Nesidioblastosis No Longer! It's All about Genetics

The past two decades have witnessed rapid advances in elucidating genetic forms of hyperinsulinemic hypoglycemia that have dramatically eclipsed our previous concepts about congenital hyperinsulinism in infants and children. Before this time, persistent hypoglycemia in infants, a disorder first described as “idiopathic hypoglycemia of infancy” in the mid-1950s (1), was attributed to nesidioblastosis, a supposed disturbance in embryonic morphogenesis in which insulin cells continued to bud off pancreatic ducts beyond the time of birth (2). This increase in β-cell mass was presumed to explain hypoglycemia in infants affected with nesidioblastosis. By the mid-1980s, however, studies by Rahier et al. (3) and other pathologists had convincingly proven that nesidioblastosis is merely a normal feature of the pancreas in early infancy. In pediatrics, use of the term has fortunately been abandoned. Interestingly, nesidioblastosis continues to be mentioned in adults as if it were still a specific entity. For example, patients with hypoglycemia post gastric bypass surgery have recently been labeled as having nesidioblastosis (4). The recognition of a genetic, as opposed to an embryological, basis for congenital hyperinsulinism evolved from the recognition of familial cases with either dominant or recessive patterns of inheritance (5). Currently, thanks to work by many investigators, eight different loci have been associated with hyperinsulinism: ABCC8, KCNJ11, HADH1, GCK, GLUD1, SLC16A1, UCP2, and HNF4a (Table 1). Mutations of these loci have significant differences in phenotype and inheritance pattern that can be helpful in differentiation and management. The first defects to be discovered, and still the most common genes associated with hyperinsulinism, involve the ABCC8 and KCNJ11 genes that are located together on the short arm of chromosome 11 and encode the two subunits of the β-cell ATP-dependent potassium channel: SUR1 and Kir6.2. Recessive mutations of these genes cause a severe form of neonatal hypoglycemia that frequently requires near-total pancreatectomy. In addition, these mutations are responsible for the focal form of congenital hyperinsulinism, caused by isodisomy for a paternally transmitted ABCC8 or KCNJ11 mutation. Dominant missense mutations of these genes have also been discovered that can be associated with either medically controllable or medically unresponsive hyperinsulinism (6, 7). Mutations in HADH1, which encodes SCHAD, a fatty acid oxidation enzyme, cause the only other recessive form of hyperinsulinism (8). Dominant, activating mutations of GCK, which encodes glucokinase, can cause hyperinsulinism that responds poorly to medical therapy. Other dominant forms of hyperinsulinism include activating mutations of GLUD1, which encodes glutamate dehydrogenase, a key step in amino acid-stimulated insulin secretion; inactivating mutations of UCP2, a mitochondrial uncoupling protein; activating mutations of SLC16A1 (encodes MCT1), a plasma membrane pyruvate transporter; and inactivating mutations of HNF4a, a nuclear transcription factor also associated with maturity-onset diabetes of the young 1, a form of monogenic diabetes of youth. Congenital hyperinsulinism genes Dom, Dominant; Rec, recessive. Congenital hyperinsulinism genes Dom, Dominant; Rec, recessive. Phenotypic features that distinguish some of these disorders can be helpful in suggesting the likely gene involved. Large birth weight is common in ABCC8, KCNJ11, GCK, and HNF4a defects. Hyperammonemia is a distinctive feature of GLUD1 mutations [“hyperinsulinism/hyperammonemia (HI/HA) syndrome”]. Protein-sensitive hypoglycemia occurs in HI/HA and SCHAD hyperinsulinism, due to hypersensitivity to leucine stimulation of insulin release. Protein sensitivity also occurs with mutations of ABCC8 and KCNJ11, but it is not due to leucine sensitivity. Hypoglycemia after anaerobic exercise is a dramatic feature in mutations of SLC16A1. Diazoxide, a drug that acts as an agonist of the ATP-dependent potassium channel to suppress insulin secretion, is effective in defects associated with mutations of GLUD1, UCP2, HNF4a, and HADH1. Diazoxide is often ineffective in mutations of the ATP-dependent potassium channel (ABCC8 and KCNJ11) and may not adequately control hypoglycemia in GCK or SLC16A1 mutations. Abnormal plasma acyl-carnitine mass spectrometry profiles are a feature of SCHAD hyperinsulinism (elevation of 3-hydroxy-butyrylcarnitine). A family history of maturity-onset diabetes of the young-type diabetes and the evolution of neonatal hyperinsulinism to diabetes later in adult life is a feature of HNF4a hyperinsulinism. Increased risk of diabetes in adulthood has also been suggested to be associated with mutations of ABCC8 and KCNJ11, although this remains controversial (9). In this issue of JCEM, Flanagan et al. (10) describe the ingenious application of homozygosity mapping to discover cases with SCHAD mutations and family histories of consanguinity. They point out that mutations of SCHAD may be a more common cause of hyperinsulinism responsive to diazoxide treatment than has been recognized previously. Abnormalities of acyl-carnitine profiles were not apparent in some of their cases. Although this finding was not verified, it is possible that metabolite testing may not be entirely reliable in detecting SCHAD hyperinsulinism. The mechanism of insulin dysregulation associated with SCHAD deficiency has recently been investigated using a SCHAD knockout mouse. Like affected children, SCHAD knockout mice are hypersensitive to protein-stimulated insulin secretion and show evidence of increased activity of glutamate dehydrogenase, similar to children with hyperinsulinism due to activating mutations of GLUD1 (11). These studies have revealed that SCHAD has a “moonlighting protein” function of binding to and inhibiting glutamate dehydrogenase; in the absence SCHAD, glutamate dehydrogenase activity is increased, resulting in hyperinsulinism and protein-sensitive hypoglycemia. Although these genetic discoveries about hyperinsulinism are of obvious interest for pediatric endocrinologists, they are also important for endocrinologists treating adults. Patients with mutations of GLUD1, GCK, SLC16A1, and some of the milder mutations of ABCC8 and KCNJ11 may escape recognition in infancy and first be discovered to have hypoglycemia as adults. Not infrequently, affected individuals have been diagnosed only after the birth of an affected child, or even a grandchild. Some cases might be sent for surgery, searching for an acquired insulinoma. Rather than labeling these cases as nesidioblastosis, careful consideration should be given to the possibility of a genetic disorder of insulin regulation. Disclosure Summary: The authors have nothing to declare. For article see page E498