Abstract
Background and Objectives Coenzyme Q10 (CoQ10)–deficient cerebellar ataxia can be due to pathogenic variants in genes encoding for CoQ10 biosynthetic proteins or associated with defects in protein unrelated to its biosynthesis. Diagnosis is crucial because patients may respond favorably to CoQ10 supplementation. The aim of this study was to identify through whole-exome sequencing (WES) the pathogenic variants, and assess CoQ10 levels, in fibroblasts from patients with undiagnosed cerebellar ataxia referred to investigate CoQ10 deficiency.
Methods WES was performed on genomic DNA extracted from 16 patients. Sequencing data were filtered using a virtual panel of genes associated with CoQ10 deficiency and/or cerebellar ataxia. CoQ10 levels were measured by high-performance liquid chromatography in 14 patient-derived fibroblasts.
Results A definite genetic etiology was identified in 8 samples of 16 (diagnostic yield = 50%). The identified genetic causes were pathogenic variants of the genes COQ8A (ADCK3) (n = 3 samples), ATP1A3 (n = 2), PLA2G6 (n = 1), SPG7 (n = 1), and MFSD8 (n = 1). Five novel mutations were found (COQ8A n = 3, PLA2G6 n = 1, and MFSD8 n = 1). CoQ10 levels were significantly decreased in 3/14 fibroblast samples (21.4%), 1 carrying compound heterozygous COQ8A pathogenic variants, 1 harboring a homozygous pathogenic SPG7 variant, and 1 with an unknown molecular defect.
Discussion This work confirms the importance of COQ8A gene mutations as a frequent genetic cause of cerebellar ataxia and CoQ10 deficiency and suggests SPG7 mutations as a novel cause of secondary CoQ10 deficiency.
Glossary
- ACMG=
- American College of Medical Genetics;
- AF=
- allele frequency;
- atypical NAD=
- atypical neuroaxonal dystrophy;
- cDNA=
- complementary DNA;
- CLN7=
- ceroid lipofuscinosis type 7;
- CoQ10=
- Coenzyme Q10;
- IRB=
- Institutional Review Board;
- OxPhos=
- oxidative phosphorylation;
- PLAN=
- PLA2G6-associated neurodegeneration;
- SCA=
- spinocerebellar ataxia;
- WES=
- whole-exome sequencing
Coenzyme Q10 (CoQ10) deficiency is a clinically and genetically heterogeneous mitochondrial disorder,1 defined by the low CoQ10 levels in muscle, fibroblasts, or other tissues. CoQ10 is a lipid molecule important for several biological processes including oxidative phosphorylation (OxPhos), reactive oxygen species detoxification, sulfide oxidation, and synthesis of pyrimidine nucleotides.2 Cerebellar ataxia, variably associated with other neurologic and systemic symptoms, is one of the most common clinical presentations of CoQ10 deficiency.1 It was initially described in 6 patients in 2001.3 In 2003, a study4 reported that approximately 10% of patients with cerebellar ataxia of undefined etiology displayed low levels of CoQ10 in muscle. Based on these findings, we collected biological samples from patients with genetically undiagnosed cerebellar ataxia referred to investigate CoQ10 deficiency.
CoQ10 deficiency in cerebellar ataxia can be primary as well as secondary to mutations in genes unrelated to its biosynthesis.1,5 The most frequent cause of primary CoQ10-deficient cerebellar ataxia is pathogenic variants in COQ8A gene (previously known as ADCK3),6 while secondary CoQ10 deficiency has been reported in association with mutations in ANO10,7 APTX,8,-,10 SLC25A26,11 GLUT1,12 and in several other OxPhos and non-OxPhos disorders.12 Remarkably, CoQ10 supplementation is partially effective in both primary and secondary cases.13
In this study, we report the results of the biochemical assessment of CoQ10 content and whole-exome sequencing (WES) study in a cohort of fibroblasts from patients with undiagnosed cerebellar ataxia and referred to investigate CoQ10 deficiency.
Methods
Standard Protocol Approvals, Registrations, and Patient Consents
The Ethics Committee of the IRCCS Foundation Ca' Granda Ospedale Maggiore Policlinico (Milan, Italy) and the Institutional Review Board (IRB) of Columbia University (New York) approved the study. Written informed consent for genetic analyses, anonymous publication of the patients' clinical features, and analyses of biological samples was obtained from all involved patients or their legal guardians/parents under Columbia University Medical Center IRB–approved protocol. Because most samples were sent from patients followed up by outside physicians, complete clinical data were not available for all of them (eTable 1, links.lww.com/NXG/A583). Fibroblasts were collected from 16 patients referred to investigate CoQ10 deficiency based on the referring physicians' clinical suspicion and/or favorable response to CoQ10 supplementation.
