| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrinology and Metabolism Service (S.T., J.L., M.H., Y.N., N.K., E.B.-S., O.S., E.N., T.P., B.G.), Internal Medicine Department, and Department of Pediatrics (D.Z., H.L.), Hadassah-Hebrew University Medical School, Jerusalem 91120, Israel; Departments of Medicine and Cellular and Molecular Biology (A.C., L.A.-B.), Baylor College of Medicine, Houston, Texas 77030; School of Biological Science (K.E.C., M.J.D.), The University of Manchester, Manchester M13 9PT, United Kingdom; The London Centre for Childhood Pancreatic Disease (K.H., K.L.), Great Ormond Street Hospital for Children National Health Service Trust, London WC1N 3JH, United Kingdom; The Institute of Child Health (K.H., K.L.), London WC1N 3JH, United Kingdom; Pediatric Endocrine Unit (Y.T.-R.), Ha’ Emek Medical Center, Afula 18101, Israel; Institute for Endocrinology and Diabetes (L.D.V.), National Center for Childhood Diabetes, Schneider Children’s Medical Center of Israel, Petah Tikva 49202 and Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel; Hospital Infantil Universitario Niño Jesús (J.A.), E-28009 Madrid, Spain; Hospital Infantil Universitario La Paz (R.G.), E-28034 Madrid, Spain; and Pediatric Endocrinology Clinic (A.E.), Pediatric Department, Meir General Hospital, Kfar-Saba, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel
Address all correspondence and requests for reprints to: Benjamin Glaser, M.D., Endocrinology and Metabolism Service, Hadassah-Hebrew University Medical Center, Jerusalem 91120, Israel. E-mail: beng{at}cc.huji.ac.il.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We selected a cohort of HI patients and, using direct sequencing, systematically searched the entire coding and intron/exon boundaries and 5'-upstream regions of ABCC8 and KCNJ11. If no mutation was identified, then the coding region and intron/exon boundaries of the other three previously identified HI candidate genes were sequenced. Twelve ABCC8 and KNCJ11 mutations were found in 11 patients. No mutations were identified in four patients.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Fifteen index patients with the clinical diagnosis of HI were selected for this study. Patients were chosen from a very heterogeneous cohort of more than 150 HI individuals from throughout Israel and Europe (DNA donated by patients and physicians). Selection criteria included clear clinical diagnosis of HI based on accepted criteria, which included inappropriate insulin levels at the time to hypoglycemia and increased glucose requirements to prevent hypoglycemia (19, 20); significant clinical data available to allow correlation between genetic and clinical findings; availability of sufficient DNA to complete the project; and whenever possible, ethnic isolates with multiple cases so that further genetic studies could be accomplished if indicated. The study population was limited to 15 due to technical reasons, and the first 15 patients that met the inclusion criteria were selected; thus, this population cannot be considered representative of the HI population in general. Informed consent was obtained from the parents of all patients, and the study was performed in accordance with the ethical guidelines of participating institutions. The basic clinical characteristics of the index patients are described in Table 1
.
|
All coding exons and exon/intron boundaries of ABCC8, KCNJ11, GLUD1, GCK, and SCHAD were PCR amplified, purified, and sequenced using the Thermo Sequenase Radiolabelled Terminator Cycle Sequencing Kit (Amersham Pharmacia Biotech, Buckinghamshire, UK). PCR primers are available from the authors upon request. The regions of the ABCC8 and KCNJ11 promoters previously shown to contain major promoter activity were also amplified and sequenced (21). For ABCC8, this included the entire 5'-untranslated region (UTR) and 220 bases 5' to the transcription start site (forward primer, CAC CCC ACT CCC CAT CTT AG; reverse primer, ACC CCC TGG TCC ACC CGG TA). The KCNJ11 promoter was divided into two amplicons. The first included 70 bases 5' to the transcription start site and 250 bases of exon 1 (forward primer, GTC TCG AAC TCC TGA CCT AGT; reverse primer, CAG GCG TGT CAG CAC GTA TT). The second segment, which contained nucleotides 712-1020 bases 5' to the transcription start site, had to be amplified using a nested PCR (first primer set: forward, CTC CAA AGG CCA GGT TGT GAG TCC; reverse, GGT GGC TCA CGC CTG TAA TCC CAG TAC GTT; and nested primers: forward, GTG CGC CCC CCT CCC GCC GTC CTA; reverse, ATC CTG CGT TCT CTG GAG TC). These two segments covered the entire region with known major promoter activity (21).
