A biallelic pathogenic variant in the OGDH gene results in a neurological disorder with features of a mitochondrial disease

Abstract

2-Oxoglutarate dehydrogenase (OGDH) is a rate-limiting enzyme in the mitochondrial TCA cycle, encoded by the OGDH gene. α-Ketoglutarate dehydrogenase (OGDH) deficiency was previously reported in association with developmental delay, hypotonia, and movement disorders and metabolic decompensation, with no genetic data provided. Using whole exome sequencing, we identified two individuals carrying a homozygous missense variant c.959A>G (p.N320S) in the OGDH gene. These individuals presented with global developmental delay, elevated lactate, ataxia and seizure. Fibroblast analysis and modeling of the mutation in Drosophila were used to evaluate pathogenicity of the variant. Skin fibroblasts from subject # 2 showed a decrease in both OGDH protein and enzyme activity. Transfection of human OGDH cDNA in HEK293 cells carrying p.N320S also produced significantly lower protein levels compared to those with wild-type cDNA. Loss of Drosophila Ogdh (dOgdh) caused early developmental lethality, rescued by expressing wild-type dOgdh (dOgdh^WT) or human OGDH (OGDH^WT) cDNA. In contrast, expression to the mutant OGDH (OGDH^N320S) or dOgdh carrying homologous mutations to human OGDH p.N320S variant (dOgdh^N324S) failed to rescue lethality of dOgdh null mutants. Knockdown of dOgdh in the nervous system resulted in locomotion defects which were rescued by dOgdh^WT expression but not by dOgdh^N324S expression. Collectively, the results indicate that c.959A>G variant in OGDH leads to an amino acid change (p.N320S) causing a severe loss of OGDH protein function. Our study establishes in the first time a genetic link between an OGDH gene mutation and OGDH deficiency.

1 |. INTRODUCTION

Mutations in genes encoding for proteins involved in the tricarboxylic acid (TCA) cycle have been reported to cause severe pediatric onset neurological disorders including; ACO2 (MIM: 100850)^1,2 SUCLG1 (MIM: 611224),³ SUCLA2 (MIM: 603921),^4,5 SDHA (MIM: 600857),⁶ SDHD (MIM: 602690),⁷ FH (MIM: 606812),⁸ and MDH2 (MIM: 154100).⁹ For example, ACO2 mutations lead to a neurological phenotype consisting of hypotonia, athetosis, ataxia, severe developmental delay, seizures, optic nerve atrophy and has been called, infantile cerebellar-retinal degeneration (ICRD [MIM: 614519]).¹ Mitochondrial DNA depletion syndrome 9 (encephalomyopathy type with methylmalonic aciduria) (MIM: 245400) is caused by SUCLG1 mutations.¹⁰ Mitochondrial DNA depletion syndrome 5 (encephalomyopathy type +/− methylmalonic aciduria) (MIM: 612073) results from SUCLA2 mutations.¹¹ Mutations in the SDHA and SDHD genes result in complex II deficient Leigh syndrome (psychomotor regression associated with basal ganglia and brainstem lesions) (MIM: 256000 and 252011, respectively). Interestingly, mutations in SDHA, SDHB, SDHC and SDHD can predispose to a variety of tumors including paragangliomas, pheochromocytomas and GIST tumors.^12–15 Mutations in the malate dehydrogenase gene (MDH2) result in early infantile epileptic encephalopathy (EIEE51 [MIM: 617339]); while, FH mutations lead to fumarase deficiency associated with a progressive encephalopathy and dystonia (MIM: 606812).¹⁶ All the above TCA disorders are autosomal recessive (often first cousin scenarios) and not surprisingly, showed mitochondrial dysfunction as common biochemical (ie, lactic acidemia, mtDNA depletion, respiratory chain deficiency) and phenotypic (Leigh disease, epilepsy, psychomotor regression, encephalopathy, epilepsy, and dystonia) features.

Although not as severe as the above disorders, retinitis pigmentosa type 46 (MIM: 612572) is caused by mutations the TCA cycle enzyme isocitrate dehydrogenase 3, beta-subunit, gene, IDH3B (MIM: 604526).¹⁷ To date, specific gene mutations leading to ophthalmological or neurological disorders have not been reported in association with other TCA cycle enzymes including; citrate synthase (CS, [MIM: 118950]) or oxoglutarate dehydrogenase (OGDH [MIM: 613022]). Oxoglutarate dehydrogenase (OGDH) is also known as α-ketoglutarate dehydrogenase (α-KGDH) and mutations in OGDH have been implicated, but not confirmed, in four publications describing α-ketoglutarate (2-oxo-glutarate) dehydrogenase enzymatic deficiency associated with hypotonia, hypertonia, developmental delay, movement disorders and hyper-lactic acidemia (MIM: 203740).^18–21 The disorder was presumed to be autosomal recessive given that three of the four publications described consanguinity with the parents of the affected children.^18,20,21 Herein, we describe two siblings with neurological features similar to the four publications of “α-ketoglutarate dehydrogenase deficiency.”^18–21 Functional studies in fibroblasts from affected individuals and Drosophila melanogaster revealed that the variant leads to a significant decrease in the levels and function of OGDH both in vitro and in vivo. Hence, we confirm the molecular basis of “α-ketoglutarate dehydrogenase deficiency” to be an autosomal recessive disorder due to pathogenic mutations in the OGDH gene.

