Autosomal recessive pathogenic MSTO1 variants in hereditary optic atrophy

Gerber et al report 2 autosomal recessive pathogenic Misato homolog 1 (MSTO1) variants causing hereditary optic atrophy and raise concerns about a previously identified dominant variant of MSTO1 by Gal et al (2017).

M isato homolog 1 (MSTO1) is a nuclear gene encoding a mitochondrial morphology-regulating protein. Two forms of the MSTO1 gene-related disease have been reported: the recessive form was reported in more than 20 families segregating biallelic variants including large genomic deletions, changes affecting the initiation codon, nonsense variants, splice site variants, frameshift variants and missense variants (Human Gene Mutation database; https://www.hgmd.cf.ac.uk/ac/index.php), while the dominant form was reported in one family of four affected individuals (Val8Met, c.22G > A; Gal et al, 2017). The recessive form is characterized by an earlyonset proximal muscle weakness, elevated plasma creatine kinase (CK) levels, and additional symptoms, including nonprogressive cerebellar atrophy/ataxia, visual impairment, motor developmental delay, cognitive alteration, corticospinal tract involvement, and skeletal abnormalities (Donkervoort et al, 2019;Nasca et al, 2021), while the affected individuals from the dominant family suffered from cognitive and/or psychiatric disorders with distal muscle weakness and normal plasma CK levels.
Here, we report 2 biallelic MSTO1 variants in a family presenting with early-onset muscle weakness and adult-onset optic neuropathy. Since the present study added MSTO1 to the list of hereditary optic neuropathy (HON)-causing genes (Gerber et al, 2021;Zeviani & Carelli, 2021), we further sequenced the MSTO1 gene in a cohort of 49 patients with undiagnosed hereditary optic neuropathy.
The proband (III-2) is a 63-year-old individual born to healthy unrelated parents of Jewish Ashkenazi origin ( Fig 1A). He displayed childhood-onset motor impairment, which remained relatively stable over time, and he subsequently developed progressive visual impairment by the age of 20 years. At 30 years of age, he presented with mild proximal limb weakness, and increased stretch reflex and positive Hoffman sign (Table 1). Neuro-ophthalmic evaluation revealed a severe bilateral optic atrophy with no retinal involvement (Table 1). Further investigations showed moderately elevated CK levels (365 UI/l), myogenic features, and complex repetitive discharges (CRD) on electroneuromyography (ENMG), diffuse fat infiltration of the pelvic girdle and a peculiar pattern of relative sparing of several muscles on muscle MRI. Cardiac examination found no cardiomyopathy or heart conduction abnormalities (Fig 1B-D). Brain MRI revealed mild cerebellar atrophy. Next generation sequencing failed to detect pathogenic variants in the OPA1 gene, the Leber hereditary optic neuropathy-associated genes, nor in 384 nuclearly encoded mitochondrial genes. Two sibs of the proband (III-1 and III-3) also developed severe, adult-onset optic neuropathy and complained of childhood-onset muscle weakness. Whole exome sequencing of the trio, analyzed according to both dominant and recessive models, identified 2 allelic variants in the MSTO1 gene: a missense variant NM_018116: c.65C > A (p.Ala22Glu) absent from public databases, and an ultra-rare splicing variant c.220 + 5G > C (rs1187504822; minor allele frequency, MAF: 0.0032%) predicted to disrupt the splice donor site of intron 2 ( Fig 1E). Both variants were predicted as "pathogenic" according to the American College of Medical Genetic (ACMG) classification (Richard et al, 2015). Biallelism was confirmed by Sanger sequencingbased segregation analysis (Fig 1A), and cDNA analysis from patient's cultured fibroblasts showed apparent homozygosity of the c.65C > A variant (Fig 1E and F), suggesting instability or degradation of the mRNA transcribed from the allele carrying the c.220 + 5G > C substitution.
While retinitis pigmentosa has previously been reported in MSTO1 gene-related disease, this study first reports on severe, bilateral optic neuropathy in the disease and suggests that MSTO1 should be added to the list of hereditary optic neuropathy (HON)causing genes (Gerber et al, 2021;Zeviani & Carelli, 2021) especially when associated with muscular involvement mimicking congenital muscle dystrophy.
Interestingly, we retrospectively screened MSTO1 in 49 patients with unexplained HON and we detected the previously reported dominant variant (c.22G > A, p.Val8Met ;Gal et al, 2017) in 16/49 patients (32.6%), questioning the relevance of this variant. We thus performed further analyses to investigate this high frequency; pairwise alignment of the MSTO1 RNA with the non-coding RNA transcribed from the MSTO1 pseudogene (MSTO2P, located 133 kb downstream) revealed a perfect match of MSTO1 Ex1-2 Fw and Rev primers (Gal et al, 2017) and a unique mismatch in the 563 bp amplified region (Fig 1G), showing a lack of specificity of these primers. In this region, the MSTO1 nucleotide involved in the reportedly dominant c.22G > A substitution aligns to the MSTO2P r.83 position ( Fig 1G) where a frequent G > A polymorphism is described (NR_024117.2: n.83G > A, rs11264409, minor allele frequency 32.6%). We amplified MSTO2P and MSTO1 individually using specific primers ( Fig 1G) and cloned the products in the pCR TM II-TOPO R Vector. Sanger sequencing of the constructs using a vector-specific primer showed the presence of the MSTO2P n.83G > A variant in the 16 patients of the HON cohort and the proband (III-2), and confirmed the location of c.65C > A and c.220 + 5G > C substitutions on each of the two MSTO1 alleles in the proband (III-2), respectively ( Fig 1H). Knowing the elevated frequency of the n.83G > A polymorphism, it is intriguing that in the dominant family, the authors reported dramatic silencing or decay of MSTO1 mRNA while there was a notable detection of a G > A change at the RNA level (Gal et al, 2017). Whether these observations describe the presence of the G > A substitution on MSTO2P instead of MSTO1 needs consideration before the dominant MSTO1 c.22G > A substitution referred to in the CLINVAR catalog (RCV000505814.2) is definitely validated.
To conclude, MSTO1 should be considered in patients with optic atrophy and muscular involvement, especially if mimicking congenital muscle dystrophy, and geneticists/clinicians should consider the presence of a highly homologous pseudogene in MSTO1 diagnosis, questioning the reality of dominant forms of the disease.