For each novel variant identified, a control group consisting of at least 50 individuals was tested to exclude the possibility that the variant is a common polymorphism. The control group was either Caucasian or Arab, depending on the ethnic background of the proband. Two hundred ten Ashkenazi Jewish control subjects were screened for the presence of the ABCC8 promoter variant (–64 c
g, see, Mutation analysis).
Electrophysiology
All data were obtained from isolated preparations of viable cells after surgery using the inside-out recording configurations of the patch-clamp technique as described previously (5, 22). The pipette contained a standard NaCl-rich bathing solution containing 140 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 1.13 mM MgCl2, 10 mM HEPES, and 2.5 mM glucose (pH 7.4 with NaOH), and the bath solution contained 140 mM KCl, 10 mM NaCl, 1.13 mM MgCl2, 1 mM EGTA, 2.5 mM glucose, and 10 mM HEPES (pH 7.2 with KOH) for all recordings. In patches with a suitable number of KATP channel events, open-state probability was assessed as described previously (5, 22).
Studies of promoter function
ABCC8. PCR products of the wild-type and mutant ABCC8 minimal promoter region and the 5'-UTR were subcloned into the pGEM-T Easy vector (Promega, Madison, WI) and then into the pGL2-Basic vector (TP99; Promega). All clones were sequenced to confirm orientation and the integrity of the sequence. Vector constructs were transiently transfected (calcium phosphate-DNA coprecipitation) into hamster insulinoma tumor cells (HIT-T15) together with cytomegalovirus ß-Gal control plasmid as follows. Cells (1 x 106) were plated on a 2-cm2 plate and incubated overnight at 37 C. The following day, transfections were performed for each plate. The transfection mixture contained 1 µg of test construct or control plasmid (CMV-ß-Gal), 125 mM (final concentration) CaCl2 in HEPES-buffered saline (final concentration, 140 mM NaCl, 0.75 mM Na2HPO4, and 25 mM HEPES; pH 7.05; total volume, 240 µl). After 5 min in ice, this mixture was added to the cells, which were then incubated for 4 h. Glycerol diluted in 1x PBS (1:5) was then added for 1 min, after which the cells were washed three times in 1x PBS, covered with 2 ml of medium, and incubated for 48 h. Cells were then harvested and assayed for ß-Gal (using standard methods) and luciferase activity according to the Luciferase Assay System (Promega).
KCNJ11. The 5'-UTR was amplified using the following primers modified to introduce Mlu1 (forward) and HindIII (reverse) restriction sites (forward, TAA TTA ACG CGT AGG TGG AGA GCC GGC AGG GT; reverse, TTA TAA GCT TCT CGG ACG TGG CCT AGG). This PCR fragment was cut with the appropriate restriction enzymes and inserted directly into the pGL2-Basic vector (TP99; Promega). Afterward, 200 bases were removed from the 5' end of the insert using the restriction enzymes FspaI and Mlu1. Similarly, the second part of the promoter was amplified using primers to introduce new Mlu1 (forward) and EcoRV (reverse) restriction sites (forward, TAT TAC GCG TTG CGC CCC CCT CCC GCC GTC CTA; reverse, CAT CGA TAT CCT GCG TTC TCT GGA GTC). This amplicon was cut with the appropriate restriction enzymes and ligated to the vector containing the 5'-UTR region. The integrity of the final constructs (wild type and mutant) was confirmed by direct sequencing of the final vectors. The activity of the mutant and wild-type KCNJ11 promoter construct was assayed as described earlier for the ABCC8 promoter construct, using the same transfection protocol, except that 1 µg of test plasmid and 0.5 µg of control plasmid were used to transfect 0.5 x 106 INS1r9 cells (kindly provided by Wang and Wollheim, Geneva, Switzerland).
Construction of ABCC8 and KCNJ11 plasmids for in vitro activity studies
Plasmids containing ABCC8 cDNA plus a myc-epitope and KCNJ11 cDNA were made as previously described (23). Point mutations in ABCC8 were introduced into hamster cDNA in the pECE vector (24), using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). The integrity of the constructs was confirmed by restriction mapping and sequencing. KATP channels were reconstituted by coelectroporation of 8 µg ABCC8 and 1 µg KCNJ11 plasmid into 7 x 106 COSm6 cells, as previously described (23).