2 |. MATERIALS AND METHODS

2.1 |. Patient and data collection

Patient #1 was referred at 3 years of age for an unsteady gait. He was the fifth child born by spontaneous vaginal delivery at term to first cousin consanguineous parents of Syrian decent. He showed delays in gross motor function, not walking until 2 years of age and now has multiple falls and cannot run. Fine motor skills were delayed and he is still unable to use zippers or buttons and has difficulty putting a spoon in his mouth. His speech was delayed and he did not start talking until 2½ years of age but is now speaking a few words of English and can communicate in short sentences and understand simple commands in Arabic. Cranial nerve examination was normal and there were no dysmorphic features. Motor exam showed normal bulk, power with mild hypotonia. Knee and ankle reflexes were normal with no clonus and plantar response was flexor. Coordination examination demonstrated definite finger nose dysmetria and a broad-based ataxic gait. His examination was unchanged at 6 years of age with the exception of some upper extremity dystonia. Investigations showed variably elevated lactate (2.9, 1.8; NR = 2.2 mmol/L), with normal values for; blood tests (creatine kinase, copper, vitamin B12, homocysteine, acyl-carnitines, ceruloplasmin, liver function tests, iron, ferritin, ammonia) and urine (oligosaccharides, organic acids, creatine, guanidinoacetate [GAA], and mucopolysaccharide screen). Magnetic resonance imaging (MRI) scanning of the brain at age 3 years found only mild ventriculomegaly with no parenchymal changes on T1, T2 and STIR images.

Patient #2 is the eldest sister of patient # 1 and was born at term by spontaneous vaginal delivery. Development appeared normal at 8 months when she was starting to crawl and pull herself to stand but had a fall and hit her head. There was no loss of consciousness and a CT scan showed no bleed. Co-temporal with this event she became hypotonic, developed dystonic movements, lost the ability to sit and crawl and did not develop speech and has remained wheelchair bound. At age 11 years she developed generalized tonic-clonic seizures and was put on phenobarbital and has remained seizure free for 3 years. On examination (age 17 years), she responded to her parents by smiling and appeared to have some stranger anxiety. She fixed and followed well and fundi were normal. She had orofacial dystonia and made some nonspecific noises but had no audible words. She had generalized muscle atrophy while lying in her wheelchair with arms, hips and knees in a flexed position with rigidity and dystonia but no spasticity or clonus and a flexor plantar response. All the above tests that her brother had plus plasma amino acids and acylcarnitine were normal. Lactate was very high when tested months apart (6.8, 7.2 mmol/L). MRI scan showed symmetric atrophy of the fronto-temporal lobes and basal ganglia with cystic changes noted in the basal ganglia and high FLAIR signal in the lentiform nuclei, internal and external capsule (Figure 1). A neurological examination was completed on the three clinically unaffected children (ages 11, 14, 16 years) and that was normal.

FIGURE 1.

FLAIR imaging MRI from patient # 2. A, White arrows = widening of the Sylvian fissure indicating cortical atrophy; black arrows = high FLAIR signal in the lentiform nuclei, internal, and external capsule. B, White arrows = simplified gyral pattern in frontal lobes; black arrows = cystic changes in the basal ganglia

The parents signed informed consent for all aspects of the testing in both patients under Hamilton Integrated Ethics Board Project #15-266-T. Patient # 2 had a skin biopsy completed from the right volar forearm under local anesthetic and this was cultured for subsequent fibroblast analysis (see below). Briefly, the skin sample was placed into fibroblast media (10% FBS, MEM, 1x HEPES, 1x L-glutamine, 1% penicillin/streptomycin; Gibco, Thermo Fisher Scientific, Waltham, Massachusetts) at 4°C for less than 12 hours. The samples were cut into smaller pieces and placed in a humidified incubator at 37°C and 5% CO₂. Every 4 days, 4 mL of media were added until the well was 90% confluent. Cells were allowed a second passage for out-growth (to reduce keratinocyte contamination) were frozen in 10% DMSO/ fibroblast media and placed into liquid nitrogen for long term storage.

2.2 |. Genetic testing

A 180 k oligonucleotide microarray for patient #2 showed no copy number variations. Mitochondrial DNA sequencing of blood DNA from patient # 2 was normal as was a 406 gene nuclear encoded mitochondrial disease next generation sequencing panel that did not include the OGDH gene (https://www.newbornscreening.on.ca/sites/default/files/mitochondrial_testing_gene_list_2019_1.pdf). Given the negative testing, parental consanguinity, and similar phenotype in two individuals, we completed whole exome sequencing (WES) on blood DNA from both patients and their parents at a CLIA certified commercial laboratory (GeneDx, Gaithersburg, Maryland), using genomic DNA from the patient #2, her similarly affected brother (patient #2) and parents. The exonic regions and flanking splice junctions of the genome were captured using the Clinical Research Exome kit (Agilent Technologies, Santa Clara, California). Massively parallel sequencing was done on an Illumina system with 100 bp or greater paired-end reads. Reads were aligned to human genome build GRCh37/UCSC hg19, and analyzed for sequence variants using a custom-developed analysis tool. Additional sequencing technology and variant interpretation protocol has been previously described.²² The general assertion criteria for variant classification are publicly available on the GeneDx ClinVar submission page (http://www.ncbi.nlm.nih.gov/clinvar/submitters/26957/). Targeted Sanger sequencing was further carried out for variant confirmation and segregation of the variant in other unaffected siblings in this family. The minimum read depth for 100% of the OGDH gene was covered at 100% at a minimum of ×10 reads and no other rare variants were reported.