CoQ10 Measurement
Levels of total CoQ10, reduced (ubiquinol-10) and oxidized (ubiquinone-10), were measured by high-performance liquid chromatography in 14 patients (2 lines were not available), and 3 control cultured skin fibroblasts as previously described.14 Cells were grown in triplicate, unless otherwise specified. CoQ10 levels were compared with our laboratory's normal values (48.7 ± 9.6, n = 45). The Mann-Whitney nonparametric U test was used to compare gene expression and CoQ10 levels between patient and control fibroblasts, and a p value of <0.05 was considered statistically significant.
Genetic Analyses
Genomic DNAs were extracted from 16 skin fibroblasts. Friedreich ataxia, spinocerebellar ataxia (SCA)1, SCA2, and SCA3 were excluded before undergoing subsequent analyses. WES was performed using the Nextera Exome Library protocol (Illumina) and an Illumina NextSeq500 sequencing platform. Variant calling and annotation were performed by GATK4 and ANNOVAR tools. Sequencing data were filtered using a virtual gene panel containing the genes associated with CoQ10 deficiency and/or cerebellar ataxia, looking for rare variants (minor allele frequency <0.0002) with protein impact (eTable 2, links.lww.com/NXG/A584). Variants were confirmed by Sanger sequencing.
Total RNA was isolated from cultured skin fibroblast using NucleoSpin RNAII (Macherey-Nagel Supplier), and complementary DNA (cDNA) was synthesized with SuperScript1 VILO cDNA Synthesis Kit (Invitrogen). The mRNA levels of COQ8A and PLA2G6 were measured by reverse transcription quantitative real-time PCR using an Applied Biosystems StepOne TM RealTime PCR System Thermal Cycling Block. Gene expression levels were determined in triplicate using TaqMan Fast Advanced Master Mix and normalized to GAPDH. Patients' haplotypes were resolved by cloning a single allele with TOPO-TA Cloning Kit for Sequencing (Invitrogen) and subsequent Sanger sequencing. In brief, cDNA was amplified by PCR with appropriate primer pairs depending on the set of mutations present. The alleles were cloned into pCR4-TOPO vector and transformed into OneShot competent cells. Bacteria were plated on lysogeny broth agar plate supplemented with 50 μL/mL of kanamicin; after overnight growth at 37°C, single colonies were picked and grown in lysogeny broth supplemented with 50 μg/mL of kanamicin for at least 3 hours at 37°C in agitation. Colony PCR was used to screen proper transformants for subsequent sequencing. PCR products were cleaned by Exonuclease I and FastAP treatment (Thermo Scientific) before amplification with BigDye terminator v3.1 Cycle Sequencing kit using as sequencing primer the former pair from the first amplification or the pair M13F and T3. Sequences were precipitated by acetate/ethanol procedure and sequenced on 3130xl GeneticAnalyzer (ABI Prism). The biallelic status of mutations was confirmed if at least 1 allele sequenced over the positions of both mutations was found to carry only one of them. Data not provided in the article because of space limitations may be shared (anonymized) at the request of any qualified investigator for purposes of replicating procedures and results.
Results
A definite genetic etiology was identified in 8 samples of 16 (diagnostic yield = 50%). The identified genetic causes were pathogenic variants affecting COQ8A (ADCK3) (n = 3 samples), ATP1A3 (n = 2), PLA2G6 (n = 1), SPG7 (n = 1), and MFSD8 (n = 1). The genetic results and available clinical data are summarized in Table 1. Causative genetic mutations were found in the following samples: CU19001, CU19004, CU19008, CU19010, CU19012, CU19013, CU19014, and CU19015. Conversely, a definitive genetic diagnosis could not be obtained in samples CU19002, CU19003, CU19005, CU19006, CU19009, CU19011, CU19016, and CU19017.
Clinical and Genetic Summary of the Samples in Which the Pathogenic Variants Were Identified
We found significantly decreased levels of CoQ10 in 3 of 14 cell lines (21.4%) (i.e., CU19001, CU19002, and CU19008) (eTable 1, links.lww.com/NXG/A583). CoQ10 levels were decreased also in CU19003 (eTable 1, links.lww.com/NXG/A583); however, the decrease was not statistically significant because we could measure CoQ10 only in 1 biological replicate due to a severe growth defect of this fibroblast line.