Photolabeling and surface expression
After an overnight incubation after electroporation, whole-cell photolabeling (25) and surface expression assays were performed as previously described (23). Briefly, for whole-cell photolabeling, cells were incubated for 30 min with 1 nM [125I]-azido-glibenclamide in the presence and absence of 1 µM cold glibenclamide and cross-linked using UV light. Cells were then washed with PBS, solubilized with loading buffer [pH 6.8; 30 mM Tris, 1.25% sodium dodecyl sulfate (SDS), 1 mM EDTA, 50 mM dithiothreitol; 10% glycerol], and resolved in a 7–10% SDS gel.
Surface expression was quantified by using a luminometer-based assay to measure the SUR1c-myc-tagged subunit. Transfected COSm6 cells were gently washed in PBS and incubated for 3 h at 4 C using a mouse monoclonal IgG1 c-myc antibody (9E10; Santa Cruz Biotechnology, Santa Cruz, CA) diluted in DMEM plus 10% fetal bovine serum to a concentration of 0.3 µg/ml. After incubation, the cells were washed three to four times with PBS containing 1 mM CaCl2 and 1 mM MgCl2 and incubated for 1 h with horseradish peroxidase-conjugated goat antimouse IgG (Sigma Co., St. Louis, MO). Chemiluminescence was then measured using the Wallac 1429 VictorII (Perkin-Elmer, Wellesley, MA) and Luminol (Santa Cruz Biotechnology) as the horseradish peroxidase substrate.
Rubidium efflux
To determine channel activity, KATP channels were reconstituted into COSm6 cells, and [86Rb+] efflux assays were performed 48 h after the electroporation as previously described (26, 27). Briefly, after overnight incubation with 1 µCi/ml of [86Rb+] at 37 C in DMEM supplemented with 10% fetal bovine serum, the cells were preincubated for 30 min at room temperature in Krebs Ringer bicarbonate buffer containing 1 µCi/ml [86Rb+], the metabolic inhibitor oligomycin (2.5 µg/ml), and 2-deoxyglucose (1 µM) in the presence or absence of 1 µM glibenclamide. The supernatant was aspirated at the end of the preincubation and at the indicated time points and replaced with fresh Krebs Ringer bicarbonate buffer containing oligomycin and 2-deoxyglucose but no [86Rb+]. At the end of the time course, the reaction was stopped with 0.3% SDS. Radioactivity in the cells and supernatants from each time point was determined using a 
-counter, and the results are reported as the cumulative counts in the aspirated solution divided by the total of the counts in the lysate and the solutions.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
ABCC8 (Table 2).
Two splice-site mutations, seven coding sequence mutations, and one promoter mutation were identified in ABCC8. None of these were present in any of the control subjects tested.
|
a) was previously described in patients from the United States, Germany, and Saudi Arabia and was shown to be responsible for 69% of HI-associated chromosomes in the Ashkenazi Jewish population (4, 28, 29). Previous studies, using intragenic single-nucleotide polymorphism (SNP) haplotype analysis, suggested that the mutation occurred de novo in each population (29). Haplotype analysis in patient 1 failed to show any similarity with any of the other reported haplotypes associated with this mutation (data not shown), thus indicating another de novo event.
The novel splice mutation in the splice donor site of intron 15 (2154 + 3ag) was found in two unrelated index patients of Irish origin (patients 6 and 10). Both were heterozygous for the mutation. Patient 6 inherited this mutation on the maternal allele and was found to have a second mutation (G1342E) on the paternal allele (Table 2
). Patient 10 has a haploidentical, affected sibling (Table 1). Both inherited this mutation on the paternal allele, and both inherited the same maternal allele on which no mutation was identified.
Two of the seven coding mutations were previously reported. One, R836X, was identified in a proband from a Bedouin family and had previously been found in three families of Mexican origin (30). Haplotype analysis using flanking SNPs revealed that a single founder mutation was present in all three Mexican probands, two with diffuse HI and one with focal HI. In contrast, in the Bedouin family, the SNP haplotype was entirely different, suggesting that there is no genetic connection between the founder mutations in the two ethnic groups (Fig. 1). After identifying this mutation in the index patient, two other affected members of the same Bedouin tribe were evaluated and found to be homozygous for the same mutation. This mutation was not identified in any of the 100 Israeli Arab control chromosomes tested.
|
One novel single base insert was identified in an Israeli Arab patient (1113 ins T). This mutation causes a frame shift and a stop codon at position 1113. If translated, the resulting truncated protein would not be expected to be functional.
Four novel missense mutations were identified and characterized in vitro (Table 2, and in Functional studies of missense mutations). In addition, a novel mutation was identified in the ABCC8 promoter (–64 gc) on the paternal allele of a single HI patient of Spanish descent. This variant has not been previously described in any HI or control subjects and was not found in 100 control chromosomes tested.