2.3 |. Cell culture and transfection

Primary skin fibroblast cultures were obtained from patient # 2, two controls (control fibroblasts (C1) (female, 11 years, healthy individual, GM2036, Coriell) and control fibroblasts (C2) (S38, female, 20 years, healthy individual)) and HEK 293 cells were culture in Dulbecco’s modified Eagle’s medium (DMEM, Life Technologies) supplemented with 10% fetal bovine serum (FBS, Sigma, F0926), 1% penicillin and streptomycin (Life Technologies) at 37° C in a 5% CO₂ atmosphere. HEK293 cells were cultured in 6 well plate up to 60% to 70% of confluency. Thirty minutes before transfection DMEM medium was replaced with serum free Opti-MEM medium. pcDNA3.1-OGDH^WT or OGDH^N320S construct together with GFP expression construct −pk2xGFP were transfected into HEK293 cells using Lipofectamine 3000 (Thermo fisher) and incubated for 4 hours. Then medium was changed back to DMEM medium and after 48 hours, the cells were harvested in RIPA buffer with protease and phosphatase inhibitors (GenDEPOT).

For construction of pcDNA3.1-OGDH(WT)-Flag, the OGDH coding sequence was amplified by PCR from the pUASTattb-OGDH^WT-Flag vector (Yoon et al. 2017) and cloned into pcDNA3.1(+) vector. PCR primers were 5′-GCGGCCGCAAA ATGTTTCATTTAAGGACTTGT-3′ and 5′-CTCGAG TTACTTGTCGTCATCGTCCTTGTAAT CAATA TCGTGGTCCTTGTAGTC GCCGCTTCC CGA GAAGTTCTTGAAGACGTC-3′. pcDNA3.1-OGDH(N320S)-Flag was generated by site-directed mutagenesis PCR using the following primers:

OGDH (N320S)-F 5′-cacagagggcggctgaacgtgcttgcaAGTgt catcaggaaggagctggaacagatc-3′.

OGDH (N320S)-R. 5′-gatctgttccagctccttcctgatgacACTtgca agcacgttcagccgccctctgtg-3′.

2.4 |. Biochemical analysis

Control and patient’s fibroblasts were cultured in 10 cm dishes up to full confluency. Next cells were washed once with 10 mL cold PBS, scraped with 750 μL cold PBS and collected into one microcentrifuge tubes. Samples were then centrifuged for 5 minutes at 1200 rpm at 4°C, the supernatant was discarded, and the pellet was snap frozen in liquid nitrogen and stored at −80°C. Cellular OGDH and mitochondrial NADH:ubiquinone oxidoreductase (complex I) activity were measured as previously described^23,24 with minor modifications. Briefly, the frozen cell pellets were resuspended in 25 mM MOPS at pH 7.4, thawed, and snap-frozen and thawed again. OGDH activity was assayed spectrophotometrically as the rate of NAD⁺ reduction to NADH (340 nm, ε = 6200 M⁻¹ cm⁻¹) in the presence of 2.5 μM rotenone and 0.05% Triton X-100 upon addition of 5.0 mM MgCl₂, 2.5 mM α-ketoglutarate, 0.1 mM CoASH, 0.2 mM thiamine pyrophosphate (TPP), and 1.0 mM NAD⁺. Complex I activity was measured spectrophotometrically as the rate of NADH oxidation in the presence of 2.5 μM antimycin A and 100 μM ubiquinone-1 following the addition of 150 μM NADH. The specificity for complex I activity was confirmed by inhibition with a Fe-S cluster inhibitor p-hydroxymercuribenzoate. Sample protein concentrations were determined after the assay by the BCA (bicinchoninic acid) method using BSA as a standard.

2.5 |. Immunoblotting analysis

Western blots were performed using 4% to 20% Mini-Protein TGX Protein gels (BioRad). Primary antibodies used were: rabbit anti-OGDH (1:1000; ab137773, Abcam), Rabbit anti-DLST (1:1000; HPA003010, Sigma), mouse anti-ATP5A (1:2000; ab14748, Abcam), and mouse-anti-actin (clone:C4) (1:10 000; 8691002, MP Biomedicals). For quantification of proteins expressed from the pcDNA3.1-OGDH (WT or N320S) in HEK293 cells, we used a capillary electrophoresis-based protein analysis system (WES; ProteinSImple, San Jose, California). Cell extracts (0.1 mg/mL) were separated and visualized using the standard instrumental protocol. Primary antibodies used were: rabbit anti-OGDH (1:10, ab137773, Abcam), and rabbit anti-GFP (1:500, A-11122, ThermoFisher).

