Long-Term Seizure and Neurodevelopmental Outcomes of Three Children with Developmental Delay, Epilepsy, and Neonatal Diabetes (DEND) Syndrome after Early Initiation of Sulphonylurea

Abstract

Developmental delay, epilepsy, and neonatal diabetes (DEND) syndrome, caused by de novo mutations in ABCC8/KCNJ11 genes encoding Adenosine Triphosphate (ATP)-sensitive potassium channels, is an unusual cause of infantile-onset developmental/epileptic encephalopathy. Here, we report the long-term seizure and neurodevelopmental outcomes of three children with pathogenic variants in ABCC8/KCNJ11 genes and the phenotypic spectrum of DEND syndrome. Genetic confirmation was followed by an immediate therapeutic switch from insulin to sulfonylurea in all three children. At the last visit, all these children had good seizure control. However, all of them had residual neurodevelopmental impairments of varying clinical severity. Further large-scale, prospective, multicenter cohorts might be needed to clearly estimate the effect of early initiation of sulphonylurea on long-term seizure and neurodevelopmental outcomes of this rare syndrome.

Introduction

Neonatal Diabetes Mellitus (NDM) is defined as persistent hyperglycemia in the first 6 months of life, lasting for more than 2 weeks and requiring insulin therapy.[1] The incidence of NDM is estimated to vary from 1/300,000 to 1/500,000 live births.[2] Specific mutations in genes encoding Adenosine Triphosphate (ATP)-sensitive potassium (K⁺ ATP) channels might lead to transient NDM (TNDM), permanent NDM (PNDM), and a multisystemic clinical syndrome characterized by developmental delay, epilepsy, and neonatal diabetes (DEND).[3,4,5]

DEND syndrome is a rarely described entity.[6] It is caused by mutations in KCNJ11 or ABCC8 genes encoding K⁺ ATP channels. The dysfunction of these genes, which are expressed both in the endocrine pancreas and brain, might have multiple downstream effects, resulting in severe cognitive, behavioral, and neurological deficits. Most of the earlier case reports in the endocrine literature described the baseline clinical phenotypes and the genetic defects, without describing the long-term seizure and neurodevelopmental outcomes. We describe the baseline clinical characteristics and long-term seizure and neurodevelopmental outcomes of three mutation-proven cases of DEND syndrome with variable clinical severity.

Methods

The ethical clearance for publication was obtained from the Institution Ethics Committee (Reference no: IEC-AIMS-2018-PEDNEURO-109, dated 7/8/2018). Informed consent was obtained from patients. The study presented the long-term seizure and neurodevelopmental outcomes of three children with pathogenic variants in ABCC8/KCNJ11 genes and the phenotypic spectrum of DEND syndrome.

Case 1

A 5-month-old boy, born of a nonconsanguineous marriage, was initiated on Adrenocorticotrophic Hormone (ACTH) therapy at an outside hospital for infantile spasms. He was referred to our centre in view of very high blood glucose (>400 mg/dl) after the initial control of spasms. He was started on insulin, with brittle glycemic control. He also had recurrence of spasms later, and a better control of spasms was obtained after the initiation of vigabatrin. Genetic analysis (Exeter, UK) revealed an ABCC8 missense pathogenic variant, p.Q211K (clinvar VCV000021169.2), confirming the clinical diagnosis of DEND.[7] He was changed over to glibenclamide (0.7 mg/kg), after which he had very good glycemic control. He also became seizure-free, with normalization of the interictal Electroencephalogram (EEG). He had recurrence of seizures at 21 months of age, following which he was re-initiated on sodium valproate. His EEG showed bilateral frontotemporal epileptiform abnormalities. He had no further seizures on follow-up. He started walking by around 2 years with language delay and autistic features. During the last follow-up, at 5½ years of age, he was on valproate and glibenclamide and was seizure-free, with good glycemic control (HbA1c 6.6%). Bilateral frontotemporal spikes were noted on the EEG. He needed minimal help in activities of daily living and was undergoing occupational and speech therapies. On the Vineland Social Maturity Scale, his social age was 24 months, with a social quotient of 38, indicating moderate deficits in social-adaptive functioning. On the Child Behaviour Checklist, high scores on attentional problems and aggressive behavior were noted.

Case 2

This girl was admitted to the local hospital with excessive irritability, seizure, and lethargy at 45 days of life. She was diagnosed with diabetic ketoacidosis and was started on insulin. Magnetic Resonance Imaging (MRI) of the brain showed white matter hyperintensities in T2W images suggesting periventricular leukomalacia. ABCC8 missense pathogenic variant, p.R1183W (clinvar VCV000210076.4), was detected.[8] Her father, who was diagnosed with diabetes at the age of 26 years, was also found to have the same variant. She was shifted from insulin to sulphonylurea (glibenclamide) at an initial dosage of 0.05 mg/kg/day. She presented to our department with a history of one episode of seizure at 16 months of age. EEG showed bilateral parieto-temporal spikes, and she was started on levetiracetam. She had significant delay in all milestones in the initial phase, with a later catch-up. She had recurrence of seizure at 4.5 years of age, and her EEG then showed bilateral temporoparietal spikes. She remained seizure-free thereafter, and levetiracetam was discontinued at 7 years of age. Glibenclamide was also stopped at 5 years of age. During the last follow-up, at 8 years of age, she remained normoglycemic and seizure-free. She had below-average scholastic performance, with an Intelligence Quotient (IQ) of 85.

