Ryanodine Receptor Type 3 (RYR3) As a Novel Gene Associated with a Myopathy with Nemaline Bodies

ORIGINAL ARTICLE

Ryanodine receptor type 3 (RYR3) as a novel gene associated with a myopathy with nemaline bodies

Y. Nilipoura, S. Naﬁssib, A. E. Tjustc,d, G. Ravenscrofte, H. Hossein Nejad Nedaif, R. L. Taylore, V. Varastehf, F. Pedrosa Domellof€ c,d, M. Zangig, S. H. Tonekabonia, M. Oliveh, K. Kiiskii, L. Sagathi , M. R. Davisj, N. G. Lainge and H. Tajsharghie,k aPediatric Pathology Research Centre, Moﬁd Children’s Hospital, Shahid Beheshti University of Medical Sciences, Tehran; bDepartment of Neurology, Tehran University of Medical Sciences, Tehran, Iran; cDepartment of Integrative Medical Biology, Umea University, Umea; dDepartment of Clinical Sciences, Umea University, Umea, Sweden; eCentre for Medical Research, University of Western Australia and Harry Perkins Institute for Medical Research, Nedlands, WA, Australia; fDepartment of Pathology, Shahid Beheshti University of Medical Sciences, Tehran; gTracheal Diseases Research Center (TDRC), National Research Institute of Tuberculosis and Lung Diseases (NRITLD), Shahid Beheshti University of Medical Sciences, Tehran, Iran; hDepartment of Pathology and Neuromuscular Unit, IDIBELL-Hospital de Bellvitge, Barcelona, Spain; iDepartment of Medical and Clinical Genetics, Folkhalsan€ Institute of Genetics, Medicum, University of Helsinki, Helsinki, Finland; jDepartment of Diagnostic Genomics, Pathwest, QEII Medical Centre, Nedlands, WA, Australia; and kDivision of Biomedicine, School of Health and Education, University of Sko¨vde, Sko¨vde, Sweden

Keywords: Background and purpose: Nemaline myopathy (NEM) has been associated intracellular Ca2+ with mutations in 12 genes to date. However, for some patients diagnosed channels, nemaline with NEM, definitive mutations are not identified in the known genes, suggest- myopathy, ryanodine ing that there are other genes involved. This study describes compound receptors, RYR3 heterozygosity for rare variants in ryanodine receptor type 3 (RYR3) gene in one such patient. Received 17 August 2017 Methods and results: Clinical examination of the patient at 22 years of age Accepted 26 February 2018 revealed a long narrow face, high arched palate and bilateral facial weakness. She had proximal weakness in all four limbs, mild scapular winging but no European Journal of scoliosis. Muscle biopsy revealed wide variation in fibre size with type 1 fibre Neurology 2018, 0: 1–7 predominance and atrophy. Abundant nemaline bodies were located in perinu- doi:10.1111/ene.13607 clear and subsarcolemmal areas, and within the cytoplasm. No likely pathogenic mutations in known NEM genes were identified. Copy number variation in known NEM genes was excluded by NEM-targeted comparative genomic hybridization array. Next-generation sequencing revealed compound heterozygous missense variants in the RYR3 gene. RYR3 transcripts are expressed in human fetal and adult skeletal muscle as well as in human brain and cauda equina samples. Immunofluorescence of human skeletal muscle revealed a ‘single-row’ appearance of RYR3, interspersed between the ‘double rows’ of ryanodine receptor type 1 (RYR1) at each A–I junction.

Conclusion: The results suggest that variants in RYR3 may cause a recessive EUROPEAN JOURNAL OF NEUROLOGY muscle disease with pathological features including nemaline bodies. We charac- terize the expression pattern of RYR3 in human skeletal muscle and brain, and the subcellular localization of RYR1 and RYR3 in human skeletal muscle.

Introduction Congenital myopathies (CMs) are a heterogeneous group of muscle diseases characterized by the presence Correspondence: H. Tajsharghi, Division of Biomedicine, School of of speciﬁc morphological features on skeletal muscle Health and Education, University of Sko¨vde, Sko¨vde, Sweden (tel.: +46 500 448612; fax: +46 500 416325; biopsy, including central cores, central nuclei and e-mail: [email protected]). nemaline bodies [1]. CMs demonstrate overlapping

(a) AB

C D

(b)

(c)

(d)