2.6 |. Immunostaining

Fibroblasts were cultured on poly-D-lysine precoated glass coverslips (NC0301187, Neuvitro Corp.) and fixed with 4% formaldehyde in PBS for 10 minutes at room temperate. Cells were washed three times with PBS-T (Triton X-100, 0.3%). The cells were incubated with 5% goat serum in PBS-T for 1 hour for blocking. Then the cells were incubated in the primary antibody—ATP5A antibody (1:500; ab14748, Abcam) in blocking solution for overnight. Next day, the cells were washed 5 times with PBST and then incubated with secondary antibody—donkey anti-Mouse IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (IF) (1:500; A21202, Life Technologies Corp.) for 1 hour. DAPI (1:1000, 1 mg/mL, Sigma) was applied for 30 minutes.

For fly larvae staining, third instar larvae dissection was performed as described in Bellen and Budnik.²⁵ Briefly, third instar larvae were fixed in 4% formaldehyde in PBS for 30 minutes at room temperature, rinsed with PBS, and washed in PBS-T three times. The antibodies used at the following concentration: chick anti-GFP (1:5000; ab13970, Abcam), Alexa Fluor 488 AffiniPure goat anti-chicken IgY (1:1000; Jackson Immunoresearch Lab). The coverslips and larvae samples were mounted in Vectashield (Vector Labs). Imaging was performed using LSM710 confocal miscroscope (Zeiss).

2.7 |. Drosophila strains and maintenance

yw; dOgdh-T2A-Gal4, UAS-dOgdh^WT-Flag (II), and UAS-OGDH^WT-Flag (II) were generated as previously described.²⁶ The following stocks were obtained from the Bloomington Stock Center at Indiana University (BDSC) and VDRC: yw; n-syb-Gal4 (BDSC 51635), w; 20xUAS-IVS-mCD8:GFP (BDSC 32194), dOgdh RNAi (v12778). All flies were maintained at room temperature (21°C). All crosses were kept at 25°C.

2.8 |. Cloning and transgenesis

To generate pUASTattb-dOgdh(N324S)-Flag and pUASTattb-OGDH (N320S)-Flag, we performed site-directed mutagenesis using the pUASTattB-dOgdh^WT-Flag, and pUASTattB-OGDH^WT-Flag as templates²⁶ and the following primers:

OGDH (N320S)-F. 5′-cacagagggcggctgaacgtgcttgcaAGTgtcatcaggaaggagctggaacagatc-3′

OGDH (N320S)-R. 5′-gatctgttccagctccttcctgatgacACTtgcaagcacgttcagccgccctctgtg-3′

dOgdh (N324S)-F. ggacgtcttaacaccttggccAGTgtatgccgcaagcccttgaac

dOgdh (N324S)-R. gttcaagggcttgcggcatacACTggccaaggtgttaagacgtcc

The mutagenized clones were confirmed by Sanger sequencing. The pUASTattB constructs were injected into yw pC31; VK37 embryos by Phi-C31-mediated transgenesis²⁷ and transgenic flies were selected and balanced.

2.9 |. Drosophila climbing assay

Flies were collected using CO₂ and allowed to rest in food vials for at least 18 hours prior to assay. Male and female flies were kept separately as gender difference on behavior might be significant. To prepare the climbing apparatus, measure a distance of 8 cm from the bottom surface of an empty polystyrene vial and mark the distance by drawing a line around the entire circumference of the vial. Twenty-five flies were transferred without using CO₂ into different climbing apparatus for each genotype. The apparatus was closed off by vertically joining to another empty polystyrene vial using tape and the flies were allowed to acclimatize to the surrounding for at least 10 minutes. Then, the apparatus was gently tapped five times rapidly to allow all the flies fall to the bottom of the apparatus and a video was recorded for 20 seconds. After a 5 minutes rest, the assay was repeated. Three trials were conducted. After testing, the flies were transferred back into food vials and maintained until the next test.

3 |. RESULTS

3.1 |. Identification of the OGDH variant

Next Generation Sequencing (NGS) testing revealed a homozygous missense variant in exon 8 of the OGDH gene (NM_002541.3), c.959A>G; p.Asn320Ser (N320S), in both affected children, and confirmed heterozygosity in each parent (Figure 2A). Sanger sequencing (GeneDx, Gaithersburg, MD) was completed in these unaffected siblings and all three were heterozygous for the N320S missense variant in OGDH (Figure 2A). An atomic model of the ODGH protein was predicted by SWISS-Model Server (https://swissmodel.expasy.org) using the PDB code: 2Y0P as a template.²⁸ Further characterization of the potential pathogenic variant was carried out in silico, in vitro (fibroblasts), and in vivo (Drosophila) (see below).

FIGURE 2.

