Hyperinsulinism Resulting from Mutations in ABCC8 or KCNJ11 (Hyperinsulinemic Hypoglycemia, Familial Types 1 and 2) – K _ATP -HI

ABCC8 and KCNJ11 encode SUR1 and Kir6.2, respectively, which together form the subunits of the heterooctameric K _ATP channel in the β-cell plasma membrane. This channel is comprised of four inner Kir6.2 subunits, which form the channel’s ion pore, and four regulatory SUR1 subunits, which surround the Kir6.2 core. These subunits are assembled in the endoplasmic reticulum. Normal insulin secretion results from a fuel stimulated increase in the ATP/ADP ratio within the β-cell leading to closure of the K _ATP channel with resultant depolarization of the β-cell membrane and opening of a voltage-gated calcium channel with subsequent Ca ²⁺ influx. This influx leads to insulin degranulation and exocytosis. Inactivating mutations of either the K _ATP or SUR1 moieties of the K _ATP channel result in reduced or absent K _ATP channel activity, leading to permanent depolarization, spontaneous and independent opening of the voltage-gated Ca ²⁺ channels, and consequent constitutive insulin secretion irrespective of serum glucose.

Over 300 mutations in ABCC8 and over 30 mutations in KCNJ11 have been described in association with HI. Mutations may lead to an inability of the altered channel to exit the endoplasmic reticulum and be trafficked to the cell membrane, a disruption of the nucleotide-binding domain of SUR1 (i.e., lack of response to increased Mg-ADP concentrations ), an alteration of K _ATP channel density, or a complete abolishment of channel activity.

Recessively inherited K _ATP -HI is both the most common and most severe form of HI. Clinically, children present with severe hyperinsulinemic hypoglycemia, often, though not always, are in the neonatal period. The initial medical management for a child diagnosed with HI is with diazoxide, a medication that acts on the K _ATP channel to prevent closure of the channel. Diazoxide unresponsive HI suggests a K _ATP channel defect and a potential need for surgery. Genetic testing of the child and parents is recommended to further delineate the type of HI. K _ATP -HI is divided into two histopathological morphologies: diffuse disease and focal disease. Diffuse K _ATP -HI is, in most cases, due to autosomal recessive mutations in ABCC8 or, less commonly, KCNJ11 .

Children with a homozygous or compound heterozygous recessive mutation will have diffuse disease, whereas children with a paternally inherited heterozygous recessive mutation most likely have focal disease. Focal HI occurs as a result of inheriting a paternally derived mutation in ABCC8 or KCNJ11 and experiencing a somatic loss of the maternal 11p15 allele, resulting in paternal uniparental isodisomy. The maternal loss of heterozygosity entails loss of tumor suppressor genes, including CDKN1c [p57(kip2)] and H19; there is also upregulation of paternal IGF2. The combination of loss of tumor suppressor genes and upregulation of IGF2 leads to adenomatous expansion of the affected region. Thus, somatic regions within the pancreas in which a paternal mutation has been inherited and maternal loss of heterozygosity has occurred become foci of HI (focal adenomatosis). While focal and diffuse HI resemble each other clinically much of the time, a recent article reviewed 223 children with diffuse or focal HI and reported several significant clinical differences between the two (all values in parentheses are means and all differences reached statistical significance). Compared to children with focal disease, children with diffuse disease were born at an earlier age (38 weeks vs. 39 weeks) and were born with larger birth weights (3,963 vs. 3,717 g). Additionally, children with diffuse HI had higher insulin levels at the time of diagnosis (31.8 μU/mL vs. 12 μU/mL) and had higher maximal GIR requirements (19.2 mg/kg/min vs. 16.1 mg/kg/min). Children with focal HI presented at a later age compared to those with diffuse disease (0.3 months vs. 0 months) and were more likely to present with a seizure (50% vs. 25%). Distinguishing focal from diffuse disease is of utmost importance as focal disease can be cured by surgical resection of the lesion, whereas diffuse disease requires a near-total pancreatectomy. In addition to genetic testing, an F-DOPA PET scan can help in localization of a focal lesion. In the review of 223 patients, those with diffuse disease had a median percent pancreatectomy of 98%, compared to 27% in those with focal disease. Of the children that had surgery for diffuse disease, 41% had persistent hypoglycemia requiring treatment with enteral dextrose and/or octreotide. Comparatively, 94% of the children that had surgery for focal disease required no further management.

Autosomal dominant mutations in KCNJ11 and ABCC8 have been reported. They are characterized by normal subunit trafficking but impaired K _ATP channel activity, which overall can result in a milder hypoglycemia compared to the recessive mutations earlier, and many are responsive to therapy with diazoxide, unlike the recessively inherited mutations.

Hyperinsulinism Resulting from Mutations in GLUD-1 (Hyperinsulinemic Hypoglycemia, Familial 6) – GDH-HI

GLUD1 encodes glutamate dehydrogenase (GDH), a mitochondrial matrix enzyme that catalyzes the oxidative deamination of glutamate to alpha-ketoglutarate and ammonia with the assistance of either NAD ⁺ or NADP ⁺ as a cofactor. GDH is an important regulator of amino acid and ammonia metabolism; it is expressed at high levels in the brain, hepatocytes, pancreatic β-cells, and the kidneys, but not in muscle. GDH enzyme activity is allosterically inhibited by ATP and GTP and allosterically activated by ADP, GDP, and leucine.