Figure 1 Analysis of the histopathology of the patient’s muscle and genetics. Expression analysis of ryanodine receptor type 1 (RYR1) and ryanodine receptor type 3 (RYR3) at transcript level in controls. (a) Muscle morphology of the patient. (A) Haematoxylin and eosin stain of the muscle biopsy showing marked fibre size variability and some internal nuclei, some of which are enlarged. (B) On modified trichrome staining, several fibres showed collections of nemaline bodies. (C) Succinic dehydrogenase reaction revealed uneven distribution of oxidative enzyme activity with some core-like areas. (D) ATPase at pH 9.4 revealed prominent type I fibre predominance and atrophy. (b) Pedigree and recessive inheritance of RYR3. Sequence analysis confirmed the segregation of the variants. The affected individual is represented by the black symbol. (c) Amino acid sequence alignment shows that the affected amino acids (red) have been highly conserved during evolution. The affected amino acid residues in RYR3 are conserved when compared with its ortho- logues RYR1 and ryanodine receptor type 2 (RYR2) (red). The neighbouring amino acid residues in paralogous RYR1 and RYR2 (green) are associated with central core disease/malignant hyperthermia and cardiac diseases, respectively. (d) RYR1 and RYR3 are coexpressed in primary human brain and skeletal muscle (SkM) tissues. Transcript abundance of RYR1 and RYR3 in human fetal SkM (n = 2) (□) and adult SkM (n = 2) ( ), adult cortex (n = 3) ( ) and adult cauda equina (CE) (n = 3) (■) was assessed by quantitative polymerase chain reaction. Expression of RYR1 and RYR3 was normalized to the geometric mean of three endogenous control genes (ACTB, EEF2 and TBP) using the delta Ct method. Graphed data represent the mean + SEM of two independent replicates for SkM samples and three independent replicates for brain and CE samples. [Colour figure can be viewed at wileyonlinelibrary.com].

clinical and genetic heterogeneity with multiple genes There are two other ryanodine receptor genes, i.e. associated with one pathology and one gene associ- ryanodine receptor type 2 (RYR2) and ryanodine recep- ated with multiple pathologies [1]. Examples of these tor type 3 (RYR3). RYR2 is expressed predominantly in are nemaline myopathy (NEM) and the ryanodine the heart and mutation of RYR2 has been associated receptor type 1 (RYR1) gene. with cardiomyopathies [10]. In mice, RYR3 is tran- Nemaline myopathy is a congenital myopathy with a siently expressed during the post-natal phase of muscle large spectrum of clinical severity, ranging from fetal development, and has a wide pattern of expression in akinesia to almost asymptomatic adult patients with adult tissues especially in brain, but no detectable mild muscle weakness [2]. The diagnostic hallmark of expression in adult muscles, with the exception of the NEM is the presence of numerous nemaline bodies that diaphragm and soleus muscles [11]. The role of RYR3 may form clusters at the periphery of the myofibres or in striated myofibres remains elusive and, to date, no be disseminated throughout the myofibres [1]. Currently, disease has been associated with RYR3 mutations. mutations in 12 genes are known to cause NEM, includ- Here, we report a patient of Iranian–Caucasian des- ing ACTA1, CFL2, KBTBD13, KLHL40, KLHL41, cent diagnosed with NEM compound heterozygous LMOD3, NEB, TNNT1, TPM2, TPM3 [3,4] and, more for rare variants in RYR3. In addition, we character- recently, MYO18B [5] and MYPN [6]. Nevertheless, for ize the expression pattern of RYR3 in human skeletal some patients with NEM, mutations are not found in muscle and brain and its subcellular localization in any of the known genes, suggesting that other causative human skeletal muscle. genes remain to be found. Mutations in RYR1 (OMIM 180901), encoding the Patient skeletal muscle ryanodine receptor involved in excita- tion–contraction coupling, are one of the commonest The patient was a 22-year-old female, with healthy causes of congenital myopathy [1]. The CMs associ- parents and one healthy 31-year-old sister. The preg- ated with RYR1 display a spectrum of pathological nancy and vaginal delivery were uncomplicated and features spanning central core disease (OMIM 117000), no muscular weakness or hypotonia were noted in the congenital fibre-type disproportion (OMIM 255310), neonatal period. Her developmental milestones were multiminicore/minicore/multicore disease (OMIM achieved on time. The first signs of muscle weakness 117000, 602771 and 255320), NEM (OMIM 117000), were observed at the age of 5 years, when she ran and CMs with prominent nuclear internalization and large climbed stairs with difficulty and rose up from the areas of myofibrillar disorganization (OMIM 160150) floor with a positive Gower’s manoeuvre. She also [7], and overlapping disorders such as core–rod disease had some weakness in her arms. She did not complain [8]. Congenital myopathy with onset in utero, including of any cranial or bulbar symptoms. lethal multiple pterygium syndrome (OMIM 253290), Clinical examination at 22 years of age revealed a and fetal akinesia deformation sequence (OMIM long narrow face, high arched palate, bilateral facial 208150) are considered as the most severe end of the weakness, but no ptosis or ophthalmoplegia, slender RYR1 clinical spectrum [9], whereas anaesthesia-related tongue and micrognathia. She had proximal weakness malignant hyperthermia may be considered the mildest in four limbs, mild scapular winging but no scoliosis. clinical manifestation [1]. She rose up from the floor with a Gower’s manoeuvre.