Case 3

This girl was born at 34 weeks of gestation to nonconsanguineous parents after treatment for infertility. Her mother had gestational diabetes and was on metformin in the second trimester. On the 40^th day of life, she had a seizure following a respiratory illness. She was found to have hyperglycemia and metabolic acidosis on evaluation. USG of the cranium, EEG, septic, and metabolic studies did not show any abnormalities. Her blood sugars were controlled with insulin. She was also initiated on phenobarbitone. Genetic evaluation showed a definitive KCNJ11 gene exon1, c.601C>T; p. Arg201 Cys (p.R201C) heterozygous missense pathogenic variant (clinvar VCV000008668.15-4/6). She was switched over to glibenclamide. On follow-up at 6 months, her EEG and MRI brain were normal. She had no seizure recurrence; hence, phenobarbitone was discontinued after 6 months of age. However, significant global developmental delay was noted in early childhood. At the last follow-up at 8 years of age, she had below-average intelligence, with an IQ of 82. She had scholastic difficulties, with problems in attention, concentration, and perceptual functions. She was seizure-free without regular intake of any antiseizure medications. She was continued to be on sulphonylurea, with good glycemic control. A summary of the clinical details and outcomes of these three patients is provided in Table 1.

Table 1.

The baseline clinical characteristics of the cohort with genetic variants and the long-term outcomes

	Case 1	Case 2	Case 3
Age at presentation/sex	5 months/male	1.5 months/female	40 days/female
Seizure onset	5 months	1.5 month	40 days
EEG findings (interictal)	Parieto temporal spikes	Parietotemporal spikes	Normal
MRI findings	Normal	White matter hyperintensities	Normal
Genetic mutation analysis	ABCC8 missense mutation, p.Q211K, of SUR1 subunit of K+ATP channel	ABCC8 missense pathogenic variant, p.R1183W	KCNJ11 heterozygous missense mutation in exon 1, c.601 C>T; p.Arg201Cys (p.R201C)
Pathogenicity	Pathogenic	Pathogenic	likely pathogenic
Family history	Maternal grandmother and paternal grandfather Type 2 DM. Mother had gestational diabetes	Father diabetic from 26 years of age. Same genetic mutation detected in father.	Father had type 2 DM from 46 years of age, Mother had gestational diabetes.
Developmental delay	Motor and language delay Autistic features/hyperactivity	Motor and language delay Mild hyperactivity/inattention	Motor delay, mild language delay. Mild hyperactivity with good social interactions.
Gross motor	Walking by 1 year 10 months	Started walking by 2 years of age.	Predominantly motor delay; attained head control by 7 months, and started walking by 2.5 years.
Language	Vocabulary of 150 words, able to combine two words, but occasionally told short sentences	She started using two-word sentences by 2.5 years of age. and short sentences by 3.5 years of age.	Attained language milestones after initial delay. Vocabulary 100–150 words by 2.5 years, short sentences by 3.5–4 years
Social and adaptive	Good bladder and bowel control by 4–5 years of age	She became toilet trained at 3–3.5 years of age	Toilet trained by 2.5–3 years of age
Last seizure (age)	1 year 9 months	4 years 6 months	2 months
Antiseizure medication	Sodium valproate	Levetiracetam, stopped at 7 years of age	Phenobarbitone, stopped at 6 months of age
Age at last follow up	7 years	8 years	8 years
Medications at last follow-up: OHA	Sulphonylurea–glibenclamide	Sulfonylurea (glibenclamide), stopped after 5 years of age; at last follow-up not on OHA	Sulphonylurea – glibenclamide
Antiseizure medication at last follow up	Sodium valproate	Nil	Nil
Developmental outcome	SQ of 38 at 5.5 years of age Moderate deficit in social-adaptive functioning	IQ 85	IQ 82 (assessed at 8 years of age)

DEND: developmental delay, epilepsy, and neonatal diabetes, DM: diabetes mellitus, EEG: Elecroencephalogram, IQ: Intelligence Quotient, OHA: Oral hyopglycemic agents, SQ: Social Quotient, SUR1: sulfonylurea receptor 1

Discussion

Developmental/epileptic encephalopathies (DEE) are conceptually defined as a group of predominantly monogenic disorders, characterized by a varying clinical spectrum of seizures, EEG abnormalities, and neurodevelopmental impairments. In this framework, developmental impairment is primarily attributed to the underlying, often genetic, etiology. Ongoing epilepsy might also contribute to adverse neurodevelopmental outcomes.[9] DEND syndrome is a rare form of DEE, associated with NDM.[5] Most of these children have been found to have underlying mutations in the genes encoding the functioning of K⁺ ATP channels.

