddle's syndrome is transmitted in an autosomal dominant fashion (Botero-Velez et al., 1994). The severeness of Liddle's syndrome is illustrated by the fact that the youngest patient diagnosed was only 10 months old and at that time had already blood pressure (BP) of 135/100 mmHg (Vania et al., 1997). Although Liddle's syndrome is a rare and extreme form of hypertension it represents a paradigm also for more common forms of salt-sensitive hypertension (Lifton, 1996). Physiology of the epithelial sodium channel Luminal uptake of Na 1 in the distal convoluted tubule (DCT) and cortical collecting duct (CCD) of the kidney occurs via the epithelial sodium channel (ENaC; Fig. 1). The energy for sodium entry comes from the basolaterally located Na/K-ATPase. Experiments in the Ussing chamber suggest that ENaC is the rate limiting step in this vectorial transport of Na 1. Both ENaC and Na/K-ATPase are regulated by the mineralocorticoid aldosterone (Rossier and Palmer, 1992; GruÈnder and Rossier, 1997). ENaC can be speci®cally blocked by amiloride and triamterene. Sodium reabsorption in the distal nephron creates a transepithelial potential difference which provides the energy for K 1 secretion through apically located K 1 channels. These physiological features make ENaC an ideal candidate gene for Liddle's disease. Constitutive activity of this channel would account for excessive Na 1

Liddle's syndrome and PHA type I

279

Fig. 1. Mechanisms of Na 1 reabsorption in a principal cell of the CCD. The epithelial Na 1 channel (ENaC) is located at the apical membrane and can be blocked by amiloride and triamterene. The Na/KATPase is located at the basolateral membrane, K 1 secretion is mediated by an apically located K 1 channel, ROMK. Both ENaC and Na/K-ATPase are under the control of the mineralocorticoid aldosterone. Aldosterone secretion is initiated by cleavage of angiotensinogen through the protease renin. MR, mineralocorticoid-receptor; AIPs, aldosterone induced proteins.

reabsorption and hypertension. The increased Na 1 reabsorption would in turn lead to increased K 1 secretion. Aldosterone and renin levels would be low to compensate these features. And ®nally, mineralocorticoid-antagonists would be without effect but channel blockers should ameliorate the symptoms. Molecular biology of the epithelial sodium channel Molecular cloning of the epithelial sodium channel Molecular cloning of ENaC has made use of the high af®nity block by the diuretic amiloride. Amiloride (5 mM) almost completely blocks the channel. A cDNA library was constructed from the

280

S. GruÈnder

distal colon of rats which were maintained on a low salt diet to increase the abundance of mRNA for this channel. Pools of RNA from the library were injected in Xenopus oocytes and ENaC expression monitored by amiloride-blockable Na 1 current. By consecutively reducing the size of the pools, ultimately three different clones were isolated ± a, b, and gENaC (Canessa et al., 1995). These three clones showed distinct sequences but were about 35% identical to each other, demonstrating that they belong to the same gene family. Injection of aENaC alone elicits only a small amiloride-blockable Na 1 current. Co-injection with either b or g increases the current amplitude but only co-expression of all three subunits leads to a fully active channel. b and gENaC alone do not produce any signi®cant amiloride-sensitive Na 1 current. Binding studies on ENaC expressing oocytes later showed that this difference in current amplitude is due to differences in surface expression of channels (Firsov et al., 1996), suggesting that only the three subunits together form an ef®ciently targeted channel. Expression of a, b, and gENaC together reconstituted all of the hallmarks of the epithelial sodium channel from the distal nephron: high selectivity for Na 1 over K 1 (.10:1), high af®nity for amiloride (Ki < 0.1±0.5 mM) and a single channel conductance of about 4.5 pS (Canessa et al., 1995). The cloned subunits are expressed in the distal nephron, colon, sweat and salivary glands and lung (Duc et al., 1994). Secondary structure of the epithelial sodium channel The three ENaC subunits share a common structure (for a recent review see Horisberger, 1998 and references herein; Fig. 2a). They have two hydrophobic transmembrane domains (M1 and M2), a large extracellular loop with 16 conserved cysteines and rather short intracellular N- and C-termini. The conserved cysteines are important for proper folding and targeting of the channel (Firsov et al., 1999), but the actual signi®cance of the big extracellular loop, parts of which are highly conserved, is not known. Expression studies using mutant channels suggest that the M2 segment is critical for the formation of the channel pore. A pre-M2 segment was identi®ed in which mutations change the af®nity for block by amiloride and ion selectivity, suggesting that this part of the channel forms or is very close to the outer mouth of the channel pore. The N-terminal half of the channel is suf®cient to mediate association between subunits. The C-terminal tail of all three subunits contains proline-rich motifs which often mediate protein±protein interactions. The motif of aENaC was shown to bind the SH3 domain of a-spectrin and may be important for proper targeting of the channel to the apical membrane of epithelial cells (Rotin et al., 1994). The active ENaC is a heterotetramer composed of two a, one b and one g subunit with the four subunits contributing probably in a symmetric fashion to the formation of the channel pore and amiloride binding site (Fig. 2b).

Liddle's syndrome and PHA type I

281

Fig. 2. (a) Membrane topology of ENaC subunits. The speci®c blocker amiloride is shown. Modi®ed from Horisberger (1998). (b) Subunit stoichiometrie. For details see text.

Identi®cation of mutations in ENaC subunits in patients with Liddle's syndrome A truncation of b ENaC in Liddle's original kindred Using the rat bENaC as a probe Lifton and co-workers could show complete linkage of the locus for human bENaC to Liddle's syndrome in the original kindred examined by Liddle (Shimkets

282

S. GruÈnder

et al., 1994). This analysis showed that both the gene for human bENaC and the gene responsible for Liddle's syndrome localized to a very short region on chromosome 16p. Screening for single strand conformational polymorphism's (SSCPs) revealed a variant in the C-terminal part of the bENaC coding sequence which was found only in subjects who had been diagnosed with Liddle's syndrome but not in unaffected members of the family or unrelated control subjects. And ®nally, direct sequence analysis revealed a single base substitution introducing a premature stop codon in the cytoplasmic C-terminus of bENaC (Fig. 3). Thus, 31 years after the initial description of this family by Liddle the disease causing mutation had been identi®ed (Shimkets et al., 1994). Disease causing mutations increase channel activity Analyzing four other kindreds with Liddle's syndrome the group of Lifton identi®ed three more mutations all affecting the cytoplasmic C-terminus of bENaC by either introducing a premature stop codon or a frameshift mutation altering the encoded protein sequence (Shimkets et al., 1994; Fig. 3). The formal proof that these truncations/frameshift mutations cause Liddle's syndrome requires functional analysis demonstrating the effect on channel activity. Schild et al. (1995), therefore, co-expressed the mutant b subunit with wild-type a and g subunits in Xenopus oocytes. They could show an approximately 3-fold increase in the amiloride-sensitive Na 1 current for the mutant compared to wild-type channels. These results provided direct evidence that the mutation from Liddle's kindred causes constitutive channel hyperactivity explaining the disease's pathophysiology. Moreover, deletion of the cytoplasmic C-terminus of gENaC proved to be functionally equivalent to the deletion of the bENaC C-terminus (Schild et al., 1995). These results from Xenopus oocytes suggested that gENaC is another candidate gene for Liddle's disease and found a con®rmation when Lifton's group identi®ed a Liddle's syndrome kindred in which the cytoplasmic C-terminus of gENaC was truncated (Hansson et al., 1995a; Fig. 3). The locus for human gENaC is in close proximity to human bENaC on chromosome 16p (Hansson et al., 1995a). This result showed genetic heterogeneity of Liddle's syndrome and indicated roles of b and g subunits in the negative regulation of channel activity. Disease causing mutations affect a proline-rich motif In order to identify the precise amino acids in which mutation leads to increased channel activity Schild et al. (1996) performed a systematic mutagenesis study scanning the cytoplasmic Cterminus of the three ENaC subunits and expressing mutants in Xenopus oocytes. They identi®ed a short proline-rich motif (PPPxY) present in each subunit as the target sequence for Liddle's disease. Again, these results from Xenopus oocytes were con®rmed by identi®cation of Liddle's syndrome kindreds who show a

Liddle's syndrome and PHA type I

283

Fig. 3. Identi®ed mutations in Liddle's syndrome. Linear sequences from the M2 transmembrane segment are schematically shown. Black bars indicate wild-type sequences; the approximate position of the proline-rich segment is shown. Altered sequences caused by frameshift mutations are indicated by gray bars, missense mutations by gray letters. Liddle' s original kindred has the R564* mutation.

single missense mutation in one of the identi®ed amino acids (Hansson et al., 1995b; Tamura et al., 1996; Inoue et al., 1998; Fig. 3). Diagnosis of Liddle's syndrome by genetic or physiological analysis of ENaC Knowing the molecular cause of Liddle's syndrome, molecular analysis of the carboxy-terminus of b and gENaC can now be used as a diagnostic tool (Jackson et al., 1998). This is especially useful since Liddle's syndrome has a variable pene-

284

S. GruÈnder

trance, probably modi®ed by other factors in¯uencing blood pressure (Jeunemaitre et al., 1997). More recently it has been reported that nasal potential difference (PD), which is used for detecting increased sodium channel activity in cystic ®brosis, is changed in patients with Liddle's syndrome as compared to normotensive control persons (Baker et al., 1998a). The change in nasal PD after application of amiloride was greater in Liddle's patients than in controls (Baker et al., 1998a), providing the ®rst in vivo demonstration of increased ENaC activity in Liddle's syndrome. Molecular mechanism of Liddle's syndrome There are two questions which are raised by the identi®cation of mutations in Liddle's syndrome. First, what is the mechanism by which mutations at the cytoplasmic C-terminus lead to increased ENaC activity? And second, what is the signi®cance of the cytoplasmic proline-rich motif? Mutations in Liddle's patients increase open probability and surface expression For the ®rst question there are two principal possibilities: either the channel open probability (Po) could be increased or the number of active channels at the cell surface could be increased. In a ®rst study to address this question, Po of mutants versus wild-type channels was not found to be statistically different and a semiquantitative approach pointed to an increased surface expression of mutant channels (Snyder et al., 1995). A different study used a quantitative approach based on the binding of an iodinated antibody directed against an extracellular epitope to determine channel number in Xenopus oocytes. The authors found both an increase in surface expression of mutant channels and a predominant change in channel gating (Firsov et al., 1996). A deeper understanding, though, had to await a more detailed understanding of the underlying molecular processes. Mutations in Liddle's patients impair degradation of ENaC The cytoplasmic PY motif of ENaC interacts with Nedd4 Deletion of the cytoplasmic proline-rich motif as the disease-causing mutation suggested interaction with a regulatory repressive protein. In order to identify such a putative protein, Staub and Rotin performed a two-hybrid screen using the C-terminus of bENaC as a bait (Staub et al., 1996). They identi®ed clones coding for neural precursor cells expressed developmentally down-regulated (Nedd4). Nedd4 is a cytoplasmic protein containing three or four copies of the WW domain, a domain mediating interaction with proline-rich sequences with a consensus of XPPXY (PY motif), a Ca lipid binding domain (C2 domain) and a domain showing homology to a ubiquitin ligase (Hect domain). The authors could show that Nedd4 interacts with the PY motif of ENaC subunits via its WW domains (Staub et al., 1996). It was, therefore,

Liddle's syndrome and PHA type I

285

intriguing to speculate that the putative repressive action of Nedd4 is mediated by its ubiquitin-ligase domain. Nedd4 mediates ubiquitination of ENaC Ubiquitin-ligases mediate covalent attachment of ubiquitin, a 76 amino acid protein, or a poly-ubiquitin tree to lysine residues of the target protein thereby tagging it for rapid degradation by the proteasome, a cytosolic protease complex. Staub and colleagues could indeed show that ENaC is ubiquitinated on N-terminal lysine residues of its a and g subunits in vivo (Staub et al. 1997). Mutation of these lysine residues lead to channel hyperactivity by increasing the amount of surface expressed channels (Staub et al. 1997). Moreover, the authors suggested that the assembled abgENaC complex is degraded by the lysosomal degradation system. This interesting ®nding could reconcile data from Canessas group showing that ENaC normally is retrieved from the cell surface via endosomes and that this retrieval mechanism is impaired by Liddle's mutants (Shimkets et al., 1997). Finally it could be shown that overexpression of Nedd4 leads to decreased channel activity due to decreased surface expression, an effect which relies on the presence of the Cterminal PY motif (Goulet et al., 1998; Abriel et al., 1999). Together, these results suggest a model in which Nedd4 tags surface expressed ENaC molecules with ubiquitin for endosomal retrieval and ultimately lysosomal degradation. As expected with this model ENaC wild-type has a high turnover and a short halflife, a status not uncommon for a highly regulated molecule, and mutants which are less ef®ciently tagged by ubiquitin, show increased half-life and increased surface expression (Shimkets et al., 1997; Staub et al., 1997) leading to constitutive hyperreabsorption of Na 1 (Fig. 4). Mutations in Liddle's patients impair Na +-dependent downregulation of ENaC What about the initial ®nding that mutations increase predominantly channel open probability (Firsov et al., 1996)? A study by Kellenberger et al. (1998) using Xenopus oocytes suggests that the PY motif is not only implicated in channel degradation but also in feedback-regulation by intracellular Na 1 (Nai1). Intracellular Na 1 at concentrations .25 mM inhibit the activity of wild-type but not of mutant ENaC, an effect which was predominantly due to increased mean open probability and only to a smaller extent to increased surface expression (Kellen berger et al., 1998). A different study showed that an epithelial Na 1 channel from salivary duct cells is down-regulated by increased Nai1 and that this inhibition is mediated by Nedd4 in a ubiquitin-dependent fashion (Dinudom et al., 1998). Although the exact molecular nature of the channel from salivary glands is not known, these results suggest that at least the effect of Nai1 on surface expression is mediated by Nedd4. The molecular mechanism for the increase in channel open probability is not clear. Moreover, it was suggested that channel activity is downregulated also by an

286

S. GruÈnder

Fig. 4. Hypothetical model of the interaction of ENaC with Nedd4. Only one ENaC subunit is shown; Nedd4 may interact with three different subunits at the same time. For details see text.

increase in intracellular Ca 2+ (Ishikawa et al., 1998). Interestingly, it has been shown that Nedd4 is targeted to the plasma membrane by means of its C2 domain in a Ca 2+dependent fashion (Plant et al., 1997). Thus, Nedd4 could mediate several feedback loops to regulate ENaC activity (Fig. 4). Molecular analysis of the mechanism by which mutations found in patients with Liddle's syndrome increase ENaC activity has already lead to the identi®cation of some exciting and unexpected mechanisms for the regulation of ENaC activity. Most of the cited studies have worked with the Xenopus oocyte expression system which has proven to be very useful also for the quantitative analysis of channel mutants. Future experiments should concentrate on the functional analysis in systems which more closely resemble the in vivo situation. It would, nevertheless, not be surprising if in the future, advances in our understanding of the negative regulation of ENaC activity would come from experiments in systems which more closely resemble the in vivo situation, such as epithelial tissue and mouse models.

Liddle's syndrome and PHA type I

287

Outlook: Liddle's syndrome and the general hypertensive population One of the most promising features of Liddle's syndrome is that it might represent a model also for more common forms of salt-sensitive hypertension (Lifton, 1996). Among several polymorphisms in ENaC subunits identi®ed in the general hypertensive population (Chang and Fujita, 1996; Su et al., 1996; Melander et al., 1998; Persu et al., 1998), so far only one could be shown to co-segregate with the hypertensive phenotype in black people (Baker et al., 1998b). This mutation which changes threonine 594 at the cytoplasmic C-terminus of bENaC in a methionine (bT594M), though, does not lead to increased ENaC activity in Xenopus oocytes (Persu et al., 1998), pointing to the limits of this system for the evaluation of polymorphisms contributing to the phenotype in a multifactorial disease. Genetic analysis of rat models for general hypertension in which several genetic loci contribute to the hypertensive phenotype have so far failed to demonstrate any mutation in the coding sequence of ENaC subunits co-segregating with blood pressure (Huang et al., 1995; GruÈnder et al., 1997a; Kreutz et al., 1997). Thus, although genetic analysis of Liddle's syndrome has deeply increased our understanding of Na 1-reabsorption in the distal nephron, the analysis of polymorphisms in ENaC subunits and their contribution to general hypertension is still at its beginning. Pseudohypoaldosteronism type I Mutations in ENaC cause the recessive form of pseudohypoaldosteronism type I The mineralocorticoid-receptor and the ENaC are candidate genes for pseudohypoaldosteronism type I Clinical phenotype Pseudohypoaldosteronism type I (PHA 1) is an uncommon inherited disease characterized by renal salt wasting associated with hyperkalaemia and metabolic acidosis, failure to thrive and weight loss (Cheek and Perry, 1958). These features suggest suppressed aldosterone levels but actually aldosterone as well as renin levels in patients are elevated, explaining the classi®cation as pseudohypoaldosteronism. The diagnostic feature for PHA 1 is the resistance to administered mineralocorticoids. Other forms of mineralocorticoid resistance include pseudohypoaldosteronism type II (PHA II, also known as Gordon's syndrome; OMIM #145260), an inherited hypertension associated with hyperkalaemia (Gordon et al., 1970). PHA II can be effectively treated by thiazides, suggesting a renal defect proximal to the ENaC. Mode of inheritance of PHA 1 is either autosomal dominant (OMIM #177735) or autosomal recessive (OMIM #264350; Hanukoglu, 1991). The recessive form is

288

S. GruÈnder

generally more severe and can be life-threatening when not diagnosed whereas the dominant form is usually not life-threatening and carriers can even be asymptomatic (Kuhnle et al., 1996). Moreover, the recessive form is characterized by the persistence of the symptoms into adulthood and by multi-organ resistance to mineralocorticoids, whereas symptoms in patients affected by the dominant form disappear with age and are con®ned to the kidney (Hanukoglu, 1991; Kuhnle, 1997). PHA 1 can be effectively treated by salt supplementation; therapy in patients with the dominant form can usually be discontinued after the ®rst 2±3 years of age, whereas patients with the multisystem form of PHA 1 require very high amounts of salt in the diet life-long (as high as 45 g NaCl/day; Hanukoglu et al., 1997). Resistance to mineralocorticoids in PHA 1 had for a long time implied the mineralocorticoid-receptor (MR) as a candidate gene for the disease. PHA 1, however, represents in many respects the mirror-image of Liddle's syndrome, suggesting also ENaC as a candidate gene (see also Fig. 1). As it turned out, mutations in both genes account for either the recessive or the dominant form of this salt-losing syndrome. Mutations in ENaC or the MR cause pseudohypoaldosteronism Mutations in ENaC cause the recessive form of PHA 1 Testing ENaC as a candidate gene for PHA 1 Lifton and co-workers ®rst screened the a subunit for new variants co-segregating with the disease (Chang et al., 1996). In four out of seven kindreds investigated, they could identify new variants which were homozygous in affected persons but not in the parents or unaffected relatives and co-segregated with the disease. These mutations lead either to a frameshift prior to the ®rst transmembrane domain, leaving only the cytoplasmic N-terminus up to isoleucine 68 (aI68), or to a premature stop codon (aR508*), resulting in a protein missing the second transmembrane domain (Chang et al., 1996; Fig. 5). As the a subunit is essential for channel function and the second transmembrane domain is critical for the formation of the channel pore, these mutations can explain loss-of-function, explaining the disease's pathophysiology. aENaC maps to human chromosome 12p. In one of the remaining kindreds Lifton's group could identify a mutation in bENaC, homozygous in affected patients and co-segregating with the disease (Chang et al., 1996). This mutation changes a highly conserved glycine in a serine (bG37S; Fig. 5) and reduces channel activity in the Xenopus oocyte expression system (Chang et al., 1996). In a different kindred, Strautnieks et al. (1996) identi®ed a homozygous splice-site mutation in gENaC, leading to two new mRNAs arising from activation of an adjacent cryptic splice site or from skipping of the downstream exon. The ®rst mRNA is predicted to have three highly conserved amino acids in the extracellular domain replaced by a new one (gKYS106-108N; Fig. 5) and the second to change the protein sequence after the ®rst 105 amino acids (Strautnieks et al., 1996). Although the functional effects on ENaC activity have not been reported, these

Liddle's syndrome and PHA type I

289

Fig. 5. Identi®ed mutations in ENaC subunits in pseudohypoaldosteronism type I.

mutations most likely account for loss-of-function. More recently, a new missense mutation in aENaC has been identi®ed changing one of the completely conserved cysteines in the ectodomain into a tyrosine (aC133Y) and leading to temperaturedependent reduction of channel activity (GruÈnder et al., 1998; Fig. 5). Thus, mutations in ENaC subunits can account for PHA 1. In contrast to Liddle's syndrome where gain-of-function mutations are focused on the PY motif at the Cterminus of either b or gENaC, loss-of-function mutations, not unexpectedly, are scattered throughout the coding sequence of all three subunits (Fig. 5). These mutations all affected offspring from consanguineous union which suffered from the severe, autosomal recessive form of PHA 1. Because ENaC is a multimeric channel, it was, though, possible, that dominant negative mutations in one of the subunits could also account for the dominant form of PHA 1. Mutations in the MR cause the dominant form of PHA 1 Studying six families with the milder form of PHA 1, Lifton's group could not identify any mutation in ENaC subunits (Geller et al., 1998). Therefore, they screened the mineralocorticoid receptor (MR) gene and identi®ed mutations in ®ve families which delete the DNA binding domain and/or the hormone binding domain. These mutations are heterozygous in affected persons and co-segregate with the disease, thus accounting for the dominant form of PHA 1 (Geller et al., 1998). Loss of MR function is very likely due to haploinsuf®ciency. Molecular analysis of disease-causing mutations Loss-of-function mutations often disrupt the protein structure. Missense mutations, though, introduce more subtle changes and may provide information on amino acids critical for normal protein function. Missense mutations in the PY motif of ENaC subunits are one example and

290

S. GruÈnder

the bG37S mutation in autosomal recessive PHA 1 provides another one. In an attempt to understand the mechanism by which this missense mutation reduces channel activity it could be shown that this highly conserved glycine is important for proper gating of ENaC (GruÈnder et al., 1997b). As the gating of this channel might be a means to regulate channel activity, molecular analysis of PHA 1 mutations may contribute to a better understanding of ENaC function and regulation. Genotype/phenotype relationships in PHA 1 Lessons from animal models Disruption of the a ENaC gene locus results in a severe pulmonary phenotype in mice Disruption of the gene for aENaC by gene targeting leads to respiratory distress and death during the ®rst 40 h after birth in aENaC (2/2) mice (Hummler et al., 1996), demonstrating the importance of ENaC for clearing the lungs from liquid during the adaptation to air breathing. Rescue of the pulmonary phenotype in mice by introduction of an aENaC transgene in the genetic (2/2) background (aENaC (2/2)Tg) leads to a phenotype similar to PHA 1 with metabolic acidosis, urinary salt-wasting, growth retardation, and 50% mortality (Hummler et al., 1997). Disruption of the b or g ENaC gene locus leads to PHA 1 Disruption of the gene for gENaC also leads to premature death but in this case very likely due to hyperkalaemia (Barker et al., 1998). Homozygous (2/2) pups cleared their lungs nearly normal from water but serum K 1 levels were almost twice as high in (2/2) mice compared to (1/1) wild-type animals (Barker et al., 1998). Moreover, mice exhibited hyponatraemia despite increased serum aldosterone levels, mimicking the phenotype of humans suffering from the recessive form of PHA 1. As expression of a and b, but not of b and gENaC alone confers activity in heterologous expression systems, these results suggest that residual ENaC activity in gENaC knock-out mice is suf®cient to support lung liquid clearance but not to maintain electrolyte balance in the distal nephron. A very similar phenotype with early death due to severe hyperkalaemia is seen in bENaC (2/2) mice (McDonald et al., 1999). Interestingly, mice with very low but signi®cant levels ( < 1%) of bENaC mRNA become symptomatic only under challenge with a low-salt diet (0.1g Na 1/ kg; Pradervand et al., 1999). Disruption of the MR gene locus leads to PHA 1 Disruption of the mineralocorticoid receptor (MR) gene in mice (MR 2/2) results in a phenotype very similar to the one seen in b and gENaC (2/2) mice, that is a dramatic dehydration due to loss of salt and water associated with hyperkalaemia and premature death at around postnatal day 10 (Berger et al., 1998). Mice

Liddle's syndrome and PHA type I

291

heterozygous for the mutation (MR 1/2), though, grew normally and were largely asymptomatic. Plasma levels of aldosterone and renin as well as urinary Na 1 loss were, nevertheless, signi®cantly increased in these animals (Berger et al., 1998), resembling the situation in patients with the mild form of PHA 1. Whether challenge by a low salt diet can induce an overt PHA 1 phenotype in heterozygotes was not reported. Genotype/phenotype relationships in PHA 1 The obvious phenotypic difference in aENaC (2/2) mice compared to humans with a homozygous mutation in aENaC is most easily explained by a species difference which was previously found for mice de®cient for the cystic ®brosis transmembrane conductance regulator gene. Either, the mechanism to clear the lungs from liquid is signi®cantly different in both species or another subunit, for example dENaC (Waldmann et al., 1995), could substitute for aENaC in humans but not in mice. It is in this context interesting that dENaC has been cloned so far only from humans but seems not to be present in rat (Drummond et al., 1998; and S. GruÈnder unpublished results). Another possibility is that the mutant a subunits in humans still confer some residual activity which would be suf®cient to overcome the pulmonary phenotype. This hypothesis is supported by the phenotype of b and gENaC (2/2) mice with no pulmonary phenotype. In fact, very low, but signi®cant channel activity could be detected in the Xenopus expression system with the aI68frameshift mutation (Bonny et al., 1997). Interestingly, some patients with the severe form of PHA 1 suffer from respiratory tract infections and other symptoms similar to cystic ®brosis (Hanukoglu et al., 1994; Marthinsen et al., 1998). Whether this pulmonary phenotype correlates with a mutation in aENaC is to the knowledge of the author not established. Life-threatening hyperkalaemia in b and gENaC (2/2) mice as well as in aENaC (2/2) Tg mice supports the ®nding that the severe recessive form of PHA 1 is caused by mutations in ENaC. Heterozygous animals for a loss in either of the three ENaC subunits were all asymptomatic (Hummler et al., 1996; Barker et al., 1998; McDonald et al., 1999), con®rming that one allele confers suf®cient ENaC activity. This is also nicely illustrated by the fact that even mice with very low levels ( < 1%) of bENaC mRNA become symptomatic only under a low salt diet (Pradervand et al., 1999). Dominant negative mutations in one of the ENaC subunits which could account for a dominant phenotype in patients have so far not been identi®ed but still remain a possibility. Largely asymptomatic mice with one defective allele for the MR support the ®nding that the mild, dominant form of PHA 1 is caused by mutations in the MR. The mild phenotype can be explained by residual receptor activity due to the second allele and/or receptor-independent ENaC activity. Improvement with age in this

292

S. GruÈnder

form of PHA 1 points to the crucial role of aldosterone in salt homeostasis early in life that diminishes with age. Outlook Is this the end of the story? Probably not. There are still some PHA 1 patients who did not show a mutation in either ENaC or the MR (Geller et al., 1998). And why do some patients with mutations in the MR develop PHA 1 while others remain asymptomatic? What are the other important factors controlling salt balance and how do they interact? In the same direction, it remains to be established to which degree variants with uncertain functional signi®cance identi®ed in ENaC could account for sporadic cases of PHA 1 (Zennaro, 1998). This all leads to the important question: to which degree do variants in ENaC or the MR could predispose ± or protect ± from other multifactorial diseases such as hypertension?

