Congenital lactic acidosis may result from any of numerous defects in oxidative metabolism, including the conversion of lactate to pyruvate, the oxidation of pyruvate in the tricarboxylic acid (TCA) cycle, and the mitochondrial respiratory chain [1]. The pyruvate dehydrogenase complex (PDC) is responsible for entry of pyruvate into the TCA cycle via formation of acetyl CoA, and PDC deficiency is a major cause of congenital lactic acidosis [2]. The PDC is composed of five components, all of which have known congenital deficiencies: the pyruvate dehydrogenase (PDH) E1, PDH E2, PDH E3, X-lipoate, and PDH phosphatase [2, 3]. The most commonly described PDC defect is due to deficiency of the PDH E1ï¡ subunit [4, 5]. The PDH E1ï¡ subunit (MIM 312170) is encoded by the 16 kb gene PDHA1 comprising 11 exons on Xp22.12 [6-8].
Given the location of PDHA1 on the X chromosome, diagnosis of PDH E1ï¡ deficiency (OMIM #312170) is usually straightforward in males, with severe neurologic manifestations and clear laboratory evidence of lactic acidosis and PDC enzyme deficiency in all tissues due to obligate hemizygosity [9-12]. However, PDH E1ï¡ deficiency in female patients often presents a diagnostic dilemma due to variable X chromosome inactivation [6, 13]. Clinical manifestations thus span a wide spectrum, from neonatal death with severe lactic acidosis and necrotizing encephalopathy to normal development with intermittent peripheral neuropathy, depending on the tissues affected [9, 14-16].
Even in the presence of clearly elevated serum lactate and pyruvate levels, the results of PDC enzyme activity assays from easily obtainable samples may not reflect PDC activity in clinically relevant tissues [17, 18]. While sequencing of PHDA1 often uncovers mutations explaining an observed lactic acidosis, sequencing does not reveal deletions of whole exons or larger regions in female patients [19-21]. Microarray based methods of comparative genomic hybridization (CGH) reveal gene dosage information which may be missed by other methods and, in several cases, have explained biochemical abnormalities in the setting of seemingly normal gene sequencing results [22, 23].
Here, we present the case of a girl with clinical and laboratory signs of congenital lactic acidosis and normal lymphocyte PDC activity assay. Sequencing of PDHA1 in a blood sample did not reveal significant deleterious mutations. Presence of additional dysmorphic features prompted further analysis including chromosomal microarray. Gene dosage analysis with array based oligonucleotide CGH revealed deletion of a chromosomal band at Xp22.12-22.13, including the entire PDHA1 locus.
Clinical findings
The patient is a # year old female born at 38 weeks gestational age to a 33 year old, gravida 4 para 3 mother. Her mother had 2 healthy daughters and 1 healthy son from the same union as the patient, and had voluntarily terminated 1 pregnancy from a previous union. Her prenatal course was uncomplicated, but ultrasound examination performed at 5 months gestational age demonstrated ventriculomegaly. At birth, the girl weighed 2.8 kg and was 49.5 cm long. She failed her newborn hearing screen and was noted to have several orthopedic anomalies, including scoliosis, left vertical talus and right talipes equinovarus. Soon after birth, she was noted to have elevated serum lactic acid and pyruvate levels. At 2 months of age she continued to have an elevated lactic acid and pyruvate, with lactic acid of 5.1 mM (reference range 0.7-2.1 mM), pyruvate of 0.30 mM (reference range 0.03-0.12 mM), and lactate/pyruvate ratio of 17. In her second year of life, she had some staring spells which were treated for less than 1 year with oxcarbazepine. EEG showed bilateral temporoparietal sharp wave discharges and background slowing, but no spontaneous or induced seizures. At that time, brain MRI demonstrated agenesis of the corpus callosum with remnants of the genu and anterior third of the body, lateral ventriculomegaly and underdevelopment of the hippocampi without focal areas of necrosis or encephalomalacia. She also underwent evaluation of auditory brainstem responses, which showed left sided hearing loss, but present hearing on the right. Due to the persistence of elevated lactate and pyruvate, she was evaluated for PDC deficiency at 2 years of age with PDC activity assay in lymphocytes, mitochondrial DNA analysis of a blood sample, and chromosomal microarray analysis utilizing a bacterial artificial chromosome based platform (SignatureChip 4.0, Signature Genomics, Spokane, WA), none of which demonstrated an abnormality.