Muscle strength was 4/5 on shoulder and arm basophilic degenerative fibres. Abundant nemaline muscles and 5/5 in distal upper limbs. Hip flexion, knee bodies were located in perinuclear and subsarcolem- flexion and extension were 4–5/5 and hip adductor, mal areas as well as in the cytoplasm. Reduced abductor and foot dorsiflexors were 4/5. Deep tendon nicotinamide adenine dinucleotide dehydrogenase- reflexes were generally decreased and sensory examina- tetrazolium reductase and succinic dehydrogenase tion was intact. Nerve conduction studies were normal. staining revealed uneven distribution of oxidative Low-frequency repetitive nerve stimulation from abduc- activity and some core-like areas in those regions tor pollicis brevis and nasalis muscles did not show any occupied by collections of nemaline bodies (Fig. 1a). decrement. Electromyography showed myopathic motor unit potentials without spontaneous activity or Genetic findings abnormal discharge. Laboratory investigations, including serum creatine kinase and serum lactate levels, were No rare likely pathogenic mutations were identified in normal. Echocardiography and electrocardiography the 12 NEM-associated genes included in the targeted were normal. Forced vital capacity was 75%. panel of the 336 neuromuscular disease genes. The targeted comparative genomic hybridization array did not reveal any copy number variations in the NEM genes. Materials and methods We then performed WES on the patient’s DNA. Following a filtering strategy, we were left, as likely Histochemical analyses of muscle biopsy candidate variants, with only compound heterozygous An open muscle biopsy specimen was obtained from missense variants in exon 40 (c.6208A>G) (p.Met2070- the biceps brachii muscle at the age of 18 years. Val) and exon 63 (c.8939G>T) (p.Arg2980Leu) of RYR3 Unfortunately, no muscle tissue was available for in the patient (Fig. 1b). Both variants are rare heterozy- immunoblot and ultrastructural analyses. gous variants in gnomAD (ENSG00000198838) (p.Met2070Val, rs769938343, frequency 0.0001381; p.Arg2980Leu, rs200346049, frequency 0.00006873). Genetic analysis Segregation analysis showed that the compound Next-generation sequencing of a targeted neuromuscu- heterozygous RYR3 c.6208A>G (p.Met2070Val) and lar panel consisting of 336 known neuromuscular dis- c.8939G>T (p.Arg29- ease genes, including the 12 genes associated with 80Leu) variants detected in the proband had been inher- NEM, was performed [9]. ited from the unaffected father and mother, respectively. A custom comparative genomic hybridization array Neither variant was found in the unaffected sister analysis, targeting known NEM genes, was used to (Fig. 1b). The two amino-acid residues affected are con- exclude the presence of large copy number variations. served across species and across all three ryanodine Whole-exome sequencing (WES) was carried out, as receptor isoforms in humans (Fig. 1c). previously described [12]. Data from WES were analysed by using the Ingenuity Variant Analysis software Analyses of RYR1 and RYR3 in control human skeletal (Qiagen, Hilden, Germany). muscle and brain tissues The RYR1 and RYR3 transcripts were expressed at Analyses of human skeletal muscle and brain tissues detectable levels in all tissue samples examined from controls (Fig. 1d). Sections of human skeletal muscle and brain from con- Immunofluorescence analysis indicated clear expres- trols were processed for quantitative polymerase chain sion of both RYR1 and RYR3 isoforms in all investi- reaction, immunofluorescence and confocal microscopy gated muscle specimens of the adult control subjects assessments. Detailed methods are provided in Data S1. (Fig. 2a). RYR1 staining demonstrated variable intensity between myofibres, particularly in temporalis and masseter muscles. RYR3 staining intensity was, how- Results ever, more similar between different myofibres and did not display the chequered labelling pattern normally Morphological analysis seen in proteins that are preferably expressed in different Muscle biopsy from the patient revealed wide varia- fibre types (Fig. 2a.B). Confocal microscopy revealed a tion in fibre size with type 1 fibre predominance and striated pattern appearance of RYR3 (Fig. 2b.A1 and atrophy and hypertrophic type 2 fibres. There were A3). Although RYR1 delineated ‘double rows’ [13] at increased numbers of internal nuclei and rare each A–I junction (Fig. 2b.a2 and a3), RYR3 was

(a) (b)

Figure 2 Confocal microscopy of muscle specimens of the adult control subjects. (a) Confocal microscopic images depicting ryanodine receptor type 3 (RYR3) (A1, B1, C1 and D1), ryanodine receptor type 1 (RYR1) (A2, B2, C2 and D2) or overlaid RYR3 and RYR1 staining. (A1, A2 and A3) Longitudinal muscle fibres from the depressor labii inferioris muscle with RYR3 and RYR1 staining muscle fibres in similar, but not identical, myofibrillar structures. (B1, B2 and B3) Cross-sections of small, circular muscle fibres of the temporalis and masseter muscles (white arrowheads). Compared with other muscles, these muscle fibres stained very strongly for RYR1, with normal levels of RYR3. (C1, C2 and C3) Cross-sectioned nerve fascicle (nf) with RYR3 staining present in the axons (open arrowheads) and RYR1 staining being completely absent. (D1, D2 and D3) An artery (a) and a vein (v) where the smooth muscle cells of lamina media (arrows) were strongly stained for RYR3, but not RYR1. Bar = 50 lm. (b) Longitudinal sections from human depressor labii inferioris (A and a), psoas major (B) and buccinator (C) muscles, analysed by confocal microscopy with a 9100 objective. The upper panels (A1, a1, B1 and C1) show RYR3 staining, rendered in black and white. The middle panels (A2, a2, B2 and C2) show RYR1 staining alone. The lower panels (A3, a3, B3 and C3) show overlaid staining of RYR3 (green) and RYR1 (red). The inset pic- ture (a) is a digital magnification of the striation in A. In all muscles, RYR1 staining exhibited a ‘double-row’ appearance at the A–I junction (white arrowhead), whereas the RYR3 staining localized more centrally in the stained bands, between the rows of RYR1 staining. Bar = 10 lm. [Colour figure can be viewed at wileyonlinelibrary.com].