Note added in proof Recently it was shown that systolic but not diastolic blood pressure co-segregates with markers on chromosome 16p12, the locus for b and gENaC, in an Australian population (Wong et al., 1999). A different study suggested that US blacks, who have a higher risk of developing hypertension compared to whites, could have an increased ENaC activity (Ambrosius et al., 1999). Molecular variants in ENaC which are more frequent in blacks showed no co-segregation with blood pressure (Ambrosius et al., 1999). On the other hand, two other studies did not ®nd a cosegregation of polymorphisms in the gene for gENaC with hypertension (Persu et al., 1999) or an increase in ENaC activity in whites with essential hypertension as assessed by nasal potential difference measurements (Baker et al., 1999), respectively. Together, these results suggest that molecular variants in ENaC may contribute to the hypertensive phenotype in subsets of the general hypertensive population. Studying pulmonary characteristics in patients suffering from pseudohypoaldosteronism type 1 Kerem and co-workers could show that patients fail to absorb liquid from airway surfaces resulting in an increased liquid volume in the airways (Kerem et al., 1999). Moreover, Malagon-Rogers reported a patient suffering from respiratory distress syndrome and PHA-1 (Malagon-Rogers, 1999). These results con®rm the importance of ENaC-mediated sodium transport for regulation of extracellular volume in the lung.

Liddle's syndrome and PHA type I

293

Acknowledgements The author is grateful to L. Schild for suggestions, to M. Mark and J. Reichl for improving style and grammar, and to E. Steinhilber for bibliographic work. References Abriel, H., Lof®ng, J., Rebun, J.F., Pratt, J.H., Schild, L., Horisberger, J.-D., Rotin, D., Staub, O., 1999. Defective regulation of the epithelial Na 1 channel by Nedd4 in Liddle's syndrome. J. Clin. Invest. 103 (5), 667±673. Ambrosius, W.T., Bloem, L.J., Zhou, L., Rebhun, J.F., Snyder, P.M., Wagner, M.A., Guo, C., Pratt, J.H., 1999. Genetic variants in the epithelial sodium channel in relation to aldosterone and potassium excretion and risk for hypertesion. Hypertension 34, 631±637. Baker, E.H., Jeunemaitre, X., Portal, A.J., Grimbert, P., Markandu, N., Persu, A., Corvol, P., MacGregor, G., 1998a. Abnormalities of nasal potential difference measurement in Liddle's syndrome. J. Clin. Invest. 102 (1), 10±14. Baker, E.H., Dong, Y.B., Sagnella, G.A., Rothwell, M., Onipinla, A.K., Markandu, N.D., Capuccio, F.P., Cook, D.G., Persu, A., Corvol, P., Jeunemaitre, X., Carter, N.D., MacGregor, G.A., 1998b. Association of hypertension with T594M mutation in b subunit of epithelial sodium channels in black people resident in London. Lancet 351, 1388±1392. Baker, E.H., Portal, A.J., McElvaney, T.A., Blackwood, A.M., Miller, M.A., Markandu, N.D., MacGregor, G.A., 1999. Epithelial sodium channel activity is not increased in hypertension in whites. Hypertension 33, 1031±1035. Barker, P.M., Nguyen, M.S., Gatzy, J.T., Grubb, B., Norman, H., Hummler, E., Rossier, B., Boucher, R., Koller, B., 1998. Role of gENaC subunit in lung liquid clearance and electrolyte balance in newborn mice. J. Clin. Invest. 102 (8), 1634±1640. Berger, S., Bleich, M., Schmid, W., Cole, T.J., Peters, J., Watanabe, H., Kriz, W., Warth, R., Greger, R., SchuÈtz, G., 1998. Mineralocorticoid receptor knockout mice: pathophysiology of Na 1 metabolism. Proc. Natl. Acad. Sci. 95, 9424±9429. Bonny, O., Rossier, B.C., Hummler, E., 1997. Functional analysis of human aENaC mutation (I68fr) causing PHA 1. J. Am. Soc. Nephrol. 8, 30A. Botero-Velez, M., Curtis, J.J., Warnock, D.G., 1994. Liddle's syndrome revisited ± a disorder of sodium reabsorption in the distal tubule. N. Engl. J. Med. 330 (3), 178±181. Canessa, C.M., Horisberger, J.-D., Schild, L., Rossier, B.C., 1995. Expression cloning of the epithelial sodium channel. Kidney Int. 48, 950±955. Chang, H., Fujita, T., 1996. Lack of mutations in epithelial sodium channel b-subunit gene in human subjects with hypertension. J. Hypertens. 14, 1417±1419. Chang, S.S., GruÈnder, S., Hanukoglu, A., RoÈsler, A., Mathew, P.M., Hanukoglu, I., Schild, L., Lu, Y., Shimkets, R.A., Nelson-Williams, C., Rossier, B.C., Lifton, R.P., 1996. Mutations in subunits of the epithelial sodium channel cause salt wasting with hyperkalaemic acidosis, pseudohypoaldosteronism type 1. Nature Genet. 12, 248±253. Cheek, D., Perry, J.W., 1958. A salt wasting syndrome in infancy. Arch. Dis. Child. 33, 252±256. Dinudom, A., Harvey, K.F., Komwatana, P., Young, J.A., Kumar, S., Cook, D.I., 1998. Nedd4 mediates control of an epithelial Na 1 channel in salivary duct cells by cytosolic Na 1. Proc. Natl. Acad. Sci. USA 95, 7169±7173. Drummond, H.A., Price, M.P., Welsh, M.J., Abboud, F.M., 1998. A molecular component of the arterial baroreceptor mechanotransducer. Neuron 21, 1435±1441.

294

S. GruÈnder

Duc, C., Farman, N., Canessa, C.M., Bonvalet, J.-P., Rossier, B.C., 1994. Cell-speci®c expression of epithelial sodium channel a, b, g subunits in aldosterone-responsive epithelia from the rat: localization by in situ hybridization and immunocytochemistry. J. Cell Biol. 127 (6), 1907±1921. Firsov, D., Schild, L., Gautschi, I., Merillat, A.-M., Schneeberger, E., Rossier, B.C., 1996. Cell surface expression of the epithelial Na channel and a mutant causing Liddle syndrome: A quantitative approach. Proc. Natl. Acad. Sci. USA 93, 15370±15375. Firsov, D., Robert-Nicoud, M., GruÈnder, S., Schild, L., Rossier, B.C., 1999. Mutational analysis of cyteine-rich domains of the epithelium sodium channel (ENaC). J. Biol. Chem. 274 (5), 2743±2749. Geller, D.S., Rodriguez-Soriano, J., Boado, A.V., Schifter, S., Bayer, M., Chang, S.S., Lifton, R.P., 1998. Mutations in the mineralocorticoid receptor gene cause autosomal dominant pseudohypoaldosteronism type I. Nature Genet. 19, 279±281. Gordon, R.D., 1995. Heterogeneous hypertension. Nature Genet. 11, 6±9. Gordon, R.D., Geddes, R.A., Pawsey, C.G., O'Halloran, M.W., 1970. Hypertension and severe hyperkalaemia associated with suppression of renin and aldosterone and completely reversed by dietary sodium restriction. Aust. Ann. Med. 4, 287±294. Goulet, C.C., Volk, K.A., Adams, C.M., Prince, L.S., Stokes, J.B., Snyder, P.M., 1998. Inhibition of the epithelial Na 1 channel by interaction of Nedd4 with a PY motif deleted in Liddle's syndrome. J. Biol. Chem. 273 (45), 30012±30017. GruÈnder, S., Rossier, B.C., 1997. A reappraisal of aldosterone effects on the kidney: new insights provided by epithelial sodium channel cloning. Curr. Opin. Nephrol. Hypertens. 6, 35±39. GruÈnder, S., Zagato, L., Yagil, C., Yagil, Y., Sassard, J., Rossier, B.C., 1997a. Polymorphisms in the carboxy-terminus of the epithelial sodium channel in rat models for hypertension. J. Hypertens. 15, 173±179. GruÈnder, S., Firsov, D., Chang, S.S., Fowler Jaeger, N., Gautschi, I., Schild, L., Lifton, R.P., Rossier, B.C., 1997b. A mutation causing pseudohypoaldosteronism type 1 identi®es a highly conserved glycine that is involved in the gating of the epithelial sodium channel. EMBO J. 16, 899±907. GruÈnder, S., Chang, S.S., Lifton, R.P., Rossier, B.C., 1998. PHA 1: a novel thermosensitive mutation in the ectodomain of alpha ENaC. J. Am. Soc. Nephrol. 9, S069. Hansson, J.H., Nelson-Williams, C., Suzuki, H., Schild, L., Shimkets, R., Lu, Y., Canessa, C., Iwasaki, T., Rossier, B.C., Lifton, R.P., 1995a. Hypertension caused by a truncated epithelial sodium channel b subunit: genetic heterogeneity of Liddle syndrome. Nature Genet. 11, 76±82. Hansson, J.H., Schild, L., Lu, Y., Wilson, T.A., Gautschi, I., Shimkets, R., Nelson-Williams, C., Rossier, B.C., Lifton, R.P., 1995b. A de novo missense mutation of the b subunit of the epithelial sodium channel causes hypertension and Liddle syndrome, identifying a proline-rich segment critical for regulation of channel activity. Proc. Natl. Acad. Sci. USA 92, 11495±11499. Hanukoglu, A., 1991. Type I pseudohypoaldosteronism includes two clinically and genetically distinct entities with either renal or multiple target organ defects. J. Clin. Endocrin. Metab. 73, 936±944. Hanukoglu, A., Bistritzer, T., Rakover, Y., Mandelberg, A., 1994. Pseudohypoaldosteronism with increased sweat and saliva electrolyte values and frequent lower respiratory tract infections mimicking cystic ®brosis. J. Pediatr. 125 (5), 752±755. Hanukoglu, A., Joy, O., Steinitz, M., RoÈsler, A., Hanukoglu, I., 1997. Pseudohypoaldosteronism due to renal and multisystem resistance to mineralocorticoids respond differently to carbenoxolone. J. Steroid Biochem. Mol. Biol. 60 (2), 105±112. Horisberger, J.-D., 1998. Amiloride-sensitive Na channels. Curr. Opin. Cell Biol. 10, 443±449. Huang, H., Pravenec, M., Wang, J.-M., Kren, V., StLezin, E., Szpirer, C., Szpirer, J., Kurtz, T.W., 1995. Mapping and sequence analysis of the gene encoding the beta subunit of the epithelial sodium channel in experimental models of hypertension. J. Hypertens. 13, 1247±1251. Hummler, E., Barker, P., Gatzy, J., Beermann, F., Verdumo, C., Schmidt, A., Boucher, R., Rossier, B.C.,

Liddle's syndrome and PHA type I

295

1996. Early death due to defective neonatal lung liquid clearance in aENaC-de®cient mice. Nature Genet. 12, 325±328. Hummler, E., Barker, P., Talbot, C., Wang, Q., Verdumo, C., Grubb, B., Gatzy, J., Burnier, M., Horisberger, J.-D., Beermann, F., Boucher, R., Rossier, B.C., 1997. A mouse model for the renal saltwasting syndrome pseudohypoaldosteronism. Proc. Natl. Acad. Sci. USA 94, 11710±11715. Inoue, J., Iwaoka, T., Tokunaga, H., Takamune, K., Naomi, S., Araki, M., Takahama, K., Yamaguchi, K., Tomita, K., 1998. A family with Liddle's syndrome caused by a new missense mutation in the beta subunit of the epithelial sodium channel. J. Clin. Endocrinol. Metab. 83, 2210±2213. Ishikawa, T., Marunaka, Y., Rotin, D., 1998. Electrophysiological characterization of the rat epithelial Na 1 channel (rENaC) expressed in MDCK cells. J. Gen. Physiol. 111, 825±846. Jackson, S.N.J., Williams, B., Houtman, P., Trembath, R.C., 1998. The diagnosis of Liddle syndrome by identi®cation of a mutation in the b subunit of the epithelial sodium channel. J. Med. Genet. 35, 510± 512. Jeunemaitre, X., Bassilana, F., Persu, A., Dumont, C., Champigny, G., Lazdunski, M., Corvol, P., Barbry, P., 1997. Genotype±phenotype analysis of a newly discovered family with Liddle's syndrome. J. Hypertens. 15, 1091±1100. Kellenberger, S., Gautschi, I., Rossier, B.C., Schild, L., 1998. Mutations causing Liddle's syndrome reduce sodium-dependent downregulation of the epithelial sodium channel in the Xenopus oocyte expression system. J. Clin. Invest. 101 (12), 2741±2750. Kerem, E., Bistritzer, T., Hanukoglu, A., Hofmann, T., Zhou, Z., Bennett, W., MacLaughlin, E., Barker, P., Nash, M., Quittell, L., Boucher, R., Knowles, M.R., 1999. Pulmonary epithelial sodium channel dysfunction and excess airway liquid in pseudohypoaldosteronism. N. Engl. J. Med. 341, 156±162. Kreutz, R., Struk, B., Rubattu, S., HuÈbner, N., Szpirer, J., Szpirer, C., Ganten, D., Lindpaintner, K., 1997. Role of the a-, b -, g-subunits of epithelial sodium channel in a model of polygenic hypertension. Hypertension 29, 131±136. Kuhnle, U., 1997. Pseudohypoaldosteronism: mutation found, problem solved?. Mol. Cell. Endocrinol. 133, 77±80. Kuhnle, U., Hinkel, G.K., Hubl, W., Reichelt, T., 1996. Pseudohypoaldosteronism: family studies to identify asymptomatic carriers by stimulation of the renin-aldosterone system. Horm. Res. 46, 124± 129. Liddle, G.W., Bledsoe, T., Coppage Jr., W.S., 1963. A familial renal disorder simulating primary aldosteronism but with negligible aldosterone secretion. Trans. Assoc. Am. Phys. 76, 199±213. Lifton, R.P., 1996. Molecular genetics of human blood pressure variation. Science 272, 676±680. Malagon-Rogers, M., 1999. A patient with pseudohypoaldosteronism type 1 and respiratory distress syndrome. Pediatr. Nephrol. 13, 484±486. Marthinsen, L., KornfaÈlt, R., Aili, M., Andersson, D., Westgren, U., Schaedel, C., 1998. Recurrent Pseudomonas bronchopneumonia and other symptoms as in cystic ®brosis in a child with type I pseudohypoaldosteronism. Acta Paediatr. 87, 472±474. McDonald, F., Yang, B., Hrstka, R.F., Drummond, H.A., Tarr, D.E., McCray Jr., P.B., Stokes, J.B., Welsh, M.J., Williamson, R.A., 1999. Disruption of the b subunit of the epithelial Na 1 channel in mice: hyperkalaemia and neonatal death associated with a pseudohypoaldosteronism phenotype. Proc. Natl. Acad. Sci. USA 96, 1727±1731. Melander, O., Orho, M., Fagerudd, J., Bengtsson, K., Groop, P.-H., Mattiasson, I., Groop, L., HultheÂn, U.L., 1998. Mutations and variants of the epithelial sodium channel gene in Liddle's syndrome and primary hypertension. Hypertension 31, 1118±1124. Persu, A., Barbry, P., Bassilana, F., Houot, A.-M., Mengual, R., Lazdunski, M., Corvol, P., Jeunemaitre, X., 1998. Genetic analysis of the b subunit of the epithelial Na 1 channel in essential hypertension. Hypertension 32, 129±137.

296

S. GruÈnder

Persu, A., Coscoy, S., Houot, A.-M., Corvol, P., Barbry, P., Jeunemaitre, X., 1999. Polymorphisms of the g subunit of the epithelial Na + channel in essential hypertension. J. Hypertens. 17, 639±645. Plant, P., Yeger, H., Staub, O., Howard, P., Rotin, D., 1997. The C2 domain of the ubiquitin protein ligase Nedd4 mediates Ca 21-dependent plasma membrane localization. J. Biol. Chem. 272 (51), 32329± 32336. Pradervand, S., Barker, P.M., Wang, Q., Ernst, S.A., Beermann, F., Grubb, B.R., Burnier, M., Schmidt, A., Bindels, R.J., Gatzy, J.T., Rossier, B.C., Hummler, E., 1999. Salt restriction induces pseudohypoaldosteronism type 1 in mice expressing low levels of the b-subunit of the amiloride-sensitive epithelial sodium channel. Proc. Natl. Acad. Sci. USA 96, 1732±1737. Rossier, B.C., Palmer, L.G. (1992). Mechanisms of aldosterone action on sodium and potassium transport. In: Seldin, D.W., Giebisch G. (Eds.). The Kidney, Physiology and Pathophysiology. Raven Press, New York, pp. 1373±1409. Rotin, D., Bar-Sagi, D., O'Brodovich, H., Merilainen, J., Lehto, V.P., Canessa, C.M., Rossier, B.C., Downey, G.P., 1994. An SH3 binding region in the epithelial sodium channel (arENaC) mediates its localization at the apical membrane. EMBO J. 13 (19), 4440±4450. Schild, L., Canessa, C.M., Shimkets, R., Gautschi, I., Lifton, R.P., Rossier, B.C., 1995. A mutation in the epithelial sodium channel causing Liddle disease increases channel activity in the Xenopus laevis oocyte expression system. Proc. Natl. Acad. Sci. USA 92, 15699±15703. Schild, L., Lu, Y., Gautschi, I., Schneeberger, E., Lifton, R.P., Rossier, B.C., 1996. Identi®cation of a PY motif in the epithelial Na channel subunits as a target sequence for mutations causing channel activation found in Liddle syndrome. EMBO J. 15 (10), 2381±2387. Shimkets, R.A., Warnock, D.G., Bositis, C.M., Nelson-Williams, C., Hansson, J.H., Schambelan, M., Gill Jr., J.R., Ulick, S., Milora, R.V., Findlin, J.W., Canessa, C.M., Rossier, B.C., Lifton, R.P., 1994. Liddle's syndrome: heritable human hypertension caused by mutations in the b subunit of the epithelial sodium channel. Cell 79, 407±414. Shimkets, R.A., Lifton, R.P., Canessa, C.M., 1997. The activity of the epithelial sodium channel is regulated by clathrin-mediated endocytosis. J. Biol. Chem. 272 (41), 25537±25541. Snyder, P.M., Price, M.P., McDonald, F.J., Adams, C.M., Volk, K.A., Zeiher, B.G., Stokes, J.B., Welsh, M.J., 1995. Mechanism by which Liddle's syndrome mutations increase activity of a human epithelial Na 1 channel. Cell 83, 969±978. Staub, O., Dho, S., Henry, P.C., Correa, J., Ishikawa, T., McGlade, J., Rotin, D., 1996. WW domains of Nedd4 bind to the proline-rich PY motifs in the epithelial Na 1 channel deleted in Liddle's syndrome. EMBO J. 15 (10), 2371±2380. Staub, Gautschi, Ishikawa, Breitschopf, Ciechanover, Schild, Rotin, 1997. Regulation of stability and function of the epithelial Na 1 channel (ENaC) by ubiquitination. EMBO J. 16 (21), 6325±6336. Strautnieks, S.S., Thompson, R.J., Gardiner, R.M., Chung, E., 1996. A novel splice-site mutation in the b subunit of the epithelial sodium channel gene in three pseudohypoaldosteronism type 1 families. Nature Genet. 13, 248±250. Su, Y.R., Rutkowski, M.P., Klanke, C.A., Wu, X., Cui, Y., Pun, R.Y.K., Carter, V., Reif, M., Menon, A.G., 1996. A novel variant of the b subunit of the amiloride-sensitive sodium channel in african americans. J. Am. Soc. Nephrol. 7, 2543±2549. Tamura, H., Schild, L., Enomoto, N., Matsui, N., Marumo, F., Rossier, B.C., Sasaki, S., 1996. Liddle disease caused by a missense mutation of b subunit of the epithelial sodium channel gene. J. Clin. Invest. 97, 1780±1784. Vania, A., Tucciarone, L., Mazzeo, D., Capodaglio, P.F., Cugini, P., 1997. Liddle's syndrome: a 14-year follow-up of the youngest diagnosed case. Pediatr. Nephrol. 11, 7±11. Waldmann, R., Champigny, G., Bassilana, F., Voilley, N., Lazdunski, M., 1995. Molecular cloning and functional expression of a novel amiloride-sensitive Na 1 channel. J. Biol. Chem. 270, 27411±27414.

Liddle's syndrome and PHA type I

297

Wong, Z.Y.H., Stebbing, M., Ellis, J.A., Lamantia, A., Harrap, S.B., 1999. Genetic linkage of b and g subunits of epithelial sodium channel to systolic blood pressure. Lancet 353, 1222±1225. Zennaro, M.-C., 1998. Syndromes of glucocorticoid and mineralocorticoid resistance. Eur. J. Endocrinol. 139, 127±138.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

299

Chapter 12

The role of ATP-sensitive K 1 channels in familial hyperinsulinism Benjamin Glaser a, Lydia Aguilar-Bryan b a Department of Endocrinology and Metabolism, Hadassah University Hospital, Jerusalem, Israel Department of Medicine, Division of Endocrinology, Baylor College of Medicine, Houston, TX, USA

b

Abstract Familial hyperinsulinism, also known as persistent hyperinsulinemic hypoglycemia of infancy (HI or PHHI), is a very heterogeneous disease characterized by the co-existence of severe hypoglycemia and inappropriately elevated serum insulin levels. Although mutations in at least four different genes may cause a similar clinical syndrome, the most common form of this disease is caused by mutations in the two sub-units that form the b-cell ATP-sensitive K 1 channel (KATP). This channel regulates insulin secretion in pancreatic b-cells by linking glucose metabolism with the electrical activity of the cell. q 2000 Elsevier Science B.V. All rights reserved.

Introduction Case 1 The patient was born of unrelated, Ashkenazi Jewish parents after a normal 40 week pregnancy. At birth he was plethoric and hypotonic and weighed 4.5 kg. Two hours after birth, he was found to be unresponsive, with shallow and irregular breathing. Severe hypoglycemia was documented and the diagnosis of hyperinsulinsim was con®rmed biochemically. After the hypoglycemia was controlled with very high doses of intravenous glucose, treatment was attempted with diazoxide, somatostatin analog, glucagon and verapamil. However, the response was insuf®cient to achieve normoglycemia, and at 1 month of age an 80% pancreatectomy was performed. The histologic ®ndings were consistent with diffuse-hyperinsulinism of infancy (HI). After surgery hypoglycemia recurred, but was controllable using somatostatin analog treatment and frequent feedings. Clinical disease improved

300

B. Glaser, L. Aguilar-Bryan

slowly until the child entered clinical remission at the age of 3 years. However, at 15 years of age insulin-requiring diabetes mellitus was diagnosed. Case 2 The patient was born of unrelated Ashkenazi Jewish patients, with normal birth weight after an uneventful pregnancy. At the age of 1 year, epilepsy was diagnosed and he was treated with antiepileptic medication with good control of symptoms. At the age of 17 years, while on a strenuous hike, he had an epileptic seizure and was taken to a local emergency room where a routine glucose level was found to be low. Subsequent evaluation documented hyperinsulinemic hypoglycemia. In retrospect, most if not all of the `epileptic' events diagnosed since the age of 1 year were associated with increased physical activity and/or with decreased oral intake. Once hypoglycemia was controlled, all anti-epileptic medications could be stopped and no further seizures were observed. Hepatic venous sampling documented what has now been de®ned as diffuse pancreatic disease, and a 95% pancreatectomy was performed (Glaser et al., 1981). The patient recovered after a stormy post-operative course, but has been an insulin-requiring diabetic ever since. Persistent hyperinsulinemic hypoglycemia of infancy (PHHI), or simply HI is a clinically heterogeneous disease. Until a few years ago, nothing was known about the molecular etiology of the disease, but recent discoveries have proven that the clinical syndrome is also heterogeneous from the genetic point of view. In this chapter we will give an overview of our current knowledge of the clinical aspects of this disease, which is characterized by dysregulation of insulin secretion, and of the known molecular mechanisms that can cause this syndrome. Control of insulin secretion ± a short review To understand the pathophysiology and etiology of hyperinsulinism a basic understanding of how the b-cell regulates insulin secretion is necessary. Glucose control of insulin secretion is a very complex and incompletely understood process. A schematic representation of the major players in the regulation of what is known as the KATP dependent, glucose stimulated insulin secretion pathway (Aguilar-Bryan and Bryan, 1996; Aizawa et al., 1998; Straub et al., 1998) is shown in Fig. 1. Plasma glucose enters the b-cell through a speci®c glucose transporter (GLUT-2), a membrane-bound high Km glucose-speci®c transporter found primarily in the pancreatic b-cell, the liver and the kidney. Once in the b-cell, glucose is rapidly metabolized. The rate-limiting step in the metabolic process is glucokinase (GK), the ®rst enzyme in the cascade. Further metabolism results in changes in the nucleotide concentrations within the cytosol, speci®cally increased ATP/ADP ratio, that reduce the potassium conductance through the KATP channels, resulting in plasma membrane depolarization. This causes voltage dependent calcium channels to open

ATP-sensitive K 1 channels in familial hyperinsulinism

301

Fig. 1. Schematic illustration of the most important players in the KATP dependent pathway of glucosestimulated insulin secretion in pancreatic b-cells (see text for explanation and abbreviations).

and the increase in intracellular Ca 21 triggers the oscillatory behavior that drives insulin release. Nucleotides like ATP and compounds such as sulfonylureas, close the KATP channel and stimulate insulin secretion. MgADP and diazoxide inhibit secretion by opening these channels, keeping the membrane hyperpolarized, preventing the opening of the voltage-dependent calcium channels and blocking the subsequent increase in cytosolic calcium that triggers insulin release (AguilarBryan and Bryan, 1999). Clinical features Diagnostic criteria The differential diagnosis and treatment of HI has recently been reviewed (Glaser et al., 1999a). Brie¯y, the clinical diagnosis must be suspected as soon as signs of hyperinsulinism or of hypoglycemia are apparent. Typically, hyperinsulinism begins before birth causing macrosomy, clinically identical to that seen in infants of poorly controlled diabetic mothers. Although not universally present, this seems to correlate with the severity of the clinical phenotype. In neonates, hypoglycemia may be missed, since it can produce non-speci®c symptoms including apathy, hypotonia, epilepsy, apnea and even cardiac arrest. As emphasized by the two cases presented in the introduction, the severity of the disease and hence the clinical symptomatology can be extremely variable. This can

302

B. Glaser, L. Aguilar-Bryan

make the clinical diagnosis dif®cult, especially if the treating physician does not have a high index of suspicion. In the ®rst case described above, the treating physician was alert to the symptoms and made the initial diagnosis of hypoglycemia at the age of 2 h, thus preventing permanent brain damage. In contrast, in the second case the hypoglycemia was missed and the symptoms interpreted as being caused by idiopathic epilepsy. The correct diagnosis was not made for 17 years despite continued medical supervision. Delayed diagnosis frequently results in irreversible brain damage, since the brain uses glucose as its primary energy source. The biochemical diagnosis of hyperinsulinism is based on the ®nding of elevated insulin levels in the face of symptomatic hypoglycemia. The absolute insulin level may not be elevated when compared to normal fasting levels, but it is invariably elevated relative to the glucose level. b-Cells from HI patients continue to secrete insulin at all glucose levels, whereas normal b-cells completely shut down insulin secretion during hypoglycemia (Kaiser et al., 1990). Surrogate measures of insulin action are often helpful for making a rapid diagnosis. Inappropriately elevated insulin levels increase glucose uptake in muscle and fat, increase hepatic glycogen storage and decrease lipolysis and hepatic glucose production. The net result is low plasma glucose levels, inappropriately low plasma levels of ketone bodies, high hepatic glucose response to glucagon stimulation and very high glucose requirements (typically .15 mg/kg per min) needed to overcome the massive glucose uptake by muscle and fat. The latter is pathognomonic for hyperinsulinism of any etiology. In virtually all cases of moderate to severe hyperinsulinism, the treating physician will be alerted to the correct diagnosis by the simple observation that highdose glucose is required to prevent recurrent hypoglycemia. Therefore, it is usually quite easy to distinguish between non-hyperinsulinemic and hyperinsulinemic hypoglycemia. However, the differential diagnosis of the latter is more complex. The most common cause of this entity in the newborn period is poorly controlled maternal diabetes. The fetus of a poorly controlled diabetic mother is exposed to abnormally high concentrations of glucose, resulting in compensatory b-cell hyperplasia and hyperinsulinism. This increases somatic growth and results in macrosomia. Immediately after birth, the glucose supply from the mother is abruptly stopped. The hyperplastic neonatal b-cells are not able to completely shut off insulin secretion, resulting in hypoglycemia. Initially this condition can mimic HI exactly, however it is transient and insulin regulation of glucose normalizes within a few days or weeks. Histological ®ndings HI was initially called `nesidioblastosis' since the typical histology described at that time, involved apparent proliferation of poorly formed islets, isolated b-cells scattered throughout the exocrine tissue, and b-cells budding from exocrine ducts.

