Clinical, biochemical, and genetic features of four patients with short‐chain enoyl‐CoA hydratase (ECHS1) deficiency

Short‐chain enoyl‐CoA hydratase (SCEH or ECHS1) deficiency is a rare inborn error of metabolism caused by biallelic mutations in the gene ECHS1 (OMIM 602292). Clinical presentation includes infantile‐onset severe developmental delay, regression, seizures, elevated lactate, and brain MRI abnormalities consistent with Leigh syndrome (LS). Characteristic abnormal biochemical findings are secondary to dysfunction of valine metabolism. We describe four patients from two consanguineous families (one Pakistani and one Irish Traveler), who presented in infancy with LS. Urine organic acid analysis by GC/MS showed increased levels of erythro‐2,3‐dihydroxy‐2‐methylbutyrate and 3‐methylglutaconate (3‐MGC). Increased urine excretion of methacrylyl‐CoA and acryloyl‐CoA related metabolites analyzed by LC‐MS/MS, were suggestive of SCEH deficiency; this was confirmed in patient fibroblasts. Both families were shown to harbor homozygous pathogenic variants in the ECHS1 gene; a c.476A > G (p.Gln159Arg) ECHS1variant in the Pakistani family and a c.538A > G, p.(Thr180Ala) ECHS1 variant in the Irish Traveler family. The c.538A > G, p.(Thr180Ala) ECHS1 variant was postulated to represent a Canadian founder mutation, but we present SNP genotyping data to support Irish ancestry of this variant with a haplotype common to the previously reported Canadian patients and our Irish Traveler family. The presence of detectable erythro‐2,3‐dihydroxy‐2‐methylbutyrate is a nonspecific marker on urine organic acid analysis but this finding, together with increased excretion of 3‐MGC, elevated plasma lactate, and normal acylcarnitine profile in patients with a Leigh‐like presentation should prompt consideration of a diagnosis of SCEH deficiency and genetic analysis of ECHS1. ECHS1 deficiency can be added to the list of conditions with 3‐MGA.


| I N TR ODU C TI ON
Biallelic mutations in the ECHS1 gene (OMIM 602292) lead to a deficiency of short-chain enoyl-CoA hydratase (SCEH/crotonase EC4.2.1.17) resulting in a rare inborn error of valine metabolism (Haack et al., 2015;Peters et al., 2014;Sakai et al., 2015;Tetreault et al., 2015). Clinical presentation includes infantile-onset of severe developmental delay and regression and seizures, with elevated plasma lactate and brain MRI abnormalities consistent with Leigh syndrome (LS) as described by Denis Leigh in 1951(Leigh, 1951 and more recently by Baertling et al. in 2014(Baertling, 2014. Genetic diagnosis of SCEH deficiency can be complicated by the clinical, radiological, and biochemical overlap with mitochondrial encephalopathic disorders; often occurring with respiratory chain complex deficiencies Sakai et al., 2015) or pyruvate dehydrogenase (PDH) deficiency (Bedoyan et al., 2017;Ferdinandusse et al., 2015;Peters et al., 2014).
Accumulation of toxic methacrylyl-CoA and acryloyl-CoA, two highly reactive intermediates that spontaneously react with sulfhydryl groups of, for example, cysteine, and cysteamine, is suspected to cause brain pathology and the biochemical pattern found in HIBCH and SCEH deficiencies (Ferdinandusse et al., 2013;Loupatty et al., 2007;Peters et al., 2014). SCEH also catalyzes the second step of the b-oxidation of short-chain fatty acids, that is, the hydration of a,b-unsaturated enoyl-CoAs to produce b-hydroxyacyl-CoAs (Kanazawa et al., 1993). It has been speculated that it is both the b-oxidation defect and the block in L-valine metabolism that contribute to the pathology in SCEH deficiency (Haack et al., 2015). However, Ferdinandusse et al. (2013) presented evidence that despite its broad substrate specificity, SCEH appears to be crucial in valine catabolism, but only of limited importance for mitochondrial fatty acid oxidation and may not play a significant role in isoleucine metabolism. Peters et al. (2014) Table 4, which are abnormal in SCEH and HIBCH deficiency. The metabolic origin of 2,3-dihydroxy-2-methylbutyrate is currently unclear. Enzymes are numbered: 2 5 propionyl-CoA carboxylase; 3 5 (R)-methylmalonyl-CoA mutase; 5 5 isobutyryl CoA dehydrogenase. [Color figure can be viewed at wileyonlinelibrary.com] carboxylase deficiency due to biotinidase deficiency (OMIM 609019).
Despite complete deficiencies of the enzymatic activities in fibroblasts, Ferdinandusse and co-workers could not detect 2,3-dihydroxy-2-methylbutyrate in the supernatant of patients' cultured fibroblasts . 2,3-dihydroxy-2-methylbutyrate exists as two isomers, erythro and threo. Erythro 2,3-dihydroxy-2-methylbutyrate is elevated in urine of patients with pathogenic ECHS1 and HIBCH variants and propionate defects and clearly segregates when plotted against the threo isomer, while the threo isomer appears to be dietary related (James Pitt, unpublished observation;Peters et al., 2015).
Given the segregation of erythro 2,3-dihydroxy-2-methylbutyrate with metabolic dysfunction, we retrospectively reviewed urine organic acid chromatograms analyzed in the Metabolic Laboratory, Department of Paediatric Laboratory Medicine in Temple Street Children's University Hospital from 2007 to 2010 and identified eight patients with increased erythro-2,3-dihydroxy-2-methylbutyrate without a diagnosis; of these, two were possible candidates for either SCEH or HIBCH deficiencies (Patient 1 and Patient 2). Two siblings of Patient 2 born after 2010 were also included (Patients 3 and 4). We describe four patients from two consanguineous families (Patient 1 from a Pakistani family and Patients 2-4 from an Irish Traveler family) who presented in infancy with LS. All had similar clinical features including developmental delay/regression, faltering growth, hypotonia, seizures, and apneic episodes (Tables 1 and 2). All had previously undergone extensive investigations without reaching a diagnosis.

