Abstract
Bartter syndrome type 3 is a clinically heterogeneous hereditary salt-losing tubulopathy caused by mutations of the chloride voltage-gated channel Kb gene (CLCNKB), which encodes the ClC-Kb chloride channel involved in NaCl reabsorption in the renal tubule. To study phenotype/genotype correlations, we performed genetic analyses by direct sequencing and multiplex ligation-dependent probe amplification and retrospectively analyzed medical charts for 115 patients with CLCNKB mutations. Functional analyses were performed in Xenopus laevis oocytes for eight missense and two nonsense mutations. We detected 60 mutations, including 27 previously unreported mutations. Among patients, 29.5% had a phenotype of ante/neonatal Bartter syndrome (polyhydramnios or diagnosis in the first month of life), 44.5% had classic Bartter syndrome (diagnosis during childhood, hypercalciuria, and/or polyuria), and 26.0% had Gitelman-like syndrome (fortuitous discovery of hypokalemia with hypomagnesemia and/or hypocalciuria in childhood or adulthood). Nine of the ten mutations expressed in vitro decreased or abolished chloride conductance. Severe (large deletions, frameshift, nonsense, and essential splicing) and missense mutations resulting in poor residual conductance were associated with younger age at diagnosis. Electrolyte supplements and indomethacin were used frequently to induce catch-up growth, with few adverse effects. After a median follow-up of 8 (range, 1–41) years in 77 patients, chronic renal failure was detected in 19 patients (25%): one required hemodialysis and four underwent renal transplant. In summary, we report a genotype/phenotype correlation for Bartter syndrome type 3: complete loss-of-function mutations associated with younger age at diagnosis, and CKD was observed in all phenotypes.
Bartter syndromes (BS) and Gitelman syndrome (GS) are autosomal recessive salt-losing tubulopathies caused by defective salt reabsorption. They are characterized by hypokalemia, metabolic alkalosis, and secondary aldosteronism, with normal or low BP.1,2 BS are classified by phenotype (antenatal or classic) or genotype (types 1–5). Antenatal BS (ABS) is the most severe form, characterized by polyhydramnios, premature birth, life-threatening episodes of neonatal salt and water loss, hypercalciuria, and early-onset nephrocalcinosis.3 Classic BS (CBS) occurs in infancy or early childhood and is characterized by marked salt wasting and hypokalemia, leading to polyuria, polydipsia, volume contraction, muscle weakness, growth retardation and, sometimes, nephrocalcinosis.4 BS types 1, 2, and 3 are caused by mutations of genes expressed in the thick ascending limb (TAL) of the loop of Henle encoding the luminal Na+–K+–2Cl− cotransporter (SLC12A1; OMIM #601678), the luminal K+ channel ROMK (KCNJ1; OMIM #241200), and the basolateral chloride channel ClC-Kb (CLCNKB; OMIM #607364), respectively.5–7 Loss-of-function mutations of BSND encoding barttin, an essential β subunit for chloride channels, cause BS type 4a with sensorineural deafness (OMIM #602522).8 Simultaneous mutations of CLCNKB and CLCNKA cause type 4b BS (OMIM #613090).9 Finally, severe gain-of-function mutations of the extracellular Ca2+-sensing receptor gene can result in a Bartter-like syndrome (BS type 5, OMIM #601199).10,11 GS (OMIM #263800) is a milder disease frequently associated with hypomagnesemia and hypocalciuria. GS is often asymptomatic or associated with mild symptoms, such as muscle weakness, salt craving, paresthesia, and tetany. GS is related to loss-of-function mutations of the SLC12A3 gene encoding the apically expressed thiazide-sensitive NaCl cotransporter of the distal convoluted tubule (DCT).12
The first described patients with BS type 3 had a clinical phenotype corresponding to CBS.7 Considerable phenotypic variability has since been described: CLCNKB mutations can also underlie the ABS, neonatal BS (NBS), and Gitelman-like (GLS) phenotypes.13–15 This study aimed to shed light on the phenotypic heterogeneity of BS type 3 by investigating phenotype/genotype correlations in a very large French cohort, and by evaluation of published results and original data for in vitro expression.