COQ8A (ADCK3)
Sample CU19001 carried 2 heterozygous COQ8A variants (National Center for Biotechnology Information transcript identifier: NM_020247.5): c.827A > G (p.Lys276Arg) and c.1799_1800insC (p.Val600fs*127), which were demonstrated to be in trans. The c.827A > G is a known (rs1223030341) but extremely rare missense variant (gnomAD allele frequency [AF] = 0.000014). It is predicted to be pathogenic by all the in-silico prediction tools tested (i.e., CADD, DANN, MutationTaster, Polyphen2, and SIFT). This variant has been reported already as a likely pathogenic variant in a single report15 and was identified in another patient of our cohort (i.e., CU19015). It meets the following pathogenicity criteria of the American College of Medical Genetics (ACMG): PM1, PM2, PP5, and PP3 (pathogenic). The second variant (c.1799_1800insC) is a novel frameshift mutation absent from population databases (e.g., gnomAD and ExAC). It is a null variant predicted as pathogenic by all in silico tools. It meets the following ACMG criteria: PVS1, PM2, and PP3 (pathogenic). Unsurprisingly, low levels of CoQ10 were found in sample CU19001.
Three heterozygous COQ8A variants were called by bioinformatic tools in sample CU19014: c.125delC (p.Leu43Cysfs*166), c.128T > A (p.Leu43Gln), and c.895C > T (p.Arg299Trp). However, the c.125delC and the c.128T > A were associated in cis as demonstrated by their co-occurrence in the same sequencing reads; therefore, the correct nomenclature of the mutation was c.127_128delinsA (Leu43Serfs*166). Hence, the patient carried only 2 heterozygous variants (c.127_128delinsA and c.895C > T), which were proved to be associated in trans. The first variant is a novel frameshift mutation absent from population databases. It is a null variant predicted as pathogenic by in silico tools. It meets the PVS1, PM2, and PP3 ACMG criteria (pathogenic). The second variant (c.895C > T) is a known pathogenic variant already associated with recessive cerebellar ataxia and reported in genetic databases (e.g., ClinVar).
Sample CU19015 was found to carry 2 compound heterozygous COQ8A variants: c.827A > G (p.Lys276Arg) and c.1748delC (p.Thr584Profs*7). The first variant was already characterized for sample CU19001. The second variant is a novel frameshift mutation absent from population databases and predicted to be pathogenic by in silico prediction tools. Regarding ACMG criteria, it meets the following: PVS1, PM2, and PP3 (pathogenic). The 3 patients from which the fibroblasts CU19001, CU19014, and CU19015 were derived had low levels of CoQ10 in muscle, but only CU19001 displayed decreased levels of CoQ10 in fibroblasts (eTable 1, links.lww.com/NXG/A583).
A single heterozygous c.1751C > T (p.Thr584Ile) COQ8A variant was identified in sample CU19006. It is a known (rs779649324) but extremely rare genetic variation (gnomAD AF = 0.000008). It is predicted pathogenic by most in silico tools; however, a single heterozygous mutation is not sufficient to cause the disease. No reduction of COQ8A mRNA levels in patient fibroblasts compared with controls was observed, making less likely the hypothesis of a highly deleterious cryptic variant involving the other allele. In the absence of another causative COQ8A mutation, at the moment, the c.1751C > T should be considered a variant of unknown significance.
ATP1A3
Samples CU19004 and CU19012 carried the known pathogenic c.2452G > A (p.Glu818Lys) ATP1A3 mutation (NM_152,296.5). Heterozygous ATP1A3 mutations are associated with at least 3 phenotypes: rapid-onset dystonia parkinsonism, alternating hemiplegia of childhood, and cerebellar ataxia, areflexia, pes cavus, optic atrophy, and sensorineural hearing loss (CAPOS syndrome).16 So far, the p.Glu818Lys is the only mutation associated with CAPOS syndrome.16 The clinical presentation of the patient from which CU19004 sample derived was compatible with CAPOS syndrome. He was seen at the Columbia Neuromuscular Center at the age of 12 years, presenting with obvious hearing loss, mild distal weakness, moderate cerebellar dysarthria, oculomotor dyspraxia with broken pursuit, and bilateral dysmetria of upper limbs. Reflexes were absent. Romberg sign was present. He had an ataxic gait and was unable to walk tandem. The neurologic abnormalities started at the age of 2 years, when a febrile illness caused the patient to be weak and flaccid and led to admission to an intensive care unit. Strength gradually improved partially over the following years. A second similar episode of generalized weakness occurred at the age of 5 years in association with another febrile illness, probably due to a Salmonella infection. After this second event, hearing loss and a slurred speech were first documented, and he progressively developed the aforementioned clinical picture, predominated by cerebellar ataxia, areflexia, and sensorineural hearing loss. No clinical information was available for sample CU19012. Patients' muscle biopsies were not available. CoQ10 levels in fibroblasts CU19004 and CU19012 were normal.
