POLRMT mutations impair mitochondrial transcription causing neurological disease

While >300 disease-causing variants have been identified in the mitochondrial DNA (mtDNA) polymerase γ, no mitochondrial phenotypes have been associated with POLRMT, the RNA polymerase responsible for transcription of the mitochondrial genome. Here, we characterise the clinical and molecular nature of POLRMT variants in eight individuals from seven unrelated families. Patients present with global developmental delay, hypotonia, short stature, and speech/intellectual disability in childhood; one subject displayed an indolent progressive external ophthalmoplegia phenotype. Massive parallel sequencing of all subjects identifies recessive and dominant variants in the POLRMT gene. Patient fibroblasts have a defect in mitochondrial mRNA synthesis, but no mtDNA deletions or copy number abnormalities. The in vitro characterisation of the recombinant POLRMT mutants reveals variable, but deleterious effects on mitochondrial transcription. Together, our in vivo and in vitro functional studies of POLRMT variants establish defective mitochondrial transcription as an important disease mechanism.


Case reports
Patient 1 Patient 1 (P1) is a 17-year-old female of mixed ancestry born to healthy, non-consanguineous unrelated parents. Around one year of age, she presented with genu varum secondary to hypophosphatemic rickets and renal Fanconi syndrome. She was also noted to have mild developmental delays, hypotonia and short stature. Brain MRI at the age 5 years demonstrated mild white matter loss with thinning of the corpus callosum. She was treated with phosphate and calcitriol for her rickets with improved skeletal outcome. She attained final height at age 13 years, with height Z score of -4.1. This patient and her laboratory investigations have been described extensively before in the literature (1) (see also Table 1 ).

Patient 2
Patient 2 (P2) is an 20-year-old Caucasian female. She displayed mild hypotonia and breath holding spells at one month of age and initial examination found bilateral ptosis (left > right). Her gross motor skills were mildly delayed and fine motor and speech skills were found to be behind 1 year. Brain MRI at 8 years was essentially normal and follow up MRI at 17 years of age showed a few scattered nonspecific FLAIR signal hyperintense lesions within the frontoparietal white matter. At age 13 years she was found to have a mild left ear sensorineural hearing loss (SNHL), primarily at higher frequencies. She continues to have mild intellectual disability and a significant impairment in speech and fine motor skills. At age 18 years her neurological examination was normal other than the mild bilateral ptosis, severe left sided SNHL requiring a hearing aid and mild left eye esotropia. She also developed psychiatric symptoms in her teenage years with anxiety and depression and at age 19 years began having psychotic features with severe emotional outbursts and anger and stopped participating in self care.

Patient 3
Patient 3 (P3) is a 58-year-old Caucasian man (weight 72.9 kg, height 1.79 m) who initially presented with a 20-year history of progressive ptosis. Clinical examination revealed bilateral ptosis (<1/3rd of the pupil involved; with evidence of surgical corrective ptosis scars bilaterally) and ophthalmoplegia (restriction of gaze: abduction <60%: adduction: <30%: down gaze 30%; up gaze 10-30%). Proximal myopathy was conspicuously absent. A maternal uncle was reported to have ptosis but this was never clinically-confirmed. Muscle creatine kinase levels were normal at 154 U/L (range 0-220); and lactate levels were normal in blood and CSF (1.1 mmol/l; normal range 0.7-2.1 mmol/l). Brain MRI was normal. Muscle creatine kinase and lactate levels in blood and CSF were normal.

Patient 4
Patient 4 (P4) is a 3-year-old Caucasian boy born to healthy, non-consanguineous parents. Prenatal history was unremarkable. He was born full term at 40 weeks gestational age. Birth weight 3315 g, length 52 cm, and head circumference 36.5 cm. No dysmorphic features on examination. At his last evaluation, he was normocephalic (OFC: 51 cm, 75th centile), height 83 cm (<3rd centile), and weight 12.3 kg (10th centile). He presents with global mild to moderate developmental delay, expressive-receptive language difficulties, short stature, and hypotonia. He was found to have incidentally renal hypomagnesemia (0.6 mg/dl; normal range 1.7-2.4 mg/dl) during workup at age 10 months for febrile urinary tract infection. Magnesium levels have remained low despite supplementation (range 0.6 to 2.1 mg/dl). Phosphorus was also low at 1.5 mg/dl but responded well to supplementation. Renal ultrasound showed medullary nephrocalcinosis. He has persistent elevation of lactate ranging from 2.4 to 3.8 mmol/l (normal range, 0.7-2.1 mmol/l). He developed one episode of anaemia requiring a blood transfusion attributed to transient erythroblastopenia of childhood. No recurrence of anaemia during follow up. He has generalized hypotonia and stable walk. The patient points to communicate and has a vocabulary of just a few words. He is making vocalizations and using few signs with speech therapy. He seems to understand better simple commands at age 3 years.