Methods
Family 1 (Fig 1A) was selected for the study due to negatively targeted exome sequencing-based molecular diagnosis of mitochondrial DNA and a panel of 384 genes, including genes known to be associated with mitochondrial disorders and genes for complex I subunits and assembly factors.
The Comit e de Protection des Personnes Ilede-France II Institutional Review Board approved all aspects of the study herein described (CPP:2015-03-03/DC2014-2272). Informed consent was obtained from all subjects involved in the study, and experiments conformed to principles set out in the WMA Declaration of Helsinki and Human Services Belmont Report. The proband gave his permission to publish the present work.

Whole exome sequencing analysis
Genomic DNA was extracted from whole blood using standard protocols. The DNA (1 lg) of the proband and his affected brother and sister were sequenced using the HiSeq 2000 (Illumina) at the Genomic ◀ Figure 1. Family, muscular findings, MSTO1 and MSTO2P sequences alignment and variants analysis.
(A) Pedigree of the family and segregation analysis of MSTO1 c.65C > A and c.220 + 5G > C variations. The proband (III-2) and his 2 siblings (III-1 and III-3) developed early-onset myopathy and severe optic neuropathy. Segregation analysis of MSTO1 variants using the MSTO1 Fw/MSTO1 Rev primer pair (specific to MSTO1; see panel G) shows heterozygosity for the c.65C > A substitution in unaffected subject IV-4, demonstrating biallelism of the c.65C > A and c.220 + 5G > C variations in the affected trio III-1, -2, and -3. (B-D) Muscle MRI. Axial T1-weighted imaging on pelvis (B), upper thigh (C), and lower thigh (D) showing major fat infiltration of the thigh muscles. AB, adductor brevis muscle; AL, adductor longus muscle; AM, adductor magnus muscle; BF, biceps femoris muscle; GM, gluteus maximus muscle; GR, gracilis muscle; IL, ilio-psoas muscle; PE, pectineus muscle; RF, rectus femoris muscle; SA, sartorius muscle; SM, semimembranosus muscle; ST, semitendinosus muscle; TFL, tensor fasciae latae; VL, vastus lateralis muscle; VM, vastus medialis muscle; VI, vastus intermedius muscle. (E) 5 0 Donor Splice Site Prediction (SSP) by varSEAK SSP tool. The varSEAK Online splice site prediction tool was used to analyze the c.220 + 5G > C variant. The table shows a strong decrease in the score for authentic Splice Site (Pos 220 + 1, highlighted in purple in the table as well as on the graph on the right). The score corresponds to the likelihood for the variant to be functional (positive values) or not functional (negative values). The score column corresponds to the score of the splice site on the reference sequence (upper value) and on the variant sequence (lower value). The DScore corresponds to the difference between the two scores. There was a strong decrease in the score and DScore for the 5 0 authentic Splice Site, thus predicting the loss of function and exon skipping of the variant. The overall prediction is given as a splice site prediction class ranging from 1 (No splicing effect) to 5 (Splicing effect). The c.220 + 5G > C variant was classified as affecting splicing (class 5). No cryptic splice sites were predicted to be activated. The MaxEntScan scoring algorithm also found a decrease in the score and negative values for the DMaxEntScan, thus predicting a deleterious effect. The sequence graph (on the right) depicts the reference and variant sequence and the 5 0 authentic Splice Site. The variant is highlighted in red, the HGVS nomenclature is given above.  Core Facility of Imagine (Paris). Libraries were prepared using the SureSelect Human All Exon Kit v3 (Agilent Technologies) according to the manufacturer's protocols and run in pair-ended mode (2 × 75). Image analysis and base calling were performed with Real Time Analysis (RTA) Pipeline version 1.9 with default parameters (Illumina). Sequences were aligned to the human genome reference sequence (UCSC Genome Browser hg19 assembly), and single nucleotide polymorphisms were called on the basis of allele calls and read depth with the use of the Consensus Assessment of Sequence and Variation pipeline (v.1.8, Illumina). Genetic variation annotation was performed by an in-house pipeline. Only the variants whose positions were covered ≥ 10× were further considered. We selected compound heterozygous variants with a minor allele frequency < 1% (recessive model) and heterozygous variants with a minor allele frequency < 0.1% (dominant model) shared by the three affected members. The pathogenicity of the selected variants was assessed using the Alamut Mutation Interpretation Software (http://www. interactive-biosoftware.com), a decision support system for mutation interpretation based on Align DGVD, MutationTaster, PolyPhen-2, SIFT, SpliceSiteFinder-like, MaxEntScan, NNSPLICE, GeneSplicer, Human Splicing Finder, ESEfinder, and RESCUE-ESE. The effect of the c.220 + 5G > C substitution on splicing was further evaluated using the varSEAK online splice site prediction tool (https://varseak.bio/) and MaxEntScan, respectively.