GDH-HI results from gain-of-function mutations in GLUD1. The insulin secretion may be triggered by glucose or by amino acids, and is potentiated by leucine. Eighty percent of GDH-HI cases are due to de novo mutations in GLUD1; the remaining 20% are inherited in an autosomal dominant pattern. Over 20 disease-causing mutations have been identified. Most heterozygous missense amino acid substitutions in GLUD1 affect the GTP-binding site, decreasing the sensitivity of GDH to inhibition by GTP to varying degrees, increased sensitivity to leucine stimulation, and leading to excess insulin secretion; the degree of sensitivity to GTP inhibition is inversely related to severity of hypoglycemia phenotype. Other mutations affect the antenna region of GDH, which communicates with adjacent enzyme subunits. Increased GDH activity results in increased glutamate oxidation to alpha-ketoglutarate, which enters the Krebs cycle and leads to increased ATP levels. The increased ATP/ADP ratio leads to K _ATP channel closing and insulin secretion as discussed earlier. The mutations result in increased enzyme activity in pancreatic β-cells, hepatocytes, in the kidneys, and in the brain. These children also manifest asymptomatic persistent elevations in serum ammonia levels due to increased hepatic GDH activity (and with some contribution from the kidney ); thus, this syndrome is referred to as hyperinsulinism-hyperammonemia syndrome, or HI-HA.

Clinically, children present recurring episodes of both fasting and postprandial hypoglycemia, especially after consuming protein-rich meals. This hypoglycemia is typically milder than that seen in K _ATP -HI and is thus often not diagnosed at birth. It is responsive to diazoxide. Although ammonia levels are usually (though not always) quite elevated, typically two to five times above the usual normal limits, affected patients typically do not manifest hyperammonemia symptoms. That said, it should be noted that generalized seizures, especially absence seizures, may be seen in patients with GDH-HI even without concurrent hypoglycemia; this may be due to hypoglycemic brain damage, chronically elevated ammonia levels, or CNS hyperexcitability due to GDH mutations. The exact etiology of the seizures, however, is unclear. Additionally, children with HI-HA have a higher incidence of developmental delays and behavior problems.

Hyperinsulinism Resulting from Mutations in GCK (Hyperinsulinemic Hypoglycemia, Familial 3) – GK-HI

GCK encodes glucokinase (GK), an isoform of hexokinase expressed in pancreatic β-cells, hepatocytes, glucose-sensing hypothalamic neurons (predominantly in the ventromedial nucleus and arcuate nucleus), and some enteroendocrine cells. GK acts as a glucose sensor and catalyzes the initial and rate-limiting step in glycolysis, the conversion of D -glucose to D -glucose-6-phosphate at physiologic glucose concentrations using MgATP as a second substrate. GK exists in three different conformations, which control catalytic function, generating a sigmoidal response to plasma glucose concentrations; the transitions between these conformations are controlled by glucose concentrations and by allosteric interactions with modulators of GK activity such as biotin, cyclic AMP, and insulin, as well as glucokinase regulatory protein (GKRP). GK acts as the β-cell’s glucose sensor because it has a relatively low affinity for its substrate, glucose; thus, it determines the glucose threshold for insulin release. The typical set-point for glucose-induced GK expression in β-cells is ∼5 mmol/L, or 72 mg/dL.

Autosomal dominant activating mutations in GCK are relatively rare causes of congenital HI; as of this writing, only 12 activating mutations have been identified. That said, one recent multicenter study found that GCK mutations were responsible for up to 1.2% of cases of congenital HI and nearly 7% of non-K _ATP -channel cases of HI. These activating missense mutations typically occur in or near the allosteric binding sites for either GK activators or the competitive inhibitor, GKRP; the net result of activating mutations is to increase the affinity of GK for its substrate glucose, effectively lowering the threshold for insulin release, sometimes to as low as 1 mmol/L or 9 mg/dL. Of interest, mutations that conversely decrease GK’s binding affinity for glucose and increase the threshold for insulin secretion are responsible for a form of monogenic diabetes, MODY2; it is noteworthy that inactivating mutations are far more numerous than activating GCK mutations.

Clinically, the phenotype of GK-HI ranges from severe neonatal HI to mild fasting hypoglycemia, which may not be recognized until later in childhood or even into adulthood. Many patients will respond to diazoxide, but the degree of responsiveness can be quite variable. Patients may demonstrate an initially good response, but over time become less responsive. Those that do not respond to diazoxide or lose responsiveness over time may require a partial or near-total pancreatectomy in order to treat the HI.

Hyperinsulinism Resulting from Mutations in HADH (Hyperinsulinemic Hypoglycemia, Familial 4) – SCHAD-HI

HADH encodes medium/short-chain-3-hydroxyacyl-CoA dehydrogenase (SCHAD), a mitochondrial matrix enzyme that catalyzes the third and penultimate step in β-oxidation of fatty acids into acetyl-CoA, the conversion of 3-hydroxyacyl-CoA to 3-ketoacyl-CoA using NAD ⁺ as a cofactor. There are different isoforms of the various enzymes that catalyze the β-oxidation of medium- and short-chain fatty acids. In addition to its actions in the β-oxidation pathway, SCHAD also acts as an allosteric inhibitor of GDH, resulting in decreased affinity of GDH for its substrate, exerting an overall inhibitory effect on GDH in pancreatic islets resulting in excessive insulin secretion.