Results
Population
We retrospectively analyzed results for 115 patients (56 men and 59 women) from 111 families with CLCNKB mutations evaluated at the Genetics Department of Georges Pompidou European Hospital over the last 15 years. A history of consanguinity was recorded for 22 families; the geographic origin is shown in Supplemental Tables 1–3.
Initial Clinical Presentation
Thirty-four patients from 32 families (29.5%) presented with ABS/NBS, 51 patients from 49 families (44.5%) presented with CBS, and 30 patients from 30 families (26%) presented with GLS.
Mutations and Large Rearrangements
Genetic status and mutation type were determined for each initial phenotype group (Table 1). The detailed genotypes of each patient are summarized in Supplemental Tables 1–3. The deletion of a single allele was excluded in patients with homozygous point mutations and no consanguinity, and molecular abnormalities of the other genes implicated in GS and BS were excluded in patients with only one heterozygous mutation. The breakpoints of large rearrangements were not characterized; in consequence, we cannot exclude the possibility that patients with homozygous deletions from nonconsanguineous families harbored two different deletions. Testing was carried out for both parents in 22 families and only the mother in seven families. In all cases, parents were heterozygous for the homozygous mutation detected in the proband or for one of the two mutations detected in compound heterozygous probands.
Genetic status of patients according to initial phenotype, and percentages of mutated alleles by phenotype and mutation type
Sixty different mutations were detected: 55% missense, 13% frameshift, 12% nonsense, 10% large deletions, and 10% splice-site mutations (Figure 1). Twenty-seven of these mutations were previously unknown (Figure 2, A and B, Supplemental Tables 1–3). Two of the three splice-site mutations disrupt the obligatory consensus donor or acceptor splice site and were considered pathogenic as likely to cause exon skipping and frameshift. The variant at position −6 in the acceptor site of exon 14 is a known rare variant (rs369329893, allele frequency in black populations of 0.02%) for which MaxEntScan predicts a 100% decrease in splice-site score and SpliceSiteFinder predicts activation of an intronic cryptic acceptor site. Unfortunately, no mRNA from this patient was available for analysis.
Type of mutations of CLCNKB detected in patients with BS type 3 (n=60).
Locations of novel CLCNKB mutations and of mutations expressed in vitro. (A) CLCNKB gene structure, showing the newly discovered large deletions and splicing mutations. (B) Schematic topological model of the ClC-Kb protein: the lower part of the model corresponds to the intracellular region, and the upper part is extracellular. Each rectangle represents one of the 18 α-helices and the two cystathionine-β-synthase (CBS) domains. The α-helices involved in the selectivity filter, those interacting with Barttin, and those located at the dimer interface are shown in blue, green, and pink, respectively. Previously unknown missense () and nonsense (
) mutations are shown in red; previously described mutations are shown in blue (
); mutations expressed in vitro are underlined.
Two of the 13 previously unreported missense mutations were present in the same allele in patient BR050 (p.Arg395Trp and p.Ala469Pro), and p.Gly465Arg was detected in the same allele as the known mutation p.Pro124Leu in three patients (BR116–1, BR157–1, and GT657–1). Eight of the 13 missense variations affected conserved amino acids and were predicted by at least four out of five tools used for in silico analysis as potentially pathogenic. The remaining five missense changes (p.Ser218Asn, p.Ala254Val, p.Arg395Trp, p.Ile447Thr, and p.Ala469Pro) were classed as variations of unknown significance (VOUS) (Supplemental Table 4). Among these changes, only the p.Arg395Trp has been described in databases (rs34255952) with an allelic frequency of 2% in blacks and has not been detected in whites. Of the 33 missense mutations detected in our population, 13 were previously shown to result in loss of function.16 In silico predictions are presented in Supplemental Table 5 for missense mutations for which in vitro analysis was not performed.