Glucose metabolism generates ATP, which results in the closure of K⁺ ATP channels in the cell membrane of the beta-cells of the pancreas. This prevents K⁺ ions from leaving the intracellular compartment, causing membrane depolarization, which in turn opens voltage-gated calcium channels and allows calcium to enter the cell. The resultant increase in cytosolic calcium initiates exocytosis of insulin granules. KCNJ11 encodes the inner subunit (Kir6.2) of the K⁺ ATP channel, while ABCC8 encodes the outer subunit (SUR1). Mutations in any one of these genes cause the K⁺ ATP channels to remain inappropriately “stuck open,” even in the presence of hyperglycemia. As a result, the cell membrane is not able to depolarize effectively, leading to impaired release of insulin from the beta-cells.[2,4]

K⁺ ATP channels are also expressed in the cerebral neocortex and hippocampus. Mutations affecting the normal activity of these channels might also result in a spectrum of neurodevelopmental impairments and early-onset seizures.[2] However, there is very little data on the long-term neurological and seizure outcomes of these affected children.

In one of the largest studies of NDM across 21 countries, involving 239 unrelated patients, 31 (26%) were noted to have a mutation in KCNJ11 gene. Of these, six children were observed with developmental delays, and five belonged to the clinical spectrum of DEND.[10] Several isolated cases were also reported from other populations.[5,11,12,13,14,15,16,17,18,19,20] The most common mutations associated with developmental delay in NDM were V59M and R201H in the KCNJ11 gene.[11,15,21] The most common mutation observed in DEND syndrome was C166Y in KCNJ11 gene.[11,12,14] Several novel mutations have also been reported recently.[11,17,21] In another recently reported multicentric cohort study of NDM, 64% of patients had neurological abnormalities, which included developmental delay, learning difficulties, Attention Deficit Hyperactivity Disorder (ADHD), autistic features, and epilepsy, on a median follow-up of 10.2 years. Seventy-five of 81 subjects were on sulphonylurea in this cohort.[22]

This case series reports the long-term seizure and neurological outcomes of three children with the clinical syndrome of DEND, followed up prospectively after the relatively early introduction of sulphonylurea therapy. Sulphonylureas bind to the SUR1 component of K⁺ ATP channels and help better modulate the channel function.[10,11] Insulin therapy might control blood glucose levels to a certain extent. However, it might not ameliorate the consequences of the enhanced activity of extra-pancreatic K⁺ ATP channels, such as epilepsy, hypotonia, and developmental delay. Theoretically, sulfonylurea have the potential to modulate K⁺ ATP channels in extra-pancreatic tissues, in addition to beta cells. However, further studies are needed to determine the exact extent of blood–brain barrier permeability of the various suphonylurea compounds.

Sulfonylurea (most commonly glibenclamide or its equivalent) is usually started at a dose of 0.05–0.1 mg/kg/day in two divided doses, overlapping with insulin, and gradually uptitrated to a dose of 1 mg/kg/day, with gradual down titration of insulin over days or weeks, depending on whether the transition is done as an outpatient or inpatient. If the baby does not respond to the dosage of 1 mg/kg/day, it is usually considered a nonresponsive case, and the patient is switched back to insulin. PNDM will require lifelong treatment, whereas TNDM might initially remit by around 1 year of age and may later relapse in adolescence or young adulthood.[23]

The current cohort reports the long-term seizure and neurodevelopmental outcomes of three children with monogenic variations encoding K⁺ ATP channels, presenting with the clinical spectrum of DEND syndrome after the early initiation of sulphonylurea. This cohort adds to the existing literature on DEND, clearly demonstrating the phenotypic heterogeneity of this DEE, with diverging seizure, neurodevelopmental, and diabetic outcomes on long-term follow-up.

Conclusion

DEND syndrome is one of those DEEs where there is currently a possible option for mechanistically driven, targeted therapy. The current cohort reports the long-term seizure and neurodevelopmental outcomes of three children with monogenic variations encoding K⁺ ATP channels, presenting with the clinical spectrum of DEND syndrome after early initiation of sulphonylurea. Prospective, large, multicentric cohorts might be needed to exactly assess the impact of such early interventions on the long-term seizure, neurodevelopmental, and diabetic outcomes in this rare syndrome.

Author contributions

NB: Data curation; literature search writing initial draft; review and editing. KPV: Conceptulalization; writing initial draft; review and editing and will act as guarantor of the article. PP: conceptualization; review and editing. KB: Data curation. NB: Conceptulaization; review and editing. YBL: data curation. AGR: Conceptualization.

Conflicts of interest

There are no conflicts of interest.

Availability of data

Original data available and will be provided, if any, upon further request.

Funding Statement

Nil.