Abstract
Introduction: This is a report of a child with congenital hyperinsulinism associated with a loss-of-function variant in KCNE1. KCNE1 encodes a human potassium channel accessory (beta) subunit that modulates potassium channel Kv7.1 (encoded by KCNQ1). Loss-of-function pathogenic variants in either the KCNQ1 or KCNE1 genes result in long QT syndrome by causing prolongation in the action potential duration at the cellular level. In addition to long QT syndrome, the phenotype associated with loss-of-function pathogenic variants in KCNQ1 is characterized by postprandial hyperinsulinemic hypoglycemia. Case Presentation: Clinical data for the proband were extracted from the medical records. The proband presented with fasting hypoglycemia due to hyperinsulinism in early childhood as well as postprandial hypoglycemia triggered by carbohydrates and by protein. Whole-exome sequencing was undertaken in genomic DNA isolated from proband and both parents. Whole-exome sequencing revealed a variant in KCNE1 inherited from the father, who also has a history of hyperinsulinism. Both the patient and father were subsequently diagnosed with long QT syndrome. The proband and father underwent phenotype testing including fasting test, oral glucose tolerance test, oral protein tolerance test, and exercise tolerance test. Conclusions: This case illustrates that loss-of-function variants in KCNE1, similar to KCNQ1, are associated with a cardiac and a beta cell phenotype, and thus, this patient population should be screened for hypoglycemia, particularly in the postprandial state.
Genetic diagnosis for congenital hyperinsulinism (HI) remains unknown for slightly over half of all patients with HI who are unresponsive to the only FDA approved therapy.
The protein encoded by KCNE1 is an ion channel partner of Kv7.1, encoded by KCNQ1; KCNQ1 is reported to cause both long QT syndrome and postprandial hyperinsulinemic hypoglycemia.
Loss-of-function variants in KCNE1 may put patients at risk for subclinical hypoglycemia; a low threshold for a hyperinsulinemic hypoglycemia workup is warranted in these patients.
The phenotype described in this case report of both protein- and carbohydrate-induced hypoglycemia may be of diagnostic utility in patients with KCNE1.
Introduction
Congenital hyperinsulinism (HI) is the most frequent cause of persistent hypoglycemia in infants and children with an estimated incidence of 1:28,000 live births [1]. Among the known genetic causes of HI, inactivating mutations in the genes encoding the two subunits of the beta cell ATP-sensitive potassium channel are the most common defects [2]. Despite advances in testing, including multigene next generation sequencing panels that include all known HI-associated genes, a molecular diagnosis is not always identified, with negative genetic testing in 58% of diazoxide-responsive cases and 8% of diazoxide-unresponsive cases [3]. This suggests that there are unidentified genes that can cause HI. We report a phenotype of HI and long QT syndrome in a proband and her father associated with a loss-of-function variant in KCNE1.
Case Report
Consent
Written informed consent was provided by subjects or their parents for genetic and phenotype testing. These studies were approved by the Children’s Hospital of Philadelphia Institutional Review Board.
Case Description
The proband was a female born at 41 weeks gestation with a birth weight of 3.397 kg (87th percentile). Gestation was remarkable for polyhydramnios and excessive maternal weight gain of 25 kg, which distinguished this pregnancy from mother’s two previous pregnancies. Delivery was uncomplicated. Hypoglycemia was noted in the neonatal period but was managed with frequent feedings and was not evaluated further at that time. Developmental delays were first noted at 6 months of age. By 9 months old, a four-to-five month delay in global milestones was noted. Also at 9 months, the child was noted to be having episodes of unresponsiveness suspected to be due to seizures. At 11 months of age, she presented to an emergency room for evaluation of rotavirus-associated gastroenteritis. A point-of-care plasma glucose was 20 mg/dL (1.1 mmol/L). A blood sample was obtained at the time of spontaneous hypoglycemia and was consistent with a diagnosis of HI: plasma glucose 46 mg/dL (2.6 mmol/L), serum bicarbonate 17 mmol/L, plasma insulin 4.4 µU/mL (26.4 pmol/L), urine ketones negative, and a positive glucagon stimulation test with rise in glucose from 37 mg/dL (2.1 mmol/L) to 166 mg/L (9.2 mmol/L) in 30 min [4]. Therapy with diazoxide was initiated.
