|Home | About | Journals | Submit | Contact Us | Français|
Five children from two consanguineous families presented with epilepsy beginning in infancy and severe ataxia, moderate sensorineural deafness, and a renal salt-losing tubulopathy with normotensive hypokalemic metabolic alkalosis. We investigated the genetic basis of this autosomal recessive disease, which we call the EAST syndrome (the presence of epilepsy, ataxia, sensorineural deafness, and tubulopathy).
Whole-genome linkage analysis was performed in the four affected children in one of the families. Newly identified mutations in a potassium-channel gene were evaluated with the use of a heterologous expression system. Protein expression and function were further investigated in genetically modified mice.
Linkage analysis identified a single significant locus on chromosome 1q23.2 with a lod score of 4.98. This region contained the KCNJ10 gene, which encodes a potassium channel expressed in the brain, inner ear, and kidney. Sequencing of this candidate gene revealed homozygous missense mutations in affected persons in both families. These mutations, when expressed heterologously in xenopus oocytes, caused significant and specific decreases in potassium currents. Mice with Kcnj10 deletions became dehydrated, with definitive evidence of renal salt wasting.
Mutations in KCNJ10 cause a specific disorder, consisting of epilepsy, ataxia, sensorineural deafness, and tubulopathy. Our findings indicate that KCNJ10 plays a major role in renal salt handling and, hence, possibly also in blood-pressure maintenance and its regulation.
The study of molecular defects in rare inherited renal tubular diseases has substantially advanced both our understanding of renal salt and water handling and the management of common disorders such as systemic hypertension.1–7 Two well-known disorders of renal tubular salt handling, Bartter’s syndrome and the Gitelman syndrome, are characterized by polyuria and normotensive hypokalemic metabolic alkalosis. We describe five children with similar clinical findings, as well as infantile-onset seizures, ataxia, and sensorineural deafness. We determined the genetic basis of this disease and the pathophysiological mechanism accounting for its seemingly divergent clinical manifestations.
Genetic studies were approved by the Institute of Child Health–Great Ormond Street Hospital Research Ethics Committee, and the parents provided written informed consent. Genotypes from DNA of the four affected children in Family 1 and the parents of the three affected siblings within this pedigree (Fig. 1A) were generated with the use of single-nucleotide polymorphism (SNP) chip arrays (GeneChip Human Mapping 10K Array Xba142 2.0, Affymetrix) according to the manufacturer’s recommendations. Genotypes were examined with the use of a multipoint parametric linkage analysis and haplotype reconstruction performed with SimWalk (version 2.91) for an autosomal recessive model with complete penetrance and a disease allele frequency of 0.001 (deCode SNP map with Asian allele frequencies).8 The data were formatted with Mega2 (version 4.0) through ALOHOMORA (version 0.30, Win32).9,10 Mendelian inconsistencies were checked with the use of PedCheck (version 1.1); unlikely genotypes were filtered with the use of Merlin (version 1.1, alpha 3).11,12 The SimWalk haplotype output files were visualized with HaploPainter.13
We sequenced the complete coding region and splice sites of KCNJ10, a gene encoding a potassium channel expressed in the brain, inner ear, and kidney, in all the affected children and their unaffected parents, as previously reported.14 The presence of sequence variations was checked in at least 100 ethnically matched control alleles.
Electrophysiological measurements were performed with the use of a classic two-electrode voltage-clamp microelectrode approach. Heterologous expression of mutant and wild-type KCNJ10 in Xenopus laevis oocytes was produced as previously reported.15 Studies conformed to relevant U.K. Home Office regulations. In short, oocytes were injected with 4 to 20 ng of complementary RNA (cRNA), transcribed with the use of mMessage-Machine (Applied Biosystems), from human KCNJ10 cloned into the vector pTLB. Measurements were performed in 20 mM of potassium chloride at −80 mV in at least four batches of oocytes. KCNJ10-specific currents in the injected oocytes, as compared with uninjected oocytes, were confirmed by means of blockade with barium.
Kcnj10 knockout mice were generated as previously described16 and were provided by Drs. Clemens Neusch and Frank Kirchhoff, Max Planck Institute for Experimental Medicine, Göttingen, Germany. The experiments were approved by the local councils for animal care and conducted according to German regulations governing animal care.
Urine was collected from the mice when they were 3 days old and was analyzed as previously reported.17 Urine electrolytes were measured by Katja Huggel and Dr. François Verrey, University of Zürich, Zurich, Switzerland, with the use of ion chromatography (Metrohm).