Functional Expression of ClC-Kb Mutants in Xenopus Oocytes
We investigated the effect of two new missense mutations predicted to be pathogenic (p.Gly345Ser and p.Ala510Thr), two new VOUS (p.Arg395Trp and p.Ala469Pro), four previously described missense mutations (p.Gly296Asp, p.Ser297Arg, p.Gly424Arg, and p.Gly433Glu) and two nonsense mutations (p.Trp391Ter and p.Arg595Ter) on chloride conductance in Xenopus oocytes; p.Gly424Arg and p.Gly433Glu are located in α-helix N of ClC-Kb, which is involved in the selectivity filter; p.Gly296Asp and p.Ser297Arg are located in the α-helix J, which interacts with barttin; p.Ala510Thr is found in the α-helix Q, involved in the dimer interface; and Arg595Ter is present in the CBS1 domain involved in channel common gating and trafficking. The p.Gly345Ser and p.Ala469Pro mutations affect α-helices K and O, respectively, and the p.Trp391Ter and p.Arg395Trp mutants affect the L-M linker (Figure 2B). Nine of these ten mutations significantly decreased or abolished normalized conductance (Figure 3). The p.Trp391Ter, p.Gly296Asp, p.Gly424Arg, p.Gly433Glu, p.Ala469Pro, and p.Arg595Ter mutations abolished conductance, whereas p.Ser297Arg, p.Gly345Ser, and p.Arg395Trp decreased conductance to 61%, 57%, and 65% of wild-type values, respectively (significantly different from oocytes expressing wild-type ClC-Kb and noninjected oocytes). By contrast, p.Ala510Thr had no influence on channel conductance. Finally, the p.Arg395Trp/p.Ala469Pro double mutation decreased conductance to 37% of wild-type values.
Functional studies of selected ClC-Kb mutants (n=10). Conductance at +60 mV for noninjected oocytes (NI) and for oocytes into which mutant ClC-Kb cRNA was injected, normalized with respect to the mean value for wild-type (WT) ClC-Kb and expressed as the mean±SEM. The mutants were exposed to a solution at pH 7.4 containing 10 mM Ca2+ (A) or to a solution at pH 9.0 containing 20 mM Ca2+ (B). As ClC-Kb current increases at high external Ca2+ concentration or high pH, these solutions were chosen to obtain a submaximal current. Number of measurements for (A): NI (n=63), WT (n=109), p.Trp391Ter (n=12), p.Arg395Trp (n=16), p.Arg395Trp/p.Ala469Pro (n=20), p.Gly424Arg (n=16), and p.Ala469Pro (n=10). Number of measurements for (B): NI (n=9), WT (n=16), p.Gly122Val (n=6), p.Gly296Asp (n=5), p.Ser297arg (n=7), p.Gly345Ser (n=8), p.Gly433Glu (n=8), p.Ala510Thr (n=8), and p.Arg595Ter (n=3). $P<0.05 for the difference between NI or mutant ClC-Kb and WT; *P<0.05 for the difference between WT or mutant ClC-Kb and NI.
Clinical Data at Diagnosis
Table 2 summarizes clinical and biochemical characteristics at birth and at diagnosis. As expected, gestational age (GA) at birth was significantly lower in the ABS/NBS group than in the CBS and GLS groups, but similar between the CBS and GLS groups. Age at diagnosis was significantly lower in the ABS/NBS group than in the other two groups and in the CBS group than in the GLS group. Polyhydramnios was found in 29 patients with ABS/NBS (85%), at mean GA of 28 weeks, and amniotic fluid had to be drained in four patients. Ten patients (five ABS/NBS and five CBS) had birth weights below the 10th percentile, and four patients (two ABS/NBS and two CBS) had birth heights below the 10th percentile for GA at birth.
Clinical and biologic manifestations at diagnosis
Plasma sodium and chloride concentrations were significantly lower and plasma renin and magnesium concentrations were significantly higher in CBS and ABS/NBS groups than in the GLS group; plasma potassium and total CO2 concentrations were similar in all groups. Strong hypochloremia is a known phenotypic hallmark of BS type 3.17–19 We therefore compared the relationship between plasma sodium and chloride concentrations between patients with BS types 1 and 2 (n=21) and patients with BS type 3 (n=51). This curve was shifted downward in the BS type 3 group, indicating that plasma chloride depletion could not be accounted for by hyponatremia (Figure 4). No difference was observed in parameters at diagnosis when confirmed homozygous patients are compared with compound heterozygous patients (data not shown).