Hot
Start Polymerase (0.015 units; Promega). A touchdown PCR program was used to increase the specificity of annealing, which consisted of initial denaturation at 98°C for 3 min; 8 cycles of denaturation at 98°C for 20 s, annealing from 66°C to 58°C for 15 s (À1°C each cycle), extension at 72°C for 30 s; 27 cycles of denaturation at 98°C for 20 s, annealing at 58°C for 15 s and extension at 72°C for 30 s, and a final extension at 72°C for 1 min.
Cloning and Sanger sequencing of MSTO1 or MSTO2P-containing colonies MSTO1 or MSTO2P amplifications (2 ll) were verified on a 2% agarose gel and cloned (2 ll) in the pCR TM II-TOPO R Vector using the TOPO TM TA Cloning TM Kit, dual promoter, with pCR TM II-TOPO TM Vector and One Shot TM Mach1 TM T1 phage-resistant chemically competent Escherichia coli according to the manufacturer's protocol Invitrogen, Life Technologies, St Aubin, France.
MSTO1 and MSTO2P sequences were amplified by PCR (10 ll) directly from bacteria colonies in 1X Green Go Taq Flexi Buffer (Promega) containing the vectorspecific M13F and M13R primers (0.2 lM each), MgCl2 (1.5 mM, Promega), dNTP (0.04 mM, Promega), and Go Taq G2 Hot Start Polymerase (0.015 units; Promega). The PCR program consisted of an initial denaturation at 98°C for 3 min, 30 cycles of denaturation at 98°C for 20 s, annealing at 58°C for 20 s, extension at 72°C for 30 s, and a final extension at 72°C for 1 min.
Amplifications (2 ll) were verified on a 2% agarose gel and purified (8 ll) from single-strand DNA by adding exonuclease I (4 units; Thermo Fisher) and FastAP thermosensitive alkaline phosphatase (0.4 unit; Thermo Fisher) for 10 min at 37°C. The reactions were arrested by heating the samples at 85°C for 10 min.
Sequencing reactions (10 ll) were performed on purified PCR products (2.5 ll) using the vector-specific M13F primer and the Applied Biosystems â BigDye Terminator v3.1 Cycle Sequencing Kit on an Applied Biosystems â 2720 thermal cycler with the following cycling conditions: an initial denaturation at 98°C for 3 min, 35 cycles of denaturation at 98°C for 20 s, annealing at 58°C for 20 s, extension at 60°C for 40, and a final extension at 60°C for 5 min.
Sequencing products in nuclease-free water (10 ll/10 ll) were purified from oligonucleotides and unincorporated ddNTPs and dNTPs by spin column size exclusion chromatography using Sephadex G-50 (GE Healthcare) on 96-well multiscreen 0.45 lm Durapore â Membrane plates (Merck Millipore; Guyancourt, France) for 4 min at 18 g. Purified products were analyzed on an Applied Biosystems â 3500XL genetic analyzer.

RNA preparation, cDNA synthesis, and RT-PCR analysis of MSTO1
Total RNA was prepared from fibroblasts using the RNeasy Mini Kit (QIAGEN, Les Ulysses, France) in accordance with the manufacturer's protocol. The samples were DNase-treated with the RNase-free DNase Set (QIAGEN). The concentration was assessed using the NanoDrop spectrophotometer (ThermoFisher) before storage at À80°C. First-strand cDNA synthesis (20 ll) was performed from total RNA (500 ng) using the Verso cDNA Kit (ThermoFisher Scientific) with random hexamer:anchored oligo (dT) primers in a 3:1 (vol:vol) ratio, in accordance with the manufacturer's instructions. A non-RT reaction (without an enzyme) for one sample was prepared as a control. MSTO1cDNA (2 ll) was amplified by PCR (10 ll) using the MSTO1 Fw and MSTO1 Rev primers specific to MSTO1 (Fig 1G) and Sanger sequenced using the MSTO1 Fw or MSTO1 Rev primers individually, following the protocol described in "Cloning and Sanger sequencing of MSTO1 and MSTO2P clones."

Data availability
Whole exome sequencing .bam files are deposited in the access-controlled Interactive Polyweb interface (https://polyweb.fr/) of the Institute of Genetic Diseases, Imagine -University Paris Descartes. They can be accessed upon a duly substantiated request. The MSTO1 c.65C > A and c.220 + 5G > C variants have been deposited in the CLINVAR database with the accession references SCV002526430 and SCV002526431, respectively.
Expanded View for this article is available online.