At 15 months of age, the proband underwent a planned hospitalization to evaluate the efficacy of diazoxide treatment and for additional phenotype testing, including an oral protein tolerance test. On a diazoxide dose of 9.8 mg/kg/day, she fasted for 16 h maintaining the plasma glucose above 70 mg/dL (3.9 mmol/L). An oral protein tolerance test revealed a decrease in plasma glucose of 30 mg/dL (1.7 mmol/L) consistent with protein-induced hyperinsulinemic hypoglycemia [5]. A repeat protein tolerance test done at 3 years of age on a diazoxide dose of 5.4 mg/kg/day confirmed protein-induced hypoglycemia with plasma glucose decreasing from 100 mg/dL (5.6 mmol/L) to 58 mg/dL (3.2 mmol/L), and the diazoxide dose was increased to 8 mg/kg/day. Evaluation at 5 years of age demonstrated ability to maintain plasma glucose greater than 70 mg/dL (3.9 mmol/L) while fasting for 18 h on a diazoxide dose of 5 mg/kg/day, with no protein-induced hypoglycemia. At 10 years of age, the proband was admitted for hypoglycemia in the setting of intermittent diazoxide adherence. Diazoxide was resumed at 5 mg/kg/day and later increased to 10 mg/kg/day due to multiple hospital readmissions for hypoglycemia. On the higher diazoxide dose, she had hyperglycemia with plasma glucoses greater than 200 mg/dL (11.1 mmol/L) 60 and 120 min postprandially, followed by hypoglycemia; therefore, the diazoxide dose was reduced again to 5 mg/kg/day. However, the pattern of postprandial hypoglycemia persisted unrelated to the composition of meals or diazoxide dose, with plasma glucose values as low as 30 mg/dL (1.7 mmol/L). The proband’s mother noted that treating her hypoglycemia with food resulted in initial resolution or hyperglycemia, though subsequently led to recurrence of hypoglycemia. In response to this pattern, her parents adjusted her diet to eat small frequent meals with an attempt to limit simple carbohydrates and use low glycemic carbohydrates, which helped mitigate hypoglycemia.
Family History
The proband’s father had been diagnosed with HI at age three (Fig. 1). He had a history of persistent hypoglycemia in the neonatal period, which manifested with seizures. He began treatment with diazoxide at 3 years of age and continued medication until age twelve. Treatment was stopped by his family due to concern about potential side effects. Off treatment, he had persistent hypoglycemia and had a seizure when his plasma glucose was recorded in the 20–30 mg/dL (1.1–1.7 mmol/L) range. As an adult, the proband’s father was overweight (BMI 36 kg/m2 at the time of phenotype testing) and had developed type 2 diabetes, corroborated by an oral glucose tolerance test completed under a research protocol with a fasting glucose of 191 mg/dL (10.6 mmol/L) and a 2-h glucose of 266 mg/dL (14.7 mmol/L). He is currently on treatment for diabetes with empagliflozin and dulaglutide.
Family pedigree. Squares, males; circles, females; black symbols, diagnosis of hypoglycemia and borderline prolonged QT interval; gray symbols, symptomatic, but hypoglycemia diagnosis not confirmed; n/M, KCNE1 variant positive; n/n, KCNE1 variant negative.
Family pedigree. Squares, males; circles, females; black symbols, diagnosis of hypoglycemia and borderline prolonged QT interval; gray symbols, symptomatic, but hypoglycemia diagnosis not confirmed; n/M, KCNE1 variant positive; n/n, KCNE1 variant negative.
The paternal grandmother had no diagnosis of hypoglycemia, but she reported symptoms consistent with hypoglycemia, specifically severe headaches if she did not eat frequently. There were no other family members with a history of hypoglycemia. The proband’s mother and both maternal grandparents had type 2 diabetes.
Genetic Evaluation
Sanger sequencing analyses for genes known to be associated with HI (ABCC8, KCNJ11, GLUD1, GCK, HADH, UCP2, HNF1A, HNF4A) were negative in peripheral blood DNA. Because of the paternal family history suggestive of a pattern of dominant inheritance, whole-exome sequencing was performed.