Patient 5
Patient 5 (P5) is a 10-year-old Iranian male, the only affected child of healthy consanguineous parents. He was born at term with unremarkable pregnancy and neonatal period with OFC 31 cm, weight 2800 g, and length 48 cm. He has a severe global developmental delay including head control: 13months, sitting: 4 years, walking: 8 years and never developed speech. He has no history of seizure. His vision and hearing are normal. At the age of 10 years, weight was 23 kg (-1.6SD), OFC was 45 cm (-6.0 SD) and height was 112 cm (-4 SD). He had normal facial appearance except of strabismus and large ear. He was unable to walk without assistance. Brain MRI showed ventriculomegaly. Developmental testing at age 10 years by Wechsler Intelligence Scale for Children (HAWIK-IV) showed an IQ 20 in the range of severe to profound intellectual disability.

Patient 6
Patient 6 (P6) is a 14-year-old Estonian female, born to healthy non-consanguineous parents. She has healthy younger brother. She was born from first pregnancy on 37th week of gestation with birth weight 2690g (-1 SD), length 47 cm (-1 SD), OFC 32 cm (-1.5 SD) and Apgar score 10/10. On 32nd week of pregnancy, the mother had threat of premature delivery. After birth, bilateral clubfoot was diagnosed and operated and presented with significant muscular hypotonia and feeding difficulties in infancy. At the age of 8 months she presented with developmental delay, muscular hypotonia and some dysmorphic features -epicanthal folds, strabismus, upturned nose, opened mouth appearance. Her height was 66 cm (-2.5 SD), weight 7 kg (-2 SD), OFC 40 cm (-2.5 SD). Metabolic investigations showed mildly elevated serum lactate 2.4 mmol/L and unspecific organic aciduria (increased excretion of Krebs cycle metabolites and methylmalonate). Brain MRI showed cavum septi pellucidi. At the age of 5 years focal epilepsy was diagnosed. She received antiepileptic treatment for five years, but it was stopped due to normal EEG. Presently she has severe intellectual disability and no speech. She has behavioral problems -irritability and aggressiveness. At the age of 14 years, she presented short stature -her height was 139 cm (-4 SD), weight 37,3 kg (-2 SD) and microcephaly -OFC 49.5 cm (-3.5 SD), BMI=19.3 (50 percentile). Hearing is normal. She has normal signs of puberty. She is not able to dress herself, has partial toileting skills and is able to understand some commands. Periodically low folate level was detected and treated with supplementation. Since 14 years of age additional treatment with coenzyme Q10, riboflavin and L-carnitine was started. This treatment showed some positive effect -better sleep at night and less irritability.

Patient 7
Patient 7 (P7) is a 59-year-old male from a non-consanguineous family of Italian ancestry who has an affected male sibling (Patient 8). His weight was 69.9kg, height 1.72 metres and BMI 23.8. He presented for evaluation in his early 40's, when he started experiencing difficulty climbing stairs. Since that time, he developed progressive proximal more than distal weakness of his arms and legs, accompanied by muscle atrophy. He recalls normal early development and was very active in sports during childhood. Neurological examination was notable for sparing of facial and neck muscles, atrophy of the biceps and quadriceps muscles, and diffuse, proximally predominant muscle weakness. There were no elbow or knee contractures, and no calf pseudohypertrophy. Serum creatine kinase levels were elevated 873 U/L (normal 175U/L).

Patient 8
Patient 8 (P8) is the affected 72-year-old male sibling of P7 who has similar clinical and histopathological features. His weight was 65kg, height 1.66 metres and BMI 24.5. He initially presented in his early 40's when he started experiencing difficulty climbing stairs and rising from the floor. Since then his weakness has been slowly progressive with proximal muscle involvement more prominent than distal involvement. There is no facial involvement, but he has neck flexion and extension weakness. He uses a cane for walking short distances and a wheelchair for longer distances. Neurological examination is notable for right exotropia and a few beats of nystagmus on lateral gaze, no facial weakness, head drop with neck flexion and extension weakness, atrophied biceps muscles, no calf pseudohypertrophy, and diffuse weakness that was more prominent in the proximal muscles. Reflexes were depressed throughout and absent at the biceps. There was no calf pseudohypertrophy. Serum creatine kinase levels were mildly elevated to 307 U/L (normal 175U/L).