OGDH variant in silico analysis. A, A pedigree of two affected probands, three unaffected siblings and parents show segregation of the homozygous OGDH variant c.959A>G (p.Asn320Ser) in subject 1 and subject 2. B, The p.Asn320Ser substitution alters the evolutionarily conserved Asn residues in multiple species including Homo sapiens and Drosophila. C, A schematic of protein domains of OGDH and position of Asn320Ser mutation. D, Homology model of OGDH shows the Asn320 residue contributes to the stability of the α-helix which is connected to the loop carrying histidine 311 and arginine 312 that is required for thiamine pyrophosphate (TPP) binding

The p.N320S variant in OGDH gene has not been reported previously as a pathologic variant or a benign variant and is not found in large population databases including EXAC and gnomAD dataset²⁹ (https://www.biorxiv.org/content/10.1101/531210v3). The asparagine at position 320 in OGDH is highly conserved in various species from human to yeast (Figure 2B). In silico prediction programs including PROVEAN (Protein Variation Effect Analyzer), SIFT,^30,31 and polyphen2^32,33 predicted this change to be deleterious.

The asparagine at residue 320 is located in the TPP binding domain that is required for binding to thiamine pyrophosphate (TPP), a cofactor for the OGDH enzymatic reaction (Figure 2C). Homology modeling for OGDH protein revealed that the N320S substitution may destabilize the α-helix that is connected to the loop containing histidine 311 and arginine 312 that bind to the phosphate group of TPP (Figure 2D). Collectively, although this rare missense, c.959A>G; p.N320S, was classified as a variant of uncertain significance in the clinical report due to limited information of this gene and variant, the molecular evidence including the domain it resides prompted further functional characterization to clarify the significance of this variant.

3.2 |. Functional studies in fibroblasts from patient # 2

To investigate the functional effects of the variant in OGDH, we sought to determine whether the levels and function of OGDH were affected in the fibroblasts obtained from subject #2. We used fibroblasts obtained from age and sex matched healthy individuals as controls. Western blot analysis revealed that patient # 2 (S2) fibroblasts have reduced levels of OGDH protein, the E1 subunit of the α-KGDH complex, as compared to fibroblasts from controls (Figure 3A). In contrast, the protein levels for DLST (dihydrolipoamide S-succinyltransferase), the E2 subunit of the α-KGDH complex, were comparable to controls (Figure 3A). We also found that levels of ATP5A (ATP synthase, H+ transporting, mitochondrial F1 complex, β-subunit), a subunit of complex V, was also similar to controls. Hence, the data suggested that the variant (p.N320S) in OGDH affected the expression and/or stability of the OGDH protein, but not other mitochondrial proteins such as DLST and ATP5A.

FIGURE 3.

The p.N320S variant in OGDH impairs the protein stability and activity of OGDH. A, Western blots and quantification for protein levels of OGDH, DLST, ATP5A, and actin in protein extracts from subject 2 cultured fibroblasts and those from two healthy controls. (n = 10). B, WES capillary electrophoresis-based protein assays detected the expression of Flag-tagged OGDH expression and GFP expression in HEK293 cells. The OGDH protein levels were normalized to GFP. (n = 6). C and D, The activities of OGDH and complex I were measured from total cell extracts from subject 2 cultured fibroblasts and control fibroblasts 1 (OGDH activity, n = 6; complex I activity, n = 5). E, Confocal micrographs of subject 2 cultured fibroblasts and control fibroblasts 1. ATP5A antibody labels mitochondria (green) and DAPI labels nuclei (blue). Scale bars, 20 μm. Error bars indicate SEM. P values were calculated using Student’s t test. ***P < .001

To exclude the genetic background effects in the patient fibroblasts, we sought to confirm that the p. N320S variant directly affects OGDH protein levels. We therefore transfected a mammalian expression construct carrying C-terminally Flag-tagged wild-type OGDH cDNA (pcDNA3.1-OGDH^WT-Flag) or OGDH cDNA carrying p. N320S variant (pcDNA3.1-OGDH^N320S-Flag) into HEK293 cells. A GFP expression construct was co-transfected, which served as normalized control for transfection efficiency. To quantify the protein amount, we performed capillary immunoassay and found that the OGDH protein levels from OGDH^N320S-Flag transfected cells was approximately half of that of OGDH^WT-Flag transfected cells (Figure 3B). This result confirmed that the N320S variant of OGDH was sufficient to lower protein levels.

To test whether a decrease in OGDH protein has a functional consequence, we assayed OGDH complex enzyme activity from the patient # 2 (S2) and control fibroblasts. As shown in Figure 3C, the OGDH complex activity was 50% in S2 as compared to controls. This effect was specific to OGDH complex activity, as the activity of complex I (CI) of the mitochondrial electron transport chain was comparable to those in controls (Figure 3D). Finally, we assessed mitochondrial morphology by performing immunostaining and found that the S2 fibroblasts exhibited normal mitochondria morphology (Figure 3E), showing that OGDH deficiency does not alter mitochondrial morphology. Collectively, the above results indicated that the p.N320S variant in OGDH causes a decrease in OGDH protein content and activity, thus, confirming the underlying pathophysiology in the patients.