SPG7
Sample CU19008 harbored a homozygous c.1715C > T (p.Ala572Val) SPG7 variant (NM_003119.4), which is a known pathogenic mutation reported in genetic databases (e.g., ClinVar) and already associated with an ataxic presentation of spastic paraplegia 7.17 Muscle biopsy was not available, but fibroblasts showed significantly decreased CoQ10 levels (36.9 ± 5.1, 76%) (eTable 1, links.lww.com/NXG/A583).
PLA2G6
In sample CU19010, 2 PLA2G6 variants were found (NM_003560.4): c.2222G > A (p.Arg741Gln) and c.1186C > T (p.Leu396Phe). Biallelic PLA2G6 mutations are the genetic cause of PLA2G6-associated neurodegeneration (PLAN), which comprises a clinical continuum of 3 phenotypes: infantile neuroaxonal dystrophy, atypical neuroaxonal dystrophy (atypical NAD), and PLA2G6-related dystonia-parkinsonism.18 Cerebellar ataxia and cerebellar atrophy can be predominant clinical features in PLAN, especially in atypical NAD. The c.2222G > A is a known PLA2G6 pathogenic variant already associated with PLAN and reported in genetic databases (e.g., ClinVar).19 The c.1186C > T is a novel missense variant absent from population databases (transcript NM_003560.4). Of interest it represents a truncating variant (p.Gln396Ter) in the PLA2G6 NM_001199562 transcript, which is predominant in brain, particularly in the cerebellum (Figure). Transcripts analyses demonstrated nonsense-mediated RNA decay of the allele carrying this variant and in-trans association with the c.2222G > A. The variant meets the following ACMG criteria: PVS1, PM2, and PM3 (pathogenic). The clinical presentation of the patient from which the CU19010 sample was derived resembled atypical NAD, being characterized by early-onset cerebellar ataxia with unsteady gait and pyramidal signs followed by progressive psychomotor and cognitive decline. Muscle biopsy of this patient was not available, and CoQ10 levels in fibroblasts were normal.
The upper panel represents wild-type transcripts while the lower panel displays mutated transcripts. On transcript NM_003560, the variant is predicted to cause a missense change (p.Leu396Phe) while on transcript NM_001199562, it generates a premature stop codon causing (p.Gln396Ter) nonsense-mediated RNA decay (NMD) of the allele carrying it in fibroblasts.
MFSD8
A very rare homozygous c.753A > G (p.Glu251 = ) variant of MFSD8 gene was found in sample CU19013 (NM_152,778). Biallelic MFSD8 mutations cause neuronal ceroid lipofuscinosis type 7 (CLN7). The phenotype of CLN7 is characterized by infantile-onset psychomotor developmental regression, cerebellar ataxia, seizures, visual impairment, and myoclonus.20,21 The synonymous c.753A > G is reported at an extremely low frequency in population databases (GnomAD AF = 0.000014), but it is absent from clinical databases, and it has never been associated with human disease. It is located at the splice site, and it is expected to impair the normal transcript splicing. All the in-silico tools for splicing defect prediction (e.g., SpliceAI, VarSeak, and AdaBoost) indicate this variant to be splice disruptive (i.e., splicing donor loss and exon skipping). The patient presented clinically with cerebellar ataxia, myoclonus, and visual defect. Muscle biopsy was not available, and CoQ10 levels in fibroblasts were normal.
Discussion
In this study, using WES, we genetically characterized a collection of fibroblasts derived from patients with cerebellar ataxia referred to investigate CoQ10 deficiency. We identified the genetic cause of ataxia in 8 of 16 patients. The results of the genetic analysis were unexpected because only 3 patients had genetic mutations in a gene (i.e., COQ8A) known to cause cerebellar ataxia and CoQ10 deficiency. The remaining genetically diagnosed samples carried mutations in ataxia-causative genes never previously associated with CoQ10 deficiency.
Low levels of CoQ10 were found in 21.4% of fibroblasts, confirming the results of the previous screening by Lamperti et al.4 who measured CoQ10 concentration in muscle biopsies from 135 patients with genetically undefined cerebellar ataxia and found it decreased in approximately 10%. The higher proportion of CoQ10 deficiency in our cohort was expected because it was composed of fibroblasts of patients with cerebellar ataxia specifically referred for the investigation of CoQ10 deficiency.