The patient was evaluated at our institution at 4 years of age in the context of critical illness brought on by pneumonia. On clinical exam, she weighed 10.4 kg. She was microcephalic with head circumference 42 cm and had a metopic ridge and plagiocephaly. Her pupils were reactive and her eyes moved spontaneously to all quadrants, and she inconsistently tracked objects in her visual field. She had a large, protruding tongue. She moved all four extremities spontaneously and had normal muscle tone and reflexes. Her psychomotor development was profoundly delayed, and she did not say words, sit independently, or make goal-directed movements, though she did vocalize. She had been taking her nutrition orally prior to her hospitalization. She had undergone bilateral below the knee amputations after arterial thromboemboli secondary to femoral line placement. Hypercoagulability workup demonstrated heterozygosity for C677T mutation of the methyltetrahydrofolate reductase, though plasma homocysteine was normal. There were no abnormalities of protein C, S, and no prothrombin 20210 or factor V Leiden mutation. She had several shaking episodes while hospitalized. These were revealed by long-term EEG monitoring to be nonepileptic, but she was found to have subclinical seizures of left temporal onset. She was also found to have mixed central and obstructive sleep apnea. MRI imaging again demonstrated ex-vacuo dilatation and congenital abnormalities but no focal lesions. Her blood lactate levels were high (2.3-11.8 mM, normal < 2.0 mM) and lactate/pyruvate molar ratio of 10:1. Plasma amino acid profile showed elevated alanine without other significant abnormalities, and urine organic acid profile showed only elevated lactic acid. Plasma acylcarnitine profile was normal. Electron transport chain enzyme assay in muscle and fibroblasts, as well as mitochondrial DNA point mutation and deletion analysis were not consistent with an underlying mitochondrial disorder. Chromosomal microarray (CMA) analyses were performed using the custom-designed EMArray Cyto6000 array platform [Baldwin et al, 2008; The International Standard Cytogenomic Array (ISCA) Consortium], and revealed a 1.1 Mb loss of genomic material from the short arm of chromosome X, at Xp22.13-p22.12
The patient's course continued to be complicated over the next several months by multiple episodes of apnea and respiratory distress, eventually requiring tracheostomy and home mechanical ventilation.
Methods
Pyruvate dehydrogenase complex activity.
We will Get Dr. Kerr's input here
PDC activity testing was performed at the Center for Inherited Disorders of Energy Metabolism (Rainbow Babies Children's Hospital, Cleveland, OH). Testing was performed using a thiamine pyrophosphate, NAD and CoA dependent decarboxylation of 1-14C-pyruvate after activation or inactivation by incubation with dichloroacetate or fluoride according to previously described methods [24, 25]. Analysis was performed both in blood lymphocytes and skin fibroblasts.
PDH E1ï¡ sequencing
Dr. Kerr
Sequencing of the PDH E1ï¡ gene was carried out by the Baylor College of Medicine Medical Genetics Laboratory. Sequencing included the exons and immediately adjacent intronic regions of the PDHA1 gene located at Xp22.2-p22.1 using GenBank NM_000284 as a reference sequence.
Chromosomal microarray analysis.
Chromosomal microarray (CMA) analysis was performed in the University of Michigan's, Michigan Medical Genetics Laboratories' (MMGL) Molecular Genetics Laboratory using the custom-designed EMArray Cyto6000 chip, implemented on the Agilent 44K platform [Baldwin et al., 2008]. [26]. This array contains 43,103 features with 95% of the probes distributed at an average interval of 75 Kb, with higher density, targeted coverage of known genomic disease loci, telomeric boundaries, and loci that exhibit copy number changes associated with known Mendelian disorders. This array design facilitates detection of copy number imbalances at a minimal resolution of ~500 kb in the genomic backbone, and at significantly greater resolution in targeted regions.
For CMA analysis, patient genomic DNA was isolated from peripheral blood samples using a standard, semi-automated method (Biorobot M48 workstation, Qiagen Inc, Valencia, CA). Patient DNA was digested with AluI and RsaI, labeled with the fluorescent dye Cy3, mixed together with a similarly processed, sex-mismatched, pooled reference DNA differentially labeled with the Cy5 dye, and hybridized to the EmArray Cyto6000 array. DNA digestion, labeling and hybridization were performed as recommended by the manufacturer of the array (Agilent Oligonucleotide-Based Array CGH for Genomic DNA Analysis, Protocol version 4.0 June 2006; Agilent Technologies, Inc., CA), with some minor modifications as described in Baldwin et al , 2008. Slides were scanned and fluorescent images were captured using a GenePix 4200A scanner and GenePix-Pro 6.1 software (Axon Instruments/Molecular Devices Corp., Union City,CA). Array images were then imported and evaluated by Agilent Feature Extraction 9.5 software. The Cy3/Cy5 ratios were calculated and analyzed using Agilent CGH Analytics 3.5 software to determine copy number differences and/or aberrations between the patient DNA and the sex mismatched DNA. The location of oligonucleotide probes on the EMArray Cyto6000 array, and the demarcation of regions exhibiting copy number changes was according to Genome Build UCSC hg 17 assembly (Build 35, May 2004). Copy number changes were confirmed by Fluorescent In Situ Hybridization (FISH) analysis when needed and possible.
Results
PDC Activity
Persistently elevated pyruvate and lactic acidosis prompted evaluation of PDC activity. Analysis of PDC activity in peripheral blood lymphocytes showed normal activity, with activated PDC activity within one standard deviation of the previously established mean values (Table 1). However, activity in skin fibroblasts was 1.6 standard deviation below the mean of prior controls, and the ration of PDC to E3 activity was depressed, as well (Table 2).