Immunolocalization studies of kidneys from 12-week-old C57BL/6 mice (Charles River) were performed according to published methods.18 Incubation and imaging with the use of primary antibodies to KCNJ10 (Alomone), aquaporin-2 (St. Cruz), calbindin (Sigma), NKCC2,19 and NCC20 were performed according to standard procedures.
Electrophysiological in vitro experiments and mouse urinalyses were interpreted with the use of an unpaired, two-sided Student’s t-test comparing effect and appropriate controls. Data are reported as means and standard error of the mean. A P value of less than 0.05 was considered to indicate statistical significance.
We identified five children from two consanguineous families (Fig. 1A) who had the clinical features of epilepsy, ataxia, sensorineural deafness, and a renal salt-losing tubulopathy. All other members of these families were clinically unaffected. None of the five children (four in Family 1 [Patients 1-1, 1-2, 1-3, and 1-4] and one in Family 2 [Patient 2-1]) had been born prematurely.
All affected patients initially presented with generalized tonic–clonic seizures in infancy (Table 1). These seizures were easily controlled with a broad-spectrum anticonvulsant agent, but focal seizures subsequently occurred in Patient 1-2. Routine electroencephalographic recordings in Patients 1-1 and 1-2 were normal at the ages of 7 and 8 years, respectively. All five patients had speech and motor delay, and they had pronounced gait ataxia, intention tremor, and dysdiadochokinesis, findings that are consistent with cerebellar dysfunction; Patients 1-2 and 2-1 were unable to walk. Ataxia was apparent from an early age and was nonprogressive. The results of magnetic resonance imaging (MRI) of the brain, performed in Patients 1-1 and 1-2, were normal (Fig. 1B). Electromyographic studies, performed in three patients, and nerve conduction velocities, assessed in all five patients, were also normal.
Hearing impairment was noted in Patient 1-1 at the age of 5 years (Fig. 1C) and in Patient 2-1 at 1 year. The grade of hearing impairment in Patient 1-1 remained stable over the subsequent 8 years. Patients 1-3 and 1-4 also had sensorineural hearing impairment; Patient 1-2 was not tested.
All affected patients had evidence of stimulated renin systems, with hypokalemic metabolic alkalosis, and they also had hypomagnesemia and hypocalciuria (Table 2).21 All patients were receiving potassium and magnesium supplements. Urinary concentrating ability, assessed by means of spot urine osmolality measurements, did not appear to be grossly affected (Table 2). Proteinuria and glycosuria were not present. Ultrasonography showed that the kidneys were normal in size, position, and structure. Blood pressure was at the low end of the normal range, typically at the 25th percentile for age and sex. Other pertinent clinical details are listed in Tables 1 and and22.
Family 1 was of Pakistani origin; the affected patients were the offspring of first-cousin marriages who also shared a common ancestor five generations earlier (Fig. 1A). Family 2 was of Arabic descent, and the affected patient’s parents were first cousins (Fig. 1A). Whole-genome linkage analysis and a subsequent haplotype reconstruction identified a single region of interest of 1.9 centimorgans (cM), equaling approximately 800,000 bases, with a lod score of 4.98 on chromosome 1 (Fig. 2A and 2B). This region localized between SNPs rs726640 and rs1268524 and contained a total of 31 annotated genes, including KCNJ10. Haplotype reconstruction suggested that the affected patients in Family 1 were homozygous by descent for the alleles in the critical region. KCNJ10, also known as Kir4.1, constituted an attractive candidate gene, since mice in which this gene has been deleted have seizures, ataxia, and sensorineural deafness16,22,23; kidney function in these mice has not been studied. Variations in KCNJ10 have been reported to be associated with seizure susceptibility in humans,24 although to our knowledge, no mutations in KCNJ10 have been recorded.
Sequencing of the complete coding region of KCNJ10 revealed a homozygous missense mutation, c.194G→C (p.R65P), in the four affected patients in Family 1 and another homozygous missense mutation, c.229G→C (p.G77R), in Patient 2-1 (Fig. 3A and 3B). Parents were heterozygous for the respective mutations (data not shown). Sequencing of 192 ethnically matched control alleles did not reveal the sequence variation p.R65P, identified in Family 1. Likewise, p.G77R, identified in Family 2, was not seen in 108 matched control alleles. Protein-homology analysis revealed that R65 and G77 were conserved among all 21 species studied (Fig. 3C). Residue R65 is predicted to localize to the beginning of the first transmembrane helix of this two-transmembrane–domain potassium channel; G77 also lies within this first transmembrane helix (Fig. 3D).