HADH loss-of-function mutations are a very rare cause of HI, although they are more common in children born to consanguineous parents and in children of Irish or Turkish populations, wherein there is a founder mutation. These mutations have an autosomal recessive inheritance mode. Several mutations have now been reported. The mechanism by which SCHAD mutations cause hyperinsulinemic hypoglycemia (as opposed to noninsulin mediated hypoketonemic hypoglycemia) entails loss of its inhibitory effect on GDH, leading to increased GDH affinity for leucine, alanine, and glutamate. Mechanisms by which the mutations cause SCHAD loss of function include causing deletions or alternate splice variants within the gene, mutations affecting protein folding, or deep intronic mutations introducing a frameshift in the gene.

Clinically, children present with hyperinsulinemic hypoglycemia ranging from mild to severe in degree, accompanied by elevated levels of plasma 3-hydroxybutyrylcarnitine in both the fasting and fed states and elevated urine 3-hydroxyglutaric acid. Unlike most other fatty acid oxidation disorders, children with SCHAD-HI typically do not have myopathies of skeletal and/or cardiac muscle and also typically do not have hepatotoxicity. SCHAD-HI is typically responsive to diazoxide.

Hyperinsulinism Resulting from Mutations in SLC16A1 (Hyperinsulinemic Hypoglycemia, Familial 7) – Exercise-Induced-HI

Proton-linked monocarboxylate transporter, member 1 (MCT-1) is an enzyme that catalyzes the movement of several monocarboxylates (e.g., lactate, pyruvate, valine, leucine-derived branched-chain oxo-acids, and ketone bodies) across the plasma membrane. It is widely expressed across multiple tissues, but its expression in pancreatic α- and β-cell membranes is typically low to absent. In addition, expression of lactate dehydrogenase-A, which is responsible for the conversion of pyruvate to lactate as the final step of anaerobic metabolism, is relatively low in β-cells; this helps channel pyruvate preferentially toward mitochondrial oxidative metabolism rather than toward lactate, and keeps glucose rather than pyruvate or lactate as the preferential trigger for insulin secretion. As a result of the low β-cell MCT-1 expression and LDH-A activity, MCT-1 activity does not ordinarily contribute to β-cell insulin secretion – in other words, changes in extracellular concentrations of lactate or pyruvate (e.g., during exercise or during a catabolic state) do not trigger insulin secretion.

In exercise-induced hyperinsulinism (EIHI, or MCT-HI), a gain-of-function mutation in the regulatory regions of SLC16A1 in the binding sites of several transcription factors allows for unusually high levels of expression of MCT-1 in β-cell membranes, enabling β-cell pyruvate uptake and pyruvate-stimulated insulin release in the presence of elevated lactate or pyruvate levels, such as in the setting of intense exercise or anaerobic exercise, leading to triggering of the usual K _ATP -mediated insulin secretory pathway and thus to ensuing hypoglycemia. Three different autosomal dominant mutations have been identified – two in Finnish families and one in a German family.

Clinically, patients present with hyperinsulinemic hypoglycemia of varying degrees in the setting of exercise; response to diazoxide is partial. Treatment involves limiting anaerobic exercise and frequent intake of carbohydrate-containing snacks during aerobic exercise.

Hyperinsulinism Resulting from Mutations in INSR (Hyperinsulinemic Hypoglycemia, Familial 5) – Insulin Receptor Activating Mutations

INSR encodes the insulin tyrosine kinase receptor, which is a heterotetrameric protein composed of two extracellular α subunits to which insulin binds and two transmembrane kinase β subunits. The binding of insulin to the α subunits stimulates the kinase, which triggers autophosphorylation of the β subunits on tyrosine residues, which in turn leads to the phosphorylation of a number of intracellular substrates, including the insulin receptor substrates. These phosphorylated substrates in turn activate two signaling cascades: the phosphotydilinositol-3-kinase-AKT/protein kinase B pathway, which mediates the majority of insulin’s metabolic actions, and the Ras-mitogen-activated protein (MAP) kinase pathway.

In 2004, an activating, autosomal dominant mutation in INSR was reported in three generations of a Danish family. This was a missense mutation in the tyrosine kinase domain of the insulin receptor gene. This is the only reported mutation in INSR linked to hypoglycemia – all other known pathogenic mutations cause variable degrees of insulin resistance and a diabetic phenotype.

Affected children and adults present with fasting hyperinsulinemia, elevated insulin: C-peptide ratios, and moderate to severe postprandial hypoglycemia coexisting in some cases with impaired glucose tolerance (first a significant rise in blood glucose levels postglucose challenge, then a lower-than-normal nadir). All affected patients manifested insulin resistance as measured by hyperinsulinemic-euglycemic clamp. Thus, the mutation caused a divergent phenotype of insulin resistance and insulin-mediated hypoglycemia. Patients responded to treatment with a long-acting formulation of octreotide, a somatostatin analog that acts downstream of the voltage-gated calcium channel in the β-cell to inhibit insulin secretion.

Hyperinsulinism Resulting from Mutations in HNF1A (Hepatocyte Nuclear Factor-1α)

HNF1A encodes the protein HNF-1α, a transcription factor required for the expression of several liver-specific genes. More specifically, the homodimer protein binds to a promoter sequence that is required for hepatocyte-specific transcription of hepatic-specific genes, including fibrinogen, alpha-1-antitrypsin and others. HNF-1α’s dimerization affects gene transcription.