Whole-exome sequencing revealed sequence variants in KCNE1 and PKM2, present in both the proband and her affected father (Table 1; Fig. 1). A six nucleotide deletion with a six nucleotide insertion was identified in KCNE1: c.172_177delACCCTGinsCCCCCT which results in amino acid substitutions at residues 58 (p.Thr58Pro) and 59 (p.Leu59Pro), also referred to as p.Thr58_Leu59delinsPro-Pro and reported to result in a loss of KCNE1 function [6]. The KCNE1: c.172_177delACCCTGinsCCCCCT variant is not reported in gnomAD although missense variants affecting the same residues (p.Thr58Pro and p.Leu59Pro) are present in the database at a low allele frequency of <0.000014 and occur in cis in 4 individuals: three of European, non-Finnish, descent and one classified as “other” [7]. Of note, the proband’s family is also of European, non-Finnish, origin. Both the amino acid substitutions at residues 58 and 59 are predicted to be damaging by SIFT and PolyPhen2 [8, 9]. These two amino acid substitutions have been reported to be associated with long QT syndrome [10]. The PKM2 variant identified in this family results in a missense change in a minor transcript (NM_001206798/ENST00000449901.2, c.155c>t/p.Ile52Thr) that is not detected in RNA sequencing datasets from either infant or adult beta cells [11‒13], suggesting that it is not expressed in beta cells. The variant is a silent change in the canonical and all other transcripts. This variant is identified in individuals of European descent in population databases at a low allele frequency of 0.00008 [7] and is predicted to be benign/tolerated by SIFT and PolyPhen [8, 9]. Thus, the PKM2 variant can be excluded as a cause of the proband’s phenotype.
Proband and father gene variants
Gene . | Chr . | Position (GRCh37) . | Nucleotide substitution . | Amino acid substitution . | dbSNP Ref . | Allele frequency (gnomAD) . | SIFT . | PP2 . | Ensemble transcript ID . |
---|---|---|---|---|---|---|---|---|---|
PKM2 | chr15 | 72511331 | c.155t>c | p.Ile52Thr | rs764572218 | 23/282,794 = 0.00008 | Tolerated | Bening | ENST00000449901.2 |
KCNE1 | chr21 | 35821761 | c.172a>c | p.Thr58Pro | rs147187721 | 4/282,768 = 0.00001415 | Damaging | Probably damaging | ENST00000337385.3 |
KCNE1 | chr21 | 35821757 | c.176t>c | p.Leu59Pro* | rs141813529 | 4/282,814 = 0.00001414 | Damaging | Probably damaging | ENST00000337385.3 |
KCNE1 | chr21 | 35821756 | c.177g>t | p.Leu59Leu* | rs748857341 | 4/282,792 = 0.00001414 | Synonymous | Synonymous | ENST00000337385.3 |
Gene . | Chr . | Position (GRCh37) . | Nucleotide substitution . | Amino acid substitution . | dbSNP Ref . | Allele frequency (gnomAD) . | SIFT . | PP2 . | Ensemble transcript ID . |
---|---|---|---|---|---|---|---|---|---|
PKM2 | chr15 | 72511331 | c.155t>c | p.Ile52Thr | rs764572218 | 23/282,794 = 0.00008 | Tolerated | Bening | ENST00000449901.2 |
KCNE1 | chr21 | 35821761 | c.172a>c | p.Thr58Pro | rs147187721 | 4/282,768 = 0.00001415 | Damaging | Probably damaging | ENST00000337385.3 |
KCNE1 | chr21 | 35821757 | c.176t>c | p.Leu59Pro* | rs141813529 | 4/282,814 = 0.00001414 | Damaging | Probably damaging | ENST00000337385.3 |
KCNE1 | chr21 | 35821756 | c.177g>t | p.Leu59Leu* | rs748857341 | 4/282,792 = 0.00001414 | Synonymous | Synonymous | ENST00000337385.3 |
*Two nucleotide substitutions were identified in codon 59 CTG>CCT in cis. The effect of both nucleotides combined is a Leucine to Proline substitution at codon 59.
The proband’s paternal half-brother and her paternal grandmother carry the KCNE1 variant. The paternal grandmother also shares the PKM2 variant. The proband’s two full siblings do not share either the KCNE1 or the PKM2 variants.