| Patient reports
Patient 1 was the first child (male) born to consanguineous Pakistani parents. He was born by emergency cesarean section at 37 weeks gestation because of intrauterine growth retardation. Birth weight was 2.6 kg (5th percentile). Initially, he was slow to feed and his parents noted there was excessive crying. Global developmental delay was apparent at 5 months of age with central hypotonia, peripheral hypertonia, nystagmus, and failure to visually fix or follow. EEG was normal at 8 months; however, seizures commenced by 1 year of age and proved difficult to control. He was not dysmorphic but had hypertrichosis. Growth was suboptimal with weight less than the 0.4th percentile for age and sex.
At 3 years, he had profound global developmental delay with regression of feeding skills, poor swallow, and chronic vomiting. A percutaneous endoscopic gastrostomy (PEG) tube was inserted to improve feeding which led to some weight gain. Apneic episodes started and became frequent. He had episodes of aspiration pneumonia and he died just before his fourth birthday. His parents subsequently had two further unaffected children (Tables 1 and 2).
Patient 2 was the eldest child (male) born to consanguineous Irish Traveler parents by vaginal delivery at full term following an uneventful pregnancy. Birth weight was 3.7 kg. From 3 months of age motor developmental delay was noted. He had central hypotonia and At 20 months of age a PEG feeding tube was inserted. She was noted to be developing apneic episodes of increasing frequency and duration. At 2 years and 3 months of age, she was intubated and ventilated following a respiratory arrest and passed away 11 days later.
Patient 4 was the third affected sibling (male) in this Irish Traveler family. He was born at 34 weeks gestation by spontaneous vaginal delivery. His birth weight was 2.31 kg (9th percentile corrected for gestation). He was noted to have feeding problems with oxygen desaturation and irritability from the outset.
Investigations at 5 weeks of age showed raised lactate and abnormal urinary organic acids, with a similar profile to his sister (Table 3). On review at 12 weeks of age he was found to be irritable, hypotonic and globally developmental delayed. He was not fixing and following and had poor head control. He had significant faltering growth with weight less than 0.4th percentile. At 5 months of age a nasogastric tube was inserted. At 7 months of age he was having increasing feeding difficulties and loose stools. A PEG tube was inserted at 10 months of age.
From 9 months of age seizures were noted which became more frequent and prolonged over the following months. He was admitted to hospital for a "breath holding" episode at 11 months of age. Further developmental regression was noted at this stage; he had become increasingly hypotonic. At 12 months of age he had a significant apneic episode and an upper gastrointestinal bleed. He was intubated and ventilated and had a prolonged hospital stay. He passed away shortly after extubation at 13 months of age. This patient series involves four patients from two different families. All four patients had similar clinical features and progression. They all had significant developmental delay and central hypotonia was a common early sign. Faltering growth and progressive feeding problems were prominent in all patients. As the clinical syndrome progressed, seizures and apneic episodes became more frequent and severe.