HNF1A mutations are typically associated with a form of autosomal dominant monogenic diabetes (MODY3); however, in 2011, one case was described where an adult with MODY3 was found to have had fetal macrosomia and childhood hypoglycaemia, which self-resolved prior to developing diabetes mellitus at the age of 19. This mutation was heterozygous (implying dominant inheritance) and affected the HNF1A gene’s DNA-binding domain. Two other mutations have since been reported, one a heterozygous nonsense mutation affecting HNF1A ’s dimerization domain and another a heretozygous missense mutation. The mechanism by which loss of function mutations in HNF1A lead to hypoglycemia alone or the dual phenotype of hypoglycemia in infancy/childhood and monogenic diabetes in adulthood has not been well elucidated.

Clinically, the reported hypoglycemia phenotypes range from isolated incidents in childhood to full-blown neonatal HI. Variable degrees of responsiveness to diazoxide have been reported.

Hyperinsulinism Resulting from Mutations in HNF4A (Hepatocyte Nuclear Factor 4-Alpha)

HNF4A encodes the protein HNF4-α, a nuclear transcription factor that binds DNA (to HNF4-α response elements and with the involvement of coactivator peptides) as a homodimer. HNF4-α is the most abundant hepatic DNA-binding protein, influencing the expression of approximately 40% of actively transcribed hepatic genes, including HNF1A and other genes involved in gluconeogenesis and hepatic liver metabolism. HNF4-α is also expressed in the pancreas, and controls the expression of approximately 11% of pancreatic islet genes.

Most pathogenic mutations in HNF4A cause monogenic diabetes (MODY1). Heterozygous loss-of-function mutations in HNF4A have also been linked with hyperinsulinemic hypoglycemia. Most of these mutations affect the DNA-binding domain; some affect the ligand-binding domain. In one recent series, HNF4A mutations were found to be the third most common mutations to cause diazoxide-responsive HI – mutations were seen in 11/220 cases, with each case being caused by a different mutation. Inheritance is generally autosomal dominant, though de novo mutations were seen in 4/11 of the cases above. MODY1 and the hyperinsulinemic hypoglycemia may not be entirely unrelated – in one report, ∼14% of patients with MODY1 had a dual phenotype, having experienced transient hypoglycemia in infancy, in addition to fetal macrosomia (which was also reported in a larger subset of patients with HNF4A mutations but without hypoglycemia ).

Clinically, affected children often manifest fetal macrosomia. The infantile hyperinsulinemic hypoglycemia has a range of severity, from mild and transient hypoglycemia requiring brief intravenous glucose support to more severe, persistent hypoglycemia requiring pharmacotherapy for several years. Cases of HNF4A-HI are typically responsive to diazoxide. One reported case also developed hypophosphatemic rickets, renal Fanconi syndrome, and hepatic glycogenosis. HNF4A mutations may have a dual phenotypic manifestation – patients experiencing infantile hypoglycemia are at risk of developing MODY1, a form of monogenic diabetes, as described earlier.

Hyperinsulinism Resulting from Mutations in UCP2 (Uncoupling Protein 2)

UCP2 encodes a mitochondrial uncoupling protein that facilitates anion transfer from the inner to the outer mitochondrial membrane in exchange for protons. It is part of a larger family of uncoupling proteins that separate (“uncouple”) oxidative phosphorylation from ATP synthesis; the resulting energy is dissipated as heat. UCP2 is widely expressed in many tissues, including pancreatic islets. UCP2 acts as a negative regulator of insulin secretion in pancreatic β-cells by decreasing ATP content and thus inhibiting glucose-stimulated insulin secretion. Two heterozygous, autosomal dominant loss-of-function UCP2 mutations have been reported to cause congenital HI. Both cases were diazoxide responsive.

Other Rare Forms of Hypoglycemia Due to Hyperinsulinemia or Related Causes

• 1.
  

  Hexokinase 1 ( HK1 ): Hexokinase 1 is another member of the mitochondrial hexokinase family, like glucokinase; it catalyzes the first step in glycolysis, namely, the conversion of glucose into glucose-6-phosphate. Several cases of fasting and postprandial HI due to autosomal dominant HK1 mutations have been reported in a single family. All cases presented with moderate to severe hypoglycemia in infancy; 40% of those presented with a seizure. All responded well to diazoxide treatment. In another report, a form of HI involved islet cell mosaicism, with some localized hyperfunctioning islets with low insulin storage along with high proinsulin production, and a larger number of atrophied islets; the hyperfunctioning was found to be due to either elevated expression of low-Km HK1 (which is normally expressed at minimal levels in β-cells) substituting for GCK in β-cell glycolysis, or to an activating mutation in GCK . These cases of HI responded to diazoxide and/or were cured by focal pancreatectomy.

  
  
• 2.
  