Cardiac Phenotype
Once the proband’s KCNE1 variant was identified at 14 years of age, an electrocardiogram was performed and demonstrated a calculated QTc of 487 ms (normal <440 ms), consistent with the diagnosis of borderline long QT syndrome. Her echocardiogram was normal. She was started on beta blocker therapy with nadolol, though this was ultimately discontinued by her cardiologist approximately 1 year later when balancing the risks of side effects with her borderline prolonged QT interval. She now avoids QT prolonging medications but otherwise does not follow any cardiac restrictions.
The proband’s father had an electrocardiogram which revealed a borderline prolonged QT interval, but he has not been on treatment, and he has not experienced cardiac symptoms including syncope or arrhythmias. The paternal grandmother now has insulin-dependent type 2 diabetes. She has no known history of long QT syndrome or cardiac symptoms including syncope or arrhythmias, though it is unknown if she was ever evaluated. The proband’s paternal half-brother who also carries the KCNE1 variant has not been formally evaluated despite genetic results. It is unclear if he is symptomatic. There are no other family members with a known diagnosis of long QT syndrome, or history of family members with syncope, arrhythmias, or sudden death in the family.
Follow-Up
On follow-up at age 19 years, the proband had been off diazoxide for the prior 3 years. She reported that her plasma glucoses are largely above 70 mg/dL (3.9 mmol/L) with occasional postprandial hypoglycemia. Her BMI at that time was 25.3 kg/m2. She had mild protein-induced hypoglycemia (nadir plasma glucose 65 mg/dL [3.7 mmol/L], peak insulin 22.6 μIU/mL) and glucose-induced hypoglycemia (nadir plasma glucose 67 mg/dL [3.8 mmol/L], peak insulin 67.6 μIU/mL). She also had exercise-induced hypoglycemia (nadir plasma glucose 50 mg/dL [2.8 mmol/L], peak insulin 23.9 μIU/mL [166 pmol/L]) (Table 2). She was able to fast for 18 h maintaining glucose above 70 mg/dL (3.9 mmol/L), ending with a plasma glucose 131 mg/dL (7.3 mmol/L) and beta-hydroxybutyrate 0.3 mmol/L. Diazoxide was resumed at 7.5 mg/kg/day and nutritional guidance against eating protein in isolation was given. With these interventions, her HI is well controlled with only occasional hypoglycemia, no lower than 60 mg/dL (3.3 mmol/L).
Proband phenotype testing results
Elapsed time, min . | Plasma glucose, mg/dL (mmol/L) . | Plasma insulin, μU/mL (pmol/L) . |
---|---|---|
Oral glucose tolerance test | ||
0 | 81 (4.5) | 9.1 (63.2) |
30 | 158 (8.8) | 57.9 (402.1 |
60 | 164 (9.1) | 46.5 (322.9) |
90 | 176 (9.8) | 49.1 (341) |
120 | 194 (10.8) | 65.5 (454.9) |
150 | 176 (9.8) | 67.6 (469.4) |
180 | 148 (8.2) | 62.4 (433.3) |
210 | 113 (6.3) | 26.6 (184.7) |
240 | 95 (5.3) | 17.5 (121.5) |
270 | 70 (3.9) | 10.2 (70.8) |
300 | 67 (3.7) | 13.1 (91) |
Oral protein tolerance test | ||
−15 | 75 (4.2) | <2 (<13.9) |
0 | 74 (4.1) | <2 (<13.9) |
15 | 74 (4.1) | 22.6 (156.9) |
30 | 75 (4.2) | 17.1 (118.8) |
45 | 72 (4) | 10.5 (72.9) |
60 | 65 (3.6) | N/A |
90 | 72 (4) | 13 (90.3) |
120 | 73 (4.1) | 21.7 (150.7) |
150 | 69 (3.8) | 12 (83.3) |
180 | 68 (3.8) | 10.6 (73.6) |
Exercise tolerance test | ||
−10 | 66 (3.7) | 8.5 (59) |
0 | 61 (3.4) | 7.2 (50) |
10 | 58 (3.2) | 23.9 (166) |
20 | 50 (2.8) | 11 (76.4) |
30 | 50 (2.8) | 10.5 (72.9) |
40 | 50 (2.8) | 13.4 (93.1) |
60 | 53 (2.9) | 12.3 (85.4) |
Elapsed time, min . | Plasma glucose, mg/dL (mmol/L) . | Plasma insulin, μU/mL (pmol/L) . |
---|---|---|
Oral glucose tolerance test | ||
0 | 81 (4.5) | 9.1 (63.2) |
30 | 158 (8.8) | 57.9 (402.1 |
60 | 164 (9.1) | 46.5 (322.9) |
90 | 176 (9.8) | 49.1 (341) |
120 | 194 (10.8) | 65.5 (454.9) |
150 | 176 (9.8) | 67.6 (469.4) |
180 | 148 (8.2) | 62.4 (433.3) |
210 | 113 (6.3) | 26.6 (184.7) |
240 | 95 (5.3) | 17.5 (121.5) |
270 | 70 (3.9) | 10.2 (70.8) |
300 | 67 (3.7) | 13.1 (91) |
Oral protein tolerance test | ||