ATP-sensitive K 1 channels in familial hyperinsulinism

303

Subsequent studies showed that a very similar picture is also seen in normal fetal and neonatal pancreas, as well as in diseases such as cystic ®brosis, multiple endocrine neoplasia type 1 and pancreatitis, among others. Attempts at quantifying different islet cell types failed to demonstrate any consistent abnormality. For this reason the name 'nesidioblastosis', although still in common use it is not correct. Recent studies have shown that the only consistent histologic ®nding in HI is that the b-cells have large nuclei and abundant cytoplasm, signs of metabolic hyperactivity (Rahier et al., 1984; Ariel et al., 1988). For this reason, most authors now refer to the disease as HI or one of the many variations of this name, such as PHHI, congenital hyperinsulinism (CHI), etc. Two histopathological forms of the disease have recently been described, diffuse and focal (Goossens et al., 1989; Sempoux et al., 1998). In the diffuse form, all bcells appear affected, whereas in the focal form, a discrete region of the pancreas shows these characteristics of metabolic hyperactivity, while the remainder of the pancreas has small, compact b-cells with small nuclei and sparse cytoplasm, characteristic of resting cells. This focal region is not a well circumscribed adenoma as usually seen in adults with isolated insulinomas, but rather a poorly de®ned region of hyperplasia. Therapy The primary goal of treatment is to prevent hypoglycemia and thus prevent irreversible brain damage. A sequential approach to the treatment of HI has been recently published (Glaser et al., 1999a). Initial treatment consists of intravenous glucose administration at a rate suf®cient to maintain adequate glucose levels. Inhibitors of insulin action such as glucocorticoids and glucagon may be added, but these alone are rarely helpful. Drugs that inhibit insulin secretion are frequently used, either alone or in combination. Diazoxide, the most commonly used of the K 1 channel openers, binds to the sulfonylurea sub-unit of the b-cell KATP channel. By increasing the mean open time of the channel, diazoxide maintains the membrane hyperpolarized, preventing opening of voltage dependent calcium channels, and thus inhibiting insulin release. The long-acting somatostatin analog, octreotide, also suppresses normal insulin secretion, and appears to be useful in some cases. Recently, calcium channel blockers such as nifedipine have been tried, but with very limited success (Lindley et al., 1996; Bas, 1999; Lawson, 2000). In patients with severe clinical disease, medical management is usually of only limited ef®cacy. Partial pancreatectomy is an alternative option, but an extensive resection must be performed to prevent immediate recurrence. This is associated with a very high incidence of diabetes months to years after surgery (Leibowitz et al., 1995; Shilyansky et al., 1997; Cade, 1998; Dacoa-Voutetakis, 1998). The treat-

304

B. Glaser, L. Aguilar-Bryan

ing physician is thus left with a dilemma as to which therapeutic approach is most appropriate for a given patient. There is no consensus in the literature to aid in the decision making process, and the clinical decision must be based on the current knowledge of the long-term outcome of the various treatment options (Table 1). Some centers place avoidance of hypoglycemia as the single most important goal, and therefore recommend early extensive pancreatectomy in all patients who fail to respond completely to medical therapy (Daneman and Ehrlich, 1993; Thornton et al., 1993). These groups view diabetes as a lesser evil that can be dealt with when the time comes. Other groups, like our own, have been impressed with the clinical improvement over time seen in medically treated patients, and therefore make a great effort to avoid surgery if this is possible without undue risk of hypoglycemia (Glaser et al., 1993). We ®nd that very intensive treatment in the ®rst few weeks to months is rewarded by clinical improvement, easier long-term management and avoidance or delay of diabetes. We feel that this approach provides better long-term results, but only if the family is able to successfully comply with the necessary treatment regimen. Clearly, some patients fail to respond even to the most aggressive medical treatment, and some families are unable to cope with the requirements of intensive medical management. In these cases, extensive, partial pancreatectomy is mandatory. Genetics Given the clinical complexity of the disease, it is not surprising that little progress was made in understanding its pathophysiology using a standard biochemical approach. It was not possible to study each step of the insulin secretory cascade in HI patients and it is equally dif®cult to systematically search for mutations in all of the genes involved in the control of insulin secretion. No animal model for this disease was available. Alternative methods were needed to understand the etiology of this disease, and in 1993 a pure genetic approach was used to try and answer these questions. Incidence and mode of transmission The incidence of HI appears to vary greatly between populations. An estimated incidence of 1:50 000 and 1:40 000 has been proposed for unselected Northern European populations (Bruining, 1990; Otonkoski et al., 1999). However, an incidence of 1:2500 has been reported in Saudi Arabia (Mathew et al., 1988). More recently, a similar incidence (1:3200) has been reported for a small region of Finland (Otonkoski et al., 1999). Most familial cases of HI appear to be inherited in a recessive manner, with both parents being clinically asymptomatic and about 25% of the siblings affected (Glaser et al., 1990; Thornton et al., 1991). Also suggesting recessive inheritance, the

ATP-sensitive K 1 channels in familial hyperinsulinism

305

Table 1 Long-term implications of therapeutic options for HI

1. Recurrent hypoglycemia can cause irreversible brain damage and therefore must be avoided at all costs 2. Even in patients with severe disease, intensive medical treatment, often with the aid of a percutaneous gastrostomy, can usually effectively prevent hypoglycemia. However this requires a very intensive effort on the part of the patients, physicians and family 3. Medical control of hypoglycemia becomes easier with time, and most medically treated patients enter clinical remission months to years after diagnosis. However, glucose intolerance and perhaps diabetes may be part of the long-term natural history of the disease 4. In diffuse disease, partial pancreatectomy usually results in persistent hypoglycemia the severity of which is roughly related to the degree of resection a. Less than 80% resection usually does not signi®cantly ameliorate the severity of hypoglycemia, so re-operation is frequently needed b. Greater than 95% resection is associated with marked improvement of hypoglycemia but a very high incidence of early-onset insulin-requiring diabetes c. 80±85% resection results in marked improvement, but not cure of hypoglycemia, and later-onset diabetes, usually becoming clinically evident at or soon after puberty 5. In focal disease, if the lesion can be identi®ed at surgery, limited resection results in complete cure. However, patients with focal disease may also enter complete clinical remission after 1±2 years of medical management

disease frequency is higher in small, isolated populations, particularly where consanguineous marriages are common (Mathew et al., 1988). However, dominant forms of the disease have also been reported (Kukuvitis et al., 1997; Thornton et al., 1998). Except for isolated, consanguineous populations, most cases are sporadic, meaning that a single case appears in a family with no previous background. These cases could be due to rare recessive mutations, new dominant mutations or some other form of genetic transmission. At the time of preparation of this manuscript, one recessive form, due to mutations in the KATP channel (HI-SUR1 and HI-KIR6.2), and two dominant forms, due to mutations in the glucokinase (HI-GK) or the glutamate dehydrogenase genes (HIGDH), have been reported. The relative prevalence of severe and mild cases depends entirely on the population studied. In central Finland, among Ashkenazi Jews and in the Arab patient populations of Saudi Arabia and Israel very severe clinical disease is the rule, since the prevalent KATP channel mutations in these populations cause severe channel dysfunction. In contrast, the non-Arab, non-Ashkenazi HI population in Israel

306

B. Glaser, L. Aguilar-Bryan

appears to have much milder disease, although mutations have not yet been identi®ed in all patients. Chromosomal linkage Initial studies attempted to identify chromosomal regions containing genes responsible for the recessive form of the disease. Linkage studies were accomplished using a highly selected population from Israel and from Philadelphia, chosen to include only familial cases with severe, neonatal hyperinsulinism, and multiple affected siblings with clinically unaffected parents. Interestingly, this initial population included 13 families of Ashkenazi Jewish origin, one large Arab family and one non-Jewish Caucasian family. Linkage to a 30 cM region of chromosome 11p 1415.1 was found (Glaser et al., 1994), and con®rmed in another population composed primarily of Saudi Arabian patients (Thomas et al., 1995a). By genotyping additional markers in a larger number of families, the critical region was reduced to about 0.8 cM (Glaser et al., 1995). These ®ndings suggested that the disease in these particular populations was caused by a gene at this locus, but the data gave no indication as to what speci®c gene(s) may be involved, since at the time, no bcell-speci®c genes had been identi®ed at this locus. Molecular etiology As mentioned before, mutations in the glucokinase and glutamate dehydrogenase genes are responsible for two different dominant forms of HI. These two enzymes regulate the rate of nutrient metabolism in the pancreatic b-cell. Although these genes do not code for channel proteins, they will be discussed brie¯y to emphasize the genetic heterogeneity of the disease. Glucokinase mutations A dominant mutation in the GK gene that increases the enzyme's af®nity for glucose was recently described (Glaser et al., 1998). This functional alteration results in increased glucose metabolism in the presence of low glucose levels, increased intracellular ATP/ADP ratio, and insulin secretion by the mechanism described above. As predicted from in vitro kinetic studies, patients with this mutation have qualitatively normal control of glucose-mediated insulin secretion. However, the set-point for insulin secretion is much lower than normal and is below the threshold of neuroglycopenia. Therefore the patients with this mutation suffer from mild to moderate fasting hypoglycemia. With no abnormal KATP channel function, the hyperinsulinism can be suppressed by treatment with diazoxide. To date, only one family with HI-GK has been reported. Glutamate dehydrogenase mutations Recently, several families with hyperinsulinemic hypoglycemia and hyperammonemia have been identi®ed (Zammarchi et al., 1996; Weinzimer et al., 1997). Dominant, activating mutations in the glutamate

ATP-sensitive K 1 channels in familial hyperinsulinism

307

dehydrogenase (GDH) gene (GLUD1) have been associated with this disorder (Stanley et al., 1998). Increased GDH activity results in increased a-ketoglutarate production, subsequent increase in ¯ux through the Krebs cycle and excessive ATP production; all unrelated to the ambient glucose concentrations. The hyperammonemia in these patients appears to be moderate and asymptomatic. As expected, in most cases the hyperinsulinism responds to treatment with the KATP channel opener diazoxide. Eight patients with this disorder, called HI-GDH, were reported. In two families autosomal dominant inheritance was documented, whereas in six, a de novo, dominant-acting mutation was identi®ed in the proband. We have screened 114 HI cases and found GLUD-1 mutations in only four cases, including one Ashkenazi Jewish patient with a new mutation, implying HI-GDH is a rare cause of neonatal hyperinsulinism. KATP channel mutations Soon after the publication of the initial linkage data, the sulfonylurea receptor gene (SUR1) was cloned (Aguilar-Bryan et al., 1995). It was mapped to 11p15.1, within the same chromosomal region that the gene responsible for HI had been localized by linkage studies, and two mutations were identi®ed in HI patients (Thomas et al., 1995b). The KIR6.2 gene was subsequently cloned and found to be physically adjacent to SUR1 (Fig. 2) (Aguilar-Bryan and Bryan, 1996; Inagaki et al., 1995; Permutt et al., 1996). Co-expression of KIR6.2 and SUR1 reconstituted KATP channel activity (Inagaki et al., 1995). These ®ndings provided the ®rst evidence for a possible link between this particular insulin secretory abnormality and KATP channels, and the hypothesis that HI may be a new channel disease was proposed (Aguilar-Bryan et al., 1996). The ®nal proof that mutations within SUR1 were responsible for HI came from Dunne et

Fig. 2. Schematic representation of the SUR1 and SUR2 genes, with their corresponding inward recti®ers, both localized at the 3 0 end of the SURs. As indicated by the number of boxes, SUR1 has 39 exons and SUR2 38. They are both similar in size and the deletion of one exon (18) determines the difference in number. SUR2 A and B differ in the differential usage of the last two exons (2A and 2B). NBF, nucleotide binding folds.

308

B. Glaser, L. Aguilar-Bryan

al. (1997), who showed that b-cells from an HI patient with a particular SUR1 mutation in exon 35, lacked KATP channel activity. This lack of channel activity, was observed by electrophysiological analysis of the patient's b-cells and in parallel, of the engineered mutation expressed in COS cells. Since the initial observations, more than 50 mutations have been identi®ed in these two genes (Table 2). All of the mutations identi®ed to date are recessive, since obligate heterozygous parents are clinically unaffected. Focal-HI In about 40% of HI patients, focal lesions with b-cell hyperplasia can be identi®ed and, if selectively removed, total clinical cure can be achieved (Sempoux et al., 1998). The molecular etiology of this entity was recently determined (deLonlay et al., 1997). Loss of the maternal allele on the short arm of chromosome 11 was observed in these lesions. The degree of loss was variable, but in all cases the segment lost included the SUR1/KIR6.2 locus and loci telomeric to it. In addition to containing the genes for the b-cell KATP channel, this chromosomal region contains several imprinted genes, including maternally expressed tumor suppressor genes and a paternally expressed growth factor gene. Three recent publications (Ryan et al., 1998; Verkarre et al., 1998; Glaser et al., 1999c), have identi®ed seven patients in whom a paternally inherited SUR1 mutation is associated with focal-HI and loss of the maternal short arm of chromosome 11 within the focal lesion. Although current data is sparse, it seems likely that as many as 40% of HI patients may have disease caused by this unique genetic constellation. b-cells will proliferate in the focal lesion, because expressed tumor suppressor genes are lost, and growth factor genes are retained. Having retained the (paternal) mutant SUR1 allele and lost the normal (maternal) allele, these cells will have no functioning KATP channels. This series of events fully explains the clinical ®ndings in these patients: (1) severe HI is not responsive to diazoxide treatment, and (2) complete clinical cure if the focal lesion can be selectively removed by surgery. The latter is expected since remaining b-cells are heterozygous for a recessive SUR1 mutation and hence have normally functioning KATP channels. Summary of HI genetics We have reviewed the results of mutational analysis in a large, heterogeneous cohort of HI patients (Glaser, Aguilar-Bryan, unpublished data). Glucokinase mutations appear to be extremely rare, since only a single family has been identi®ed. Glutamate dehydrogenase mutations account for ,5% of the disease. Mutations on the KATP channel appear to account for at least 50% of the cases, and in about 40% of these, focal-HI may be present. For the remaining ,45% of HI patients, the genetic etiology is still unknown.

ATP-sensitive K 1 channels in familial hyperinsulinism Table 2 HI-associated KATP channel mutations

Mutation Exon 2 R 74 Q Exon 3 A 116 P H 125 Q Exon 4 V187D

Sequence

Restriction site

References

CTG CGG TGG CTG CAG

1Pst1

Nestorowicz et al. (1998)

ATG GCT GCT ATG CCT TAT CAC AAC TAT CAA

2Bbv1

Aguilar-Bryan and Bryan, (1999) Nestorowicz et al. (1998)

GAG GTC AAT GAG GAC

Exon 5 679 ins 18 bp

Ins 6 aa

731 del A

fs/ter after 15aa

R 248 X

TTG CGA GCC TTG TGA

Exon 6 949 del C N 406 D C 418 R 1260 ins 31 bases Exon 10 L 508 P Intron 10 163011 g ! t Intron 11 1672220 a ! g Exon 12 F 591 L

GGG CCA CTG GGG -CA TCC AAC CTG TCC GAC ATC TGT AAT ATC CGT ins 31 bp ter after 11aa

No site 1TTH111I

Otonkoski et al. (1999) Aguilar-Bryan and Bryan, (1999) Aguilar-Bryan and Bryan, (1999) Aguilar-Bryan and Bryan (1999)

Bsp1286I XcmI 2Bgl II

Nestorowicz et al. (1998) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999)

AAG CTG CTG AAG CCG

1MspA1 2NspBII

Aguilar-Bryan and Bryan (1999)

CCA GTG AGT CCA TTG

Bsr1

Nestorowicz et al. (1998)

CCT CAC TTG CCT CGC

No site

Thomas et al. (1996b)

CTG TTC CTG CTG TTA

1Mae III

Aguilar-Bryan and Bryan (1999)

309

310

B. Glaser, L. Aguilar-Bryan

Table 2 (continued) Mutation

Sequence

Restriction site

References

F 591 L

CTG TTC CTG CTG TTG

BsoF1

Nestorowicz et al. (1998)

ATC CGT GAG CTG TGT CCC CAT GAG CTG -AT ACA CCT CAG CTG CC-

2XhoII

Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Nestorowicz et al. (1998)

Exon 13 R 620 C H 627 X 1893 del T Intron 15 211721 g ! a Exon 16 G 716 V Intron 18 229521 g ! a Exon 21 R 837 X R 842 G Exon 22 K 890 T Exon 24 Q 954 X S 957 F Exon 28 T1139M Exon 29 R 1215 Q Intron 29 365121 g ! t

1BstN1

CTG CAG GCC CTG CAA

2Pst1

Aguilar-Bryan and Bryan (1999)

GTG GGC TGC CTG GTC

1Bbv1

Thomas et al. (1996b)

TCC AGG AAG CTG AAG

2BstN1

Thomas et al. (1996b)

CAG CGA ATC CAG TGA GCC CGA GCC CAG CGC

Bsr1

Aguilar-Bryan and Bryan (1999) Juniene et al., unpublished data

CAC AAG CTA CTG ACG

Aguilar-Bryan and Bryan (1999)

CCC CAG GGC CTG TAG CTA TCT CGT CTA TTT

2BstN1

Nestorowicz et al. (1998) Aguilar-Bryan and Bryan (1999)

TCC ACG CTG CTG ATG

1NlaIII

Nestorowicz et al. (1998)

ATC CGG GCC ATC CAG

1Stu1

Nestorowicz et al. (1998)

TCT CAG GTA CTG CAT

Aguilar-Bryan and Bryan (1999)

ATP-sensitive K 1 channels in familial hyperinsulinism

Table 2 (continued) Mutation Intron 32 399229 g ! a 399223 c ! g Exon 33 K 1337 N W 1339 X R 1353 P V 1361 M Exon 34 4144 CG ! GT del CA C ! G G 1379 R G 1382 S S 1387 F Del F 1388 R 1394 H Exon 35 4310 g ! a R 1419 C R 1421 C R 1437 Q Exon 37 4415213 g ! a

Sequence

Restriction site

References

CCC CTG CCC CTG

1NCI1

Thomas et al. (1996b) Nestorowicz et al. (1998)

GGC CCC AGC CAG CAC GAG

CCA AAG AAC CCA AAC AAC TGG CCA AAC TAG GTG CGC TAC CTG CCC CCG GTG CTG CTG ATG CGC ACC GGC GT- -GC GGC TGC GGC CGC CTG CGC ACC GGC GCC CTG AGC TCC TCC TTC CTG TTC TCC TTC TCT TCC - - - TCT TTC CGC ATG CTG CAC CAT CTG CTG CTG TCA CTG ATC CTG

CCG GTG CCA CGC TCA TGC CGC CTC TGC CGA GTG CAA

CCC CGG GCT CTG CAG

AVAI

2Bstx1 2Hha1 2BsrFI

2EagI 2Bgl1 2Mnl1 1BSeR1 1P¯M1 2Msp1 2Hha1 2NlaIII 1Msp1 2Nci1

Juniene et al., unpublished data Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Nestorowicz et al. (1998) Nestorowicz et al. (1998) Aguilar-Bryan and Bryan (1999) Nestorowicz et al. (1996) Nestorowicz et al. (1998) Thomas et al. (1995b) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999) Aguilar-Bryan and Bryan (1999)

311

312

B. Glaser, L. Aguilar-Bryan

Table 2 (continued) Mutation

Sequence

Restriction site

References

G 1479 R

GGC GGG GAG CTG GGA CTG GCC CGG CTG GAC GCC CGG GGC CTG TGG GCC CGG GGC CTG CAG CAG GAG GCC CTG AAG ACG GCT T- - - - - CCACG GCT TCG GCT TCC

2AciI

Aguilar-Bryan and Bryan (1999) Kutchinski, unpublished data Aguilar-Bryan and Bryan (1999) Kutchinski, unpublished data Aguilar-Bryan and Bryan (1999) Nestorowicz et al. (1998)

A 1493 T R 1494 W R 1494 Q E 1507 K 4525 ins cggctt Exon 39 L 1544 P

ATC CTG AGT ATC CCG Mutations in KIR 6.2 T 12 X GAA TAC GTG GAA TAG W 91 R TGG TGG CTC CTG CGG L 147 P ATC CTG AGC CTG CCG

2Sma1

2Mnl1 PVUII

Aguilar-Bryan and Bryan (1999) 1BsaA1 1AcII 1Ava1

Nestorowicz et al. (1997) Aguilar-Bryan and Bryan (1999) Thomas et al. (1996a)

What are KATP channels? ATP-sensitive potassium channels, or what have come to be termed KATP channels, were identi®ed simultaneously by Cook and Hales (1984) in the US, by Noma (1983) in Japan, and by Trube and Hescheler (1984) in Germany approximately 15 years ago. These channels have been identi®ed in several tissues, including heart and skeletal muscle, vascular and non-vascular smooth muscle, brain and pancreatic islets. KATP channels have been de®ned by their kinetic properties, sensitivity to ATP, and stimulation by ADP in the presence of Mg 21, as well as a de®ned pharmacology that includes channel blockers (sulfonylureas) and openers (diazoxide) (Ashcroft, 1988; Aguilar-Bryan et al., 1998). The role that KATP channels play in the regulation of insulin secretion, is determined by the link they establish between glucose metabolism and the electrical activity of the cell. The KATP channel is made up of two sub-units. The high af®nity sulfonylurea receptor is responsible for the regulation of channel activity, whereas the inward

ATP-sensitive K 1 channels in familial hyperinsulinism

313

Fig. 3. The top part of the ®gure, shows the schematic representation of the predicted topology for the two sub-units that form a KATP channel, the sulfonylurea receptor and the inward recti®er. We are using the model suggested by Tusnady et al. (1997). For SUR1, the characteristic two sets of transmembrane spanning domains with an extended N-terminal are shown. The two nucleotide binding folds (NBF) with the Walker motifs and the glycosylation sites are also illustrated. For KIR6.x, the two transmembrane domains separated by the H5 loop are shown. The bottom ®gure shows the proposed hetero-octameric architecture of the channel.

recti®er forms the actual ion pore (Fig. 3). It has become very clear that the expression of both sub-units is necessary to reproduce the characteristics of the native KATP channel (Inagaki et al., 1995).

314

B. Glaser, L. Aguilar-Bryan

The high af®nity sulfonylurea receptor (SUR1) is a member of the ATP binding cassette (ABC) superfamily of proteins. These proteins have the characteristic of having an ATP-binding domain containing Walker A and B motifs. Some of the proteins that form membrane-bound channels have two sets of transmembrane spanning domains and two nucleotide binding folds (Fig. 3). SUR1 has binding sites for sulfonylureas and diazoxide, as well as the site for stimulatory nucleosides (MgATP and MgADP). Two sulfonylurea receptor genes have been identi®ed (Inagaki et al., 1996). The high af®nity receptor, SUR1, and the highly homologous low af®nity receptor, SUR2. We have now cloned, mapped and characterized both human SUR genes (SUR1 and SUR2). The characterization of the SUR1 gene has been completed, the gene spans ,90 kb of DNA and has 39 exons. The SUR1 gene is localized on the short arm of chromosome 11 (11p15.1), and the SUR2 gene on the short arm of chromosome 12 (12p11.12). There are two variants of the SUR2 gene, SUR2A and SUR2B, which result from different splicing of the last exon (AguilarBryan and Bryan, 1999, Fig. 2). The second sub-unit of the KATP channel is a member of the inwardly rectifying family of K 1 channels (KIR6.2), with two transmembrane spanning domains and the P or H5 loop (Fig. 3). This protein forms a ~75p S selective K 1 pore, is the site where polyamines bind and presents a low af®nity site for binding ATP. Two different, highly homologous KIR genes have been identi®ed, KIR6.1 and KIR6.2 (Inagaki et al., 1996). Interestingly, the SURs and the KIRs are paired on the same chromosome, with KIR6.2 localizing 4.5 Kb 3 0 of SUR1 and KIR3.1 located 3 0 of SUR2 (Fig. 2). This observation has raised some intriguing questions about the evolutionary relationship of these two proteins and suggests interesting possibilities for coordinated regulation of these genes through transcription of a polycistronic mRNA. We have previously shown that the two sub-units that form the KATP channel are physically associated in a 1:1 stoichiometry (Clement et al., 1997), that form the hetero-octameric architecture (Fig. 3). Using a molecular approach we constructed two different kinds of fusion proteins, one that had one SUR1 and one KIR6.2, and a second one that had one SUR and two KIR6.2. Expression of the 1:1 stoichiometry was suf®cient to obtain functional KATP channels in COS cells. To test if this was necessary, the second fusion was used. KATP channel activity was not present, but reestablishing the 1:1 stoichiometry by adding SUR1 monomers resulted in active channels. The current model for KATP topology is shown in Fig. 3. HI-associated KATP channel mutations Nestorowicz et al. (1996, 1997, 1998) scanned the entire coding sequence of both the KIR6.2 and the SUR1 genes in 45 probands including representatives of a wide spectrum of ethnic groups . Sharma and Aguilar-Bryan also looked for mutations in

ATP-sensitive K 1 channels in familial hyperinsulinism

315

Fig. 4. (A) Channel density measured by patch-clamp recordings. Channel activities of wild-type and mutant channels were measured by the inside-out patch-clamp method in the absence of ATP. Channel densities are expressed as average currents (with the sign reversed) at 2 50 mV. (B) Alterations of sensitivity to MgADP stimulation by SUR1 mutations. Currents in 0.1 mmol/l ATP and 1.1 mmol/l Mg 21 relative to currents in the absence of ATP with (®lled columns) and without (empty columns) 0.5 mmol/l ADP. *P , 0:005. This ®gure is a modi®cation of Figs. 3 and 4 from Shyng et al. (1998).

the entire coding sequence of both proteins in a group of 50 patients from different ethnic backgrounds (unpublished data, described in Table 2). Interestingly, in both sets of patients, mutations were identi®ed in less than 50% of the probands. However, in these two studies, only the coding regions were screened using single

316

B. Glaser, L. Aguilar-Bryan

strand conformational polymorphism (SSCP) analysis. Therefore, any mutations in the promoter or the non-coding region of the gene, would not have been identi®ed. Furthermore, SSCP may miss as many as 10±20% of point mutations even within the regions screened.

Fig. 5. Ion channel recordings from control and HI b-cells. (A) Typical intact cell data from control human b-cells and b-cells isolated from the HI patients after pancreatectomy. In control cells, spontaneously active KATP channels are further activated by the addition of diazoxide to the intact cell (0.5 mmol/l) and inhibited by the presence of the sulfonylurea, tolbutamide (0.2 mmol/l) in the presence of the KATP channel agonist. In contrast, there are no actions of diazoxide or somatostatin (100 nmol/l) on K 1 channels in HI b-cells. (B) The effect of removing patches of membrane from either HI b-cells (open column) or control tissue (black column). In control cells, the procedure causes the appearance of a marked and sustained KATP channel current due to wash out of cytosolic ATP. In the representative current trace, the arrow indicates the point of formation of the inside-out patch. All data were obtained at 0 mV voltage-clamp with a 140 mmol/l NaCl-rich solution in the bath. In direct comparison with control b-cells, under exactly the same experimental conditions, the data reveal a marked loss of functional KATP channels in HI b-cells (re-printed from Otonkoski et al. (1999)).