© 2018 EAN 6 Y. NILIPOUR ET AL. present as a single line interspersed between the double genes with muscle diseases further defines the expres- rows of the RYR1 staining pattern (Fig. 2b.a1 and a3). sion pattern of RYR1 and RYR3 in human tissues as well as its subcellular localization in skeletal muscle, suggesting that RYR3 may play a role in skeletal Discussion muscle that is distinct from that of RYR1. Targeted neuromuscular panel sequencing or WES in Ryanodine receptor type 1 is known to bridge the our patient did not identify pathogenic variants in all membranes of the terminal cisterns, a voluminous part 12 genes known to be mutated in NEM, and the com- of the sarcoplasmic reticulum, and the T-tubule at the parative genomic hybridization array confirmed the I–H junctions. Our double immunostaining of longitu- absence of large copy number variations in NEM genes, dinal muscle sections demonstrated the typical ‘double- including NEB. Both healthy parents were found to be row’ staining pattern of RYR1 [13], whereas RYR3 heterozygous carriers of one of the variants and nei- revealed a ‘single-row’ appearance, interspersed ther variant was found in the unaffected sister. The between the double rows of RYR1 (Fig. 2). Further likely pathogenicity of the variants was supported by studies will be needed to discern whether this finding the results of the prediction tools PolyPhen-2, scale- reflects differences in epitope binding of antibodies or invariant feature transform and combined annotation an actual difference in membranous localization. How- dependent depletion. The mutated RYR3 residues ever, compartmentalized localizations of RYR isoforms (Met2070Val and Arg2980Leu) are conserved in orthol- in smooth muscle cells have previously been clearly ogous as well as paralogous RYR isoforms. The RYR3 demonstrated and inferred to correlate closely with a Met2070 residue is paralogous to Met2208 in RYR1 need for the cell to manage Ca2+ flow both spatially and Met2172 in RYR2. Similarly, the RYR3 residue and temporally during different conditions. In this con- Arg2980 is paralogous to Arg3119 in RYR1 and text, RYR3 is thought to function primarily not in the Arg3084 in RYR2. Variants of residues that are close initial, voltage-dependent Ca2+ release, but rather at to the paralogous residues in RYR1 and RYR2 cause later stages as a maintainer of Ca2+ release during pro- central/mini core disease [14], malignant hyperthermia longed periods of Ca2+ influx [22]. This role of RYR3 [15,16] and arrhythmogenic heart disease or sudden car- might be more evident in slow-twitch fibres, where it diac death, respectively [10]. Interestingly, RYR1 vari- can be observed that high cytoplasmic Ca2+ is main- ant at position 3119 (Arg3119His) is associated with tained for markedly longer periods in slow-twitch com- malignant hyperthermia [17]. pared with fast-twitch fibres during continuous activity. In addition, an integrated data-mining strategy However, further studies will be needed to elucidate the ranking genes with known or potential importance for physiological role of RYR3 in myofibres. skeletal muscle has previously highlighted RYR3 as a candidate gene for CMs, myotonic syndromes and ion Acknowledgements channel muscle diseases [18]. Some of the other candidates from Neto et al. [18] are now known disease The study was supported by grants from the Swedish genes, i.e. MYO18B and CASQ1. Research Council (H.T., F.P.D.), the Swedish Society Taken together, our findings suggest RYR3 as a cau- of Medicine (H.T.) and European Union’s Seventh sative gene of a myopathy with nemaline bodies. How- Framework Programme for research, technological ever, as this study is based on a single family, development and demonstration under grant agree- € additional reports are required to confirm the proposed ment no. 608473 (H.T.). The Vasterbotten’s County link between RYR3 mutations and NEM. We have Council (F.P.D). M.O. is supported by a grant from reported the result of this family at the Annual Con- Fondo de Investigaciones Sanitarias (FIS), Instituto gress of the World Muscle Society in 2016 [19] and have de Salud Carlos III (ISCIII PI14/00738) and FEDER attempted to obtain further families through match- funds ‘a way to achieve Europe’. Australian National maker data sites, but no other families have come to Health and Medical Research Council Fellowships our attention. Hence, it should be borne in mind that APP1002147, APP1035955, European Union Colla- there have been a total of only four descriptions of con- borative Grant APP1055295 and Project Grant genital myopathy associated with CFL2 [4,20]. APP1080587 (N.L.) Ryanodine receptor type 3 may represent an additional example of an increasing number of genes Supporting Information encoding components of Ca2+ release-activated chan- Additional Supporting Information may be found in nels, recently found to be involved in human myopa- the online version of this article: thies, and also dihydropyridine receptor causing myopathy [21]. Association of mutations in these Data S1. Supplementary methods and results.