Correlation between plasma sodium and plasma chloride concentrations in patients with BS type 1 and 2 (black symbols, gray area) and in patients with BS type 3 (white symbols, orange area).
Clinical and Biologic Data during Follow-Up
Clinical manifestations during follow-up and treatment were recorded for 77 patients (Table 3). Median follow-up was 8 years and was similar in the three groups. The main treatments administered to these patients were NaCl and KCl supplementation and nonsteroidal anti-inflammatory drugs (mainly indomethacin). The main adverse effects were abdominal pain (n=5), weight gain (n=1), esophagitis (n=1), and diarrhea (n=1). One of the main criteria of a successful treatment is a normal growth; in this cohort, 63 out of 77 patients (82%) had a height between −2 SD and +1 SD or a normal height as adults. Fourteen patients had height below −2 SD: five patients with ABS/NBS, including one with a growth hormone (GH) deficiency (IGF1=38 ng/ml before GH initiation at 15 years of age); eight patients with CBS, including two with GH deficiency and two with CKD; and one patient with GLS with no identified cause of failure-to-thrive.
Associated clinical manifestations and treatment during follow-up, for 77 patients with CLCNKB mutations
Abnormalities in psychomotor and neurologic development included psychomotor retardation in four patients with ABS/NBS and four patients with CBS (Table 3). Five patients with CBS required psychiatric follow-up (hyperactivity, anorexia, or eating disorders).
Irregular heart rate or ECG abnormalities were documented in six patients: one patient with ABS/NBS had premature ventricular beats with prolonged QT interval, three patients with CBS had a right bundle branch block or U wave, and one patients with GLS presented with torsade de pointe attacks. None of the patients in this cohort had high BP.
Fourteen patients developed nephrolithiasis or nephrocalcinosis during follow-up. None of the patients required shockwave lithotripsy. Other renal and urological abnormalities diagnosed in eight patients with ABS/NBS and nine patients with CBS are detailed in Table 3. Proteinuria data were available for 43 patients, nine of whom displayed glomerular proteinuria >50 mg/dl.
Nineteen (ten women and nine men) of the 77 patients (25%) presented with CKD (Table 4). Ten patients presented with stage 2 CKD: four patients with ABS/NBS, four patients with CBS, and two patients with GLS. Two patients reached stage 3 CKD (one patient with ABS/NBS and one patient with GLS). One patient with CBS reached stage 4 CKD. Six patients reached stage 5 CKD (three patients with ABS/NBS and three patients with CBS), at a mean age of 25 (6–49) years. Renal biopsies were performed in five out of 19 patients with CKD. Four patients with stage 5 CKD had diffuse glomerular and tubulointerstitial lesions with enlarged glomeruli presenting with FSGS. One patient with stage 2 CKD had minimal glomerular and tubular alterations (Supplemental Table 6, Table 4). Patients with CKD were older than patients without CKD; they did not differ in terms of birth weight, AINS treatment, urologic or renal abnormalities, or hypokalemia severity (Table 5).
Clinical characteristics of patients with BS type 3 and CKD
Patients with BS type 3: Comparison between patients with CKD and patients with normal GFR
The last eGFR follow-up data for 30 patients with BS type 1, 34 patients with BS type 2, and 11 patients with BS type 4a were compared with eGFR for the 77 patients with BS type 3. These groups had similar age distribution, and eGFR decreased with age (Supplemental Figure 1). In patients with BS type 1 and BS type 4, eGFR decrease was more severe and there were higher proportions of patients with CKD 3–5 than in patients with BS type 2 and BS type 3 (Supplemental Figures 1 and 2).