ATP-sensitive K 1 channels in familial hyperinsulinism

317

More than 50 KATP channel mutations have now been identi®ed in patients with this insulin secretory abnormality (Table 2). Three were found in the small intronless gene that codes for the inward recti®er, while the rest were found in the gene that codes for the high af®nity sulfonylurea receptor. Many of these mutations have been engineered and tested for channel activity by electrophysiological means as well as 86 Rb ef¯ux. Most of the mild cases whose mutations have been analyzed for KATP channel activity by electrophysiology and 86Rb ef¯ux, con®rm the defect initially observed with the G1479R mutation (Nichols et al., 1996). Brie¯y, the lack of normal KATP channel activity is characterized by poor channel opening in response to MgADP in the presence of ATP. The 86Rb experiments con®rm that the number of channels that open after metabolic inhibition is very small, indicating that very low levels of ATP, are not suf®cient to activate the channel. The response to MgADP appears to be critical in allowing enough channels to open and therefore maintain membrane hyperpolarization. Fig. 4A shows signi®cant decrease in channel density, analyzed by the patch-clamp technique in the absence of ATP, for the H125Q and the R1215Q mutations but not for the other missense mutations (Shyng et al., 1998). Fig. 4B illustrates the defective response to MgADP in the several mutations (F591L, R1215Q, G1382S and G1479R) all of which had a normal sensitivity to ATP. Dunne et al. (1997) demonstrated that b-cells from a severe case of HI showed no KATP channel activity. Recently, Otonkoski et al. (1999) con®rmed this observation, describing a novel mutation, V187D, associated with a severe phenotype and lack of clinical response to treatment with diazoxide. No KATP channel activity was present in the b-cells obtained at surgery from one homozygous patient (Fig. 5) or in the engineered mutation expressed with wild-type KIR6.2, consistent with the clinical picture. One splice mutation (3992-9 g ! a), that is particularly common in the Ashkenazi population and is also found in patients from other ethnic groups, seems to have a different behavior. Most patients homozygous for this mutation have severe clinical disease. However, we have identi®ed two families with multiple siblings in which the proband was homozygous for this mutation and had severe HI, whereas haploidentical siblings had very mild disease or were clinically unaffected. These are both Ashkenazi families in which this is a founder mutation so that severe and mild cases were all genetically identical at this locus (Glaser et al., 1999b). This ®nding suggests that some splice mutations may be variably expressed and that some normal protein may be produced in some individuals. With different genetic backgrounds, a different percentage of normal protein may be produced. A similar phenomenon has been found in patients with a similar splice-site mutation in the 17 hydroxylase enzyme (Witchel et al., 1996). Most patients homozygous for this mutation have severe congenital adrenal hyperplasia, however asymptomatic homozygous patients with apparent near-normal enzyme function have been identi®ed.

318

B. Glaser, L. Aguilar-Bryan

Only three mutations have been identi®ed in KIR6.2. Nestorowicz et al. (1997) found an early truncation of KIR6.2, where after only 12 amino acids (Y12X), a stop codon was introduced. As with the other mutations, the patient was homozygous and had severe clinical manifestations. When the mutation was engineered and expressed, no functional channels were observed, as was expected, since the protein was missing all the elements needed to form the pore. Thomas et al. (1996a) reported a KIR6.2 mutation, L147P, in the second transmembrane domain. The patient was the product of a consanguineous marriage of Iranian origin. The biochemical diagnosis was based on the presence of low glucose levels (,30 mg/dl), in the presence of inappropriate levels on insulin (.30 mU/ml) and high glucose requirements to maintain euglycemia (.15 mg glucose/kg min). We engineered and expressed the mutation and found no channel activity by electrophysiology or 86Rb ef¯ux (Aguilar-Bryan et al, unpublished data). Sharma and Aguilar-Bryan have identi®ed a third mutation, W91R, present on the opposite side of L147P, just at the beginning of the H5 loop (unpublished data). The patient was the product of a consanguineous marriage of Palestinian descent, homozygous for the mutation and severely affected. She had a pancreatectomy 2 weeks after birth and a second resection 4 weeks latter. The mutation was expressed in COS cells and KATP channel activity was absent. Comparison to animal models No naturally occurring animal models for HI have been described, so attempts have been made to bioengineer an animal model. One of the more puzzling new ®ndings is the discovery that KATP channel knockout mice, missing either KIR6.2 (Miki et al., 1998) or SUR1 (Seghers et al., 2000), and clearly lacking KATP channels in their b-cells, have nearly normal glucose levels, a result remarkably different from human HI neonates. Uncovering the compensatory mechanism(s) should prove interesting and provide new insight into glucose homeostasis. Understanding how mice are able to regulate their glucose levels without KATP channels may prompt new therapies for HI. Perspectives Although limited, the recent discoveries described above have already had some impact on the diagnosis and treatment of HI. Because of the many mutations that have been identi®ed, the role of genetic testing in the clinical diagnosis is still limited. However, in speci®c population isolates, where the majority of disease is caused by a few, easily identi®ed mutations, rapid mutational analysis may provide clinically useful information. We now offer routine testing for the two common mutations present in Ashkenazi Jews, and have found this test to be clinically useful for the rapid diagnosis of this disease and for genetic counseling. The diagnosis is

ATP-sensitive K 1 channels in familial hyperinsulinism

319

de®nitive if two mutant alleles are identi®ed. If a single mutation is identi®ed in a hypoglycemic infant, the probability of HI is very high and if this mutation is on the paternal allele, the probability of focal disease is high, albeit not proven. In contrast, if neither mutation is present in an Ashkenazi Jewish patient with hyperinsulinemic hypoglycemia, then another genetic etiology is likely and further clinical evaluation is needed. HI is a rare disease, and as such it may be questioned whether such an uncommon disease warrants the attention that has been afforded. From the discussion above, it is clear that the study of this disease has provided new data whose importance far exceeds its contribution towards the understanding of this one disease. The ®nding of SUR1/KIR6.2 mutations in patients with HI, and the subsequent observation that these patients have abnormal or absent KATP channel activity in their b-cells, proved conclusively that these two proteins form a channel that is very important for the regulation of insulin secretion in humans. Identi®cation of novel mutations and their in vitro study has enhanced our knowledge of how this channel functions. Although most of these studies could have been and in fact were done in animal models, the availability of human patients with the mutations provides important correlation regarding the relevance of animal ®ndings for humans. Non-insulin dependent diabetes is a common disease associated with abnormal insulin secretion and commonly treated with sulfonylureas. Cloning and structure/ function studies of the receptor for this class of drug will be extremely helpful in the development of a new generation of drugs for the treatment of diabetes mellitus. Therefore, although HI is a very rare disease, lessons learned from the study of the molecular biology of this disease may have very important rami®cations for our understanding and treatment of common human insulin secretory abnormalities. Future research The studies reviewed here open opportunities for additional research in several different directions. 1. Mutations for about 45% of patients with this disease have not been identi®ed. Therefore the identi®cation of new genes that will be responsible for producing HI will be very important. So far, mutations in four different genes have been shown to cause this syndrome. In each case, the identi®cation of the mutation and the physiological repercussions have improved our understanding of b-cell physiology and pathophysiology. It is likely that ®nding additional genes that can cause this disease will be similarly rewarding. 2. As stated above, the SUR1 sub-unit of the KATP channel is an important target for drug development. Better understanding of its structure and function will clearly aid in this endeavor. More detailed information is needed for many aspects of channel function.

320

B. Glaser, L. Aguilar-Bryan

(a) De®ning the precise topology of SUR1 and the exact binding sites for endogenous and exogenous ligands could prove useful. (b) The mechanism by which the sub-units interact will be important for the understanding of channel activity in general. (c) Understanding the mechanisms that regulate traf®cking of the KATP channel to the plasma membrane, will be important for the understanding of channel expression at this level. (d) Do other peptides that regulate insulin secretion such as GLP-I, galanin, leptin have any direct effect on channel activity? (e) Are there other isoforms of SUR and/or KIR6.X? If so, what is their physiological signi®cance? 3. It is intriguing that the SUR1 and KIR6.2 are physically associated. This may be important for coordinated gene transcription and may have more general implications in human molecular biology. 4. The ®nding of SUR1 and its association to HI has led to the discovery of SUR2. Could mutations in SUR2 cause a novel, as yet undiscovered clinical syndrome?

Note added in proof

1. Zerangue et al. (1999): This paper identi®ed an ER retention signal, -RKR-, present in both SUR1 and KIR6.2. These signals are masked in some unspeci®ed manner during assembly and serve as a quality control mechanism to insure that only complete octameric channels reach the plasma membrane. 2. Sharma et al. (1999): This paper identi®ed a second quality control signal within the C-terminal 25 amino acids of SUR1 which is required for the exit of assembled KATP channels from the ER. Lack of this signal in the receptors of patients with mutations that truncate this protein was proposed to account for loss of KATP channel activity. 3. Seghers et al. (2000): This paper, brie¯y cited in the text, describes the phenotype of Sur1 knockout mice. The results underscore the wide differences between the quite mild changes in glucose homeostasis observed when KATP channel activity is lost in mice versus the severe effects observed in humans. The importance of a KATP-independent pathway for regulation of glucose-stimulated insulin release is discussed. 4. Bryan and Aguilar-Bryan (1999): This review article provides an update on the composition and stoichiometry of KATP channels and their traf®cking, and the use of chimeric channels to identify functionally important domains including those

ATP-sensitive K 1 channels in familial hyperinsulinism

321

involved in the action of the hypoglycemic sulfonylureas and potassium channel openers. 5..±8. Aynsley-Green et al. (2000), Glaser et al. (2000), Rahier et al. (2000), Shepherd et al. (2000): These four articles provide a comprehensive and up-to-date review of all aspects of the clinical syndrome, hyperinsulinism of infancy. 9. Tanizawa et al. (2000): The disease HI has not been previously described in Oriental populations. This article describes the clinical syndrome and SUR1 mutations in Japanese HI patients. Acknowledgements We gratefully acknowledge the contribution of all of the patients, their families and their physicians who provided samples and clinical information for these studies. This work was supported by grant #93/00191/2 from the US±Israel Binational Science Foundation (B.G.), grant nos. 2677 and 4201 from the Israel Ministry of Health (B.G.), a grant from the Israel Science Foundation founded by the Academy of Sciences and Humanities (B.G.); a grant from the ADA (L.A.-B.), and the Texas Endowment (L.A.-B.). The authors would like to thank Dr Nidhi Sharma for helpful critical discussion and all the members from the Baylor group that worked on this project for the last several years. Dr Aguilar-Bryan would like to invite anybody who is interested, to visit http://www.SUR1.com, a web site about PHHI that two families with affected children with this disease and herself have developed. References Aguilar-Bryan, L., Bryan, J., 1996. ATP-sensitive potassium channels, sulfonylurea receptors, and persistent hyperinsulinemic hypoglycemia of infancy. Diabetes Rev. 4, 336±346. Aguilar-Bryan, L., Bryan J., 1999. The molecular biology of ATP-sensitive K + channels. Endocrine Rev. 20, 101±135. Aguilar-Bryan, L., Nichols, C.G., Wechsler, S.W., Clement IV, J.P., Boyd, A.E., Gonzalez, G., Herrera Sosa, H., Nguy, K., Bryan, J., Nelson, D.A., 1995. Cloning of the beta cell high-af®nity sulfonylurea receptor: a regulator of insulin secretion. Science 268, 423±426. Aguilar-Bryan, L., Clement IV, J.P., Gonzalez, G., Kunjilwar, K., Babenko, A., Bryan, J., 1998. Toward understanding the assembly and structure of KATP channels. Physiol. Rev. 78, 227±245. Aizawa, T., Komatsu, M., Asanuma, N., Sato, Y., Sharp, G.W., 1998. Glucose action `beyond ionic events' in the pancreatic beta cell. TIPS 19, 496±499. Ariel, I., Kerem, E., Schwartz-Arad, D., Bartfeld, E., Ron, N., Pizov, G., Zajicek, G., 1988. Nesidioblastosis ± a histologic entity? Hum. Pathol. 19, 1215±12128. Ashcroft, F.M., 1988. Adenosine 5 0 -triphosphate-sensitive potassium channels. Ann. Rev. Neurosc. 11, 97±118. Aynsley-Green, A., Hussain, K., Hall, J., Saudubray, J.M., Nihoul-Fekete, C., De Lonlay-Debeney, P., Brunelle, F., Otonkoski, T., Thornton, P., Lindley, K.J., 2000. Practical management of hyperinsulinism in infancy. Arch. Dis. Child Fetal Neonatal Ed. 82, F98±F107.

322

B. Glaser, L. Aguilar-Bryan

Bas, F., Darendeliler, F., Demirkol, D., Bundak, R., Saka, N., Gunoz H., 1999. Successful therapy with calcium channel blocker (nifedipine) in persistent neonatal hyperinsulinemic hypoglycemia of infancy. J. Pediatr. Endocrinol. Metab. 12, 873±878. Bruining, G.J., 1990. Recent advances in hyperinsulinism and the pathogenesis of diabetes mellitus. Curr. Opin. Pediatr. 2, 758±765. Bryan, J., Aguilar-Bryan, L., 1999. Sulfonylurea receptors: ABC transporters that regulate ATP-sensitive K(+) channels. Biochim. Biophys. Acta 1461, 285±303. Cade, A., Walters, M., Puntis, J.W., Arthur, R.J., Stringer, M.D., 1998. Pancreatic exocrine and endocrine function after pancreatectomy for persistent hyperinsulinaemic hypoglycaemia of infancy. Arch. Dis. Child 79, 435±439. Clement IV, J.P., Kunjilwar, K., Gonzalez, G., Schwanstecher, M., Panten, U., Aguilar-Bryan, L., Bryan, J., 1997. Association and stoichiometry of KATP channel subunits. Neuron 18, 827±838. Cook, D.L., Hales, C.N., 1984. Intracellular ATP directly blocks K 1 channels in pancreatic B-cells. Nature 311, 271±273. Dacou-Voutetakis, C., Psychou, F., Maniati-Christidis, M., 1998. Persistent hyperinsulinemic hypoglycemia of infancy: long-term results. J. Pediatr. Endocrinol. Metab, 11 (Suppl. 1), 131±141. Daneman, D., Ehrlich, R.M., 1993. The enigma of persistent hyperinsulinemic hypoglycemia of infancy. J. Pediatr. 123, 573±575. deLonlay, P., Fournet, J.C., Rahier, J., Gross-Morand, M.S., Poggi-Travert, F., Foussier, V., Bonnefont, J.P., Brusset, M.C., Brunelle, F., Robert, J.J., Nihoul-Fekete, C., Saudubray, J.M., Junien, C., 1997. Somatic deletion of the imprinted 11p15 region in sporadic persistent hyperinsulinemic hypoglycemia of infancy is speci®c of focal adenomatous hyperplasia and endorses partial pancreatectomy. J. Clin. Invest. 100, 802±807. Dunne, M.J., Kane, C., Shepherd, R.M., Sanchez, J.A., James, R.F., Johnson, P.R., Aynsley-Green, A., Lu, S., Clement IV, J.P., Lindley, K.J., Seino, S., Aguilar-Bryan, L., 1997. Familial persistent hyperinsulinemic hypoglycemia of infancy and mutations in the sulfonylurea receptor. N. Engl. J. Med. 336, 703±706. Glaser, B., Valtysson, G., Fajans, S.S., Vinik, A.I., Cho, K., Thompson, N., 1981. Gastrointestinal/ pancreatic hormone concentrations in the portal venous system of nine patients with organic hyperinsulinism. Metabolism 30, 1001±1010. Glaser, B., Phillip, M., Carmi, R., Lieberman, E., Landau, H., 1990. Persistent hyperinsulinemic hypoglycemia of infancy (`nesidioblastosis'): autosomal recessive inheritance in seven pedigrees. Am. J. Med. Genet. 37, 511±515. Glaser, B., Hirsch, H.J., Landau, H., 1993. Persistent hyperinsulinemic hypoglycemia of infancy: longterm octreotide treatment without pancreatectomy. J. Pediatr. 123, 644±650. Glaser, B., Chiu, K.C., Anker, R., Nestorowicz, A., Landau, H., Ben Bassat, H., Shlomai, Z., Kaiser, N., Thornton, P.S., Stanley, C.A., et al., 1994. Familial hyperinsulinism maps to chromosome 11p14-15.1, 30 cM centromeric to the insulin gene. Nat. Genet. 7, 185±188. Glaser, B., Chiu, K.C., Liu, L., Anker, R., Nestorowicz, A., Cox, N., Landau, H., Kaiser, N., Thornton, P.S., Stanley, C.A., Cerasi, E., Baker, L., Donis-Keller, H., Permutt, M.A., 1995. Recombinant mapping of the familial hyperinsulinism gene to an 0.8 cM region on chromosome 11p15.1 and demonstration of a founder effect in Ashkenazi Jews. Hum. Mol. Genet. 4, 879±886. Glaser, B., Kesavan, P., Heyman, M., Davis, E., Cuesta, A., Buchs, A., Stanley, C.A., Thornton, P.S., Permutt, M.A., Matschinsky, F.M., Herold, K.C., 1998. Familial hyperinsulinism caused by an activating glucokinase mutation. N. Engl. J. Med. 338, 226±230. Glaser, B., Landau, H., Permutt, M.A., 1999a. Neonatal hyperinsulinism. Trends Endocrinol. Metabol. 10, 55±61. Glaser, B., Furth, J., Stanley, C.A., Baker, L., Thornton, P.S., Landau, H., Permutt, M.A., 1999b. Intra-

ATP-sensitive K 1 channels in familial hyperinsulinism

323

genic single nucleotide polymorphism haplotype analysis of SUR1 mutations in familial hyperinsulinism. Hum. Mutat. 14, 23±29. Glaser, B., Ryan, F., Donath, M., Landau, H., Stanley, C.A., Baker, L., Barton, D.E., Thornton, P.S., 1999c. Hyperinsulinism caused by paternal-speci®c inheritance of a recessive mutation in the sulfonylurea-receptor gene. Diabetes 48, 1652±1657. Glaser, B., Thornton, P., Otonkoski, T., Junien, C., 2000. Genetics of neonatal hyperinsulinism. Arch. Dis. Child Fetal Neonatal Ed. 82, F79±F86. Goossens, A., Gepts, W., Saudubray, J.M., Bonnefont, J.P., Nihoul, F., Heitz, P.U., Kloppel, G., 1989. Diffuse and focal nesidioblastosis. A clinicopathological study of 24 patients with persistent neonatal hyperinsulinemic hypoglycemia. Am. J. Surg. Pathol. 13, 766±775. Inagaki, N., Gonoi, T., Clement IV, J.P., Namba, N., Inazawa, J., Gonzalez, G., Aguilar Bryan, L., Seino, S., Bryan, J., 1995. Reconstitution of IKATP: an inward recti®er subunit plus the sulfonylurea receptor. Science 270, 1166±1170. Inagaki, N., Gonoi, T., Clement IV, J.P., Wang, C.Z., Aguilar-Bryan, L., Bryan, J., Seino, S., 1996. A family of sulfonylurea receptors determines the pharmacological properties of ATP-sensitive K 1 channels. Neuron 16, 1011±1017. Kaiser, N., Corcos, A.P., Tur, S.A., Ariav, Y., Glaser, B., Landau, H., Cerasi, E., 1990. Regulation of insulin release in persistent hyperinsulinaemic hypoglycaemia of infancy studied in long-term culture of pancreatic tissue. Diabetologia 33, 482±488. Kukuvitis, A., Deal, C., Arbour, L., Polychronakos, C., 1997. An autosomal dominant form of familial persistent hyperinsulinemic hypoglycemia of infancy, not linked to the sulfonylurea receptor locus. J. Clin. Endocrinol. Metab. 82, 1192±1194. Lawson, K., 2000. Potassium channel openers as potential therapeutic weapons in ion channel disease [In Process Citation]. Kidney Int. 57, 838±845. Leibowitz, G., Glaser, B., Higazi, A.A., Salameh, M., Cerasi, E., Landau, H., 1995. Hyperinsulinemic hypoglycemia of infancy (nesidioblastosis) in clinical remission: Incidence of diabetes mellitus and persistent b-cell dysfunction at long-term follow-up. J. Clin. Endocrinol. Metabol. 80, 386±392. Lindley, K., Dunne, M., Kane, C., Shepherd, R., Squires, P., James, R., Johnson, P., Eckhardt, S., Wakeling, E., Dattani, M., Milla, P., Aynsley-Green, A., 1996. Ionic control of beta cell function in nesidioblastosis. A possible therapeutic role for calcium channel blockade. Arch. Dis. Child. 74, 373±378. Mathew, P.M., Young, J.M., Abu, O.Y., Mulhern, B.D., Hammoudi, S., Hamdan, J.A., Sa'di, A.R., 1988. Persistent neonatal hyperinsulinism. Clin. Pediatr. Phil. 27, 148±151. Miki, T., Nagashima, K., Tashiro, F., Kotake, K., Yoshitomi, H., Tamamoto, A., Gonoi, T., Iwanaga, T., Miyazaki, J., Seino, S., 1998. Defective insulin secretion and enhanced insulin action in KATP channelde®cient mice. Proc. Natl. Acad. Sci. USA 95, 10402±14046. Nestorowicz, A., Wilson, B.A., Schoor, K.P., Inoue, H., Glaser, B., Landau, H., Stanley, C.A., Thornton, P.S., Clement IV, J.P., Bryan, J., Aguilar-Bryan, L., Permutt, M.A., 1996. Mutations in the sulfonylurea receptor gene are associated with familial hyperinsulinism in Ashkenazi Jews. Hum. Mol. Genet. 5, 1813±1822. Nestorowicz, A., Inagaki, N., Gonol, T., Schoor, K.P., Wilson, B.A., Glaser, B., Landau, B., Stanley, H., Thornton, C.A., Seino, P.S., Permutt, M.A., 1997. A nonsense mutation in the inward recti®er potassium channel gene, KIR6.2, is associated with familial hyperinsulinism. Diabetes 46, 1743±1748. Nestorowicz, A., Glaser, B., Wilson, B.A., Shyng, S.-L., Nichols, C.G., Stanley, C.A., Thornton, P.S., Permutt, M.A., 1998. Genetic heterogeneity in familial hyperinsulinism. Hum. Mol. Genet. 7, 1119± 1128. Nichols, C.G., Shyng, S.-L., Nestorowicz, A., Glaser, B., Clement IV, J.P., Gonzalez, G., Aguilar-Bryan, L., Permutt, M.A., Bryan, J., 1996. Adenosine diphosphate as an intracellular regulator of insulin secretion. Science 272, 1785±1787.

324

B. Glaser, L. Aguilar-Bryan

Noma, A., 1983. ATP-regulated K 1 channels in cardiac muscle. Nature 305, 147±148. Otonkoski, T., Ammala, C., Huopio, H., Cote, G.J., Chapman, J., Cosgrove, K., Ash®eld, R., Huang, E., Komulainen, J., Ashcroft, F.M., Dunne, M.J., Kere, J., Thomas, P.M., 1999. A point mutation inactivating the sulfonylurea receptor causes the severe form of persistent hyperinsulinemic hypoglycemia of infancy in Finland. Diabetes 48, 408±415. Permutt, M.A., Nestorowicz, A., Glaser, B., 1996. Familial hyperinsulinsim: an inherited disorder of spontaneous hypoglycemia in neonates and infants. Diabetes Rev. 4, 347±355. Rahier, J., Falt, K., Muntefering, H., Becker, K., Gepts, W., Falkmer, S., 1984. The basic structural lesion of persistent neonatal hypoglycaemia with hyperinsulinism: de®ciency of pancreatic D cells or hyperactivity of B cells? Diabetologia 26, 282±289. Rahier, J., Guiot, Y., Sempoux, C., 2000. Persistent hyperinsulinaemic hypoglycaemia of infancy: a heterogeneous syndrome unrelated to nesidioblastosis. Arch. Dis. Child Fetal Neonatal Ed. 82, F108±F112. Ryan, F.D., Devaney, D., Joyce, C., Nestorowicz, A., Permutt, M.A., Glaser, B., Barton, D.E., Thornton, P.S., 1998. Hyperinsulinism: the molecular aetiology of focal disease. Arch. Dis. Child. 79, 445±447. Seghers, V., Nakazaki, M., DeMayo, F., Aguilar-Bryan, L., Bryan J., 2000. Sur1 knockout mice. A model for k(atp) channel-independent regulation of insulin secretion. J. Biol. Chem. 275, 9270±9277. Seghers, V., Nakazaki, M., DeMayo, F., Aguilar-Bryan, L., Bryan, J., 2000. Sur1 knockout mice. A model for k(atp) channel-independent regulation of insulin secretion. J. Biol. Chem. 275, 9270±9277. Sempoux, C., Guiot, Y., Lefevre, A., Nihoul-FeÂkeÂteÂ, C., Jaubert, F., Saudubray, J.-M., Rahier, J., 1998. Neonatal hyperinsulinemic hyperglycemia: heterogeneity of the syndrome and keys for differential diagnosis. J. Clin. Endocrinol. Metabol. 83, 1455±1461. Sharma, N., Crane, A., Clement, J.T., Gonzalez, G., Babenko, A., Bryan, J., Aguilar-Bryan, L., 1999. The C terminus of SUR1 is required for traf®cking of KATP channels. J. Biol. Chem. 274, 20628±20632. Shepherd, R.M., Cosgrove, K.E., O'Brien, R.E., Barnes, P.D., Ammala, C., Dunne, M.J., 2000. Hyperinsulinism of infancy: towards an understanding of unregulated insulin release. Arch. Dis. Child Fetal Neonatal Ed. 82, F87±F97. Shilyansky, J., Fisher, S., Cutz, E., Perlman, K., Filler, R.M., 1997. Is 95% pancreatectomy the procedure of choice for treatment of persistent hyperinsulinemic hypoglycemia of the neonate? J. Pediatr. Surg. 32, 342±346. Shyng, S.L., Ferrigni, T., Shepard, J.B., Nestorowicz, A., Glaser, B., Permutt, M.A., Nichols, C.G., 1998. Functional analyses of novel mutations in the sulfonylurea receptor 1 associated with persistent hyperinsulinemic hypoglycemia of infancy. Diabetes 47, 1145±1151. Stanley, C.A., Lieu, Y.K., Hsu, B.Y., Burlina, A.B., Greenberg, C.R., Hopwood, N.J., Perlman, K., Rich, B.H., Zammarchi, E., Poncz, M., 1998. Hyperinsulinism and hyperammonemia in infants with regulatory mutations of the glutamate dehydrogenase gene. N. Engl. J. Med. 338, 1352±1357. Straub, S.G., James, R.F., Dunne, M.J., Sharp, G.W., 1998. Glucose activates both KATP channel-dependent and KATP channel-independent signaling pathways in human islets. Diabetes 47, 758±763. Tanizawa, Y., Matsuda, K., Matsuo, M., Ohta, Y., Ochi, N., Adachi, M., Koga, M., Mizuno, S., Kajita, M., Tanaka, Y., Tachibana, K., Inoue, H., Furukawa, S., Amachi, T., Ueda, K., Oka, Y., 2000. Genetic analysis of Japanese patients with persistent hyperinsulinemic hypoglycemia of infancy: nucleotidebinding fold-2 mutation impairs cooperative binding of adenine nucleotides to sulfonylurea receptor 1. Diabetes 49, 114±120. Thomas, P.M., Cote, G.J., Hallman, D.M., Mathew, P.M., 1995a. Homozygosity mapping, to chromosome 11p, of the gene for familial persistent hyperinsulinemic hypoglycemia of infancy. Am. J. Hum. Genet. 56, 416±421. Thomas, P.M., Cote, G.J., Wohllk, N., Haddad, B., Mathew, P.M., Rabl, W., Aguilar-Bryan, L., Gagel, R.F., Byran, J., 1995b. Mutations in the sulfonylurea receptor gene in Familial Hyperinsulinemic Hypoglycemia of Infancy. Science 268, 426±429.