Expression of wild-type KCNJ10 resulted in robust currents with the typical characteristics of an inward-rectifier potassium channel (data not shown). In contrast, currents from mutant R65P KCNJ10 were reduced to approximately 25% and those from mutant G77R KCNJ10 to about 5% of wild-type controls (Fig. 3E).
KCNJ10 is expressed in the distal tubule of the kidney25; its presence in the thick ascending limb has also been suggested.26 We used a KCNJ10-specific antibody in wild-type mice to demonstrate the presence of Kcnj10 distal to the macula densa (i.e., on the basolateral membrane of the distal convoluted tubule, the connecting tubule, and the early cortical collecting duct) (Fig. 4A). The absence of signal in Kcnj10 knockout mice verified the specificity of the antibody (data not shown).
Kcnj10 knockout mice die very early as a result of central nervous system symptoms such as seizures.16 However, a conditional knockout of Kcnj10 in the brain leads to death later in life, suggesting that renal salt loss is an aggravating factor in the mice with complete knockout.23 To our knowledge, renal involvement in Kcnj10 knockout mice has not previously been recognized, although these mice clearly have diminished growth and no weight gain after birth (Fig. 4B).
In our studies, Kcnj10 knockout mice died after day 8. In the neonatal period, all seven knockout mice had a significantly lower urinary creatinine concentration than did normal mice, indicating polyuria; the urinary sodium concentration was significantly elevated in the knockout mice, indicating renal salt loss. These neonatal knockout mice also had significantly reduced calcium excretion (Fig. 4C), as do patients with the Gitelman syndrome. In addition, the tubulopathy observed in the knockout mice closely resembled that seen in our patients. Urinary findings in 18 heterozygous mice were not significantly different from those in 11 wild-type mice (data not shown).
We report a unique constellation of multiorgan signs and symptoms — epilepsy, ataxia, sensorineural deafness, and a renal salt-losing tubulopathy, which we call the EAST syndrome — associated with mutant KCNJ10. The multiplicity of symptoms reveals the important role of KCNJ10 in these various organ systems.
We detected homozygous missense mutations in our patients that substantially impair the function of KCNJ10, a potassium-channel gene. The identified mutations reside in a highly conserved area — namely, a transmembrane region that probably determines the overall function of this type of channel.27 Indeed, in a heterologous expression system, these mutations nearly abrogated potassium current. Mice lacking this potassium channel died as neonates; homozygous nonsense mutations or gene deletions of this channel may also be fatal in humans. Alternatively, it is possible that KCNJ10 has a more important role in mouse kidneys than in human kidneys.28 In any case, our patients are probably in a state of compensated volume, whereas knockout mice are not.
The currently accepted model of renal epithelial salt transport posits that a favorable electrochemical gradient drives the influx of sodium (often transported with other substances) from the tubular lumen into the cell. This gradient is established by the basolateral sodium–potassium pump (Na+/K+-ATPase), which also provides an exit mechanism for sodium. In order for this primary active Na+/K+-ATPase to function, potassium must be able to leave the cell and must be readily available basolaterally.29 The task of KCNJ10 is to recycle potassium, which is necessary for the function of the primary active Na+/K+-ATPase. This “pump coupling” was postulated in 1958 by Koefoed-Johnsen and Ussing,30 and experimental evidence has been subsequently corroborated by many investigators.31,32 Our findings appear to constitute genetic proof for this basic physiological principle. A functional basolateral potassium channel also translates the potassium concentration gradient into a cell-negative transmembrane potential. Since uptake of luminal sodium chloride in the distal convoluted tubule is electroneutral, impairment of the negative membrane voltage itself would not be problematic. However, because the functions of other channels are critically dependent on membrane voltage, impairment of other transport processes (e.g., for chloride or magnesium) can occur.