KCNJ11 (Table 2).
Two novel mutations were identified, a missense mutation at codon 254 (P254L) and a mutation in the 5'-UTR, 88 bases 5' of the transcription start site (+88 gt). Neither of these mutations was present in any of the control subjects tested. The missense mutation, P254L, was identified in the homozygous state in a single Arab patient, and the 5'-UTR mutation was found in the homozygous state in an Israeli Bedouin proband. The functional significance of each mutation is discussed in Functional studies of missense mutations and in Regulatory variants in ABCC8 and KCNJ11. Neither was identified in 100 Israeli Arab control chromosomes.
GLUD1, GCK, SCHAD. In four patients, no mutations were found in ABCC8 or in KCNJ11. In these four patients, the coding sequences of GLUD1, GCK, and SCHAD were sequenced, and no mutations were identified.
Functional studies
ß-cell electrophysiology studies.
Figure 2 summarizes the KATP channel activity in ß-cells obtained from normal human controls (islet transplant donors) and from HI patients for whom appropriate tissue was available. Patients 1, 3, 5, 6, and 10 showed marked defects in KATP channel activity. The recorded average peak currents were (percentage of control given in parenthesis) 0.4 ± 0.25 pA (n = 8, 1.8%), 0.8 ± 0.45 pA (n = 4, 3.7%), 0.7 ± 0.19 pA (n = 19, 3.5%), 1.4 ± 0.35 pA (n = 4, 6.6%), and 1.6 ± 0.04 pA (n = 5, 7.4%), respectively, compared with control values of 21.2 ± 5.5 pA (n = 134, 100%). Interestingly, in the ß-cells from patient 14, in whom no mutation was identified, the level of KATP channel activity was similar to that found in normal cells, 18.2 ± 4.2 pA (n = 4, 85.9%). In all the other patients, ß-cells were not available either because the patients did not undergo surgery (patients 9, 12, and 13) or the surgical specimen could not be transported to the laboratory in the United Kingdom (patients 2, 4, 7, 8, and 11).
|
g mutation on splicing in vivo, we studied frozen pancreatic tissue from patient 10 and preserved, cultured islets from both patient 10 and her haploidentical affected sibling. RT-PCR was performed, and all PCR products were sequenced. RT-PCR of cDNA generated from the frozen pancreas spanning the region including exons 14–22 revealed two distinct bands. One of these bands represented the normal expected fragment, whereas the other band was found to contain a fragment with a 410-base deletion that started at codon 703 at the end of exon 15 and ended at codon 840 in the middle of exon 21 (Fig. 3A). RT-PCR of cDNA from cultured islets gave variable results. In one experiment, two bands were seen, one was the normal transcript, whereas the second band contained a 258-base deletion extending from within exon 15 (codon 687) to within exon 19 (codon 773) (Fig. 3B). This variant was found only in the cultured islets obtained from the sibling. When the experiment was repeated, three bands were seen (Fig. 3C). Again, one of the bands was the normal transcript. The third band showed a deletion of all of exon 15, whereas the middle band on that gel contained a mixture of the two splice variants. This pattern was seen on RT-PCR of cultured islets from both patients (patient 10 and her sibling). On other occasions, only the normal transcript was amplified. The same PCR reaction on cDNA from control pancreases and from other HI patients consistently revealed only the normal transcript. It was not possible to determine the expression level of the normal ABCC8 transcript relative to that of the abnormal splice variants or to the transcript level normally found in healthy ß-cells prepared and processed the same way.
|
g in ABCC8 and +88 gt in KCNJ11. To determine the functional significance of these variants, each was PCR amplified and cloned into a luciferase reporter vector. Each variant significantly reduced reporter gene expression (40 ± 7.8% and 44 ± 5% of the wild type for the ABCC8 and the KCNJ11 variants, respectively; Fig. 4).
|
).