3.3 |. Functional studies in vivo using Drosophila

We next determined the functional effects of the p. N320S variant upon the OGDH protein in vivo. In our previous published work, we utilized a recombination-mediated cassette exchange (RMCE) technique and created a loss-of-function mutant for Drosophila Ogdh (dOgdh) by introducing an artificial exon carrying a splicing acceptor (SA)-T2A-Gal4-polyA into the coding intron of dOgdh genomic locus (dOgdh-T2A-Gal4) (Figure 4A).^26,34,35 These mutant flies express Gal4 in the proper temporal and spatial fashion under the control of endogenous dOgdh cis-regulatory elements. Gal4 is a transcriptional activator that induces expression of transgenes by binding to UAS (Upstream Activating Sequence) (Figure 4A). Flies carrying dOgdh-T2A-Gal4 and UAS-mCD8:GFP showed neuronal and muscle expression of dOgdh in both larvae and adult flies (Figure 4B). Homozygous mutants for dOgdh-T2A-Gal4 or transheterozygous for dOgdh-T2A-Gal4/Df(3 L) ED4674 died in embryonic stages, but this lethality was rescued by expression of wild-type dOgdh or wild-type human OGDH²⁶ (Figure 4C). Thus, these known functional effects of OGDH depletion in Drosophila make them an excellent model for OGDH mutation studies.

FIGURE 4.

Biallelic OGDH variant impairs the in vivo function of dOgdh in Drosophila. A, A schematic of transgene expression mediated by dOgdh-T2A-Gal4 allele. A T2A-Gal4 cassette consisting of a splice acceptor (SA), a linker (L), a ribosomal skipping T2A peptide (T2A), and a Gal4 coding sequence, and a polyadenylation signal (pA), was inserted to a coding intron of dOgdh genomic locus by RMEC. B, Expression of dOGDH in adults and larvae. dOGDH is highly expressed in ventral nerve cord (VNC) in larva, and head in adult. C, The lethality caused by dOgdh loss (homozygous for dOgdh-T2A-Gal4) was rescued by OGDH^WT or dOgdh^WT expression, but not by cDNA carrying the patient variant (OGDH^N320S or dOgdh^N324S). D, dOgdh knockdown in neurons exhibited progressive defects in climbing ability, which was fully rescued by expressing dOgdh^WT, but not by dOgdh^N324S. Error bars indicate SEM (n = 3). P values were calculated using Student’s t test. ***P < .001. N.S. indicates not statistically significant

To test the effect of the p.N320S, we generated transgenic flies harboring UAS-human OGDH cDNA carrying N320S (UAS-OGDH^N320S), and fly cDNA carrying a homologous mutation for N320S (UAS-dOgdh^N324S). We found that expression of OGDH^N320S or dOgdh^N324S failed to rescue the lethality caused by dOgdh loss, indicating that N to S variant is a loss of function allele. To further test the effect of p. N320S on the nervous system, we utilized an RNA interference (RNAi) strategy. We found that knockdown of dOgdh in the nervous tissues by a neuronal specific Gal4 driver (n-syb-Gal4>UAS-dOgdh RNAi) resulted in progressive locomotion defects, which were fully rescued by expression of wild-type dOgdh, but not by dOgdh^N324S (Figure 4D). Hence, our results indicate that the p.N320S variant is a loss of function pathogenic variant causing the pathology seen in the patients.

4 |. DISCUSSION

This appears to be the first reported case of a pathogenic variant (c.959A>G) in the OGDH gene resulting in a pathogenic amino acid substitution (p.N320S), leading to an infantile-onset neurologic disease. Importantly, we have confirmed the pathogenicity both in vivo and in vitro. Similar to most of the other TCA cycle disorders associated with neurological disease^1–9 OGDH pathogenic variants also display clinical, brain MRI and metabolic features of a primary mitochondrial disease. This report brings to total number of TCA cycle enzymes associated with disease to seven of the eight, leaving citrate synthase (CS) as the only enzyme not yet associated with an OMIM disease (https://www.omim.org). Given that CS is the enzyme catalyzing entry to the TCA cycle it is possible that pathogenic biallelic pathogenic variants are embryologically lethal.

Pathogenic variants in the OGDH gene have been implied, but not confirmed, in three reports of children with a severe infantile onset neurological disorder associated with α-ketoglutarate dehydrogenase deficiency (MIM: 203740).^18,20,21 One report described two children born to consanguineous parents with a reduction in 2-oxoglutarate dehydrogenase activity at 25% of control values in both children.²¹ The children both had slight elevations of 2-oxoglutarate in blood but no urine values were presented. Clinically, the children were born at term with no issues during pregnancy or at delivery. The boy developed global developmental delays but did have the ability to walk with an ataxia gait but later lost the ability to walk. There was dyscoordination with dystonia, rigidity, no spasticity, no clonus, and toes were downgoing. The sister had normal early development with slight delay in walking (18 months) but then developed stiffness, dystonia, and “signs of spasticity were absent.” Genetic testing was not completed.²¹ Two other sibling cases of α-ketoglutarate dehydrogenase deficiency were reported in a boy and a girl born to consanguineous parents who developed encephalopathy, hypotonia, failure to thrive and demonstrated lactic acidosis, which got worse with superimposed emotional stress and infections.²⁰ The girl was doing well until 7 months of age when she developed more hypotonia and at 12 months developed encephalopathy, ataxia, chorioathetotic movements and died at 10 years of age with severe acidosis after general anesthesia. The boy was born following normal pregnancy and delivery. He developed hypotonia at 5 months of age and showed deterioration starting at 10 months of age with an inability to walk and at 4 years of age had severe encephalopathy, psychotic behavior, severe axial hypotonia with some pyramidal signs. He was also able to fix and follow and smile but could not speak. Brain CT scanning showed bilateral symmetric decreased signal of the periventricular white matter. α-ketoglutarate dehydrogenase activity (OGDH) was 15% of control values in fibroblasts and 7.5% of control values in skeletal muscle.²⁰ One case series described three children with congenital lactic acidosis and hypotonia with two dying before 3 years of age and the other child severely growth retarded and spastic at 20 months of age.¹⁸ Finally, a case was reported of a 14 month old male with opisthotonos, hypertonia and a normal brain MRI with variably elevated urinary 2-oxo-glutarate and an OGDH activity that was 5% of controls.¹⁹ Given the biochemical evidence and neurological phenotype in the latter four cases we strongly suspect that the reports of “2-oxoglutaric aciduria and 2-ketoglutarate dehydrogenase deficiency” (collectively called α-ketoglutarate dehydrogenase deficiency, MIM: 203740), were likely due to biallelic pathogenic variants in the OGDH gene.