  KCNQ1: KCNQ1 encodes a voltage-gated potassium channel with six transmembrane regions. It is located on a region of chromosome 11 along with a cluster of other genes involved in Beckwith–Wiedemann syndrome. KCNQ1 is primarily found in the plasma cell membrane of both cardiac myocytes, where activity of this channel is required for the repolarization phase of the cardiac action potential; loss-of-function mutations cause long QT syndrome, a disease of prolonged cardiac repolarization predisposing to reentry arrhythmias. However, KCNQ1 is also expressed in other tissues, including pancreatic β-cells, where it regulates potassium current and thus, potentially, regulates insulin secretion. Single-nucleotide polymorphisms associated with increased KCNQ1 activity have been linked to susceptibility to type 2 diabetes mellitus due to reduced insulin secretion in several Asian populations. The converse has very recently been reported – fourteen individuals from six families with long QT syndrome who were discovered to have postprandial and symptomatic hyperinsulinemic hypoglycemia along with hypokalemia after oral glucose challenge. It is unknown how people with hypoglycemia would respond to medical therapy; however, the mechanism of the hypoglycemia suggests that they may be diazoxide unresponsive.

  
  
• 3.
  

  AKT2 : AKT2 encodes one of three serine/threonine protein kinases that regulate a number of processes – cell proliferation, metabolism, survival, and angiogenesis. AKT2 is also involved in insulin signal processing downstream of the insulin receptor. Three known cases of hypoglycemia due to gain-of-function AKT2 mutations (heretozygous, meaning autosomal dominant, in two-thirds of cases, and mosaic in the remaining case) resulting in hypoglycemia have been reported. All three presented with severe hypoglycemia beginning at around 6 months of age. As the hypoglycemia in their cases was not caused by insulin but rather by a mediator of insulin’s effects, it was not amenable to treatment with diazoxide or octreotide; all required gastrostomy tube placement for frequent feeds to prevent hypoglycemia.

Summary

As the most common cause of hypoglycemia in infants and children, it is imperative that the diagnosis of HI be made swiftly in order to prevent further hypoglycemia and possible brain damage. Once the diagnosis has been made, the child should be started on dextrose-containing intravenous fluids while initiating therapy with diazoxide. Genetic testing will further aid in the management of the patient with HI, as the results can suggest possible focal disease, which can be cured with surgery, versus results that confirm diffuse disease. It is critical to understand the different genetic forms of HI in order to manage the patient with HI correctly. Lastly, for the surgical HI patient, it is crucial that they be cared for in a center that specializes in the management of HI and that has a team of endocrinologists, radiologists, pathologists, and surgeons that are all trained in the management of this disease.

Glycogen Storage Disease Type “0”: Glycogen Synthase Deficiency

Glycogen synthase, as discussed earlier, catalyzes the rate-limiting step in glycogen synthesis in the liver and in skeletal muscle, namely, the transfer of glucose monomers from UDP-glucose to the terminal branch of the growing glycogen chain via the formation of α(1→4) glycosidic bonds. Several glycogen synthase isoforms exist – one specific to skeletal muscle (encoded by GYS1 ), and one specific to the liver (encoded by GYS2 ). Mutations in GYS2 cause glycogen synthase deficiency leading to GSD 0. Of note, the appellation of “glycogen storage disease” for GSD 0 is actually a misnomer; this is a disorder of glycogen formation rather than of aberrant glycogen breakdown.

In GSD 0, the lack of glycogen synthase activity leads to a decrease in hepatic glycogen content. Since excess dietary carbohydrates cannot be converted into glycogen, shunting into the glycolytic pathway and consequent lactate formation occur, resulting in postprandial lactic acid elevation, hyperglycemia, and sometimes postprandial glycosuria. However, once all dietary carbohydrate is consumed, gluconeogenesis and ketosis are triggered. Initially, gluconeogenesis blunts the impact of the hypoglycemia, but eventually, the ketosis and free fatty acid elevation due to lipolysis, fatty acid oxidation, and ketogenesis inhibit alanine release from skeletal muscle. Thus, the end result is postprandial ketotic hypoglycemia accompanied by elevated triglycerides and low alanine. To date, 15 different missense mutations in GYS2 leading to glycogen synthase deficiency have been reported to date; all are autosomal recessive in inheritance.

This disease may first manifest as infants transition to longer periods of overnight fasting in between feeds or may remain asymptomatic until a period of illness (e.g., gastroenteritis) leads to a prolonged fast, which is why some cases remain undiagnosed well outside of infancy and the toddler years. Fasting tolerance typically improves as children get older – adolescents may be able to fast as long as 18 h – but hyperketonemia is still frequently seen in the fasting state if the individual is untreated. Hypoglycemia can also occur even into adulthood in circumstances of prolonged fasting, pregnancy, gastrointestinal, or other illness leading to poor enteral carbohydrate intake, and upon strenuous exercise. Other manifestations in untreated individuals include short stature and osteopenia (frequently seen), cognitive delay in 22% of cases, and rarely seizures. These sequelae are preventable if hypoglycemia and hyperketonemia are prevented. Hepatomegaly is absent, as are other complications seen in other GSDs.

The diagnosis of GSD 0 is suggested by the presence of postprandial hyperglycemia, elevated lactates, and postprandial hyperketotic hypoglycemia on oral glucose tolerance testing. Definitive diagnosis of this condition can be done noninvasively via mutation analysis of the GYS2 gene. Liver biopsies are unnecessary.

Treatment of GSD 0 entails prevention of hypoglycemia by frequent daytime snacking (every two to four h) and, in young children, administering uncooked cornstarch (1–1.5 g/kg) at bedtime. Uncooked cornstarch should also be administered every 6 h during times of intercurrent illness and prior to strenuous or prolonged exercise. These interventions also have the effect of minimizing the ketosis and accumulation of free fatty acids to prevent hyperlipidemia and the suppression of skeletal muscle alanine release. Finally, affected individuals should consume a low-glycemic index diet with complex carbohydrates (as simple carbohydrates are more prone to trigger lactate elevation) and with protein supplementation to provide a greater quantity of gluconeogenic precursors in between meals.

