3-Hydroxyacyl-Coenzyme A Dehydrogenase Deficiency and Hyperinsulinemic Hypoglycemia: Characterization of a Novel Mutation and Severe Dietary Protein Sensitivity

Abstract

Background

HADH encodes for the enzyme 3-hydroxyacyl-coenzyme A dehydrogenase (HADH) and catalyses the penultimate reaction in the β-oxidation of fatty acids. All previously reported patients with mutations in HADH gene and hyperinsulinemic hypoglycemia (HH) showed raised plasma hydroxybutyrylcarnitine and urinary 3-hydroxyglutarate.

Aims

The aims of the study were: 1) to report a novel HADH gene mutation not associated with abnormal acylcarnitine or urinary organic acid profile; and 2) to report the novel observation of severe protein-sensitive HH in three patients with HADH gene mutations.

Research Design and Methods

The index case presented at 4 months of age with hypoglycemic seizures. Her HH responded to diazoxide, but she continued to have episodes of hypoglycemia even on diazoxide, especially when consuming high-protein foods.

Results

Investigations confirmed HH (blood glucose level of 1.8 mmol/liter with simultaneous serum insulin level of 58 mU/liter) with normal acylcarnitines and urine organic acids. Sequencing of the HADH gene identified a homozygous missense mutation (c.562A>G; p.Met188Val). Hydroxyacyl-coenzyme A dehydrogenase activity was significantly decreased compared with controls (index patient, mean ± SEM, 26.8 ± 4.8 mU/mg protein; controls, 48.0 ± 8.1 mU/mg protein; P = 0.029) in skin fibroblasts. This patient was severely protein sensitive. Two other children with HH due to HADH gene mutations also demonstrated marked protein sensitivity.

Conclusions

Mutations in the HADH gene are associated with protein-induced HH, and patients with HH due to HADH gene mutations may have normal acylcarnitines and urine organic acids. (J Clin Endocrinol Metab 94: 2221–2225, 2009)

Hyperinsulinemic hypoglycemia (HH) is a cause of persistent hypoglycemia in the newborn and infancy period. Biochemically, it is characterized by the inappropriate secretion of insulin from pancreatic β-cells in the presence of a low blood glucose level. The commonest genetic cause of persistent HH are inactivating mutations in the genes (ABCC8 and KCNJ11) encoding the two subunits of the pancreatic β-cell AMP (ATP)-sensitive potassium (K_ATP) channel (1, 2).

Activating mutations in GLUD1 [encoding glutamate dehydrogenase (GDH)] are the second commonest cause of HH and cause the hyperinsulinism-hyperammonemia syndrome (3). Patients with hyperinsulinism-hyperammonemia syndrome typically demonstrate leucine sensitivity (4). In contrast, patients with HH due to defective K_ATP channels develop protein sensitivity without leucine sensitivity (5). It is thought that amino acids other than leucine (such as glutamine or glutamate) are involved in provoking HH in these patients, possibly via pathways independent of GDH.

Fatty acids play a pivotal role in regulating insulin secretion (6). However, the molecular mechanisms of fatty acid-induced insulin secretion are unclear. 3-Hydroxyacyl-coenzyme A dehydrogenase [HADH; previously known as short-chain L-3-hydroxyacyl-CoA dehydrogenase (SCHAD)], encoded by the HADH gene, is an intramitochondrial enzyme that catalyzes the penultimate reaction in the β-oxidation of fatty acids, the nicotinamide adenine dinucleotide (NAD⁺)-dependent dehydrogenation of 3-hydroxyacyl-CoA to the corresponding 3-ketoacyl-CoA (7). Mutations in HADH are a rare cause of recessively inherited HH (8–10). The clinical presentation of all these reported patients was heterogeneous, with either mild late onset hypoglycemia or severe neonatal hypoglycemia. The acylcarnitine profile in all reported patients demonstrated raised hydroxybutyrylcarnitine, and urine organic acids showed raised 3-hydroxyglutarate with decreased expression and function of the HADH enzyme.

We now report the fifth case of HH associated with a novel missense mutation in the HADH gene. Unlike all other previously reported cases, our patient has normal acylcarnitines and urine organic acids profile. In addition, we examined the cause for intermittent episodes of HH in this patient while on diazoxide therapy and found her to be markedly protein sensitive. In this report, we describe the phenotype of this patient with the novel mutation and describe our novel observation of severe protein sensitivity in other patients with mutations in the HADH gene.