Genotype/Phenotype Correlation
Large deletions were more frequent in patients with earlier onset, and severe phenotypes and missense mutations were more common in the GLS phenotype (Table 1). Similar results were obtained if other potentially severe mutations (frameshift, nonsense, and essential splicing) were considered with large deletions: severe mutated alleles were more frequent in patients with ABS/NBS and CBS (74 and 66% respectively) than in patients with GLS (42%). Further, missense mutations were more frequent in patients with less severe phenotypes: 58% in patients with GLS versus 34% and 26% in patients with CBS and ABS/NBS, respectively.
We classified mutations into two groups: complete loss-of-function (CL) and partial loss-of-function (PL) groups (Table 6). We included p.Trp610Ter, the only C-terminus–truncating mutation expressed in vitro and yielding a residual current.20 Each mutated allele was classified separately, independently of the initial phenotype, and only patients for whom both alleles could be classified were analyzed (n=85) (Supplemental Tables 1–3). ClC-Kb functions as a homodimer. The residual activity of the CL/PL genotypes may therefore correspond to homodimers of PL mutants, with similar consequences to the PL/PL genotype. We therefore analyzed the CL/PL genotype together with the PL/PL genotype. With this classification, 56 patients had a CL/CL genotype, and 29 patients had a CL/PL or PL/PL genotype. CL/CL genotypes were associated with a significantly younger age at diagnosis than CL/PL and PL/PL genotypes. No difference was observed for the other biologic parameters analyzed (Table 6).
Characteristics of patients according to genotype severity
Discussion
BS are phenotypically and genotypically heterogeneous. Phenotype/genotype correlations highlighting the link between particular traits and genetic types (i.e., transitory hyperkalemia in BS type 2, severe hypochloremia in BS type 3, and hearing loss in BS type 4) have been identified in previous studies.17–19 BS type 3 is particularly heterogeneous in terms of clinical presentation, accounting for the diverse initial diagnoses attributed (ABS/NBS, CBS, or GLS). We investigated the basis of this variability in a cohort of 115 patients harboring CLCNKB mutations, and studied the phenotype/genotype correlation on the basis of clinical presentation and follow-up as well as on in vitro functional studies of missense mutants.
More than 54 mutations of this gene have been reported in free access HGMD (www.hgmd.cf.ac.uk) and scientific publications.4,7,17,19,21–25 They include a high frequency of large rearrangements favored by the close location of the homologous CLCNKA. We detected 60 different mutations, 27 of which had not been previously reported (13 missense, five frameshift, three nonsense, three splice-site mutations, and three large deletions). Thirteen of these mutations (frameshift, nonsense, splice-site mutations, and large deletions) were predicted to result in the production of unstable mRNAs or truncated or absent proteins. Eight of the 13 previously unknown missense mutations were predicted to be pathogenic in silico (Supplemental Table 4). Three out of the other five, classified as VOUS, were expressed in Xenopus laevis oocytes (p.Arg395Trp, p.Ala469Pro, and p.Gly345Ser) as were two previously described mutations (p.Gly424Arg and p.Gly433Glu) detected as the only heterozygous mutation in two patients. All of these mutations significantly decreased chloride conductance. The p.Ala510Thr, predicted in silico as pathogenic, had a chloride conductance similar to that of the wild-type channel. The molecular abnormality of patient heterozygous for this variant thus remains unidentified.
In this large BS type 3 cohort, we confirmed the phenotypic variability, consisting of about 30% ABS/NBS, 45% CBS, and 25% GLS (Table 2). In order to determine if the type of mutation influences the phenotype, we first correlated them with initial clinical presentation. Large deletions and severe mutations were associated with all clinical presentations, but were more frequent in ABS/NBS and CBS. Next, 85 patients with two mutated alleles were analyzed for phenotype/genotype correlations, taking into account the type of mutation and in vitro expression results, regardless of initial clinical presentation. Patients with complete loss-of-function (CL/CL) were significantly younger at diagnosis than patients harboring one or two alleles with a partial loss-of-function (CL/PL or PL/PL), suggesting that the type of mutation may influence the clinical presentation of BS type 3.