In addition to the biochemical and molecular data, the similarity of the clinical phenotype of our patients and those described above with OGDH deficiency further supports that there is a genotype/phenotype correlation for the OGDH gene variant (Table 1). The older sibling in our cohort (patient #2) showed a rapid decline in development following a minor head trauma and became mute, spastic and dystonic and this neurological clinical picture is similar to the OGDH enzyme deficiency patients previously described.^18,20,21 The younger sibling in our cohort (patient # 1) has a milder clinical phenotype with developmental and speech delays and dystonia and this fits with some of the initial features in patients with OGDH deficiency prior to their clinical decline,^19–21 and with the relatively mild features of the child with intermittent elevations of urinary 2-oxo-glutarate.¹⁹ The significant changes on MRI for the more affected patient (#2, Figure 1) is also congruent with the more severe clinical picture as compared to patient #1 who displayed a milder phenotype and a fairly normal MRI (mild ventriculomegaly). Although the urine organic acids were reported as “normal” in the two patients presented in our study, we had the biochemical geneticist who reported the results review the data for each patient and he did say that the TCA cycle intermediates that were proximal to the OGDH catalyzed step were higher than expected (2-oxo-glutarate and cis-aconitate) relative to the lower abundance for the distal compounds (succinate and fumarate). Furthermore, quantitative urinary 2-oxoglutarate measurements can be normal,¹⁸ or variably elevated,¹⁹ even with severe reductions in OGDH activity.

TABLE 1.

Summary of clinical features of patients with OGDH deficiency

Subject	Onset	Last examination	Abnormal tone	Developmental delay	Ref.
1—male	8 mo	5 y	Rigidity, spastic, ataxic, athetosis	Y-moderate	21
2—female	18 mo	3 y	Rigidity	Y-moderate	21
3—male	Birth	10 mo	Rigidity, hypotonia, athetosis	Y-severe, died—32 mo	18
4—male	Birth	5 mo	Spastic, rigidity, hypotonia	Y-severe, died—30 mo	18
5—male	Birth	20 mo	Hypotonia, spasticity, dystonia (face)	Y-severe	18
6—female	7 mo	12 mo	Hypotonia, athetosis	Y- ? severe died—10 y	20
7—male	5 mo	4 y	Hypotonia, spasticity	Y-severe	20
8—male	Birth	14 mo	“Hypertonia” (likely spastic and rigid)	N	19
9—female	8 mo	17 y	Hypotonia, dystonia (face), rigidity	Y-severe	Current article
10—male	2 y	6 y	Hypotonia, ataxia, dystonia (arm)	Y-moderate	Current article

N = Normal development to date.

Y = Yes they had developmental delay.

It is not clear whether the pathology of OGDH deficiency is due the accumulation of toxic intermediates proximal to the OGDH enzyme and/or a mitochondrial energy deficiency. Arguing against a simple mitochondrial energy deficiency is our previous observation that a decrease in OGDH activity did not cause reduced levels of ATP, mitochondrial membrane potential, or NADH levels.²⁶ This study showed that the accumulation of 2-oxoglutarate leads to higher mTORC1 activity in Drosophila and murine cells with reduced OGDH activity, and treating Ogdh mutant flies with rapamycin, an mTOR inhibitor partially rescues the neurodegenerative phenotypes.²⁶ Thus, these data indicated that a significant part of the pathology of OGDH mutations is 2-oxoglutarate accumulation leading to higher mTOR activity. However, anecdotal evidence for at least some contribution from mitochondrial energy deficiency in vivo comes from the very similar neurological features shared by most of the reported TCA enzyme defects (developmental delay/psychomotor regression, basal ganglia associated movement disorders, ataxia, exacerbation by infection or other stressor, lactic acidemia, basal ganglia, and other Leigh-like MRI imaging changes) implicating that a final common pathophysiological process is a reduction in electron transport chain flow with lower aerobic ATP production. An interesting feature that could explain some variance in the timing of the psychomotor regression and/or the severity of the disorders is the presence of trauma/stressors.²⁰ For example, patient # 2 in our pedigree did not appear to have any symptoms until a fall and a head trauma and then had a very rapid psychomotor regression. It is known that head trauma is associated with mitochondrial dysfunction and oxidative stress,³⁶ and these could be a precipitating trigger for the rapid decline observed in patient # 2. A sudden decline in function is a common feature in most of the patients reported with OGDH enzyme deficiency.^18–21