−15 | 75 (4.2) | <2 (<13.9) |
0 | 74 (4.1) | <2 (<13.9) |
15 | 74 (4.1) | 22.6 (156.9) |
30 | 75 (4.2) | 17.1 (118.8) |
45 | 72 (4) | 10.5 (72.9) |
60 | 65 (3.6) | N/A |
90 | 72 (4) | 13 (90.3) |
120 | 73 (4.1) | 21.7 (150.7) |
150 | 69 (3.8) | 12 (83.3) |
180 | 68 (3.8) | 10.6 (73.6) |
Exercise tolerance test | ||
−10 | 66 (3.7) | 8.5 (59) |
0 | 61 (3.4) | 7.2 (50) |
10 | 58 (3.2) | 23.9 (166) |
20 | 50 (2.8) | 11 (76.4) |
30 | 50 (2.8) | 10.5 (72.9) |
40 | 50 (2.8) | 13.4 (93.1) |
60 | 53 (2.9) | 12.3 (85.4) |
The initial developmental delays noted in infancy and early childhood appeared to improve upon treatment of hypoglycemia. She now attends college and does well in school. She has no noted cognitive deficits.
Mutation Analysis
Sanger Sequencing
Mutation screening was performed in a commercial laboratory for ABCC8, KCNJ11, GLUD1, and GCK. Additional mutation screening (UCP2, HADH, HNF1A, HNF4A) was performed on a research basis. Genomic DNA was isolated from peripheral blood (5 PRIME, Gaithersburg, MD, USA) or from saliva (Oragene DNA self-collection kit; DNA Genotek, Kanata, ON, Canada). Coding sequences and intron/exon splice junctions were amplified and directly sequenced on an ABI 3730 capillary DNA analyzer (Applied Biosystems, Carlsbad, CA, USA).
Whole Exome Sequencing
Genomic DNA was isolated from blood samples of the proband and both parents. Three micrograms of genomic DNA were enriched using the Agilent SureSelect Human All Exon 44 Mb kit (Agilent Technologies, Santa Clara, CA, USA) following the manufacturer’s protocol and subsequently sequenced using Illumina HiSeq2000 following the manufacturer’s instructions (Illumina, San Diego, CA, USA). Raw reads (paired-end, 2 × 90 bp) were aligned to the reference human genome (UCSC hg19) using the Burrows-Wheeler Aligner (BWA, 0.6.2, http://bio-bwa.sourceforge.net/). Optical and PCR duplicates were marked and removed with Picard. Local realignment of reads in the insertion/deletion (indel) sites and quality recalibration were performed with the Genome Analysis Tool Kit (version 2.1, http://www.broadinstitute.org/gatk/). Single-nucleotide polymorphisms (SNPs) and small indels were called with GATK UnifiedGenotyper and annotated with ANNOVAR (http://www.openbioinformatics.org/annovar/) and SnpEff (http://snpeff.sourceforge.net/). Initial variants were filtered using ANNOVAR following GATK best practice by excluding SNPs with “QD <2.0,” “MQ <40.0,” “FS >60.0,” “HaplotypeScore >13.0,” “MQRankSum <−12.5,” “ReadPosRankSum <−8.0” and indels with “QD <2.0,” “ReadPosRankSum <−20.0,” “InbreedingCoeff <−0.8,” “FS >200.0.” De novo variants were also filtered for read depth <10. Variants were filtered out if they did not fit the model of dominant inheritance from the father or being a de novo variant in the child (done in the case of possible mosaicism in the affected father). Variants were also removed if they were present in public databases at an allele frequency greater than 0.5% (dbSNP135, 1000 genomes database, HapMap database) or greater than 10% in an in-house database which includes over 1,000 samples. Candidate genes were then prioritized using known HI genes (ABCC8, KCNJ11, GCK, HADH, INSR, GLUD1, SLC16A1) using Endeavour (https://endeavour.esat.kuleuven.be/), ToppGene (http://toppgene.cchmc.org/prioritization.jsp) and GeneDistiller 2 (http://www.genedistiller.org/). The functional consequences of identified variants were predicted with bioinformatics software SIFT [8], PolyPhen2 [9], Grantham, PhastCons, and Gerp. Next, variants were filtered based on the genes’ known biological functions to determine if mutations in these genes may be involved in insulin secretion or glucose metabolism. Variants identified by whole-exome sequencing were searched against the gnomAD Browser (v2.1) [7]. Variants were confirmed in the proband and parents by Sanger sequencing and screened in additional family members as available.