ATP-sensitive K 1 channels in familial hyperinsulinism

325

Thomas, P., Ye, Y., Lightner, E., 1996a. Mutation of the pancreatic islet inward recti®er KIR6.2 also leads to familial persistent hyperinsulinemic hypoglycemia of infancy. Hum. Mol. Genet. 5, 1809±1812. Thomas, P.M., Wohllk, N., Huang, E., Kuhnle, U., Rabl, W., Gagel, R.F., Cote, G.J., 1996b. Inactivation of the ®rst nucleotide-binding fold of the sulfonylurea receptor, and familial persistent hyperinsulinemic hypoglycemia on infancy. Am. J. Hum. Genet. 59, 510±518. Thornton, P.S., Sumner, A.E., Ruchelli, E.D., Spielman, R.S., Baker, L., Stanley, C.A., 1991. Familial and sporadic hyperinsulinism: histopathologic ®ndings and segregation analysis support a single autosomal recessive disorder. J. Pediatr. 119, 721±724. Thornton, P.S., Alter, C.A., Katz, L.E., Baker, L., Stanley, C.A., 1993. Short- and long-term use of octreotide in the treatment of congenital hyperinsulinism. J. Pediatr. 123, 637±643. Thornton, P.S., Satin-Smith, M.S., Herold, K., Glaser, B., Chiu, K.C., Nestorowicz, A., Permutt, M.A., Baker, L., Stanley, C.A., 1998. Familial hyperinsulinism with apparent autosomal dominant inheritance: Clinical and genetic differences from the autosomal recessive variant. J. Pediatr. 132, 9±14. Trube, G., Hescheler, J., 1984. Inward-rectifying channels in isolated patches of the heart cell membrane: ATP-dependence and comparison with cell-attached patches. P¯ug. Arch. Eur. J. Physiol. 401, 178± 184. Tusnady, G.E., Bakos, E., Varadi, A., Sarkadi, B., 1997. Membrane topology distingushes a subfamiy of the ATP-binding cassette (ABC) transporters. FEBS Lett. 401, 1±3. Verkarre, V., Fournet, J.C., de Lonlay, P., Gross-Morand, M.S., Devillers, M., Rahier, J., Brunelle, F., Robert, J.J., Nihoul-Fekete, C., Saudubray, J.M., Junien, C., 1998. Paternal mutation of the sulfonylurea receptor (SUR1) gene and maternal loss of 11p15 imprinted genes lead to persistent hyperinsulinism in focal adenomatous hyperplasia. J. Clin. Invest. 102, 1286±1291. Weinzimer, S.A., Stanley, C.A., Berry, G.T., Yudkoff, M., Tuchman, M., Thornton, P.S., 1997. A syndrome of congenital hyperinsulinism and hyperammonemia. J. Pediatr. 130, 661±664. Witchel, S.F., Bhamidipati, D.K., Hoffman, E.P., Cohen, J.B., 1996. Phenotypic heterogeneity associated with the splicing mutation in congenital adrenal hyperplasia due to 21-hydroxylase de®ciency. J. Clin. Endocrinol. Metab 81, 4081±4088. Zammarchi, E., Filippi, L., Novembre, E., Donati, M.A., 1996. Biochemical evaluation of a patient with a familial form of leucine-sensitive hypoglycemia and concomitant hyperammonemia. Metabolism 45, 957±960. Zerangue, N., Schwappach, B., Jan, Y.N., Jan, L.Y., 1999. A new ER traf®cking signal regulates the subunit stoichiometry of plasma membrane KATP channels. Neuron 22, 537±548.

Channelopathies ± Common Mechanisms in Aura, Arrhythmia and Alkalosis F. Lehmann-Horn and K. Jurkat-Rott, editors q 2000 Elsevier Science B.V. All rights reserved.

327

Chapter 13

Hereditary hypokalemic salt-losing tubulopathies N. Jeck, M. Konrad, H.W. Seyberth Department of Pediatrics, Deutschhausstrasse 12, 35037 Marburg, Germany

Abstract Autosomal recessively transmitted Bartter syndrome is characterized by normotensive hyperreninism and hyperaldosteronism with hypokalemic metabolic alkalosis. Three main phenotypical variants have been described: the classic Bartter syndrome, the hypomagnesemic hypocalciuric Gitelman syndrome, and the antenatal, hypercalciuric hyperprostaglandin E syndrome. The latter of which is the most severe variant with polyhydramnios due to fetal polyuria and life-threatening episodes of salt- and water-wasting postnatally. Four causative genes have been identi®ed each indicating different mechanisms for disease pathogenesis. In two of the genes, SLC12A1 and KCNJ1, encoding the Na±K±2Cl cotransporter type 2 and the inwardly rectifying potassium channel ROMK, loss of function mutations are found that exert urine concentration disturbances corresponding to the diuretic effects of furosemide. Mutations in the third gene, SLC12A3, encoding the NaCl cotransporter of the distal convoluted tubule, resemble thiazide effects, while the phenotype due to defective chloride channel ClC-Kb can mimic both diuretic substances. The combination of pharmacology and genetics now implicates a new terminology: furosemide-like salt-losing tubulopathy for the hyperprostaglandin E syndrome, thiazide-like salt-losing tubulopathy for Gitelman syndrome, and Bartter-like saltlosing tubulopathy for the intermediate type presenting with elements of both phenotypical variants. q 2000 Elsevier Science B.V. All rights reserved.

Introduction In 1957, the American pediatricians Rosenbaum and Hughes reported a new syndrome in a 2 month old patient who presented with persistent hyokalemic alkalosis, diarrhea, hyposthenuria, dehydration, and malnutrition with lethal outcome at the age of 7.5 months (Rosenbaum and Hughes, 1957). Five years later, Bartter observed two cases with hypokalemic salt loss, elevated aldosterone synthesis, and hyperplasia of the juxtaglomerular cells (Bartter et al., 1962). Since this case report the combination of hypokalemic alkalosis and normotensive hyperaldosteronism is

328

N. Jeck et al.

termed Bartter syndrome. Susceptibility to tetany initially described by Bartter in these patients was con®rmed by Gitelman et al. (1966) in a case with hypokalemic alkalosis and concomitant hypomagnesemia and hypocalciuria, later termed Gitelman syndrome. The pathogenetic homogeneity of the hypokalemic tubulopathy was questioned as early as 1971 when cases with extreme hypercalciuria and nephrocalcinosis were described (Fanconi et al., 1971; McCredie et al., 1971, 1974). This variant displayed severe hyposthenuric polyuria and potentially lethal salt loss (Ohlsson et al., 1984). Because this clinical picture was associated with elevated prostaglandin E2 (PGE2) activity and therefore responded well to treatment with prostaglandin synthesis inhibitors, i.e. indomethacin, it was termed hyperprostaglandin E syndrome (Seyberth et al., 1985,1987). All three variants of Bartter syndrome show activation of the renin±angiotensin± aldosterone system with corresponding development of hypokalemic alkalosis but normal blood pressure. Absense of hypertension despite elevated angiotensin formation was originally explained by putative decreased sensitivity of the vascular system to this hormone (Bartter et al., 1962). Hyperplasia and hypertrophy of the juxtaglomerular apparatus were also compatible with the postulated predominant role of the renin±angiotensin system for pathogenesis. Later, this hypothesis has been corrected when treatment directed at the renin±angiotensin system turned out to have no suf®cient therapeutic potency (Trygstad et al., 1969). Moreover, activation of the renin±angiotensin system was also described secondary to other salt loss diseases of gastrointestinal or respiratory origin (Bartter et al., 1976). In the following decade, overproduction of prostaglandins was thought to play a crucial role in disease pathogenesis (Seyberth et al., 1985; Fichman et al., 1976; Gill et al., 1976; Verberckmoes et al., 1976). But, as the antenatal variant did not completely normalize despite suf®cient suppression of prostaglandin synthesis by cyclooxygenase inhibitors, such as indomethacin, and because the other two variants were even less responsive to indomethacin, a primary tubular defect was considered. Intensive clearance studies were performed on many patients with hypokalemic salt-losing tubulopathies in order to localize the putative defect in the renal tubular system (Chaimovitz et al., 1973; Sutton et al., 1988). These studies did not succeed in pinpointing functional disturbances to a distinct tubular region probably due to compensatory mechanisms of the unaffected tubular regions. Nephropharmacology was more successful in detecting the underlying defect by analogy conclusions on the effects of thiazide (hypokalemia, hypomagnesemia, hypocalciuria) and furosemide (hypokalemia, hypercalciuria, polyuria): Gitelman patients were shown not to respond upon administration of thiazide diuretics (Puschett et al., 1988; Sutton et al., 1992), while in contrast, patients with hyperprostaglandin E syndrome were found to be insensitive towards furosemide administration (Kockerling et al., 1996). Apparently, the tubular target was resistant against these drugs in each case.

Hereditary hypokalemic salt-losing tubulopathies

329

Clinical features Thiazide-like salt-losing tubulopathy Taking the pharmacological aspects into account, the hypocalciuric hypomagnesemic Gitelman syndrome can be described functionally as a thiazide-like saltlosing tubulopathy (TSLT; Table 1) (Gitelman et al., 1966; Rudin et al., 1984; Sutton et al., 1992). Because renal salt loss is relatively mild in this disorder, age at diagnosis ranges from infancy to adulthood. Affected individuals may present with fatigue, muscle weakness, tetany, carpopedal spasms, obstipation, chondrocalcinosis, and small stature (Seyberth et al., 1998; Bettinelli et al., 1999). In other cases, the diagnosis is coincidentally made by preoperative serum electrolyte measurements or within enuresis diagnostics. Furosemide-like salt-losing tubulopathy Hyperprostaglandin E syndrome or hypercalciuric antenatal Bartter syndrome may be functionally described as a furosemide-like salt-losing tubulopathy (FSLT; Table 1) (Fanconi et al., 1971; McCredie et al., 1974; Seyberth et al., 1987; Deschenes et al., 1993). The onset of this disorder is typically prenatal. Fetal polyuria leads to development of a polyhydramnios around the ®fth month of pregnancy, which naturally results in premature birth. The pre-term infants show markedly elevated diuresis and salt-wasting. In this early phase, hyponatremic dehydration is the main symptom so that differential diagnosis to pseudohypoaldosteronism type 1 or, following the new nomenclature, the salt-losing tubulopathy of the amiloride type has to be made. During the further course, hypercalciuria and hypokalemic alkalosis of variable severity are regularly found in this syndrome. Nephrocalcinosis which is partly attributable to hypercalciuria occurs already within the ®rst weeks of life (Seyberth et al., 1985; Shoemaker et al., 1993). Polyuria and salt loss correlate well with the elevation of renal excretion of PGE2 and its main metabolite PGE-M (Seyberth et al., 1987). Frequently ± probably Table 1 Proposal of a new terminology of the hereditary hypokalemic salt-losing tubulopathies

Conventional terminology

New terminology

Gitelman syndrome Classic Bartter syndrome Hyperprostaglandin E-syndrome/antenatal Bartter syndrome

Thiazide-SLT (TSLT) Bartter-SLT (BSLT) Furosemide-SLT (FSLT)

330

N. Jeck et al.

mediated by PGE2 ± systemic symptoms such as fever, vomiting, secretory diarrhea, and osteopenia are found (Seyberth et al., 1998). Consequently, severe dystrophia and growth retardation can be observed under inadequate therapy (Seidel et al., 1995). The coincidental occurrence of an FSLT resembling phenotype with inner ear deafness reported in a single Israeli Beduin family is probably a distinct entity (Landau et al., 1995). The association of a tubular disorder with sensorineural deafness also found in autosomal recessive distal renal tubular acidosis (Karet et al., 1999), is of particular interest, because it provides further evidence that the functions of kidney and inner ear may be mediated by similar regulatory pathways. Bartter-like salt-losing tubulopathy Classic Bartter syndrome with intermediate clinical features may be classi®ed in the new terminology as Bartter-like salt-losing tubulopathy (BSLT; Table 1; Clive, 1995; Schwartz and Alon, 1996; KaÂrolyi et al., 1998; Scheinman et al., 1999). This variant with isolated hypokalemic alkalosis without major disturbances of renal calcium and magnesium handling does not present with such a uniform clinical picture as the other two tubulopathies. Post-natal salt and water loss is less pronounced than in the FSLT-type, however, failure to thrive is frequently observed from early infancy on. Beyond infancy, marked hypokalemic alkalosis is the predominant clinical feature. Without further biochemical evaluation, it is sometimes dif®cult to separate this entity from TSLT. Clinically, the only apparent difference is the occurrence of the Chvostek sign and carpopedal spasms exclusively in the TSLT variant.

Genetics Candidate genes In order to explain the pathogenesis of congenital hypokalemic salt-losing tubulopathies, various channels and cotransporters participating in tubular salt reabsorption were particularly attractive candidates (Fig. 1). An important hint was given by demonstration that loss of function mutations in the amiloride-sensitive sodium channel ENaC cause pseudohypoaldosteronism type 1, an inverse salt-losing tubulopathy characterized by low serum sodium but high potassium levels together with metabolic acidosis (Chang et al., 1996a).

Hereditary hypokalemic salt-losing tubulopathies

Fig. 1. Transepithelial NaCl reabsorption in the TAL and DCT. (a) Model of the electrogenic Cl 2 reabsorption in the TAL: chloride is reabsorbed across the apical cell membrane by the furosemide-sensitive Na±K±2Cl cotransporter (NKCC2). This electroneutral transporter is driven by the low intracellular Na 1 concentration generated by the Na±K-ATPase. Chloride ef¯ux through the basolateral membrane is mediated by channel molecules, most likely by ClC-Kb, and the electroneutral K-Cl-cotransporter. With respect to the stoichiometry, ef®cient functioning of the NKCC2 is dependent on apical K 1 recycling via the ROMK. In addition, apical secretion of positively charged K 1 as well as basolateral secretion of negatively charged Cl 2 establish a transepithelial potential difference that drives reabsorption of Na 1 and divalent cations, such as Ca 21 and Mg 21, through a paracellular pathway. (b) Model of NaCl reabsorption in the DCT: the driving force for NaCl entry via the thiazide-sensitive cotransporter, NCCT, is an electrochemical gradient across the apical membrane with low intracellular Na 1 concentration. Na 1 continuously leaves the cells into the blood stream through the Na±K-ATPase and the basolateral Cl 2 exit is maintained by channel molecules, probably by the ClC-Kb.

331

332

N. Jeck et al.

Thiazide-sensitive NaCl cotransporter Early microperfusion studies on dissected nephrons allowed the postulation of a sodium-dependent thiazide-blockable chloride transport in the distal tubule (Ellison et al., 1987). Originally, the gene of the corresponding NaCl cotransporter (NCCT) was cloned in ¯ounder (Gamba et al., 1993). Soon thereafter, the human gene SLC12A3 encoding NCCT was cloned and its structure clari®ed (Simon et al., 1996c). It is located on chromosome 16q13 and consists of 26 exons. Human NCCT (hNCCT) has 1021 amino acids with 12 putative transmembrane domains and a long cytosolic C-terminus. This structure is characteristic for the whole family of electroneutral cation-chloride cotransporters (Mount et al., 1997). The protein sequence of hNCCT shows 89% sequence identity to the rat homologue and 63% identity to the ¯ounder protein. In contrast to rat NCCT, hNCCT contains additional 17 C-terminal amino acids, which are encoded by a separate exon and include a potential proteinkinase A phosphorylation site indicating a possible regulatory function. Maybe, an alternative human splice variant of NCCT exists lacking these 17 amino acids as observed in rats (Gamba, 1999). In kidney, hNCCT is speci®cally expressed in the distal convoluted tubule (Obermuller et al., 1995; Mastroianni et al., 1996b). NCCT transcripts are also found in spleen, smooth and cardiac muscle, gallbladder, ileum, prostate gland, and placenta as well as in an osteoblast cell line (Chang et al., 1996b; Barry et al., 1997; Gamba, 1999). Furosemide sensitive Na±K±2Cl cotransporter type 2. Like NCCT, the furosemide-sensitive Na±K±2Cl cotransporters belong to the family of electroneutral cation-chloride cotransporters. Two types can be distinguished: (1) the ubiquitously expressed secretory type 1 (NKCC1) and (2) the kidney-speci®c absorptive type 2 (NKCC2). The genes encoding these cotransporters are localized on different chromosomes but the proteins show over 60% sequence homology. Long before cloning of the gene encoding NKCC2, the mechanism of potassium dependent NaCl reabsorption in TAL was known, the stoichiometry of this symport clari®ed (Greger, 1985), and the pharmacological blockage by furosemide demonstrated (Muschaweck and Hadju, 1964; Burg et al., 1973). Rodent gene homologues were cloned in the mid 1990s from kidney tissue of rat, mouse, and rabbit (Gamba et al., 1994; Payne and Forbush, 1994; Igarashi et al., 1995). Shortly after, the human gene, SLC12A1, was cloned (Simon et al., 1996a). It is situated on the long arm of chromosome 15 and comprises 26 exons. Among these, exon 4 exists in three variants sequentially following one another within the gene (Vargas-Poussou et al., 1998). This exon variability, which mainly codes for the second transmembrane domain of the protein, is also conserved in rat, mouse, and rabbit. Alternative

Hereditary hypokalemic salt-losing tubulopathies

333

splicing of the different exon 4 variants results in three NKCC2 isoforms designated A, B, and F. They differ in their expression along the nephron and in their functional characteristics (Payne and Forbush, 1995; Yang et al., 1996). In rat mTAL, only the isoforms A and F are found whereas the isoforms A and B are predominant in the cortical region of the ascending loop of Henle and in the macula densa. More distally, in the outer medullary collecting duct, the isoform F is expressed. Each isoform is thought to have its own characteristic physiological signi®cance because of their different af®nities for cations and varying transport capacity. Commom to all hNKCC2 isoforms are 1099 amino acids with twelve putative membrane-spanning segments like NCCT. Luminal potassium channel The renal outer medullary potassium channel ROMK belongs to a family of structurally and functionally related K 1 channels designated as Kir channels (Doupnik et al., 1995). Kir channels are assembled from four subunits. Hydropathy and sequence analyses predict cytoplasmic N- and C-termini with a well conserved core region that consists of two transmembrane domains, ¯anking the pore-forming segment containing the selectivity ®lter for potassium ions (G-Y/F-G-motif). This con®guration has been con®rmed by the crystal structure of the potassium channel from Streptomyces lividans (Doyle et al., 1998). ROMK was cloned in the early 1990s from rat kidney (Ho et al., 1993) and subsequently from human kidney (Shuck et al., 1994). Next to kidney, transcripts of this channel are also found in brain (Kenna et al., 1994) and in the hair cells of the organ of Corti (Glowatzki et al., 1995). The encoding gene, KCNJ1, resides on chromosome 11q24 and consists of 5 exons. Only exons 2, 4, and 5 contain information directly translated into the protein sequence, while the signi®cance of exons 1 and 3 are unknown to date. Alternative splicing can generate at least ®ve transcripts that code for only three different proteins, ROMK 1±3, because of inclusion of noncoding exons (Fig. 2). The isoproteins differ in length and sequence of the initial Nterminus (Shuck et al., 1994). Little is known about the functional difference of the three isoforms until now, but it was demonstrated that ROMK isoforms are differentially expressed along the nephron: ROMK1 is speci®cally expressed in the apical membrane of the collecting duct, and ROMK3 is only expressed in the ascending loop of Henle and the distal convolute, while ROMK2 with the shortest N-terminal region is wedely distributed along the loop of Henle and distal nephron (Boim et al., 1995). The isoforms consist of 372±391 amino acids and, analogous to other Kir channels, form homo- and heterotetramers that may contribute to diversity of tissuespeci®c function (Duprat et al., 1995; Derst et al., 1998).

334

N. Jeck et al.

Fig. 2. Genomic organization of KCNJ1 and illustration of an unique molecular ®nding in a consanguineous Indian FSLT kindred (Feldmann et al. 1998). KCNJ1 consists of ®ve exons, with exon 5 encoding the main part of the protein, and at least two promotor elements 5 0 of exon 1 and 4, respectively. Alternative splicing accounts for ®ve distinct transcripts, and three protein isoforms. The promotor region 5 0 of exon 1 is utilized from ROMK isoforms 2 and 3, whereas transcription of the ROMK1 isoform is probably initiated from the second promotor 5 0 of exon 4, and may not be dependent on an additional upstream promotor. In one kindred, deletion of exon 1 and 2 together with the upstream promotor region has been demonstrated to account for the FSLT phenotype, suggesting a predominant role of ROMK2 and ROMK3 for K 1 recycling in the TAL.

Basolateral chloride channel The kidney-speci®c chloride channel ClC-Kb belongs to an expanding family of voltage-gated chloride channels, originally discovered by the cloning of the ClC-0 chloride channel from Torpedo electric organ (Jentsch et al., 1990). Presumably, ClC channels function as homodimers containing two chloride-conduction pores (Ludewig et al., 1996). The monomers consists of 13 hydrophobic helical domains with the fourth domain situated in the exterior and not crossing the membrane. Both its N- and C-termini reside in the cytoplasm. To date, the genes for nine human voltage-gated chloride channels have been cloned, and among these, mutations in the genes CLCN1 and CLCN5 have been shown to cause human diseases, such as myotonia congenita Becker and Thomsen (Koch et al., 1992; Steinmeyer et al., 1994) and X-linked recessive nephrolithiasis or Dent's disease (Lloyd et al., 1996). The gene encoding ClC-Kb, CLCNKB, consists of 19 exons and is localized on chromosome 1p36. In close vicinity of approximately 11 kb genomic DNA resides the gene for the 90% homologous kidney-speci®c chloride channel, ClC-Ka, suggesting evolutionary gene duplication. Recently, the generation of a ClC-Ka knockout mouse was successful (Matsumura et al., 1999), showing a diabetes insipidus phenotype,

Hereditary hypokalemic salt-losing tubulopathies

335

promoting the human homologue as an attractive candidate for NDI cases without mutations in the vasopressin receptor or in the aquaporin water channel. ClC-Kb consists of 687 amino acids (Kieferle et al., 1994). Its membrane topology is mainly deduced by homology conclusions from studies on other voltage-dependent chloride channels because heterologous expression and functional characterization of ClC-Kb turned out to be extremely tricky (Jentsch et al., 1995; Waldegger and Jentsch, 2000). Only 11 of 13 hydrophobic stretches seems to be integrated into the membrane. Apart from the extracellular D4, the C-terminal D13 helix does not cross the membrane but is located within the cytoplasm (Schmidt-Rose and Jentzch, 1997). ClC-Kb channels are likely homodimers (Middleton et al., 1996; Waldegger and Jentzch, 2000), even though the possibility of multimerization has not been di®nitely ruled out (Steinmeyer et al., 1994). Moreover, it remains to be clari®ed, whether ClC-Kb and ClC-Ka might also constitute heteromultimers. The distribution of CIC-Kb in the kidney and its targeting into the membrane of the epithelial cells are controversial. The only study performed on human kidney is suggestive for expression in the glomeruli, proximal tubule, and collecting duct (Takeuchi et al., 1995) with targeting of the protein to the apical membrane. In contrast, the rat homologue is predominantly expressed in the distal parts of the nephron (Vandewalle et al., 1997), polarized to the basolateral membrane of tubular epithelia from mTAL to the cortical collecting duct. A con®rmation of any of these ®ndings may be obtained from thorough clinical evaluation of patients with loss of function mutations in the CLCNKB gene. Linkage studies As a result of the previous pharmacological studies, the understanding of the pathophysiology was well advanced, reducing the number of candidate genes to a minimum. Therefore, linkage analysis followed a selective candidate gene approach rather than a genome-wide screening. Very soon, the genetic heterogeneity of the autosomal recessively transmitted hypokalemic salt-losing tubulopathies became apparent (KaÂrolyi et al., 1996; Simon and Lifton, 1996). Families with the TSLT phenotype almost always showed linkage to the SLC12A3 locus on chromosome 16q13, whereas pedigrees with FSLT and BSLT did not. Among the FSLT population, further genetic heterogeneity was found. Nearly half of the FSLT families were linked to the cytogenetic region 15q15±21 harbouring the SLC12A1 gene, while the other half were compatible with linkage to the KCNJ1 locus on chromosome 11q24. Finally, BSLT turned out to be associated with the gene encoding ClC-Kb, CLCNKB, on chromosome 1p36 (Table 2). In addition the association of a FSLT-like tubulopathy with sensorineural deafness was described in a large Beduine family with seven affected individuals (Landau et al., 1995). This syndrome did not cosegregate with any of the known

336

N. Jeck et al.

Table 2 Aetiology of the hereditary hypokalemic salt-losing tubulopathies

Affected tubular site

Defective protein

Thiazide-like SLT (TSLT) Distal convoluted tubule NCCT Bartter-like SLT (BSLT) Distal tubule ClC-Kb Furosemide-SLT (FSLT) Thick ascending loop of Henle NKCC2 ROMK

Mutant gene SLC12A3 CLCNKB SLC12A1 KCNJ1

gene loci. Instead, linkage to chromosome 1p31 was established recently (Brennan et al., 1998). A causative gene has not been identi®ed yet. Aside from this exception, nearly all cases of inherited hypokalemic salt-losing tubulopathies could be explained by mutations in one of the above genes. Mutational analysis NCCT was the ®rst tubular membrane protein in which mutations related to a hypokalemic salt-losing tubulopathy were identi®ed (Simon et al., 1996c). To date, about 120 point mutations in SLC12A3 have been described (Mastroianni et al., 1996a; Simon et al., 1998), with clustering of mutations in the cytoplasmic Cterminus (Lemmink et al., 1998) (Fig. 3). In the majority, these are missense mutations leading to an exchange of a single amino acid residue. More rare are

Fig. 3. Structural model of NCCT and localization of the amino acid residues affected by mutations in patients with thiazide-like salt-losing tubulopathy.

Hereditary hypokalemic salt-losing tubulopathies

337

deletions or insertions of single bases that lead to frameshifts and premature stop codons or splice-site mutations, which are expected to disturb correct mRNA processing. Missense mutations are also predominant in SLC12A1 (Simon et al., 1996a; Vargas-Poussou et al., 1998) (Fig. 4). Surprisingly, no missense mutations in the cytoplasmic C-terminus of NKCC2 have been described until now. Instead, all mutations in that region were splice-site or truncating mutations. Almost all mutations found in KCNJ1 affect exon 5 which encodes the main part of the ROMK protein (Simon et al., 1996b; KaÂrolyi et al., 1997; Vollmer et al., 1998) (Fig. 5). As in the two cotransporters, the majority of ROMK mutations represents nonconservative amino acid exchanges. The localization of the ROMK mutations is in accordance with regions of functional relevance, e.g. several mutations affect the putative channel pore, others cluster around the cytosolic components of the pHsensor or the PKA binding motif in the C-terminus. To date, 15 missense mutations are known in the CLCNKB gene distributed throughout the channel protein (Simon et al., 1997; Konrad et al., 2000) (Fig. 6). Additionally, splice-site and frameshift mutations were described. There is an exceptionally high frequency of deletions of the entire gene, suggesting that the close vicinity of the almost identical CLCNKA gene predisposes to inhomologous recombinations with breakpoints in the non-coding regions 3 0 of the genes. In rare cases, an unequal crossing over event with breakpoints within the two genes has been identi®ed, probably leading to a chimeric chloride channel with N-terminal sequences of ClC-Ka and C-terminal sequences of ClC-Kb (Fig. 7).

Fig. 4. Structural model of NKCC2 and distribution of missense and non-sense mutations linked to furosemide-like salt-losing tubulopathy.

338

N. Jeck et al.

Fig. 5. Putative topology of the ROMK subunit; four subunits assemble into a functional channel forming a central K 1-selective aqueous pore. Missense mutations may be subdivided in cytosolic mutations that disturb pH-dependent gating and mutations in the core region that alter the channel pore and subunit assembly.

Fig. 6. Predicted topology of ClC-Kb and distribution of missense and non-sense mutations associated with Bartter-like salt-losing tubulopathy.

Hereditary hypokalemic salt-losing tubulopathies

339

Fig. 7. Schematic model of an unequal crossing over between CLCNKA and CLCNKB leading to a chimera of both genes. The close vicinity of the almost identical chloride channel genes may predispose to a high rate of non-homologous recombinations.