PDH E1ï¡ sequencing
Decreased PDC activity suggested a mutation of one of the PDC components, of which PHDA1 is the most commonly affected. Sequencing of the PDHA1 gene did not reveal any known deleterious mutations. Several variants were detected. Three variants in the intronic regions are previously reported polymorphisms (NCBI rs11278403, rs7058209, couldn't find rs for intron 10). A variant in exon 8 was also detected, but did not result in any amino acid change (NCBI rs5955757). All detected variants were homozygous.
Genomic analysis
CMA analysis revealed a female chromosomal profile with a 1.1 Mb loss of genomic material at Xp22.13-p22.12 (ChrX:18588606-19671947) represented by 18 oligonucleotide probes . The region of loss encompasses 4 genes, PHKA2, GPR64, PDHA1, and SH3KBP1, and partially overlaps several 5' exons of the PPEF1 gene (see Figure 1).
The presence of a genomic deletion in Xp was confirmed by FISH analysis using a labeled bacterial artificial chromosome (BAC) probe, RP11-574D3, located at Xp22.13 (data not shown).
Discussion
The patient presented here illustrates common features of the psychomotor-retardation subtype of PDC deficiency [2]. Loss of any component of the PDC may produce a spectrum of findings from infantile death with severe lactic acidemia and cystic degeneration of the cerebral cortex to intermittent ataxia with mild carbohydrate-sensitive lactic acidemia. In this case, PDC deficiency was suggested from the neonatal period by ventriculomegaly and elevated lactic acid levels with normal lactate/pyruvate ratio in blood, without other significant abnormalities in amino or organic acids upon repeated testing. As she has grown, the patient has illustrated common features of the psychomotor-retardation subtype of PDC deficiency, including agenesis of the corpus callosum, cortical atrophy and ventriculomegaly, profound developmental delay, and lactic acidemia exacerbated by infection and illness. Though her seizures and central apneas suggested focal pathology, no necrotizing lesions suggesting Leigh encephalopathy were found on repeated MRI imaging.
These signs of clinical PDC deficiency prompted biochemical evaluation of PDC function. Assays demonstrated normal PDC function in lymphocytes. PDC activity in fibroblasts, however, was reduced. Disparate biochemical function in different tissues in a female patient is consistent with an X-linked disease. Indeed, the most common disorder of the PDC, PDH E1ï¡ deficiency, is encoded by PDHA1 on the X chromosome, and is often difficulty to diagnose in female patients [18, 27].
Sequencing of the PDHA1 gene did not reveal any known deleterious mutations. Numerous mutations of PDHA1 have been identified, including missense mutations of the coding sequence , exonic mutations leading to exon-skipping, insertions, and small deletions [5, 19, 20, 28-33]. Although none of these abnormalities were detected in this case, homozygosity for all SNPs that were detected did suggest that only a single allele was sequenced, due either to deletions of entire exons with deletion breakpoints outside the sequenced region, or deletion of the entire PDHA1 gene. A deletion of multiple exons has been previously described: a 4.2 kb deletion spanning intron 5 to intron 9 in a female patient with developmental delay, cortical atrophy, and partial agenesis of the corpus callosum [21].
Since sequencing of PDHA1 did not reveal any abnormality, we assessed copy number variation through array CGH. Array CGH utilizing a relatively low resolution bacterial artificial chromosome (BAC) based array containing 1887 BAC probes covering 622 loci platform had been performed when the patient was two years of age, and was normal. Thus, array CGH was repeated with an oligonucleotide array offering both wider genomic coverage and higher resolution. Oligonucletide array CGH revealed a loss of 1.1 Mb at Xp22.13-p22.12, which was verified by FISH (data not shown). The patient's family declined parental FISH analysis, so it is unknown if this deletion was inherited or arose de novo. Given the heterogeneity of PDH E1ï¡ deficiency in female patients, it is possible that the patient's mother is an asymptomatic carrier of this deletion.
In addition to PDHA1, the region of loss also encompasses the PHKA2 (Phosphorylase kinase ï¡2) gene. Mutations in PHKA2 cause glycogen storage disease 9A (also known as X-linked glycogenosis, OMIM #306000), which presents with generally benign hepatomegaly, elevated transaminases, and hypercholesterolemia even in affected males [36, 37]. These biochemical abnormalities were not observed in our patient, most likely due to X-inactivation.
In addition to demonstrating a whole gene deletion of PDHA1, this case further underscores the utility of array-based CGH, and particularly newer oligonucleotide arrays, in detecting abnormalities in gene dosage. Array-based CGH has recently been used to detect deletions in cases where characteristic clinical presentation strongly suggests a genetic disorder, but gene sequencing fails to demonstrate an abnormality, or a specific gene has not yet been identified [22, 23, 38-41]. Future application of this technology to the cohort of patients with biochemical evidence of PDC deficiency but negative genetic testing may reveal additional microdeletions of the Xp22.12 region as a common cause of PDH E1ï¡ deficiency. (Maybe Dr. Kerr has something more specific to say based on his large case series).