Phenotype Testing
Oral Glucose Tolerance Test
The proband’s father underwent an oral glucose tolerance test (oGTT) under a research protocol. A high concentration glucose drink (NERL™ Trutol™; ThermoFisher Scientific™, East Providence, RI, USA) containing 75 g of glucose was consumed after at least 3 h of fasting. Blood samples were taken at baseline and every 30 min for 4 h to measure plasma glucose and insulin. The proband underwent an oGTT as part of her clinical evaluation following a similar protocol. Results were interpreted according to standards from the American Diabetes Association [14].
Oral Protein Tolerance Test
The proband underwent an oral protein tolerance test (oPTT) as part of her clinical evaluation. For the test, she consumed a high protein load (Resource Protein Powder 1.16 g/kg, maximum 60 g) after at least 3 h of fasting. Plasma glucose and insulin were measured at −15, 0, +15, +30, +45, +60, +90, +120, +150, +180 min.
Exercise Tolerance Test
The proband underwent an exercise tolerance test (ETT) as part of her clinical evaluation. For the test, she performed moderately strenuous bicycle exercise for 10 min. Plasma glucose and insulin were measured at −10, 0, +10, +20, +30, +40, and +60 min.
Discussion
This report describes a phenotype of HI and long QT syndrome attributable to a sequence variant in KCNE1. The protein encoded by KCNE1 is an ion channel partner of Kv7.1, encoded by KCNQ1, which has previously been described to cause both long QT syndrome and postprandial hyperinsulinemic hypoglycemia [15, 16]. Interestingly, the proband and her father had both fasting and postprandial hypoglycemia. We demonstrated that the proband’s postprandial hypoglycemia is triggered by carbohydrates and by protein. This is in contrast to other forms of HI with postprandial hypoglycemia in which the trigger is either protein (including ATP-dependent potassium channel HI, glutamate dehydrogenase HI, and short-chain 3-hydroxyacyl-CoA dehydrogenase HI [5, 17, 18]) or glucose (including UCP2-HI [19, 20], hexokinase 1 HI [21], and phosphoglucomutase 1 deficiency HI [22]). Protein-induced hypoglycemia may be a more common manifestation of HI than currently perceived, which often persists long after improvement in fasting tolerance. Testing for protein-induced hypoglycemia with an oral protein tolerance test is a useful diagnostic tool that can also lead to effective therapeutic dietary modification by avoidance of ingesting protein in isolation. This is particularly true in cases of diazoxide-responsive HI without an identified genetic etiology, but with unexplained postprandial hypoglycemia. Similar to other forms of HI, the severity of the HI in these two individuals ameliorated with age; however, the proband continues to have postprandial episodes of hypoglycemia. The proband’s father now has type 2 diabetes; however, the association between his diabetes and the sequence variant in KCNE1 is not known.