12. Kariminejad A, Dahl-Halvarsson M, Ravenscroft G, References et al. TOR1A variants cause a severe arthrogryposis 1. Romero NB, Clarke NF. Congenital myopathies. Handb with developmental delay, strabismus and tremor. Brain Clin Neurol 2013; 113: 1321–1336. 2017; 140: 2851–2859. 2. Ryan MM, Schnell C, Strickland CD, et al. Nemaline 13. Flucher BE, Conti A, Takeshima H, Sorrentino V. Type myopathy: a clinical study of 143 cases. Ann Neurol 3 and type 1 ryanodine receptors are localized in triads 2001; 50: 312–320. of the same mammalian skeletal muscle fibers. J Cell 3. Ravenscroft G, Miyatake S, Lehtokari VL, et al. Muta- Biol 1999; 146: 621–630. tions in KLHL40 are a frequent cause of severe autoso- 14. Duarte ST, Oliveira J, Santos R, et al. Dominant and mal-recessive nemaline myopathy. Am J Hum Genet recessive RYR1 mutations in adults with core lesions 2013; 93: 6–18. and mild muscle symptoms. Muscle Nerve 2011; 44: 4. Ong RW, AlSaman A, Selcen D, et al. Novel cofilin-2 102–108. (CFL2) four base pair deletion causing nemaline myopa- 15. Sambuughin N, Holley H, Muldoon S, et al. Screening thy. J Neurol Neurosurg Psychiatry 2014; 85: 1058–1060. of the entire ryanodine receptor type 1 coding region for 5. Malfatti E, Bohm J, Lacene E, et al. A premature stop sequence variants associated with malignant hyperther- codon in MYO18B is associated with severe nemaline mia susceptibility in the North American population. myopathy with cardiomyopathy. J Neuromuscul Dis Anesthesiology 2005; 102: 515–521. 2015; 2: 219–227. 16. Manning BM, Quane KA, Ording H, et al. 6. Lornage X, Malfatti E, Cheraud C, et al. Recessive Identification of novel mutations in the ryanodine-recep- MYPN mutations cause cap myopathy with occasional tor gene (RYR1) in malignant hyperthermia: genotype- nemaline rods. Ann Neurol 2017; 81: 467–473. phenotype correlation. Am J Hum Genet 1998; 62: 7. Monnier N, Romero NB, Lerale J, et al. An autosomal 599–609. dominant congenital myopathy with cores and rods is 17. Robinson R, Carpenter D, Shaw MA, Halsall J, Hop- associated with a neomutation in the RYR1 gene encod- kins P. Mutations in RYR1 in malignant hyperthermia ing the skeletal muscle ryanodine receptor. Hum Mol and central core disease. Hum Mutat 2006; 27: 977–989. Genet 2000; 9: 2599–2608. 18. Abath Neto O, Tassy O, Biancalana V, et al. Integrative 8. Scacheri PC, Hoffman EP, Fratkin JD, et al. A novel data mining highlights candidate genes for monogenic ryanodine receptor gene mutation causing both cores myopathies. PLoS ONE 2014; 9: e110888. and rods in congenital myopathy. Neurology 2000; 55: 19. Nilipour Y, Nafissi S, Varasteh V, et al. Ryanodine 1689–1696. receptor type 3 (RYR3) as a novel gene associated with 9. Todd EJ, Yau KS, Ong R, et al. Next generation nemaline myopathy and fibre type disproportion. Gran- sequencing in a large cohort of patients presenting with ada: World Muscle Society, 2016. neuromuscular disease before or at birth. Orphanet J 20. Fattori F, Fiorillo C, Rodolico C, et al. Congenital Rare Dis 2015; 10: 148. myopathy with protein aggregates and nemaline bodies 10. van der Werf C, Nederend I, Hofman N, et al. Familial related to CFL2 mutations. St. Malo: World Muscle evaluation in catecholaminergic polymorphic ventricular Society, 2017. tachycardia: disease penetrance and expression in car- 21. Schartner V, Romero NB, Donkervoort S, et al. Dihy- diac ryanodine receptor mutation-carrying relatives. Circ dropyridine receptor (DHPR, CACNA1S) congenital Arrhythm Electrophysiol 2012; 5: 748–756. myopathy. Acta Neuropathol 2017; 133: 517–533. 11. Bertocchini F, Ovitt CE, Conti A, et al. Requirement 22. Jeyakumar LH, Copello JA, O’Malley AM, et al. for the ryanodine receptor type 3 for efficient contrac- Purification and characterization of ryanodine receptor tion in neonatal skeletal muscles. EMBO J 1997; 16: 3 from mammalian tissue. J Biol Chem 1998; 273: 6956–6963. 16011–16020.

Content

1. Supplementary Methods

2. Supplementary Results

3. Supplementary References

1. Supplementary Methods

Ethical approval

The study was approved by the ethical standards of the relevant institutional review boards. The study was approved of the Ethics Review committee in the Gothenburg Region (Dn1: 842-14).