Genotype±phenotype correlations Prerequisite of a genotype±phenotype correlation study is a representative size of the patient cohort and careful clinical evaluation of all individuals. Frequently, data are collected retrospectively using a questionnaire inquiring the cardinal symptoms of the patients treated in external clinics. A standardized procedure is not very well achievable because diagnostic and therapeutic standards vary from one clinic to the next. This situation should be taken into account when studying publications on this topic. The following genotype±phenotype correlation is based on data recruited in several clinical centers but including only such patients in which either a homozygous mutation or two heterozygous mutations (compound heterozygous) could be identi®ed. Ninety-eight genotyped patients were included. Of these, 24 displayed mutations in NCCT, 14 in NKCC2, 33 in ROMK and 27 in CLC-Kb. Patients with NCCT mutations almost invariably revealed discordant renal handling of magnesium and calcium responsible for hypomagesemia and hypocalciuria, which is per de®nition the pathognomonic feature of TSLT. Patients with NKCC2 or ROMK mutations displayed the trias of polyhydramnios, iso-/hyposthenuria and nephrocalcinosis. In contrast, patients with mutations in ClC-Kb presented with variable clinical features, and may correlate best with Bartter's original description

340

N. Jeck et al.

(onset in early childhood with isolated hypokalemic alkalosis). Typical signs of impaired TAL function, such as isosthenuria and hypercalciuria with consecutive nephrocalcinosis usually did not occur. Three of the 27 patients with CLC-Kb mutations had a TSLT phenotype. Only one patient developed the clinical characteristics of FSLT. The wide phenotypic spectrum related to CLCNKB mutations might be explained with the interindividually different ability to compensate the disturbances in basolateral chloride outward current by activating alternative pathways for chloride exit from the cell into the blood. Taken together, NCCT mutations cause a phenotype associated with dysfunction of the distal convoluted tubule, whereas molecular defects in either NKCC2 or ROMK are associated with impaired TAL function. The precise localization of the tubular dysfunction due to ClC-Kb mutations is more dif®cult to assess because of the variability of related phenotypes. As ClC-Kb is ubiquitously expressed in the distal parts of the nephron, the clinical presentation might be in¯uenced by the individual capacity to compensate dysfunction of the TAL system or the distal tubule. Diagnosis All three salt-losing tubulopathies result in secondary hyperaldosteronism responsible for hypokalemic alkalosis with serum potassium values between 1.8 and 3.3 mmol/l. Metabolic alkalosis may be determined by elevated serum concentration of bicarbonate within a range of 26±33 mmol/l corresponding to a positive base excess of 5±15 mmol/l. An indication of a primary defect in the distal tubule is a signi®cantly elevated renal sodium and chloride loss with high excretion rates of 5±10 mmol/kg per day. In patients with neonatal manifestation the excretion rate can even reach up to 50 mmol/kg per day (Konrad et al., 1999). In contrast, elder patients with long-term negative NaCl balance may show excretion rates of only 4±6 mmol/kg per day. Fractional excretion rates for electrolytes calculated from the quotient of electrolyte clearance/creatinine £ 100 can serve as an additional indicator for renal electrolyte loss. Usually, the fractional excretion rate of sodiuim (FeNa) is more than 1% and that of potassium (FeK) more than 30% for all SLTs. In the neonatal period and early infancy, marked hyponatremia and hypochloremia are common ®ndings due to the extremely high renal salt losses. Later, serum concentrations of sodium and chloride tend to be only few lower compared to normal subjects. A more precise parameter for determining the extent of NaCl loss are plasma concentrations of renin and aldosterone which are often extremely elevated in untreated patients. Histological examination revealing hypertrophy of the macula densa, the main source of renin synthesis, gives no further insights for diagnosis, because it unspeci®cally re¯ects body NaCl depletion and volume contraction only.

Hereditary hypokalemic salt-losing tubulopathies

341

In contrast, the determination of the renal excretion rates of prostaglandin E2 (PGE2) and its main metabolite, PGE-M, has direct therapeutic implications even though the elevated renal (PGE2) and systemic (PGE-M) prostaglandin synthesis is not the primary pathophysiological event. PGE2 overproduction drives the clinical course of FSLT and, to a lesser extent, that of BSLT. The mechanisms leading to increased synthesis are not yet fully understood. Hypokalemia has been excluded as a possible initiating trigger (Konrad et al., 1999; overview in Seyberth et al., 1998). Signi®cant differences distinguishing the hypokalemic salt-losing tubulopathies from each other are age of onset, degree of polyuria, and renal handling of calcium and magnesium. Table 3 gives an overview of the decisive clinical and laboratory parameters of diuretic-like and Bartter-resembling salt-losing tubulopathies. No polyhydramnios is observed in TSLT and children are born at term. Important for the mild clinical picture is that the kidney has almost normal capacity to concentrate urine. Signi®cant for TSLT diagnosis is symptomatic hypomagnesemia with subsequent tetany and carbopedal spasms distinguishing it from both BSLT and FSLT. FSLT is characterized by antenatal onset and severe polyuria and saliuresis. Urine osmolality ranges from hyposthenuric to isosthenuric values, very rarely slightly above 400 mmol/kg. Urinary calcium excretion is high leading to nephrocalcinosis within a few weeks after birth. BSLT represents an intermediate disorder between TSLT and FSLT. Birth is usually at term. During the ®rst year of life, hypochloremia and marked hypokalemic alkalosis secondary to activation of the renin±angiotensin±aldosterone system might be observed. Polyuria is not as predominant as in FSLT, and urine concentration ability is only slightly impaired. Similar to TSLT, serum magnesium levels are frequently at the lower normal limit or even below. Urinary calcium excretion is normal and nephrocalcinosis is rarely found in this tubulopathy. Urinary calcium excretion can be determined in 24 h urine or, if not available, by the quotient of urinary calcium and creatinine concentrations in morning urine. Urinary excretion of calcium is age-dependent, with higher level in early infancy. In children aged 13±24 months, hypercalcuria is diagnosed when values of more than 1.6 mol/mol creatinine or 0.6 mg/mg creatinine are found corresponding to an absolute daily excretion rate of approximately 7 mg/kg per day or 175 mmol/kg per day. Hypomagnesemia is de®ned by serum levels below 0.65 mmol/l. Therapy and prognosis Manifestation in the neonatal period and early infancy requires supplementation of ¯uid and sodium chloride suf®cient for compensating the high renal excretion rates (up to 20 ml/kg per h and 50 mmol/kg per day, respectively). Careful monitoring of urinary ¯uid and electrolyte excretion is mandatory for correct adjustment of the replacement therapy. Inappropriate supply of ¯uid and electrolytes can result in

342

Table 3 Clinical and biochemical characteristics of diuretic- and Bartter-like salt losing tubulopathies (SLTs) a

Serum-potassium (mmol/l)

Nephrocalcinosis Serum-magnesium (mmol/l) PGE2/PGE-M a

Bartter-SLT (classic Bartter syndrome)

Furosemide-SLT (hyperprostaglandin E syndrome) NKCC2-type

ROMK-type

# (1.9±3.2) 2 39±41 2 500±900 # 0.1±2 ± # 0.3±0.75 $ (")

# (1.7±3.4) 2 (1) 33±41 2 (1) 500±900 $ 0.5±6 ± (1) $/# 0.5±0.9 "

# (1.8±3.2) 1 27±36 1 100±350 " 10±25 1 $ (#) 0.6±1.0 "

# ($) (2.0±3.7) 1 28±36 1 100±450 " 10±25 1 $ (#) 0.5±1.1 "

Parentheses indicate that this characteristic has (1) or has not (2) been observed in single cases.

N. Jeck et al.

Polyhydramnios leading to prematurity (weeks of gestation) Polyuria Urine osmolality (mmol/kg) Calcium excretion (mg/kg per day)

Thiazide-SLT (Gitelman syndrome)

Hereditary hypokalemic salt-losing tubulopathies

343

acute hypovolemic renal failure, fever, and epileptic seizures. Potassium supplementation, preferably as potassium chloride, becomes necessary as soon as hypokalemic alkalosis has developed as a consequence of secondary hyperaldosteronism (Reinalter et al., 1998a, Konrad et al., 1999) For FSLT and BSLT, the standard therapeutic approach is based on the ®nding of renal and systemic overproduction of prostaglandin E2. Renal prostaglandin E2 may reduce transepithelial electrolyte transport in the ascending loop of Henle, by this aggravating of renal salt losses (Fernandez-Llama et al., 1999), while increased systemic PGE2 activity drives the clinical picture including fever, vomiting, diarrhea, and severe failure to thrive. Consequently, suppression of prostaglandin E2 formation with the unselective cyclooxygenase inhibitor indomethacin, has bene®cial effects on reduction of saliuretic polyuria and amelioration of systemic symptoms. Usually, renal loss of sodium chloride, calcium, and water may be reduced by approximately 50%. Moreover, indomethacin therapy restores normal physical development and even induces catch-up growth. The therapeutic success is sustained by careful clinical and laboratory monitoring. Serum concentration of indomethacin has to be checked regularly, especially, in the pre-term infants and in the initial treatment phase. The therapeutic range for newborns 10 h after the last dose and for children 4 h after the last dose is between 0.5 and 1.5 mg/ml. Renal excretion of PGE2 and PGE-M as well as plasma renin activity should normalize under this treatment. The individual dose for a child can vary between 0.2 and 5 mg/ kg per day. After 15 years experience, the risk-bene®t ratio of a long-term treatment with indomethacin (10±20 years) is considered to be very favorable. Other unspeci®c prostaglandin synthesis inhibitors such as acetylsalicylic acid, ibuprofen, and naproxen have not been as successful. Either they cause more severe gastrointestinal side-effects or show less ef®cacy on suppression of renal and systemic PGE2 synthesis. Because the renal prostaglandin synthesis, especially PGE2 and PGI2, has an important protective effect on renal perfusion, particularly under conditions of volume contraction, suf®cient dietary intake of NaCl, potassium, and water should be guaranteed. This is very important in conditions of additional extrarenal salt loss due to diarrhea or increased perspiration. Additionally, the indomethacin intake should be as regular as possible, because even a missed dose for 1±2 days duration can severely increase renal salt loss and lead to hypovolemia. The continuation of indomethacin therapy without preceding salt and liquid substitution may, under certain circumstances, cause acute renal failure. Even under careful monitoring, indomethacin may cause a reversible decrease of the glomerular ®ltration rate (GFR) by 15% on average. The reversibility of this effect has been demonstrated by follow-up studies on patients who were continuously treated over 15 years with indomethacin (Reinalter et al., 1998b).

344

N. Jeck et al.

Administration of hydrochlorothiazide for patients with FSLT decreases renal calcium excretion in addition to the effect of indomethacin, but the hypokalemia and hyperuricemia as well as the GFR can be in¯uenced negatively. For BSLT, indomethacin treatment is only partially effective requiring additional potassium administration, in contrast to the ROMK type of FSLT, which virtually needs no potassium supplementation when on indomethacin. In single BSLT cases, ACE inhibitors have been bene®cial (Meyburg et al., 1999), but experience with this substance group is fairly poor. In the TSLT variant, administration of amiloride, which blocks the aldosteronedependent apical Na + channel ENaC, can partially compensate the negative potassium balance. A major challenge in therapy is ameliorating clinical symptoms such as carpopedal spasms, tetany, abdominal and joint pain attributable to hypomagnesemia. Generally, no standard drug therapy can be recommended. Because of dramatically impaired renal conservation of magnesium, even high magnesium intake, reaching the limit of intestinal tolerability, can hardly provide an adaquate supply for normalization of serum magnesium levels. Correction of hypokalemia might have some bene®cial effect on reduction of renal magnesium loss. Pathogenesis The abnormal response of FSLT patients to furosemide and of TSLT patients to thiazide administration pointed to an impaired salt reabsorption in the thick ascending limb (TAL) of the loop of Henle and in the distal convoluted tubule (DCT), respectively (Sutton et al., 1992; KoÈckerling et al., 1996; Colussi et al., 1997). Even before the era of molecular biology, important determinants of NaCl reabsorption in TAL and DCT were characterized by means of perfusion studies on isolated nephron segments (Murer and Greger, 1982; Burg and Good, 1983; Greger, 1985; Stokes, 1989). Physiology of salt reabsorption in TAL and DCT Approximately 30% of the ®ltered load of sodium is reabsorbed in TAL. Salt reabsorbtion by the medullary TAL is essential for countercurrent multiplication and the generation of medullary hypertonicity, a major prerequisite for excretion of a concentrated urine. Transepithelial NaCl reabsorption follows an electrochemical gradient across the apical membrane with low Na 1 concentration in the cell. Continuous Na 1 exit from the cell into the blood is sustained by the basolateral Na±KATPase, thereby allowing the furosemide-sensitive cotransporter NKCC2 to passively conduct Na 1 together with two chloride ions and one potassium ion through the apical membrane into the cytosol (see also Fig. 1a). With respect to the stoichiometry of this symport and to the fact that K 1 concentration in the lumen

Hereditary hypokalemic salt-losing tubulopathies

345

is much lower than Na 1, permanent K 1 recycling via ROMK is required, ensuring a suf®cient supply of K 1 that enables the continuous operation of NKCC2 (Greger, 1985; Giebisch, 1998). The apical potassium conductance is also essential for the generation of the lumen-positive potential difference, which provides a signi®cant driving force for passive absorption of Ca 21, Mg 21 and Na 1 along the paracellular pathway. Chloride ions entering the cytosol via NKCC2 may leave the cell through channel molecules or a K±Cl cotransporter in the basolateral membrane. It remains elusive which of these two mechanisms is critical by means of quantity. The DCT accounts for approximately 7% of Na 1 reabsorption in the nephron. The activity of the Na±K-ATPase is comparable to that in the TAL, establishing a suf®cient electrochemical gradient for passive Na 1 entry through the apical membrane (see also Fig. 1b). The major Na 1 reabsorption pathway in the apical membrane is the NaCl cotransporter NCCT, which is the major target for thiazide diuretics. Chloride ef¯ux into the blood is mainly mediated by channel molecules in the basolateral membrane. In addition to Na 1 transport, the DCT participates in renal Mg 21 and Ca 21 homeostasis. Sodium and Ca 21 reabsorption are inversely related. Functional consequences of mutations linked to hypokalemic SLTs Analyses of the mechanisms by which naturally occurring mutations lead to loss or reduction of protein activity are most advanced in ROMK. Various heterologous expression systems were applied, such as COS 7 cells (Derst et al., 1997), SF9 cells (Schwalbe et al., 1998), and Xenopus oocytes (Derst et al., 1998; Schulte et al., 1999). Several mutations affecting the channel pore exhibited no measurable potassium currents or only very small currents (Derst et al., 1997). Mechanisms leading to the reduced potassium conductance still have to be clari®ed in detail, possibly the mutations lead to changes of the pore structure. But, also mutations in the ®rst transmembrane helix, which is predicted not to contribute to pore lining (Minor et al., 1999), as well as mutations in the extracellular loop are associated with reduced currents. The N124K mutation within the extracellular loop displayed 12-fold decreased macroscopic potassium currents when compared to wild-type currents. As single channel conductance and open probability were normal, the loss-of-function is likely related to disturbance of channel assembly and decreased number of expressed functional channels (Derst et al., 1998). Mutations in the cytosolic N- and C-termini were found to affect the structural element for pH-induced channel gating. ROMK channels are naturally gated by intracellular pH with acidi®cation leading to channel closure and with half-maximal activation at intracellular pH (pH0.5) 6.8 (Fakler et al., 1996). The molecular determinant responsible for pH-gating comprises a triad of amino acid residues Arg49Lys80-Arg311. Protonization of the lysine residue blocks the channel, while disso-

346

N. Jeck et al.

ciation of the proton opens the pore. Interestingly, the majority of cytosolic mutations linked to FSLT cluster around the triad, and exhibited no potassium current under standard conditions. However, alkalinization of the cytosol did restore channel function with shifts in pH0.5 from 0.4 pH units to .2 pH units. (Schulte et al., 1999). Therefore, these cytosolic mutants encode functional channels with pHgating shifted to more alkaline values. Another subset of cytosolic ROMK mutants revealed either a de®cient processing with lack of glycosylation and failure to insert into the cell membrane (Schwalbe et al., 1998). The description of a homozygous deletion of the 5 0 part of the KCNJ1 gene in a consanguineous Indian family (Feldmann et al., 1998) is unique. The loss of transcription initiation sites is the most likely pathogenic factor. This is of particular interest, because despite the loss of the promotor 5 0 of exon 1 the ROMK isoform 1 should be synthesized (Fig. 1). However, ROMK1 is speci®cally expressed in the collecting duct, and, obviously, it cannot be recruited for potassium recycling in TAL. Naturally occurring mutations in NKCC2 related to the FLST phenotype have not been functionally characterized as yet. Site-directed mutagenesis experiments, however, identi®ed residues in transmembrane domains 2 and 7 as playing an important role for Na 1 af®nity (Isenring et al., 1998b). Transmembrane domain 7 also in¯uences the af®nity to Cl 2 in cooperation with residues in domain 4. These residues may be within the translocation pocket, interacting directly with transported ions (Isenring et al., 1998a). Furosemide-binding appears to be more complex and involves interaction with multiple regions of the transporter in the large central domain. The NKCC2 knockout mouse shows a lethal phenotype with typical symptoms of a TAL lesion and obvious parallels to FSLT (Takahashi et al., 1999). Diuresis was extremely elevated and the ability to excrete concentrated urine was impaired even in states of severe volume depletion. Plasma renin activity was stimulated and urinary Ca 21 excretion was high. Within 2 weeks, the mice died of dehydration. Indomethacin was effective in terms of amelioration of symptoms and extending the life of these animals. To date, only one study was published focussing on the mechanisms by which TSLT mutations impair function of the NCCT (Kunchaparty et al., 1999). Eight mutations distributed throughout the cotransporter were analyzed in Xenopus oocytes, showing that in each case the mutant protein was synthesized normally, but was not glycosylated and did not appear to be expressed at the plasma membrane. This effect was independent of the location of the mutation, i.e. intraor extracellularly or within a transmembrane domain. Detective protein folding is considered as the underlying pathomechanism leading to retention of the mutant protein in the endoplasmic reticulum. The NCCT knockout mice appear healthy and are normal with respect to plasma electrolyte concentrations and acid±base balance (Schultheis et al., 1998). However,

Hereditary hypokalemic salt-losing tubulopathies

347

renin release was enhanced, and the animals exhibited hypomagnesemia and hypocalciuria, indicating that the loss of NCCT activity in mouse causes only subtle perturbations of sodium and ¯uid volume homeostasis, but alters renal handling of Mg 21 and Ca 21, as observed in TSLT Little is known about the functional consequences of mutations in ClC-Kb. Very recently, several ClC-Kb mutants were analyzed in a chimeric chloride channel, containing predominantly ClC-Kb sequences except the last transmembrane domains and the initial C-terminus of ClC-Ka, which were found to be essential for stable heterologous expression (Waldegger and Jentsch, 2000). All tested ClCKb mutants linked to Bartter-like SLT revealed complete loss-of-function which was not related to decreased expression in the membrane.

Perspectives As the hypokalemic salt-losing tubulopathies represent monogenic inherited diseases, somatic gene therapy is a tempting perspective. This option, however, is still far from becoming a reality as known from research in cystic ®brosis where gene therapy is studied most intensively (reviews by Davies et al., 1998; Rosenecker et al., 1998; Riordan, 1999) without succeeding in stable expression of the wild-type CFTR gene in the target cells. For the renal tubulopathies, gene therapy appears even more dif®cult, because selective expression in speci®c segments of the nephron and polarization to either the apical or basolateral membrane of the epithelial cells have been taken into account. As basically only a single organ is affected by the disease, another curative approach is conceivable: kidney transplantation. This is the most invasive type of treatment, as it is associated with surgical risks and life-long immunosuppression, an unwarranted side-effect when taking into consideration that congenital salt-losing tubulopathies are not life-threatening diseases once the problematic neonatal period of FSLT patients has been overcome. However, there are few case reports, in which the salt-losing tubulopathy led to end stage renal disease, giving the justi®cation for renal transplantation (Arant et al., 1970; Rudin, 1988; Takahashi et al., 1996). A third possibility for therapy is directed at restoration of the mutant protein by targeting the underlying molecular defect. In this regard, the experiences made again in cystic ®brosis research may point the way for the future. Once the precise molecular pathomechanisms underlying the loss of function of the protein are clari®ed, drugs could be searched that compensate these pathogene factors. For example, in vitro experiments for some of the stop mutations in the CFTR gene showed that aminoglycosides are effective in overcoming the translation termination induced by the mutant stop codon (Howard et al., 1996). For the most frequent CFTR mutant, delF508, which is hypothesized to cause misfolding of the nascent chain and impair-

348

N. Jeck et al.

ment of protein traf®cking application of butyrates turned out to correct processing of the mutant (Rubinstein et al., 1997). A clinical pilot study demonstrated that administration of a 20 g daily dose of phenylbuturate led to detectable chloride currents in the nasal epithelia of CF patients carrying the delF508 mutation (Rubinstein and Zeitlin, 1998). These results are encouraging for novel therapy of other monogenic diseases where processing defects have been demonstrated, such as familial hypercholesterolemia, or, with respect to the kidney, renal diabetes insipidus. Indeed, the processing defect of a aquaporin 2 mutant could be corrected by chaperones, (Tamarappoo and Verkman, 1998). Pharmacological therapy directed at restoration of channel or cotransporter function in SLTs seems promising. An important prerequisite is the exact functional characterization of the mutants. Analyses of naturally occuring mutation have been performed in NCCT, CIC-Kb, and, most intensively, in ROMK. In NCCT and ROMK, the processing defect was shown to be a leading pathogenetic mechanism (Schwalbe et al., 1998; Kunchaparty et al., 1999), stimulating the search for pharmaceuticals, such as chaperones, which correct the abnormal processing defect. Moreover, several cytosolic ROMK mutations disturb the regulation of open probability by intracellular pH (Schulte et al., 1999). Therefore, pharmaceuticals, which in¯uence the pH in the submembrane environment of tubular epithelia leading to moderate alkalinization into the range of pH-gating of the individual mutant, may be allowed to, restore channel function. For NCCT, NKCC2, and ClC-Kb, studies of the functional consequences of mutants are still rare, or not available at all. However, the new insights into channel physiology, given from the analyses of naturally occurring ROMK mutants, which also give rise to new therapeutic strategies are certainly an impulse to continue genetic and functional studies in this research ®eld.

References Arant, B., Brackett, N., Young, R., Still, W., 1970. Case studies of siblings with juxtaglomerular hyperplasia and secondary aldosteronism associated with severe azotemia and renal rickets ±Bartter syndrome or disease. Pediatrics 46, 344±361. Barry, E.L., Gesek, F.A., Kaplan, M.R., Hebert, S.C., Friedman, P.A., 1997. Expression of the sodiumchloride cotransporter in osteoblast-like cells: effect of thiazide diuretics. Am. J. Physiol. 272, C109± 1C16. Bartter, F.C., Pronove, P., Gill, J.R., MacCardle, R.C., 1962. Hyperplasia of the juxtaglomerular complex with hyperaldosteronism and hypokalemic alkalosis ± a new syndrome. Am. J. Med. 33, 811±828. Bartter, F.C., Gill Jr., J.R., Frolich, J.C., Bowden, R.E., Holli®eld, J.W., Radfar, N., Keiser, H.R., Oates, J.A., Seyberth, H., Taylor, A.A., 1976. Prostaglandins are overproduced by the kidneys and mediate hyperreninemia in Bartter syndrome. Trans. Assoc. Am. Phys. 89, 77±91. Bettinelli, A., Rusconi, R., Ciarmatori, S., Righini, V., Zammarchi, E., Donati, M.A., Isimbaldi, C., Bevilacqua, M., Cesareo, L., Tedeschi, S., Garavaglia, R., Casari, G., 1999. Gitelman disease asso-

Hereditary hypokalemic salt-losing tubulopathies

349

ciated with growth hormone de®ciency, disturbances in vasopressin secretion and empty sella: a new hereditary renal tubular-pituitary syndrome?. Pediatr. Res. 46, 232±238. Boim, M.A., Ho, K., Shuck, M.E., Bienkowski, M.J., Block, J.H., Slightom, J.L., Yang, Y., Brenner, B.M., Hebert, S.C., 1995. ROMK inwardly rectifying ATP-sensitive K 1 channel. II. Cloning and distribution of alternative forms. Am. J. Physiol. 268, 1132±1140. Brennan, T.M., Landau, D., Shalev, H., Lamb, F., Schutte, B.C., Walder, R.Y., Mark, A.L., Carmi, R., Shef®eld, V.C., 1998. Linkage of infantile Bartter syndrome with sensorineural deafness to chromosome 1p. Am. J. Hum. Genet. 62, 355±361. Burg, M., Good, D., 1983. Sodium chloride coupled transport in mammalian nephrons. Annu. Rev. Physiol. 45, 533±547. Burg, M., Stoner, J., Cardinal, N., Green, N., 1973. Furosemide effect on isolated perfused tubules. Am. J. Physiol. 225, 119±124. Chaimovitz, C., Levi, J., Better, O.S., Oslander, L., Benderli, A., 1973. Studies on the site of renal salt loss in a patient with Bartter syndrome. Pediatr. Res. 7, 89±94. Chang, H., Tashiro, K., Hirai, M., Ikeda, K., Kurokawa, K., Fujita, T., 1996a. Identi®cation of a cDNA encoding a thiazide-sensitive sodium-chloride cotransporter from the human and its mRNA expression in various tissues. Biochem. Biophys. Res. Commun. 223, 324±328. Chang, S.S., Grunder, S., Hanukoglu, A., Rosler, A., Mathew, P.M., Hanukoglu, I., Schild, L., Lu, Y., Shimkets, R.A., Nelson-Williams, C., Rossier, B.C., Lifton, R.P., 1996b. Mutations in the epithelial sodium channel causes salt wasting with hyperkalaemic acidosis, pseudohypoaldosteronism type 1. Nat. Genet. 12, 248±253. Clive, D.M., 1995. Bartter syndrome: the unsolved puzzle. Am. J. Kidney Dis. 25, 813±823. Colussi, G., Rombola, G., Brunati, C., De Ferrari, M.E., 1997. Abnormal reabsorption of Na 1/Cl 2 by the thiazide-inhibitable transporter of the distal convoluted tubule in Gitelman syndrome. Am. J. Nephrol. 17, 103±111. Davies, J.C., Geddes, D.M., Alton, E.W., 1998. Prospects for gene therapy for cystic ®brosis. Mol. Med. Today 4, 292±299. Derst, C., Konrad, M., Kockerling, A., Karolyi, L., Deschenes, G., Daut, J., Karschin, A., Seyberth, H.W., 1997. Mutations in the ROMK gene in antenatal Bartter syndrome are associated with impaired K 1 channel function. Biochem. Biophys. Res. Commun. 230, 641±645. Derst, C., Wischmeyer, E., Preisig-MuÈller, R., Spauschus, A., Konrad, M., Hensen, P., Jeck, N., Seyberth, H.W., Daut, J., Karschin, A., 1998. A hyperprostaglandin E syndrome mutation in Kir1.1 (renal outer medullary potassium) channels reveals a crucial residue for channel function in Kir1.3 channels. J. Biol. Chem. 273, 23884±23891. Deschenes, G., Burguet, A., Guyot, C., Hubert, P., Garabedian, M., Dechaux, M., Loirat, C., Broyer, M., 1993. Forme anteÂnatale du syndrome de Bartter. Ann. PeÂdiatr. (Paris) 40 (2), 95±101. Doupnik, C.A., Davidson, N., Lester, H.A., 1995. The inward recti®er potassium channel family. Curr. Op. Neurobiol. 5, 268±277. Doyle, D.A., Morais Cabral, J., Pfuetzner, R.A., Kuo, A., Gulbis, J.M., Cohen, S.L., Chait, B.T., MacKinnon, R., 1998. The structure of the potassium channel: molecular basis of K 1 conduction and selectivity. Science 280, 69±76. Duprat, F., Lesage, F., Guillemare, E., Fink, M., Hugnot, J.P., Bigay, J., Lazdunski, M., Romey, G., Barhanin, J., 1995. Heterologous multimeric assembly is essential for K 1 channel activity of neuronal and cardiac G-protein-activated inward recti®ers. Biochem. Biophys. Res. Commun. 212, 657±663. Ellison, D.H., Velazquez, H., Wright, F.S., 1987. Thiazide-sensitive sodium chloride cotransport in early distal tubule. Am. J. Physiol. 253, F546±F554. Fakler, B., Schultz, J.H., Yang, J., Schulte, U., Brandle, U., Zenner, H.P., Jan, L.Y., Ruppersberg, J.P., 1996. Identi®cation of a titratable lysine residues that determines sensitivity of kidney potassium channels (ROMK) to intracellular pH. EMBO J. 15, 4093±4099.