KCNE1 encodes a potassium channel accessory (beta) subunit that modulates the voltage gated potassium channel Kv7.1 (encoded by KCNQ1) [23]. Loss-of-function pathogenic variants in both KCNE1 and KCNQ1 are associated with long QT syndrome. Torekov et al. [15] previously reported postprandial hyperinsulinemic hypoglycemia in individuals with long QT syndrome due to loss-of-function variants in KCNQ1. They found that study subjects experienced lower plasma glucose 3 h post glucose ingestion compared to controls, with significantly increased insulin response to oral glucose stimulation and significantly higher C-peptide and proinsulin levels. Subjects in that study who underwent a prolonged (6 h) oral glucose tolerance test demonstrated delayed postprandial hypoglycemia (plasma glucose below 70 mg/dL [3.9 mmol/L]) between 3.5 and 5 h post glucose ingestion, with nadirs 25 mg/dL (1.4 mmol/L) to 65 mg/dL (3.6 mmol/L), which was significantly different from matched controls. Continuous glucose monitoring revealed that the subjects with inactivating KCNQ1 variants spent more than 1 h per day with glucose below 70 mg/dL (3.9 mmol/L), and more than half an hour with glucose below 50 mg/dL (2.8 mmol/L), versus no hypoglycemia in controls. The hypoglycemic episodes occurred postprandially rather than in the fasting state, and were accompanied by autonomic and neurogenic symptoms including jitteriness, sweating, dizziness, and weakness [15]. In summary, KCNQ1 HI is characterized by postprandial hypoglycemia, which is often severe and symptomatic, and typically unrecognized prior to the evaluation.
While postprandial hypoglycemia is clearly described in patients with KCNQ1 HI, specific protein-induced hypoglycemia has not been reported. However, given our observations consistent with protein-induced hypoglycemia in KCNE1 HI, the possibility of protein-induced hypoglycemia in patients with KCNQ1 HI should be considered.
KCNE1 is expressed in pancreatic beta cells, as is KCNQ1 [24]. A study by Rosengren demonstrated that inhibition of KCNQ1 in human beta cells leads to enhanced insulin exocytosis, indicating KCNQ1 plays a direct role in insulin secretion [24]. Previous studies have reported conflicting findings regarding KCNE1 expression, some suggesting that KCNE1 is not expressed in beta cells [25]. However, recent evidence supports that KCNE1 is expressed in beta cells both during infancy and adulthood [11, 12]. There is heterogeneity regarding the relative amounts of KCNE1 expression, both among individuals and among beta cells within the same individual [11].
KCNE1 HI likely shares a similar mechanism to that of KCNQ1. This is supported by the similar phenotype involving postprandial hypoglycemia, particularly triggered by carbohydrates. In the beta cell, the closure of beta cell ATP-sensitive potassium channels resulting from an increased ATP/ADP ratio, leads to membrane depolarization, which triggers calcium influx to stimulate insulin secretion. We postulate that dysfunctional Kv7.1 channels resulting from mutations in either KCNQ1 or KCNE1 affect the capacity of beta cells to repolarize after fuel stimulation (Fig. 2). In fact, functional analysis of the KCNE1: c.172-177delACCCTGinsCCCCCT variant demonstrates a small defect in trafficking and a severely weakened interaction of KCNE1 with KCNQ1, resulting in an inability to modulate channel gating [26]. Similarly, in cardiac myocytes, loss-of-function variants in both KCNE1 and KCNQ1 lead to prolonged QT syndrome due to abnormal repolarization [10]. Other extracardiac manifestations of KCNE1 mutations have previously been described, as autosomal recessive inactivating mutations in KCNE1 result in Jervell and Lange-Nielsen syndrome 2, characterized by bilateral congenital deafness in addition to abnormal cardiac ventricular repolarization with prolonged QT interval [27].
Proposed mechanism of KCNE1 modulation of insulin release. Increased ATP/ADP in the beta cell resulting from glucose or amino acids metabolism leads to inhibition/closure of ATP-sensitive potassium channels, which triggers membrane depolarization and opening of voltage-dependent calcium channels. Increased cytosolic calcium triggers insulin secretion. The voltage gated potassium channel Kv7.1 (encoded by KCNQ1 with beta subunit encoded by KCNE1) is activated and plays a role in plasma membrane repolarization and cessation of insulin release. Positive effects noted as arrow, negative effects noted as minus arrow. Dashed arrows denote a multistep pathway. GDH, glutamate dehydrogenase; GCK, glucokinase; ATP, adenosine triphosphate; ADP, adenosine diphosphate; G6P, glucose-6-phosphate.