To determine the effect of the missense mutations on KATP channel expression, trafficking, and/or function, mutant SUR1 and KIR6.2 were transiently transfected alone or with their respective wild-type partner in COSm6 cells and analyzed by photolabeling, surface expression of tagged SUR1c-myc subunit, and rubidium efflux (30, 33). When the mutations in ABCC8 (G70E, G111R, R836X, G1343E, R1419H, and R1494W) were expressed alone, all except R836X were photolabeled with the high-affinity sulfonylurea ligand [125I]-azido-glibenclamide (data not shown). This indicated integrity of the glibenclamide-binding site of these mutant constructs. The missense mutation, R836X, caused a truncation that resulted in loss of the binding site for glibenclamide. All myc-tagged wild-type and mutant constructs, including the missense truncation mutation, expressed similar protein levels as determined by Western blot analysis (data not shown).
|
shows photolabeling of the reconstituted mutant and wild-type KATP channels. Wild-type KATP channels appear at the plasma membrane only if SUR1 and KIR6.2 are coexpressed and if the core protein is glycosylated to produce the mature form of the receptor. The upper two bands represent the core (140 kDa) and mature (150–170 kDa) form of the receptor, and the lower band corresponds to the inward rectifier (38–40 kDa), which is co-photolabeled when both subunits are coexpressed. Although this is not a quantitative assay, the photolabeling of the core form of the wild-type and mutant receptors appears to be very similar. In contrast, the mature form, which is indicative of channels trafficking at least to the Golgi, was present only in the G70E homozygote, in the G70E/R1419H compound heterozygote, and to a lesser extent, in G111R. In mutants G1343E, R1419H, and R1494W in ABCC8 and P254L of KCNJ11, only the core band was detected, indicating that these mutant channels are retained in the endoplasmic reticulum (ER). As shown in the lower part of Fig. 5A, all the mutants co-photolabeled the inward rectifier, indicating that although not present in the plasma membrane, both subunits do associate.
Figure 5B shows the results of the surface expression studies. This experiment clearly shows that there is a parallel correlation between the loss of the mature band and the level of channels that are expressed at the cell surface. The first three SUR1 mutant channels, G70E homozygote, G70E/R1419H compound heterozygote, and G111R, were expressed at the plasma membrane, albeit at much lower levels than the wild-type channel. The remaining mutants, G1343E, R1419H, R1494W in ABCC8 and P254L in KCNJ11, did not reach the cell surface at all, confirming the retention pattern observed by photolabeling. We then assayed for recombinant channel activity by analyzing glibenclamide-inhibitable [86Rb+] efflux (Fig. 5C). We confirmed that channel activity of the G70E and G111R mutants was decreased when compared with wild-type controls. No channel activity was detected from any of the mutant subunits retained in the ER. Patient 5 was a compound heterozygote for mutations G70E and R1419H. Coexpression of an equimolar mixture of the two mutants resulted in a proportional reduction in channel activity, as determined by [86Rb+] efflux and surface expression, compared with the control.
Patients in whom no mutant allele was identified
In four patients, no mutation was identified in ABCC8, KCNJ11, GCK, GLUD1, or SCHAD. Two of these patients (patients 12 and 14) had typical severe HI. Patient 12 presented both clinically and biochemically with a classical phenotype of severe HI and was unresponsive to diazoxide and only partially responsive to octreotide. Due to the difficulty in preventing recurrent hypoglycemia using medical therapy, near total pancreatectomy was recommended but was refused by the parents. The patient was successfully managed using an intensive medical treatment protocol that included continuous sc infusion of octreotide and glucagon, along with frequent feedings and continuous enteral feeding during the night. At the age of 20 months, another trial of diazoxide treatment was performed, and a partial response was obtained. Currently, at the age of 3.5 yr, the patient is developmentally normal, but he still requires treatment with continuous sc infusions of both glucagon (4 µg/kg·h) and octreotide (12 µg/kg·d), frequent feedings during the day, and continuous feeding during the night.
We were also not able to identify a mutation in patient 14. His presentation was also severe, drug unresponsive, and required two partial pancreatectomies to prevent hypoglycemia. ß-cells from this patient were isolated and shown to have normal KATP channel function. We also found that cytosolic Ca2+ concentrations in isolated ß-cells were not significantly elevated when compared with normal values (data not shown).
Patient 13 had a relatively mild form of the disease, which presented with general convulsions associated with hypoglycemia at 6 months of age. He showed good response to diazoxide and has been well controlled with treatment ever since.