Next generation sequencing analysis
Whole exome sequencing (WES), whole genome sequencing (WGS) or nuclear gene panel sequencing, followed by filtering and candidate variant analysis was undertaken in all patients.
For P1 DNA was fragmented by sonication performed on a Covaris S2 system (Covaris, Woburn, MA). Sequencing libraries were constructed using the New England BioLabs NEBNext® Ultra™ DNA Library Prep Kit. Exomes were captured by the SeqCap Kit V3 (NimbleGen) according to the manufacturer's protocol. Libraries were sequenced on an Illumina HiSeq 2500 system at the University Hospital in Motol as described in (2) For P3, WES was performed as previously described (3). Briefly, exome capture was attained using the Illumina TruSeq 62 Mb capture kit, sequenced using the Illumina HiSeq 2000 system in 100 bp reads and aligned to the human reference genome (UCSC hg19). Variants were restricted to those with a minor allele frequency of ≤ 0.01% from in-house controls (n = 378) and external variant databases GnomAD (http://gnomad.broadinstitute.org/), National Heart, Lung, and Blood Institute Exome Sequencing Project (http://evs.gs.washington.edu/EVS/) and 1000 Genomes Project (http://www.internationalgenome.org/). On the basis of no confirmed family history, autosomal dominant (heterozygous) and autosomal recessive (homozygous and compound heterozygous) variants were equally considered. First, pathogenic or likely pathogenic variants in known nuclear genes associated with mitochondrial DNA maintenance disorders were excluded including genes that were analysed diagnostically (POLG, SLC25A4, TWNK, RRM2B, POLG2, TK2 and RNASEH1). Next, variants were filtered using Gene Ontology (GO)-Terms to prioritise nuclear genes encoding mitochondrially-targeted proteins and/or DNA repair or replication machinery. GO-Terms employed were the wildcard term 'mitochondr*', DNA repair', 'replication', 'transcription', 'nucleotide', 'purine', 'pyrimidine', 'exonuclease', 'polymerase', 'topoisomerase', 'ligase', 'helicase' and 'nucleoside'. Copy number variants (CNVs) were also examined using the same GO-Terms. In silico tools Polyphen-2 (http://genetics.bwh.harvard.edu/pph2/), Align GVGD (http://agvgd.hci.utah.edu/agvgd_input.php) and SIFT (http://sift.jcvi.org/) were used to assess the pathogenicity of identified candidate variants. The identified POLRMT variant in P3 was confirmed by Sanger sequencing of PCR products amplified using custom forward and reverse primers flanked with universal M13-derived tags, using the BigDye Terminator v3.1 Cycle Sequencing kit on the ABI 3130xl Genetic Analyser, according to manufacturer's guidelines. Total RNA was extracted from P3 and control cultured fibroblasts using the ReliaPrep RNA Cell MiniPrep system (Promega). To inhibit nonsense-mediated decay, cultured fibroblast medium was supplemented with 100μg ml -1 emetine for 10 hours. RNA was reverse transcribed to cDNA using the GoScript Reverse Transcription system (Promega). Custom forward and reverse primers flanked with universal M13-derived tags were used to amplify POLRMT exons 10-13. PCR products were separated by gel electrophoresis and sequenced as described above. Primers available upon request. For completeness, we also performed WGS (30X coverage) using a TruSeq DNA library and the Illumina HiSeq 2000 platform; no additional candidate variants were identified.
For P4, WES performed at GeneDx. Using genomic DNA from the submitted specimen, the Agilent Clinical Research Exome kit was used to target the exonic regions and flanking splice junctions of the genome. These target regions were sequenced simultaneously by massively parallel (NextGen) sequencing on an Illumina HiSeq sequencing system with 100bp or greater paired-end reads. Reads are aligned to human genome build GRCh37/UCSC hg19, and analyzed for sequence variants using a custom-developed analysis tool (Xome Analyzer). Capillary sequencing or another appropriate method is used to confirm all potentially pathogenic variants identified in the individual and relative samples, if submitted. Sequence alterations are reported according to the Human Genome Variation Society (HGVS) nomenclature guidelines. Analysis of exome for the proband includes evaluation of variants that are identified to be de novo, compound heterozygous, homozygous, heterozygous and Xlinked in addition to relevant analysis based on the family structure and reported phenotype.
Trio WES analysis was performed on genomic DNA extracted from white blood cells of P5 and parents as previously described in (4). The Agilent SureSelectXT Human All Exon V5 enrichment kit was used to enrich exonic sequences and 100 bp paired-end sequencing was performed on an Illumina HiSeq sequencer. MERAP pipeline was employed for sequence analysis.
Trio WES analysis was performed in the family of P6 using genomic DNA extracted from blood. Sequencing libraries were made using SureSelect Human All Exon V5 kit (Agilent) and target regions were sequenced on Illumina HiSeq platform. The variant analysis was focused on coding nonsynonymous recessive (homozygous and compound heterozygous) and de novo variants with allele frequencies below 1% in population databases ExAC and gnomAD. This analysis revealed a de novo stop-gain variant c.1525C>T p.Arg509Ter in SLC22A31, which was not prioritized due to having 23 heterozygous carriers in gnomAD and for showing a low probability of being loss-of-function intolerant (pLI score = 0 in gnomAD). Recessive variants search revealed putative variants in OR6A2, KMT2D, TGM7, MMP25, OR7G3, POLRMT and P2RY11. While only KMT2D is a known disease gene, both detected variants were classified as benign in ClinVar, and the established inheritance pattern (dominant with de novo variants) does not fit with known disease association of Kabuki syndrome.
Whole exome sequencing was performed on genomic DNA samples from P7 and P8 using the Illumina platform using previously described methods (5)(6)(7)(8). The exome-sequencing pipeline included sample plating, library preparation using the Illumina Nextera kit (2-plexing of samples per hybridization), hybrid capture using the Illumina Rapid Capture Enrichment kit (37 Mb target), sequencing (150 bp paired reads), sample identification QC check, and data storage. The BWA aligner was used for mapping reads to the human genome build 37 (hg19). Single Nucleotide Polymorphism (SNPs) and insertions/deletions (indels) were jointly called across all samples using Genome Analysis Toolkit (GATK) HaplotypeCaller package version 3.4. Default filters were applied to SNP and indel calls using the GATK Variant Quality Score Recalibration (VQSR) approach. Lastly, the variants were annotated using Variant Effect Predictor (VEP). To select the variants predicted to have high likelihood of pathogenicity, the variant call set was uploaded to the Seqr platform (https://seqr.broadinstitute.org) and filtered using standard online databases as previously described, based on a presumed autosomal recessive inheritance pattern (5)(6)(7)(8). Exons 3 and 13 of POLRMT were amplified by PCR from genomic DNA samples to confirm specific variants by Sanger sequencing and determine segregation patterns of these variants among affected and unaffected family members.