The muscle samples from control individuals were collected and studied with the approval of the Regional Ethical Review Board in Umeå, Sweden. The samples from human fetal muscle obtained at autopsy from stillbirths and human cortex and cauda equina obtained via the

Australian Brain Bank were collected and studied with the approval of the University of

Western Australia (UWA) Human Research Ethics Committee. Informed consent was obtained from the family after appropriate genetic counselling. Blood samples were obtained from the patient, her parents and sister.

Clinical evaluation

Medical history, physical examination and clinical follow-up were performed as part of routine clinical workup.

Histochemical analyses of muscle biopsy

For conventional histochemical techniques, 8-µm-thick cryostat sections of fresh-frozen

1 muscle tissue were stained with hematoxylin and eosin (HE), modified trichrome, Congo Red, periodic acid-Schiff-diastase (PAS), Oil Red O, reduced nicotinamide adenine dinucleotide dehydrogenase-tetrazolium reductase (NADH-TR), succinic dehydrogenase (SDH), cytochrome oxidase (COX), COX + SDH and adenosine triphosphatase (ATPase) preincubated at pH 9.4, 4.63, and 4.35, according to standard techniques.

Genetic analysis

DNA isolation

Extraction of genomic DNA was performed from whole blood from family members, using

DNeasy Blood & Tissue kit (Qiagen, Hilden Germany), according to the manufacturer’s instructions.

Next-generation sequencing

Neuromuscular sub-exomic sequencing (NSES) was initially performed on DNA from the patient using the Ion ProtonTM sequencer (Life Technologies), as previously described [1, 2].

The NSES panel comprised of 336 known and candidate neuromuscular disease genes, including those listed within the December 2012 freeze of Neuromuscular Disorders gene table

[3]. The NEM genes included: ACTA1, CFL2, KBTBD13, KLHL40, KLHL41, LMOD3, NEB,

TNNT1, TPM2, TPM3, MYO18B and MYPN. NSES data were analysed using the Cartegenia software package (Agilent Technologies). Only variants with a minor allele frequency of <2% in control population databases ExAC, 1000 Genome Project and dbSNP [4] were retained.

As no pathogenic variants were identified within the known neuromuscular disease genes, including the twelve known NEM genes, whole exome sequencing (WES) was performed on

DNA from the patient, as previously described [5]. Briefly, target enrichment was performed with 3 µg genomic DNA using the SureSelectXT Human All Exon Kit version 5 (Agilent

2 Technologies, Santa Clara, CA, USA) to generate barcoded whole-exome sequencing libraries.

Libraries were sequenced on the HiSeq2000 platform (Illumina, San Diego, CA, USA) as paired-end 2 ×100-bp reads with 60x coverage. Quality assessment of the sequence reads was performed by generating QC statistics with FastQC

(http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc). Read alignment to the reference human genome (hg19, UCSC assembly, February 2009) was done using BWA [6] with default parameters. After removal of PCR duplicates (Picard tools, http://picard.sourceforge.net) and file conversion (SAMtools) [7], quality score recalibration, indel realignment and variant calling were performed with the HaplotypeCaller algorithm in the GATK package [8], based on established best practices [9].

WES variant annotation, filtering and prioritisation

Variants were annotated with ANNOVAR [10] using a wide range of databases such as dbSNP build 135, dbNSFP, KEGG, the Gene Ontology project and tracks from the UCSC. A filtering strategy, directed to disease gene candidates, was performed by QIAGEN’s Ingenuity® Variant

Analysis™ software (www.qiagen.com/ingenuity) from QIAGEN Redwood City. Ingenuity

Variant Analysis combines analytical tools which annotates variants and displays data including Polyphen, SIFT and CADD scores. We focused initially on coding variants in genes known to be expressed in skeletal muscle, e.g. where the mutation produced a missense change, stop gain or stop loss, frameshift or essential splicing change. Only those changes that were predicted to be damaging or with unknown impact were analysed. We excluded variants that were frequent in control datasets (>1% in dbSNP, [4], the Exome Variant Server (EVS)

(NHLBI) (http://evs.gs.washington.edu/EVS/), the 1000 Genome Project Database

(http://browser.1000genomes.org/index.html), the Exome Aggregation Consortium (ExAC)

3 (http://exac.broadinstitute.org/) and the Human Background Variant Database

(http://neotek.scilifelab.se/hbvdb/).

Verification of the identified RYR3 variants by Polymerase chain reaction (PCR) and Sanger sequencing

Polymerase chain reaction (PCR) and bi-directional Sanger sequencing of RYR3 was performed on DNA samples from the patient and her unaffected parents and sister. PCR primers are available on request.

Array Comparative Genomic Hybridization (aCGH)

A custom 4x180k Comparative Genomic Hybridization array (array-CGH) (manuscript in preparation), was used to search for copy number variations (CNVs) that could be disease- causing [11]. The array design is based on the previously published NEM-CGH-array [12], which targets the published NM genes. In addition to the NM genes, the updated 4x180k array targets 176 additional genes related to neuromuscular diseases [11]. The labelling, hybridization, scanning and data analysis was performed as previously described [12, 13].