Glycogen Storage Disease Type 1: Von Gierke’s Disease (Glucose-6-Phosphatase Deficiency)

G6Pase, as discussed earlier, catalyzes the final step of both glycogenolysis and gluconeogenesis in hepatocytes, intestinal mucosa, and renal epithelial cells – namely, the dephosphorylation of D -G6P to D -glucose and orthophosphate. G6Pase is a 35-kD endoplasmic reticulum (ER) integral membrane protein with nine transmembrane domains; the active domain faces into the ER. G6Pase is comprised of a catalytic subunit, which performs the dephosphorylation, and a transporter subunit encoded by SLC37A4 (chromosome 11q23.3), which handles the transport of G6P and of inorganic phosphate across the ER membrane. There are three G6Pase catalytic subunit-encoding genes in humans – G6PC , G6PC2 , and G6PC3 . Mutations in G6PC , a 12.5-kb single-copy gene comprised of five exons located on chromosome 17q21.31, are responsible for GSD 1a, while mutations in SLC37A4 are responsible for GSD 1b. GSD type 1 accounts for ∼25% of all GSDs diagnosed in the United States and Europe; overall prevalence is between 1/100,000 and 1/400,000 live births, with increased prevalence in some populations such as Ashkenazi Jews (prevalence 1/20,000). GSD 1a is far more common than 1b; 80% of GSD 1 patients have mutations in G6PC .

In both GSD 1a and 1b, hepatic production of glucose from both glycogenolysis and gluconeogenesis is impaired, because G6P cannot be converted to glucose. Since G6P cannot freely cross out of the liver and into the systemic circulation, the liver effectively cannot provide glucose to other tissues to use for energy. This results in severe hypoglycemia – as low as 40 mg/dL – within two to three h of feeding, as soon as the last bit of ingested carbohydrates has been metabolized. The hypoglycemia is accompanied by increased production of lactate, triglycerides, and uric acid. Onset of hypoglycemia typically occurs in infancy once the interval between feeding is lengthened, the infant starts feeding through the night, or an intercurrent illness interrupts the feeds. As glycogen synthesis is not impaired, glycogen accumulates in the liver, resulting in massive hepatomegaly and a protruberant abdomen as well as motor delay. Poor linear growth is frequently seen. Cognitive delay is infrequent, but may occur if recurrent hypoglycemia leads to cerebral damage.

GSD III (Cori Disease) – Glycogen Debrancher Enzyme Deficiency

As discussed earlier, glycogenolysis entails the actions of several different enzymes. The major one is glycogen phosphorylase, which cleaves α(1→4) glycosidic bonds between adjacent glucose moieties in glycogen. However, 1 in 10 glucose residues in glycogen are branch points, existing as α(1→6 linkages). As glycogen phosphorylase stops cleaving when it reaches a point four residues from a branching point, the actions of glycogen phosphorylase alone do not suffice to mobilize glycogen stores. To mobilize the rest, the actions of another enzyme are required – glycogen debranching enzyme, encoded by the AGL gene and located on chromosome 1p21.2. The AGL gene is 85 kbp long and is composed of 35 exons. Six major mRNA isoforms have been identified. All isoforms share the same translation initiation site in exon 3, but differ in the 5′-untranslated regions due to differential transcription relating to specific cryptic splice sites located upstream from the translation initiation site and which enable exon removal or retention. Isoform 1, the major isoform, contains exon 1 and lacks exon 2; it is the most widely expressed, and is found in liver, kidney, and lymphoblastoid cells, as well as in myocytes; it encodes a 1532-amino acid residue protein. Isoforms 2–4 lack exon 1 and are exclusively expressed in cardiac and skeletal myocytes. All major isoforms contain exons 4–35.

Debranching enzyme is a 165-kD protein that acts as a 1,4-alpha- d -glucan:1,4-alpha- d -glucan 4-alpha- d – glycosyltransferase and amylo-1,6-glucosidase in glycogen. The enzyme has two different, independent catalytic activities that take place at different sites and can function independently of each other – 4-alpha-glucotransferase activity and amylo-1,6-glucosidase activity. GSD III results from an autosomal recessive mutation in AGL, due either to homozygous or to compound heterozygous mutations or deletions. Missense, splice site and nonsense mutations, small frameshift deletions and insertions, and large gene deletions and duplications have all been described in the AGL gene. The majority of GSD III patients have deficient enzyme activity in both liver and muscle (GSD IIIa), while ∼15% have insufficient enzyme activity in the liver only (GSD IIIb). GSD IIIa mutations possess considerable allelic heterogeneity; GSD IIIb, on the other hand, is primarily associated with two mutations in exon 3 of AGL – c.18_19delGA (p.Gln6HisfsX20), formerly described as c.17_18delAG, and c.16C > T (p.Gln6X). Most patients with GSD III lack enzymatic activity at both catalytic sites (both the 4-α-glucotransferase and the amylo-1,6-glucosidase sites); deficiencies only of the glucosidase activity (GSD IIIc) or of the glucotransferase activity (GSD IIId) have been reported, but are rare. The incidence of GSD III in the US population is 1/100,000, although it is more common in certain subpopulations, such as individuals of North African Jewish descent, where the prevalence is estimated to be 1/5400.