Finally, patient 15 had neonatal-onset HI associated with intermittent, mild hyperammonemia that was responsive to diazoxide. The patient’s father was diagnosed with HI at the age of 30 yr, although in retrospect, he appears to have had episodes of mild, episodic, nonfasting hypoglycemia since birth. Despite the fact that the clinical presentation of diazoxide-responsive hyperinsulinism with hyperammonemia suggests the presence of a GLUD1 mutation, no sequence alteration was identified. At the age of 9.5 yr, patient 15 is treated with diazoxide (1.5 mg/kg·d). No hypoglycemic event has been documented in the past 4 yr.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
a, R836X, and R1494W) were previously reported in other ethnic groups. In the first two cases, intragenic SNP haplotype analysis excluded the possibility of founder mutation affecting multiple ethnic groups, whereas the third case was located in a region where two other mutations were previously described (30, 31, 32). Thus, these three regions of ABCC8 appear to be mutation hot spots. Eleven of the 15 patients studied required pancreatectomy to alleviate hypoglycemia. Channel function was studied in the ß-cells from six of these patients, and in five, all with ABCC8 gene mutations, a marked decrease in the expression and activity of KATP channel was demonstrated. In contrast, the sixth patient had normal channel function, and no ABCC8 or KCNJ11 mutation was identified. These findings confirm the correlation between ABCC8 and KCNJ11 mutations and defects in the ß-cell KATP channel.
One novel splice-site mutation was identified in two patients. By studying mRNA extracted from one patient’s pancreas and from cultured ß-cells from the two patients, we demonstrated that this mutation results in the production of at least three different splice variants as well as the normal transcript. Based on known characteristics of the mammalian splice mechanism, the specific splice variants that we observed cannot be readily explained. To exclude mutations on the second ABCC8 allele, exons 15–21 and their intron-exon boundaries were PCR amplified and sequenced from genomic DNA, and no evidence of additional mutations or deletions was discovered. Therefore, we conclude that this single-base change in the exon 15/intron 15 splice junction may have caused major changes in the splicing process, resulting in several alternative splice variants in the two probands. Apparently normal ABCC8 transcript was also consistently identified by PCR amplification of pancreas and ß-cell mRNA from these patients. However, it was not possible to determine the amount of normal ABCC8 transcript present, and thus it was not possible to compare it to the amount present in normal ß-cells prepared the same way. Similarly, it was not possible to determine whether this apparently normal mRNA was translated into any functional protein. However, because this mutation was associated with a complete loss of KATP channel function in patients 6 and 10 (22), we conclude that either very little normal ABCC8 transcript was produced or it was not correctly translated into functional protein.
Two novel mutations were identified in regulatory regions, one in the promoter of ABCC8, –64 cg, and the other in the 5'-UTR of KCNJ11, +88 gt. This is the first time that mutations in the regulatory region of these genes have been associated with HI. Functional analysis of these sequence changes revealed that each mutation decreased transcriptional activity in the model system by approximately 55–60%. Patient 9 was homozygous for the KCNJ11 promoter mutation. A decrease in transcriptional activity of only 55–60% may not be sufficient to cause disease; however, this activity was evaluated using a small fragment of the promoter region in a cell line derived from rodent insulinoma. In the human ß-cell, the degree of suppression of transcriptional activity may be quite different. Neither mutation was identified in 100 normal chromosomes or in other HI-associated chromosomes, indicating that these are not common polymorphisms; however, additional studies are needed to prove unequivocally that these mutations are, in fact, the cause for HI in these patients.
In patient 11, the mutation –64 cg was located on the paternal allele, and no mutation was identified on the maternal allele. This mutation is located in a putative SP1/AP2 binding site. Further studies are needed to determine how it affects binding to these, and perhaps other, transcription factors. It remains unclear how a single mutation with only a partial effect on transcription efficiency can cause HI because heterozygosity for mutations that completely eliminate channel function is not associated with clinical disease (34). This discrepancy may be partially explained by the limitations of the in vitro method used. Thus, it is possible that this mutation has a much more dramatic effect in the normal human ß-cell. Alternatively, the patient may have focal HI, although this seems unlikely because he underwent three surgeries, and greater than 95% of the pancreas was removed. No focal lesion was identified, and the histology was reported to be consistent with diffuse HI. However, if the focal lesion is not removed and identified, the histological differentiation between normal neonatal pancreas outside a focal lesion and diffuse HI, although possible, can be quite difficult (35). Unfortunately, the pathological specimen is not available for reevaluation. A third possibility, which can never be entirely excluded, is that the maternal allele contains a mutation outside the regions tested or carries a deletion of one or more exons.
We used photolabeling in combination with surface expression studies and rubidium efflux assays to determine the effect of missense mutations on protein expression, posttranslational processing, and cell surface expression. In the pancreatic ß-cell, the KATP channel is a hetero-octameric complex, [(SUR1:KIR6.2)]4, formed by four inward rectifiers and four sulfonylurea receptors that physically associate in a 1:1 stoichiometry (26). This requirement is tightly regulated by the presence of ER retention (36) and exit (23) signals that ensure that only properly assembled, full-length KATP channels reach the plasma membrane. All of the six novel missense mutations on ABCC8 and KCNJ11 exhibit altered trafficking of the KATP channel to the plasma membrane, suggesting that this is a common mechanism by which missense mutations result in altered or lack of channel activity.