Thermal stability measurements
The fluorescent dye Sypro Orange (Invitrogen; λex = 490 nm, λem = 570 nm) was used to monitor the temperature-induced unfolding of POLRMT variants. Proteins were set up in 96 well PCR plates at a final concentration of 1.6 µM protein and 2.5X dye in assay buffer (25 mM Tris-HCl pH 8.0, 0.2 M NaCl, 0.5 mM EDTA, 10 % (w/v) glycerol and 1 mM DTT). Analyses were performed in a C1000 Thermal Cycler using the CFX96 real time software (BioRad). Scans were recorded using the HEX emission filter (560-580 nm) between 4 to 95 °C in 0.5 °C increments with a 5 s equilibration time. Data were normalised between 0-1 and the melting temperature (Tm) was determined from the first derivative of a plot of normalised fluorescence intensity versus temperature (9).

Mitochondrial network and nucleoid analysis
Control and POLRMT patient fibroblasts (9 000 cm -2 ) were grown in 35 mm µ-Dishes (Ibidi) overnight. For analysis of mitochondrial networks, cells were incubated in 5 nM TMRM (Invitrogen) for 30 minutes. Nucleoids were visualised by incubation in 1.5 µl.ml -1 PicoGreen (Invitrogen) for 30 minutes prior to staining with 10 nM TMRM. Z-stack images (1024 x 1024, 0.063 µm per pixel) comprising 0.1 µm steps were captured on an inverted array-scanning confocal iSIM microscope (VisiTech) with a TiE body (Nikon) using a 100 X magnification oil immersion objective, and two Hamamatsu Orca Flash 4 sCMOS monochrome cameras. For network analysis using imaging of TMRM alone, a 561 nm laser was used at 85% power for 300 ms. For dual imaging of mitochondrial nucleoids and network, a 488 nm laser (7%) and 561 nm laser (50%) were used to excite PicoGreen and TMRM, respectively, for 400 ms. A 525/50 filter was used for the 488 nm channel, and a 590L filter was used for the 561 nm channel. For quantification of mitochondrial networks, triplicate experiments were undertaken with microscope and laser settings consistent throughout and on each experimental day 10 cells were imaged for each cell line (at least 8 fields). Images were then processed using Richardson-Lucy deconvolution (10 iterations) on NIS-Elements (Nikon). Maximum intensity projections of Z-stacks (opened in Fiji using the Autoscale function) were generated using Fiji then analysed using the MiNA 2.0.0 toolset (10). For pre-processing the following settings were used: blocksize (127), histogram bins (256), maximum slope (3), median filter: radius (2), unsharp mask: radius (3) and mask strength (0.85). Tophat was not applied. For each image, mean branch length and mean number of branches per mitochondrial network were recorded, and then multiplied to calculate mean mitochondrial network length per image.