To confirm the pathogenicity of the c.959A>G (p. N320S) variant in the OGDH gene we first examined fibroblasts obtained from patient #2. We found that the patient fibroblasts carrying p.N320S exhibited lower protein levels of OGDH, which in turn lead to lower OGDH enzymatic activity (50% normal). Although the enzyme activity was higher than those reported in other cases of OGDH deficiency (2%-25%),^18–21 patient # 2 is clinically stable (albeit mute and spastic) now at 17.5 years of age and this vastly exceeds the ages at last examination,^19,21 or death,^18,20 in other reports. These results were confirmed in HEK293 cells expressing OGDH carrying the N320S variant. The levels of OGDH preprotein (upper bands in Figure 3A) in patient fibroblasts (S2) are comparable to those for control fibroblasts, whereas the levels of the processed OGDH (lower bands in Figure 3A) in the patient fibroblasts are significantly lower than those for controls. Nuclear-encoded mitochondrial proteins including OGDH are translated in preprotein forms containing a mitochondrial targeting sequence (MTS). Upon mitochondrial import, the MTS is cleaved by the mitochondrial processing peptidases, and processed proteins are folded to form functional proteins by help with the mitochondrial chaperone system^37,38 Misfolded proteins are typically degraded by the mitochondrial peptidase system.³⁹ Our homology model of human OGDH showed that the side chain of the asparagine 320 (N320) contributes to the stability of the α-helix by forming hydrogen bonds with the neighboring asparagine 316 residue. (Figure 2D). Thus, substitution of asparagine to serine in position 320 would be expected to weaken these hydrogen bonds formation, which may affect the α-helix stability and OGDH folding. Hence, our data suggest that the N320S variant does not affect translation in the cytosol, but may lead to defective protein folding and subsequent degradation of the protein in the mitochondria.

To test the functionality of p.N320S variant with in vivo models, we utilized Drosophila. RNA expression dataset showed that dOgdh is expressed ubiquitously (http://marrvel.org/search/gene/ogdh).⁴⁰ Using the dOgdh-T2A-Gal4 allele, we found that dOgdh expression was higher in the ventral nerve cord in larvae brain and adult heads as compared to other tissues (Figure 4B). A complete loss of dOgdh led to embryonic lethality, indicating that dOgdh is an essential gene in development. The lethality of dOgdh loss was rescued by wild-type dOgdh or OGDH expression. However, expression of cDNA carrying N320S (or N324S in dOgdh) variant could not rescue the embryonic lethality. Patients carrying recessive variant for OGDH (N320S) exhibited early-onset neuromuscular and metabolic manifestation, but it did not cause lethality. Drosophila have a single OGDH homolog—dOgdh, whereas vertebrates, including humans, have two OGDH paralogs—OGDH and OGDH-like (OGDHL). The redundant OGDH genes may partially compensate for a deficiency in the other.

Abnormal activities of TCA cycle enzymes are associated with many types of cancer. High levels of OGDH is shown to associated with gastric cancers and OGDH knockdown interferes with cancer cell proliferation through reducing β-catenin signaling.⁴¹ In the other hands, loss of OGDH was shown to stabilize hypoxia-inducible transcription factor 1α (HIF1α), a key protein for cancer proliferation in hypoxia condition.⁴² Burr et al. showed that OGDH defects leads to L-2-hydroxygularate (L-2-HG) formation, which inhibits the activity of HIFα prolyl hydroxylases (PHDs), an enzyme that is required for HIF1α degradation in normoxia condition.⁴² PHDs belongs to the α-KG-dependent dioxygenases family that includes histone and DNA demethylases. These enzymes require α-KG for their activities, whereas L-2-HG that is derived from α-KG, succinate, and fumarate inhibit the activities of the enzymes.⁴³ Germline mutations in fumarate hydratase (FH) and succinate dehydrogenases (SDHA, SDHB, SDHC, and SDHD) are associated with renal cell cancer and gastric cancers where the levels of HIF1α is overexpressed probably due to high levels of fumarate, and succinate⁴⁴; hence, changes of TCA cycle enzyme activities play a key role in cancer progression by modulating HIF1 and probably DNA and histone demethylases. Whether and how genetic mutations in TCA cycle genes lead to the neurological diseases through altering epigenetic landscapes and/or HIF1α pathway will be needed in future studies. Furthermore, it is unclear whether TCA cycle mutations that result in severe neurological phenotypes (such as OGDH) increase the risk of carcinogenesis for the emergence of a clinically detectable tumor is less likely to appear given a shortened lifespan.

In summary, our data confirms that biallelic pathogenic variants in the OGDH gene alters the activity of the OGDH enzyme and leads to an autosomal recessive disorder associated with some or all features of infantile and pediatric onset basal ganglia associated movement disorders, hypotonia, developmental delays, ataxia, and seizures.