The first five transmembrane domains (TMD) in the N-terminal region of SUR1 (TMD0) include the region that specifically associates with KIR6.2, enhancing surface expression and forming a mini KATP channel capable of bursting similar to native channels (37). Two HI-associated TMD0 missense mutations have been previously studied and have shown to cause a severe trafficking defect. Chan et al. (38) demonstrated that these mutations prevent the association of SUR1 with KIR6.2; however, this finding was not confirmed by others (39). In the current study, we describe two different missense mutations in the same domain, G70E and G111R. Although both have reduced trafficking through the Golgi, neither appears to affect association with KIR6.2. Thus, the specific mechanism that causes this functional defect is still unknown. Our recombinant approach did not allow us to quantitate the number of channels at the plasma membrane or to define functionality. However, in ß-cells derived from these patients at surgery, 44% of all experiments failed to identify any channel activity, and in the remaining 56%, only brief single-channel events were seen that were unresponsive to diazoxide or the intracellular addition of nucleotides. This confirms that few channels actually traffic to the plasma membrane and those that do are functionally abnormal.
In contrast, the three other ABCC8 missense mutations, G1343E, R1419H, and R1494W, cluster in the second nucleotide-binding fold. These mutations exert a profound effect by confining the channel to the ER. It is interesting that none of these mutations affect the association of SUR1 with KIR6.2, as shown in the photoaffinity labeling studies. Other mutations, such as a single amino acid deletion (40) or missense mutations (41, 42, 43), with altered trafficking cluster in the second nucleotide-binding fold and C terminus of SUR1. Several mechanisms have been proposed to explain the trafficking defect, including increased degradation (44), unmasking of the retention signal (42), or retention in the trans-Golgi network (41). At the present time, we cannot ascertain which mechanism accounts for the phenotype observed with our mutants. The KCNJ11 mutation, P254L, which is 137 amino acids from the C terminus, is located in the cytoplasmic part of the protein. It interacts with SUR1 but cannot exit from the ER. The mechanism is not yet understood.
Defects in intracellular trafficking appear to be common mechanisms by which missense mutations cause recessive HI. This can be explained by the hetero-octomeric structure of the mature channel because, if mutant proteins would be readily incorporated into the channels and transported to the plasma membrane, then in the heterozygous state, 15 of 16 membrane-bound channels would be expected to contain mutant SUR1. In this situation, a dominant-negative effect would be expected, and dominant or codominant inheritance would be observed (45).
In nine probands, two mutant alleles were identified, thus fully explaining the disease in these probands. In two probands, only a single mutant allele was found. In patient 10, the mutation 2154 + 3ag was found on the paternal allele, suggesting that the patient may have had focal HI; however, histological analysis revealed diffuse disease that was confirmed by studies of ion channel electrophysiology (22). Furthermore, the proband had a haploidentical sibling with similar clinical, histological, and functional findings. A defect in splicing was observed in the proband and the sister. A second proband (patient 6), who was heterozygous for the same mutation on the maternal allele, had a different mutation on the paternal allele, confirming that this is a recessive mutation. Dominant inheritance of disease was further excluded by the fact that the carrier parent of each proband was clinically healthy. The possibility that this mutation can express a phenotype in the heterozygous state (autosomal dominant with variable penetrance) cannot be rigorously excluded; however, for this patient, the most likely conclusion is that there is a mutation present on the maternal allele that failed to be detected, despite direct sequencing of the entire gene.
In four patients, no mutation was identified in any of the five genes. These patients may be the most interesting patients of the group because some or all of them may have disease caused by mutations in other genes. Identifying the genetic cause of disease in these patients will be particularly challenging because they do not have affected relatives, and thus, linkage analysis is not possible. Mutation analysis of all genes that could potentially be responsible for unregulated insulin release is not feasible because the candidate list is too long. None of the patients had unique clinical or biochemical features that might direct the search to a particular candidate gene or signaling pathway. Patients 12 and 14 are phenotypically similar to most patients with KATP channel mutations; they have severe HI that failed to respond to diazoxide treatment. Patient 14 required pancreatectomies, whereas patient 12 responded to very intensive medical management, including continuous sc infusion of octreotide and glucagon as well as strict dietary management. ß-cells isolated from the resected pancreas of patient 14 had normal KATP channels, confirming our genetic findings but, unfortunately, not providing evidence for any specific alternative mechanism causing hyperinsulinism.