Surprisingly, the milder GLS phenotype did occur in patients harboring severe mutations or deletions, suggesting that the phenotype severity is not only driven by CLCNKB allelic variability. Phenotypic heterogeneity of BS type 3 has been attributed to distribution of the ClC-Kb channel along the nephron and to possible compensatory function of the ClC-Ka channel. ClC-Kb channel is expressed in the TAL, DCT, and collecting duct, where it transfers chloride (Cl−) ions to the basolateral side.26 Impaired ClC-Kb function in the TAL results in lower levels of Cl− exit, NaCl reabsorption through the Na-K-2Cl cotransporter, and divalent cation reabsorption, accounting for the Bartter phenotype. Defective basolateral Cl− exit in the DCT decreases NaCl reabsorption via the NaCl cotransporter, accounting for the GLS phenotype in other patients. Two recent studies in which the mouse Clcnk2 gene (corresponding to CLCNKB in humans) was disrupted confirmed that ClC-K2 is the principal chloride channel in all three nephron segments and that TAL impairment is not compensated by ClC-K1 (corresponding to the human ClC-Ka channel, which is also expressed in the TAL).27,28 Nevertheless, it cannot be excluded that allelic variants of genes encoding KCl cotransporters or other chloride channels may also compensate for renal sodium loss, accounting for phenotypic variability.27,29
Next generation sequencing (NGS) approaches allow parallel analysis of several genes, which is particularly useful in diseases with genetic heterogeneity, such as ABS, or in diseases with phenotypic variability such as BS type 3. A genetic confirmation is important for the follow-up as well as to improve our knowledge of the natural history of these syndromes. Therefore, we recommend the use of NGS panels to diagnosis confirmation. NGS could also be useful to determine whether additional genes are involved in the observed clinical variability. In the future, in vitro studies of additional mutant proteins can contribute to improving our understanding of the phenotype/genotype correlation and of the precise pathogenic mechanism of mutants. These studies have the potential interest to define targeted therapeutic approaches, such as channel openers or pharmacologic chaperones.30,31
Despite missing data for some phenotypic criteria because of the retrospective nature of this study, several patterns emerged from our analysis. First, growth retardation was common but frequently improved with treatment. Fourteen patients presented with persistent growth retardation; two of these patients had CKD, a common cause of growth retardation due to a combination of abnormalities of the growth hormone axis, vitamin D deficiency, hyperparathyroidism, inadequate nutrition, and drug toxicity; and three patients presented with GH deficiency.32 BS and potassium deficiency have already been reported to be associated with GH deficiency.33–35 One previous study showed that GH and IGF1 did not stimulate longitudinal growth unless hypokalemia was corrected.36 In two patients with hypokalemia (median, 2.5 mmol/L), growth improved after GH supplementation but remained below −2 SD.
Second, hypochloremia is a hallmark of BS type 3: an analysis of the data available at diagnosis showed that hypochloremia was more severe in patients with ABS/NBS and CBS than in patients with GLS. ClC-Kb is expressed not only in the diluting segment but also in the intercalated cells of the collecting duct. Defects in this segment may impair chloride exit and transepithelial chloride reabsorption through the pendrin Cl/HCO3 exchanger, potentially accounting for the stronger chloride depletion in patients with BS type 3 than in patients with BS type 1 or BS type 2.19 We found a downward shift of the relationship between plasma chloride and sodium concentrations consistent with a defect in adaptation to chloride depletion in patients with BS type 3 as compared with patients with BS type 1 and BS type 2 patients. These results are consistent with the phenotype of mice with Clcnk2 disruption,27 and suggest that sodium and potassium supplementation should be provided as chloride salts in patients with BS type 3.