Regarding COQ8A mutations, the results of this study appear to be in line with the previous literature, confirming their role as the most frequent cause of primary CoQ10 deficiency.5 COQ8A gene is located on chromosome 1q42.13 and comprises 15 exons. It is ubiquitously expressed, with greater abundance in the heart, skeletal muscle, and cerebellum.22 The encoded COQ8A protein is an atypical kinase involved in ubiquinone biosynthesis. It probably acts as a lipid kinase that phosphorylates a prenyl lipid intermediate in the CoQ10 biosynthesis pathway.23 More than 20 COQ8A pathogenic variants have been reported, including missense variants, null mutations, and a large deletion (from exon 3 to exon 15). All the pathogenic variants are private, and no founder effect has been identified so far.5 In this study, we provide 3 novel pathogenic COQ8A mutations: c.1799_1800insC (p.Val600fs*127), c.127_128delinsA (Leu43Serfs*166), and c.1748delC (p.Thr584Profs*7).
A single sample carried a homozygous pathogenic SPG7 mutation, known to cause spastic paraplegia 7. Surprisingly, CoQ10 levels were significantly decreased in the fibroblasts of this patient. SPG7 protein is an inner mitochondrial membrane transmembrane protein, which is part of the m-AAA metalloproteinase complex that plays a role in mitochondrial protein quality control.24,25 Muscle biopsy performed in patients with unexplained ataxia, who were later found to have mutations in SPG7, showed cytochrome c oxidase–negative fibers, multiple mtDNA deletions, and, interestingly, CoQ10 deficiency in a single patient.26 Therefore, a pathogenic link between SPG7 mutations and CoQ10 deficiency can be hypothesized.
Two samples carried a pathogenic ATP1A3 mutation (e.g., p.Glu818Lys) previously associated with CAPOS syndrome, a very rare inherited ataxia form. Another sample carried compound heterozygous mutations of PLA2G6, the causative gene of PLAN. One of the identified mutations was novel (i.e., c.1186C > T, p.Leu396Phe). Finally, a novel homozygous MFSD8 mutation was found in 1 sample (i.e., c.753A > G, p.Glu251 = ). Biallelic pathogenic variants of this gene are associated with CLN7, a rare form of ceroid neuronal lipofuscinosis. In these cases, CoQ10 deficiency was suspected based on clinical suspicion and/or a favorable response to oral CoQ10 supplementation, but no biochemical evidence of CoQ10 deficiency was present in any of these cases. Therefore, links between mutations in these genes (i.e., ATP1A3, PLA2G6, and MFSD8) and CoQ10 deficiency are not supported by our findings.
CoQ10 measurement in skin fibroblasts as a diagnostic tool to identify ataxic patients with CoQ10 deficiency is less sensitive than the same measurement in muscle. Indeed, 2 patients with COQ8A mutations displayed normal CoQ10 values in fibroblasts while showing reduced CoQ10 levels in muscle. Given the invasiveness of this procedure, the first-line approach for the diagnosis of inherited cerebellar ataxias with suspected CoQ10 deficiency, in the absence of other specific clinical manifestations such as myopathic signs, should be the same as for other undiagnosed inherited ataxias, namely direct genetic testing, especially with a next-generation sequencing approach due to its low cost, the relatively high diagnostic yield, and the minimal invasiveness. Molecular diagnosis of a known CoQ10-deficient cerebellar ataxia represents a strong indication for CoQ10 supplementation.
In conclusion, this study describes 5 novel pathogenic mutations (COQ8A n = 3, PLA2G6 n = 1, MFSD8 n = 1) and confirms the importance and prevalence of COQ8A (ADCK3) gene mutations as a genetic cause of cerebellar ataxia and CoQ10 deficiency and further suggests that SPG7 gene mutations are a possible novel genetic determinant of secondary CoQ10 deficiency. Although further studies are necessary to understand the causes and role of secondary CoQ10 deficiency in cerebellar ataxias, diagnosis of CoQ10 deficiency is important because patients with primary and secondary CoQ10-deficient cerebellar ataxia may clinically respond to CoQ10 supplementation.6,-,8,13,27
Study Funding
This study was partially supported by a National Ataxia Foundation (NAF) Research Seed-Money, DoD GRANT12894332 (CMQ) and by the Italian Ministry of Health (Ricerca Corrente 2022 to A. Di Fonzo).
Disclosure
The authors report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/NG.
Acknowledgment
The authors are grateful to all the patients and relatives for their participation. The authors thank all the clinicians who referred patients and samples to us.
Appendix Authors
Footnotes
Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/NG.
The Article Processing Charge was funded by the authors.
Submitted and externally peer reviewed. The handling editor was Editor Stefan M. Pulst, MD, Dr med, FAAN.
- Received October 11, 2022.
- Accepted in final form January 4, 2023.
- Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology.
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