KCNJ10 activity, according to this view, provides a mechanism for indirectly regulating re-absorption of renal tubular sodium, which modulates volume homeostasis and maintains blood pressure. Indeed, our patients, lacking normal KCNJ10 activity, had low blood pressure. Moreover, chromosome 1q23.2, where the locus for KCNJ10 resides, has been implicated repeatedly as being linked to blood-pressure variation in different ethnic groups.33–35 In addition, a whole genome scan for the identification of blood-pressure modifiers (as a quantitative trait) in hypertensive and normotensive Lyon rat strains showed linkage to the region syntenic to human chromosome 1q23.2.36
Although the presence of basolateral potassium channels in the renal tubule, including the distal convoluted tubule, has long been known from electrophysiological studies, the molecular identity of these channels has remained unresolved.37 Our results should establish KCNJ10 as a critical component of basolateral potassium conductance in the distal convoluted tubule. Indeed, similarities in the electrophysiological properties of KCNJ10 and other known basolateral potassium channels have led to speculation that basolateral potassium conductance is achieved by the KCNJ10 protein in heteromerization with other potassium-channel proteins.38
Thus, loss of KCNJ10 function probably leads to a compensated state of salt loss, resulting in stimulation of the renin–angiotensin–aldosterone system and the respective renal tubular activation of potassium secretion in the aldosterone-sensitive nephron (collecting duct). The concomitant proximal tubular increase in bicarbonate (resulting in metabolic alkalosis) and calcium absorption (with consequent hypocalciuria) causes additional signs and symptoms of the EAST syndrome (Tables 1 and and22).
Seizures and ataxia develop in mice with Kcnj10 deletion, and they die shortly after birth, reflecting the critical role of KCNJ10 in the functioning of the central nervous system.16 Even mice with a conditional knockout of Kcnj10 in glial cells alone die at approximately 3 weeks of age.23 In humans, KCNJ10 is expressed in glial cells in the cerebral cortex and cerebellar cortex and in the caudate and putamen, and it is believed to establish the neuronal cell’s resting membrane potential through a process called potassium spatial buffering.39 With repeated excitation and repolarization, a neuron takes up substantial amounts of sodium and loses potassium. Thus, potassium accumulates extracellularly, decreasing the membrane potential, facilitating further excitations, and creating a diathesis toward seizures. Glial cells presumably take up the extruded potassium and distribute it through gap junctions; KCNJ10 has been implicated in this process.16 We propose that the absence of fully functional KCNJ10 removes this protective “potassium sink,” accounting for the seizures in our patients.
Investigations in mice have shown that Kcnj10 is expressed in intermediate cells in the stria vascularis and is required for the generation of the endocochlear potential, suggesting that it contributes indirectly to potassium enrichment of the endolymph.22 This explains why Kcnj10 knockout mice have markedly impaired hearing. Our patients’ hearing was only moderately impaired, and in two patients (1-3 and 1-4), hearing impairment was noted only on specific testing — findings that are consistent with the presence of residual channel function,
Our observations also show further genetic heterogeneity among the salt-losing tubulopathies and establish KCNJ10 as a basolateral potassium channel that is necessary for proper salt handling in the distal convoluted tubule of the kidney. These observations illustrate the critical role of KCNJ10 in the human brain and inner ear and provide the basis for further studies of the pathogenesis and potential treatment of epilepsy, movement disorders, and sensorineural deafness. Mutations in the transport genes of the renal tubular lumen often lead to kidney-specific phenotypes, as in the Gitelman syndrome and Bartter’s syndrome types I and II.2–4,40 Mutations in the basolateral subunit of a renal tubular chloride channel result in clinical findings in the kidney and the ear, as seen in Bartter’s syndrome type IV.6 We now know that mutations in KCNJ10 lead to distinct epithelial-transport abnormalities in the kidney and the ear. In other organs and systems, the KCNJ10 channel plays a major role in modulating resting membrane potentials in excitable cells, causing epilepsy if mutated.
The EAST syndrome should be suspected in patients presenting with any cardinal signs or symptoms of epilepsy, ataxia, or sensorineural deafness, especially if a concurrent electrolyte disorder, such as hypokalemia or hypomagnesemia, is diagnosed.
We speculate that the identification of the genetic basis of the EAST syndrome reveals a key role of KCNJ10 in the modification of volume homeostasis. Reevaluation of genomewide association studies may identify KCNJ10 as a candidate gene associated with blood pressure and its regulation.
Supported by the Intramural Research Programs of the National Human Genome Research Institute, National Institutes of Health, the Special Trustees of the Great Ormond Street Hospital, St. Peter’s Trust for Kidney, Bladder, and Prostate Research, the Grocers’ Charity, the David and Elaine Potter Charitable Foundation, and Deutsche Forschungsgemeinschaft (SFB699).
Dr. Sheridan reports receiving grant support from the Well-being of Women. No other potential conflict of interest relevant to this article was reported.