Third, 19 patients presented with CKD, and seven of these patients also had proteinuria (Table 4). Five patients underwent renal biopsy, which revealed diffuse glomerular and tubulointerstitial lesions with enlarged glomeruli in four patients, suggesting compensatory hypertrophy to nephron reduction (Supplemental Table 6, Table 4). Six patients diagnosed before the age of 8 years displayed progression to ESRD at a median age of 24 years, associated with FSGS in four patients. Proteinuria, a low GFR, and FSGS have been reported in patients with BS and GS.22,37–39 It has been suggested that FSGS is a secondary lesion because of adaptation to salt loss, resulting in chronic stimulation of the renin-angiotensin system.37,39,40 In this study, FSGS occurred in late-stage CKD, suggesting a large contribution of nephron reduction. We failed to identify other risk factors of CKD progression, including birth weight, age at diagnosis, long-term nonsteroidal anti-inflammatory drug treatment, persistent hypokalemia, and other renal abnormalities (Table 5). CKD has been described in other types of BS.19,22 In our BS cohort, CKD was observed in all BS types but the proportion of patients with preserved renal function (i.e., eGFR>90 ml/min per 1.73 m2) was higher in patients with BS types 2 and 3 and the proportion of patients with moderate to severe kidney disease (i.e., eGFR<60 ml/min per 1.73 m2) in patients with BS types 1 and 4, suggesting that the later BS subtypes 1 and 4 are associated with more severe renal prognosis. The mechanism of CKD development is probably multifactorial, and its elucidation will require prospective studies. Case reports are rare for patients with BS undergoing renal transplantation. The post-transplantation period was uneventful in our four patients, with the complete disappearance of BS and no recurrence of FSGS, as previously described.41–43
In conclusion, BS type 3 syndrome, which is caused by CLCNKB mutations, is highly variable phenotypically. We show, for the first time, that there is a correlation between severe mutations and a significantly younger age at diagnosis, suggesting that milder defects of ClC-Kb function may account for some of this variability. We also confirm the severe chloride depletion previously observed in patients with BS type 3 and report that 25% of patients suffer from CKD. Long-term prospective follow-up of this cohort will identify other severity parameters involved in this genotype/phenotype correlation, and will allow us to evaluate whether early diagnosis and treatment have an influence on the evolution to CKD.
Concise Methods
Patients
The study included 115 patients (from 111 families) with CLCNKB mutations referred to the Genetics Department of Georges Pompidou European Hospital (Paris, France) from January of 2001 to December of 2014 for genetic analysis after the diagnosis of BS or GS. The study was approved by the Comité de Protection des Personnes, Paris-Île de France XI (reference no. 09069) and informed consent for genetic studies was obtained from each proband or from their parents (for minors). Genetic investigations were performed after the clinical and biologic diagnosis of salt-losing tubulopathy. Patients with a history of polyhydramnios or clinical manifestations in the first month of life were considered to have ABS/NBS. Patients diagnosed during childhood, with hypercalciuria and/or polyuria, were considered to have CBS, and children, adolescents, or adults for whom hypokalemia and hypomagnesemia and/or hypocalciuria were discovered fortuitously were considered to have GLS. Genetic investigations were extended to both parents in 22 families and to the mother only in another seven families. Twenty-three patients from this cohort have been described before19,23,25 (Supplemental Tables 1–3).
Detection of Point Mutations
DNA was extracted with a salt-based method or with blood DNA midi kits (Qiagen, Venlo, The Netherlands). CLCNKB exons and flanking intron sequences were amplified by PCR, sequenced with BigDye Terminator v3.1 cycle sequencing kits, and run on an ABI Prism 3730XL DNA Analyzer (Perkin Elmer Applied Biosystems, Foster City, CA), as previously described.19,23 DNA mutations were identified with Sequencher software, by comparison with the reference sequence for CLCNKB: NM_000085.4. Each mutation was confirmed by sequencing a second independent PCR product.
Detection of Large Rearrangements
Large rearrangements were detected by quantitative multiplex PCR of short fluorescent fragments before June of 2010, and by multiplex ligation-dependent probe amplification (MLPA) thereafter. We adapted the quantitative multiplex PCR of short fluorescent fragments method for the detection of large deletions of CLCNKB.44 The procedure is described in detail in the Supplemental Material and the primers used, covering all exons, are listed in Supplemental Table 7. For MLPA, we used the SALSA MLPA P266-B1 CLCNKB Kit (MRC Holland, Amsterdam, The Netherlands). The P136 Kit contains 29 probes: probes for 14 of the 20 exons of CLCNKB (exons 4, 7, 9, 12, 16, and 20 are not represented), four probes for upstream genes (PRM2, CASP9, and the homologous CLCNKA gene), and 11 reference probes. The procedure is described in detail in the Supplemental Material.