Quantitative polymerase chain reaction (qPCR) analysis of RYR1 and RYR3 transcripts of human tissue samples

Human adult control samples were obtained as surplus tissue from in vitro contracture testing for malignant hyperthermia (MH) susceptibility, following informed consent and were normal by MH testing. Human fetal muscle was obtained at autopsy from stillbirths, following informed consent for use of excess tissue for research. The healthy human cortex and cauda equina samples were obtained via the Australian Brain Bank.

4 RNA was extracted from 30 mg frozen tissue using TRIzol reagent (Thermo Fisher Scientific).

The SuperScript III First-Strand Synthesis System (Thermo Fisher Scientific) was used to synthesize cDNA from 1 µg total RNA using random hexamers according to the manufacturer’s protocol. The Rotor-Gene SYBR Green PCR Kit (Qiagen) was used to set up 10 µL reactions containing 1 µL diluted cDNA and 0.8 µM each of forward and reverse primers (RYR1, RYR3,

TBP, ACTB, EEF2; primer sequences are available on request). Thermal cycling was performed on the Rotor-Gene Q real-time PCR cycler and data analysis was performed with the associated software (Qiagen). Graphed data represent the mean ± SEM and were generated using

GraphPad Prism (V6.02).

Immunofluorescence and confocal microscopy analyses of muscle sections from controls

To evaluate the presence, and intracellular localization of RYR1 and RYR3, frozen muscle sections from human subjects were analysed by immunofluorescence. Nine muscle species (one from each muscle: biceps, psoas major, diaphragm, depressor labii inferior, depressor anguli oris, buccinator, zygomaticus major, temporalis and masseter) were collected at autopsy from one female and two male human donors. None of the donors were known to suffer from any neuromuscular disease.

Cryostat sections, 5- or 20-µm-thick, were processed for immunofluorescence analysis, as previously described,[14] using antibodies against RYR1 and RYR3. In brief, muscle sections were incubated overnight at 4ºC with mouse monoclonal anti-RYR1 (1:1000, Clone 43-C

MA3-925; Thermo Fischer Scientific. Waltham, MA, USA) and subsequently incubated for two hours at 37 °C with a rabbit polyclonal anti-RYR3 antibody (1:100, AB9082; EMD

Millipore. Darmstadt, Germany). RYR1 and RYR3 were visualized with a donkey anti-mouse

Rhobodamine RedX (1:500 dilution, Cat. No 715-295-151) and a donkey anti-rabbit Alexa 488

(1:400 dilution, Cat. No 711-545-152), respectively. Slides were mounted with Prolong Gold

5 (Molecular Probes®, Thermo Fisher Scientific Inc).

For 5-µm sections, the staining protocol was shortened and simplified in the following ways: post-fixation was reduced to 8 minutes; Triton X-100 incubation was omitted; normal serum treatment was done for 15 minutes and with 5 % sera in PBS; incubation with secondary antibodies and the RYR3 primary antibody was reduced to half and normal serum concentration was reduced to 5%. To validate staining specificity, sections were processed as above, except the primary antibody was replaced with either non-immune rabbit serum (1:100, Code X0902;

Dako. Glostrup, Denmark) or just PBS. Muscle sections were examined using a Nikon A1R-

LSM confocal microscope (Nikon Instruments Europe BV, Amsterdam, Netherlands) with 60× or 100× objectives. The stacks acquired were exported to Image J (version 2.0.0, Build

49b667f9aa, Open source software licensed under Creative Commons) for analysis.

Representative pictures of myofibres were generated with 3-slice, averaging Z-projections.

Assembly and layout of images was finalized using Adobe Photoshop CS6 v.13.0.6 (Adobe systems Inc. San Jose, CA, USA).

2. Supplementary Results

Pathological relevance of the RYR3 variants

The pathological relevance of the variants was supported by application of computational prediction tools. The CADD scores of p.Met2070Val and p.Arg2980Leu variants were 23.300 and 34.000, respectively. The PolyPhen-2 and SIFT function predictions of p.Met2070Val and p.Arg2980Leu scored probably damaging and damaging, respectively.

Transcript analysis of RYR1 and RYR3 in human skeletal muscle and brain tissues from controls

6 Transcript abundance of RYR1 and RYR3 were measured by qPCR in human fetal skeletal muscle (n = 2), adult skeletal muscle (n = 2), adult cortex (n = 3) and adult cauda equina (CE; n = 3) (Fig 1d). RYR1 were expressed at detectable levels in all tissue samples examined. As expected, expression of RYR1 was highest in adult skeletal muscle followed by fetal skeletal muscle. However, RYR1 transcript was also detectable in adult cortex and CE samples at levels comparable to RYR3 in both tissues (Fig 1d). RYR3 was expressed at comparable levels in all tissues. Thus, RYR1 and RYR3 are co-expressed at the transcript level in fetal skeletal muscle, adult skeletal muscle, adult cortex and CE.

Immunofluorescence and confocal microscopy analyses of skeletal muscle tissue from controls

Longitudinally sectioned myofibres were readily stained for both RYR1 and RYR3 isoforms.

