Family with autosomal dominant hyperinsulinism associated with A456V mutation in the glucokinase gene

Dear Sir, Autosomal dominant persistent hyperinsulinaemic hypoglycaemia of infancy is a rare disorder with only a few families described so far [1]. One of the genetic abnormalities underlying this disorder is a gain of function mutation in the gene encoding for glucokinase, one of the enzymes that regulates glucose-mediated insulin secretion in the pancreatic β-cell [2]. As yet, three gain of function mutations in this gene have been reported to be associated with familial autosomal dominant hyperinsulinism. First, Glaser et al. discovered a heterozygous single base-pair mutation at codon 455 of the glucokinase gene, resulting in the substitution of methionine for valine (V455M) [3]. Secondly, Christesen et al. reported another heterozygous mutation also in exon 10 of the glucokinase gene (A456V) [4]. Recently, a mutation (T65I) in exon 2 of the glucokinase gene was identified in a mother and son with the disease [5]. These gain of function mutations are remote from the glucose and adenosine triphosphate (ATP) binding site of this protein, and it has been suggested that glucokinase contains an allosteric activator domain [2, 4]. Here we report on a fourth nonconsanguineous family with autosomal dominant hyperinsulinism in association with a mutation in the glucokinase gene. Persistent hyperinsulinaemic hypoglycaemia of infancy was diagnosed in a 4-year-old, moderately obese boy (body mass index 16.7 kg m−2, i.e. between the 90th and 97th percentile for age) evaluated for convulsions that first occurred at an age of 3 years. On repeated occasions, fasting plasma insulin (124–173 pmol L−1, normal <69 pmol L−1) and fasting plasma C-peptide levels (0.267–0.374 nmol L−1, normal <0.112 nmol L−1) were elevated in the presence of nonketotic hypoglycaemia (blood glucose 0.9–2.1 mmol L−1). Plasma carnitine and ammonia levels were normal, and ketogenesis defects could be excluded. Until now, 7 years after diagnosis, the boy is succesfully treated with diazoxide (8 mg kg−1). His mother and grandmother suffered from occasional convulsions, whereas the deceased great-grandfather was said to have attacks of inappropriate behaviour that were relieved by eating. In the 29-year-old moderately obese mother (body mass index 27.5 kg m−2) blood glucose was 1.6 mmol L−1 with a plasma insulin level of 193 pmol L−1 and a C-peptide level of 0.424 nmol L−1, respectively, after a 40 h fast. The nonobese grandmother (body mass index 20.3 kg m−2), aged 57 years, had a blood glucose level of 1.8 mmol L−1, whilst plasma insulin was only 21 pmol L−1 and plasma C-peptide was mildly elevated (0.165 nmol L−1) after a 62 h fast. They were both diagnosed to have hyperinsulinaemic hypoglycaemia, and were treated by dietary counselling only. None of the 14 other family members over three generations suffered from convulsions, although some had vague symptoms of sweating and tremor. In 12 of them, blood glucose obtained after a 12 h fast ranged from 3.9 to 6.3 mmol L−1, whereas in two remaining subjects nonfasting blood glucose was 4.6 and 5.2 mmol L−1. Thus, the pedigree analysis and blood glucose measurement indicated an autosomal dominant mode of inheritance of persistent hyperinsulinaemic hypoglycaemia of infancy in this family. After written informed consent was obtained from each family member, analysis of exon 10 of the glucokinase gene was performed by cycle sequencing [3] on a Pharmacia ALF-Express sequencer. This analysis revealed a heterozygous single base C/T substitution at codon 456, resulting in a substitution of valine for alanine in the affected boy as well as in his mother and grandmother, but in none of the 14 other family members. Re-analysis confirmed the finding. Obviously, this single base substitution results in an aminoacid substitution (A456V) which is similar to that described by Christesen et al. [4]. Further endocrine studies were performed in the mother, the grandmother and an unaffected aunt. These studies were not performed in the affected boy, because no consent was obtained from his parents. As shown in Fig. 1, the mother showed increased plasma insulin and C-peptide levels obtained after 120 min moderate (8 mmol L−1) followed by 120 min severe hyperglycaemia (32 mmol L−1), achieved with a hyperglycaemic clamp protocol as described [6]. In the grandmother, plasma insulin and C-peptide responses during hyperglycaemia were within the reference range. However, it should be noted that plasma insulin and C-peptide levels were rather high with 55 pmol L−1 and 0.757 nmol L−1, respectively, in relation to a baseline blood glucose of 2.8 mmol L−1. From the plasma insulin and C-peptide response curve in the mother and grandmother, that also showed a first phase insulin and C-peptide increment, it is apparent that glucose-mediated insulin secretion by the pancreatic β-cell remained regulated. In comparison, in the report by Christesen et al. [4] the mother of their proband carrying the A456V mutation did not suffer from clinically manifest hypoglycaemia, nor showed exaggerated plasma insulin and C-peptide responses after an oral glucose load [4]. As the mother of our proband was moderately obese, it cannot be excluded that coexisting insulin resistance contributed to her enhanced insulin and C-peptide responses to hyperglycaemia. In addition, a C-peptide suppression test was performed by infusion of insulin (100 mU kg−1 h−1) during 120 min with blood glucose clamped at 2.1 mmol L−1 (i.e. around the predicted glucose threshold for stimulation of insulin release for the V455M mutation [3]). This test showed elevated plasma C-peptide levels in both the mother (0.281 nmol L−1, normal <0.08 nmol L−1 [7]) and the grandmother (0.212 nmol L−1). Plasma insulin and C-peptide levels in response to hyperglycaemia (Fig. 1), as well as plasma C-peptide suppression following insulin infusion (0.07 nmol L−1) were within the reference range in the unaffected aunt. Blood glucose, plasma insulin and C-peptide concentrations in response to 120 min of moderate (8 mmol L−1) and 120 min of severe (32 mmol L−1) hyperglycaemia in the affected mother and grandmother, and in an unaffected aunt of the proband. The stippled lines of plasma insulin and C-peptide responses represent the 95% confidence interval, obtained from previous experiments [5]. Note that the first phase insulin response is preserved in the mother and grandmother. The present investigations extend limited data with respect to the pathogenetic role of mutations in the glucokinase gene in the autosomal dominant form of persistent hyperinsulinaemic hypoglycaemia of infancy. Our results also underscore the phenotypic variability of this disorder [3–5], in which the clinical manifestations generally appear to be less severe than those associated with the more common autosomal recessive forms of this disease [8] that are due to mutations in the SUR1 and Kir6.2 genes [1]. No conflict of interest was declared.