| Brain magnetic resonance imagining (MRI) and spectroscopy (MRS)
Patient 1 had a total of three MRI Brain examinations. On the initial imaging at 6 months of age, there was pronounced symmetrical T2W prolongation of the globus pallidi and cerebral crura with corresponding restricted diffusion on apparent diffusion coefficient (ADC). Less marked hyperintensity was evident in the putamen and caudate nuclei.
There was a lactate peak on MRS. The thalami, pons, and cerebellum were not involved.
Progressive volume loss was evident on subsequent imaging, most pronounced in the cerebellum. The corpus callosum was markedly thinned on the patient's last imaging at 35 months. There was no restricted diffusion on ADC or an abnormal lactate peak on MRS on the subsequent imaging.
Patient 2 had one single MRI Brain examination at 9 months of age.
There was bilateral symmetrical T2W increased signal intensity involving the corpus striatum, cerebral peduncles, and the periaqueductal gray matter. The thalami were not involved. Diffusion weighted imaging was not performed. A lactate peak was present on MRS (Figure 2a,b).
Patient 3 had a total of three MRI brain scans. Her initial MRI at 5 weeks was normal. A subsequent MRI performed at 10 months demonstrated T2W prolongation of the caudate heads, lentiform nuclei, and cerebral crura with corresponding restricted diffusion on ADC. No abnormality was detected on MRS. Generalized brain volume loss was also evident.
There was further progressive brain volume loss, quite marked in the cerebellum on her last MR Imaging at 28 months. There was T2W hyperintensity involving the corpus striatum, which demonstrated volume loss in the interval. There were new areas of restricted diffusion, involving the hypothalamus (Figure 2c). MRS was normal.
Patient 4, the youngest of the three siblings had three MRI brain examinations. The initial MRI at 6 weeks corrected gestational age, showed mild delay of myelination. On subsequent MRI at 10 months of age, there was extensive abnormal symmetrical hyperintensity in caudate nuclei, lentiform nuclei, cerebral crura, periaqueductal gray matter, and bilateral thalami. T2W prolongation was also noted in the posterior tegmental tracts of the medulla.
On his last imaging at 12 months of age, there was marked brain volume loss with prominent sulci and extra-axial spaces. There was also abnormal T2W hyperintense signal involving the cerebellar gray matter, posterior medulla, periaqueductal gray matter, cerebral crura, thalami, and corpus striata. There were new areas of restricted diffusion seen in the caudate heads. A lactate peak was evident on MRS.

| SCEH enzyme activity measurement
Measurement of SCEH enzyme activity and immunoblotting, were performed in cultured skin fibroblasts from Patient 1 and Patient 3 as described in Peters et al. (2014).

| Genetic analyses
All exons and flanking intronic sequences of the ECHS1 gene of Patient  Lactate/pyruvate ratio was measured in perchloric acid deproteinised blood and values ranged from 6.9 to 15.2 (reference range 12-20). PDH complex (PDHc) activity in cultured skin fibroblasts (performed as described by Wicking, Scholem, Hunt, and Brown (1986) was slightly decreased at 0.58 nmol/mg protein/min (reference range 0.7-1.1). Analysis of MT-ATP8 and MT-ATP6 genes associated with LS and analysis of PDH genes and AUH gene revealed no known or candidate pathogenic variants.

| Patient 2
Plasma lactate was marginally increased with essentially normal plasma alanine and proline. Other amino acids were essentially normal including BCAA. Urine organic acids performed on one occasion showed an abnormal increase in excretion of (3MGC) and erythro-2,3-dihydroxy-2methylbutyrate. DBS acylcarnitines were not analyzed. He had no mutations in the SURF1 gene. Skin or muscle biopsies were not consented and no further investigations were performed prior to death (Table 3).

| Patient 3
Plasma lactate increased with progression of disease while plasma alanine and proline levels remained within reference range. Other amino acids were essentially normal including branched-chain amino acids.
There was a significant increase in plasma lactate from 1.71 to 3.  (Table 3).

| Patient 4
Plasma lactate levels increased with disease progression while plasma alanine and proline levels remained within reference range similar to those of his sister though they were only measured in the neonatal period. Other amino acids were also normal including branched-chain probably secondary to resolving ketosis. Excretion of erythro-2,3-dihydroxy-2-methylbutyrate while detectable in all six urines was not considered significant enough for comment on qualitative analysis during earlier sampling and was only considered to be increased when he was approximately 5 months of age. The acylcarnitine profile was normal and neither C 4 OH/isoC 4 OH-carnitine nor methylmalonyl/succinyl-carnitine (C 4 DC) were increased. As with his two siblings, analysis of mtDNA genes commonly associated with LS and WES did not identify a genetic cause. Skin and muscle biopsies were not consented (Table 3).
All four patients had overlapping biochemical abnormalities; elevated plasma lactate, with or without increased plasma alanine and proline, 3-methylglutaconic aciduria with elevated erythro-2,3-dihydroxy-2methylbutyrate and usually normal C 4 OH/isoC 4 OH-carnitine and normal branched-chain amino acid levels.
Quantitative analysis of erythro-2,3-dihydroxy-2-methylbutyrate confirmed that levels were significantly increased in the urine of Patient 1 and Patient 4 ( Table 3)   patients (Haack et al., 2015). Unfortunately, parental DNA samples were not available to confirm the apparent homozygosity in Patient 1.