350

N. Jeck et al.

Fanconi, A., Schachenmann, G., NuÈssli, R., Prader, A., 1971. Chronic Hypokalaemia with growth retardation, normotensive hyperrenin-hyperaldosteronism (``Bartter syndrome''), and hypercalciuria. Helv. Paediatr. Acta, 144±163. Feldmann, D., Alessandri, J.L., Deschenes, G., 1998. Large deletion of the 5 0 end of the ROMK1 gene causes antenatal Bartter syndrome. J. Am. Soc. Nephrol. 9, 2357±2359. Fernandez-Llama, P., Ecelbarger, C.A., Ware, J.A., Andrews, P., Lee, A.J., Turner, R., Nielsen, S., Knepper, M.A., 1999. Cyclooxygenase inhibitors increase Na±K±2Cl cotransporter abundance in thick ascending limb of Henle's loop. Am. J. Physiol. 277, F219±F226. Fichman, M.P., Telfer, N., Zia, P., Speckart, P., Golub, M., Rude, R., 1976. Role of prostaglandins in the pathogenesis of Bartter syndrome. Am. J. Med. 60, 785±797. Gamba, G., 1999. Molecular biology of distal nephron sodium transport mechanisms. Kidney Int. 56, 1606±1622. Gamba, G., Saltzberg, S.N., Lombardi, M., Miyanoshita, A., Lytton, J., Hediger, M.A., Brenner, B.M., Hebert, S.C., 1993. Primary structure and functional expression of a cDNA encoding the thiazidesensitive, electroneutral sodium-chloride cotransporter. Proc. Natl. Acad. Sci. U.S.A. 90, 2749±2753. Gamba, G., Miyanoshita, A., Lombardi, M., Lytton, J., Lee, W.S., Hediger, M.A., Hebert, S.C., 1994. Molecular cloning, primary structure, and characterization of two members of the mammalian electroneutral sodium-(potassium)-chloride cotransporter family expressed in kidney. J. Biol. Chem. 269, 17713±17722. Giebisch, G., 1998. Renal potassium transport: mechanism and regulation. Am. J. Physiol. 274, F817± F833. Gill Jr., J.R., Frolich, J.C., Bowden, R.E., Taylor, A.A., Keiser, H.R., Seyberth, H.W., Oates, J.A., Bartter, F.C., 1976. Bartter syndrome: a disorder characterized by high urinary prostaglandins and a dependence of hyperreninemia on prostaglandin synthesis. Am. J. Med. 61, 43±51. Gitelman, H.J., Graham, J.B., Welt, L.G., 1966. A new familial disorder characterized by hypokalemia and hypomagnesemia. Trans. Assoc. Am. Phys. 79, 221±235. Glowatzki, E., Fakler, G., Brandle, U., Rexhausen, U., Zenner, H.P., Ruppersberg, J.P., Fakler, B., 1995. Subunit-dependent assembly of inward-recti®er K 1 channels. Proc. R. Soc. Lond. B. Biol. Sci. 261, 251±261. Greger, R., 1985. Ion transport mechanisms in thick ascending limb of Henle's loop of mammalian nephron. Physiol. Rev. 65, 760±797. Ho, K., Nichols, C.G., Lederer, W.J., Lytton, J., Vassilev, P.M., Kanazirska, M.V., Hebert, S.C., 1993. Cloning and expression of an inwardly rectifying ATP-regulated potassium channel. Nature 362, 31± 38. Howard, M., Frizzel, R.A., Bedwell, D.M., 1996. Aminoglycoside antibiotics restore CFTR function by overcoming premature stop mutations. Nature Med. 2, 467±469. Igarashi, P., van den Heuvel, G.B., Payne, J.A., Forbush III, B., 1995. Cloning, embryonic expression, and alternative splicing of a murine kidney-speci®c Na 1±K 1±Cl 2 cotransporter. Am. J. Physiol. 269, F405±F418. International Collaborative Study Group for Bartter-like Syndromes, KaÂrolyi, L., Konrad, M., Kockerling, A., Ziegler, A., Zimmermann, D.K., Roth, B., Wieg, C., Grzeschik, K.H., Koch, M.C., Seyberth, H.W., 1997. Mutations in the gene encoding the inwardly-rectifying renal potassium channel, ROMK, cause the antenatal variant of Bartter syndrome: evidence for genetic heterogeneity. Hum. Mol. Genet. 6, 17±26. Isenring, P., Jacoby, S.C., Chang, J., Forbush III, B., 1998a. Mutagenic mapping of the Na±K±Cl cotransporter for domains involved in ion transport and bumetanide binding. J. Gen. Physiol. 112, 549±558. Isenring, P., Jacoby, S.C., Forbush III, B., 1998b. The role of transmembrane domain 2 in cation transport by the Na±K±Cl cotransporter. Proc. Natl. Acad. Sci. U.S.A. 95, 7179±7184.

Hereditary hypokalemic salt-losing tubulopathies

351

Jentsch, T.J., Steinmeyer, K., Schwarz, G., 1990. Primary structure of Torpedo marmorata chloride channel isolated by expression cloning in Xenopus oocytes. Nature 348, 510±514. Jentsch, T.J., GuÈnther, W., Pusch, M., Schwappach, B., 1995. Properties of voltage-gated chloride channels of the ClC gene family. J. Physiol. 482, 19S±25S. Karet, F.E., Finberg, K.E., Nelson, R.D., Nayir, A., Mocan, H., Sanjad, S.A., Rodriguez-Soriano, J., Santos, F., Cremers, C.W., Di Pietro, A., Hoffbrand, B.I., Winiarski, J., Bakkaloglu, A., Ozen, S., Dusunsel, R., Goodyer, P., Hulton, S.A., Wu, D.K., Skvorak, A.B., Morton, C.C., Cunningham, M.J., Jha, V., Lifton, R.P., 1999. Mutations in the gene encoding B1 subunit of H 1-ATPase cause renal tubular acidosis with sensorineural deafness. Nat. Genet. 21 (1), 84±90. KaÂrolyi, L., Ziegler, A., Pollak, M., Fischbach, M., Grzeschik, K.H., Koch, M.C., Seyberth, H.W., 1996. Gitelman syndrome is genetically distinct from other forms of Bartter syndrome. Pediatr. Nephrol. 10, 551±554. KaÂrolyi, L., Koch, M.C., Grzeschik, K.H., Seyberth, H.W., 1998. The molecular approach to Bartter syndrome. J. Mol. Med. 76, 317±325. Kenna, S., Roper, J., Ho, K., Hebert, S., Ashcroft, S.J., Ashcroft, F.M., 1994. Differential expression of the inwardly rectifying K 1 channel ROMK1 in rat brain. Mol. Brain. Res. 24, 353±356. Kieferle, S., Fong, P., Bens, M., Vandewalle, A., Jentsch, T.J., 1994. Two highly homologous members of the ClC chloride family in both rat and human kidney. Proc. Natl. Acad. Sci. U.S.A. 91, 6943± 6947. Koch, M.C., Steinmeyer, K., Lorenz, C., Ricker, K., Wolf, F., Otto, M., Zoll, B., Lehmann-Horn, F., Grzeschik, K.H., Jentsch, T.J., 1992. The skeletal muscle chloride channel in dominant and recessive human myotonia. Science 257, 797±800. KoÈeckerling, A., Reinalter, S.C., Seyberth, H.W., 1996. Impaired response to furosemide in hyperprostaglandin E syndrome: evidence for a tubular defect in the loop of Henle. J. Pediatr. 129, 519±528. Konrad, M., Vollmer, M., Lemmink, H.H., van den Heuvel, L.P.W., Jeck, N., Vargas-Poussou, R., Lakings, A., Ruf, R., DescheÃnes, G., Antignac, C., Guay-Woodford, L., Knoers, N.V., Seyberth, H.W., Feldmann, D., Hildebrandt, F., 2000. Mutations in the chloride channel gene. CLCNKB, as a cause of classic Bartter syndrome. J. Am. Soc. Nephrol. in press. Konrad, M., Leonhardt, A., Hensen, P., Seyberth, H.W., Kockerling, A., 1999. Prenatal and postnatal management of Hyperprostaglandin E syndrome after genetic diagnosis from amniocytes. Pediatrics 103, 678±683. Kunchaparty, S., Palcso, M., Berkman, J., Velazquez, H., Desir, G.V., Bernstein, P., Reilly, R.F., Ellison, D.H., 1999. Defective processing and expression of thiazide-sensitive NaCl cotransporter as a cause of Gitelman syndrome. Am. J. Physiol. 277, F643±F649. Landau, D., Shalev, H., Ohaly, M., Carmi, R., 1995. Infantile variant of Bartter syndrome and senorineural deafness: a new autosomal recessive disorder. Am. J. Med. Genet. 59, 454±459. Lemmink, H.H., Knoers, N.V., Karolyi, L., van Dijk, H., Niaudet, P., Antignac, C., Guay-Woodford, L.M., Goodyer, P.R., Carel, J.C., Hermes, A., Seyberth, H.W., Monnens, L.A., van den Heuvel, L.P., 1998. Novel mutations in the thiazide-sensitive NaCl cotransporter gene in patients with Giteman syndrome with predominant localization to the C-terminal domain. Kidney Int. 54, 720±730. Lloyd, S.E., Pearce, S.H., Fisher, S.E., Steinmeyer, K., Schwappach, B., Scheinman, S.J., Harding, B., Bolino, A., Devoto, M., Goodyer, P., Rigden, S.P., Wrong, O., Jentsch, T.J., Craig, I.W., Thakker, R.V., 1996. A common molecular basis for three inherited kidney stone diseases. Nature 379, 445± 449. Ludewig, U., Pusch, M., Jentsch, T.J., 1996. Two physically distinct pores in the dimeric ClC-0 chloride channel. Nature 383, 340±343. Mastroianni, N., Bettinelli, A., Bianchetti, M., Colussi, G., De Fusco, M., Sereni, F., Ballabio, A., Casari, G., 1996. Novel molecular variants of the NaCl cotransporter gene are responsible for Giteman syndrome. Am. J. Hum. Genet. 59, 1019±1026.

352

N. Jeck et al.

Mastroianni, N., De Fusco, M., Zollo, M., Arrigo, G., Zuffardi, O., Bettinelli, A., Ballabio, A., Casari, G., 1996. Molecular cloning, expression pattern, and chromosomal localization of the human NaCl thiazide-sensitive cotransporter (SLC12A3). Genomics 35, 486±493. Matsumura, Y., Uchida, S., Kondo, Y., Miyazaki, H., Ko, S.B., Hayama, A., Morimoto, T., Liu, W., Arisawa, M., Sasaki, S., Marumo, F., 1999. Overt nephrogenic diabetes insipidus in mice lacking the ClC-K1 chloride channel. Nat. Genet. 21, 95±98. McCredie, D.A., Blair-West, J.R., Scoggins, B.A., Shipman, R., Chir, B., 1971. Potassium-losing nephropathy of childhood. Med. J. Aust. 1, 129±135. McCredie, D.A., Rotenberg, E., Williams, A.L., 1974. Hypercalciuria in potassium-losing nephropathy: a variant of Bartter syndrome. Aust. Paediatr. J. 10, 286±295. Meyburg, J., Mayatepek, E., Riester, U., Himbert, U., Zilow, E.P., Hilgenfeldt, U., Bremer, H.J., Linderkamp, O., 1999. Clinical symptoms, biochemical studies and therapeutic approaches in a sibship with a new congenital tubulopathy. Eur. J. Pediatr. 158, 673±678. Middleton, R.E., Pheasant, D.J., Miller, C., 1996. Homodimeric architecture of a ClC-type chloride ion channel. Nature 383, 337±343. Minor Jr., D.L., Masseling, S.J., Jan, Y.N., Jan, L.Y., 1999. Transmembrane structure of an inwardly rectifying potassium channel. Cell 96, 879±891. Mount, D.B., Hoover, R.S., Hebert, S.C., 1997. The molecular physiology of electroneutral cation-chloride cotransport. J. Membr. Biol. 158, 177±186. Murer, H., Greger, R., 1982. Membrane transport in the proximal tubule and thick ascending limb of Henle's loop: mechanisms and their alterations. Klin. Wochenschr. 60, 1103±1113. Muschaweck, R., Hadju, P., 1964. Die salidiuretische Wirksamkeit der chlor-N-(2-furylmethyl)-5-sulfamyl-anthranilsaÈure. Arzneimittelforschung 14, 44±47. Obermuller, N., Bernstein, P., Velazquez, H., Reilly, R., Moser, D., Ellison, D.H., Bachmann, S., 1995. Expression of the thiazide-sensitive NaCl-cotransporter in rat and human kidney. Am. J. Physiol. 269, F900±F910. Ohlsson, A., Sieck, U., Cumming, W., Akhtar, M., Serenius, F., 1984. Case report: a variant of Bartter syndrome. Acta Paediatr. Scand. 73, 868±874. Payne, J.A., Forbush III, B., 1994. Alternatively spliced isoforms of the putative renal Na 1±K 1±Cl 2 cotransporter are differentially distributed within the rabbit kidney. Proc. Natl. Acad. Sci. U.S.A. 91, 4544±4548. Payne, J.A., Forbush III, B., 1995. Molecular characterization of the epithelial Na±K±Cl cotransporter isoforms. Curr. Opin. Cell Biol. 7, 493±503. Puschett, J.B., Greenberg, A., Mitro, R., Piraino, B., Wallia, R., 1988. Variant of Bartter syndrome with a distal tubular rather than loop of Henle defect. Nephron 50, 205±211. Reinalter, S., Devlieger, H., Proesmans, W., 1998a. Neonatal Bartter syndrome: spontaneous resolution of all signs and symptoms. Pediatr. Nephrol. 12 (3), 186±188. Reinalter, S.C., Schmitt, J., Groene, H.J., Jeck, N., Soergel, M., Seyberth, H.W., Klaus, G., 1998b. Reversible renal impairment and mild histological alterations in patients with hereditary salt-wasting tubulopathies under long-term treatment with indomethacin. J. Am. Soc. Nephrol. 9, A501 Abstract. Riordan, J.R., 1999. Therapeutic strategies for treatment of CF based on knowledge of CFTR. Pediatr. Pulmonol. 18, 83±87. Rosenbaum, P., Hughes, M., 1957. Persistent, probably congenital, hypokalemic metabolic alkalosis with hyaline degeneration of renal tubules and normal urinary aldosteron. Am. J. Dis. Child. 94, 560. Rosenecker, J., Schmalix, W.A., Schindelhauer, D., Plank, C., Reinhardt, D., 1998. Towards gene therapy of cystic ®brosis. Eur. J. Med. Res. 3, 149±156. Rubinstein, R.C., Egan, M.E., Zeitlin, P.L., 1997. In vitro pharmacological restoration of CFTR-mediated chloride transport with sodium 4-phenylbutyrate in cystic ®brosis epithelial cells containing delta F508-CFTR. J. Clin. Invest. 100, 2457±2465.

Hereditary hypokalemic salt-losing tubulopathies

353

Rubinstein, R.C., Zeitlin, P.L., 1998. A pilot clinical trial of sodium 4-phenylbutyrate (Buphenyl) in delta F508-homozygous cystic ®brosis patients: evidence of restoration of nasal epithelial CFTR function. Am. J. Resp. Crit. Care Med. 157, 484±490. Rudin, A., 1988. Bartter syndrome: a review of 28 patients followed for 10 years. Acta Med. Scand. 224, 165±171. Rudin, A., SjoÈgren, B., Aurell, M., 1984. Low urinary calcium excretion in Bartter syndrome. N. Engl. J. Med. 310, 1190. Scheinman, S.J., Guay-Woodford, L.M., Thakker, R.V., Warnock, D.G., 1999. Genetic disorders of renal electrolyte transport. N. Engl. J. Med. 340, 1177±1187. Schmidt-Rose, T., Jentsch, T.J., 1997. Transmembrane topology of a ClC chloride channel. Proc. Natl. Acad. Sci. U.S.A. 94, 7633±7638. Schulte, U., Hahn, H., Konrad, M., Jeck, N., Derst, C., Wild, K., Weidemann, S., Ruppersberg, J.P., Fakler, B., Ludwig, J., 1999. pH gating of ROMK (K(ir)1.1) channels: control by an Arg-Lys-Arg triad disrupted in antenatal Bartter syndrome. Proc. Natl. Acad. Sci. U.S.A. 96, 15298±15303. Schultheis, P.J., Lorenz, J.N., Meneton, P., Nieman, M.L., Riddle, T.M., Flagella, M., Duffy, J.J., Doetschman, T., Miller, M.L., Shull, G.E., 1998. Phenotype resembling Gitelman syndrome in mice lacking the apical Na 1±Cl- cotransporter of the distal convoluted tubule. J. Biol. Chem. 273, 29150±29155. Schwalbe, R., Bianchi, L., Accili, E.A., Brown, A., 1998. Functional consequences of ROMK mutants linked to antenatal Bartter syndrome and implications for treatment. Hum. Mol. Genet. 7, 975±980. Schwartz, I.D., Alon, U.S., 1996. Bartter syndrome revisited. J. Nephrol. 9, 81±87. Seidel, C., Reinalter, S., Seyberth, H.W., SchaÈrer, K., 1995. Prepubertal growth in the hyperprostaglandin E syndrome. Pediatr. Nephrol. 9, 723±728. Seyberth, H.W., Rascher, W., Schweer, H., Kuhl, P.G., Mehls, O., Scharer, K., 1985. Congenital hypokalemia with hypercalciuria in pre-term infants: A hyperprostaglandinuric tubular syndrome different from Bartter syndrome. J. Pediatr. 107, 694±701. Seyberth, H.W., Koniger, S.J., Rascher, W., Kuhl, P.G., Schweer, H., 1987. Role of postaglandins in hyperprostaglandin E syndrome and in selected tubular disorders. Pediatr. Nephrol. 1, 491±497. Seyberth, H.W., Soergel, M., Kockerling, A., 1998. HypoKalemic tubular disorders: The Hyperprostaglandin E syndrome and the Bartter-Gitelman syndrome. In: Davison, A.M., Cameron, J.S., GruÈnfeld, J.P. (Eds.). Oxford Textbook of Clinical Nephrology, Oxford Medical Publications, Oxford, pp. 1085± 1094. Shoemaker, L., Welch, T.R., Bergstrom, W., Abrams, S.A., Yergey, A.L., Vieira, N., 1993. Calcium kinetics in the hyperprostaglandin E syndrome. Pediatr. Res. 33, 92±96. Shuck, M.E., Bock, J.H., Benjamin, C.W., Tsai, T.D., Lee, K.S., Slightom, J.L., Bienkowski, M.J., 1994. Cloning and characterization of multiple forms of the human kidney ROMK potassium channel. J. Biol. Chem. 269, 24261±24270. Simon, D.B., Lifton, R.P., 1996a. The molecular basis of inherited hypokalemic alkalosis: Bartter and Gitelman syndromes. Am. J. Physiol. 40, F961±F966. Simon, D.B., Karet, F.E., Hamdan, J.M., DiPietro, A., Sanjad, S.A., Lifton, R.P., 1996b. Bartter syndrome, hypokalemic alkalosis with hypercalciuria, is caused by mutations in the Na±K±2Cl cotransporter NKCC2. Nat. Genet. 13, 183±188. Simon, D.B., Karet, F.E., Rodriguez-Soriano, J., Hamdan, J.H., DiPietro, A., Trachtman, H., Sanjad, S.A., Lifton, R.P., 1996c. Genetic heterogeneity of Bartter syndrome revealed by mutations in the K 1 channel, ROMK. Nat. Genet. 14, 152±156. Simon, D.B., Nelson-Williams, C., Bia, M.J., Ellison, D., Karet, F.E., Molina, A.M., Vaara, I., Iwata, F., Cushner, H.M., Koolen, M., Gainza, F.J., Gitelman, H.J., Lifton, R.P., 1996. Gitelman variant of Bartter syndrome, inherited hypokalemic alkalosis, is caused by mutations in the thiazide-sensitive Na±Cl cotransporter. Nat. Genet. 12, 24±30.

354

N. Jeck et al.

Simon, D.B., Bindra, R.S., Mans®eld, T.A., Nelson-Williams, C., Mendonca, E., Stone, R., Schurman, S., Nayir, A., Alpay, H., Bakkaloglu, A., Rodriguez-Soriano, J., Morales, J.M., Sanjad, S.A., Taylor, C.M., Pilz, D., Brem, A., Trachtman, H., Griswold, W., Richard, G.A., John, E., Lifton, R.P., 1997. Mutations in the chloride channel gene, CLCNKB, cause Bartter syndrome type III. Nat. Genet 17, 171±178. Simon, D.B., Crut, D.N., Lu, Y., Lifton, R.P., 1998. Genotype±phenotype correlation of the NCCT and Giteman syndrome. J. Am. Soc. Nephrol. 9, A0573 Abstract. Steinmeyer, K.C., Pusch, L.M., Koch, M.C., Jentsch, T.J., 1994. Multimeric structure of ClC-1 chloride channel revealed by mutations in dominant myotonia congenita (Thomsen). EMBO J. 13, 737±743. Stokes, J.B., 1989. Electroneutral NaCl transport in the distal tubule. Kidney Int. 36, 427±433. Sutton, R.A., Mavichak, V., Wilkins, G.E., 1988. Bartter syndrome with hypocalciuria ± a distal tubule defect?. Clin. Res. 33, 85A. Sutton, R.A., Mavichak, V., Halabe, A., Wilkins, G.E., 1992. Bartter syndrome: evidence suggesting a distal tubular defect in a hypocalciuric variant of the syndrome. Miner. Electrolyte Metab. 18, 43± 51. Takahashi, M., Yanagida, N., Okano, M., Ishizaki, A., Meguro, J., Kukita, K., Tamaki, T., Yonekawa, M., Kawamura, A., Yokoyama, T., 1996. A ®rst report: living related kidney transplantation on a patient with Bartter syndrome. Transplant. Proc. 28, 1588. Takahashi, N., Igarashi, P., Smithies, O., 1999. A mouse model of neonatal Bartter syndrome. J. Am. Soc. Nephrol. 10, A230 Abstract. Takeuchi, Y., Uchida, S., Marumo, F., Sasaki, S., 1995. Cloning, tissue distribution, and intrarenal localization of ClC chloride channels in human kidney. Kidney Int. 48, 1497±1503. Tamarappoo, B.K., Verkman, A.S., 1998. Defective aquaporin 2 traf®cking in nephrogenic diabetes insipidus and correction by chemical chaperones. J. Clin. Invest. 101, 2257±2267. Trygstad, C.W., Mangos, J.A., Bloodworth, J.M.B., Lobeck, C.C., 1969. A sibship with Bartter syndrome: failure of total adrenalectomy to correct the potassium wasting. Pediatrics 44, 234. Vandewalle, A., Cluzeaud, F., Bens, M., Kieferle, S., Steinmeyer, K., Jentsch, T.J., 1997. Localization and induction by dehydration of ClC±K chloride channels in the rat kidney. Am. J. Physiol. 272, F678± F688. Vargas-Poussou, R., Feldmann, D., Vollmer, M., Konrad, M., Kelly, L., van den Heuvel, L.P., Tebourbi, L., Brandis, M., Karolyi, L., Hebert, S.C., Lemmink, H.H., Deschenes, G., Hildebrandt, F., Seyberth, H.W., Guay-Woodford, L.M., Knoers, N.V., Antignac, C., 1998. Novel molecular variants of the Na± K±2Cl cotransporter gene are responsible for antenatal Bartter syndrome. Am. J. Hum. Genet. 62, 1332±1340. Verberckmoes, R., van Damme, B.B., Clement, J., Amery, A., Michielsen, P., 1976. Bartter syndrome with hyperplasia of renomedullary cells: successful treatment with indomthacin. Kidney Int. 9, 302± 307. Vollmer, M., Koehrer, M., Topaloglu, R., Strahm, B., Omran, H., Hildebrandt, F., 1998. Two novel mutations of the gene for Kir 1.1 (ROMK) in neonatal Bartter syndrome. Pediatr. Nephrol. 12, 69±71. Waldegger, S., Jentsch, T.J., 2000. Functional characterization of the Bartter syndrome chloride channel ClC-Kb. Eur. J. Physiol. 438, R316. Yang, T., Huang, Y.G., Singh, I., Schnermann, J., Briggs, J.P., 1996. Localization of bumetanide- and thiazide-sensitive Na±K±Cl cotransporters along the rat nephron. Am. J. Physiol. 271, F931±F939.

355

Table of known channelopathies The table lists the known channelopathies in man and various other species. The names of the genes and their chromosomal location are given in columns 1 and 2 (human genes in capitals); the names and synonyms of the corresponding proteins are listed in column 3 for transporters and voltage-gated (vg), voltage-sensitive (vs) and voltage-insensitive (vi) channels; vs and vi channels are activated by endagenous ligands. Columns 4 and 5 list the diseases and their transmission. The effects of the mutations on the function of the channel complex are categorized in column 6 as gain- or loss-of-fuction. In general, gain-of-function mutations exert dominant effects if they dominate cell function whereas loss-of-function mutations only cause dominant inheritance if the channel complex is multimeric (dominant negative effect) or the second allele haploinsuf®cient. The last column informs about the book chapter dealing with this topic; skm ˆ skeletal muscle.