Case Report

The index case, born to consanguineous parents of Bangladeshi origin presented at 4 months of age with hypoglycemic seizures. Investigations revealed a biochemical picture consistent with hypoketotic HH with raised serum insulin levels of 58 mU/liter and low serum fatty acids during hypoglycemia (blood glucose level of 1.8 mmol/liter). The hypoglycemia responded to diazoxide, but she continued to have episodes of HH especially when taking a meal rich in protein. Given the unpredictable nature of the hypoglycemic episodes and the absence of a mutation in ABCC8 and KCNJ11, we examined the HADH gene. Interestingly, despite repeated measurements of the acylcarnitine and urinary organic acids, no abnormal metabolites were detected.

To investigate the intermittent nature of the hyperinsulinism, we evaluated our patient for protein specifically.

Patients and Methods

HADH sequencing

Genomic DNA was extracted from peripheral leukocytes using standard procedures. The eight exons of the HADH gene were amplified, and the products were sequenced using a BigDye Terminator cycler sequencing Kit v3.1 (Applied Biosystems, Warrington, UK). The sequencing reactions were analyzed on an ABI 3730 capillary sequencer (Applied Biosystems), and sequences were compared with the published sequence (NM_005327.2) using Mutation Surveyor version 2.61 (Softgenetics, State College, PA). PCR primers are available on request. The study was approved by the regional ethical committee, and written, valid consent was obtained from the patients and the families.

Measurement of HADH activity

Short-, medium-, and long-chain HADH activity was measured in fibroblast homogenates with acetoacetyl-CoA, 3-ketooctanoyl-CoA, and 3-ketohexadecanoyl-CoA and normalized to protein as described previously (8). Each measurement was performed on six fibroblast pellets from the index patient and on 12 different controls, and data were compared by Student’s t test.

Protein tolerance test

We performed an oral protein load to investigate the possibility of protein-induced hypoglycemia. After a positive result, we then proceeded to investigate our two previous patients with HH due to mutations in the HADH gene (8, 9). The results were compared with three control subjects with ketotic hypoglycemia. Patients with ketotic hypoglycemia were selected as controls because the cause of hypoglycemia is not related to insulin action in these patients (as evidenced by the presence of elevated ketone bodies during hypoglycemia). All the patients had previously demonstrated normal fasting tolerance on diazoxide. The diazoxide dose for patients with HADH mutations was between 13 and 15 mg/kg · d. Diazoxide was not discontinued for the purpose of the test because stopping diazoxide in patients with hyperinsulinism may itself lead to hypoglycemia and false-positive results. After the observation of severe protein sensitivity, we sequenced the GLUD1 gene in all our patients and found no mutations.

The patients and controls were fasted for 4 h before the test. Baseline bloods (insulin, glucose) were collected before oral administration of the protein mixture. The oral protein load was given in the form of a standard protein powder (Vitapro; Vita-Mix, Cleveland, OH) dissolved in water. The amount of amino acids present in 100 g of this mixture is as follows: L-alanine 3.6 g, L-arginine 1.9 g, L-cysteine 1.7 g, L-glutamine 14.8 g, L-glycine 1.4 g, L-histidine 1.2 g, L-isoleucine 4.7 g, L-leucine 7.5 g, L-lysine 6.6 g, L-methionine 1.7 g, L-phenylalanine 2.2 g, L-proline 4.9 g, L-serine 3.2 g, L-threonine 5.9 g, L-tryptophan 1.8 g, L-tyrosine 1.9 g, L-valine 4.0 g, and L-asparagine 8.9 g. A total of 1.5 g/kg of the mixture was administered, and bloods were collected at 30-min intervals after load. The test was stopped at 180 min or earlier if hypoglycemia developed; hypoglycemia was defined as blood glucose level below 3.0 mmol/liter for the purpose of the test.

Results

HADH mutation analysis

A novel homozygous missense mutation, M188V (c.562A>G, p.Met188Val), was identified in exon 5 of the patient’s HADH gene. Her parents and two unaffected siblings were heterozygous for the M188V mutation. Both methionine and valine are nonpolar amino acids, but the methionine residue at position 188 is conserved from human to Xenopus tropicalis. We sequenced 72 control ethnically matched chromosomes, and the homozygous missense mutation M188V (c.562A>G, p.Met188Val) was not detected.

Biochemistry

No abnormal acylcarnitines were detected, and repeated urine organic acids showed no abnormality.

Activity of L-HADH

Short-chain L-HADH activity was significantly decreased compared with controls (index patient, mean ± SEM, 26.8 ± 4.8 mU/mgprotein; controls, 48.0 ± 8.1 mU/mgprotein;P = 0.029), although the residual activity was much higher than in the other reported cases (8, 9). Activities of medium- and long-chain L-HADH were also mildly decreased compared with controls, although this difference was not significant (medium-chain index patient, 25.8 ± 4.2 mU/mg, vs. controls, 40.1 ± 6.1 mU/mg; long-chain indexpatient,28.8 ± 5.0mU/mg, vs. controls, 43.8 ± 6.2 mU/mg).