| Family 2
Initial WES analysis of Patients 3 and 4 did not identify any potentially pathogenic variants, but following a biochemical diagnosis of SCEH deficiency in Patient 3, reanalysis of the exome data at the ECHS1 locus revealed exon 5 was not captured in this exome assay, and had no corresponding sequencing data, despite an average 1003 coverage exome-wide. Sanger sequencing of exon 5 identified a previously

| D I SCUSSION
This case series was part of a retrospective study performed following the publication that the presence of erythro-2,3-dihydroxy-2-methylbutyrate in urine had been identified in some patients with SCEH and HIBCH deficiency. This review involves four patients from two families where no definitive cause of death was made at the time of death in two of the index cases in each family.
All four patients were diagnosed with SCEH deficiency and had similar clinical features and progression. They all had significant developmental delay. Central hypotonia was a common early sign. Faltering growth and progressive feeding problems were prominent in all cases.
As that levels may be unreliable shortly after birth (Ganetzky et al., 2016) which was the case in Patient 4. Two out of four patients had increased methylmalonic acid which has been reported previously in a patient with SCEH deficiency (Tetreault et al., 2015). although his AUH gene sequencing was negative. In primary 3-MGA, deficiency of 3-methyl glutaconyl Co A hydratase due to mutations in the AUH gene are directly responsible for the accumulation of 3MGC while in secondary 3MGA, no defect in leucine catabolism exists and the metabolic origin of 3MGC is unknown (Su & Ryan, 2014). Increased 3-HIVA has been previously reported in a 1-year-old male patient  and increased excretion of 3-MGC has been reported previously in at least one case of SCEH deficiency, a 7- year-old female and may represent a later biochemical feature in milder cases  and noted to be increased transiently following long-term follow-up (Huffnagel et al., 2017). All of our patients were from consanguineous unions so it is possible that excretion of 3-MGC is not related to ECHS1 mutations but to a separate condition or secondary to mitochondrial dysfunction.
There is a rapidly growing group of IEMs with a syndromic phenotype reviewed by Wortmann, Kluijtmans, Engelke, Wevers, and Morava (2012) and Wortmann et al. (2013) causing secondary 3methylglutaconic aciduria (3-MGA) due to defective phospholipid remodeling or mitochondrial membrane associated disorders. The excretion of 3-MGC can be highly variable or intermittently normal in secondary 3-MGA and is generally less than levels seen in primary 3-MGA due to AUH. Clinical features of secondary 3-MGA are heterogeneous but distinctive, rare, but highly characteristic, neurometabolic syndromes (Saunders et al., 2015;Wortmann et al., 2013Wortmann et al., , 2015. 3-MGC is also is a marker not only for mitochondrial dysfunction in general, but for specific mitochondrial disorders (Mandel et al., 2016;Ol ahov a et al., 2017;Wortmann et al., 2013).
Our patients had extensive mitochondrial molecular genetic studies performed. WES data showed no evidence for an alternative etiology in the second family so it is less likely that 3-MGC excretion is due to any of the reported conditions causing secondary 3-MGA and more likely secondary to non-specified mitochondrial dysfunction.  (Bedoyan et al., 2017;Ferdinandusse et al., 2013;Sakai et al., 2015).
A wide phenotypic spectrum is now emerging for SCEH deficiency, ranging from lethality in the first days of life to adult patients who may not fulfill all criteria for LS. Urine metabolite levels correlate with clinical severity and specific separation of isoC4OH and OH-C4-carnitine isomers can distinguish between SCEH and HIBCH deficiency (Al Mutairi et al., 2017;Bedoyan et al., 2017;Ganetzky et al., 2016;Haack et al., 2015;Nair et al., 2016).
The increased excretion of methacrylyl-CoA, acryloyl-CoA adducts, and erythro-2,3-dihydroxy-2-methylbutyrate in the two patients quantitated by LC-MS/MS are consistent with SCEH deficiency; however, they are less than those previously reported in clinically severe cases (James Pitt, unpublished observations;Peters et al., 2014Peters et al., , 2015. Evidence is emerging that metabolite levels may correlate with disease severity, being subtle or normal for some metabolites in clinically milder cases (Haack et al., 2015;Yamada et al., 2015) and retrospective analysis of S-(2-carboxypropyl)cysteamine, S-(2-carboxypropyl)cysteine, and N-acetyl-S-(2-carboxypropyl) cysteine can be a diagnostic clue in the disease spectrum of