Gene

Locus

Sodium channel SCN1A 2q24

SCN1B

19q13.1

SCN4A

17q23.1±25.3

SCN5A

3p21

Channel protein Disease

Inheritance

Change

vg neuronal a subunit

Dominant

Gain

7

Dominant

Gain

7

Dominant

Gain

1

Dominant

Loss

1

Dominant

Gain

5

Generalized epilepsy with febrile seizures plus (GEFS1 2) Generalized epilepsy vg b 1 subunit with febrile seizures plus (GEFS1 1) Skm1, vg skm a Hyperkalemic subunit periodic paralysis (also in horses), paramyotonia, potassium aggravated myotonia Hypokalemic periodic paralysis 2 vg cardiac Brugada syndrome; a subunit long QT syndrome 3

Ch

356

Table of known channelopathies

Gene

Locus

SCN8A

12q13

SCNN1B

SCNN1G

Channel protein Disease

vg neuronal a subunit 16p12.2±p12.1 ENaC, vi epithelial b subunit

16p12.2±p12.1 ENaC, vi epithelial g subunit

Inheritance

Change

Mice: MED, jolting

Recessive

Loss

±

Liddle's syndrome

Dominant

Gain

11

Pseudohypoaldosteronism I (PHA1)

Dominant

Gain

11

Recessive Dominant

Loss Gain

11

Dominant

Loss

11 11 ±

Liddle's syndrome

Pseudohypoaldosteronism I (PHA1) mec-4, mec10, deg-1 Calcium channel CACNA1A 19p13.1

CACNA1F

Xp11.23

CACNA1S

1q31±32

CACNB4

CACNG2

2q22±23

Ch

vi ENaC homolog

C. elegans: mec-4, mec-10, deg-1

Recessive Dominant and recessive

Loss Unclear

vg neuronal P/Q-type a1 subunit

Episodic ataxia 2

Dominant

Loss

6

Familial hemiplegic migraine, spinocerebellar ataxia 6 Mice: tottering, leaner, rolling Congenital stationary night blindness (CSNB2) Hypokalemic periodic paralysis 1, malignant hyperthermia 5

Dominant

Unclear

6

Recessive

Loss

6

Recessive

Loss

±

Dominant

Unclear

Mice: muscular dysgenesis vg neuronal b 4 Generalized epilepsy, subunit episodic ataxia 3 Mice: lethargic vg neuronal g 2 Mice: stargazer, subunit, waggler stargazin

Recessive

Loss

±

Dominant

Gain

±

Recessive Recessive

Loss Loss

± ±

vg retinal L-type a1 subunit vg skm L-type a1 subunit, dihydropyridine (DHP) receptor

1, 3

Table of known channelopathies

357

Gene

Locus

Channel protein Disease

Inheritance

Change

RYR1

19q13.1

vi skm ryanodine receptor 1, calcium release channel vg neuronal a1 subunit

Malignant hyperthermia 1 (also in pigs), central core disease

Dominant

Gain

3

Drosophila: nightblindA, lethal(1)L13, cacophony Knockout mice: dyspedic Polycystic kidney disease

Recessive

Loss

±

vi ryanodine receptor Polycystin-1, similar to polycystin-2 and vg calcium channels vi IP3 homolog Mice: opisthotonus

Recessive

Loss

±

Dominant

Unclear

±

Recessive

Unclear

±

Vg neuronal a1 Episodic ataxia 1 subunit, A-type, Kv1.1 Drosophila: Shaker Knock out mice: epilepsy Vg b subunit, Knock out mice: impaired learning Kvb 1.1 Vg b subunit of Long QT syndrome 5 KVLQT1, MinK, ISK Jervell and LangeNielsen Vi b subunit of Long QT syndrome HERG, MiRP1 inducible Vg cardiac/ Long QT syndrome 1 neuronal a subunit, KCNQ1, KVLQT1 Jervell and LangeNielsen

Dominant

Loss

6

Loss

± ±

DmCa1D

Ryr1 PKD1

16p13.3

Ip3r1 Potassium channel KCNA1 12p13

KCNA1B

3q26.1

KCNE1

21q22.1±22.2

KCNE2

21q22.1

KCNQ1

11p15.5

Ch

± Dominant

Loss

Recessive

Loss

5

5 Dominant

Loss

Recessive

Loss

5

358

Table of known channelopathies

Gene

Locus

Channel protein Disease

Inheritance

Change

KCNQ2

20q13.3

Dominant

Loss

7

KCNQ3

8q24.22±24.3

Dominant

Loss

7

KCNQ4

1p34

Dominant

Loss

±

HERG

7q35±36

Long QT syndrome 2 Dominant

Loss

5

KCNJ1

11q24

Vg neuronal a subunit, KCNQ2 Vg neuronal a subunit, KCNQ3 Vg neuronal a subunit, KCNQ4 Vs cardiac a subunit, HERG, ether aÁ go go (eag) related, Ikr vi a subunit, ROMK1, Kirl.1

Dominant

Loss

13

KCNJ6

21q22.1

Antenatal variant of Bartter ˆ hyperprostaglandin E syndrome mice: weaver

Recessive

Gain

±

KCNJ11

11p15.1

Hyperinsulinemic Recessive hypoglycemia (PHHI)

Loss

12

SUR1

11p15.1

Hyperinsulinemic Recessive hypoglycemia (PHHI)

Loss

12

Recessive

Loss

±

Recessive Recessive

Loss Loss

± ±

Thomsen myotonia Dominant (also in goats) Becker myotonia (also Recessive in mice)

Loss

2

Loss

2

vi a subunit, GIRK2, Kir3.2 vi KATP a subunit of pancreatic islet, Kir6.2 vi KATP b subunit of pancreatic islet, sulfonylurea receptor, SUR1

Benign familial neonatal convulsions 1 (BFNC1) Benign familial neonatal convulsions 2 (BFNC2) Dominant deafness

Less selective cation channels CNGA1 4p12-ce Vi retinal Retinitis pigmentosa cGMP-gated a subunit CNGA3 2q11 alpha-3 subunit Achromatopsia-2 CNGB3 8q21-22 beta-3 subunit Achromatopsia-3 Chloride channel CLCN1 7q32-qter

Vs skm dimer, ClC1

Ch

Table of known channelopathies

359

Gene

Locus

Channel protein Disease

Inheritance

Change

CLCN5

Xp11.22

Vs kidney epithelial dimer, ClC5

Recessive

Loss

10

CLCNKB

1p36

Dominant

Loss

13

CFTR

7q31.2

Vs kidney epithelial dimer, ClCKb, ClCK2 CF transCystic ®brosis (CF) membrane conductance regulator, ATPbinding cassette (sub-family C, member 7)

Recessive

Loss

9

Neuronal a1 subunit

Dom/recess

Loss

8

Recessive

Loss

Recessive

Loss

8

Dom/recess

Gain, loss Loss

4

Gain, loss Gain, loss Gain, loss Gain

4

Glycine receptor GLRA1 5q31.2

GLRB

4q31.3

Neuronal b 1 subunit

Dent's disease ˆ X-linked nephrolithiasis or hypophosphatemic rickets Classical Bartter

Hyperekplexia ˆ startle disease ˆ stiff baby syndrome (STHE) Mice: spasmodic, oscillator Mice: spastic

Nicotinic acetylcholine receptor CHRNA1 2q24±32 Skm a1 subunit Congenital myasthenic syndrome CHRNA4 20q13.3 Neuronal a4 Nocturnal frontal subunit lobe, ADNFLE CHRNB1 17p12-11 Skm b 1 subunit Congenital myasthenic syndrome CHRND 2q33±34 Skm d subunit Congenital myasthenic syndrome CHRNE 17 Skm 1 1 subunit Congenital myasthenic syndrome deg-3 vi AChR a C. elegans: deg-3 homolog

Dominant Dom/recess Dom/recess Dom/recess Dominant

Ch

7

4 4 ±

360

Gene

Table of known channelopathies

Locus

Glutamate receptor GRIND1 and 2

GABA receptor Gabara5, Gabarb3, Gabarg3 Transporter SLC12A1 15q15-q21.1

SLC12A3

Slc12a2 Slc9a1

16q13

Channel protein Disease

Inheritance

Change

vi glutamate receptor d1 and d2 subunits

Semidominant

Unclear

±

Mice: lurcher

Ch

vi GABA receptor subunits

Natural knockout mice: pink-eyed cleftpalate

Recessive

Loss

±

Na/K/2Cl transporter, NKCC2

Antenatal Bartter ˆ hyperprostaglandin E syndrome, knock-out mice lethal Gitelman

Dominant

Not measured

13

Dominant

Loss

13

Recessive

Loss

±

Recessive

Loss

±

Na/Cl transporter, NCCT Na/K/2Cl coMice: shaker/waltzer transporter Na/H exchanger Mice: slow wave NHE epilepsy

Frank Lehmann-Horn and Karin Jurkat-Rott

361

Subject Index 4-aminopyridine (4-AP) 142 9-anthracen carboxylic acid (9-AC) 43, 265 absence 175, 268 acetazolamide 13, 17, 23, 30, 157±162, 172±175 acetylcholine receptor, nicotinic (nAChR) motor endplate 85±114, 359 neuronal 181±187, 359 acetylcholinesterase (AChE) 85, 92, 110± 114 action potential 4±18, 38±41, 89±109, 115± 142, 172, 188, 190, 262 ADNFLE, see autosomal dominant nocturnal frontal lobe epilepsy ADR, see myotonic mice albuminuria 258 aldosterone 23, 229, 277±298, 327±328, 340±344 alkalosis 327±330, 340±353 amiloride 229±250, 259, 278±284, 329, 344 aminoaciduria 257±258 amiodarone 121, 145, 148 anesthetic drug general 55, 57, 66, 139 local 22, 49 angiotensin sensitivity 328, 349 animal models 355±360 (overview) anthopleurin-A 137, 151 antiarrhythmic drug 9, 33, 49, 118, 137±140 apical membrane 235, 271, 279±296, 331± 344 arrhythmia, cardiac 22, 73, 115±145 Ashkenazi Jews 305, 318, 322, 323 ataxia 356±357 (overview) episodic type 1 (EA-1) 172±174, 357 episodic type 2 (EA-2) 155±172, 356 progressive cerebellar 155, 159, 161, 169, 175 spinocerebellar 155, 158, 176, 177, 178 ATP-sensitive potassium channels, see potassium channels

ATP/ADP ratio 300, 306 ATPase 71, 260, 279, 344 aura 155±160, 183 autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) 181±187, 359 Bartter's syndrome 235, 264, 327±354, 358±360 antenatal variant 327±354, 358, 360 classic form 327±330, 342, 359 Gitelman's syndrome 327±329, 349±353, 360 hyperprostaglandin E syndrome see antenatal Becker's myotonia 37±51, 358 benign familial neonatal convulsions (BFNC) 182, 193±196, 357±358 blepharospasm 36 blocker 52, 109, 232, 281, 322 Brugada syndrome 115, 142±145, 355 4-chloro-m-cresol (4-CmC) 58, 70 CACNA1A, and hemiplegic migraine or episodic ataxia 155, 161±179, 356 CACNA1S, and hypokalemic periodic paralysis or malignant hyperthermia 20±24, 64, 78, 356 caffeine 57±58, 68±84 CAG repeat expansion, and episodic ataxia 155, 162±177 CTG repeat expansion, and myotonic dystrophy 34 calcium channel 356±357 (overview) calmodulin 68, 244 calpains 76 calsequestrin 68 candidate gene 64, 65, 81, 259, 278, 282, 287, 288 approach 10, 93, 335

362

Subject Index

cell volume 252 central core disease, CCD 55, 59, 81, 84 cerebellum 164, 171, 174, 194, 217 cerebrospinal ¯uid 159 channel blocker 93, 121, 142, 279, 303, 312, 322 chaperones 241, 348, 354 chloride channel gene 40, 253, 274, 276, 339, 351, 354 chloride channel, ClC 33±53, 262, 274, 334 (overview) CHRNA1/B1/D/E, and myasthenic syndrome 85±114, 359 (overview) CHRNA4, and nocturnal frontal lobe epilepsy 181±187, 359 (overview) CLCN1, and myotonia congenita 33, 40± 45, 49, 334, 358 CLCN5, and Dent's disease 42, 256±275, 334, 358 CLCNKB, and classic Bartter 42, 235, 253, 276, 334±340, 351, 354, 359 clinical variability 162, 169 clo®lium 124 ColQ 85, 110, 112, 114 cold sensitivity 3, 13, 25 coma 156, 175, 177 congenital myasthenic syndrome 85±114, 359 creatine kinase 56, 57, 73 C-terminus 68, 72, 93, 98, 126, 127, 128, 129, 133, 143, 144, 190, 243, 267, 282, 284, 287, 332, 337, 347 cyclooxygenase 350 cystic ®brosis 227±253, 359 dantrolene 57, 69, 81 deactivation 14, 15, 30, 129, 131, 173, 265 deafness 118, 179, 256, 330, 349, 351, 358 degradation 76, 98, 99, 237, 251, 255, 260, 271, 284, 285 dehydration 290, 327, 329, 346, 354 denervation 121, 140, 150

Dent's disease 255±276, 334, 358 dephenylamine-2-carboxylic acid 265 DIDS 68, 244, 265, 269 dihydropyridine receptor (DHPR) 20±31, 64±68, 81±84, 116, 147, 164 disopyramide 141 distal convolute 327, 332, 333, 340, 344, 349, 353 diuretic 17, 259, 278, 279, 327, 342 dofetilide 124, 130, 137 dysarthria 157, 158 dyspedic mice 68, 357 dystrophic myotonia, see myotonic dystrophy early afterdepolarization 119, 147, 153 electroencephalography 159, 195 electromyography 31, 89, 90, 91, 92 endosome 270, 271 endplate 87±95, 101±114, 203 endplate myopathy 92, 108±111 epilepsy, see also ADNFLE and GEFS+ 13, 154, 169±172, 181±198, 256, 300±302, 344, 357±360 episodic 4, 21, 33, 118, 155, 157, 169, 172± 179, 192, 195, 211 estrogen 140 etiology 55, 72, 300, 302, 304, 306, 308, 319 European malignant hyperthermia protocol 58, 64, 78 excitation±contraction coupling 83, 84 excitotoxic 108, 109, 111 exercise 3, 12±17, 22, 28, 29, 60, 76, 85, 91, 92, 150, 157, 173, 174 expression cloning 40, 51, 274, 351 external intercostal muscles 44 familial hemiplegic migraine 4, 155, 161, 175±177, 356 familial hyperinsulinism 299, 322±325 Fanconi's syndrome 257 fast channel syndrome 85, 95, 106, 110 fast inactivation 6, 9, 14, 18±20, 26, 27, 29, 30, 31, 165

Subject Index

febrile convulsions, febrile seizures 182, 184, 194, 197 fever 156, 157, 159, 241, 330, 343 Finland 304, 305, 324 FK binding protein 12 (FKBP12) 68 Focal adenomatous hyperplasia 322, 325 founder effect 27, 28, 61, 322 functional studies 52, 70, 72, 74, 113, 143, 147, 194, 219, 286, 296, 350 furosemide-like salt losing tubulopathy 327, 329, 338 furosemide-sensitive 331, 332, 344 gain-of-function mutations 355±360 (overview) gating modes 43, 134 GEFS+, see generalized epilepsy with febrile seizures gene therapy 236, 239 generalized epilepsy with febrile seizures (GEFS+) 193, 194, 355 genetic background 78, 181, 193, 217, 223 genetic heterogeneity 24, 58, 77, 152, 187, 193, 196, 282, 294, 306, 335, 350 genotype 27, 55, 80, 82, 101, 119, 143, 153, 171, 175, 235, 242, 249, 339 genotype±phenotype 161, 168 gephyrin 201, 204, 216, 217, 219 Gitelman's syndrome, see Bartter's syndrome GLRA1, and hyperekplexia 199, 200, 211, 213, 223, 359 Glra1, and spasmodic and oscillator mice 214, 215, 218, 220, 359 Glrb, and spastic mice 215, 359 glucokinase 306, 308 glucosuria 257, 258, 261 glycine receptor 199±224, 359 glycine receptor sensitivity ethanol 210 glycine 213 neurotransmitter 214 picrotoxinin 201, 213 strychnin 206

363

GPI locus, porcine 61 guidelines 67, 79 halothane 55±84 haploinsuf®ciency 122, 123 head retraction jerk 214 headache 156, 176 hematuria 257 HERG potassium channel 118, 119, 129± 153, 194, 357, 358 heteromeric receptor 189, 190 high-risk haplotype 67 homeostasis 55, 69, 73±80, 318, 320, 345, 347 homology 130, 188, 200, 202, 219, 284, 332, 335 human embryonic kidney cells 11, 14, 125, 201 human kidney 51, 255, 261, 274, 333±354 hyperactivity 183, 282, 285, 303 hyperammonemia 306, 324, 325 hypercalciuria 255±263, 273, 276, 328, 329, 340, 350, 353 hyperekplexia 211, 212, 216, 219, 223, 359 hyperexcitability 4, 14, 18, 27, 33, 35, 51, 173, 192 hyperinsulinemic hypoglycemia (PHHI) 299±326, 358 hyperinsulinism, familial, see familial hyperinsulinism hyperkalemic periodic paralysis (HyperPP) 4, 5, 27±33, 192, 195, 355 hyperprostaglandin E syndrome see Bartter's syndrome, antenatal variant hypertension 223, 277±279, 287, 292±296, 328 hypoglycemia, see hyperinsulinemic hypoglycemia hypokalemic salt-losing tubulopathies, familial 327±354, 359±360 hypokalemic periodic paralysis (HypoPP) 3, 15, 20±31, 64, 83, 355, 356 hypomagnesemia 328, 341, 347, 350 hyponatraemia 290

364

Subject Index

hypophosphatemic rickets, see rickets hypovolaemia 277

KCNQ3, and benign familial neonatal convulsions (BFNC) 181±196, 358

idiopathic hypercalciuria 259, 261 idiopathic ventricular ®brillation (IVF), see also Brugada syndrome 141±143 Ikr potassium current 116, 118, 129±151 IKs potassium current 116, 118, 124±151 inactivation gate 6, 10, 14, 30, 42, 135, 265 incoordination 157, 158 indomethacin 231, 233, 328, 343, 344, 352 insulin 299±322 intracellular free calcium 61 inward recti®er potassium channel (Kir) 121, 150, 152, 307±325, 349, 358 in-vitro contracture test (IVCT) 57, 58, 59, 64, 66, 67, 71, 72, 73, 77, 78, 79, 82 sensitivity and speci®city 58 ion selectivity 217, 264, 280 ionic control 323 isosthenuria 340 IVF, see idiopathic ventricular ®brillation Jervell±Lange±Nielsen syndrome (JLN) 118, 126, 137, 357 junctional folds 92, 107, 108 juvenile myoclonic epilepsy 182, 194, 197 juxtaglomerular apparatus 328

Lafora disease 193 leaner mice 155, 170, 171, 172, 176, 356 leanness 60 Liddle's syndrome 277±298, 356 ligand-gated ion channel 85, 200 ligands 68, 206, 219, 221, 223, 320, 355 linkage 10, 24, 26, 40, 55, 61, 62, 64, 66, 67, 77, 78, 80, 82, 123, 130, 134, 140, 142, 143, 148, 161, 182, 184, 187, 190, 195, 196, 197, 211, 259, 275, 281, 297, 306, 307, 335, 336 liposomes 239 loop of Henle 235, 333, 336, 343, 344, 351, 352 loops 68, 82, 163, 199, 207, 212, 286 loss-of-function mutations 355±360 (overview) low-molecular weight proteins 260 long QT (LQT) syndrome 115±154, 188, 355, 357, 358 lysosomes 260 L-type calcium channel, see dihydropyridine receptor

kaliuresis 257 K channels, see potassium channels KATP channels, see potassium channels K channel genes 357±358 (overview) KCNA1, and episodic ataxia 172±173, 357 KCNE1, and long QT syndrome 118, 119, 124±129, 138±151, 357 KCNE2, and inducible long QT syndrome 141, 357 KCNJ1, and antenatal variant of Bartter's syndrome 327, 333±337, 346, 358 KCNJ11, and hyperinsulinemic hypoglycemia (PHHI) 229±325, 358 KCNQ1, and long QT syndrome 118, 119, 123, 143, 188, 357 KCNQ2, and benign familial neonatal convulsions (BFNC) 181±196, 357

male prevalence 142 malignant hyperthermia (MH) 55±83, 356, 357 malignant hyperthermia susceptibility (MHS) loci 64±67, 77, 78 malignant neuroleptic syndrome (MNS) 59 malignant ventricular arrhythmia 115 M-current 189, 190 meconium ileus 228, 229, 234 membrane depolarization 10±26, 162, 163, 192, 218, 300 meningismus 156 mental retardation 157, 158 MEPP, miniature endplate potentials 87, 90, 92 MgADP 301, 314, 315, 317 migraine 4, 155±179, 192, 195, 356

Subject Index

mineralocorticoid 293 minK potassium channel, and long QT syndrome 118, 124±129, 138±152, 189, 193, 357 minor head trauma 156, 161 modulator 68, 239 motor control 199 motor de®cit 155 mouse models 355±369 (overview) MRI 159, 177 MTAL 261 mucociliary clearance 237, 239, 247 myasthenia gravis 85, 113 myasthenic syndromes 85±114, 359 myokymia 51, 160, 173±178, 192, 194 myopathy 15, 16, 17, 22, 26, 34, 59, 92, 108, 109, 111 myotonia 3±53, 82, 145, 264, 274, 276, 334, 351, 354, 355, 358 Becker-type, see Becker's myotonia dominant, see Thomsen's myotonia electrical 15, 45 ¯uctuans 13, 14 latent 35 levior 44, 47, 51 paradoxical 3, 7, 12, 16, 19, 33, 35 percussion 35, 36 permanens 13, 14 potassium-aggravated 3±5 recessive, see Becker's myotonia Thomsen, see Thomsen's myotonia myotonic dystrophy 33±35 goat 35, 39, 40, 41, 45, 46, 50, 358 mice (ADR, MTO) 35, 46, 48, 358 Na channels, see sodium channels nasal transepithelial voltage 229, 242 N-box 88, 95, 97 nephrocalcinosis 257±276, 328, 339±341 nephrolithiasis 23, 160, 255±263, 274±276, 334, 358 nephron 252±290, 333, 335, 340, 344±347, 350, 354

365

nerve impulse 89 nerve terminal 85, 89, 90, 92, 110 neuromuscular transmission 85, 90, 93, 104, 106±109 neuromyotonia 52 neuron 27±52, 113, 145, 152, 174, 176, 218±223, 274, 293, 322±325 neuronal 155, 159, 164, 165, 171±208, 223, 274, 349 neuronal nicotinic acetylcholine receptor, see acetylcholine receptor 185 neurotransmitter receptor 200, 212, 220 neurotransmitter release 164 ni¯umic acid 265 non-insulin dependent diabetes 319 normokalemic periodic paralysis 16 North American malignant hyperthermia protocol 58, 78 NPPB 265, 269 nucleotide binding domain 243 nystagmus 157±159, 162, 175, 179 open probability of channels 47, 170, 265, 284, 285, 345 oscillator mice, see glra1 osteomalacia 257, 259, 272 P/Q-type calcium channel, and migraine or episodic ataxia 161, 164, 170, 177, 178, 356 pancreatectomy 299±305, 316, 318, 322, 324 pancreatic insuf®ciency 228, 232, 242, 250 pancreatic beta cell 321, 323 paralysis 3±5, 12±33, 64, 81, 83, 138, 145, 192, 195, 355, 356 paramyotonia congenita (PC) 3±5, 27±35, 145, 355 paresthesias 155 patch-clamp 209, 265, 269, 315, 317 penetrance 7, 15±17, 21, 55, 59, 66, 78, 161±165, 176, 184, 187 permeation 24, 43, 213, 218, 267 persistent current 10±19, 134, 138 pH 18, 29, 40, 50, 56, 73, 255, 260, 266, 271±275, 337, 345±349, 353

366

Subject Index

PHA1, see pseudohypoaldosteronism type I phenotype 126±128, 133, 135, 166, 212, 353 PHHI, see hyperinsulinemic hypoglycemia phosphaturia 257, 261 photometry 69±71 polyhydramnios 327, 329, 339, 341 polymorphism 49, 282, 316, 323 polyuria 257, 258, 327±329, 341, 343 porcine stress syndrome (PSS), see ryanodine receptor postsynaptic 87±95, 107±113, 187, 203, 216, 221 postsynaptic membrane 89, 90, 95, 108, 111 potassium channels 357, 358 (overview) current 25, 145±151, 189, 193, 345, 346 gene 146±155, 174±197, 357, 358 secretion 278 sensitivity 3, 13, 25 prenatal diagnosis 240 presynaptic 92, 110, 164, 187, 197 proband 66, 67, 79, 307, 317 progressive renal failure 257, 276 proline-rich motif 280, 282, 284 promoter mutations 113 protein kinase A 273, 275 proteinuria 255, 256, 257, 258, 260, 273, 275, 276 proton pump 260, 262, 270, 271, 274 proximal myotonic myopathy (PROMM) 34 proximal tubule 42, 255, 260, 261, 262, 269, 270, 271, 272, 273, 275, 335, 352 pseudohypoaldosteronism type I (PHA1) 277±298, 356 Purkinje cells 159, 164, 165, 170, 171, 176 QT interval, of ECG 131, 139, 140, 144, 147, 148 QT syndrome, see long QT syndrome quantal release 110

quinidine 109±113, 121, 141, 145, 149±151 recovery from inactivation 11, 12, 19, 131, 142, 169, 170, 192 recurrence 160, 181, 303 regulatory domain 243, 245 renal failure 23, 256±259, 263, 274, 276, 343 respiratory tract infections 236 resting membrane potential 10 rickets, hypophosphatemic 257, 258, 259, 273, 348, 358 right bundle branch block 142 risk 13, 22, 58, 60, 67, 79, 115, 143, 148, 188, 292, 293, 304 Rolando±Ward syndrome, see long QT syndromes ROMK1 potassium channel, and antenatal variant of Bartter's syndrome 246, 279, 327±354, 358 ryanodine receptor, human skeletal muscle (RyR1), and MH and CCD 20, 21, 55± 84, 357 ryanodine receptor, porcine skeletal muscle (ryr1), and MH or PSS 55±62, 70, 357 S-100 protein 68 S4 segment, see voltage sensor salt homeostasis 277, 292 salt reabsorption 344 sarcolemma 3±31, 263 sarcoplasmic reticulum (SR) 20, 21, 57±61, 69, 73±76, 81±84 Saudi Arabia 304±306 Schwartz±Jampel syndrome 34 SCN4A, and myotonias and periodic paralyses 3±32, 355 SCN5A, and long QT syndrome 115, 143, 153 screening 26, 40, 62, 66, 68, 79±83, 161, 165, 168, 184, 217, 240, 335 genome-wide 23, 336 homology 40 mutation 26, 55, 62±68, 79, 88, 160±183

Subject Index

population 240 single strand conformational polymorphism (SSCP) 282 urine 257 secretion hormone 162 insulin 299±322 K, aldosteron 277±292, 331 NaCl, water 227±234 proton 271 segregation 59 sensorineural deafness 349, 351 SERCA2b 71, 72, 273 Shaker gene 172±173, 357 seizures, see febrile seizures and epilepsy site-directed mutagenesis 94, 206, 244, 282, 346 SLC12A1, and Bartter's syndrome 335, 360 SLC12A3, and Bartter's syndrome 327, 332±336, 360 slow channel syndrome 85, 89, 91, 100, 101, 104, 108, 114 sodium channel genes and proteins 355±356 (overview) epithelial, EnaC, and Liddle's syndrome and pseudohypoaldosteronism I (PHA1) 253, 277±297 voltage-gated, and myotonias and periodic paralyses 3-32, 39, 49 and endplate potential 89, 108, 111 and long QT syndrome 147±152 and epilepsies 181, 190±194 persistent current 10±19, 108, 131± 142 slowed inactivation 3, 14, 19 steady-state inactivation 11±14, 19, 25, 134 sodium reabsorption 277, 278 spasmodic mice 215, 359 spastic mice 215, 216, 359 SR, see sarcoplasmic reticulum ST segment, of ECG 142, 144, 145 startle disease, see hyperekplexia STHE, see hyperekplexia

367

stiff baby syndrome, see hyperexplexia stiffness 3, 4, 7, 9, 10, 13, 14, 15, 30, 33, 35, 36, 37, 39, 44, 48, 49, 200, 211, 264 stress-induced symptoms 15, 22, 55, 56, 118, 129, 140, 156±158, 174, 184 stress syndrome, see porcine stress syndrome (PSS) strychnine 200, 206±224 succinylcholine 13, 55, 56 sulfonylurea 322 sulfonylurea receptors 322 suxamethonium, see succinylcholine sweat test 228, 229, 232, 236, 242 sympathetic imbalance 140 synapse 95, 104, 109, 201 terfenadine 141, 148±151 tetrodotoxin (TTX) 10, 137, 146 therapy of cardiac arrhythmia 137, 146, 150±152 congenital myasthenic syndrome 109 cystic ®brosis 227, 236±239, 248, 249 familial hyperinsulinism 304, 322 Liddle's syndrome and pseudohypoaldosteronism 288 malignant hyperthermia 57 migraine and ataxias 160 myotonias and periodic paralyses 3, 10, 13, 17, 23, 28, 49 tubulopathies 330, 341, 343±349, 352 thiazide diuretics 17, 259, 327, 328, 344± 348, 350 thiazide-like salt losing tubulopathy 327± 348 Thomsen's myotonia 4, 10, 14, 33±53, 274, 334, 354, 358 TM2 movement 205 tolbutamide 316 torpedo marmorata 40, 51, 263, 274, 351 torsade de pointes 118, 152 tottering mice 170, 171, 176 transfection 72, 125, 239 transgenic 109±112, 137, 139, 145, 235, 239, 241, 248, 250, 272

368

Subject Index

transplantation 248, 278, 347, 354 transporter, and disease 360 (overview) transverse (t-) tubular membrane 24, 39, 336 triadic junctions 20 triadin 68 triamterene 278, 279 truncation 98, 131, 135, 281, 318 trypsin test 229 TTX, see tetrodotoxin tubulopathy, see hypokalemic salt-losing tubulopathy two-hybrid screen 284 t-tubular, see transverse tubular membrane uricosuria 257 ventricular ®brillation, see also idiopathic ventricular ®brillation 115±119, 133, 139

vermis 159, 160 vertigo 157, 158, 174 visual signs 155 voltage sensor 6, 8, 9, 10, 19, 24, 26, 28, 29, 43, 188, 189, 27, 30, 64, 163 voltage-gated channels 355±360 (overview) wheel-chair 158 whole-cell mode, of patch clamp technique 11, 47, 169, 211, 263, 265 WW domain 284, 296 X-linked recessive hypophosphatamic rickets, see also Dent's disease 257, 259, 358 X-linked recessive nephrolithiasis, see also Dent's disease 257, 275, 276, 334, 358