Protein-induced HH

The blood glucose response to the oral protein load in the patients with HADH deficiency is illustrated in Fig. 1. As shown, all three subjects with HADH deficiency demonstrated a dramatic decline in blood glucose level. In contrast, the control subjects had no change in blood glucose levels in response to protein. Table 1 shows the blood glucose and insulin responses to the oral protein load in subjects and controls.

Fig. 1.

Blood glucose levels in response to oral protein load in patients and controls. All patients with mutations in the HADH gene demonstrated marked HH in response to a standard protein load. In contrast, there was no hypoglycemia in the control patients.

Table 1. Comparison of blood glucose and insulin levels between patients (Pt 1, Pt 2, and Pt 3) and controls (C 1, C 2, and C 3) during protein load.

Subject	Age at testing (yr)	HADH mutation	Baseline blood glucose (mmol/ liter)	Nadir blood glucose (mmol/liter)	Baseline insulin (mU/liter)	Peak insulin (mU/liter)	Time to nadir (min)	Fasting tolerance (h)
Pt 1	10.3	P258 liter	4.5	2.7	6.1	67.7	180	16
Pt 2	8.4	IVS6–2a>g	5.2	2.9	<2.0	24	60	14
Pt 3	1.2	M188V	4.0	2.8	4.7	7.8	120	16
C 1	1.1	Nil	4.0	4.1	3.1	6.1		10
C 2	8.8	Nil	3.9	4.4	2.6	5.1		16
C 3	11	Nil	4.6	4.6	4.1	12.7		18

All three subjects with mutations in the HADH gene developed hypoglycemia with a concomitant rise in plasma insulin levels. In contrast, the control subjects had no change in blood glucose levels in response to a protein load.

Discussion

Mitochondrial HADH is involved in the second dehydrogenation step of the β-oxidation pathway. It catalyzes the conversion of L-3-hydroxyacyl CoAs to 3-ketoacyl CoAs, and this reaction also uses NAD⁺ as cofactor producing reduced nicotinamide adenine dinucleotide. HADH is highly expressed in pancreatic β-cells, suggesting that it has an important role in insulin secretion (7). The normal β-cell phenotype is characterized by a relatively high expression of HADH and a low expression of other β-oxidation enzymes (such as acyl-CoA dehydrogenase short, medium, and long chain and acetyl-CoA acyltransferase 2) (11).

Several recent studies have shown that HADH has a pivotal role in regulating insulin secretion (11, 12) and that it interacts with other genes that are known to be important for β-cell development and function (13, 14). In one study, suppression of HADH activity using small interfering RNA caused a significant increase in basal insulin secretion compared with untreated cells (12) . This demonstrates for the first time that HADH is required directly in β-cells for the regulation of basal insulin release. The addition of diazoxide did not alter the enhanced basal insulin secretion caused by suppression of HADH, indicating that HADH functions directly in β-cells to regulate a K_ATP-independent pathway to regulate insulin secretion.

In another study (11) based on rat β-cells and in the β-cell line INS1 832-13, HADH silencing resulted in elevated insulin release at low and high glucose concentrations, which appeared not to be caused by increased rates of glucose metabolism or an inhibition in fatty acid oxidation. Down-regulation of HADH caused an elevated secretory activity, suggesting that this enzyme protects against inappropriately high insulin levels and hypoglycemia (11). More importantly, the study by Martens et al. (11) suggests that the increased insulin secretion observed in HADH knockout cells may not correlate with the degree of β-oxidation inhibition but correlates with the actual level of HADH protein.

HADH expression is regulated by transcription factors that are crucial for proper β-cell differentiation and function such as Foxa2 (13). β-Cell-specific Foxa2 knockout mice show a 3-fold downregulation of HADH. These mice have severe HH due to unregulated insulin secretion (14). Foxa2 encodes a transcription factor (forkhead box a2) that has been postulated to play a central role in β-cell development due to its ability to bind to and transactivate pancreatic duodenal homeobox 1 (Pdx1) cis-regulatory elements in vitro (15). Hence, it is clear from these recent studies that HADH has an extremely important but unknown role in regulating insulin secretion. Understanding the molecular mechanisms of how mutations in HADH lead to unregulated insulin secretion will provide novel insights into pancreatic β-cell physiology.