Hypoglycemia in GSD III occurs due to an inability to fully utilize glycogen. Glycogen debrancher enzyme (GDE) deficiency leads to accumulation of abnormal glycogen possessing short outer chains in the liver, phosphorylase limit dextrin, in affected tissues. It is typically seen after more prolonged fasts than in GSD I, is accompanied by ketosis, and is usually milder than in GSD I because gluconeogenesis is not impacted, and hepatic glycogen can be metabolized to a certain extent via glycogen phosphorylase. However, some patients with GSD III have hypoglycemia equal in severity to patients with GSD I. Other biochemical abnormalities include hyperlipidemia and transaminase elevation of over 500 U/L are often seen (the latter primarily in childhood). Uric acid and lactate concentrations are typically normal, in contrast to GSD 1a and 1b. Profound ketosis can develop after overnight fast.

The clinical phenotype of GSD III is somewhat variable. However, the main underlying clinical features in childhood are hepatomegaly, fasting hypoglycemia as above, and hyperlipidemia; if treatment is not initiated, growth failure and pubertal delay can result. Renal disease is not typically seen. In GSD IIIa, where skeletal and/or cardiac muscle is involved, a vacuolar myopathy may develop, and serum creatine kinase (CK) elevations of 5–45x above the upper limit of normal may be seen. Long-term complications of GSD III can include:

• •
  

  Hepatic cirrhosis: The hepatomegaly associated with GSD III usually resolves postpuberty, which may be attributable to a combination of lower glucose requirements but also to possible development of livery cirrhosis, which can progress to end-stage liver disease.

  
  
• •
  

  Hepatic adenomas, whose prevalence in GSD III is estimated at between 4% and 25%, and rarely hepatocellular carcinomas.

  
  
• •
  

  Skeletal myopathy: Motor delays may be the initial signs, followed by slowly developing weakness and wasting of both proximal and distal muscles. Nerve conduction studies may be abnormal. Muscles involved in respiration are typically spared. Skeletal myopathy may occur in isolation, or in combination with cardiomyopathy.

  
  
• •
  

  Osteoporosis: Seen primarily in GSD III patients with skeletal myopathy.

  
  
• •
  

  In women, polycystic ovaries may develop at higher rates; however, fertility is not typically impacted.

  
  
• •
  

  Cardiomyopathy: Ventricular hypertrophy is commonly seen; overt cardiac dysfunction is rare, but congestive heart failure may occasionally develop.

  
  
• •
  

  Sudden death: Rarely seen; thought to be secondary to cardiomyopathy-related arrhythmias.

The diagnosis of GSD III is suggested by fasting hypoglycemia accompanied by ketosis in the setting of hepatomegaly and hyperlipidemia, especially if CK elevation and/or transaminitis is seen. The absence of uric acid and lactate elevation help to differentiate from GSD I (though uric acid elevations may be seen following exertion) muscle. Glucagon administration after an overnight fast does not trigger a rise in blood glucose levels, but if glucagon is administered 2 h following a carbohydrate-rich meal, blood glucose levels do increase normally, again in contrast to GSD I. In most GSD III individuals, liver biopsy (or muscle biopsy in those patients with myopathy) shows cytosolic vacuolar accumulation of nonmembrane-bound glycogen. The excess glycogen that accumulates is structurally abnormal (short outer branches); however, lipid vacuoles are less common than in GSD I. Hepatic fibrosis ranging in severity from periportal fibrosis to micronodular cirrhosis may be seen. Diagnosis of GSD III may be based on demonstration of deficient GDE activity in biopsied hepatocytes and/or myocytes. However, biopsy is no longer necessary to establish a diagnosis; molecular genetic testing for pathogenic AGL mutations on both alleles is sufficient to establish the diagnosis noninvasively .

Treatment of GSD III entails avoidance of prolonged fasting, typically involving uncooked cornstarch (1.75 g/kg), though at longer intervals than in GSD I – every 6 h. This avoids hypoglycemia, improves linear growth, and reduces the transaminitis. For patients too young for cornstarch or who have significant growth failure and/or myopathy, continuous tube feeds overnight of a mixture of glucose and protein (or glucose oligosaccharides and amino acids) is beneficial. There is no need to restrict fructose or galactose in GSD III as in GSD I, as gluconeogenesis is intact. Serum alpha-fetoprotein and hepatic ultrasound are performed annually to biannually to screen for hepatic adenomas; liver transplants are performed, but rarely. Patients with skeletal myopathies should have intermittent electrocardiograms and echocardiograms. The prognosis for patients with myopathy is less favorable than for those with isolated hepatic involvement. Physical therapy and aerobic training may be beneficial for patients with myopathies.

Glycogen Storage Disease VI (Hers Disease, Glycogen Phosphorylase Deficiency) and GSD IX (Phosphorylase Kinase Deficiency)

In the liver, glycogen phosphorylase cleaves α(1→4) glycosidic bonds, yielding G1P and a linear glycogen chain shortened by one glucose residue. Glycogen phosphorylase has two configurations – the inactive configuration, phosphorylase B, and the active configuration, phosphorylase A. The conformational shift to the more active configuration is triggered by phosphorylation of two serine residues by phosphorylase kinase (PhK). Glycogenoses caused by reduced glycogen phosphorylase activity (GSD VI and IX) are a heterogeneous group, clinically indistinguishable but of different etiologies; some result from loss of function of glycogen phosphorylase itself (GSD VI), while others are due to hepatic phosphorylase kinase deficiency (GSD IX) resulting from mutations in one of the genes encoding PhK subunits.