Patients 13 and 15 have clinically milder disease that was responsive to diazoxide. Patient 15 had mild hyperammonemia. The father has a similar, albeit milder, clinical picture, suggesting autosomal dominant inheritance. This syndrome has previously been associated with mutations in GLUD1; however, no mutation was detected in this patient. MacMullen et al. (46) reported finding GLUD1 mutations in 50 of 65 probands with HI/hyperammonemia, suggesting that mutations in other genes can cause this apparently discrete clinical syndrome. Therefore, it is likely that mutations in another gene, perhaps one coding for another enzyme in the same metabolic pathway as GLUD1, could be responsible for this syndrome.
In conclusion, we performed sequence analysis on the entire coding region of ABCC8 and KCNJ11 in a cohort of 15 HI patients. In patients in whom no mutation was identified in these two genes, the three other genes known to be associated with hyperinsulinemic hypoglycemia were similarly sequenced. This is the first study to report mutation screening, using direct sequencing, of all five genes (ABCC8, KCNJ11, GCK, GLUD1, and SCHAD) in the same patient cohort. Nine novel mutations were identified, including two in regulatory regions, one in ABCC8, and the others in KCNJ11, demonstrating for the first time mutations in the regulatory regions of these genes associated with HI. Mutations were found in KATP channels in 11 of 15 patients. Four patients with no apparent mutations in any of the five genes tested were thus identified. These patients represent a new cohort of patients that will be studied to identify new genes responsible for this disease.
|
Footnotes |
|---|
Abbreviations: ER, Endoplasmic reticulum; HI, hyperinsulinism of infancy; KATP, ATP-sensitive K+ channel; SDS, sodium dodecyl sulfate; SNP, single-nucleotide polymorphism; TMD, transmembrane domain; UTR, untranslated region.
Received June 28, 2004.
Accepted September 9, 2004.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C James, R R Kapoor, D Ismail, and K Hussain The genetic basis of congenital hyperinsulinism J. Med. Genet., May 1, 2009; 46(5): 289 - 299. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. B T Christesen, N. D Tribble, A. Molven, J. Siddiqui, T. Sandal, K. Brusgaard, S. Ellard, P. R Njolstad, J. Alm, B. Brock Jacobsen, et al. Activating glucokinase (GCK) mutations as a cause of medically responsive congenital hyperinsulinism: prevalence in children and characterisation of a novel GCK mutation. Eur. J. Endocrinol., July 1, 2008; 159(1): 27 - 34. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Abdulhadi-Atwan, J. D. Bushman, S. Tornovsky-Babaey, A. Perry, A. Abu-Libdeh, B. Glaser, S.-L. Shyng, and D. H. Zangen Novel De Novo Mutation in Sulfonylurea Receptor 1 Presenting as Hyperinsulinism in Infancy Followed by Overt Diabetes in Early Adolescence Diabetes, July 1, 2008; 57(7): 1935 - 1940. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Hussain, S. E. Flanagan, V. V. Smith, M. Ashworth, M. Day, A. Pierro, and S. Ellard An ABCC8 Gene Mutation and Mosaic Uniparental Isodisomy Resulting in Atypical Diffuse Congenital Hyperinsulinism Diabetes, January 1, 2008; 57(1): 259 - 263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F.-F. Yan, Y.-W. Lin, C. MacMullen, A. Ganguly, C. A. Stanley, and S.-L. Shyng Congenital Hyperinsulinism Associated ABCC8 Mutations That Cause Defective Trafficking of ATP-Sensitive K+ Channels: Identification and Rescue Diabetes, September 1, 2007; 56(9): 2339 - 2348. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F.-F. Yan, C.-W. Lin, E. A. Cartier, and S.-L. Shyng Role of ubiquitin-proteasome degradation pathway in biogenesis efficiency of {beta}-cell ATP-sensitive potassium channels Am J Physiol Cell Physiol, November 1, 2005; 289(5): C1351 - C1359. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Marthinet, A. Bloc, Y. Oka, Y. Tanizawa, B. Wehrle-Haller, V. Bancila, J.-M. Dubuis, J. Philippe, and V. M. Schwitzgebel Severe Congenital Hyperinsulinism Caused by a Mutation in the Kir6.2 Subunit of the Adenosine Triphosphate-Sensitive Potassium Channel Impairing Trafficking and Function J. Clin. Endocrinol. Metab., September 1, 2005; 90(9): 5401 - 5406. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