While RYR1 staining was found exclusively in striated muscle tissues (Fig 2a.A and B) and showed no reactivity for other structures such as connective tissue, nerves, sarcolemma, or nuclei, RYR3 staining was, in addition, detectable in axons of nerve fascicles (Fig 2a.C) and in lamina media smooth muscle cells of blood vessels (Fig 2a.D). This is consistent with the previous reports that RYR3 is a ubiquitous Ca2+ receptor in nerve tissue and vasculature [15-

17].

WEB RESOURCES

The following Databases were used in this study:

The Exome Variant Server: NHLBI Exome Sequencing Project (ESP), Seattle, WA;

URL: http://evs.gs.washington.edu/EVS/

1000 Genome Project Database: http://browser.1000genomes.org/index.html

7 Human Background Variant DataBase: http://neotek.scilifelab.se/hbvdb/

Greater Middle East (GME) Variome web: http://igm.ucsd.edu/gme/index.php gnomAD: http://gnomad.broadinstitute.org/

CONSENT TO PUBLISH

The family in this study provided written informed consent to publish its family tree, and family data.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

ACKNOWLEDGEMENTS

We thank the family members who provided samples and clinical information for this study.

We would like to thank Per Stål from the department of Integrative Medical Biology, Umeå

University for providing muscle samples from the control donors. We would like to thank the

Bioinformatics Core Facility platform, at the Sahlgrenska Academy, University of Gothenburg for assistance with the bioinformatics analyses. We would like to thank Irene Martinez

Carrasco, Biochemical Imaging Centre Umeå, Umeå University, for excellent assistance with confocal microscopy.

FUNDING

The study was supported by grants from the Swedish Research Council (H.T., F.P.D.), the

Swedish Society of Medicine (H.T.) and European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no. 608473

(H.T.). The Västerbotten’s County Council (F.P.D). M.O. is supported by a grant from Fondo de Investigaciones Sanitarias (FIS), Instituto de Salud Carlos III (ISCIII PI14/00738) and

8 FEDER funds “a way to achieve Europe’’. Australian National Health and Medical Research

Council Fellowships APP1002147, APP1035955, European Union Collaborative Grant

APP1055295 and Project Grant APP1080587 (N.L.)

9 3. Supplementary References

1. Todd EJ, Yau KS, Ong R, et al. Next generation sequencing in a large cohort of patients presenting with neuromuscular disease before or at birth. Orphanet J Rare Dis. 2015; 10: 148. 2. Cabrera-Serrano M, Ghaoui R, Ravenscroft G, et al. Expanding the phenotype of GMPPB mutations. Brain : a journal of neurology. 2015; 138: 836-44. 3. Kaplan JC. The 2012 version of the gene table of monogenic neuromuscular disorders. Neuromuscular disorders : NMD. 2011; 21: 833-61. 4. Sherry ST, Ward MH, Kholodov M, et al. dbSNP: the NCBI database of genetic variation. Nucleic acids research. 2001; 29: 308-11. 5. Kariminejad A, Ghaderi-Sohi S, Hossein-Nejad Nedai H, et al. Lethal multiple pterygium syndrome, the extreme end of the RYR1 spectrum. BMC musculoskeletal disorders. 2016; 17: 109. 6. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009; 25: 1754-60. 7. Li H, Handsaker B, Wysoker A, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009; 25: 2078-9. 8. McKenna A, Hanna M, Banks E, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010; 20: 1297-303. 9. DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature genetics. 2011; 43: 491-8. 10. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010; 38: e164. 11. Sagath L, V.-L. Lehtokari, C. Wallgren-Pettersson, K. Pelin and K. Kiiski. A novel copy number variation detection array for the diagnostics of neuromuscular disorders. World Muscle Society; St. Malo, France: Neuromuscul Disord; 2017. p. S182. 12. Kiiski K, Laari L, Lehtokari VL, et al. Targeted array comparative genomic hybridization--a new diagnostic tool for the detection of large copy number variations in nemaline myopathy-causing genes. Neuromuscul Disord. 2013; 23: 56-65. 13. Kiiski K, Lehtokari VL, Loytynoja A, et al. A recurrent copy number variation of the NEB triplicate region: only revealed by the targeted nemaline myopathy CGH array. Eur J Hum Genet. 2016; 24: 574-80. 14. Lindstrom M, Tjust AE, Pedrosa Domellof F. Pax7-Positive Cells/Satellite Cells in Human Extraocular Muscles. Investigative ophthalmology & visual science. 2015; 56: 6132- 43. 15. Takenaka T, Ohno Y, Hayashi K, Saruta T, Suzuki H. Governance of arteriolar oscillation by ryanodine receptors. American journal of physiology Regulatory, integrative and comparative physiology. 2003; 285: R125-31. 16. Eckert RE, Karsten AJ, Utz J, Ziegler M. Regulation of renal artery smooth muscle tone by alpha1-adrenoceptors: role of voltage-gated calcium channels and intracellular calcium stores. Urological research. 2000; 28: 122-7. 17. Hakamata Y, Nakai J, Takeshima H, Imoto K. Primary structure and distribution of a novel ryanodine receptor/calcium release channel from rabbit brain. FEBS letters. 1992; 312: 229-35.