Neurodevelopmental regression, severe generalized dystonia, and metabolic acidosis caused by POLR3A mutations

Objective To expand the clinical phenotype of POLR3A mutations by assessing the functional consequences of a missense and a splicing acceptor mutation. Methods We performed whole-exome sequencing for identification of likely pathogenic mutations in a 9-year-old female patient with severe generalized dystonia, metabolic acidosis, leukocytosis, hypotonia, and dysphagia. Brain MRI showed basal ganglia atrophy and presence of lactate and lipid peaks by [1H]-magnetic resonance spectroscopy. Expression levels of Pol III target genes were measured by quantitative real-time (qRT)-PCR to study the pathogenicity of the biallelic mutations in patient fibroblasts. Results The patient is a compound heterozygous for a novel missense c.3721G>A (p.Val1241Met) and the splicing region c.1771-6C>G mutation in POLR3A, the gene coding for the catalytic subunit of RNA polymerase III (Pol III). Aberrant splicing was observed for the c.1771-6C>G mutation. Decreased RNA expression levels of Pol III targets (HNRNPH2, ubiquitin B, lactotransferrin, and HSP90AA1) were observed in patient fibroblasts with rescue to normal levels by overexpression of the wild-type protein but not by the p.Val1241Met variant. Conclusions Mutations in the POLR3A gene cause POLR3A-related hypomyelinating leukodystrophy with or without oligodontia or hypogonadotropic hypogonadism (HLD7, OMIM: 607694) and neonatal progeroid syndrome (OMIM: 264090), both with high phenotypic variability. We demonstrated the pathogenicity of c.1771-6C>G and c.3721G>A mutations causing an early-onset disorder. The phenotype of our patient expands the clinical presentation of POLR3A-related mutations and suggests a new classification that we propose designating as Neurodevelopmental Disorder with Regression, Abnormal Movements, and Increased Lactate.

POLR3A encodes the catalytic subunit A of Pol III, which is responsible for the constitutive transcription of transfer ribonucleic acids, mitochondrial RNA-processing RNA, 5S, H1, and noncoding RNAs (ncRNAs), and involved in translation of several mRNA. 1,2 Because of its involvement in the transcription of so many different RNAs, it is not surprising that mutations in POLR3A can lead to a variety of phenotypes.
Two different conditions are associated with POLR3A mutations: HLD7 (OMIM: 607694) and neonatal progeroid syndrome (NPS, OMIM: 264090). HLD7 is an autosomal recessive early-onset leukodystrophy 3 presented by hypomyelination, hypogonadotropic hypogonadism, hypodontia, spasticity, dystonia, and neurodevelopmental regression. 1,4 Like HLD7, NPS is also an autosomal recessive early-onset disorder presenting a wide range of phenotypes including macrocephaly, progeroid appearance, triangular face, micrognathia, nystagmus, hypodontia, muscle atrophy, hypertonia, and agenesis of the corpus callosum. 5 Here, we present the c.3721G>A and c.1771-6C>G mutations in POLR3A, with features not described in HLD7 nor NPS. We propose that these mutations extend the phenotype of POLR3A deficiency and create a new classification of patients with features similar to those presented here.

Case report
The proband is a 9-year-old girl with a healthy mother who had no other pregnancies and a father diagnosed with depression. Gastroesophageal reflux and ineffective breast sucking were observed just after birth. When aged 5 months, she was apathetic and presented generalized dystonia. Four months later, she refused food with no swallowing disorder. Protein malnutrition was observed, and oral administration of hypercaloric diet was started with ensuing of metabolic acidosis (lactate 4.88, reference: 0.63-2.44 mmol/L, normal pH and decreased HCO 3 21.4 reference: 22-29 mEq/L). She could sit but neither crawl nor walk.
Brain MRI, muscle biopsy, and karyotype were normal at age 1 year. Improvement of weight gain and hydration was achieved by nasogastric tube feeding. Gastrostomy was performed but showed dumping, requiring thickening of diet.
She presented with recurrent lung infections, milestone regression, and was unable to talk at 2 years. Worsening of dystonia and hypodontia was observed, with the absence of 2 baby teeth, but permanent dentition developed.
Another MRI was performed at age 4 years (data not shown) with total brain volume reduction and alterations in striatal bodies. Four years later, new MRI showed bilateral hyperintensity in T2 (figure 1A, a-d) and FLAIR (figure 1A, e-g), atrophy of the caudate nucleus and putamen with compensatory enlargement of the lateral ventricles. Absent enhancement on postcontrast images (figure 1A, h-i) and increase of lipids/lactate on [ 1 H]-magnetic resonance spectroscopy were observed, but the noisy background could contribute to possible artifactual peaks (figure 1A, j).
EEG showed slight diffuse disorganization in background activity, reflecting a diffuse cerebral dysfunction, with no association with specific pathologic features. No epileptiform activity was observed.
Analysis of WES data DNA was extracted from blood (DNeasy Kit; Qiagen). Wholeexome sequencing (WES) was prepared (Illumina ® Nextera) and run (HiSeq 2000). Human GRCh37 reference genome and a GATK-based pipeline 6 were used. Sanger sequencing was performed by Source BioScience, United Kingdom.
qRT-PCR RNA was extracted using the Invitrogen TRIzol kit. cDNA was prepared with reverse transcriptase (Invitrogen). For POLR3A mRNA processing, cDNA was amplified by PCR, separated by agarose gel electrophoresis, gel extracted, and Sanger sequenced. qRT-PCR of target genes was performed in duplicate with Life Technologies TaqMan Assays (Applied Biosystems): ncRNA (HNRNPH2, Hs01395062_m1) and mRNAs: ubiquitin B (UBB), lactotransferrin (LTF), and HSP90AA1.