Proposed mechanism of KCNE1 modulation of insulin release. Increased ATP/ADP in the beta cell resulting from glucose or amino acids metabolism leads to inhibition/closure of ATP-sensitive potassium channels, which triggers membrane depolarization and opening of voltage-dependent calcium channels. Increased cytosolic calcium triggers insulin secretion. The voltage gated potassium channel Kv7.1 (encoded by KCNQ1 with beta subunit encoded by KCNE1) is activated and plays a role in plasma membrane repolarization and cessation of insulin release. Positive effects noted as arrow, negative effects noted as minus arrow. Dashed arrows denote a multistep pathway. GDH, glutamate dehydrogenase; GCK, glucokinase; ATP, adenosine triphosphate; ADP, adenosine diphosphate; G6P, glucose-6-phosphate.
Our findings have important implications for care of patients with KCNE1 pathogenic variants. While KCNE1 is recognized to cause long QT syndrome, patients may also have subclinical hypoglycemia, or attribute hypoglycemia symptoms to their cardiac dysrhythmia and remain unaware of and at risk for recurrent hypoglycemia. The recent report of patients with KCNQ1 long QT syndrome had both subclinical and symptomatic postprandial hypoglycemia [15], though none of the patients were previously clinically evaluated and thus none were previously aware of their hypoglycemia. Thus, physicians caring for patients with a diagnosis of long QT syndrome should maintain a high index of suspicion for concurrent hypoglycemia, and screen for it should symptoms arise. Conversely, clinicians should consider the possibility of this form of HI and its association with long QT syndrome when caring for patients with protein- and carbohydrate-induced hypoglycemia without an identified genetic cause. These patients may benefit from screening with an electrocardiogram.
This case highlights the importance of recognizing and treating hypoglycemia to prevent neurological symptoms and neurodevelopmental brain damage, as illustrated by the proband’s developmental delays as well as her father’s seizure activity. Recognizing that inactivating variants in KCNE1 cause HI, patients with these variants should be monitored for hypoglycemia, especially postprandially. Interventions can be taken to prevent recurrent hypoglycemia once detected. Furthermore, the described phenotype of both protein and carbohydrate-induced hypoglycemia, may be of diagnostic utility in cases of HI, which can prompt targeted testing for KCNE1 and KCNQ1, and ultimately uncover cardiac comorbidities if not yet established.
Acknowledgments
We would like to thank the proband and her family for their participation and contribution to this research.
Statement of Ethics
This study protocol was reviewed and approved by Committees for the Protection of Human Subjects at The Children’s Hospital of Philadelphia, Approval Nos. 07-005772 (research genetic testing for proband and father) and 09-007046 (phenotyping for father). Written informed consent was obtained from the proband’s parent and the father provided his own written informed consent to participate in the study. Written informed consent was obtained from the proband’s parent and the father provided his own written informed consent for publication of the details of their medical cases.
Conflict of Interest Statement
D.D.L. has received research funding from Hanmi Pharmaceuticals, Zealand Pharma A/S, Eiger Pharma, Twist Biosciences, Rezolute, Ultragenyx, and Crinetics Pharmaceuticals for studies not included in this manuscript. D.D.L. has received consulting fees from Zealand Pharma A/S, Crinetics Pharmaceuticals, Hanmi Pharmaceuticals, Eiger Pharma, Twist Biosciences, Spruce Biosciences, and Rhythm Pharmaceuticals not related to this manuscript.
Funding Sources
This work was supported by National Institutes of Health grant awarded to D.D.L. and A.G. [R01-DK056268]. Additional support was provided by the CHOP Center for Human Phenomic Science supported by the National Center for Advancing Translational Sciences, National Institutes of Health [UL1TR001878].
Author Contributions
W.S. acquired data and wrote the manuscript; K.E.B. and A.G. acquired and analyzed data and edited the manuscript; L.M.M. acquired data and edited manuscript; C.A.S. analyzed data and edited the manuscript; D.D.L. designed and analyzed experiments and edited the manuscript.
Data Availability Statement
Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request. Restrictions apply to the availability of some, or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.