Glycogen phosphorylase has three different isoforms in humans encoded by three different genes; the isoforms are respectively expressed primarily in the liver, brain, and muscle. The hepatic isoform is encoded by the PYGL gene, located on chromosome 14q22.1; the classic form of GSD VI results from loss-of-function mutations in or deletions of PYGL (homozygous or compound heterozygous mutations, deletions or duplications of one or more exons or whole-gene deletions or duplications; at least 12 different pathogenic mutations are known). Glycogen phosphorylase mutations are autosomally recessively inherited, and the deficiency is relatively rare except in the Mennonite population, where they are relatively common due to a founder mutation, affecting 0.1% of the population.

PhK is a serine-threonine phosphorylase comprised of a homotetramer of tetramers, for a total of 16 subunits; each of the four tetramers is composed of four different subunits, each encoded by a different gene: α (encoded by PHKA1 and PHKA2 ), β (encoded by PHKB ), γ (encoded by different PHKG genes for different tissues – the hepatic/testis-specific isoform is encoded by PHKG2 ), and δ (encoded by CALM1 ). The enzyme’s catalytic activity occurs at the γ subunit; the α and β subunits have regulatory functions; the δ subunit, a calmodulin, mediates the calcium dependence of the enzyme. PhK is encoded by several different genes:

• •
  

  PHKA1 , located on chromosome Xq13.2, encodes a muscle isoform of subunit α; mutations therein result in X-linked muscle PhK deficiency, a rare X-linked disorder.

  
  
• •
  

  PHKA2 , located on chromosome Xp22.13, also encodes a hepatic isoform of subunit α; PHKA2 mutations cause hepatic PhK deficiency, or GSD9A, also known as X-linked liver glycogenosis (XLG). XLG is the most common variant of GSD IX, responsible for ∼75% of cases of GSD IX. These mutations are subdivided into several categories, including GSD IXa1, where there is no PhK activity in hepatocytes or erythrocytes, and GSD IXa2, where the PhK activity is diminished or absent in hepatocytes but preserved in erythrocytes. At least 28 different PHKA2 mutations have been identified. Both GSD9A1 and GSD9A2 are caused by mutations in PHKA2 ; however, GSD9A2 is caused by either small in-frame deletions or insertions or by missense mutations, explaining the residual enzymatic expression in erythrocytes.

  
  
• •
  

  PHKB , located on chromosome 16q12.1, encodes subunit β in liver and muscle; mutations, which are rare, cause autosomal recessive GSD9B. Fourteen different pathogenic variants have been described, including missense and nonsense mutations, splice site variations, and small insertions or deletions.

  
  
• •
  

  PHKG2 , located on chromosome 16p11.2, encodes the catalytic δ subunit of PhK in the liver and testis. Mutations cause autosomal recessive hepatic PhK deficiency, GSD9C. Nine different pathogenic mutations have been described.

Phosphorylase kinase mutations are considerably more common than glycogen phosphorylase mutations, and are collectively responsible for ∼25% of all GSDs, with an estimated prevalence of 1/100,000. GSD9A, 9B, and 9C are clinically indistinguishable from each other, as well as from GSD XI.

GSD VI and XI have milder phenotypes than GSD I or III and a more favorable prognosis. Symptomatic hypoglycemia is uncommon, and primarily occurs after prolonged fasting or intensive exercise; it is accompanied by ketosis. Relatively mild hyperlipidemia (triglycerides and other lipids) and mild transaminase elevation may be seen; lactic acid and uric acid are typically normal, and thus metabolic acidosis is uncommon. Affected individuals usually present in infancy or early childhood due to growth retardation and hepatomegaly. If muscle as well as hepatic glycogen phosphorylase is impacted ( PHKA1 , PHKB ), hypotonia and motor delay may be present. As patients age, the laboratory abnormalities and hepatomegaly typically improve, as does linear growth – often without treatment. Adults are often asymptomatic. Long-term complications have been reported – liver cirrhosis, CNS complications such as impaired cognition, speech impairment, seizures, peripheral sensory neuropathy, and/or renal impairments such as tubular dysfunction and proximal renal tubular acidosis. However, these long-term complications are all rare.

Diagnosis of GSD VI can be made by assaying phosphorylase activity in erythrocytes. PHKB activity can also be assessed in erythrocytes or leukocytes, but normal PhK activity in blood does not definitively rule out GSD9, given the possibility of XLG due to PHKA2 . Hepatic biopsies for enzyme activity are possible, but again unnecessary; molecular genetic diagnosis of mutations or deletions in PYGL , PHKA1 , PHKA2 , PHKB , or PHKG2 is the test of choice.

GSD VI and IX are easier to treat than GSD I or III. No special diet is required, and there is no need to avoid fructose- or galactose-containing foods. Prolonged fasting should be avoided, and plenty of carbohydrates should be ingested if strenuous exercise is undertaken. Bedtime snacks often suffice to prevent nocturnal hypoglycemia and ketosis; if not, 2 g/kg of uncooked cornstarch can also be administered at bedtime. Overnight feeds are not required.