Statistical analysis
The differences among groups were calculated with 95% confidence intervals (p < 0.05) using the one-way analysis of variance and Tukey multiple comparison test. All data are presented as mean ± SD of the mean (SD) of 3 biological independent experiments in duplicates.
Data availability Data will be available upon request.  pathogenic (score 0.99). The valine residue in the 1,241 position is highly conserved in metazoans with bilateral symmetry, except in some species, which present isoleucine instead ( figure 1D).

WES detected 2 mutations in
Electrophoresis from control and proband fibroblasts detected 2 transcripts of shorter length along with the canonical full-length mRNA (figure 2A). The longest product corresponded to the full-length mRNA, the one right below it showed deletion of exon 14, whereas the smallest one displayed the combined deletion of both exons 13 + 14 ( figure  2B). An accumulation of shorter compared with full-length products was more evident in the patient cells.
We tested the dependency of the patient's fibroblast phenotype on defective POLR3A by expressing HA-tagged POLR3A, either wild type (P POLR3A_WT_1 and P POLR3A_WT_2 ) or mutant (P POLR3A_V1241M_1 , P POLR3A_V1241M_2 , and P POLR3A_V1241M_3 ), using the empty vector (P EV ) as a negative control. Compared with the control (Ctrl), patient-derived cell lines (P and P EV ) displayed low levels of POLR3A, whereas overexpressing cell lines presented higher levels of the protein, as also demonstrated by anti-HA antibody immunovisualization (figure 3A).
A significant decrease in the levels of the HNRNPH2, UBB, LTF, and HSP90AA1 was observed in the patient's fibroblasts compared with controls in all cases except for HSP90AA1 compared with one of the controls, C1 (figure 3B-E). As expected, no significant difference was detected between P cells and P EV . Furthermore, patient cells overexpressing the wild-type protein recovered basal or higher levels of Pol III target genes. This was not the case for patient cells overexpressing the p.V1241M variant, as their levels were similar to or even lower than both patient cell lines, P and P EV .

Discussion
Here, we present a case report of a compound heterozygous for c.1771-6C>G and c.3721G>A mutations in POLR3A. The patient presented clinical features associated with HLD7 and NPS, but also metabolic acidosis, leukocytosis, lipid and lactate peak.
The c.1771-6C>G mutation is located 6 nucleotides upstream exon 14 of POLR3A, resulting in the loss of this exon. It is classified as a VUS (ClinVar), and it has been reported in a family with HLD7, causing basal ganglia atrophy with no hypomyelination. 3 Loss of exon 14 has been linked to reduced expression of the splicing factor HNRNPH2, responsible for the regulation of pre-mRNA splicing. 3 Expression levels of HNRNPH2 were reduced in patient's fibroblasts, and we showed that overexpression of WT POLR3A, but not of the mutated p.Val1241Met variant, rescued the expression of HNRNPH2. Decreased expression of UBB, LTF, and HSP90AA1 has been previously found in other patients carrying the c.1771-6C>G mutation in POLR3A. 3 All the mRNAs expression levels were decreased in the patient cells, with rescue of their expression levels when WT but not p.Val1241Met POLR3A was overexpressed, demonstrating the pathogenicity of the c.3721G>A mutation.
No other VUS with correlation between the patient's phenotype and previously associated disease genes was identified in the nuclear genome. In mtDNA, the mutation m.14831G>A was found in a patient with Leber hereditary optic neuropathy, but no functional studies were performed. 9 This variant was recently classified as benign in ClinVar following the American College of Medical Genetics guidelines 10 (evidence codes: BS1, BS2, and BP4).
The clinical phenotype of our patient resembles a combination of symptoms of 2 conditions associated with POLR3A mutations and even mutations in genes coding for POL III targets, 8 highlighting the difficulty of a clinical diagnosis and reinforcing the importance of a detailed clinical report combined with molecular studies.
In conclusion, we suggest a new classification to patients carrying mutations in POLR3A, with neither leukodystrophy nor NPS features, which we propose designating as Neurodevelopmental Disorder with Regression, Abnormal Movements, and Increased Lactate.

Study funding
This study was financed in by the Core Grant from the MRC (Grant MC_UU_00015/5) and by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES) (Finance Code 001).