ECHS1 deficiency (A Mutairi et al., 2017). SCEH activity in cultured fibroblasts was markedly reduced but not completely deficient, which is in agreement with a milder clinical phenotype considering the age of death when compared to some other published cases (Bedoyan et al., 2017;Haack et al., 2015;Peters et al., 2014). in ECHS1 and results from their enzyme activity measurements and immunoblot analysis strongly suggest that there is a correlation between the residual SCEH enzyme activity and the severity of the clinical symptoms . In addition, excretion of 2,3-dihydroxy-2-methylbutyrate may not have been reported or checked in some historical cases because its significance was not known at the time; it may be a more common biochemical feature of SCEH deficiency than we are aware of. WES has accelerated the discovery of new genes and pathways involved in LS, providing novel insights into the pathophysiological mechanisms. No general curative treatment is available for this devastating disorder, although several recent studies imply that early treatment might be beneficial for some patients depending on the gene or process affected, for example, the ketogenic diet has been noted to be helpful in patients with PDHc deficiency and some patients with PDHc deficiency and LS may show small benefits from vitamin or co-factor supplementation with coenzyme Q10, thiamine or riboflavin (Baertling et al., 2014;Gerards, Sallevelt, & Smeets, 2016;Sofou et al., 2014).
Mahajan, Constantinou, and Sidiropoulos (2017) describe a case of ECHS1 deficiency-associated paroxysmal exercise-induced dyskinesias with initial symptomatic improvement after 3 months treatment with a mitochondrial cocktail. Soler-Alfonso et al. (2015) proposed that valine restriction made significant improvement in bilateral ptosis and postural tone of their one HIBCH deficient patient when plasma valine was lowered to 83 mmol/L (reference range 82-293). However, valine restriction alone may not be effective because it was shown that it did not decrease valine levels in rat brain (Hutchison, Zarghami, Cusick, Longenecker, & Haskell, 1983). Restricting leucine, isoleucine, and valine (i.e. a Maple Syrup Urine Disease diet) does decrease brain valine levels so this approach may show better benefits. CSF valine is normal in SCEH/HIBCH deficiency, and the methacrylyl-CoA and acryloyl-CoA toxic metabolites are three steps downstream in the valine pathway.
Therefore, decreasing cerebral valine may not make a significant impact on the accumulation of toxic metabolites. Treatment with cysteamine and/or N-acetylcysteine is both safe and widely used and cross the blood brain barrier. Some benefits have been shown in Huntington's disease, Parkinson's disease and Alzheimer's disease, some aspects of which involve glutathione depletion, by enhancing natural detoxification (Besouw, Masereeuw, van den Heuvel, & Levtchenko, 2013). Currently, there is no direct evidence of cerebral glutathione depletion in ECHS1 or HIBCH gene defects. However, a fully functional glutathione system may still not be efficient enough to prevent damage from accumulating methacrylyl-CoA and treatment with cysteamine or N-acetylcysteine may be a useful adjunct to the detoxification of these metabolites.

| C ONC LUSI ON
The presence of detectable erythro-2,3-dihydroxy-2-methylbutyrate in urine is a nonspecific biochemical finding. 3MGA is not a discriminative feature but a minor finding in the biochemical phenotype of ECHS1 deficiency; however, in combination with increased excretion of erythro-2,3-dihydroxy-2-methylbutyrate, elevated plasma lactate and normal acylcarnitine profile in a patient with LS should prompt consideration of SCEH deficiency.
It is a known limitation of WES that not all nucleotides of every coding exon are captured. This can be due to the nature of the primary sequence, for example GC-rich and repetitive regions are not well ascertained by hybridization-based capture assays. Additionally, multiple different gene annotation systems exist that do not show complete agreement and capture assays have been designed according to best approximations of genic structures. In this study, WES was conducted to an exome-wide average coverage of 100x, however, there were no reads present in ECHS1 exon 5. This underlines the importance of taking into account coverage when interpreting exome results.
Our report also highlights that biochemical analyses remain the gold standard for the diagnosis of some patients with inborn errors of metabolism and, where possible, a molecular genetic diagnosis should always be contextualized with functional (e.g. biochemical) data.