It has been hypothesized that fatty acids increase insulin secretion by affecting the concentrations of long-chain fatty acyl derivatives as a result of the inhibitory effect of citrate and malonyl-CoA on the rate-controlling enzyme carnitine palmitoyl-transferase-1 (16, 17). Various potential mechanisms have been put forward to explain HH associated with HADH deficiency (18). The presence of circulating abnormal acyl carnitine metabolites in the blood in all previously reported patients led to speculations that accumulation of short-chain acyl Co-A esters originating from circulating metabolites or the metabolites themselves may lead to dysregulated insulin secretion by inhibition of carnitine palmitoyltransferase-1, via K_ATP channels, or via the G protein-coupled receptor GPR40 (18).

HADH deficiency remains an ill-defined entity. Of the previously described cases, hyperinsulinism was the defining clinical feature (8–10). All patients had low residual HADH activity in fibroblasts and elevated 3-hydroxybutyrylcarnitine in blood. Another patient, described by Bennett et al. (19), had a high residual enzyme activity and absence of 3-hydroxybutyrylcarnitine and presented with a Reye-like hepatic dysfunction and no documented hyperinsulinism. The high residual enzyme activity and absence of 3-hydroxybutyrylcarnitine in the patient of Bennett et al. (19) is similar to the findings in the patient we describe here. However, our patients all presented with hyperinsulinism. Although these findings are apparently contradictory, they can be reconciled if a function of HADH other than the activity is involved in pancreatic insulin secretion. Recently, Filling et al. (20) summarized findings in six additional individuals, all of whom had HADH deficiency demonstrable in liver or muscle, but none of whom had a pathogenic HADH mutation or documented hyperinsulinism.

The molecular basis of HH due to mutations in HADH is still unclear. It is now well established that the HADH gene is important for and regulates the secretion of insulin via a mechanism that is yet unknown (11, 12). Hardy et al. (12) have now established that impaired insulin secretion in HADH deficiency occurs via mechanisms independent of the K_ATP channels. This mechanismmay or may not be unrelated to the activity of the enzyme. Indeed, the high residual enzyme activity is in keeping with the lack of 3-hydroxybutyrylcarnitine or the other metabolites measured in the previous cases of HADH with hyperinsulinism. The HADH protein has been previously shown to be associated with other proteins that may be involved in insulin secretion, e.g. GDH (20) or complex I (21). Indeed, Filling et al. (20) conclude that protein interactions rather than activity changes are likely to lead to the phenotype. Mutations in HADH may cause only small changes in enzyme activity but may cause major changes in interactions with other proteins. The study by Martens etal. (11) demonstrates that knockdown of HADH in β-cells increases insulin secretion not as a consequence of a perturbation of β-oxidation but possibly by another metabolic mechanism that is currently unknown. Thus, the accumulation of 3-hydroxybutyrylcarnitine may not be the trigger for insulin secretion in patients with HADH deficiency. Our index case now demonstrates that HH associated with HADH deficiency in fact may not always manifest itself with the presence of these abnormal metabolites. Indeed, the lack of detectable abnormal acylcarnitines is in keeping with the mild defect in enzyme activity compared with previously described patients. A patient with a missense mutation in the HADH gene and normal acylcarnitines has also been reported recently (22).

Our study also demonstrates for the first time that children with HH due to mutations in the HADH gene are severely protein sensitive (Table 1). This unique clinical observation suggests that defects in HADH (and hence HADH enzyme activity) uncover an unidentified biochemical pathway by which amino acid(s) trigger insulin secretion. Leucine-sensitive HH occurs in patients with GLUD1 mutations, whereas protein-sensitive (but not leucine-sensitive) HH is observed in patients with defects in the K_ATP channel genes (4, 5). However, protein sensitivity due to mutations in the HADH gene has not been described before. Our clinical observations suggest that HADH protects against excess amino acid-induced insulin secretion. Unraveling this biochemical pathway will provide unique novel insights into fatty acid- and amino acid-induced insulin secretion.

The study by Filling et al. (20) suggests the existence of an interaction between HADH and GDH in the pancreatic β-cell. It could be hypothesized that HADH deficiency causes proteinsensitive HH via the GDH axis. This is likely to be mediated not by leucine sensitivity but via another amino acid possibly involving glutamine. Further investigations are required to understand this link between HADH, GDH, and regulation of insulin secretion, which will provide invaluable insights into pancreatic β-cell biochemistry and physiology.

Abbreviations

CoA

Coenzyme A

HADH

3-hydroxyacyl-coenzyme A dehydrogenase

HH

hyperinsulinemic hypoglycemia

K_ATP

ATP-sensitive potassium