Bioinformatic Analysis of Mutations
The software used to interpret variants is described in the Supplemental Material.
Functional Expression in X. laevis
Voltage clamp experiments were performed as previously described in X. laevis oocytes.23,25 We injected 10 ng ClC-Kb cRNA and 5 ng barttin cRNA into defolliculated oocytes, which were then incubated in modified Barth solution at 16°C. Two-electrode voltage clamp experiments were performed at room temperature with TURBO TEC-10CX (npi electronic GmbH, Tamm, Germany) and PClamp 8 software (Axon Instruments, Union City, CA), 2 to 3 days after injection. Conductance at +60 mV (G+60 mV) was calculated by dividing the current at +60 mV by the difference in current between +60 mV and the reversal potential.
Statistical Analyses
Clinical data were analyzed with GraphPad Prism Software (La Jolla, CA) and SPSS software, release 20.0.0 (SPSS, Chicago, IL). Kruskal–Wallis tests were used to compare the three groups (ABS/NBS, CBS, and GLS). In cases of statistical significance, Mann–Whitney U tests were used to compare the groups in pairs. Dichotomous variables were compared using the chi-squared test or Fisher exact test as appropriate. Variables for which >50% of the data were missing in one group were excluded from the analysis. Clinical manifestations were analyzed by descriptive methods, using the number of subjects for each group. Data for in vitro X. laevis studies were analyzed by ANOVA and Holm–Sidak tests, with Sigma Stat software.
Disclosures
None.
Acknowledgments
We thank the patients and their families for agreeing to participate in this study. We thank all of the staff of the genetics laboratory of Georges Pompidou European Hospital and all of the doctors involved: Hans-Jacob Bangstaf (Norway), Pierre Bataille (Boulogne-sur-Mer, France), Mohammed-Najib Bensemlali (Morocco), Delphine Brouet (Lorient, France), Christophe Bouaka (Martigues, France), Marina Charbit (Paris, France), Gabriel Choukroun (Amiens, France), Renaud De La Faille (Bordeaux, France), Geneviève Dumont (Orleans, France), Olivier Dunand (La Reunion, France), Catherine Dupré-Goudable (Toulouse, France), Claire Garandeau (Nantes, France), Angela Granadas (Madrid, Spain), Steven Grange (Rouen, France), Jean-Christophe Hebert (Guadeloupe, France), Aurélie Hummel (Paris, France), Khalid Ismaïli (Brussels), Cynthia Kahil (Thonon les Bains, France), Christina Kanaka (Athens, Greece), Bruno Legallicier (Rouen, France), Renaud Leray (Montpellier, France), Bruno Maranda (Sainte Foy, Quebec, Canada), Mohammed-Diab Mahmoud (La Roche Sur Yon, France), Pierre Merville (Bordeaux, France), Bruno Moulin (Strasbourg, France), Isabelle Raingeard (Montpellier, France), Mehta Sanjay (Toronto, Canada), David Schwarz (Zurich, Switzerland), Pascale Siohan (Quimper, France), Ivan Tack (Toulouse, France), Cécile Vigneau (Rennes, France), and João Esteves (Ponta Delgada, Açores, Portugal).
J.T.’s group is funded by a grant from l’Agence Nationale de la Recherche (grant no. ANR BLANC 14-CE12-0013-01/HYPERSCREEN). This work was supported by the French Ministry of Health (Plan Maladies Rares) and the European Community (grant nos. FP7EUNEFRON 201590 and EURenOmics 2012-305608).
Footnotes
E.S. and O.A. contributed equally to this work.
Published online ahead of print. Publication date available at www.jasn.org.
This article contains supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2016101057/-/DCSupplemental.
- Copyright © 2017 by the American Society of Nephrology